[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
/*
|
2011-06-06 07:16:30 +00:00
|
|
|
* SPI init/core code
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
*
|
|
|
|
* Copyright (C) 2005 David Brownell
|
2012-04-07 20:16:53 +00:00
|
|
|
* Copyright (C) 2008 Secret Lab Technologies Ltd.
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
*
|
|
|
|
* This program is free software; you can redistribute it and/or modify
|
|
|
|
* it under the terms of the GNU General Public License as published by
|
|
|
|
* the Free Software Foundation; either version 2 of the License, or
|
|
|
|
* (at your option) any later version.
|
|
|
|
*
|
|
|
|
* This program is distributed in the hope that it will be useful,
|
|
|
|
* but WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
|
|
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
|
|
|
* GNU General Public License for more details.
|
|
|
|
*/
|
|
|
|
|
|
|
|
#include <linux/kernel.h>
|
|
|
|
#include <linux/device.h>
|
|
|
|
#include <linux/init.h>
|
|
|
|
#include <linux/cache.h>
|
2014-01-16 12:22:43 +00:00
|
|
|
#include <linux/dma-mapping.h>
|
|
|
|
#include <linux/dmaengine.h>
|
2007-07-17 11:04:16 +00:00
|
|
|
#include <linux/mutex.h>
|
2010-10-03 03:28:29 +00:00
|
|
|
#include <linux/of_device.h>
|
2012-04-07 20:16:53 +00:00
|
|
|
#include <linux/of_irq.h>
|
2014-06-18 15:29:32 +00:00
|
|
|
#include <linux/clk/clk-conf.h>
|
include cleanup: Update gfp.h and slab.h includes to prepare for breaking implicit slab.h inclusion from percpu.h
percpu.h is included by sched.h and module.h and thus ends up being
included when building most .c files. percpu.h includes slab.h which
in turn includes gfp.h making everything defined by the two files
universally available and complicating inclusion dependencies.
percpu.h -> slab.h dependency is about to be removed. Prepare for
this change by updating users of gfp and slab facilities include those
headers directly instead of assuming availability. As this conversion
needs to touch large number of source files, the following script is
used as the basis of conversion.
http://userweb.kernel.org/~tj/misc/slabh-sweep.py
The script does the followings.
* Scan files for gfp and slab usages and update includes such that
only the necessary includes are there. ie. if only gfp is used,
gfp.h, if slab is used, slab.h.
* When the script inserts a new include, it looks at the include
blocks and try to put the new include such that its order conforms
to its surrounding. It's put in the include block which contains
core kernel includes, in the same order that the rest are ordered -
alphabetical, Christmas tree, rev-Xmas-tree or at the end if there
doesn't seem to be any matching order.
* If the script can't find a place to put a new include (mostly
because the file doesn't have fitting include block), it prints out
an error message indicating which .h file needs to be added to the
file.
The conversion was done in the following steps.
1. The initial automatic conversion of all .c files updated slightly
over 4000 files, deleting around 700 includes and adding ~480 gfp.h
and ~3000 slab.h inclusions. The script emitted errors for ~400
files.
2. Each error was manually checked. Some didn't need the inclusion,
some needed manual addition while adding it to implementation .h or
embedding .c file was more appropriate for others. This step added
inclusions to around 150 files.
3. The script was run again and the output was compared to the edits
from #2 to make sure no file was left behind.
4. Several build tests were done and a couple of problems were fixed.
e.g. lib/decompress_*.c used malloc/free() wrappers around slab
APIs requiring slab.h to be added manually.
5. The script was run on all .h files but without automatically
editing them as sprinkling gfp.h and slab.h inclusions around .h
files could easily lead to inclusion dependency hell. Most gfp.h
inclusion directives were ignored as stuff from gfp.h was usually
wildly available and often used in preprocessor macros. Each
slab.h inclusion directive was examined and added manually as
necessary.
6. percpu.h was updated not to include slab.h.
7. Build test were done on the following configurations and failures
were fixed. CONFIG_GCOV_KERNEL was turned off for all tests (as my
distributed build env didn't work with gcov compiles) and a few
more options had to be turned off depending on archs to make things
build (like ipr on powerpc/64 which failed due to missing writeq).
* x86 and x86_64 UP and SMP allmodconfig and a custom test config.
* powerpc and powerpc64 SMP allmodconfig
* sparc and sparc64 SMP allmodconfig
* ia64 SMP allmodconfig
* s390 SMP allmodconfig
* alpha SMP allmodconfig
* um on x86_64 SMP allmodconfig
8. percpu.h modifications were reverted so that it could be applied as
a separate patch and serve as bisection point.
Given the fact that I had only a couple of failures from tests on step
6, I'm fairly confident about the coverage of this conversion patch.
If there is a breakage, it's likely to be something in one of the arch
headers which should be easily discoverable easily on most builds of
the specific arch.
Signed-off-by: Tejun Heo <tj@kernel.org>
Guess-its-ok-by: Christoph Lameter <cl@linux-foundation.org>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Lee Schermerhorn <Lee.Schermerhorn@hp.com>
2010-03-24 08:04:11 +00:00
|
|
|
#include <linux/slab.h>
|
2009-09-22 23:46:08 +00:00
|
|
|
#include <linux/mod_devicetable.h>
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
#include <linux/spi/spi.h>
|
2012-11-15 19:19:57 +00:00
|
|
|
#include <linux/of_gpio.h>
|
2010-12-25 14:32:27 +00:00
|
|
|
#include <linux/pm_runtime.h>
|
2014-09-29 11:58:47 +00:00
|
|
|
#include <linux/pm_domain.h>
|
2011-07-10 16:57:55 +00:00
|
|
|
#include <linux/export.h>
|
2013-02-07 15:47:07 +00:00
|
|
|
#include <linux/sched/rt.h>
|
2012-02-22 09:05:38 +00:00
|
|
|
#include <linux/delay.h>
|
|
|
|
#include <linux/kthread.h>
|
2012-11-30 11:37:53 +00:00
|
|
|
#include <linux/ioport.h>
|
|
|
|
#include <linux/acpi.h>
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
|
2013-10-07 18:33:53 +00:00
|
|
|
#define CREATE_TRACE_POINTS
|
|
|
|
#include <trace/events/spi.h>
|
|
|
|
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
static void spidev_release(struct device *dev)
|
|
|
|
{
|
2007-02-12 08:52:45 +00:00
|
|
|
struct spi_device *spi = to_spi_device(dev);
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
|
|
|
|
/* spi masters may cleanup for released devices */
|
|
|
|
if (spi->master->cleanup)
|
|
|
|
spi->master->cleanup(spi);
|
|
|
|
|
2006-01-08 21:34:25 +00:00
|
|
|
spi_master_put(spi->master);
|
2010-04-12 09:56:35 +00:00
|
|
|
kfree(spi);
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
static ssize_t
|
|
|
|
modalias_show(struct device *dev, struct device_attribute *a, char *buf)
|
|
|
|
{
|
|
|
|
const struct spi_device *spi = to_spi_device(dev);
|
ACPI: fix module autoloading for ACPI enumerated devices
ACPI enumerated devices has ACPI style _HID and _CID strings,
all of these strings can be used for both driver loading and matching.
Currently, in Platform, I2C and SPI bus, the ACPI style driver matching
is supported by invoking acpi_driver_match_device() in bus .match() callback.
But, the module autoloading is still broken.
For example, there is any ACPI device with _HID "INTABCD" that is
enumerated to platform bus, and we have a driver that can probe it.
The driver exports its module_alias as "acpi:INTABCD" use the following code
static const struct acpi_device_id xxx_acpi_match[] = {
{ "INTABCD", 0 },
{ }
};
MODULE_DEVICE_TABLE(acpi, xxx_acpi_match);
But, unfortunately, the device' modalias is shown as "platform:INTABCD:00",
please refer to modalias_show() and platform_uevent() in
drivers/base/platform.c.
This results in that the driver will not be loaded automatically when the
device node is created, because their modalias do not match.
This also applies to I2C and SPI bus.
With this patch, the device' modalias will be shown as "acpi:INTABCD" as well.
Signed-off-by: Zhang Rui <rui.zhang@intel.com>
Acked-by: Mark Brown <broonie@linaro.org>
Acked-by: Wolfram Sang <wsa@the-dreams.de>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
2014-01-14 08:46:37 +00:00
|
|
|
int len;
|
|
|
|
|
|
|
|
len = acpi_device_modalias(dev, buf, PAGE_SIZE - 1);
|
|
|
|
if (len != -ENODEV)
|
|
|
|
return len;
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
|
2012-05-20 06:08:13 +00:00
|
|
|
return sprintf(buf, "%s%s\n", SPI_MODULE_PREFIX, spi->modalias);
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
}
|
2013-10-08 01:27:38 +00:00
|
|
|
static DEVICE_ATTR_RO(modalias);
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
|
2015-06-22 13:00:36 +00:00
|
|
|
#define SPI_STATISTICS_ATTRS(field, file) \
|
|
|
|
static ssize_t spi_master_##field##_show(struct device *dev, \
|
|
|
|
struct device_attribute *attr, \
|
|
|
|
char *buf) \
|
|
|
|
{ \
|
|
|
|
struct spi_master *master = container_of(dev, \
|
|
|
|
struct spi_master, dev); \
|
|
|
|
return spi_statistics_##field##_show(&master->statistics, buf); \
|
|
|
|
} \
|
|
|
|
static struct device_attribute dev_attr_spi_master_##field = { \
|
|
|
|
.attr = { .name = file, .mode = S_IRUGO }, \
|
|
|
|
.show = spi_master_##field##_show, \
|
|
|
|
}; \
|
|
|
|
static ssize_t spi_device_##field##_show(struct device *dev, \
|
|
|
|
struct device_attribute *attr, \
|
|
|
|
char *buf) \
|
|
|
|
{ \
|
|
|
|
struct spi_device *spi = container_of(dev, \
|
|
|
|
struct spi_device, dev); \
|
|
|
|
return spi_statistics_##field##_show(&spi->statistics, buf); \
|
|
|
|
} \
|
|
|
|
static struct device_attribute dev_attr_spi_device_##field = { \
|
|
|
|
.attr = { .name = file, .mode = S_IRUGO }, \
|
|
|
|
.show = spi_device_##field##_show, \
|
|
|
|
}
|
|
|
|
|
|
|
|
#define SPI_STATISTICS_SHOW_NAME(name, file, field, format_string) \
|
|
|
|
static ssize_t spi_statistics_##name##_show(struct spi_statistics *stat, \
|
|
|
|
char *buf) \
|
|
|
|
{ \
|
|
|
|
unsigned long flags; \
|
|
|
|
ssize_t len; \
|
|
|
|
spin_lock_irqsave(&stat->lock, flags); \
|
|
|
|
len = sprintf(buf, format_string, stat->field); \
|
|
|
|
spin_unlock_irqrestore(&stat->lock, flags); \
|
|
|
|
return len; \
|
|
|
|
} \
|
|
|
|
SPI_STATISTICS_ATTRS(name, file)
|
|
|
|
|
|
|
|
#define SPI_STATISTICS_SHOW(field, format_string) \
|
|
|
|
SPI_STATISTICS_SHOW_NAME(field, __stringify(field), \
|
|
|
|
field, format_string)
|
|
|
|
|
|
|
|
SPI_STATISTICS_SHOW(messages, "%lu");
|
|
|
|
SPI_STATISTICS_SHOW(transfers, "%lu");
|
|
|
|
SPI_STATISTICS_SHOW(errors, "%lu");
|
|
|
|
SPI_STATISTICS_SHOW(timedout, "%lu");
|
|
|
|
|
|
|
|
SPI_STATISTICS_SHOW(spi_sync, "%lu");
|
|
|
|
SPI_STATISTICS_SHOW(spi_sync_immediate, "%lu");
|
|
|
|
SPI_STATISTICS_SHOW(spi_async, "%lu");
|
|
|
|
|
|
|
|
SPI_STATISTICS_SHOW(bytes, "%llu");
|
|
|
|
SPI_STATISTICS_SHOW(bytes_rx, "%llu");
|
|
|
|
SPI_STATISTICS_SHOW(bytes_tx, "%llu");
|
|
|
|
|
2015-06-22 13:02:04 +00:00
|
|
|
#define SPI_STATISTICS_TRANSFER_BYTES_HISTO(index, number) \
|
|
|
|
SPI_STATISTICS_SHOW_NAME(transfer_bytes_histo##index, \
|
|
|
|
"transfer_bytes_histo_" number, \
|
|
|
|
transfer_bytes_histo[index], "%lu")
|
|
|
|
SPI_STATISTICS_TRANSFER_BYTES_HISTO(0, "0-1");
|
|
|
|
SPI_STATISTICS_TRANSFER_BYTES_HISTO(1, "2-3");
|
|
|
|
SPI_STATISTICS_TRANSFER_BYTES_HISTO(2, "4-7");
|
|
|
|
SPI_STATISTICS_TRANSFER_BYTES_HISTO(3, "8-15");
|
|
|
|
SPI_STATISTICS_TRANSFER_BYTES_HISTO(4, "16-31");
|
|
|
|
SPI_STATISTICS_TRANSFER_BYTES_HISTO(5, "32-63");
|
|
|
|
SPI_STATISTICS_TRANSFER_BYTES_HISTO(6, "64-127");
|
|
|
|
SPI_STATISTICS_TRANSFER_BYTES_HISTO(7, "128-255");
|
|
|
|
SPI_STATISTICS_TRANSFER_BYTES_HISTO(8, "256-511");
|
|
|
|
SPI_STATISTICS_TRANSFER_BYTES_HISTO(9, "512-1023");
|
|
|
|
SPI_STATISTICS_TRANSFER_BYTES_HISTO(10, "1024-2047");
|
|
|
|
SPI_STATISTICS_TRANSFER_BYTES_HISTO(11, "2048-4095");
|
|
|
|
SPI_STATISTICS_TRANSFER_BYTES_HISTO(12, "4096-8191");
|
|
|
|
SPI_STATISTICS_TRANSFER_BYTES_HISTO(13, "8192-16383");
|
|
|
|
SPI_STATISTICS_TRANSFER_BYTES_HISTO(14, "16384-32767");
|
|
|
|
SPI_STATISTICS_TRANSFER_BYTES_HISTO(15, "32768-65535");
|
|
|
|
SPI_STATISTICS_TRANSFER_BYTES_HISTO(16, "65536+");
|
|
|
|
|
2013-10-08 01:27:38 +00:00
|
|
|
static struct attribute *spi_dev_attrs[] = {
|
|
|
|
&dev_attr_modalias.attr,
|
|
|
|
NULL,
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
};
|
2015-06-22 13:00:36 +00:00
|
|
|
|
|
|
|
static const struct attribute_group spi_dev_group = {
|
|
|
|
.attrs = spi_dev_attrs,
|
|
|
|
};
|
|
|
|
|
|
|
|
static struct attribute *spi_device_statistics_attrs[] = {
|
|
|
|
&dev_attr_spi_device_messages.attr,
|
|
|
|
&dev_attr_spi_device_transfers.attr,
|
|
|
|
&dev_attr_spi_device_errors.attr,
|
|
|
|
&dev_attr_spi_device_timedout.attr,
|
|
|
|
&dev_attr_spi_device_spi_sync.attr,
|
|
|
|
&dev_attr_spi_device_spi_sync_immediate.attr,
|
|
|
|
&dev_attr_spi_device_spi_async.attr,
|
|
|
|
&dev_attr_spi_device_bytes.attr,
|
|
|
|
&dev_attr_spi_device_bytes_rx.attr,
|
|
|
|
&dev_attr_spi_device_bytes_tx.attr,
|
2015-06-22 13:02:04 +00:00
|
|
|
&dev_attr_spi_device_transfer_bytes_histo0.attr,
|
|
|
|
&dev_attr_spi_device_transfer_bytes_histo1.attr,
|
|
|
|
&dev_attr_spi_device_transfer_bytes_histo2.attr,
|
|
|
|
&dev_attr_spi_device_transfer_bytes_histo3.attr,
|
|
|
|
&dev_attr_spi_device_transfer_bytes_histo4.attr,
|
|
|
|
&dev_attr_spi_device_transfer_bytes_histo5.attr,
|
|
|
|
&dev_attr_spi_device_transfer_bytes_histo6.attr,
|
|
|
|
&dev_attr_spi_device_transfer_bytes_histo7.attr,
|
|
|
|
&dev_attr_spi_device_transfer_bytes_histo8.attr,
|
|
|
|
&dev_attr_spi_device_transfer_bytes_histo9.attr,
|
|
|
|
&dev_attr_spi_device_transfer_bytes_histo10.attr,
|
|
|
|
&dev_attr_spi_device_transfer_bytes_histo11.attr,
|
|
|
|
&dev_attr_spi_device_transfer_bytes_histo12.attr,
|
|
|
|
&dev_attr_spi_device_transfer_bytes_histo13.attr,
|
|
|
|
&dev_attr_spi_device_transfer_bytes_histo14.attr,
|
|
|
|
&dev_attr_spi_device_transfer_bytes_histo15.attr,
|
|
|
|
&dev_attr_spi_device_transfer_bytes_histo16.attr,
|
2015-06-22 13:00:36 +00:00
|
|
|
NULL,
|
|
|
|
};
|
|
|
|
|
|
|
|
static const struct attribute_group spi_device_statistics_group = {
|
|
|
|
.name = "statistics",
|
|
|
|
.attrs = spi_device_statistics_attrs,
|
|
|
|
};
|
|
|
|
|
|
|
|
static const struct attribute_group *spi_dev_groups[] = {
|
|
|
|
&spi_dev_group,
|
|
|
|
&spi_device_statistics_group,
|
|
|
|
NULL,
|
|
|
|
};
|
|
|
|
|
|
|
|
static struct attribute *spi_master_statistics_attrs[] = {
|
|
|
|
&dev_attr_spi_master_messages.attr,
|
|
|
|
&dev_attr_spi_master_transfers.attr,
|
|
|
|
&dev_attr_spi_master_errors.attr,
|
|
|
|
&dev_attr_spi_master_timedout.attr,
|
|
|
|
&dev_attr_spi_master_spi_sync.attr,
|
|
|
|
&dev_attr_spi_master_spi_sync_immediate.attr,
|
|
|
|
&dev_attr_spi_master_spi_async.attr,
|
|
|
|
&dev_attr_spi_master_bytes.attr,
|
|
|
|
&dev_attr_spi_master_bytes_rx.attr,
|
|
|
|
&dev_attr_spi_master_bytes_tx.attr,
|
2015-06-22 13:02:04 +00:00
|
|
|
&dev_attr_spi_master_transfer_bytes_histo0.attr,
|
|
|
|
&dev_attr_spi_master_transfer_bytes_histo1.attr,
|
|
|
|
&dev_attr_spi_master_transfer_bytes_histo2.attr,
|
|
|
|
&dev_attr_spi_master_transfer_bytes_histo3.attr,
|
|
|
|
&dev_attr_spi_master_transfer_bytes_histo4.attr,
|
|
|
|
&dev_attr_spi_master_transfer_bytes_histo5.attr,
|
|
|
|
&dev_attr_spi_master_transfer_bytes_histo6.attr,
|
|
|
|
&dev_attr_spi_master_transfer_bytes_histo7.attr,
|
|
|
|
&dev_attr_spi_master_transfer_bytes_histo8.attr,
|
|
|
|
&dev_attr_spi_master_transfer_bytes_histo9.attr,
|
|
|
|
&dev_attr_spi_master_transfer_bytes_histo10.attr,
|
|
|
|
&dev_attr_spi_master_transfer_bytes_histo11.attr,
|
|
|
|
&dev_attr_spi_master_transfer_bytes_histo12.attr,
|
|
|
|
&dev_attr_spi_master_transfer_bytes_histo13.attr,
|
|
|
|
&dev_attr_spi_master_transfer_bytes_histo14.attr,
|
|
|
|
&dev_attr_spi_master_transfer_bytes_histo15.attr,
|
|
|
|
&dev_attr_spi_master_transfer_bytes_histo16.attr,
|
2015-06-22 13:00:36 +00:00
|
|
|
NULL,
|
|
|
|
};
|
|
|
|
|
|
|
|
static const struct attribute_group spi_master_statistics_group = {
|
|
|
|
.name = "statistics",
|
|
|
|
.attrs = spi_master_statistics_attrs,
|
|
|
|
};
|
|
|
|
|
|
|
|
static const struct attribute_group *spi_master_groups[] = {
|
|
|
|
&spi_master_statistics_group,
|
|
|
|
NULL,
|
|
|
|
};
|
|
|
|
|
|
|
|
void spi_statistics_add_transfer_stats(struct spi_statistics *stats,
|
|
|
|
struct spi_transfer *xfer,
|
|
|
|
struct spi_master *master)
|
|
|
|
{
|
|
|
|
unsigned long flags;
|
2015-06-22 13:02:04 +00:00
|
|
|
int l2len = min(fls(xfer->len), SPI_STATISTICS_HISTO_SIZE) - 1;
|
|
|
|
|
|
|
|
if (l2len < 0)
|
|
|
|
l2len = 0;
|
2015-06-22 13:00:36 +00:00
|
|
|
|
|
|
|
spin_lock_irqsave(&stats->lock, flags);
|
|
|
|
|
|
|
|
stats->transfers++;
|
2015-06-22 13:02:04 +00:00
|
|
|
stats->transfer_bytes_histo[l2len]++;
|
2015-06-22 13:00:36 +00:00
|
|
|
|
|
|
|
stats->bytes += xfer->len;
|
|
|
|
if ((xfer->tx_buf) &&
|
|
|
|
(xfer->tx_buf != master->dummy_tx))
|
|
|
|
stats->bytes_tx += xfer->len;
|
|
|
|
if ((xfer->rx_buf) &&
|
|
|
|
(xfer->rx_buf != master->dummy_rx))
|
|
|
|
stats->bytes_rx += xfer->len;
|
|
|
|
|
|
|
|
spin_unlock_irqrestore(&stats->lock, flags);
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(spi_statistics_add_transfer_stats);
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
|
|
|
|
/* modalias support makes "modprobe $MODALIAS" new-style hotplug work,
|
|
|
|
* and the sysfs version makes coldplug work too.
|
|
|
|
*/
|
|
|
|
|
2009-09-22 23:46:04 +00:00
|
|
|
static const struct spi_device_id *spi_match_id(const struct spi_device_id *id,
|
|
|
|
const struct spi_device *sdev)
|
|
|
|
{
|
|
|
|
while (id->name[0]) {
|
|
|
|
if (!strcmp(sdev->modalias, id->name))
|
|
|
|
return id;
|
|
|
|
id++;
|
|
|
|
}
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
|
|
|
|
const struct spi_device_id *spi_get_device_id(const struct spi_device *sdev)
|
|
|
|
{
|
|
|
|
const struct spi_driver *sdrv = to_spi_driver(sdev->dev.driver);
|
|
|
|
|
|
|
|
return spi_match_id(sdrv->id_table, sdev);
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(spi_get_device_id);
|
|
|
|
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
static int spi_match_device(struct device *dev, struct device_driver *drv)
|
|
|
|
{
|
|
|
|
const struct spi_device *spi = to_spi_device(dev);
|
2009-09-22 23:46:04 +00:00
|
|
|
const struct spi_driver *sdrv = to_spi_driver(drv);
|
|
|
|
|
2010-10-03 03:28:29 +00:00
|
|
|
/* Attempt an OF style match */
|
|
|
|
if (of_driver_match_device(dev, drv))
|
|
|
|
return 1;
|
|
|
|
|
2012-11-30 11:37:53 +00:00
|
|
|
/* Then try ACPI */
|
|
|
|
if (acpi_driver_match_device(dev, drv))
|
|
|
|
return 1;
|
|
|
|
|
2009-09-22 23:46:04 +00:00
|
|
|
if (sdrv->id_table)
|
|
|
|
return !!spi_match_id(sdrv->id_table, spi);
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
|
2009-01-06 18:44:37 +00:00
|
|
|
return strcmp(spi->modalias, drv->name) == 0;
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
}
|
|
|
|
|
2007-08-14 13:15:12 +00:00
|
|
|
static int spi_uevent(struct device *dev, struct kobj_uevent_env *env)
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
{
|
|
|
|
const struct spi_device *spi = to_spi_device(dev);
|
ACPI: fix module autoloading for ACPI enumerated devices
ACPI enumerated devices has ACPI style _HID and _CID strings,
all of these strings can be used for both driver loading and matching.
Currently, in Platform, I2C and SPI bus, the ACPI style driver matching
is supported by invoking acpi_driver_match_device() in bus .match() callback.
But, the module autoloading is still broken.
For example, there is any ACPI device with _HID "INTABCD" that is
enumerated to platform bus, and we have a driver that can probe it.
The driver exports its module_alias as "acpi:INTABCD" use the following code
static const struct acpi_device_id xxx_acpi_match[] = {
{ "INTABCD", 0 },
{ }
};
MODULE_DEVICE_TABLE(acpi, xxx_acpi_match);
But, unfortunately, the device' modalias is shown as "platform:INTABCD:00",
please refer to modalias_show() and platform_uevent() in
drivers/base/platform.c.
This results in that the driver will not be loaded automatically when the
device node is created, because their modalias do not match.
This also applies to I2C and SPI bus.
With this patch, the device' modalias will be shown as "acpi:INTABCD" as well.
Signed-off-by: Zhang Rui <rui.zhang@intel.com>
Acked-by: Mark Brown <broonie@linaro.org>
Acked-by: Wolfram Sang <wsa@the-dreams.de>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
2014-01-14 08:46:37 +00:00
|
|
|
int rc;
|
|
|
|
|
|
|
|
rc = acpi_device_uevent_modalias(dev, env);
|
|
|
|
if (rc != -ENODEV)
|
|
|
|
return rc;
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
|
2009-09-22 23:46:08 +00:00
|
|
|
add_uevent_var(env, "MODALIAS=%s%s", SPI_MODULE_PREFIX, spi->modalias);
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
struct bus_type spi_bus_type = {
|
|
|
|
.name = "spi",
|
2013-10-08 01:27:38 +00:00
|
|
|
.dev_groups = spi_dev_groups,
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
.match = spi_match_device,
|
|
|
|
.uevent = spi_uevent,
|
|
|
|
};
|
|
|
|
EXPORT_SYMBOL_GPL(spi_bus_type);
|
|
|
|
|
2006-01-08 21:34:23 +00:00
|
|
|
|
|
|
|
static int spi_drv_probe(struct device *dev)
|
|
|
|
{
|
|
|
|
const struct spi_driver *sdrv = to_spi_driver(dev->driver);
|
2015-10-09 14:45:55 +00:00
|
|
|
struct spi_device *spi = to_spi_device(dev);
|
2013-10-10 10:28:48 +00:00
|
|
|
int ret;
|
|
|
|
|
2014-06-18 15:29:32 +00:00
|
|
|
ret = of_clk_set_defaults(dev->of_node, false);
|
|
|
|
if (ret)
|
|
|
|
return ret;
|
|
|
|
|
2015-10-09 14:45:55 +00:00
|
|
|
if (dev->of_node) {
|
|
|
|
spi->irq = of_irq_get(dev->of_node, 0);
|
|
|
|
if (spi->irq == -EPROBE_DEFER)
|
|
|
|
return -EPROBE_DEFER;
|
|
|
|
if (spi->irq < 0)
|
|
|
|
spi->irq = 0;
|
|
|
|
}
|
|
|
|
|
2014-09-19 18:27:41 +00:00
|
|
|
ret = dev_pm_domain_attach(dev, true);
|
|
|
|
if (ret != -EPROBE_DEFER) {
|
2015-10-09 14:45:55 +00:00
|
|
|
ret = sdrv->probe(spi);
|
2014-09-19 18:27:41 +00:00
|
|
|
if (ret)
|
|
|
|
dev_pm_domain_detach(dev, true);
|
|
|
|
}
|
2006-01-08 21:34:23 +00:00
|
|
|
|
2013-10-10 10:28:48 +00:00
|
|
|
return ret;
|
2006-01-08 21:34:23 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
static int spi_drv_remove(struct device *dev)
|
|
|
|
{
|
|
|
|
const struct spi_driver *sdrv = to_spi_driver(dev->driver);
|
2013-10-10 10:28:48 +00:00
|
|
|
int ret;
|
|
|
|
|
2014-02-13 14:28:41 +00:00
|
|
|
ret = sdrv->remove(to_spi_device(dev));
|
2014-09-19 18:27:41 +00:00
|
|
|
dev_pm_domain_detach(dev, true);
|
2006-01-08 21:34:23 +00:00
|
|
|
|
2013-10-10 10:28:48 +00:00
|
|
|
return ret;
|
2006-01-08 21:34:23 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
static void spi_drv_shutdown(struct device *dev)
|
|
|
|
{
|
|
|
|
const struct spi_driver *sdrv = to_spi_driver(dev->driver);
|
|
|
|
|
|
|
|
sdrv->shutdown(to_spi_device(dev));
|
|
|
|
}
|
|
|
|
|
2007-05-08 07:32:21 +00:00
|
|
|
/**
|
2015-10-23 13:59:10 +00:00
|
|
|
* __spi_register_driver - register a SPI driver
|
2015-11-10 12:03:04 +00:00
|
|
|
* @owner: owner module of the driver to register
|
2007-05-08 07:32:21 +00:00
|
|
|
* @sdrv: the driver to register
|
|
|
|
* Context: can sleep
|
2015-10-22 16:59:23 +00:00
|
|
|
*
|
|
|
|
* Return: zero on success, else a negative error code.
|
2007-05-08 07:32:21 +00:00
|
|
|
*/
|
2015-10-23 13:59:10 +00:00
|
|
|
int __spi_register_driver(struct module *owner, struct spi_driver *sdrv)
|
2006-01-08 21:34:23 +00:00
|
|
|
{
|
2015-10-23 13:59:10 +00:00
|
|
|
sdrv->driver.owner = owner;
|
2006-01-08 21:34:23 +00:00
|
|
|
sdrv->driver.bus = &spi_bus_type;
|
|
|
|
if (sdrv->probe)
|
|
|
|
sdrv->driver.probe = spi_drv_probe;
|
|
|
|
if (sdrv->remove)
|
|
|
|
sdrv->driver.remove = spi_drv_remove;
|
|
|
|
if (sdrv->shutdown)
|
|
|
|
sdrv->driver.shutdown = spi_drv_shutdown;
|
|
|
|
return driver_register(&sdrv->driver);
|
|
|
|
}
|
2015-10-23 13:59:10 +00:00
|
|
|
EXPORT_SYMBOL_GPL(__spi_register_driver);
|
2006-01-08 21:34:23 +00:00
|
|
|
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
/*-------------------------------------------------------------------------*/
|
|
|
|
|
|
|
|
/* SPI devices should normally not be created by SPI device drivers; that
|
|
|
|
* would make them board-specific. Similarly with SPI master drivers.
|
|
|
|
* Device registration normally goes into like arch/.../mach.../board-YYY.c
|
|
|
|
* with other readonly (flashable) information about mainboard devices.
|
|
|
|
*/
|
|
|
|
|
|
|
|
struct boardinfo {
|
|
|
|
struct list_head list;
|
2010-08-02 07:52:15 +00:00
|
|
|
struct spi_board_info board_info;
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
};
|
|
|
|
|
|
|
|
static LIST_HEAD(board_list);
|
2010-08-02 07:52:15 +00:00
|
|
|
static LIST_HEAD(spi_master_list);
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Used to protect add/del opertion for board_info list and
|
|
|
|
* spi_master list, and their matching process
|
|
|
|
*/
|
2007-07-17 11:04:16 +00:00
|
|
|
static DEFINE_MUTEX(board_lock);
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
|
2008-05-15 22:50:22 +00:00
|
|
|
/**
|
|
|
|
* spi_alloc_device - Allocate a new SPI device
|
|
|
|
* @master: Controller to which device is connected
|
|
|
|
* Context: can sleep
|
|
|
|
*
|
|
|
|
* Allows a driver to allocate and initialize a spi_device without
|
|
|
|
* registering it immediately. This allows a driver to directly
|
|
|
|
* fill the spi_device with device parameters before calling
|
|
|
|
* spi_add_device() on it.
|
|
|
|
*
|
|
|
|
* Caller is responsible to call spi_add_device() on the returned
|
|
|
|
* spi_device structure to add it to the SPI master. If the caller
|
|
|
|
* needs to discard the spi_device without adding it, then it should
|
|
|
|
* call spi_dev_put() on it.
|
|
|
|
*
|
2015-10-22 16:59:23 +00:00
|
|
|
* Return: a pointer to the new device, or NULL.
|
2008-05-15 22:50:22 +00:00
|
|
|
*/
|
|
|
|
struct spi_device *spi_alloc_device(struct spi_master *master)
|
|
|
|
{
|
|
|
|
struct spi_device *spi;
|
|
|
|
|
|
|
|
if (!spi_master_get(master))
|
|
|
|
return NULL;
|
|
|
|
|
2013-10-14 01:31:51 +00:00
|
|
|
spi = kzalloc(sizeof(*spi), GFP_KERNEL);
|
2008-05-15 22:50:22 +00:00
|
|
|
if (!spi) {
|
|
|
|
spi_master_put(master);
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
|
|
|
|
spi->master = master;
|
2011-12-12 00:15:06 +00:00
|
|
|
spi->dev.parent = &master->dev;
|
2008-05-15 22:50:22 +00:00
|
|
|
spi->dev.bus = &spi_bus_type;
|
|
|
|
spi->dev.release = spidev_release;
|
2013-02-13 13:20:25 +00:00
|
|
|
spi->cs_gpio = -ENOENT;
|
2015-06-22 13:00:36 +00:00
|
|
|
|
|
|
|
spin_lock_init(&spi->statistics.lock);
|
|
|
|
|
2008-05-15 22:50:22 +00:00
|
|
|
device_initialize(&spi->dev);
|
|
|
|
return spi;
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(spi_alloc_device);
|
|
|
|
|
2013-11-14 12:03:53 +00:00
|
|
|
static void spi_dev_set_name(struct spi_device *spi)
|
|
|
|
{
|
|
|
|
struct acpi_device *adev = ACPI_COMPANION(&spi->dev);
|
|
|
|
|
|
|
|
if (adev) {
|
|
|
|
dev_set_name(&spi->dev, "spi-%s", acpi_dev_name(adev));
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
|
|
|
dev_set_name(&spi->dev, "%s.%u", dev_name(&spi->master->dev),
|
|
|
|
spi->chip_select);
|
|
|
|
}
|
|
|
|
|
2014-01-09 13:23:55 +00:00
|
|
|
static int spi_dev_check(struct device *dev, void *data)
|
|
|
|
{
|
|
|
|
struct spi_device *spi = to_spi_device(dev);
|
|
|
|
struct spi_device *new_spi = data;
|
|
|
|
|
|
|
|
if (spi->master == new_spi->master &&
|
|
|
|
spi->chip_select == new_spi->chip_select)
|
|
|
|
return -EBUSY;
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2008-05-15 22:50:22 +00:00
|
|
|
/**
|
|
|
|
* spi_add_device - Add spi_device allocated with spi_alloc_device
|
|
|
|
* @spi: spi_device to register
|
|
|
|
*
|
|
|
|
* Companion function to spi_alloc_device. Devices allocated with
|
|
|
|
* spi_alloc_device can be added onto the spi bus with this function.
|
|
|
|
*
|
2015-10-22 16:59:23 +00:00
|
|
|
* Return: 0 on success; negative errno on failure
|
2008-05-15 22:50:22 +00:00
|
|
|
*/
|
|
|
|
int spi_add_device(struct spi_device *spi)
|
|
|
|
{
|
2008-08-15 07:40:44 +00:00
|
|
|
static DEFINE_MUTEX(spi_add_lock);
|
2012-11-15 19:19:57 +00:00
|
|
|
struct spi_master *master = spi->master;
|
|
|
|
struct device *dev = master->dev.parent;
|
2008-05-15 22:50:22 +00:00
|
|
|
int status;
|
|
|
|
|
|
|
|
/* Chipselects are numbered 0..max; validate. */
|
2012-11-15 19:19:57 +00:00
|
|
|
if (spi->chip_select >= master->num_chipselect) {
|
2008-05-15 22:50:22 +00:00
|
|
|
dev_err(dev, "cs%d >= max %d\n",
|
|
|
|
spi->chip_select,
|
2012-11-15 19:19:57 +00:00
|
|
|
master->num_chipselect);
|
2008-05-15 22:50:22 +00:00
|
|
|
return -EINVAL;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Set the bus ID string */
|
2013-11-14 12:03:53 +00:00
|
|
|
spi_dev_set_name(spi);
|
2008-08-15 07:40:44 +00:00
|
|
|
|
|
|
|
/* We need to make sure there's no other device with this
|
|
|
|
* chipselect **BEFORE** we call setup(), else we'll trash
|
|
|
|
* its configuration. Lock against concurrent add() calls.
|
|
|
|
*/
|
|
|
|
mutex_lock(&spi_add_lock);
|
|
|
|
|
2014-01-09 13:23:55 +00:00
|
|
|
status = bus_for_each_dev(&spi_bus_type, NULL, spi, spi_dev_check);
|
|
|
|
if (status) {
|
2008-08-15 07:40:44 +00:00
|
|
|
dev_err(dev, "chipselect %d already in use\n",
|
|
|
|
spi->chip_select);
|
|
|
|
goto done;
|
|
|
|
}
|
|
|
|
|
2012-11-15 19:19:57 +00:00
|
|
|
if (master->cs_gpios)
|
|
|
|
spi->cs_gpio = master->cs_gpios[spi->chip_select];
|
|
|
|
|
2008-08-15 07:40:44 +00:00
|
|
|
/* Drivers may modify this initial i/o setup, but will
|
|
|
|
* normally rely on the device being setup. Devices
|
|
|
|
* using SPI_CS_HIGH can't coexist well otherwise...
|
|
|
|
*/
|
2009-06-17 23:26:03 +00:00
|
|
|
status = spi_setup(spi);
|
2008-05-15 22:50:22 +00:00
|
|
|
if (status < 0) {
|
2010-10-21 19:06:44 +00:00
|
|
|
dev_err(dev, "can't setup %s, status %d\n",
|
|
|
|
dev_name(&spi->dev), status);
|
2008-08-15 07:40:44 +00:00
|
|
|
goto done;
|
2008-05-15 22:50:22 +00:00
|
|
|
}
|
|
|
|
|
2008-08-15 07:40:44 +00:00
|
|
|
/* Device may be bound to an active driver when this returns */
|
2008-05-15 22:50:22 +00:00
|
|
|
status = device_add(&spi->dev);
|
2008-08-15 07:40:44 +00:00
|
|
|
if (status < 0)
|
2010-10-21 19:06:44 +00:00
|
|
|
dev_err(dev, "can't add %s, status %d\n",
|
|
|
|
dev_name(&spi->dev), status);
|
2008-08-15 07:40:44 +00:00
|
|
|
else
|
2009-01-06 18:44:37 +00:00
|
|
|
dev_dbg(dev, "registered child %s\n", dev_name(&spi->dev));
|
2008-05-15 22:50:22 +00:00
|
|
|
|
2008-08-15 07:40:44 +00:00
|
|
|
done:
|
|
|
|
mutex_unlock(&spi_add_lock);
|
|
|
|
return status;
|
2008-05-15 22:50:22 +00:00
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(spi_add_device);
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
|
2007-05-08 07:32:21 +00:00
|
|
|
/**
|
|
|
|
* spi_new_device - instantiate one new SPI device
|
|
|
|
* @master: Controller to which device is connected
|
|
|
|
* @chip: Describes the SPI device
|
|
|
|
* Context: can sleep
|
|
|
|
*
|
|
|
|
* On typical mainboards, this is purely internal; and it's not needed
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
* after board init creates the hard-wired devices. Some development
|
|
|
|
* platforms may not be able to use spi_register_board_info though, and
|
|
|
|
* this is exported so that for example a USB or parport based adapter
|
|
|
|
* driver could add devices (which it would learn about out-of-band).
|
2007-07-31 07:39:45 +00:00
|
|
|
*
|
2015-10-22 16:59:23 +00:00
|
|
|
* Return: the new device, or NULL.
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
*/
|
2007-03-27 05:32:23 +00:00
|
|
|
struct spi_device *spi_new_device(struct spi_master *master,
|
|
|
|
struct spi_board_info *chip)
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
{
|
|
|
|
struct spi_device *proxy;
|
|
|
|
int status;
|
|
|
|
|
2007-07-31 07:39:45 +00:00
|
|
|
/* NOTE: caller did any chip->bus_num checks necessary.
|
|
|
|
*
|
|
|
|
* Also, unless we change the return value convention to use
|
|
|
|
* error-or-pointer (not NULL-or-pointer), troubleshootability
|
|
|
|
* suggests syslogged diagnostics are best here (ugh).
|
|
|
|
*/
|
|
|
|
|
2008-05-15 22:50:22 +00:00
|
|
|
proxy = spi_alloc_device(master);
|
|
|
|
if (!proxy)
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
return NULL;
|
|
|
|
|
2008-07-24 04:29:55 +00:00
|
|
|
WARN_ON(strlen(chip->modalias) >= sizeof(proxy->modalias));
|
|
|
|
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
proxy->chip_select = chip->chip_select;
|
|
|
|
proxy->max_speed_hz = chip->max_speed_hz;
|
2006-06-28 14:47:15 +00:00
|
|
|
proxy->mode = chip->mode;
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
proxy->irq = chip->irq;
|
2008-07-24 04:29:55 +00:00
|
|
|
strlcpy(proxy->modalias, chip->modalias, sizeof(proxy->modalias));
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
proxy->dev.platform_data = (void *) chip->platform_data;
|
|
|
|
proxy->controller_data = chip->controller_data;
|
|
|
|
proxy->controller_state = NULL;
|
|
|
|
|
2008-05-15 22:50:22 +00:00
|
|
|
status = spi_add_device(proxy);
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
if (status < 0) {
|
2008-05-15 22:50:22 +00:00
|
|
|
spi_dev_put(proxy);
|
|
|
|
return NULL;
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
return proxy;
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(spi_new_device);
|
|
|
|
|
2010-08-02 07:52:15 +00:00
|
|
|
static void spi_match_master_to_boardinfo(struct spi_master *master,
|
|
|
|
struct spi_board_info *bi)
|
|
|
|
{
|
|
|
|
struct spi_device *dev;
|
|
|
|
|
|
|
|
if (master->bus_num != bi->bus_num)
|
|
|
|
return;
|
|
|
|
|
|
|
|
dev = spi_new_device(master, bi);
|
|
|
|
if (!dev)
|
|
|
|
dev_err(master->dev.parent, "can't create new device for %s\n",
|
|
|
|
bi->modalias);
|
|
|
|
}
|
|
|
|
|
2007-05-08 07:32:21 +00:00
|
|
|
/**
|
|
|
|
* spi_register_board_info - register SPI devices for a given board
|
|
|
|
* @info: array of chip descriptors
|
|
|
|
* @n: how many descriptors are provided
|
|
|
|
* Context: can sleep
|
|
|
|
*
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
* Board-specific early init code calls this (probably during arch_initcall)
|
|
|
|
* with segments of the SPI device table. Any device nodes are created later,
|
|
|
|
* after the relevant parent SPI controller (bus_num) is defined. We keep
|
|
|
|
* this table of devices forever, so that reloading a controller driver will
|
|
|
|
* not make Linux forget about these hard-wired devices.
|
|
|
|
*
|
|
|
|
* Other code can also call this, e.g. a particular add-on board might provide
|
|
|
|
* SPI devices through its expansion connector, so code initializing that board
|
|
|
|
* would naturally declare its SPI devices.
|
|
|
|
*
|
|
|
|
* The board info passed can safely be __initdata ... but be careful of
|
|
|
|
* any embedded pointers (platform_data, etc), they're copied as-is.
|
2015-10-22 16:59:23 +00:00
|
|
|
*
|
|
|
|
* Return: zero on success, else a negative error code.
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
*/
|
2012-12-07 16:57:14 +00:00
|
|
|
int spi_register_board_info(struct spi_board_info const *info, unsigned n)
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
{
|
2010-08-02 07:52:15 +00:00
|
|
|
struct boardinfo *bi;
|
|
|
|
int i;
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
|
2014-09-24 06:30:29 +00:00
|
|
|
if (!n)
|
|
|
|
return -EINVAL;
|
|
|
|
|
2010-08-02 07:52:15 +00:00
|
|
|
bi = kzalloc(n * sizeof(*bi), GFP_KERNEL);
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
if (!bi)
|
|
|
|
return -ENOMEM;
|
|
|
|
|
2010-08-02 07:52:15 +00:00
|
|
|
for (i = 0; i < n; i++, bi++, info++) {
|
|
|
|
struct spi_master *master;
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
|
2010-08-02 07:52:15 +00:00
|
|
|
memcpy(&bi->board_info, info, sizeof(*info));
|
|
|
|
mutex_lock(&board_lock);
|
|
|
|
list_add_tail(&bi->list, &board_list);
|
|
|
|
list_for_each_entry(master, &spi_master_list, list)
|
|
|
|
spi_match_master_to_boardinfo(master, &bi->board_info);
|
|
|
|
mutex_unlock(&board_lock);
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
}
|
2010-08-02 07:52:15 +00:00
|
|
|
|
|
|
|
return 0;
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
/*-------------------------------------------------------------------------*/
|
|
|
|
|
2013-10-05 10:50:40 +00:00
|
|
|
static void spi_set_cs(struct spi_device *spi, bool enable)
|
|
|
|
{
|
|
|
|
if (spi->mode & SPI_CS_HIGH)
|
|
|
|
enable = !enable;
|
|
|
|
|
2015-10-20 09:28:29 +00:00
|
|
|
if (gpio_is_valid(spi->cs_gpio))
|
2013-10-05 10:50:40 +00:00
|
|
|
gpio_set_value(spi->cs_gpio, !enable);
|
|
|
|
else if (spi->master->set_cs)
|
|
|
|
spi->master->set_cs(spi, !enable);
|
|
|
|
}
|
|
|
|
|
2014-05-02 04:29:34 +00:00
|
|
|
#ifdef CONFIG_HAS_DMA
|
2014-02-02 13:47:47 +00:00
|
|
|
static int spi_map_buf(struct spi_master *master, struct device *dev,
|
|
|
|
struct sg_table *sgt, void *buf, size_t len,
|
|
|
|
enum dma_data_direction dir)
|
|
|
|
{
|
|
|
|
const bool vmalloced_buf = is_vmalloc_addr(buf);
|
2015-06-30 15:48:37 +00:00
|
|
|
int desc_len;
|
|
|
|
int sgs;
|
2014-02-02 13:47:47 +00:00
|
|
|
struct page *vm_page;
|
|
|
|
void *sg_buf;
|
|
|
|
size_t min;
|
|
|
|
int i, ret;
|
|
|
|
|
2015-06-30 15:48:37 +00:00
|
|
|
if (vmalloced_buf) {
|
|
|
|
desc_len = PAGE_SIZE;
|
|
|
|
sgs = DIV_ROUND_UP(len + offset_in_page(buf), desc_len);
|
|
|
|
} else {
|
|
|
|
desc_len = master->max_dma_len;
|
|
|
|
sgs = DIV_ROUND_UP(len, desc_len);
|
|
|
|
}
|
|
|
|
|
2014-02-02 13:47:47 +00:00
|
|
|
ret = sg_alloc_table(sgt, sgs, GFP_KERNEL);
|
|
|
|
if (ret != 0)
|
|
|
|
return ret;
|
|
|
|
|
|
|
|
for (i = 0; i < sgs; i++) {
|
|
|
|
|
|
|
|
if (vmalloced_buf) {
|
2015-06-30 15:48:37 +00:00
|
|
|
min = min_t(size_t,
|
|
|
|
len, desc_len - offset_in_page(buf));
|
2014-02-02 13:47:47 +00:00
|
|
|
vm_page = vmalloc_to_page(buf);
|
|
|
|
if (!vm_page) {
|
|
|
|
sg_free_table(sgt);
|
|
|
|
return -ENOMEM;
|
|
|
|
}
|
2014-11-17 09:14:31 +00:00
|
|
|
sg_set_page(&sgt->sgl[i], vm_page,
|
|
|
|
min, offset_in_page(buf));
|
2014-02-02 13:47:47 +00:00
|
|
|
} else {
|
2015-06-30 15:48:37 +00:00
|
|
|
min = min_t(size_t, len, desc_len);
|
2014-02-02 13:47:47 +00:00
|
|
|
sg_buf = buf;
|
2014-11-17 09:14:31 +00:00
|
|
|
sg_set_buf(&sgt->sgl[i], sg_buf, min);
|
2014-02-02 13:47:47 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
buf += min;
|
|
|
|
len -= min;
|
|
|
|
}
|
|
|
|
|
|
|
|
ret = dma_map_sg(dev, sgt->sgl, sgt->nents, dir);
|
2014-07-10 13:29:32 +00:00
|
|
|
if (!ret)
|
|
|
|
ret = -ENOMEM;
|
2014-02-02 13:47:47 +00:00
|
|
|
if (ret < 0) {
|
|
|
|
sg_free_table(sgt);
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
sgt->nents = ret;
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void spi_unmap_buf(struct spi_master *master, struct device *dev,
|
|
|
|
struct sg_table *sgt, enum dma_data_direction dir)
|
|
|
|
{
|
|
|
|
if (sgt->orig_nents) {
|
|
|
|
dma_unmap_sg(dev, sgt->sgl, sgt->orig_nents, dir);
|
|
|
|
sg_free_table(sgt);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2014-05-02 04:29:34 +00:00
|
|
|
static int __spi_map_msg(struct spi_master *master, struct spi_message *msg)
|
2014-01-16 12:22:43 +00:00
|
|
|
{
|
|
|
|
struct device *tx_dev, *rx_dev;
|
|
|
|
struct spi_transfer *xfer;
|
2014-02-02 13:47:47 +00:00
|
|
|
int ret;
|
2014-01-28 20:17:03 +00:00
|
|
|
|
2014-02-02 13:47:47 +00:00
|
|
|
if (!master->can_dma)
|
2014-01-16 12:22:43 +00:00
|
|
|
return 0;
|
|
|
|
|
2015-07-23 09:10:40 +00:00
|
|
|
if (master->dma_tx)
|
|
|
|
tx_dev = master->dma_tx->device->dev;
|
|
|
|
else
|
|
|
|
tx_dev = &master->dev;
|
|
|
|
|
|
|
|
if (master->dma_rx)
|
|
|
|
rx_dev = master->dma_rx->device->dev;
|
|
|
|
else
|
|
|
|
rx_dev = &master->dev;
|
2014-01-16 12:22:43 +00:00
|
|
|
|
|
|
|
list_for_each_entry(xfer, &msg->transfers, transfer_list) {
|
|
|
|
if (!master->can_dma(master, msg->spi, xfer))
|
|
|
|
continue;
|
|
|
|
|
|
|
|
if (xfer->tx_buf != NULL) {
|
2014-02-02 13:47:47 +00:00
|
|
|
ret = spi_map_buf(master, tx_dev, &xfer->tx_sg,
|
|
|
|
(void *)xfer->tx_buf, xfer->len,
|
|
|
|
DMA_TO_DEVICE);
|
|
|
|
if (ret != 0)
|
|
|
|
return ret;
|
2014-01-16 12:22:43 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
if (xfer->rx_buf != NULL) {
|
2014-02-02 13:47:47 +00:00
|
|
|
ret = spi_map_buf(master, rx_dev, &xfer->rx_sg,
|
|
|
|
xfer->rx_buf, xfer->len,
|
|
|
|
DMA_FROM_DEVICE);
|
|
|
|
if (ret != 0) {
|
|
|
|
spi_unmap_buf(master, tx_dev, &xfer->tx_sg,
|
|
|
|
DMA_TO_DEVICE);
|
|
|
|
return ret;
|
2014-01-16 12:22:43 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
master->cur_msg_mapped = true;
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2015-05-25 10:13:10 +00:00
|
|
|
static int __spi_unmap_msg(struct spi_master *master, struct spi_message *msg)
|
2014-01-16 12:22:43 +00:00
|
|
|
{
|
|
|
|
struct spi_transfer *xfer;
|
|
|
|
struct device *tx_dev, *rx_dev;
|
|
|
|
|
2014-02-02 13:47:47 +00:00
|
|
|
if (!master->cur_msg_mapped || !master->can_dma)
|
2014-01-16 12:22:43 +00:00
|
|
|
return 0;
|
|
|
|
|
2015-07-23 09:10:40 +00:00
|
|
|
if (master->dma_tx)
|
|
|
|
tx_dev = master->dma_tx->device->dev;
|
|
|
|
else
|
|
|
|
tx_dev = &master->dev;
|
|
|
|
|
|
|
|
if (master->dma_rx)
|
|
|
|
rx_dev = master->dma_rx->device->dev;
|
|
|
|
else
|
|
|
|
rx_dev = &master->dev;
|
2014-01-16 12:22:43 +00:00
|
|
|
|
|
|
|
list_for_each_entry(xfer, &msg->transfers, transfer_list) {
|
|
|
|
if (!master->can_dma(master, msg->spi, xfer))
|
|
|
|
continue;
|
|
|
|
|
2014-02-02 13:47:47 +00:00
|
|
|
spi_unmap_buf(master, rx_dev, &xfer->rx_sg, DMA_FROM_DEVICE);
|
|
|
|
spi_unmap_buf(master, tx_dev, &xfer->tx_sg, DMA_TO_DEVICE);
|
2014-01-16 12:22:43 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
2014-05-02 04:29:34 +00:00
|
|
|
#else /* !CONFIG_HAS_DMA */
|
|
|
|
static inline int __spi_map_msg(struct spi_master *master,
|
|
|
|
struct spi_message *msg)
|
|
|
|
{
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2015-05-25 10:13:10 +00:00
|
|
|
static inline int __spi_unmap_msg(struct spi_master *master,
|
|
|
|
struct spi_message *msg)
|
2014-05-02 04:29:34 +00:00
|
|
|
{
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
#endif /* !CONFIG_HAS_DMA */
|
|
|
|
|
2015-05-25 10:13:10 +00:00
|
|
|
static inline int spi_unmap_msg(struct spi_master *master,
|
|
|
|
struct spi_message *msg)
|
|
|
|
{
|
|
|
|
struct spi_transfer *xfer;
|
|
|
|
|
|
|
|
list_for_each_entry(xfer, &msg->transfers, transfer_list) {
|
|
|
|
/*
|
|
|
|
* Restore the original value of tx_buf or rx_buf if they are
|
|
|
|
* NULL.
|
|
|
|
*/
|
|
|
|
if (xfer->tx_buf == master->dummy_tx)
|
|
|
|
xfer->tx_buf = NULL;
|
|
|
|
if (xfer->rx_buf == master->dummy_rx)
|
|
|
|
xfer->rx_buf = NULL;
|
|
|
|
}
|
|
|
|
|
|
|
|
return __spi_unmap_msg(master, msg);
|
|
|
|
}
|
|
|
|
|
2014-05-02 04:29:34 +00:00
|
|
|
static int spi_map_msg(struct spi_master *master, struct spi_message *msg)
|
|
|
|
{
|
|
|
|
struct spi_transfer *xfer;
|
|
|
|
void *tmp;
|
|
|
|
unsigned int max_tx, max_rx;
|
|
|
|
|
|
|
|
if (master->flags & (SPI_MASTER_MUST_RX | SPI_MASTER_MUST_TX)) {
|
|
|
|
max_tx = 0;
|
|
|
|
max_rx = 0;
|
|
|
|
|
|
|
|
list_for_each_entry(xfer, &msg->transfers, transfer_list) {
|
|
|
|
if ((master->flags & SPI_MASTER_MUST_TX) &&
|
|
|
|
!xfer->tx_buf)
|
|
|
|
max_tx = max(xfer->len, max_tx);
|
|
|
|
if ((master->flags & SPI_MASTER_MUST_RX) &&
|
|
|
|
!xfer->rx_buf)
|
|
|
|
max_rx = max(xfer->len, max_rx);
|
|
|
|
}
|
|
|
|
|
|
|
|
if (max_tx) {
|
|
|
|
tmp = krealloc(master->dummy_tx, max_tx,
|
|
|
|
GFP_KERNEL | GFP_DMA);
|
|
|
|
if (!tmp)
|
|
|
|
return -ENOMEM;
|
|
|
|
master->dummy_tx = tmp;
|
|
|
|
memset(tmp, 0, max_tx);
|
|
|
|
}
|
|
|
|
|
|
|
|
if (max_rx) {
|
|
|
|
tmp = krealloc(master->dummy_rx, max_rx,
|
|
|
|
GFP_KERNEL | GFP_DMA);
|
|
|
|
if (!tmp)
|
|
|
|
return -ENOMEM;
|
|
|
|
master->dummy_rx = tmp;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (max_tx || max_rx) {
|
|
|
|
list_for_each_entry(xfer, &msg->transfers,
|
|
|
|
transfer_list) {
|
|
|
|
if (!xfer->tx_buf)
|
|
|
|
xfer->tx_buf = master->dummy_tx;
|
|
|
|
if (!xfer->rx_buf)
|
|
|
|
xfer->rx_buf = master->dummy_rx;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
return __spi_map_msg(master, msg);
|
|
|
|
}
|
2014-01-16 12:22:43 +00:00
|
|
|
|
2013-10-05 10:50:40 +00:00
|
|
|
/*
|
|
|
|
* spi_transfer_one_message - Default implementation of transfer_one_message()
|
|
|
|
*
|
|
|
|
* This is a standard implementation of transfer_one_message() for
|
|
|
|
* drivers which impelment a transfer_one() operation. It provides
|
|
|
|
* standard handling of delays and chip select management.
|
|
|
|
*/
|
|
|
|
static int spi_transfer_one_message(struct spi_master *master,
|
|
|
|
struct spi_message *msg)
|
|
|
|
{
|
|
|
|
struct spi_transfer *xfer;
|
|
|
|
bool keep_cs = false;
|
|
|
|
int ret = 0;
|
2015-02-02 08:30:32 +00:00
|
|
|
unsigned long ms = 1;
|
2015-06-22 13:00:36 +00:00
|
|
|
struct spi_statistics *statm = &master->statistics;
|
|
|
|
struct spi_statistics *stats = &msg->spi->statistics;
|
2013-10-05 10:50:40 +00:00
|
|
|
|
|
|
|
spi_set_cs(msg->spi, true);
|
|
|
|
|
2015-06-22 13:00:36 +00:00
|
|
|
SPI_STATISTICS_INCREMENT_FIELD(statm, messages);
|
|
|
|
SPI_STATISTICS_INCREMENT_FIELD(stats, messages);
|
|
|
|
|
2013-10-05 10:50:40 +00:00
|
|
|
list_for_each_entry(xfer, &msg->transfers, transfer_list) {
|
|
|
|
trace_spi_transfer_start(msg, xfer);
|
|
|
|
|
2015-06-22 13:00:36 +00:00
|
|
|
spi_statistics_add_transfer_stats(statm, xfer, master);
|
|
|
|
spi_statistics_add_transfer_stats(stats, xfer, master);
|
|
|
|
|
2014-08-16 15:27:41 +00:00
|
|
|
if (xfer->tx_buf || xfer->rx_buf) {
|
|
|
|
reinit_completion(&master->xfer_completion);
|
2013-10-05 10:50:40 +00:00
|
|
|
|
2014-08-16 15:27:41 +00:00
|
|
|
ret = master->transfer_one(master, msg->spi, xfer);
|
|
|
|
if (ret < 0) {
|
2015-06-22 13:00:36 +00:00
|
|
|
SPI_STATISTICS_INCREMENT_FIELD(statm,
|
|
|
|
errors);
|
|
|
|
SPI_STATISTICS_INCREMENT_FIELD(stats,
|
|
|
|
errors);
|
2014-08-16 15:27:41 +00:00
|
|
|
dev_err(&msg->spi->dev,
|
|
|
|
"SPI transfer failed: %d\n", ret);
|
|
|
|
goto out;
|
|
|
|
}
|
2013-10-05 10:50:40 +00:00
|
|
|
|
2014-08-16 15:27:41 +00:00
|
|
|
if (ret > 0) {
|
|
|
|
ret = 0;
|
|
|
|
ms = xfer->len * 8 * 1000 / xfer->speed_hz;
|
|
|
|
ms += ms + 100; /* some tolerance */
|
2014-01-30 22:16:41 +00:00
|
|
|
|
2014-08-16 15:27:41 +00:00
|
|
|
ms = wait_for_completion_timeout(&master->xfer_completion,
|
|
|
|
msecs_to_jiffies(ms));
|
|
|
|
}
|
2014-01-30 22:16:41 +00:00
|
|
|
|
2014-08-16 15:27:41 +00:00
|
|
|
if (ms == 0) {
|
2015-06-22 13:00:36 +00:00
|
|
|
SPI_STATISTICS_INCREMENT_FIELD(statm,
|
|
|
|
timedout);
|
|
|
|
SPI_STATISTICS_INCREMENT_FIELD(stats,
|
|
|
|
timedout);
|
2014-08-16 15:27:41 +00:00
|
|
|
dev_err(&msg->spi->dev,
|
|
|
|
"SPI transfer timed out\n");
|
|
|
|
msg->status = -ETIMEDOUT;
|
|
|
|
}
|
|
|
|
} else {
|
|
|
|
if (xfer->len)
|
|
|
|
dev_err(&msg->spi->dev,
|
|
|
|
"Bufferless transfer has length %u\n",
|
|
|
|
xfer->len);
|
2014-01-18 14:05:22 +00:00
|
|
|
}
|
2013-10-05 10:50:40 +00:00
|
|
|
|
|
|
|
trace_spi_transfer_stop(msg, xfer);
|
|
|
|
|
|
|
|
if (msg->status != -EINPROGRESS)
|
|
|
|
goto out;
|
|
|
|
|
|
|
|
if (xfer->delay_usecs)
|
|
|
|
udelay(xfer->delay_usecs);
|
|
|
|
|
|
|
|
if (xfer->cs_change) {
|
|
|
|
if (list_is_last(&xfer->transfer_list,
|
|
|
|
&msg->transfers)) {
|
|
|
|
keep_cs = true;
|
|
|
|
} else {
|
2014-03-29 23:48:07 +00:00
|
|
|
spi_set_cs(msg->spi, false);
|
|
|
|
udelay(10);
|
|
|
|
spi_set_cs(msg->spi, true);
|
2013-10-05 10:50:40 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
msg->actual_length += xfer->len;
|
|
|
|
}
|
|
|
|
|
|
|
|
out:
|
|
|
|
if (ret != 0 || !keep_cs)
|
|
|
|
spi_set_cs(msg->spi, false);
|
|
|
|
|
|
|
|
if (msg->status == -EINPROGRESS)
|
|
|
|
msg->status = ret;
|
|
|
|
|
2015-04-07 18:39:19 +00:00
|
|
|
if (msg->status && master->handle_err)
|
2015-02-27 15:34:15 +00:00
|
|
|
master->handle_err(master, msg);
|
|
|
|
|
2013-10-05 10:50:40 +00:00
|
|
|
spi_finalize_current_message(master);
|
|
|
|
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
|
|
|
* spi_finalize_current_transfer - report completion of a transfer
|
2014-08-08 11:02:36 +00:00
|
|
|
* @master: the master reporting completion
|
2013-10-05 10:50:40 +00:00
|
|
|
*
|
|
|
|
* Called by SPI drivers using the core transfer_one_message()
|
|
|
|
* implementation to notify it that the current interrupt driven
|
2014-01-21 15:10:05 +00:00
|
|
|
* transfer has finished and the next one may be scheduled.
|
2013-10-05 10:50:40 +00:00
|
|
|
*/
|
|
|
|
void spi_finalize_current_transfer(struct spi_master *master)
|
|
|
|
{
|
|
|
|
complete(&master->xfer_completion);
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(spi_finalize_current_transfer);
|
|
|
|
|
2012-02-22 09:05:38 +00:00
|
|
|
/**
|
2014-12-10 13:46:33 +00:00
|
|
|
* __spi_pump_messages - function which processes spi message queue
|
|
|
|
* @master: master to process queue for
|
|
|
|
* @in_kthread: true if we are in the context of the message pump thread
|
2012-02-22 09:05:38 +00:00
|
|
|
*
|
|
|
|
* This function checks if there is any spi message in the queue that
|
|
|
|
* needs processing and if so call out to the driver to initialize hardware
|
|
|
|
* and transfer each message.
|
|
|
|
*
|
2014-12-09 21:38:05 +00:00
|
|
|
* Note that it is called both from the kthread itself and also from
|
|
|
|
* inside spi_sync(); the queue extraction handling at the top of the
|
|
|
|
* function should deal with this safely.
|
2012-02-22 09:05:38 +00:00
|
|
|
*/
|
2014-12-10 13:46:33 +00:00
|
|
|
static void __spi_pump_messages(struct spi_master *master, bool in_kthread)
|
2012-02-22 09:05:38 +00:00
|
|
|
{
|
|
|
|
unsigned long flags;
|
|
|
|
bool was_busy = false;
|
|
|
|
int ret;
|
|
|
|
|
2014-12-09 19:46:56 +00:00
|
|
|
/* Lock queue */
|
2012-02-22 09:05:38 +00:00
|
|
|
spin_lock_irqsave(&master->queue_lock, flags);
|
2014-12-09 19:46:56 +00:00
|
|
|
|
|
|
|
/* Make sure we are not already running a message */
|
|
|
|
if (master->cur_msg) {
|
|
|
|
spin_unlock_irqrestore(&master->queue_lock, flags);
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
2014-12-09 21:38:05 +00:00
|
|
|
/* If another context is idling the device then defer */
|
|
|
|
if (master->idling) {
|
|
|
|
queue_kthread_work(&master->kworker, &master->pump_messages);
|
|
|
|
spin_unlock_irqrestore(&master->queue_lock, flags);
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
2014-12-09 19:46:56 +00:00
|
|
|
/* Check if the queue is idle */
|
2012-02-22 09:05:38 +00:00
|
|
|
if (list_empty(&master->queue) || !master->running) {
|
2013-03-13 18:17:40 +00:00
|
|
|
if (!master->busy) {
|
|
|
|
spin_unlock_irqrestore(&master->queue_lock, flags);
|
|
|
|
return;
|
2012-02-22 09:05:38 +00:00
|
|
|
}
|
2014-12-10 13:46:33 +00:00
|
|
|
|
|
|
|
/* Only do teardown in the thread */
|
|
|
|
if (!in_kthread) {
|
|
|
|
queue_kthread_work(&master->kworker,
|
|
|
|
&master->pump_messages);
|
|
|
|
spin_unlock_irqrestore(&master->queue_lock, flags);
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
2012-02-22 09:05:38 +00:00
|
|
|
master->busy = false;
|
2014-12-09 21:38:05 +00:00
|
|
|
master->idling = true;
|
2012-02-22 09:05:38 +00:00
|
|
|
spin_unlock_irqrestore(&master->queue_lock, flags);
|
2014-12-09 21:38:05 +00:00
|
|
|
|
2014-01-28 20:17:03 +00:00
|
|
|
kfree(master->dummy_rx);
|
|
|
|
master->dummy_rx = NULL;
|
|
|
|
kfree(master->dummy_tx);
|
|
|
|
master->dummy_tx = NULL;
|
2013-03-13 18:17:40 +00:00
|
|
|
if (master->unprepare_transfer_hardware &&
|
|
|
|
master->unprepare_transfer_hardware(master))
|
|
|
|
dev_err(&master->dev,
|
|
|
|
"failed to unprepare transfer hardware\n");
|
2013-07-28 13:47:02 +00:00
|
|
|
if (master->auto_runtime_pm) {
|
|
|
|
pm_runtime_mark_last_busy(master->dev.parent);
|
|
|
|
pm_runtime_put_autosuspend(master->dev.parent);
|
|
|
|
}
|
2013-10-07 18:33:53 +00:00
|
|
|
trace_spi_master_idle(master);
|
2012-02-22 09:05:38 +00:00
|
|
|
|
2014-12-09 21:38:05 +00:00
|
|
|
spin_lock_irqsave(&master->queue_lock, flags);
|
|
|
|
master->idling = false;
|
2012-02-22 09:05:38 +00:00
|
|
|
spin_unlock_irqrestore(&master->queue_lock, flags);
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Extract head of queue */
|
|
|
|
master->cur_msg =
|
2014-01-09 08:03:58 +00:00
|
|
|
list_first_entry(&master->queue, struct spi_message, queue);
|
2012-02-22 09:05:38 +00:00
|
|
|
|
|
|
|
list_del_init(&master->cur_msg->queue);
|
|
|
|
if (master->busy)
|
|
|
|
was_busy = true;
|
|
|
|
else
|
|
|
|
master->busy = true;
|
|
|
|
spin_unlock_irqrestore(&master->queue_lock, flags);
|
|
|
|
|
2013-07-28 13:47:02 +00:00
|
|
|
if (!was_busy && master->auto_runtime_pm) {
|
|
|
|
ret = pm_runtime_get_sync(master->dev.parent);
|
|
|
|
if (ret < 0) {
|
|
|
|
dev_err(&master->dev, "Failed to power device: %d\n",
|
|
|
|
ret);
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2013-10-07 18:33:53 +00:00
|
|
|
if (!was_busy)
|
|
|
|
trace_spi_master_busy(master);
|
|
|
|
|
2012-05-10 13:50:41 +00:00
|
|
|
if (!was_busy && master->prepare_transfer_hardware) {
|
2012-02-22 09:05:38 +00:00
|
|
|
ret = master->prepare_transfer_hardware(master);
|
|
|
|
if (ret) {
|
|
|
|
dev_err(&master->dev,
|
|
|
|
"failed to prepare transfer hardware\n");
|
2013-07-28 13:47:02 +00:00
|
|
|
|
|
|
|
if (master->auto_runtime_pm)
|
|
|
|
pm_runtime_put(master->dev.parent);
|
2012-02-22 09:05:38 +00:00
|
|
|
return;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2013-10-07 18:33:53 +00:00
|
|
|
trace_spi_message_start(master->cur_msg);
|
|
|
|
|
2013-10-04 23:23:12 +00:00
|
|
|
if (master->prepare_message) {
|
|
|
|
ret = master->prepare_message(master, master->cur_msg);
|
|
|
|
if (ret) {
|
|
|
|
dev_err(&master->dev,
|
|
|
|
"failed to prepare message: %d\n", ret);
|
|
|
|
master->cur_msg->status = ret;
|
|
|
|
spi_finalize_current_message(master);
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
master->cur_msg_prepared = true;
|
|
|
|
}
|
|
|
|
|
2014-01-16 12:22:43 +00:00
|
|
|
ret = spi_map_msg(master, master->cur_msg);
|
|
|
|
if (ret) {
|
|
|
|
master->cur_msg->status = ret;
|
|
|
|
spi_finalize_current_message(master);
|
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
2012-02-22 09:05:38 +00:00
|
|
|
ret = master->transfer_one_message(master, master->cur_msg);
|
|
|
|
if (ret) {
|
|
|
|
dev_err(&master->dev,
|
2014-01-28 09:33:03 +00:00
|
|
|
"failed to transfer one message from queue\n");
|
2012-02-22 09:05:38 +00:00
|
|
|
return;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2014-12-10 13:46:33 +00:00
|
|
|
/**
|
|
|
|
* spi_pump_messages - kthread work function which processes spi message queue
|
|
|
|
* @work: pointer to kthread work struct contained in the master struct
|
|
|
|
*/
|
|
|
|
static void spi_pump_messages(struct kthread_work *work)
|
|
|
|
{
|
|
|
|
struct spi_master *master =
|
|
|
|
container_of(work, struct spi_master, pump_messages);
|
|
|
|
|
|
|
|
__spi_pump_messages(master, true);
|
|
|
|
}
|
|
|
|
|
2012-02-22 09:05:38 +00:00
|
|
|
static int spi_init_queue(struct spi_master *master)
|
|
|
|
{
|
|
|
|
struct sched_param param = { .sched_priority = MAX_RT_PRIO - 1 };
|
|
|
|
|
|
|
|
master->running = false;
|
|
|
|
master->busy = false;
|
|
|
|
|
|
|
|
init_kthread_worker(&master->kworker);
|
|
|
|
master->kworker_task = kthread_run(kthread_worker_fn,
|
2013-07-03 22:04:58 +00:00
|
|
|
&master->kworker, "%s",
|
2012-02-22 09:05:38 +00:00
|
|
|
dev_name(&master->dev));
|
|
|
|
if (IS_ERR(master->kworker_task)) {
|
|
|
|
dev_err(&master->dev, "failed to create message pump task\n");
|
2014-12-04 09:02:25 +00:00
|
|
|
return PTR_ERR(master->kworker_task);
|
2012-02-22 09:05:38 +00:00
|
|
|
}
|
|
|
|
init_kthread_work(&master->pump_messages, spi_pump_messages);
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Master config will indicate if this controller should run the
|
|
|
|
* message pump with high (realtime) priority to reduce the transfer
|
|
|
|
* latency on the bus by minimising the delay between a transfer
|
|
|
|
* request and the scheduling of the message pump thread. Without this
|
|
|
|
* setting the message pump thread will remain at default priority.
|
|
|
|
*/
|
|
|
|
if (master->rt) {
|
|
|
|
dev_info(&master->dev,
|
|
|
|
"will run message pump with realtime priority\n");
|
|
|
|
sched_setscheduler(master->kworker_task, SCHED_FIFO, ¶m);
|
|
|
|
}
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
|
|
|
* spi_get_next_queued_message() - called by driver to check for queued
|
|
|
|
* messages
|
|
|
|
* @master: the master to check for queued messages
|
|
|
|
*
|
|
|
|
* If there are more messages in the queue, the next message is returned from
|
|
|
|
* this call.
|
2015-10-22 16:59:23 +00:00
|
|
|
*
|
|
|
|
* Return: the next message in the queue, else NULL if the queue is empty.
|
2012-02-22 09:05:38 +00:00
|
|
|
*/
|
|
|
|
struct spi_message *spi_get_next_queued_message(struct spi_master *master)
|
|
|
|
{
|
|
|
|
struct spi_message *next;
|
|
|
|
unsigned long flags;
|
|
|
|
|
|
|
|
/* get a pointer to the next message, if any */
|
|
|
|
spin_lock_irqsave(&master->queue_lock, flags);
|
2014-01-02 07:16:40 +00:00
|
|
|
next = list_first_entry_or_null(&master->queue, struct spi_message,
|
|
|
|
queue);
|
2012-02-22 09:05:38 +00:00
|
|
|
spin_unlock_irqrestore(&master->queue_lock, flags);
|
|
|
|
|
|
|
|
return next;
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(spi_get_next_queued_message);
|
|
|
|
|
|
|
|
/**
|
|
|
|
* spi_finalize_current_message() - the current message is complete
|
|
|
|
* @master: the master to return the message to
|
|
|
|
*
|
|
|
|
* Called by the driver to notify the core that the message in the front of the
|
|
|
|
* queue is complete and can be removed from the queue.
|
|
|
|
*/
|
|
|
|
void spi_finalize_current_message(struct spi_master *master)
|
|
|
|
{
|
|
|
|
struct spi_message *mesg;
|
|
|
|
unsigned long flags;
|
2013-10-04 23:23:12 +00:00
|
|
|
int ret;
|
2012-02-22 09:05:38 +00:00
|
|
|
|
|
|
|
spin_lock_irqsave(&master->queue_lock, flags);
|
|
|
|
mesg = master->cur_msg;
|
|
|
|
spin_unlock_irqrestore(&master->queue_lock, flags);
|
|
|
|
|
2014-01-16 12:22:43 +00:00
|
|
|
spi_unmap_msg(master, mesg);
|
|
|
|
|
2013-10-04 23:23:12 +00:00
|
|
|
if (master->cur_msg_prepared && master->unprepare_message) {
|
|
|
|
ret = master->unprepare_message(master, mesg);
|
|
|
|
if (ret) {
|
|
|
|
dev_err(&master->dev,
|
|
|
|
"failed to unprepare message: %d\n", ret);
|
|
|
|
}
|
|
|
|
}
|
2015-03-18 10:27:28 +00:00
|
|
|
|
2015-05-10 07:50:45 +00:00
|
|
|
spin_lock_irqsave(&master->queue_lock, flags);
|
|
|
|
master->cur_msg = NULL;
|
2013-10-04 23:23:12 +00:00
|
|
|
master->cur_msg_prepared = false;
|
2015-05-10 07:50:45 +00:00
|
|
|
queue_kthread_work(&master->kworker, &master->pump_messages);
|
|
|
|
spin_unlock_irqrestore(&master->queue_lock, flags);
|
|
|
|
|
|
|
|
trace_spi_message_done(mesg);
|
2013-10-04 23:23:12 +00:00
|
|
|
|
2012-02-22 09:05:38 +00:00
|
|
|
mesg->state = NULL;
|
|
|
|
if (mesg->complete)
|
|
|
|
mesg->complete(mesg->context);
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(spi_finalize_current_message);
|
|
|
|
|
|
|
|
static int spi_start_queue(struct spi_master *master)
|
|
|
|
{
|
|
|
|
unsigned long flags;
|
|
|
|
|
|
|
|
spin_lock_irqsave(&master->queue_lock, flags);
|
|
|
|
|
|
|
|
if (master->running || master->busy) {
|
|
|
|
spin_unlock_irqrestore(&master->queue_lock, flags);
|
|
|
|
return -EBUSY;
|
|
|
|
}
|
|
|
|
|
|
|
|
master->running = true;
|
|
|
|
master->cur_msg = NULL;
|
|
|
|
spin_unlock_irqrestore(&master->queue_lock, flags);
|
|
|
|
|
|
|
|
queue_kthread_work(&master->kworker, &master->pump_messages);
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int spi_stop_queue(struct spi_master *master)
|
|
|
|
{
|
|
|
|
unsigned long flags;
|
|
|
|
unsigned limit = 500;
|
|
|
|
int ret = 0;
|
|
|
|
|
|
|
|
spin_lock_irqsave(&master->queue_lock, flags);
|
|
|
|
|
|
|
|
/*
|
|
|
|
* This is a bit lame, but is optimized for the common execution path.
|
|
|
|
* A wait_queue on the master->busy could be used, but then the common
|
|
|
|
* execution path (pump_messages) would be required to call wake_up or
|
|
|
|
* friends on every SPI message. Do this instead.
|
|
|
|
*/
|
|
|
|
while ((!list_empty(&master->queue) || master->busy) && limit--) {
|
|
|
|
spin_unlock_irqrestore(&master->queue_lock, flags);
|
2014-02-21 01:15:18 +00:00
|
|
|
usleep_range(10000, 11000);
|
2012-02-22 09:05:38 +00:00
|
|
|
spin_lock_irqsave(&master->queue_lock, flags);
|
|
|
|
}
|
|
|
|
|
|
|
|
if (!list_empty(&master->queue) || master->busy)
|
|
|
|
ret = -EBUSY;
|
|
|
|
else
|
|
|
|
master->running = false;
|
|
|
|
|
|
|
|
spin_unlock_irqrestore(&master->queue_lock, flags);
|
|
|
|
|
|
|
|
if (ret) {
|
|
|
|
dev_warn(&master->dev,
|
|
|
|
"could not stop message queue\n");
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int spi_destroy_queue(struct spi_master *master)
|
|
|
|
{
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
ret = spi_stop_queue(master);
|
|
|
|
|
|
|
|
/*
|
|
|
|
* flush_kthread_worker will block until all work is done.
|
|
|
|
* If the reason that stop_queue timed out is that the work will never
|
|
|
|
* finish, then it does no good to call flush/stop thread, so
|
|
|
|
* return anyway.
|
|
|
|
*/
|
|
|
|
if (ret) {
|
|
|
|
dev_err(&master->dev, "problem destroying queue\n");
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
flush_kthread_worker(&master->kworker);
|
|
|
|
kthread_stop(master->kworker_task);
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2014-12-09 21:38:05 +00:00
|
|
|
static int __spi_queued_transfer(struct spi_device *spi,
|
|
|
|
struct spi_message *msg,
|
|
|
|
bool need_pump)
|
2012-02-22 09:05:38 +00:00
|
|
|
{
|
|
|
|
struct spi_master *master = spi->master;
|
|
|
|
unsigned long flags;
|
|
|
|
|
|
|
|
spin_lock_irqsave(&master->queue_lock, flags);
|
|
|
|
|
|
|
|
if (!master->running) {
|
|
|
|
spin_unlock_irqrestore(&master->queue_lock, flags);
|
|
|
|
return -ESHUTDOWN;
|
|
|
|
}
|
|
|
|
msg->actual_length = 0;
|
|
|
|
msg->status = -EINPROGRESS;
|
|
|
|
|
|
|
|
list_add_tail(&msg->queue, &master->queue);
|
2014-12-09 21:38:05 +00:00
|
|
|
if (!master->busy && need_pump)
|
2012-02-22 09:05:38 +00:00
|
|
|
queue_kthread_work(&master->kworker, &master->pump_messages);
|
|
|
|
|
|
|
|
spin_unlock_irqrestore(&master->queue_lock, flags);
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2014-12-09 21:38:05 +00:00
|
|
|
/**
|
|
|
|
* spi_queued_transfer - transfer function for queued transfers
|
|
|
|
* @spi: spi device which is requesting transfer
|
|
|
|
* @msg: spi message which is to handled is queued to driver queue
|
2015-10-22 16:59:23 +00:00
|
|
|
*
|
|
|
|
* Return: zero on success, else a negative error code.
|
2014-12-09 21:38:05 +00:00
|
|
|
*/
|
|
|
|
static int spi_queued_transfer(struct spi_device *spi, struct spi_message *msg)
|
|
|
|
{
|
|
|
|
return __spi_queued_transfer(spi, msg, true);
|
|
|
|
}
|
|
|
|
|
2012-02-22 09:05:38 +00:00
|
|
|
static int spi_master_initialize_queue(struct spi_master *master)
|
|
|
|
{
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
master->transfer = spi_queued_transfer;
|
2013-10-05 10:50:40 +00:00
|
|
|
if (!master->transfer_one_message)
|
|
|
|
master->transfer_one_message = spi_transfer_one_message;
|
2012-02-22 09:05:38 +00:00
|
|
|
|
|
|
|
/* Initialize and start queue */
|
|
|
|
ret = spi_init_queue(master);
|
|
|
|
if (ret) {
|
|
|
|
dev_err(&master->dev, "problem initializing queue\n");
|
|
|
|
goto err_init_queue;
|
|
|
|
}
|
2014-05-01 17:47:52 +00:00
|
|
|
master->queued = true;
|
2012-02-22 09:05:38 +00:00
|
|
|
ret = spi_start_queue(master);
|
|
|
|
if (ret) {
|
|
|
|
dev_err(&master->dev, "problem starting queue\n");
|
|
|
|
goto err_start_queue;
|
|
|
|
}
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
err_start_queue:
|
|
|
|
spi_destroy_queue(master);
|
2014-05-01 17:47:52 +00:00
|
|
|
err_init_queue:
|
2012-02-22 09:05:38 +00:00
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*-------------------------------------------------------------------------*/
|
|
|
|
|
2012-12-04 14:09:38 +00:00
|
|
|
#if defined(CONFIG_OF)
|
2014-10-29 08:40:37 +00:00
|
|
|
static struct spi_device *
|
|
|
|
of_register_spi_device(struct spi_master *master, struct device_node *nc)
|
|
|
|
{
|
|
|
|
struct spi_device *spi;
|
|
|
|
int rc;
|
|
|
|
u32 value;
|
|
|
|
|
|
|
|
/* Alloc an spi_device */
|
|
|
|
spi = spi_alloc_device(master);
|
|
|
|
if (!spi) {
|
|
|
|
dev_err(&master->dev, "spi_device alloc error for %s\n",
|
|
|
|
nc->full_name);
|
|
|
|
rc = -ENOMEM;
|
|
|
|
goto err_out;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Select device driver */
|
|
|
|
rc = of_modalias_node(nc, spi->modalias,
|
|
|
|
sizeof(spi->modalias));
|
|
|
|
if (rc < 0) {
|
|
|
|
dev_err(&master->dev, "cannot find modalias for %s\n",
|
|
|
|
nc->full_name);
|
|
|
|
goto err_out;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Device address */
|
|
|
|
rc = of_property_read_u32(nc, "reg", &value);
|
|
|
|
if (rc) {
|
|
|
|
dev_err(&master->dev, "%s has no valid 'reg' property (%d)\n",
|
|
|
|
nc->full_name, rc);
|
|
|
|
goto err_out;
|
|
|
|
}
|
|
|
|
spi->chip_select = value;
|
|
|
|
|
|
|
|
/* Mode (clock phase/polarity/etc.) */
|
|
|
|
if (of_find_property(nc, "spi-cpha", NULL))
|
|
|
|
spi->mode |= SPI_CPHA;
|
|
|
|
if (of_find_property(nc, "spi-cpol", NULL))
|
|
|
|
spi->mode |= SPI_CPOL;
|
|
|
|
if (of_find_property(nc, "spi-cs-high", NULL))
|
|
|
|
spi->mode |= SPI_CS_HIGH;
|
|
|
|
if (of_find_property(nc, "spi-3wire", NULL))
|
|
|
|
spi->mode |= SPI_3WIRE;
|
|
|
|
if (of_find_property(nc, "spi-lsb-first", NULL))
|
|
|
|
spi->mode |= SPI_LSB_FIRST;
|
|
|
|
|
|
|
|
/* Device DUAL/QUAD mode */
|
|
|
|
if (!of_property_read_u32(nc, "spi-tx-bus-width", &value)) {
|
|
|
|
switch (value) {
|
|
|
|
case 1:
|
|
|
|
break;
|
|
|
|
case 2:
|
|
|
|
spi->mode |= SPI_TX_DUAL;
|
|
|
|
break;
|
|
|
|
case 4:
|
|
|
|
spi->mode |= SPI_TX_QUAD;
|
|
|
|
break;
|
|
|
|
default:
|
|
|
|
dev_warn(&master->dev,
|
|
|
|
"spi-tx-bus-width %d not supported\n",
|
|
|
|
value);
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
if (!of_property_read_u32(nc, "spi-rx-bus-width", &value)) {
|
|
|
|
switch (value) {
|
|
|
|
case 1:
|
|
|
|
break;
|
|
|
|
case 2:
|
|
|
|
spi->mode |= SPI_RX_DUAL;
|
|
|
|
break;
|
|
|
|
case 4:
|
|
|
|
spi->mode |= SPI_RX_QUAD;
|
|
|
|
break;
|
|
|
|
default:
|
|
|
|
dev_warn(&master->dev,
|
|
|
|
"spi-rx-bus-width %d not supported\n",
|
|
|
|
value);
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Device speed */
|
|
|
|
rc = of_property_read_u32(nc, "spi-max-frequency", &value);
|
|
|
|
if (rc) {
|
|
|
|
dev_err(&master->dev, "%s has no valid 'spi-max-frequency' property (%d)\n",
|
|
|
|
nc->full_name, rc);
|
|
|
|
goto err_out;
|
|
|
|
}
|
|
|
|
spi->max_speed_hz = value;
|
|
|
|
|
|
|
|
/* Store a pointer to the node in the device structure */
|
|
|
|
of_node_get(nc);
|
|
|
|
spi->dev.of_node = nc;
|
|
|
|
|
|
|
|
/* Register the new device */
|
|
|
|
rc = spi_add_device(spi);
|
|
|
|
if (rc) {
|
|
|
|
dev_err(&master->dev, "spi_device register error %s\n",
|
|
|
|
nc->full_name);
|
|
|
|
goto err_out;
|
|
|
|
}
|
|
|
|
|
|
|
|
return spi;
|
|
|
|
|
|
|
|
err_out:
|
|
|
|
spi_dev_put(spi);
|
|
|
|
return ERR_PTR(rc);
|
|
|
|
}
|
|
|
|
|
2012-04-07 20:16:53 +00:00
|
|
|
/**
|
|
|
|
* of_register_spi_devices() - Register child devices onto the SPI bus
|
|
|
|
* @master: Pointer to spi_master device
|
|
|
|
*
|
|
|
|
* Registers an spi_device for each child node of master node which has a 'reg'
|
|
|
|
* property.
|
|
|
|
*/
|
|
|
|
static void of_register_spi_devices(struct spi_master *master)
|
|
|
|
{
|
|
|
|
struct spi_device *spi;
|
|
|
|
struct device_node *nc;
|
|
|
|
|
|
|
|
if (!master->dev.of_node)
|
|
|
|
return;
|
|
|
|
|
2012-11-29 07:59:29 +00:00
|
|
|
for_each_available_child_of_node(master->dev.of_node, nc) {
|
2014-10-29 08:40:37 +00:00
|
|
|
spi = of_register_spi_device(master, nc);
|
|
|
|
if (IS_ERR(spi))
|
|
|
|
dev_warn(&master->dev, "Failed to create SPI device for %s\n",
|
2012-04-07 20:16:53 +00:00
|
|
|
nc->full_name);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
#else
|
|
|
|
static void of_register_spi_devices(struct spi_master *master) { }
|
|
|
|
#endif
|
|
|
|
|
2012-11-30 11:37:53 +00:00
|
|
|
#ifdef CONFIG_ACPI
|
|
|
|
static int acpi_spi_add_resource(struct acpi_resource *ares, void *data)
|
|
|
|
{
|
|
|
|
struct spi_device *spi = data;
|
|
|
|
|
|
|
|
if (ares->type == ACPI_RESOURCE_TYPE_SERIAL_BUS) {
|
|
|
|
struct acpi_resource_spi_serialbus *sb;
|
|
|
|
|
|
|
|
sb = &ares->data.spi_serial_bus;
|
|
|
|
if (sb->type == ACPI_RESOURCE_SERIAL_TYPE_SPI) {
|
|
|
|
spi->chip_select = sb->device_selection;
|
|
|
|
spi->max_speed_hz = sb->connection_speed;
|
|
|
|
|
|
|
|
if (sb->clock_phase == ACPI_SPI_SECOND_PHASE)
|
|
|
|
spi->mode |= SPI_CPHA;
|
|
|
|
if (sb->clock_polarity == ACPI_SPI_START_HIGH)
|
|
|
|
spi->mode |= SPI_CPOL;
|
|
|
|
if (sb->device_polarity == ACPI_SPI_ACTIVE_HIGH)
|
|
|
|
spi->mode |= SPI_CS_HIGH;
|
|
|
|
}
|
|
|
|
} else if (spi->irq < 0) {
|
|
|
|
struct resource r;
|
|
|
|
|
|
|
|
if (acpi_dev_resource_interrupt(ares, 0, &r))
|
|
|
|
spi->irq = r.start;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Always tell the ACPI core to skip this resource */
|
|
|
|
return 1;
|
|
|
|
}
|
|
|
|
|
|
|
|
static acpi_status acpi_spi_add_device(acpi_handle handle, u32 level,
|
|
|
|
void *data, void **return_value)
|
|
|
|
{
|
|
|
|
struct spi_master *master = data;
|
|
|
|
struct list_head resource_list;
|
|
|
|
struct acpi_device *adev;
|
|
|
|
struct spi_device *spi;
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
if (acpi_bus_get_device(handle, &adev))
|
|
|
|
return AE_OK;
|
|
|
|
if (acpi_bus_get_status(adev) || !adev->status.present)
|
|
|
|
return AE_OK;
|
|
|
|
|
|
|
|
spi = spi_alloc_device(master);
|
|
|
|
if (!spi) {
|
|
|
|
dev_err(&master->dev, "failed to allocate SPI device for %s\n",
|
|
|
|
dev_name(&adev->dev));
|
|
|
|
return AE_NO_MEMORY;
|
|
|
|
}
|
|
|
|
|
ACPI / driver core: Store an ACPI device pointer in struct acpi_dev_node
Modify struct acpi_dev_node to contain a pointer to struct acpi_device
associated with the given device object (that is, its ACPI companion
device) instead of an ACPI handle corresponding to it. Introduce two
new macros for manipulating that pointer in a CONFIG_ACPI-safe way,
ACPI_COMPANION() and ACPI_COMPANION_SET(), and rework the
ACPI_HANDLE() macro to take the above changes into account.
Drop the ACPI_HANDLE_SET() macro entirely and rework its users to
use ACPI_COMPANION_SET() instead. For some of them who used to
pass the result of acpi_get_child() directly to ACPI_HANDLE_SET()
introduce a helper routine acpi_preset_companion() doing an
equivalent thing.
The main motivation for doing this is that there are things
represented by struct acpi_device objects that don't have valid
ACPI handles (so called fixed ACPI hardware features, such as
power and sleep buttons) and we would like to create platform
device objects for them and "glue" them to their ACPI companions
in the usual way (which currently is impossible due to the
lack of valid ACPI handles). However, there are more reasons
why it may be useful.
First, struct acpi_device pointers allow of much better type checking
than void pointers which are ACPI handles, so it should be more
difficult to write buggy code using modified struct acpi_dev_node
and the new macros. Second, the change should help to reduce (over
time) the number of places in which the result of ACPI_HANDLE() is
passed to acpi_bus_get_device() in order to obtain a pointer to the
struct acpi_device associated with the given "physical" device,
because now that pointer is returned by ACPI_COMPANION() directly.
Finally, the change should make it easier to write generic code that
will build both for CONFIG_ACPI set and unset without adding explicit
compiler directives to it.
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Acked-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Tested-by: Mika Westerberg <mika.westerberg@linux.intel.com> # on Haswell
Reviewed-by: Mika Westerberg <mika.westerberg@linux.intel.com>
Reviewed-by: Aaron Lu <aaron.lu@intel.com> # for ATA and SDIO part
2013-11-11 21:41:56 +00:00
|
|
|
ACPI_COMPANION_SET(&spi->dev, adev);
|
2012-11-30 11:37:53 +00:00
|
|
|
spi->irq = -1;
|
|
|
|
|
|
|
|
INIT_LIST_HEAD(&resource_list);
|
|
|
|
ret = acpi_dev_get_resources(adev, &resource_list,
|
|
|
|
acpi_spi_add_resource, spi);
|
|
|
|
acpi_dev_free_resource_list(&resource_list);
|
|
|
|
|
|
|
|
if (ret < 0 || !spi->max_speed_hz) {
|
|
|
|
spi_dev_put(spi);
|
|
|
|
return AE_OK;
|
|
|
|
}
|
|
|
|
|
2013-10-10 10:28:48 +00:00
|
|
|
adev->power.flags.ignore_parent = true;
|
2013-10-10 14:19:17 +00:00
|
|
|
strlcpy(spi->modalias, acpi_device_hid(adev), sizeof(spi->modalias));
|
2012-11-30 11:37:53 +00:00
|
|
|
if (spi_add_device(spi)) {
|
2013-10-10 10:28:48 +00:00
|
|
|
adev->power.flags.ignore_parent = false;
|
2012-11-30 11:37:53 +00:00
|
|
|
dev_err(&master->dev, "failed to add SPI device %s from ACPI\n",
|
|
|
|
dev_name(&adev->dev));
|
|
|
|
spi_dev_put(spi);
|
|
|
|
}
|
|
|
|
|
|
|
|
return AE_OK;
|
|
|
|
}
|
|
|
|
|
|
|
|
static void acpi_register_spi_devices(struct spi_master *master)
|
|
|
|
{
|
|
|
|
acpi_status status;
|
|
|
|
acpi_handle handle;
|
|
|
|
|
2013-04-01 00:21:08 +00:00
|
|
|
handle = ACPI_HANDLE(master->dev.parent);
|
2012-11-30 11:37:53 +00:00
|
|
|
if (!handle)
|
|
|
|
return;
|
|
|
|
|
|
|
|
status = acpi_walk_namespace(ACPI_TYPE_DEVICE, handle, 1,
|
|
|
|
acpi_spi_add_device, NULL,
|
|
|
|
master, NULL);
|
|
|
|
if (ACPI_FAILURE(status))
|
|
|
|
dev_warn(&master->dev, "failed to enumerate SPI slaves\n");
|
|
|
|
}
|
|
|
|
#else
|
|
|
|
static inline void acpi_register_spi_devices(struct spi_master *master) {}
|
|
|
|
#endif /* CONFIG_ACPI */
|
|
|
|
|
2007-10-16 08:27:48 +00:00
|
|
|
static void spi_master_release(struct device *dev)
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
{
|
|
|
|
struct spi_master *master;
|
|
|
|
|
2007-10-16 08:27:48 +00:00
|
|
|
master = container_of(dev, struct spi_master, dev);
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
kfree(master);
|
|
|
|
}
|
|
|
|
|
|
|
|
static struct class spi_master_class = {
|
|
|
|
.name = "spi_master",
|
|
|
|
.owner = THIS_MODULE,
|
2007-10-16 08:27:48 +00:00
|
|
|
.dev_release = spi_master_release,
|
2015-06-22 13:00:36 +00:00
|
|
|
.dev_groups = spi_master_groups,
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
};
|
|
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
* spi_alloc_master - allocate SPI master controller
|
|
|
|
* @dev: the controller, possibly using the platform_bus
|
2007-05-08 07:32:21 +00:00
|
|
|
* @size: how much zeroed driver-private data to allocate; the pointer to this
|
2007-10-16 08:27:48 +00:00
|
|
|
* memory is in the driver_data field of the returned device,
|
2006-01-08 21:34:25 +00:00
|
|
|
* accessible with spi_master_get_devdata().
|
2007-05-08 07:32:21 +00:00
|
|
|
* Context: can sleep
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
*
|
|
|
|
* This call is used only by SPI master controller drivers, which are the
|
|
|
|
* only ones directly touching chip registers. It's how they allocate
|
2006-05-20 22:00:14 +00:00
|
|
|
* an spi_master structure, prior to calling spi_register_master().
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
*
|
2015-10-22 16:59:23 +00:00
|
|
|
* This must be called from context that can sleep.
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
*
|
|
|
|
* The caller is responsible for assigning the bus number and initializing
|
2006-05-20 22:00:14 +00:00
|
|
|
* the master's methods before calling spi_register_master(); and (after errors
|
2015-09-05 22:46:54 +00:00
|
|
|
* adding the device) calling spi_master_put() to prevent a memory leak.
|
2015-10-22 16:59:23 +00:00
|
|
|
*
|
|
|
|
* Return: the SPI master structure on success, else NULL.
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
*/
|
2007-03-27 05:32:23 +00:00
|
|
|
struct spi_master *spi_alloc_master(struct device *dev, unsigned size)
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
{
|
|
|
|
struct spi_master *master;
|
|
|
|
|
2006-01-08 21:34:25 +00:00
|
|
|
if (!dev)
|
|
|
|
return NULL;
|
|
|
|
|
2013-10-14 01:31:51 +00:00
|
|
|
master = kzalloc(size + sizeof(*master), GFP_KERNEL);
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
if (!master)
|
|
|
|
return NULL;
|
|
|
|
|
2007-10-16 08:27:48 +00:00
|
|
|
device_initialize(&master->dev);
|
2012-05-20 05:42:08 +00:00
|
|
|
master->bus_num = -1;
|
|
|
|
master->num_chipselect = 1;
|
2007-10-16 08:27:48 +00:00
|
|
|
master->dev.class = &spi_master_class;
|
2015-12-14 15:16:19 +00:00
|
|
|
master->dev.parent = dev;
|
2006-01-08 21:34:25 +00:00
|
|
|
spi_master_set_devdata(master, &master[1]);
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
|
|
|
|
return master;
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(spi_alloc_master);
|
|
|
|
|
2012-11-15 19:19:57 +00:00
|
|
|
#ifdef CONFIG_OF
|
|
|
|
static int of_spi_register_master(struct spi_master *master)
|
|
|
|
{
|
2013-02-12 17:48:37 +00:00
|
|
|
int nb, i, *cs;
|
2012-11-15 19:19:57 +00:00
|
|
|
struct device_node *np = master->dev.of_node;
|
|
|
|
|
|
|
|
if (!np)
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
nb = of_gpio_named_count(np, "cs-gpios");
|
2013-10-14 01:31:51 +00:00
|
|
|
master->num_chipselect = max_t(int, nb, master->num_chipselect);
|
2012-11-15 19:19:57 +00:00
|
|
|
|
2013-02-13 13:23:24 +00:00
|
|
|
/* Return error only for an incorrectly formed cs-gpios property */
|
|
|
|
if (nb == 0 || nb == -ENOENT)
|
2012-11-15 19:19:57 +00:00
|
|
|
return 0;
|
2013-02-13 13:23:24 +00:00
|
|
|
else if (nb < 0)
|
|
|
|
return nb;
|
2012-11-15 19:19:57 +00:00
|
|
|
|
|
|
|
cs = devm_kzalloc(&master->dev,
|
|
|
|
sizeof(int) * master->num_chipselect,
|
|
|
|
GFP_KERNEL);
|
|
|
|
master->cs_gpios = cs;
|
|
|
|
|
|
|
|
if (!master->cs_gpios)
|
|
|
|
return -ENOMEM;
|
|
|
|
|
2013-01-29 14:53:40 +00:00
|
|
|
for (i = 0; i < master->num_chipselect; i++)
|
2013-02-13 13:20:25 +00:00
|
|
|
cs[i] = -ENOENT;
|
2012-11-15 19:19:57 +00:00
|
|
|
|
|
|
|
for (i = 0; i < nb; i++)
|
|
|
|
cs[i] = of_get_named_gpio(np, "cs-gpios", i);
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
#else
|
|
|
|
static int of_spi_register_master(struct spi_master *master)
|
|
|
|
{
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
#endif
|
|
|
|
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
/**
|
|
|
|
* spi_register_master - register SPI master controller
|
|
|
|
* @master: initialized master, originally from spi_alloc_master()
|
2007-05-08 07:32:21 +00:00
|
|
|
* Context: can sleep
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
*
|
|
|
|
* SPI master controllers connect to their drivers using some non-SPI bus,
|
|
|
|
* such as the platform bus. The final stage of probe() in that code
|
|
|
|
* includes calling spi_register_master() to hook up to this SPI bus glue.
|
|
|
|
*
|
|
|
|
* SPI controllers use board specific (often SOC specific) bus numbers,
|
|
|
|
* and board-specific addressing for SPI devices combines those numbers
|
|
|
|
* with chip select numbers. Since SPI does not directly support dynamic
|
|
|
|
* device identification, boards need configuration tables telling which
|
|
|
|
* chip is at which address.
|
|
|
|
*
|
|
|
|
* This must be called from context that can sleep. It returns zero on
|
|
|
|
* success, else a negative error code (dropping the master's refcount).
|
2006-01-08 21:34:25 +00:00
|
|
|
* After a successful return, the caller is responsible for calling
|
|
|
|
* spi_unregister_master().
|
2015-10-22 16:59:23 +00:00
|
|
|
*
|
|
|
|
* Return: zero on success, else a negative error code.
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
*/
|
2007-03-27 05:32:23 +00:00
|
|
|
int spi_register_master(struct spi_master *master)
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
{
|
2007-06-03 20:50:40 +00:00
|
|
|
static atomic_t dyn_bus_id = ATOMIC_INIT((1<<15) - 1);
|
2007-10-16 08:27:48 +00:00
|
|
|
struct device *dev = master->dev.parent;
|
2010-08-02 07:52:15 +00:00
|
|
|
struct boardinfo *bi;
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
int status = -ENODEV;
|
|
|
|
int dynamic = 0;
|
|
|
|
|
2006-01-08 21:34:25 +00:00
|
|
|
if (!dev)
|
|
|
|
return -ENODEV;
|
|
|
|
|
2012-11-15 19:19:57 +00:00
|
|
|
status = of_spi_register_master(master);
|
|
|
|
if (status)
|
|
|
|
return status;
|
|
|
|
|
2007-07-31 07:39:45 +00:00
|
|
|
/* even if it's just one always-selected device, there must
|
|
|
|
* be at least one chipselect
|
|
|
|
*/
|
|
|
|
if (master->num_chipselect == 0)
|
|
|
|
return -EINVAL;
|
|
|
|
|
2012-12-21 19:32:09 +00:00
|
|
|
if ((master->bus_num < 0) && master->dev.of_node)
|
|
|
|
master->bus_num = of_alias_get_id(master->dev.of_node, "spi");
|
|
|
|
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
/* convention: dynamically assigned bus IDs count down from the max */
|
2006-04-03 22:49:04 +00:00
|
|
|
if (master->bus_num < 0) {
|
2007-07-31 07:39:45 +00:00
|
|
|
/* FIXME switch to an IDR based scheme, something like
|
|
|
|
* I2C now uses, so we can't run out of "dynamic" IDs
|
|
|
|
*/
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
master->bus_num = atomic_dec_return(&dyn_bus_id);
|
2006-01-08 21:34:23 +00:00
|
|
|
dynamic = 1;
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
}
|
|
|
|
|
2014-12-10 17:40:53 +00:00
|
|
|
INIT_LIST_HEAD(&master->queue);
|
|
|
|
spin_lock_init(&master->queue_lock);
|
spi/mmc_spi: SPI bus locking API, using mutex
SPI bus locking API to allow exclusive access to the SPI bus, especially, but
not limited to, for the mmc_spi driver.
Coded according to an outline from Grant Likely; here is his
specification (accidentally swapped function names corrected):
It requires 3 things to be added to struct spi_master.
- 1 Mutex
- 1 spin lock
- 1 flag.
The mutex protects spi_sync, and provides sleeping "for free"
The spinlock protects the atomic spi_async call.
The flag is set when the lock is obtained, and checked while holding
the spinlock in spi_async(). If the flag is checked, then spi_async()
must fail immediately.
The current runtime API looks like this:
spi_async(struct spi_device*, struct spi_message*);
spi_sync(struct spi_device*, struct spi_message*);
The API needs to be extended to this:
spi_async(struct spi_device*, struct spi_message*)
spi_sync(struct spi_device*, struct spi_message*)
spi_bus_lock(struct spi_master*) /* although struct spi_device* might
be easier */
spi_bus_unlock(struct spi_master*)
spi_async_locked(struct spi_device*, struct spi_message*)
spi_sync_locked(struct spi_device*, struct spi_message*)
Drivers can only call the last two if they already hold the spi_master_lock().
spi_bus_lock() obtains the mutex, obtains the spin lock, sets the
flag, and releases the spin lock before returning. It doesn't even
need to sleep while waiting for "in-flight" spi_transactions to
complete because its purpose is to guarantee no additional
transactions are added. It does not guarantee that the bus is idle.
spi_bus_unlock() clears the flag and releases the mutex, which will
wake up any waiters.
The difference between spi_async() and spi_async_locked() is that the
locked version bypasses the check of the lock flag. Both versions
need to obtain the spinlock.
The difference between spi_sync() and spi_sync_locked() is that
spi_sync() must hold the mutex while enqueuing a new transfer.
spi_sync_locked() doesn't because the mutex is already held. Note
however that spi_sync must *not* continue to hold the mutex while
waiting for the transfer to complete, otherwise only one transfer
could be queued up at a time!
Almost no code needs to be written. The current spi_async() and
spi_sync() can probably be renamed to __spi_async() and __spi_sync()
so that spi_async(), spi_sync(), spi_async_locked() and
spi_sync_locked() can just become wrappers around the common code.
spi_sync() is protected by a mutex because it can sleep
spi_async() needs to be protected with a flag and a spinlock because
it can be called atomically and must not sleep
Signed-off-by: Ernst Schwab <eschwab@online.de>
[grant.likely@secretlab.ca: use spin_lock_irqsave()]
Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
Tested-by: Matt Fleming <matt@console-pimps.org>
Tested-by: Antonio Ospite <ospite@studenti.unina.it>
2010-06-29 00:49:29 +00:00
|
|
|
spin_lock_init(&master->bus_lock_spinlock);
|
|
|
|
mutex_init(&master->bus_lock_mutex);
|
|
|
|
master->bus_lock_flag = 0;
|
2013-10-05 10:50:40 +00:00
|
|
|
init_completion(&master->xfer_completion);
|
2014-02-02 13:47:47 +00:00
|
|
|
if (!master->max_dma_len)
|
|
|
|
master->max_dma_len = INT_MAX;
|
spi/mmc_spi: SPI bus locking API, using mutex
SPI bus locking API to allow exclusive access to the SPI bus, especially, but
not limited to, for the mmc_spi driver.
Coded according to an outline from Grant Likely; here is his
specification (accidentally swapped function names corrected):
It requires 3 things to be added to struct spi_master.
- 1 Mutex
- 1 spin lock
- 1 flag.
The mutex protects spi_sync, and provides sleeping "for free"
The spinlock protects the atomic spi_async call.
The flag is set when the lock is obtained, and checked while holding
the spinlock in spi_async(). If the flag is checked, then spi_async()
must fail immediately.
The current runtime API looks like this:
spi_async(struct spi_device*, struct spi_message*);
spi_sync(struct spi_device*, struct spi_message*);
The API needs to be extended to this:
spi_async(struct spi_device*, struct spi_message*)
spi_sync(struct spi_device*, struct spi_message*)
spi_bus_lock(struct spi_master*) /* although struct spi_device* might
be easier */
spi_bus_unlock(struct spi_master*)
spi_async_locked(struct spi_device*, struct spi_message*)
spi_sync_locked(struct spi_device*, struct spi_message*)
Drivers can only call the last two if they already hold the spi_master_lock().
spi_bus_lock() obtains the mutex, obtains the spin lock, sets the
flag, and releases the spin lock before returning. It doesn't even
need to sleep while waiting for "in-flight" spi_transactions to
complete because its purpose is to guarantee no additional
transactions are added. It does not guarantee that the bus is idle.
spi_bus_unlock() clears the flag and releases the mutex, which will
wake up any waiters.
The difference between spi_async() and spi_async_locked() is that the
locked version bypasses the check of the lock flag. Both versions
need to obtain the spinlock.
The difference between spi_sync() and spi_sync_locked() is that
spi_sync() must hold the mutex while enqueuing a new transfer.
spi_sync_locked() doesn't because the mutex is already held. Note
however that spi_sync must *not* continue to hold the mutex while
waiting for the transfer to complete, otherwise only one transfer
could be queued up at a time!
Almost no code needs to be written. The current spi_async() and
spi_sync() can probably be renamed to __spi_async() and __spi_sync()
so that spi_async(), spi_sync(), spi_async_locked() and
spi_sync_locked() can just become wrappers around the common code.
spi_sync() is protected by a mutex because it can sleep
spi_async() needs to be protected with a flag and a spinlock because
it can be called atomically and must not sleep
Signed-off-by: Ernst Schwab <eschwab@online.de>
[grant.likely@secretlab.ca: use spin_lock_irqsave()]
Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
Tested-by: Matt Fleming <matt@console-pimps.org>
Tested-by: Antonio Ospite <ospite@studenti.unina.it>
2010-06-29 00:49:29 +00:00
|
|
|
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
/* register the device, then userspace will see it.
|
|
|
|
* registration fails if the bus ID is in use.
|
|
|
|
*/
|
2009-01-06 18:44:37 +00:00
|
|
|
dev_set_name(&master->dev, "spi%u", master->bus_num);
|
2007-10-16 08:27:48 +00:00
|
|
|
status = device_add(&master->dev);
|
2006-01-08 21:34:23 +00:00
|
|
|
if (status < 0)
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
goto done;
|
2009-01-06 18:44:37 +00:00
|
|
|
dev_dbg(dev, "registered master %s%s\n", dev_name(&master->dev),
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
dynamic ? " (dynamic)" : "");
|
|
|
|
|
2012-02-22 09:05:38 +00:00
|
|
|
/* If we're using a queued driver, start the queue */
|
|
|
|
if (master->transfer)
|
|
|
|
dev_info(dev, "master is unqueued, this is deprecated\n");
|
|
|
|
else {
|
|
|
|
status = spi_master_initialize_queue(master);
|
|
|
|
if (status) {
|
2013-08-31 12:25:52 +00:00
|
|
|
device_del(&master->dev);
|
2012-02-22 09:05:38 +00:00
|
|
|
goto done;
|
|
|
|
}
|
|
|
|
}
|
2015-06-22 13:00:36 +00:00
|
|
|
/* add statistics */
|
|
|
|
spin_lock_init(&master->statistics.lock);
|
2012-02-22 09:05:38 +00:00
|
|
|
|
2010-08-02 07:52:15 +00:00
|
|
|
mutex_lock(&board_lock);
|
|
|
|
list_add_tail(&master->list, &spi_master_list);
|
|
|
|
list_for_each_entry(bi, &board_list, list)
|
|
|
|
spi_match_master_to_boardinfo(master, &bi->board_info);
|
|
|
|
mutex_unlock(&board_lock);
|
|
|
|
|
2012-11-30 11:37:53 +00:00
|
|
|
/* Register devices from the device tree and ACPI */
|
2010-07-27 20:35:58 +00:00
|
|
|
of_register_spi_devices(master);
|
2012-11-30 11:37:53 +00:00
|
|
|
acpi_register_spi_devices(master);
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
done:
|
|
|
|
return status;
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(spi_register_master);
|
|
|
|
|
2013-08-31 17:50:52 +00:00
|
|
|
static void devm_spi_unregister(struct device *dev, void *res)
|
|
|
|
{
|
|
|
|
spi_unregister_master(*(struct spi_master **)res);
|
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
|
|
|
* dev_spi_register_master - register managed SPI master controller
|
|
|
|
* @dev: device managing SPI master
|
|
|
|
* @master: initialized master, originally from spi_alloc_master()
|
|
|
|
* Context: can sleep
|
|
|
|
*
|
|
|
|
* Register a SPI device as with spi_register_master() which will
|
|
|
|
* automatically be unregister
|
2015-10-22 16:59:23 +00:00
|
|
|
*
|
|
|
|
* Return: zero on success, else a negative error code.
|
2013-08-31 17:50:52 +00:00
|
|
|
*/
|
|
|
|
int devm_spi_register_master(struct device *dev, struct spi_master *master)
|
|
|
|
{
|
|
|
|
struct spi_master **ptr;
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
ptr = devres_alloc(devm_spi_unregister, sizeof(*ptr), GFP_KERNEL);
|
|
|
|
if (!ptr)
|
|
|
|
return -ENOMEM;
|
|
|
|
|
|
|
|
ret = spi_register_master(master);
|
2013-11-21 23:11:15 +00:00
|
|
|
if (!ret) {
|
2013-08-31 17:50:52 +00:00
|
|
|
*ptr = master;
|
|
|
|
devres_add(dev, ptr);
|
|
|
|
} else {
|
|
|
|
devres_free(ptr);
|
|
|
|
}
|
|
|
|
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(devm_spi_register_master);
|
|
|
|
|
2010-08-30 21:54:17 +00:00
|
|
|
static int __unregister(struct device *dev, void *null)
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
{
|
2010-08-30 21:54:17 +00:00
|
|
|
spi_unregister_device(to_spi_device(dev));
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
|
|
|
* spi_unregister_master - unregister SPI master controller
|
|
|
|
* @master: the master being unregistered
|
2007-05-08 07:32:21 +00:00
|
|
|
* Context: can sleep
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
*
|
|
|
|
* This call is used only by SPI master controller drivers, which are the
|
|
|
|
* only ones directly touching chip registers.
|
|
|
|
*
|
|
|
|
* This must be called from context that can sleep.
|
|
|
|
*/
|
|
|
|
void spi_unregister_master(struct spi_master *master)
|
|
|
|
{
|
2006-12-07 04:35:35 +00:00
|
|
|
int dummy;
|
|
|
|
|
2012-02-22 09:05:38 +00:00
|
|
|
if (master->queued) {
|
|
|
|
if (spi_destroy_queue(master))
|
|
|
|
dev_err(&master->dev, "queue remove failed\n");
|
|
|
|
}
|
|
|
|
|
2010-08-02 07:52:15 +00:00
|
|
|
mutex_lock(&board_lock);
|
|
|
|
list_del(&master->list);
|
|
|
|
mutex_unlock(&board_lock);
|
|
|
|
|
2010-12-22 01:24:31 +00:00
|
|
|
dummy = device_for_each_child(&master->dev, NULL, __unregister);
|
2007-10-16 08:27:48 +00:00
|
|
|
device_unregister(&master->dev);
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(spi_unregister_master);
|
|
|
|
|
2012-02-22 09:05:38 +00:00
|
|
|
int spi_master_suspend(struct spi_master *master)
|
|
|
|
{
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
/* Basically no-ops for non-queued masters */
|
|
|
|
if (!master->queued)
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
ret = spi_stop_queue(master);
|
|
|
|
if (ret)
|
|
|
|
dev_err(&master->dev, "queue stop failed\n");
|
|
|
|
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(spi_master_suspend);
|
|
|
|
|
|
|
|
int spi_master_resume(struct spi_master *master)
|
|
|
|
{
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
if (!master->queued)
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
ret = spi_start_queue(master);
|
|
|
|
if (ret)
|
|
|
|
dev_err(&master->dev, "queue restart failed\n");
|
|
|
|
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(spi_master_resume);
|
|
|
|
|
2013-02-01 19:40:17 +00:00
|
|
|
static int __spi_master_match(struct device *dev, const void *data)
|
2008-01-22 07:14:18 +00:00
|
|
|
{
|
|
|
|
struct spi_master *m;
|
2013-02-01 19:40:17 +00:00
|
|
|
const u16 *bus_num = data;
|
2008-01-22 07:14:18 +00:00
|
|
|
|
|
|
|
m = container_of(dev, struct spi_master, dev);
|
|
|
|
return m->bus_num == *bus_num;
|
|
|
|
}
|
|
|
|
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
/**
|
|
|
|
* spi_busnum_to_master - look up master associated with bus_num
|
|
|
|
* @bus_num: the master's bus number
|
2007-05-08 07:32:21 +00:00
|
|
|
* Context: can sleep
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
*
|
|
|
|
* This call may be used with devices that are registered after
|
|
|
|
* arch init time. It returns a refcounted pointer to the relevant
|
|
|
|
* spi_master (which the caller must release), or NULL if there is
|
|
|
|
* no such master registered.
|
2015-10-22 16:59:23 +00:00
|
|
|
*
|
|
|
|
* Return: the SPI master structure on success, else NULL.
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
*/
|
|
|
|
struct spi_master *spi_busnum_to_master(u16 bus_num)
|
|
|
|
{
|
2007-10-16 08:27:48 +00:00
|
|
|
struct device *dev;
|
2007-01-26 08:56:54 +00:00
|
|
|
struct spi_master *master = NULL;
|
2008-01-22 07:14:18 +00:00
|
|
|
|
2008-05-22 21:21:08 +00:00
|
|
|
dev = class_find_device(&spi_master_class, NULL, &bus_num,
|
2008-01-22 07:14:18 +00:00
|
|
|
__spi_master_match);
|
|
|
|
if (dev)
|
|
|
|
master = container_of(dev, struct spi_master, dev);
|
|
|
|
/* reference got in class_find_device */
|
2007-01-26 08:56:54 +00:00
|
|
|
return master;
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(spi_busnum_to_master);
|
|
|
|
|
|
|
|
|
|
|
|
/*-------------------------------------------------------------------------*/
|
|
|
|
|
2009-06-17 23:26:03 +00:00
|
|
|
/* Core methods for SPI master protocol drivers. Some of the
|
|
|
|
* other core methods are currently defined as inline functions.
|
|
|
|
*/
|
|
|
|
|
2015-08-23 14:06:30 +00:00
|
|
|
static int __spi_validate_bits_per_word(struct spi_master *master, u8 bits_per_word)
|
|
|
|
{
|
|
|
|
if (master->bits_per_word_mask) {
|
|
|
|
/* Only 32 bits fit in the mask */
|
|
|
|
if (bits_per_word > 32)
|
|
|
|
return -EINVAL;
|
|
|
|
if (!(master->bits_per_word_mask &
|
|
|
|
SPI_BPW_MASK(bits_per_word)))
|
|
|
|
return -EINVAL;
|
|
|
|
}
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2009-06-17 23:26:03 +00:00
|
|
|
/**
|
|
|
|
* spi_setup - setup SPI mode and clock rate
|
|
|
|
* @spi: the device whose settings are being modified
|
|
|
|
* Context: can sleep, and no requests are queued to the device
|
|
|
|
*
|
|
|
|
* SPI protocol drivers may need to update the transfer mode if the
|
|
|
|
* device doesn't work with its default. They may likewise need
|
|
|
|
* to update clock rates or word sizes from initial values. This function
|
|
|
|
* changes those settings, and must be called from a context that can sleep.
|
|
|
|
* Except for SPI_CS_HIGH, which takes effect immediately, the changes take
|
|
|
|
* effect the next time the device is selected and data is transferred to
|
|
|
|
* or from it. When this function returns, the spi device is deselected.
|
|
|
|
*
|
|
|
|
* Note that this call will fail if the protocol driver specifies an option
|
|
|
|
* that the underlying controller or its driver does not support. For
|
|
|
|
* example, not all hardware supports wire transfers using nine bit words,
|
|
|
|
* LSB-first wire encoding, or active-high chipselects.
|
2015-10-22 16:59:23 +00:00
|
|
|
*
|
|
|
|
* Return: zero on success, else a negative error code.
|
2009-06-17 23:26:03 +00:00
|
|
|
*/
|
|
|
|
int spi_setup(struct spi_device *spi)
|
|
|
|
{
|
2014-04-14 17:39:53 +00:00
|
|
|
unsigned bad_bits, ugly_bits;
|
2015-10-14 19:43:07 +00:00
|
|
|
int status;
|
2009-06-17 23:26:03 +00:00
|
|
|
|
2013-08-11 10:15:17 +00:00
|
|
|
/* check mode to prevent that DUAL and QUAD set at the same time
|
|
|
|
*/
|
|
|
|
if (((spi->mode & SPI_TX_DUAL) && (spi->mode & SPI_TX_QUAD)) ||
|
|
|
|
((spi->mode & SPI_RX_DUAL) && (spi->mode & SPI_RX_QUAD))) {
|
|
|
|
dev_err(&spi->dev,
|
|
|
|
"setup: can not select dual and quad at the same time\n");
|
|
|
|
return -EINVAL;
|
|
|
|
}
|
|
|
|
/* if it is SPI_3WIRE mode, DUAL and QUAD should be forbidden
|
|
|
|
*/
|
|
|
|
if ((spi->mode & SPI_3WIRE) && (spi->mode &
|
|
|
|
(SPI_TX_DUAL | SPI_TX_QUAD | SPI_RX_DUAL | SPI_RX_QUAD)))
|
|
|
|
return -EINVAL;
|
2009-06-17 23:26:04 +00:00
|
|
|
/* help drivers fail *cleanly* when they need options
|
|
|
|
* that aren't supported with their current master
|
|
|
|
*/
|
|
|
|
bad_bits = spi->mode & ~spi->master->mode_bits;
|
2014-04-14 17:39:53 +00:00
|
|
|
ugly_bits = bad_bits &
|
|
|
|
(SPI_TX_DUAL | SPI_TX_QUAD | SPI_RX_DUAL | SPI_RX_QUAD);
|
|
|
|
if (ugly_bits) {
|
|
|
|
dev_warn(&spi->dev,
|
|
|
|
"setup: ignoring unsupported mode bits %x\n",
|
|
|
|
ugly_bits);
|
|
|
|
spi->mode &= ~ugly_bits;
|
|
|
|
bad_bits &= ~ugly_bits;
|
|
|
|
}
|
2009-06-17 23:26:04 +00:00
|
|
|
if (bad_bits) {
|
2010-10-21 19:06:44 +00:00
|
|
|
dev_err(&spi->dev, "setup: unsupported mode bits %x\n",
|
2009-06-17 23:26:04 +00:00
|
|
|
bad_bits);
|
|
|
|
return -EINVAL;
|
|
|
|
}
|
|
|
|
|
2009-06-17 23:26:03 +00:00
|
|
|
if (!spi->bits_per_word)
|
|
|
|
spi->bits_per_word = 8;
|
|
|
|
|
2015-10-14 19:43:07 +00:00
|
|
|
status = __spi_validate_bits_per_word(spi->master, spi->bits_per_word);
|
|
|
|
if (status)
|
|
|
|
return status;
|
2015-08-23 14:06:30 +00:00
|
|
|
|
2014-02-09 16:08:05 +00:00
|
|
|
if (!spi->max_speed_hz)
|
|
|
|
spi->max_speed_hz = spi->master->max_speed_hz;
|
|
|
|
|
2012-11-09 09:05:22 +00:00
|
|
|
if (spi->master->setup)
|
|
|
|
status = spi->master->setup(spi);
|
2009-06-17 23:26:03 +00:00
|
|
|
|
2015-10-16 15:29:03 +00:00
|
|
|
spi_set_cs(spi, false);
|
|
|
|
|
2013-10-14 01:31:51 +00:00
|
|
|
dev_dbg(&spi->dev, "setup mode %d, %s%s%s%s%u bits/w, %u Hz max --> %d\n",
|
2009-06-17 23:26:03 +00:00
|
|
|
(int) (spi->mode & (SPI_CPOL | SPI_CPHA)),
|
|
|
|
(spi->mode & SPI_CS_HIGH) ? "cs_high, " : "",
|
|
|
|
(spi->mode & SPI_LSB_FIRST) ? "lsb, " : "",
|
|
|
|
(spi->mode & SPI_3WIRE) ? "3wire, " : "",
|
|
|
|
(spi->mode & SPI_LOOP) ? "loopback, " : "",
|
|
|
|
spi->bits_per_word, spi->max_speed_hz,
|
|
|
|
status);
|
|
|
|
|
|
|
|
return status;
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(spi_setup);
|
|
|
|
|
2013-11-13 23:44:15 +00:00
|
|
|
static int __spi_validate(struct spi_device *spi, struct spi_message *message)
|
spi/mmc_spi: SPI bus locking API, using mutex
SPI bus locking API to allow exclusive access to the SPI bus, especially, but
not limited to, for the mmc_spi driver.
Coded according to an outline from Grant Likely; here is his
specification (accidentally swapped function names corrected):
It requires 3 things to be added to struct spi_master.
- 1 Mutex
- 1 spin lock
- 1 flag.
The mutex protects spi_sync, and provides sleeping "for free"
The spinlock protects the atomic spi_async call.
The flag is set when the lock is obtained, and checked while holding
the spinlock in spi_async(). If the flag is checked, then spi_async()
must fail immediately.
The current runtime API looks like this:
spi_async(struct spi_device*, struct spi_message*);
spi_sync(struct spi_device*, struct spi_message*);
The API needs to be extended to this:
spi_async(struct spi_device*, struct spi_message*)
spi_sync(struct spi_device*, struct spi_message*)
spi_bus_lock(struct spi_master*) /* although struct spi_device* might
be easier */
spi_bus_unlock(struct spi_master*)
spi_async_locked(struct spi_device*, struct spi_message*)
spi_sync_locked(struct spi_device*, struct spi_message*)
Drivers can only call the last two if they already hold the spi_master_lock().
spi_bus_lock() obtains the mutex, obtains the spin lock, sets the
flag, and releases the spin lock before returning. It doesn't even
need to sleep while waiting for "in-flight" spi_transactions to
complete because its purpose is to guarantee no additional
transactions are added. It does not guarantee that the bus is idle.
spi_bus_unlock() clears the flag and releases the mutex, which will
wake up any waiters.
The difference between spi_async() and spi_async_locked() is that the
locked version bypasses the check of the lock flag. Both versions
need to obtain the spinlock.
The difference between spi_sync() and spi_sync_locked() is that
spi_sync() must hold the mutex while enqueuing a new transfer.
spi_sync_locked() doesn't because the mutex is already held. Note
however that spi_sync must *not* continue to hold the mutex while
waiting for the transfer to complete, otherwise only one transfer
could be queued up at a time!
Almost no code needs to be written. The current spi_async() and
spi_sync() can probably be renamed to __spi_async() and __spi_sync()
so that spi_async(), spi_sync(), spi_async_locked() and
spi_sync_locked() can just become wrappers around the common code.
spi_sync() is protected by a mutex because it can sleep
spi_async() needs to be protected with a flag and a spinlock because
it can be called atomically and must not sleep
Signed-off-by: Ernst Schwab <eschwab@online.de>
[grant.likely@secretlab.ca: use spin_lock_irqsave()]
Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
Tested-by: Matt Fleming <matt@console-pimps.org>
Tested-by: Antonio Ospite <ospite@studenti.unina.it>
2010-06-29 00:49:29 +00:00
|
|
|
{
|
|
|
|
struct spi_master *master = spi->master;
|
2012-11-09 09:06:45 +00:00
|
|
|
struct spi_transfer *xfer;
|
2014-02-28 14:03:16 +00:00
|
|
|
int w_size;
|
spi/mmc_spi: SPI bus locking API, using mutex
SPI bus locking API to allow exclusive access to the SPI bus, especially, but
not limited to, for the mmc_spi driver.
Coded according to an outline from Grant Likely; here is his
specification (accidentally swapped function names corrected):
It requires 3 things to be added to struct spi_master.
- 1 Mutex
- 1 spin lock
- 1 flag.
The mutex protects spi_sync, and provides sleeping "for free"
The spinlock protects the atomic spi_async call.
The flag is set when the lock is obtained, and checked while holding
the spinlock in spi_async(). If the flag is checked, then spi_async()
must fail immediately.
The current runtime API looks like this:
spi_async(struct spi_device*, struct spi_message*);
spi_sync(struct spi_device*, struct spi_message*);
The API needs to be extended to this:
spi_async(struct spi_device*, struct spi_message*)
spi_sync(struct spi_device*, struct spi_message*)
spi_bus_lock(struct spi_master*) /* although struct spi_device* might
be easier */
spi_bus_unlock(struct spi_master*)
spi_async_locked(struct spi_device*, struct spi_message*)
spi_sync_locked(struct spi_device*, struct spi_message*)
Drivers can only call the last two if they already hold the spi_master_lock().
spi_bus_lock() obtains the mutex, obtains the spin lock, sets the
flag, and releases the spin lock before returning. It doesn't even
need to sleep while waiting for "in-flight" spi_transactions to
complete because its purpose is to guarantee no additional
transactions are added. It does not guarantee that the bus is idle.
spi_bus_unlock() clears the flag and releases the mutex, which will
wake up any waiters.
The difference between spi_async() and spi_async_locked() is that the
locked version bypasses the check of the lock flag. Both versions
need to obtain the spinlock.
The difference between spi_sync() and spi_sync_locked() is that
spi_sync() must hold the mutex while enqueuing a new transfer.
spi_sync_locked() doesn't because the mutex is already held. Note
however that spi_sync must *not* continue to hold the mutex while
waiting for the transfer to complete, otherwise only one transfer
could be queued up at a time!
Almost no code needs to be written. The current spi_async() and
spi_sync() can probably be renamed to __spi_async() and __spi_sync()
so that spi_async(), spi_sync(), spi_async_locked() and
spi_sync_locked() can just become wrappers around the common code.
spi_sync() is protected by a mutex because it can sleep
spi_async() needs to be protected with a flag and a spinlock because
it can be called atomically and must not sleep
Signed-off-by: Ernst Schwab <eschwab@online.de>
[grant.likely@secretlab.ca: use spin_lock_irqsave()]
Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
Tested-by: Matt Fleming <matt@console-pimps.org>
Tested-by: Antonio Ospite <ospite@studenti.unina.it>
2010-06-29 00:49:29 +00:00
|
|
|
|
2013-07-10 14:05:40 +00:00
|
|
|
if (list_empty(&message->transfers))
|
|
|
|
return -EINVAL;
|
|
|
|
|
spi/mmc_spi: SPI bus locking API, using mutex
SPI bus locking API to allow exclusive access to the SPI bus, especially, but
not limited to, for the mmc_spi driver.
Coded according to an outline from Grant Likely; here is his
specification (accidentally swapped function names corrected):
It requires 3 things to be added to struct spi_master.
- 1 Mutex
- 1 spin lock
- 1 flag.
The mutex protects spi_sync, and provides sleeping "for free"
The spinlock protects the atomic spi_async call.
The flag is set when the lock is obtained, and checked while holding
the spinlock in spi_async(). If the flag is checked, then spi_async()
must fail immediately.
The current runtime API looks like this:
spi_async(struct spi_device*, struct spi_message*);
spi_sync(struct spi_device*, struct spi_message*);
The API needs to be extended to this:
spi_async(struct spi_device*, struct spi_message*)
spi_sync(struct spi_device*, struct spi_message*)
spi_bus_lock(struct spi_master*) /* although struct spi_device* might
be easier */
spi_bus_unlock(struct spi_master*)
spi_async_locked(struct spi_device*, struct spi_message*)
spi_sync_locked(struct spi_device*, struct spi_message*)
Drivers can only call the last two if they already hold the spi_master_lock().
spi_bus_lock() obtains the mutex, obtains the spin lock, sets the
flag, and releases the spin lock before returning. It doesn't even
need to sleep while waiting for "in-flight" spi_transactions to
complete because its purpose is to guarantee no additional
transactions are added. It does not guarantee that the bus is idle.
spi_bus_unlock() clears the flag and releases the mutex, which will
wake up any waiters.
The difference between spi_async() and spi_async_locked() is that the
locked version bypasses the check of the lock flag. Both versions
need to obtain the spinlock.
The difference between spi_sync() and spi_sync_locked() is that
spi_sync() must hold the mutex while enqueuing a new transfer.
spi_sync_locked() doesn't because the mutex is already held. Note
however that spi_sync must *not* continue to hold the mutex while
waiting for the transfer to complete, otherwise only one transfer
could be queued up at a time!
Almost no code needs to be written. The current spi_async() and
spi_sync() can probably be renamed to __spi_async() and __spi_sync()
so that spi_async(), spi_sync(), spi_async_locked() and
spi_sync_locked() can just become wrappers around the common code.
spi_sync() is protected by a mutex because it can sleep
spi_async() needs to be protected with a flag and a spinlock because
it can be called atomically and must not sleep
Signed-off-by: Ernst Schwab <eschwab@online.de>
[grant.likely@secretlab.ca: use spin_lock_irqsave()]
Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
Tested-by: Matt Fleming <matt@console-pimps.org>
Tested-by: Antonio Ospite <ospite@studenti.unina.it>
2010-06-29 00:49:29 +00:00
|
|
|
/* Half-duplex links include original MicroWire, and ones with
|
|
|
|
* only one data pin like SPI_3WIRE (switches direction) or where
|
|
|
|
* either MOSI or MISO is missing. They can also be caused by
|
|
|
|
* software limitations.
|
|
|
|
*/
|
|
|
|
if ((master->flags & SPI_MASTER_HALF_DUPLEX)
|
|
|
|
|| (spi->mode & SPI_3WIRE)) {
|
|
|
|
unsigned flags = master->flags;
|
|
|
|
|
|
|
|
list_for_each_entry(xfer, &message->transfers, transfer_list) {
|
|
|
|
if (xfer->rx_buf && xfer->tx_buf)
|
|
|
|
return -EINVAL;
|
|
|
|
if ((flags & SPI_MASTER_NO_TX) && xfer->tx_buf)
|
|
|
|
return -EINVAL;
|
|
|
|
if ((flags & SPI_MASTER_NO_RX) && xfer->rx_buf)
|
|
|
|
return -EINVAL;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2012-11-09 09:06:45 +00:00
|
|
|
/**
|
2013-01-04 18:47:14 +00:00
|
|
|
* Set transfer bits_per_word and max speed as spi device default if
|
|
|
|
* it is not set for this transfer.
|
2013-08-11 10:15:17 +00:00
|
|
|
* Set transfer tx_nbits and rx_nbits as single transfer default
|
|
|
|
* (SPI_NBITS_SINGLE) if it is not set for this transfer.
|
2012-11-09 09:06:45 +00:00
|
|
|
*/
|
2015-11-27 12:31:09 +00:00
|
|
|
message->frame_length = 0;
|
2012-11-09 09:06:45 +00:00
|
|
|
list_for_each_entry(xfer, &message->transfers, transfer_list) {
|
2013-07-18 10:01:25 +00:00
|
|
|
message->frame_length += xfer->len;
|
2012-11-09 09:06:45 +00:00
|
|
|
if (!xfer->bits_per_word)
|
|
|
|
xfer->bits_per_word = spi->bits_per_word;
|
2014-03-17 02:08:12 +00:00
|
|
|
|
|
|
|
if (!xfer->speed_hz)
|
2013-01-04 18:47:14 +00:00
|
|
|
xfer->speed_hz = spi->max_speed_hz;
|
2015-08-20 18:52:18 +00:00
|
|
|
if (!xfer->speed_hz)
|
|
|
|
xfer->speed_hz = master->max_speed_hz;
|
2014-03-17 02:08:12 +00:00
|
|
|
|
|
|
|
if (master->max_speed_hz &&
|
|
|
|
xfer->speed_hz > master->max_speed_hz)
|
|
|
|
xfer->speed_hz = master->max_speed_hz;
|
2013-08-14 08:25:28 +00:00
|
|
|
|
2015-08-23 14:06:30 +00:00
|
|
|
if (__spi_validate_bits_per_word(master, xfer->bits_per_word))
|
|
|
|
return -EINVAL;
|
2013-07-10 13:57:26 +00:00
|
|
|
|
2014-02-20 10:02:08 +00:00
|
|
|
/*
|
|
|
|
* SPI transfer length should be multiple of SPI word size
|
|
|
|
* where SPI word size should be power-of-two multiple
|
|
|
|
*/
|
|
|
|
if (xfer->bits_per_word <= 8)
|
|
|
|
w_size = 1;
|
|
|
|
else if (xfer->bits_per_word <= 16)
|
|
|
|
w_size = 2;
|
|
|
|
else
|
|
|
|
w_size = 4;
|
|
|
|
|
|
|
|
/* No partial transfers accepted */
|
2014-02-28 14:03:16 +00:00
|
|
|
if (xfer->len % w_size)
|
2014-02-20 10:02:08 +00:00
|
|
|
return -EINVAL;
|
|
|
|
|
2013-07-10 13:57:26 +00:00
|
|
|
if (xfer->speed_hz && master->min_speed_hz &&
|
|
|
|
xfer->speed_hz < master->min_speed_hz)
|
|
|
|
return -EINVAL;
|
2013-08-11 10:15:17 +00:00
|
|
|
|
|
|
|
if (xfer->tx_buf && !xfer->tx_nbits)
|
|
|
|
xfer->tx_nbits = SPI_NBITS_SINGLE;
|
|
|
|
if (xfer->rx_buf && !xfer->rx_nbits)
|
|
|
|
xfer->rx_nbits = SPI_NBITS_SINGLE;
|
|
|
|
/* check transfer tx/rx_nbits:
|
2014-01-12 13:00:29 +00:00
|
|
|
* 1. check the value matches one of single, dual and quad
|
|
|
|
* 2. check tx/rx_nbits match the mode in spi_device
|
2013-08-11 10:15:17 +00:00
|
|
|
*/
|
2013-08-22 15:50:48 +00:00
|
|
|
if (xfer->tx_buf) {
|
|
|
|
if (xfer->tx_nbits != SPI_NBITS_SINGLE &&
|
|
|
|
xfer->tx_nbits != SPI_NBITS_DUAL &&
|
|
|
|
xfer->tx_nbits != SPI_NBITS_QUAD)
|
|
|
|
return -EINVAL;
|
|
|
|
if ((xfer->tx_nbits == SPI_NBITS_DUAL) &&
|
|
|
|
!(spi->mode & (SPI_TX_DUAL | SPI_TX_QUAD)))
|
|
|
|
return -EINVAL;
|
|
|
|
if ((xfer->tx_nbits == SPI_NBITS_QUAD) &&
|
|
|
|
!(spi->mode & SPI_TX_QUAD))
|
|
|
|
return -EINVAL;
|
|
|
|
}
|
2013-08-11 10:15:17 +00:00
|
|
|
/* check transfer rx_nbits */
|
2013-08-22 15:50:48 +00:00
|
|
|
if (xfer->rx_buf) {
|
|
|
|
if (xfer->rx_nbits != SPI_NBITS_SINGLE &&
|
|
|
|
xfer->rx_nbits != SPI_NBITS_DUAL &&
|
|
|
|
xfer->rx_nbits != SPI_NBITS_QUAD)
|
|
|
|
return -EINVAL;
|
|
|
|
if ((xfer->rx_nbits == SPI_NBITS_DUAL) &&
|
|
|
|
!(spi->mode & (SPI_RX_DUAL | SPI_RX_QUAD)))
|
|
|
|
return -EINVAL;
|
|
|
|
if ((xfer->rx_nbits == SPI_NBITS_QUAD) &&
|
|
|
|
!(spi->mode & SPI_RX_QUAD))
|
|
|
|
return -EINVAL;
|
|
|
|
}
|
2012-11-09 09:06:45 +00:00
|
|
|
}
|
|
|
|
|
spi/mmc_spi: SPI bus locking API, using mutex
SPI bus locking API to allow exclusive access to the SPI bus, especially, but
not limited to, for the mmc_spi driver.
Coded according to an outline from Grant Likely; here is his
specification (accidentally swapped function names corrected):
It requires 3 things to be added to struct spi_master.
- 1 Mutex
- 1 spin lock
- 1 flag.
The mutex protects spi_sync, and provides sleeping "for free"
The spinlock protects the atomic spi_async call.
The flag is set when the lock is obtained, and checked while holding
the spinlock in spi_async(). If the flag is checked, then spi_async()
must fail immediately.
The current runtime API looks like this:
spi_async(struct spi_device*, struct spi_message*);
spi_sync(struct spi_device*, struct spi_message*);
The API needs to be extended to this:
spi_async(struct spi_device*, struct spi_message*)
spi_sync(struct spi_device*, struct spi_message*)
spi_bus_lock(struct spi_master*) /* although struct spi_device* might
be easier */
spi_bus_unlock(struct spi_master*)
spi_async_locked(struct spi_device*, struct spi_message*)
spi_sync_locked(struct spi_device*, struct spi_message*)
Drivers can only call the last two if they already hold the spi_master_lock().
spi_bus_lock() obtains the mutex, obtains the spin lock, sets the
flag, and releases the spin lock before returning. It doesn't even
need to sleep while waiting for "in-flight" spi_transactions to
complete because its purpose is to guarantee no additional
transactions are added. It does not guarantee that the bus is idle.
spi_bus_unlock() clears the flag and releases the mutex, which will
wake up any waiters.
The difference between spi_async() and spi_async_locked() is that the
locked version bypasses the check of the lock flag. Both versions
need to obtain the spinlock.
The difference between spi_sync() and spi_sync_locked() is that
spi_sync() must hold the mutex while enqueuing a new transfer.
spi_sync_locked() doesn't because the mutex is already held. Note
however that spi_sync must *not* continue to hold the mutex while
waiting for the transfer to complete, otherwise only one transfer
could be queued up at a time!
Almost no code needs to be written. The current spi_async() and
spi_sync() can probably be renamed to __spi_async() and __spi_sync()
so that spi_async(), spi_sync(), spi_async_locked() and
spi_sync_locked() can just become wrappers around the common code.
spi_sync() is protected by a mutex because it can sleep
spi_async() needs to be protected with a flag and a spinlock because
it can be called atomically and must not sleep
Signed-off-by: Ernst Schwab <eschwab@online.de>
[grant.likely@secretlab.ca: use spin_lock_irqsave()]
Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
Tested-by: Matt Fleming <matt@console-pimps.org>
Tested-by: Antonio Ospite <ospite@studenti.unina.it>
2010-06-29 00:49:29 +00:00
|
|
|
message->status = -EINPROGRESS;
|
2013-11-13 23:44:15 +00:00
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int __spi_async(struct spi_device *spi, struct spi_message *message)
|
|
|
|
{
|
|
|
|
struct spi_master *master = spi->master;
|
|
|
|
|
|
|
|
message->spi = spi;
|
|
|
|
|
2015-06-22 13:00:36 +00:00
|
|
|
SPI_STATISTICS_INCREMENT_FIELD(&master->statistics, spi_async);
|
|
|
|
SPI_STATISTICS_INCREMENT_FIELD(&spi->statistics, spi_async);
|
|
|
|
|
2013-11-13 23:44:15 +00:00
|
|
|
trace_spi_message_submit(message);
|
|
|
|
|
spi/mmc_spi: SPI bus locking API, using mutex
SPI bus locking API to allow exclusive access to the SPI bus, especially, but
not limited to, for the mmc_spi driver.
Coded according to an outline from Grant Likely; here is his
specification (accidentally swapped function names corrected):
It requires 3 things to be added to struct spi_master.
- 1 Mutex
- 1 spin lock
- 1 flag.
The mutex protects spi_sync, and provides sleeping "for free"
The spinlock protects the atomic spi_async call.
The flag is set when the lock is obtained, and checked while holding
the spinlock in spi_async(). If the flag is checked, then spi_async()
must fail immediately.
The current runtime API looks like this:
spi_async(struct spi_device*, struct spi_message*);
spi_sync(struct spi_device*, struct spi_message*);
The API needs to be extended to this:
spi_async(struct spi_device*, struct spi_message*)
spi_sync(struct spi_device*, struct spi_message*)
spi_bus_lock(struct spi_master*) /* although struct spi_device* might
be easier */
spi_bus_unlock(struct spi_master*)
spi_async_locked(struct spi_device*, struct spi_message*)
spi_sync_locked(struct spi_device*, struct spi_message*)
Drivers can only call the last two if they already hold the spi_master_lock().
spi_bus_lock() obtains the mutex, obtains the spin lock, sets the
flag, and releases the spin lock before returning. It doesn't even
need to sleep while waiting for "in-flight" spi_transactions to
complete because its purpose is to guarantee no additional
transactions are added. It does not guarantee that the bus is idle.
spi_bus_unlock() clears the flag and releases the mutex, which will
wake up any waiters.
The difference between spi_async() and spi_async_locked() is that the
locked version bypasses the check of the lock flag. Both versions
need to obtain the spinlock.
The difference between spi_sync() and spi_sync_locked() is that
spi_sync() must hold the mutex while enqueuing a new transfer.
spi_sync_locked() doesn't because the mutex is already held. Note
however that spi_sync must *not* continue to hold the mutex while
waiting for the transfer to complete, otherwise only one transfer
could be queued up at a time!
Almost no code needs to be written. The current spi_async() and
spi_sync() can probably be renamed to __spi_async() and __spi_sync()
so that spi_async(), spi_sync(), spi_async_locked() and
spi_sync_locked() can just become wrappers around the common code.
spi_sync() is protected by a mutex because it can sleep
spi_async() needs to be protected with a flag and a spinlock because
it can be called atomically and must not sleep
Signed-off-by: Ernst Schwab <eschwab@online.de>
[grant.likely@secretlab.ca: use spin_lock_irqsave()]
Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
Tested-by: Matt Fleming <matt@console-pimps.org>
Tested-by: Antonio Ospite <ospite@studenti.unina.it>
2010-06-29 00:49:29 +00:00
|
|
|
return master->transfer(spi, message);
|
|
|
|
}
|
|
|
|
|
2009-09-22 23:46:18 +00:00
|
|
|
/**
|
|
|
|
* spi_async - asynchronous SPI transfer
|
|
|
|
* @spi: device with which data will be exchanged
|
|
|
|
* @message: describes the data transfers, including completion callback
|
|
|
|
* Context: any (irqs may be blocked, etc)
|
|
|
|
*
|
|
|
|
* This call may be used in_irq and other contexts which can't sleep,
|
|
|
|
* as well as from task contexts which can sleep.
|
|
|
|
*
|
|
|
|
* The completion callback is invoked in a context which can't sleep.
|
|
|
|
* Before that invocation, the value of message->status is undefined.
|
|
|
|
* When the callback is issued, message->status holds either zero (to
|
|
|
|
* indicate complete success) or a negative error code. After that
|
|
|
|
* callback returns, the driver which issued the transfer request may
|
|
|
|
* deallocate the associated memory; it's no longer in use by any SPI
|
|
|
|
* core or controller driver code.
|
|
|
|
*
|
|
|
|
* Note that although all messages to a spi_device are handled in
|
|
|
|
* FIFO order, messages may go to different devices in other orders.
|
|
|
|
* Some device might be higher priority, or have various "hard" access
|
|
|
|
* time requirements, for example.
|
|
|
|
*
|
|
|
|
* On detection of any fault during the transfer, processing of
|
|
|
|
* the entire message is aborted, and the device is deselected.
|
|
|
|
* Until returning from the associated message completion callback,
|
|
|
|
* no other spi_message queued to that device will be processed.
|
|
|
|
* (This rule applies equally to all the synchronous transfer calls,
|
|
|
|
* which are wrappers around this core asynchronous primitive.)
|
2015-10-22 16:59:23 +00:00
|
|
|
*
|
|
|
|
* Return: zero on success, else a negative error code.
|
2009-09-22 23:46:18 +00:00
|
|
|
*/
|
|
|
|
int spi_async(struct spi_device *spi, struct spi_message *message)
|
|
|
|
{
|
|
|
|
struct spi_master *master = spi->master;
|
spi/mmc_spi: SPI bus locking API, using mutex
SPI bus locking API to allow exclusive access to the SPI bus, especially, but
not limited to, for the mmc_spi driver.
Coded according to an outline from Grant Likely; here is his
specification (accidentally swapped function names corrected):
It requires 3 things to be added to struct spi_master.
- 1 Mutex
- 1 spin lock
- 1 flag.
The mutex protects spi_sync, and provides sleeping "for free"
The spinlock protects the atomic spi_async call.
The flag is set when the lock is obtained, and checked while holding
the spinlock in spi_async(). If the flag is checked, then spi_async()
must fail immediately.
The current runtime API looks like this:
spi_async(struct spi_device*, struct spi_message*);
spi_sync(struct spi_device*, struct spi_message*);
The API needs to be extended to this:
spi_async(struct spi_device*, struct spi_message*)
spi_sync(struct spi_device*, struct spi_message*)
spi_bus_lock(struct spi_master*) /* although struct spi_device* might
be easier */
spi_bus_unlock(struct spi_master*)
spi_async_locked(struct spi_device*, struct spi_message*)
spi_sync_locked(struct spi_device*, struct spi_message*)
Drivers can only call the last two if they already hold the spi_master_lock().
spi_bus_lock() obtains the mutex, obtains the spin lock, sets the
flag, and releases the spin lock before returning. It doesn't even
need to sleep while waiting for "in-flight" spi_transactions to
complete because its purpose is to guarantee no additional
transactions are added. It does not guarantee that the bus is idle.
spi_bus_unlock() clears the flag and releases the mutex, which will
wake up any waiters.
The difference between spi_async() and spi_async_locked() is that the
locked version bypasses the check of the lock flag. Both versions
need to obtain the spinlock.
The difference between spi_sync() and spi_sync_locked() is that
spi_sync() must hold the mutex while enqueuing a new transfer.
spi_sync_locked() doesn't because the mutex is already held. Note
however that spi_sync must *not* continue to hold the mutex while
waiting for the transfer to complete, otherwise only one transfer
could be queued up at a time!
Almost no code needs to be written. The current spi_async() and
spi_sync() can probably be renamed to __spi_async() and __spi_sync()
so that spi_async(), spi_sync(), spi_async_locked() and
spi_sync_locked() can just become wrappers around the common code.
spi_sync() is protected by a mutex because it can sleep
spi_async() needs to be protected with a flag and a spinlock because
it can be called atomically and must not sleep
Signed-off-by: Ernst Schwab <eschwab@online.de>
[grant.likely@secretlab.ca: use spin_lock_irqsave()]
Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
Tested-by: Matt Fleming <matt@console-pimps.org>
Tested-by: Antonio Ospite <ospite@studenti.unina.it>
2010-06-29 00:49:29 +00:00
|
|
|
int ret;
|
|
|
|
unsigned long flags;
|
2009-09-22 23:46:18 +00:00
|
|
|
|
2013-11-13 23:44:15 +00:00
|
|
|
ret = __spi_validate(spi, message);
|
|
|
|
if (ret != 0)
|
|
|
|
return ret;
|
|
|
|
|
spi/mmc_spi: SPI bus locking API, using mutex
SPI bus locking API to allow exclusive access to the SPI bus, especially, but
not limited to, for the mmc_spi driver.
Coded according to an outline from Grant Likely; here is his
specification (accidentally swapped function names corrected):
It requires 3 things to be added to struct spi_master.
- 1 Mutex
- 1 spin lock
- 1 flag.
The mutex protects spi_sync, and provides sleeping "for free"
The spinlock protects the atomic spi_async call.
The flag is set when the lock is obtained, and checked while holding
the spinlock in spi_async(). If the flag is checked, then spi_async()
must fail immediately.
The current runtime API looks like this:
spi_async(struct spi_device*, struct spi_message*);
spi_sync(struct spi_device*, struct spi_message*);
The API needs to be extended to this:
spi_async(struct spi_device*, struct spi_message*)
spi_sync(struct spi_device*, struct spi_message*)
spi_bus_lock(struct spi_master*) /* although struct spi_device* might
be easier */
spi_bus_unlock(struct spi_master*)
spi_async_locked(struct spi_device*, struct spi_message*)
spi_sync_locked(struct spi_device*, struct spi_message*)
Drivers can only call the last two if they already hold the spi_master_lock().
spi_bus_lock() obtains the mutex, obtains the spin lock, sets the
flag, and releases the spin lock before returning. It doesn't even
need to sleep while waiting for "in-flight" spi_transactions to
complete because its purpose is to guarantee no additional
transactions are added. It does not guarantee that the bus is idle.
spi_bus_unlock() clears the flag and releases the mutex, which will
wake up any waiters.
The difference between spi_async() and spi_async_locked() is that the
locked version bypasses the check of the lock flag. Both versions
need to obtain the spinlock.
The difference between spi_sync() and spi_sync_locked() is that
spi_sync() must hold the mutex while enqueuing a new transfer.
spi_sync_locked() doesn't because the mutex is already held. Note
however that spi_sync must *not* continue to hold the mutex while
waiting for the transfer to complete, otherwise only one transfer
could be queued up at a time!
Almost no code needs to be written. The current spi_async() and
spi_sync() can probably be renamed to __spi_async() and __spi_sync()
so that spi_async(), spi_sync(), spi_async_locked() and
spi_sync_locked() can just become wrappers around the common code.
spi_sync() is protected by a mutex because it can sleep
spi_async() needs to be protected with a flag and a spinlock because
it can be called atomically and must not sleep
Signed-off-by: Ernst Schwab <eschwab@online.de>
[grant.likely@secretlab.ca: use spin_lock_irqsave()]
Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
Tested-by: Matt Fleming <matt@console-pimps.org>
Tested-by: Antonio Ospite <ospite@studenti.unina.it>
2010-06-29 00:49:29 +00:00
|
|
|
spin_lock_irqsave(&master->bus_lock_spinlock, flags);
|
2009-09-22 23:46:18 +00:00
|
|
|
|
spi/mmc_spi: SPI bus locking API, using mutex
SPI bus locking API to allow exclusive access to the SPI bus, especially, but
not limited to, for the mmc_spi driver.
Coded according to an outline from Grant Likely; here is his
specification (accidentally swapped function names corrected):
It requires 3 things to be added to struct spi_master.
- 1 Mutex
- 1 spin lock
- 1 flag.
The mutex protects spi_sync, and provides sleeping "for free"
The spinlock protects the atomic spi_async call.
The flag is set when the lock is obtained, and checked while holding
the spinlock in spi_async(). If the flag is checked, then spi_async()
must fail immediately.
The current runtime API looks like this:
spi_async(struct spi_device*, struct spi_message*);
spi_sync(struct spi_device*, struct spi_message*);
The API needs to be extended to this:
spi_async(struct spi_device*, struct spi_message*)
spi_sync(struct spi_device*, struct spi_message*)
spi_bus_lock(struct spi_master*) /* although struct spi_device* might
be easier */
spi_bus_unlock(struct spi_master*)
spi_async_locked(struct spi_device*, struct spi_message*)
spi_sync_locked(struct spi_device*, struct spi_message*)
Drivers can only call the last two if they already hold the spi_master_lock().
spi_bus_lock() obtains the mutex, obtains the spin lock, sets the
flag, and releases the spin lock before returning. It doesn't even
need to sleep while waiting for "in-flight" spi_transactions to
complete because its purpose is to guarantee no additional
transactions are added. It does not guarantee that the bus is idle.
spi_bus_unlock() clears the flag and releases the mutex, which will
wake up any waiters.
The difference between spi_async() and spi_async_locked() is that the
locked version bypasses the check of the lock flag. Both versions
need to obtain the spinlock.
The difference between spi_sync() and spi_sync_locked() is that
spi_sync() must hold the mutex while enqueuing a new transfer.
spi_sync_locked() doesn't because the mutex is already held. Note
however that spi_sync must *not* continue to hold the mutex while
waiting for the transfer to complete, otherwise only one transfer
could be queued up at a time!
Almost no code needs to be written. The current spi_async() and
spi_sync() can probably be renamed to __spi_async() and __spi_sync()
so that spi_async(), spi_sync(), spi_async_locked() and
spi_sync_locked() can just become wrappers around the common code.
spi_sync() is protected by a mutex because it can sleep
spi_async() needs to be protected with a flag and a spinlock because
it can be called atomically and must not sleep
Signed-off-by: Ernst Schwab <eschwab@online.de>
[grant.likely@secretlab.ca: use spin_lock_irqsave()]
Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
Tested-by: Matt Fleming <matt@console-pimps.org>
Tested-by: Antonio Ospite <ospite@studenti.unina.it>
2010-06-29 00:49:29 +00:00
|
|
|
if (master->bus_lock_flag)
|
|
|
|
ret = -EBUSY;
|
|
|
|
else
|
|
|
|
ret = __spi_async(spi, message);
|
2009-09-22 23:46:18 +00:00
|
|
|
|
spi/mmc_spi: SPI bus locking API, using mutex
SPI bus locking API to allow exclusive access to the SPI bus, especially, but
not limited to, for the mmc_spi driver.
Coded according to an outline from Grant Likely; here is his
specification (accidentally swapped function names corrected):
It requires 3 things to be added to struct spi_master.
- 1 Mutex
- 1 spin lock
- 1 flag.
The mutex protects spi_sync, and provides sleeping "for free"
The spinlock protects the atomic spi_async call.
The flag is set when the lock is obtained, and checked while holding
the spinlock in spi_async(). If the flag is checked, then spi_async()
must fail immediately.
The current runtime API looks like this:
spi_async(struct spi_device*, struct spi_message*);
spi_sync(struct spi_device*, struct spi_message*);
The API needs to be extended to this:
spi_async(struct spi_device*, struct spi_message*)
spi_sync(struct spi_device*, struct spi_message*)
spi_bus_lock(struct spi_master*) /* although struct spi_device* might
be easier */
spi_bus_unlock(struct spi_master*)
spi_async_locked(struct spi_device*, struct spi_message*)
spi_sync_locked(struct spi_device*, struct spi_message*)
Drivers can only call the last two if they already hold the spi_master_lock().
spi_bus_lock() obtains the mutex, obtains the spin lock, sets the
flag, and releases the spin lock before returning. It doesn't even
need to sleep while waiting for "in-flight" spi_transactions to
complete because its purpose is to guarantee no additional
transactions are added. It does not guarantee that the bus is idle.
spi_bus_unlock() clears the flag and releases the mutex, which will
wake up any waiters.
The difference between spi_async() and spi_async_locked() is that the
locked version bypasses the check of the lock flag. Both versions
need to obtain the spinlock.
The difference between spi_sync() and spi_sync_locked() is that
spi_sync() must hold the mutex while enqueuing a new transfer.
spi_sync_locked() doesn't because the mutex is already held. Note
however that spi_sync must *not* continue to hold the mutex while
waiting for the transfer to complete, otherwise only one transfer
could be queued up at a time!
Almost no code needs to be written. The current spi_async() and
spi_sync() can probably be renamed to __spi_async() and __spi_sync()
so that spi_async(), spi_sync(), spi_async_locked() and
spi_sync_locked() can just become wrappers around the common code.
spi_sync() is protected by a mutex because it can sleep
spi_async() needs to be protected with a flag and a spinlock because
it can be called atomically and must not sleep
Signed-off-by: Ernst Schwab <eschwab@online.de>
[grant.likely@secretlab.ca: use spin_lock_irqsave()]
Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
Tested-by: Matt Fleming <matt@console-pimps.org>
Tested-by: Antonio Ospite <ospite@studenti.unina.it>
2010-06-29 00:49:29 +00:00
|
|
|
spin_unlock_irqrestore(&master->bus_lock_spinlock, flags);
|
|
|
|
|
|
|
|
return ret;
|
2009-09-22 23:46:18 +00:00
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(spi_async);
|
|
|
|
|
spi/mmc_spi: SPI bus locking API, using mutex
SPI bus locking API to allow exclusive access to the SPI bus, especially, but
not limited to, for the mmc_spi driver.
Coded according to an outline from Grant Likely; here is his
specification (accidentally swapped function names corrected):
It requires 3 things to be added to struct spi_master.
- 1 Mutex
- 1 spin lock
- 1 flag.
The mutex protects spi_sync, and provides sleeping "for free"
The spinlock protects the atomic spi_async call.
The flag is set when the lock is obtained, and checked while holding
the spinlock in spi_async(). If the flag is checked, then spi_async()
must fail immediately.
The current runtime API looks like this:
spi_async(struct spi_device*, struct spi_message*);
spi_sync(struct spi_device*, struct spi_message*);
The API needs to be extended to this:
spi_async(struct spi_device*, struct spi_message*)
spi_sync(struct spi_device*, struct spi_message*)
spi_bus_lock(struct spi_master*) /* although struct spi_device* might
be easier */
spi_bus_unlock(struct spi_master*)
spi_async_locked(struct spi_device*, struct spi_message*)
spi_sync_locked(struct spi_device*, struct spi_message*)
Drivers can only call the last two if they already hold the spi_master_lock().
spi_bus_lock() obtains the mutex, obtains the spin lock, sets the
flag, and releases the spin lock before returning. It doesn't even
need to sleep while waiting for "in-flight" spi_transactions to
complete because its purpose is to guarantee no additional
transactions are added. It does not guarantee that the bus is idle.
spi_bus_unlock() clears the flag and releases the mutex, which will
wake up any waiters.
The difference between spi_async() and spi_async_locked() is that the
locked version bypasses the check of the lock flag. Both versions
need to obtain the spinlock.
The difference between spi_sync() and spi_sync_locked() is that
spi_sync() must hold the mutex while enqueuing a new transfer.
spi_sync_locked() doesn't because the mutex is already held. Note
however that spi_sync must *not* continue to hold the mutex while
waiting for the transfer to complete, otherwise only one transfer
could be queued up at a time!
Almost no code needs to be written. The current spi_async() and
spi_sync() can probably be renamed to __spi_async() and __spi_sync()
so that spi_async(), spi_sync(), spi_async_locked() and
spi_sync_locked() can just become wrappers around the common code.
spi_sync() is protected by a mutex because it can sleep
spi_async() needs to be protected with a flag and a spinlock because
it can be called atomically and must not sleep
Signed-off-by: Ernst Schwab <eschwab@online.de>
[grant.likely@secretlab.ca: use spin_lock_irqsave()]
Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
Tested-by: Matt Fleming <matt@console-pimps.org>
Tested-by: Antonio Ospite <ospite@studenti.unina.it>
2010-06-29 00:49:29 +00:00
|
|
|
/**
|
|
|
|
* spi_async_locked - version of spi_async with exclusive bus usage
|
|
|
|
* @spi: device with which data will be exchanged
|
|
|
|
* @message: describes the data transfers, including completion callback
|
|
|
|
* Context: any (irqs may be blocked, etc)
|
|
|
|
*
|
|
|
|
* This call may be used in_irq and other contexts which can't sleep,
|
|
|
|
* as well as from task contexts which can sleep.
|
|
|
|
*
|
|
|
|
* The completion callback is invoked in a context which can't sleep.
|
|
|
|
* Before that invocation, the value of message->status is undefined.
|
|
|
|
* When the callback is issued, message->status holds either zero (to
|
|
|
|
* indicate complete success) or a negative error code. After that
|
|
|
|
* callback returns, the driver which issued the transfer request may
|
|
|
|
* deallocate the associated memory; it's no longer in use by any SPI
|
|
|
|
* core or controller driver code.
|
|
|
|
*
|
|
|
|
* Note that although all messages to a spi_device are handled in
|
|
|
|
* FIFO order, messages may go to different devices in other orders.
|
|
|
|
* Some device might be higher priority, or have various "hard" access
|
|
|
|
* time requirements, for example.
|
|
|
|
*
|
|
|
|
* On detection of any fault during the transfer, processing of
|
|
|
|
* the entire message is aborted, and the device is deselected.
|
|
|
|
* Until returning from the associated message completion callback,
|
|
|
|
* no other spi_message queued to that device will be processed.
|
|
|
|
* (This rule applies equally to all the synchronous transfer calls,
|
|
|
|
* which are wrappers around this core asynchronous primitive.)
|
2015-10-22 16:59:23 +00:00
|
|
|
*
|
|
|
|
* Return: zero on success, else a negative error code.
|
spi/mmc_spi: SPI bus locking API, using mutex
SPI bus locking API to allow exclusive access to the SPI bus, especially, but
not limited to, for the mmc_spi driver.
Coded according to an outline from Grant Likely; here is his
specification (accidentally swapped function names corrected):
It requires 3 things to be added to struct spi_master.
- 1 Mutex
- 1 spin lock
- 1 flag.
The mutex protects spi_sync, and provides sleeping "for free"
The spinlock protects the atomic spi_async call.
The flag is set when the lock is obtained, and checked while holding
the spinlock in spi_async(). If the flag is checked, then spi_async()
must fail immediately.
The current runtime API looks like this:
spi_async(struct spi_device*, struct spi_message*);
spi_sync(struct spi_device*, struct spi_message*);
The API needs to be extended to this:
spi_async(struct spi_device*, struct spi_message*)
spi_sync(struct spi_device*, struct spi_message*)
spi_bus_lock(struct spi_master*) /* although struct spi_device* might
be easier */
spi_bus_unlock(struct spi_master*)
spi_async_locked(struct spi_device*, struct spi_message*)
spi_sync_locked(struct spi_device*, struct spi_message*)
Drivers can only call the last two if they already hold the spi_master_lock().
spi_bus_lock() obtains the mutex, obtains the spin lock, sets the
flag, and releases the spin lock before returning. It doesn't even
need to sleep while waiting for "in-flight" spi_transactions to
complete because its purpose is to guarantee no additional
transactions are added. It does not guarantee that the bus is idle.
spi_bus_unlock() clears the flag and releases the mutex, which will
wake up any waiters.
The difference between spi_async() and spi_async_locked() is that the
locked version bypasses the check of the lock flag. Both versions
need to obtain the spinlock.
The difference between spi_sync() and spi_sync_locked() is that
spi_sync() must hold the mutex while enqueuing a new transfer.
spi_sync_locked() doesn't because the mutex is already held. Note
however that spi_sync must *not* continue to hold the mutex while
waiting for the transfer to complete, otherwise only one transfer
could be queued up at a time!
Almost no code needs to be written. The current spi_async() and
spi_sync() can probably be renamed to __spi_async() and __spi_sync()
so that spi_async(), spi_sync(), spi_async_locked() and
spi_sync_locked() can just become wrappers around the common code.
spi_sync() is protected by a mutex because it can sleep
spi_async() needs to be protected with a flag and a spinlock because
it can be called atomically and must not sleep
Signed-off-by: Ernst Schwab <eschwab@online.de>
[grant.likely@secretlab.ca: use spin_lock_irqsave()]
Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
Tested-by: Matt Fleming <matt@console-pimps.org>
Tested-by: Antonio Ospite <ospite@studenti.unina.it>
2010-06-29 00:49:29 +00:00
|
|
|
*/
|
|
|
|
int spi_async_locked(struct spi_device *spi, struct spi_message *message)
|
|
|
|
{
|
|
|
|
struct spi_master *master = spi->master;
|
|
|
|
int ret;
|
|
|
|
unsigned long flags;
|
|
|
|
|
2013-11-13 23:44:15 +00:00
|
|
|
ret = __spi_validate(spi, message);
|
|
|
|
if (ret != 0)
|
|
|
|
return ret;
|
|
|
|
|
spi/mmc_spi: SPI bus locking API, using mutex
SPI bus locking API to allow exclusive access to the SPI bus, especially, but
not limited to, for the mmc_spi driver.
Coded according to an outline from Grant Likely; here is his
specification (accidentally swapped function names corrected):
It requires 3 things to be added to struct spi_master.
- 1 Mutex
- 1 spin lock
- 1 flag.
The mutex protects spi_sync, and provides sleeping "for free"
The spinlock protects the atomic spi_async call.
The flag is set when the lock is obtained, and checked while holding
the spinlock in spi_async(). If the flag is checked, then spi_async()
must fail immediately.
The current runtime API looks like this:
spi_async(struct spi_device*, struct spi_message*);
spi_sync(struct spi_device*, struct spi_message*);
The API needs to be extended to this:
spi_async(struct spi_device*, struct spi_message*)
spi_sync(struct spi_device*, struct spi_message*)
spi_bus_lock(struct spi_master*) /* although struct spi_device* might
be easier */
spi_bus_unlock(struct spi_master*)
spi_async_locked(struct spi_device*, struct spi_message*)
spi_sync_locked(struct spi_device*, struct spi_message*)
Drivers can only call the last two if they already hold the spi_master_lock().
spi_bus_lock() obtains the mutex, obtains the spin lock, sets the
flag, and releases the spin lock before returning. It doesn't even
need to sleep while waiting for "in-flight" spi_transactions to
complete because its purpose is to guarantee no additional
transactions are added. It does not guarantee that the bus is idle.
spi_bus_unlock() clears the flag and releases the mutex, which will
wake up any waiters.
The difference between spi_async() and spi_async_locked() is that the
locked version bypasses the check of the lock flag. Both versions
need to obtain the spinlock.
The difference between spi_sync() and spi_sync_locked() is that
spi_sync() must hold the mutex while enqueuing a new transfer.
spi_sync_locked() doesn't because the mutex is already held. Note
however that spi_sync must *not* continue to hold the mutex while
waiting for the transfer to complete, otherwise only one transfer
could be queued up at a time!
Almost no code needs to be written. The current spi_async() and
spi_sync() can probably be renamed to __spi_async() and __spi_sync()
so that spi_async(), spi_sync(), spi_async_locked() and
spi_sync_locked() can just become wrappers around the common code.
spi_sync() is protected by a mutex because it can sleep
spi_async() needs to be protected with a flag and a spinlock because
it can be called atomically and must not sleep
Signed-off-by: Ernst Schwab <eschwab@online.de>
[grant.likely@secretlab.ca: use spin_lock_irqsave()]
Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
Tested-by: Matt Fleming <matt@console-pimps.org>
Tested-by: Antonio Ospite <ospite@studenti.unina.it>
2010-06-29 00:49:29 +00:00
|
|
|
spin_lock_irqsave(&master->bus_lock_spinlock, flags);
|
|
|
|
|
|
|
|
ret = __spi_async(spi, message);
|
|
|
|
|
|
|
|
spin_unlock_irqrestore(&master->bus_lock_spinlock, flags);
|
|
|
|
|
|
|
|
return ret;
|
|
|
|
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(spi_async_locked);
|
|
|
|
|
2009-06-17 23:26:03 +00:00
|
|
|
|
|
|
|
/*-------------------------------------------------------------------------*/
|
|
|
|
|
|
|
|
/* Utility methods for SPI master protocol drivers, layered on
|
|
|
|
* top of the core. Some other utility methods are defined as
|
|
|
|
* inline functions.
|
|
|
|
*/
|
|
|
|
|
2006-01-11 19:23:49 +00:00
|
|
|
static void spi_complete(void *arg)
|
|
|
|
{
|
|
|
|
complete(arg);
|
|
|
|
}
|
|
|
|
|
spi/mmc_spi: SPI bus locking API, using mutex
SPI bus locking API to allow exclusive access to the SPI bus, especially, but
not limited to, for the mmc_spi driver.
Coded according to an outline from Grant Likely; here is his
specification (accidentally swapped function names corrected):
It requires 3 things to be added to struct spi_master.
- 1 Mutex
- 1 spin lock
- 1 flag.
The mutex protects spi_sync, and provides sleeping "for free"
The spinlock protects the atomic spi_async call.
The flag is set when the lock is obtained, and checked while holding
the spinlock in spi_async(). If the flag is checked, then spi_async()
must fail immediately.
The current runtime API looks like this:
spi_async(struct spi_device*, struct spi_message*);
spi_sync(struct spi_device*, struct spi_message*);
The API needs to be extended to this:
spi_async(struct spi_device*, struct spi_message*)
spi_sync(struct spi_device*, struct spi_message*)
spi_bus_lock(struct spi_master*) /* although struct spi_device* might
be easier */
spi_bus_unlock(struct spi_master*)
spi_async_locked(struct spi_device*, struct spi_message*)
spi_sync_locked(struct spi_device*, struct spi_message*)
Drivers can only call the last two if they already hold the spi_master_lock().
spi_bus_lock() obtains the mutex, obtains the spin lock, sets the
flag, and releases the spin lock before returning. It doesn't even
need to sleep while waiting for "in-flight" spi_transactions to
complete because its purpose is to guarantee no additional
transactions are added. It does not guarantee that the bus is idle.
spi_bus_unlock() clears the flag and releases the mutex, which will
wake up any waiters.
The difference between spi_async() and spi_async_locked() is that the
locked version bypasses the check of the lock flag. Both versions
need to obtain the spinlock.
The difference between spi_sync() and spi_sync_locked() is that
spi_sync() must hold the mutex while enqueuing a new transfer.
spi_sync_locked() doesn't because the mutex is already held. Note
however that spi_sync must *not* continue to hold the mutex while
waiting for the transfer to complete, otherwise only one transfer
could be queued up at a time!
Almost no code needs to be written. The current spi_async() and
spi_sync() can probably be renamed to __spi_async() and __spi_sync()
so that spi_async(), spi_sync(), spi_async_locked() and
spi_sync_locked() can just become wrappers around the common code.
spi_sync() is protected by a mutex because it can sleep
spi_async() needs to be protected with a flag and a spinlock because
it can be called atomically and must not sleep
Signed-off-by: Ernst Schwab <eschwab@online.de>
[grant.likely@secretlab.ca: use spin_lock_irqsave()]
Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
Tested-by: Matt Fleming <matt@console-pimps.org>
Tested-by: Antonio Ospite <ospite@studenti.unina.it>
2010-06-29 00:49:29 +00:00
|
|
|
static int __spi_sync(struct spi_device *spi, struct spi_message *message,
|
|
|
|
int bus_locked)
|
|
|
|
{
|
|
|
|
DECLARE_COMPLETION_ONSTACK(done);
|
|
|
|
int status;
|
|
|
|
struct spi_master *master = spi->master;
|
2014-12-09 21:38:05 +00:00
|
|
|
unsigned long flags;
|
|
|
|
|
|
|
|
status = __spi_validate(spi, message);
|
|
|
|
if (status != 0)
|
|
|
|
return status;
|
spi/mmc_spi: SPI bus locking API, using mutex
SPI bus locking API to allow exclusive access to the SPI bus, especially, but
not limited to, for the mmc_spi driver.
Coded according to an outline from Grant Likely; here is his
specification (accidentally swapped function names corrected):
It requires 3 things to be added to struct spi_master.
- 1 Mutex
- 1 spin lock
- 1 flag.
The mutex protects spi_sync, and provides sleeping "for free"
The spinlock protects the atomic spi_async call.
The flag is set when the lock is obtained, and checked while holding
the spinlock in spi_async(). If the flag is checked, then spi_async()
must fail immediately.
The current runtime API looks like this:
spi_async(struct spi_device*, struct spi_message*);
spi_sync(struct spi_device*, struct spi_message*);
The API needs to be extended to this:
spi_async(struct spi_device*, struct spi_message*)
spi_sync(struct spi_device*, struct spi_message*)
spi_bus_lock(struct spi_master*) /* although struct spi_device* might
be easier */
spi_bus_unlock(struct spi_master*)
spi_async_locked(struct spi_device*, struct spi_message*)
spi_sync_locked(struct spi_device*, struct spi_message*)
Drivers can only call the last two if they already hold the spi_master_lock().
spi_bus_lock() obtains the mutex, obtains the spin lock, sets the
flag, and releases the spin lock before returning. It doesn't even
need to sleep while waiting for "in-flight" spi_transactions to
complete because its purpose is to guarantee no additional
transactions are added. It does not guarantee that the bus is idle.
spi_bus_unlock() clears the flag and releases the mutex, which will
wake up any waiters.
The difference between spi_async() and spi_async_locked() is that the
locked version bypasses the check of the lock flag. Both versions
need to obtain the spinlock.
The difference between spi_sync() and spi_sync_locked() is that
spi_sync() must hold the mutex while enqueuing a new transfer.
spi_sync_locked() doesn't because the mutex is already held. Note
however that spi_sync must *not* continue to hold the mutex while
waiting for the transfer to complete, otherwise only one transfer
could be queued up at a time!
Almost no code needs to be written. The current spi_async() and
spi_sync() can probably be renamed to __spi_async() and __spi_sync()
so that spi_async(), spi_sync(), spi_async_locked() and
spi_sync_locked() can just become wrappers around the common code.
spi_sync() is protected by a mutex because it can sleep
spi_async() needs to be protected with a flag and a spinlock because
it can be called atomically and must not sleep
Signed-off-by: Ernst Schwab <eschwab@online.de>
[grant.likely@secretlab.ca: use spin_lock_irqsave()]
Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
Tested-by: Matt Fleming <matt@console-pimps.org>
Tested-by: Antonio Ospite <ospite@studenti.unina.it>
2010-06-29 00:49:29 +00:00
|
|
|
|
|
|
|
message->complete = spi_complete;
|
|
|
|
message->context = &done;
|
2014-12-09 21:38:05 +00:00
|
|
|
message->spi = spi;
|
spi/mmc_spi: SPI bus locking API, using mutex
SPI bus locking API to allow exclusive access to the SPI bus, especially, but
not limited to, for the mmc_spi driver.
Coded according to an outline from Grant Likely; here is his
specification (accidentally swapped function names corrected):
It requires 3 things to be added to struct spi_master.
- 1 Mutex
- 1 spin lock
- 1 flag.
The mutex protects spi_sync, and provides sleeping "for free"
The spinlock protects the atomic spi_async call.
The flag is set when the lock is obtained, and checked while holding
the spinlock in spi_async(). If the flag is checked, then spi_async()
must fail immediately.
The current runtime API looks like this:
spi_async(struct spi_device*, struct spi_message*);
spi_sync(struct spi_device*, struct spi_message*);
The API needs to be extended to this:
spi_async(struct spi_device*, struct spi_message*)
spi_sync(struct spi_device*, struct spi_message*)
spi_bus_lock(struct spi_master*) /* although struct spi_device* might
be easier */
spi_bus_unlock(struct spi_master*)
spi_async_locked(struct spi_device*, struct spi_message*)
spi_sync_locked(struct spi_device*, struct spi_message*)
Drivers can only call the last two if they already hold the spi_master_lock().
spi_bus_lock() obtains the mutex, obtains the spin lock, sets the
flag, and releases the spin lock before returning. It doesn't even
need to sleep while waiting for "in-flight" spi_transactions to
complete because its purpose is to guarantee no additional
transactions are added. It does not guarantee that the bus is idle.
spi_bus_unlock() clears the flag and releases the mutex, which will
wake up any waiters.
The difference between spi_async() and spi_async_locked() is that the
locked version bypasses the check of the lock flag. Both versions
need to obtain the spinlock.
The difference between spi_sync() and spi_sync_locked() is that
spi_sync() must hold the mutex while enqueuing a new transfer.
spi_sync_locked() doesn't because the mutex is already held. Note
however that spi_sync must *not* continue to hold the mutex while
waiting for the transfer to complete, otherwise only one transfer
could be queued up at a time!
Almost no code needs to be written. The current spi_async() and
spi_sync() can probably be renamed to __spi_async() and __spi_sync()
so that spi_async(), spi_sync(), spi_async_locked() and
spi_sync_locked() can just become wrappers around the common code.
spi_sync() is protected by a mutex because it can sleep
spi_async() needs to be protected with a flag and a spinlock because
it can be called atomically and must not sleep
Signed-off-by: Ernst Schwab <eschwab@online.de>
[grant.likely@secretlab.ca: use spin_lock_irqsave()]
Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
Tested-by: Matt Fleming <matt@console-pimps.org>
Tested-by: Antonio Ospite <ospite@studenti.unina.it>
2010-06-29 00:49:29 +00:00
|
|
|
|
2015-06-22 13:00:36 +00:00
|
|
|
SPI_STATISTICS_INCREMENT_FIELD(&master->statistics, spi_sync);
|
|
|
|
SPI_STATISTICS_INCREMENT_FIELD(&spi->statistics, spi_sync);
|
|
|
|
|
spi/mmc_spi: SPI bus locking API, using mutex
SPI bus locking API to allow exclusive access to the SPI bus, especially, but
not limited to, for the mmc_spi driver.
Coded according to an outline from Grant Likely; here is his
specification (accidentally swapped function names corrected):
It requires 3 things to be added to struct spi_master.
- 1 Mutex
- 1 spin lock
- 1 flag.
The mutex protects spi_sync, and provides sleeping "for free"
The spinlock protects the atomic spi_async call.
The flag is set when the lock is obtained, and checked while holding
the spinlock in spi_async(). If the flag is checked, then spi_async()
must fail immediately.
The current runtime API looks like this:
spi_async(struct spi_device*, struct spi_message*);
spi_sync(struct spi_device*, struct spi_message*);
The API needs to be extended to this:
spi_async(struct spi_device*, struct spi_message*)
spi_sync(struct spi_device*, struct spi_message*)
spi_bus_lock(struct spi_master*) /* although struct spi_device* might
be easier */
spi_bus_unlock(struct spi_master*)
spi_async_locked(struct spi_device*, struct spi_message*)
spi_sync_locked(struct spi_device*, struct spi_message*)
Drivers can only call the last two if they already hold the spi_master_lock().
spi_bus_lock() obtains the mutex, obtains the spin lock, sets the
flag, and releases the spin lock before returning. It doesn't even
need to sleep while waiting for "in-flight" spi_transactions to
complete because its purpose is to guarantee no additional
transactions are added. It does not guarantee that the bus is idle.
spi_bus_unlock() clears the flag and releases the mutex, which will
wake up any waiters.
The difference between spi_async() and spi_async_locked() is that the
locked version bypasses the check of the lock flag. Both versions
need to obtain the spinlock.
The difference between spi_sync() and spi_sync_locked() is that
spi_sync() must hold the mutex while enqueuing a new transfer.
spi_sync_locked() doesn't because the mutex is already held. Note
however that spi_sync must *not* continue to hold the mutex while
waiting for the transfer to complete, otherwise only one transfer
could be queued up at a time!
Almost no code needs to be written. The current spi_async() and
spi_sync() can probably be renamed to __spi_async() and __spi_sync()
so that spi_async(), spi_sync(), spi_async_locked() and
spi_sync_locked() can just become wrappers around the common code.
spi_sync() is protected by a mutex because it can sleep
spi_async() needs to be protected with a flag and a spinlock because
it can be called atomically and must not sleep
Signed-off-by: Ernst Schwab <eschwab@online.de>
[grant.likely@secretlab.ca: use spin_lock_irqsave()]
Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
Tested-by: Matt Fleming <matt@console-pimps.org>
Tested-by: Antonio Ospite <ospite@studenti.unina.it>
2010-06-29 00:49:29 +00:00
|
|
|
if (!bus_locked)
|
|
|
|
mutex_lock(&master->bus_lock_mutex);
|
|
|
|
|
2014-12-09 21:38:05 +00:00
|
|
|
/* If we're not using the legacy transfer method then we will
|
|
|
|
* try to transfer in the calling context so special case.
|
|
|
|
* This code would be less tricky if we could remove the
|
|
|
|
* support for driver implemented message queues.
|
|
|
|
*/
|
|
|
|
if (master->transfer == spi_queued_transfer) {
|
|
|
|
spin_lock_irqsave(&master->bus_lock_spinlock, flags);
|
|
|
|
|
|
|
|
trace_spi_message_submit(message);
|
|
|
|
|
|
|
|
status = __spi_queued_transfer(spi, message, false);
|
|
|
|
|
|
|
|
spin_unlock_irqrestore(&master->bus_lock_spinlock, flags);
|
|
|
|
} else {
|
|
|
|
status = spi_async_locked(spi, message);
|
|
|
|
}
|
spi/mmc_spi: SPI bus locking API, using mutex
SPI bus locking API to allow exclusive access to the SPI bus, especially, but
not limited to, for the mmc_spi driver.
Coded according to an outline from Grant Likely; here is his
specification (accidentally swapped function names corrected):
It requires 3 things to be added to struct spi_master.
- 1 Mutex
- 1 spin lock
- 1 flag.
The mutex protects spi_sync, and provides sleeping "for free"
The spinlock protects the atomic spi_async call.
The flag is set when the lock is obtained, and checked while holding
the spinlock in spi_async(). If the flag is checked, then spi_async()
must fail immediately.
The current runtime API looks like this:
spi_async(struct spi_device*, struct spi_message*);
spi_sync(struct spi_device*, struct spi_message*);
The API needs to be extended to this:
spi_async(struct spi_device*, struct spi_message*)
spi_sync(struct spi_device*, struct spi_message*)
spi_bus_lock(struct spi_master*) /* although struct spi_device* might
be easier */
spi_bus_unlock(struct spi_master*)
spi_async_locked(struct spi_device*, struct spi_message*)
spi_sync_locked(struct spi_device*, struct spi_message*)
Drivers can only call the last two if they already hold the spi_master_lock().
spi_bus_lock() obtains the mutex, obtains the spin lock, sets the
flag, and releases the spin lock before returning. It doesn't even
need to sleep while waiting for "in-flight" spi_transactions to
complete because its purpose is to guarantee no additional
transactions are added. It does not guarantee that the bus is idle.
spi_bus_unlock() clears the flag and releases the mutex, which will
wake up any waiters.
The difference between spi_async() and spi_async_locked() is that the
locked version bypasses the check of the lock flag. Both versions
need to obtain the spinlock.
The difference between spi_sync() and spi_sync_locked() is that
spi_sync() must hold the mutex while enqueuing a new transfer.
spi_sync_locked() doesn't because the mutex is already held. Note
however that spi_sync must *not* continue to hold the mutex while
waiting for the transfer to complete, otherwise only one transfer
could be queued up at a time!
Almost no code needs to be written. The current spi_async() and
spi_sync() can probably be renamed to __spi_async() and __spi_sync()
so that spi_async(), spi_sync(), spi_async_locked() and
spi_sync_locked() can just become wrappers around the common code.
spi_sync() is protected by a mutex because it can sleep
spi_async() needs to be protected with a flag and a spinlock because
it can be called atomically and must not sleep
Signed-off-by: Ernst Schwab <eschwab@online.de>
[grant.likely@secretlab.ca: use spin_lock_irqsave()]
Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
Tested-by: Matt Fleming <matt@console-pimps.org>
Tested-by: Antonio Ospite <ospite@studenti.unina.it>
2010-06-29 00:49:29 +00:00
|
|
|
|
|
|
|
if (!bus_locked)
|
|
|
|
mutex_unlock(&master->bus_lock_mutex);
|
|
|
|
|
|
|
|
if (status == 0) {
|
2014-12-09 21:38:05 +00:00
|
|
|
/* Push out the messages in the calling context if we
|
|
|
|
* can.
|
|
|
|
*/
|
2015-06-22 13:00:36 +00:00
|
|
|
if (master->transfer == spi_queued_transfer) {
|
|
|
|
SPI_STATISTICS_INCREMENT_FIELD(&master->statistics,
|
|
|
|
spi_sync_immediate);
|
|
|
|
SPI_STATISTICS_INCREMENT_FIELD(&spi->statistics,
|
|
|
|
spi_sync_immediate);
|
2014-12-10 13:46:33 +00:00
|
|
|
__spi_pump_messages(master, false);
|
2015-06-22 13:00:36 +00:00
|
|
|
}
|
2014-12-09 21:38:05 +00:00
|
|
|
|
spi/mmc_spi: SPI bus locking API, using mutex
SPI bus locking API to allow exclusive access to the SPI bus, especially, but
not limited to, for the mmc_spi driver.
Coded according to an outline from Grant Likely; here is his
specification (accidentally swapped function names corrected):
It requires 3 things to be added to struct spi_master.
- 1 Mutex
- 1 spin lock
- 1 flag.
The mutex protects spi_sync, and provides sleeping "for free"
The spinlock protects the atomic spi_async call.
The flag is set when the lock is obtained, and checked while holding
the spinlock in spi_async(). If the flag is checked, then spi_async()
must fail immediately.
The current runtime API looks like this:
spi_async(struct spi_device*, struct spi_message*);
spi_sync(struct spi_device*, struct spi_message*);
The API needs to be extended to this:
spi_async(struct spi_device*, struct spi_message*)
spi_sync(struct spi_device*, struct spi_message*)
spi_bus_lock(struct spi_master*) /* although struct spi_device* might
be easier */
spi_bus_unlock(struct spi_master*)
spi_async_locked(struct spi_device*, struct spi_message*)
spi_sync_locked(struct spi_device*, struct spi_message*)
Drivers can only call the last two if they already hold the spi_master_lock().
spi_bus_lock() obtains the mutex, obtains the spin lock, sets the
flag, and releases the spin lock before returning. It doesn't even
need to sleep while waiting for "in-flight" spi_transactions to
complete because its purpose is to guarantee no additional
transactions are added. It does not guarantee that the bus is idle.
spi_bus_unlock() clears the flag and releases the mutex, which will
wake up any waiters.
The difference between spi_async() and spi_async_locked() is that the
locked version bypasses the check of the lock flag. Both versions
need to obtain the spinlock.
The difference between spi_sync() and spi_sync_locked() is that
spi_sync() must hold the mutex while enqueuing a new transfer.
spi_sync_locked() doesn't because the mutex is already held. Note
however that spi_sync must *not* continue to hold the mutex while
waiting for the transfer to complete, otherwise only one transfer
could be queued up at a time!
Almost no code needs to be written. The current spi_async() and
spi_sync() can probably be renamed to __spi_async() and __spi_sync()
so that spi_async(), spi_sync(), spi_async_locked() and
spi_sync_locked() can just become wrappers around the common code.
spi_sync() is protected by a mutex because it can sleep
spi_async() needs to be protected with a flag and a spinlock because
it can be called atomically and must not sleep
Signed-off-by: Ernst Schwab <eschwab@online.de>
[grant.likely@secretlab.ca: use spin_lock_irqsave()]
Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
Tested-by: Matt Fleming <matt@console-pimps.org>
Tested-by: Antonio Ospite <ospite@studenti.unina.it>
2010-06-29 00:49:29 +00:00
|
|
|
wait_for_completion(&done);
|
|
|
|
status = message->status;
|
|
|
|
}
|
|
|
|
message->context = NULL;
|
|
|
|
return status;
|
|
|
|
}
|
|
|
|
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
/**
|
|
|
|
* spi_sync - blocking/synchronous SPI data transfers
|
|
|
|
* @spi: device with which data will be exchanged
|
|
|
|
* @message: describes the data transfers
|
2007-05-08 07:32:21 +00:00
|
|
|
* Context: can sleep
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
*
|
|
|
|
* This call may only be used from a context that may sleep. The sleep
|
|
|
|
* is non-interruptible, and has no timeout. Low-overhead controller
|
|
|
|
* drivers may DMA directly into and out of the message buffers.
|
|
|
|
*
|
|
|
|
* Note that the SPI device's chip select is active during the message,
|
|
|
|
* and then is normally disabled between messages. Drivers for some
|
|
|
|
* frequently-used devices may want to minimize costs of selecting a chip,
|
|
|
|
* by leaving it selected in anticipation that the next message will go
|
|
|
|
* to the same chip. (That may increase power usage.)
|
|
|
|
*
|
2006-01-08 21:34:25 +00:00
|
|
|
* Also, the caller is guaranteeing that the memory associated with the
|
|
|
|
* message will not be freed before this call returns.
|
|
|
|
*
|
2015-10-22 16:59:23 +00:00
|
|
|
* Return: zero on success, else a negative error code.
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
*/
|
|
|
|
int spi_sync(struct spi_device *spi, struct spi_message *message)
|
|
|
|
{
|
spi/mmc_spi: SPI bus locking API, using mutex
SPI bus locking API to allow exclusive access to the SPI bus, especially, but
not limited to, for the mmc_spi driver.
Coded according to an outline from Grant Likely; here is his
specification (accidentally swapped function names corrected):
It requires 3 things to be added to struct spi_master.
- 1 Mutex
- 1 spin lock
- 1 flag.
The mutex protects spi_sync, and provides sleeping "for free"
The spinlock protects the atomic spi_async call.
The flag is set when the lock is obtained, and checked while holding
the spinlock in spi_async(). If the flag is checked, then spi_async()
must fail immediately.
The current runtime API looks like this:
spi_async(struct spi_device*, struct spi_message*);
spi_sync(struct spi_device*, struct spi_message*);
The API needs to be extended to this:
spi_async(struct spi_device*, struct spi_message*)
spi_sync(struct spi_device*, struct spi_message*)
spi_bus_lock(struct spi_master*) /* although struct spi_device* might
be easier */
spi_bus_unlock(struct spi_master*)
spi_async_locked(struct spi_device*, struct spi_message*)
spi_sync_locked(struct spi_device*, struct spi_message*)
Drivers can only call the last two if they already hold the spi_master_lock().
spi_bus_lock() obtains the mutex, obtains the spin lock, sets the
flag, and releases the spin lock before returning. It doesn't even
need to sleep while waiting for "in-flight" spi_transactions to
complete because its purpose is to guarantee no additional
transactions are added. It does not guarantee that the bus is idle.
spi_bus_unlock() clears the flag and releases the mutex, which will
wake up any waiters.
The difference between spi_async() and spi_async_locked() is that the
locked version bypasses the check of the lock flag. Both versions
need to obtain the spinlock.
The difference between spi_sync() and spi_sync_locked() is that
spi_sync() must hold the mutex while enqueuing a new transfer.
spi_sync_locked() doesn't because the mutex is already held. Note
however that spi_sync must *not* continue to hold the mutex while
waiting for the transfer to complete, otherwise only one transfer
could be queued up at a time!
Almost no code needs to be written. The current spi_async() and
spi_sync() can probably be renamed to __spi_async() and __spi_sync()
so that spi_async(), spi_sync(), spi_async_locked() and
spi_sync_locked() can just become wrappers around the common code.
spi_sync() is protected by a mutex because it can sleep
spi_async() needs to be protected with a flag and a spinlock because
it can be called atomically and must not sleep
Signed-off-by: Ernst Schwab <eschwab@online.de>
[grant.likely@secretlab.ca: use spin_lock_irqsave()]
Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
Tested-by: Matt Fleming <matt@console-pimps.org>
Tested-by: Antonio Ospite <ospite@studenti.unina.it>
2010-06-29 00:49:29 +00:00
|
|
|
return __spi_sync(spi, message, 0);
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(spi_sync);
|
|
|
|
|
spi/mmc_spi: SPI bus locking API, using mutex
SPI bus locking API to allow exclusive access to the SPI bus, especially, but
not limited to, for the mmc_spi driver.
Coded according to an outline from Grant Likely; here is his
specification (accidentally swapped function names corrected):
It requires 3 things to be added to struct spi_master.
- 1 Mutex
- 1 spin lock
- 1 flag.
The mutex protects spi_sync, and provides sleeping "for free"
The spinlock protects the atomic spi_async call.
The flag is set when the lock is obtained, and checked while holding
the spinlock in spi_async(). If the flag is checked, then spi_async()
must fail immediately.
The current runtime API looks like this:
spi_async(struct spi_device*, struct spi_message*);
spi_sync(struct spi_device*, struct spi_message*);
The API needs to be extended to this:
spi_async(struct spi_device*, struct spi_message*)
spi_sync(struct spi_device*, struct spi_message*)
spi_bus_lock(struct spi_master*) /* although struct spi_device* might
be easier */
spi_bus_unlock(struct spi_master*)
spi_async_locked(struct spi_device*, struct spi_message*)
spi_sync_locked(struct spi_device*, struct spi_message*)
Drivers can only call the last two if they already hold the spi_master_lock().
spi_bus_lock() obtains the mutex, obtains the spin lock, sets the
flag, and releases the spin lock before returning. It doesn't even
need to sleep while waiting for "in-flight" spi_transactions to
complete because its purpose is to guarantee no additional
transactions are added. It does not guarantee that the bus is idle.
spi_bus_unlock() clears the flag and releases the mutex, which will
wake up any waiters.
The difference between spi_async() and spi_async_locked() is that the
locked version bypasses the check of the lock flag. Both versions
need to obtain the spinlock.
The difference between spi_sync() and spi_sync_locked() is that
spi_sync() must hold the mutex while enqueuing a new transfer.
spi_sync_locked() doesn't because the mutex is already held. Note
however that spi_sync must *not* continue to hold the mutex while
waiting for the transfer to complete, otherwise only one transfer
could be queued up at a time!
Almost no code needs to be written. The current spi_async() and
spi_sync() can probably be renamed to __spi_async() and __spi_sync()
so that spi_async(), spi_sync(), spi_async_locked() and
spi_sync_locked() can just become wrappers around the common code.
spi_sync() is protected by a mutex because it can sleep
spi_async() needs to be protected with a flag and a spinlock because
it can be called atomically and must not sleep
Signed-off-by: Ernst Schwab <eschwab@online.de>
[grant.likely@secretlab.ca: use spin_lock_irqsave()]
Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
Tested-by: Matt Fleming <matt@console-pimps.org>
Tested-by: Antonio Ospite <ospite@studenti.unina.it>
2010-06-29 00:49:29 +00:00
|
|
|
/**
|
|
|
|
* spi_sync_locked - version of spi_sync with exclusive bus usage
|
|
|
|
* @spi: device with which data will be exchanged
|
|
|
|
* @message: describes the data transfers
|
|
|
|
* Context: can sleep
|
|
|
|
*
|
|
|
|
* This call may only be used from a context that may sleep. The sleep
|
|
|
|
* is non-interruptible, and has no timeout. Low-overhead controller
|
|
|
|
* drivers may DMA directly into and out of the message buffers.
|
|
|
|
*
|
|
|
|
* This call should be used by drivers that require exclusive access to the
|
2011-03-31 01:57:33 +00:00
|
|
|
* SPI bus. It has to be preceded by a spi_bus_lock call. The SPI bus must
|
spi/mmc_spi: SPI bus locking API, using mutex
SPI bus locking API to allow exclusive access to the SPI bus, especially, but
not limited to, for the mmc_spi driver.
Coded according to an outline from Grant Likely; here is his
specification (accidentally swapped function names corrected):
It requires 3 things to be added to struct spi_master.
- 1 Mutex
- 1 spin lock
- 1 flag.
The mutex protects spi_sync, and provides sleeping "for free"
The spinlock protects the atomic spi_async call.
The flag is set when the lock is obtained, and checked while holding
the spinlock in spi_async(). If the flag is checked, then spi_async()
must fail immediately.
The current runtime API looks like this:
spi_async(struct spi_device*, struct spi_message*);
spi_sync(struct spi_device*, struct spi_message*);
The API needs to be extended to this:
spi_async(struct spi_device*, struct spi_message*)
spi_sync(struct spi_device*, struct spi_message*)
spi_bus_lock(struct spi_master*) /* although struct spi_device* might
be easier */
spi_bus_unlock(struct spi_master*)
spi_async_locked(struct spi_device*, struct spi_message*)
spi_sync_locked(struct spi_device*, struct spi_message*)
Drivers can only call the last two if they already hold the spi_master_lock().
spi_bus_lock() obtains the mutex, obtains the spin lock, sets the
flag, and releases the spin lock before returning. It doesn't even
need to sleep while waiting for "in-flight" spi_transactions to
complete because its purpose is to guarantee no additional
transactions are added. It does not guarantee that the bus is idle.
spi_bus_unlock() clears the flag and releases the mutex, which will
wake up any waiters.
The difference between spi_async() and spi_async_locked() is that the
locked version bypasses the check of the lock flag. Both versions
need to obtain the spinlock.
The difference between spi_sync() and spi_sync_locked() is that
spi_sync() must hold the mutex while enqueuing a new transfer.
spi_sync_locked() doesn't because the mutex is already held. Note
however that spi_sync must *not* continue to hold the mutex while
waiting for the transfer to complete, otherwise only one transfer
could be queued up at a time!
Almost no code needs to be written. The current spi_async() and
spi_sync() can probably be renamed to __spi_async() and __spi_sync()
so that spi_async(), spi_sync(), spi_async_locked() and
spi_sync_locked() can just become wrappers around the common code.
spi_sync() is protected by a mutex because it can sleep
spi_async() needs to be protected with a flag and a spinlock because
it can be called atomically and must not sleep
Signed-off-by: Ernst Schwab <eschwab@online.de>
[grant.likely@secretlab.ca: use spin_lock_irqsave()]
Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
Tested-by: Matt Fleming <matt@console-pimps.org>
Tested-by: Antonio Ospite <ospite@studenti.unina.it>
2010-06-29 00:49:29 +00:00
|
|
|
* be released by a spi_bus_unlock call when the exclusive access is over.
|
|
|
|
*
|
2015-10-22 16:59:23 +00:00
|
|
|
* Return: zero on success, else a negative error code.
|
spi/mmc_spi: SPI bus locking API, using mutex
SPI bus locking API to allow exclusive access to the SPI bus, especially, but
not limited to, for the mmc_spi driver.
Coded according to an outline from Grant Likely; here is his
specification (accidentally swapped function names corrected):
It requires 3 things to be added to struct spi_master.
- 1 Mutex
- 1 spin lock
- 1 flag.
The mutex protects spi_sync, and provides sleeping "for free"
The spinlock protects the atomic spi_async call.
The flag is set when the lock is obtained, and checked while holding
the spinlock in spi_async(). If the flag is checked, then spi_async()
must fail immediately.
The current runtime API looks like this:
spi_async(struct spi_device*, struct spi_message*);
spi_sync(struct spi_device*, struct spi_message*);
The API needs to be extended to this:
spi_async(struct spi_device*, struct spi_message*)
spi_sync(struct spi_device*, struct spi_message*)
spi_bus_lock(struct spi_master*) /* although struct spi_device* might
be easier */
spi_bus_unlock(struct spi_master*)
spi_async_locked(struct spi_device*, struct spi_message*)
spi_sync_locked(struct spi_device*, struct spi_message*)
Drivers can only call the last two if they already hold the spi_master_lock().
spi_bus_lock() obtains the mutex, obtains the spin lock, sets the
flag, and releases the spin lock before returning. It doesn't even
need to sleep while waiting for "in-flight" spi_transactions to
complete because its purpose is to guarantee no additional
transactions are added. It does not guarantee that the bus is idle.
spi_bus_unlock() clears the flag and releases the mutex, which will
wake up any waiters.
The difference between spi_async() and spi_async_locked() is that the
locked version bypasses the check of the lock flag. Both versions
need to obtain the spinlock.
The difference between spi_sync() and spi_sync_locked() is that
spi_sync() must hold the mutex while enqueuing a new transfer.
spi_sync_locked() doesn't because the mutex is already held. Note
however that spi_sync must *not* continue to hold the mutex while
waiting for the transfer to complete, otherwise only one transfer
could be queued up at a time!
Almost no code needs to be written. The current spi_async() and
spi_sync() can probably be renamed to __spi_async() and __spi_sync()
so that spi_async(), spi_sync(), spi_async_locked() and
spi_sync_locked() can just become wrappers around the common code.
spi_sync() is protected by a mutex because it can sleep
spi_async() needs to be protected with a flag and a spinlock because
it can be called atomically and must not sleep
Signed-off-by: Ernst Schwab <eschwab@online.de>
[grant.likely@secretlab.ca: use spin_lock_irqsave()]
Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
Tested-by: Matt Fleming <matt@console-pimps.org>
Tested-by: Antonio Ospite <ospite@studenti.unina.it>
2010-06-29 00:49:29 +00:00
|
|
|
*/
|
|
|
|
int spi_sync_locked(struct spi_device *spi, struct spi_message *message)
|
|
|
|
{
|
|
|
|
return __spi_sync(spi, message, 1);
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(spi_sync_locked);
|
|
|
|
|
|
|
|
/**
|
|
|
|
* spi_bus_lock - obtain a lock for exclusive SPI bus usage
|
|
|
|
* @master: SPI bus master that should be locked for exclusive bus access
|
|
|
|
* Context: can sleep
|
|
|
|
*
|
|
|
|
* This call may only be used from a context that may sleep. The sleep
|
|
|
|
* is non-interruptible, and has no timeout.
|
|
|
|
*
|
|
|
|
* This call should be used by drivers that require exclusive access to the
|
|
|
|
* SPI bus. The SPI bus must be released by a spi_bus_unlock call when the
|
|
|
|
* exclusive access is over. Data transfer must be done by spi_sync_locked
|
|
|
|
* and spi_async_locked calls when the SPI bus lock is held.
|
|
|
|
*
|
2015-10-22 16:59:23 +00:00
|
|
|
* Return: always zero.
|
spi/mmc_spi: SPI bus locking API, using mutex
SPI bus locking API to allow exclusive access to the SPI bus, especially, but
not limited to, for the mmc_spi driver.
Coded according to an outline from Grant Likely; here is his
specification (accidentally swapped function names corrected):
It requires 3 things to be added to struct spi_master.
- 1 Mutex
- 1 spin lock
- 1 flag.
The mutex protects spi_sync, and provides sleeping "for free"
The spinlock protects the atomic spi_async call.
The flag is set when the lock is obtained, and checked while holding
the spinlock in spi_async(). If the flag is checked, then spi_async()
must fail immediately.
The current runtime API looks like this:
spi_async(struct spi_device*, struct spi_message*);
spi_sync(struct spi_device*, struct spi_message*);
The API needs to be extended to this:
spi_async(struct spi_device*, struct spi_message*)
spi_sync(struct spi_device*, struct spi_message*)
spi_bus_lock(struct spi_master*) /* although struct spi_device* might
be easier */
spi_bus_unlock(struct spi_master*)
spi_async_locked(struct spi_device*, struct spi_message*)
spi_sync_locked(struct spi_device*, struct spi_message*)
Drivers can only call the last two if they already hold the spi_master_lock().
spi_bus_lock() obtains the mutex, obtains the spin lock, sets the
flag, and releases the spin lock before returning. It doesn't even
need to sleep while waiting for "in-flight" spi_transactions to
complete because its purpose is to guarantee no additional
transactions are added. It does not guarantee that the bus is idle.
spi_bus_unlock() clears the flag and releases the mutex, which will
wake up any waiters.
The difference between spi_async() and spi_async_locked() is that the
locked version bypasses the check of the lock flag. Both versions
need to obtain the spinlock.
The difference between spi_sync() and spi_sync_locked() is that
spi_sync() must hold the mutex while enqueuing a new transfer.
spi_sync_locked() doesn't because the mutex is already held. Note
however that spi_sync must *not* continue to hold the mutex while
waiting for the transfer to complete, otherwise only one transfer
could be queued up at a time!
Almost no code needs to be written. The current spi_async() and
spi_sync() can probably be renamed to __spi_async() and __spi_sync()
so that spi_async(), spi_sync(), spi_async_locked() and
spi_sync_locked() can just become wrappers around the common code.
spi_sync() is protected by a mutex because it can sleep
spi_async() needs to be protected with a flag and a spinlock because
it can be called atomically and must not sleep
Signed-off-by: Ernst Schwab <eschwab@online.de>
[grant.likely@secretlab.ca: use spin_lock_irqsave()]
Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
Tested-by: Matt Fleming <matt@console-pimps.org>
Tested-by: Antonio Ospite <ospite@studenti.unina.it>
2010-06-29 00:49:29 +00:00
|
|
|
*/
|
|
|
|
int spi_bus_lock(struct spi_master *master)
|
|
|
|
{
|
|
|
|
unsigned long flags;
|
|
|
|
|
|
|
|
mutex_lock(&master->bus_lock_mutex);
|
|
|
|
|
|
|
|
spin_lock_irqsave(&master->bus_lock_spinlock, flags);
|
|
|
|
master->bus_lock_flag = 1;
|
|
|
|
spin_unlock_irqrestore(&master->bus_lock_spinlock, flags);
|
|
|
|
|
|
|
|
/* mutex remains locked until spi_bus_unlock is called */
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(spi_bus_lock);
|
|
|
|
|
|
|
|
/**
|
|
|
|
* spi_bus_unlock - release the lock for exclusive SPI bus usage
|
|
|
|
* @master: SPI bus master that was locked for exclusive bus access
|
|
|
|
* Context: can sleep
|
|
|
|
*
|
|
|
|
* This call may only be used from a context that may sleep. The sleep
|
|
|
|
* is non-interruptible, and has no timeout.
|
|
|
|
*
|
|
|
|
* This call releases an SPI bus lock previously obtained by an spi_bus_lock
|
|
|
|
* call.
|
|
|
|
*
|
2015-10-22 16:59:23 +00:00
|
|
|
* Return: always zero.
|
spi/mmc_spi: SPI bus locking API, using mutex
SPI bus locking API to allow exclusive access to the SPI bus, especially, but
not limited to, for the mmc_spi driver.
Coded according to an outline from Grant Likely; here is his
specification (accidentally swapped function names corrected):
It requires 3 things to be added to struct spi_master.
- 1 Mutex
- 1 spin lock
- 1 flag.
The mutex protects spi_sync, and provides sleeping "for free"
The spinlock protects the atomic spi_async call.
The flag is set when the lock is obtained, and checked while holding
the spinlock in spi_async(). If the flag is checked, then spi_async()
must fail immediately.
The current runtime API looks like this:
spi_async(struct spi_device*, struct spi_message*);
spi_sync(struct spi_device*, struct spi_message*);
The API needs to be extended to this:
spi_async(struct spi_device*, struct spi_message*)
spi_sync(struct spi_device*, struct spi_message*)
spi_bus_lock(struct spi_master*) /* although struct spi_device* might
be easier */
spi_bus_unlock(struct spi_master*)
spi_async_locked(struct spi_device*, struct spi_message*)
spi_sync_locked(struct spi_device*, struct spi_message*)
Drivers can only call the last two if they already hold the spi_master_lock().
spi_bus_lock() obtains the mutex, obtains the spin lock, sets the
flag, and releases the spin lock before returning. It doesn't even
need to sleep while waiting for "in-flight" spi_transactions to
complete because its purpose is to guarantee no additional
transactions are added. It does not guarantee that the bus is idle.
spi_bus_unlock() clears the flag and releases the mutex, which will
wake up any waiters.
The difference between spi_async() and spi_async_locked() is that the
locked version bypasses the check of the lock flag. Both versions
need to obtain the spinlock.
The difference between spi_sync() and spi_sync_locked() is that
spi_sync() must hold the mutex while enqueuing a new transfer.
spi_sync_locked() doesn't because the mutex is already held. Note
however that spi_sync must *not* continue to hold the mutex while
waiting for the transfer to complete, otherwise only one transfer
could be queued up at a time!
Almost no code needs to be written. The current spi_async() and
spi_sync() can probably be renamed to __spi_async() and __spi_sync()
so that spi_async(), spi_sync(), spi_async_locked() and
spi_sync_locked() can just become wrappers around the common code.
spi_sync() is protected by a mutex because it can sleep
spi_async() needs to be protected with a flag and a spinlock because
it can be called atomically and must not sleep
Signed-off-by: Ernst Schwab <eschwab@online.de>
[grant.likely@secretlab.ca: use spin_lock_irqsave()]
Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
Tested-by: Matt Fleming <matt@console-pimps.org>
Tested-by: Antonio Ospite <ospite@studenti.unina.it>
2010-06-29 00:49:29 +00:00
|
|
|
*/
|
|
|
|
int spi_bus_unlock(struct spi_master *master)
|
|
|
|
{
|
|
|
|
master->bus_lock_flag = 0;
|
|
|
|
|
|
|
|
mutex_unlock(&master->bus_lock_mutex);
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(spi_bus_unlock);
|
|
|
|
|
2006-04-02 18:37:40 +00:00
|
|
|
/* portable code must never pass more than 32 bytes */
|
2013-10-14 01:31:51 +00:00
|
|
|
#define SPI_BUFSIZ max(32, SMP_CACHE_BYTES)
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
|
|
|
|
static u8 *buf;
|
|
|
|
|
|
|
|
/**
|
|
|
|
* spi_write_then_read - SPI synchronous write followed by read
|
|
|
|
* @spi: device with which data will be exchanged
|
|
|
|
* @txbuf: data to be written (need not be dma-safe)
|
|
|
|
* @n_tx: size of txbuf, in bytes
|
2009-06-17 23:26:06 +00:00
|
|
|
* @rxbuf: buffer into which data will be read (need not be dma-safe)
|
|
|
|
* @n_rx: size of rxbuf, in bytes
|
2007-05-08 07:32:21 +00:00
|
|
|
* Context: can sleep
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
*
|
|
|
|
* This performs a half duplex MicroWire style transaction with the
|
|
|
|
* device, sending txbuf and then reading rxbuf. The return value
|
|
|
|
* is zero for success, else a negative errno status code.
|
2006-01-08 21:34:23 +00:00
|
|
|
* This call may only be used from a context that may sleep.
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
*
|
2006-01-08 21:34:25 +00:00
|
|
|
* Parameters to this routine are always copied using a small buffer;
|
2007-05-08 07:32:21 +00:00
|
|
|
* portable code should never use this for more than 32 bytes.
|
|
|
|
* Performance-sensitive or bulk transfer code should instead use
|
2006-01-08 21:34:25 +00:00
|
|
|
* spi_{async,sync}() calls with dma-safe buffers.
|
2015-10-22 16:59:23 +00:00
|
|
|
*
|
|
|
|
* Return: zero on success, else a negative error code.
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
*/
|
|
|
|
int spi_write_then_read(struct spi_device *spi,
|
2011-05-10 22:09:30 +00:00
|
|
|
const void *txbuf, unsigned n_tx,
|
|
|
|
void *rxbuf, unsigned n_rx)
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
{
|
2007-12-05 07:45:09 +00:00
|
|
|
static DEFINE_MUTEX(lock);
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
|
|
|
|
int status;
|
|
|
|
struct spi_message message;
|
2009-04-13 21:39:57 +00:00
|
|
|
struct spi_transfer x[2];
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
u8 *local_buf;
|
|
|
|
|
2012-12-02 03:54:25 +00:00
|
|
|
/* Use preallocated DMA-safe buffer if we can. We can't avoid
|
|
|
|
* copying here, (as a pure convenience thing), but we can
|
|
|
|
* keep heap costs out of the hot path unless someone else is
|
|
|
|
* using the pre-allocated buffer or the transfer is too large.
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
*/
|
2012-12-02 03:54:25 +00:00
|
|
|
if ((n_tx + n_rx) > SPI_BUFSIZ || !mutex_trylock(&lock)) {
|
2013-01-27 06:35:04 +00:00
|
|
|
local_buf = kmalloc(max((unsigned)SPI_BUFSIZ, n_tx + n_rx),
|
|
|
|
GFP_KERNEL | GFP_DMA);
|
2012-12-02 03:54:25 +00:00
|
|
|
if (!local_buf)
|
|
|
|
return -ENOMEM;
|
|
|
|
} else {
|
|
|
|
local_buf = buf;
|
|
|
|
}
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
|
2006-01-08 21:34:28 +00:00
|
|
|
spi_message_init(&message);
|
2013-10-14 01:31:51 +00:00
|
|
|
memset(x, 0, sizeof(x));
|
2009-04-13 21:39:57 +00:00
|
|
|
if (n_tx) {
|
|
|
|
x[0].len = n_tx;
|
|
|
|
spi_message_add_tail(&x[0], &message);
|
|
|
|
}
|
|
|
|
if (n_rx) {
|
|
|
|
x[1].len = n_rx;
|
|
|
|
spi_message_add_tail(&x[1], &message);
|
|
|
|
}
|
2006-01-08 21:34:28 +00:00
|
|
|
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
memcpy(local_buf, txbuf, n_tx);
|
2009-04-13 21:39:57 +00:00
|
|
|
x[0].tx_buf = local_buf;
|
|
|
|
x[1].rx_buf = local_buf + n_tx;
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
|
|
|
|
/* do the i/o */
|
|
|
|
status = spi_sync(spi, &message);
|
2007-12-05 07:45:10 +00:00
|
|
|
if (status == 0)
|
2009-04-13 21:39:57 +00:00
|
|
|
memcpy(rxbuf, x[1].rx_buf, n_rx);
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
|
2009-04-13 21:39:57 +00:00
|
|
|
if (x[0].tx_buf == buf)
|
2007-12-05 07:45:09 +00:00
|
|
|
mutex_unlock(&lock);
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
else
|
|
|
|
kfree(local_buf);
|
|
|
|
|
|
|
|
return status;
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(spi_write_then_read);
|
|
|
|
|
|
|
|
/*-------------------------------------------------------------------------*/
|
|
|
|
|
2014-10-28 20:36:05 +00:00
|
|
|
#if IS_ENABLED(CONFIG_OF_DYNAMIC)
|
|
|
|
static int __spi_of_device_match(struct device *dev, void *data)
|
|
|
|
{
|
|
|
|
return dev->of_node == data;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* must call put_device() when done with returned spi_device device */
|
|
|
|
static struct spi_device *of_find_spi_device_by_node(struct device_node *node)
|
|
|
|
{
|
|
|
|
struct device *dev = bus_find_device(&spi_bus_type, NULL, node,
|
|
|
|
__spi_of_device_match);
|
|
|
|
return dev ? to_spi_device(dev) : NULL;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int __spi_of_master_match(struct device *dev, const void *data)
|
|
|
|
{
|
|
|
|
return dev->of_node == data;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* the spi masters are not using spi_bus, so we find it with another way */
|
|
|
|
static struct spi_master *of_find_spi_master_by_node(struct device_node *node)
|
|
|
|
{
|
|
|
|
struct device *dev;
|
|
|
|
|
|
|
|
dev = class_find_device(&spi_master_class, NULL, node,
|
|
|
|
__spi_of_master_match);
|
|
|
|
if (!dev)
|
|
|
|
return NULL;
|
|
|
|
|
|
|
|
/* reference got in class_find_device */
|
|
|
|
return container_of(dev, struct spi_master, dev);
|
|
|
|
}
|
|
|
|
|
|
|
|
static int of_spi_notify(struct notifier_block *nb, unsigned long action,
|
|
|
|
void *arg)
|
|
|
|
{
|
|
|
|
struct of_reconfig_data *rd = arg;
|
|
|
|
struct spi_master *master;
|
|
|
|
struct spi_device *spi;
|
|
|
|
|
|
|
|
switch (of_reconfig_get_state_change(action, arg)) {
|
|
|
|
case OF_RECONFIG_CHANGE_ADD:
|
|
|
|
master = of_find_spi_master_by_node(rd->dn->parent);
|
|
|
|
if (master == NULL)
|
|
|
|
return NOTIFY_OK; /* not for us */
|
|
|
|
|
|
|
|
spi = of_register_spi_device(master, rd->dn);
|
|
|
|
put_device(&master->dev);
|
|
|
|
|
|
|
|
if (IS_ERR(spi)) {
|
|
|
|
pr_err("%s: failed to create for '%s'\n",
|
|
|
|
__func__, rd->dn->full_name);
|
|
|
|
return notifier_from_errno(PTR_ERR(spi));
|
|
|
|
}
|
|
|
|
break;
|
|
|
|
|
|
|
|
case OF_RECONFIG_CHANGE_REMOVE:
|
|
|
|
/* find our device by node */
|
|
|
|
spi = of_find_spi_device_by_node(rd->dn);
|
|
|
|
if (spi == NULL)
|
|
|
|
return NOTIFY_OK; /* no? not meant for us */
|
|
|
|
|
|
|
|
/* unregister takes one ref away */
|
|
|
|
spi_unregister_device(spi);
|
|
|
|
|
|
|
|
/* and put the reference of the find */
|
|
|
|
put_device(&spi->dev);
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
return NOTIFY_OK;
|
|
|
|
}
|
|
|
|
|
|
|
|
static struct notifier_block spi_of_notifier = {
|
|
|
|
.notifier_call = of_spi_notify,
|
|
|
|
};
|
|
|
|
#else /* IS_ENABLED(CONFIG_OF_DYNAMIC) */
|
|
|
|
extern struct notifier_block spi_of_notifier;
|
|
|
|
#endif /* IS_ENABLED(CONFIG_OF_DYNAMIC) */
|
|
|
|
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
static int __init spi_init(void)
|
|
|
|
{
|
2006-01-08 21:34:23 +00:00
|
|
|
int status;
|
|
|
|
|
2006-12-07 04:33:17 +00:00
|
|
|
buf = kmalloc(SPI_BUFSIZ, GFP_KERNEL);
|
2006-01-08 21:34:23 +00:00
|
|
|
if (!buf) {
|
|
|
|
status = -ENOMEM;
|
|
|
|
goto err0;
|
|
|
|
}
|
|
|
|
|
|
|
|
status = bus_register(&spi_bus_type);
|
|
|
|
if (status < 0)
|
|
|
|
goto err1;
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
|
2006-01-08 21:34:23 +00:00
|
|
|
status = class_register(&spi_master_class);
|
|
|
|
if (status < 0)
|
|
|
|
goto err2;
|
2014-10-28 20:36:05 +00:00
|
|
|
|
2014-11-26 22:13:57 +00:00
|
|
|
if (IS_ENABLED(CONFIG_OF_DYNAMIC))
|
2014-10-28 20:36:05 +00:00
|
|
|
WARN_ON(of_reconfig_notifier_register(&spi_of_notifier));
|
|
|
|
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
return 0;
|
2006-01-08 21:34:23 +00:00
|
|
|
|
|
|
|
err2:
|
|
|
|
bus_unregister(&spi_bus_type);
|
|
|
|
err1:
|
|
|
|
kfree(buf);
|
|
|
|
buf = NULL;
|
|
|
|
err0:
|
|
|
|
return status;
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
}
|
2006-01-08 21:34:23 +00:00
|
|
|
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
/* board_info is normally registered in arch_initcall(),
|
|
|
|
* but even essential drivers wait till later
|
2006-01-08 21:34:23 +00:00
|
|
|
*
|
|
|
|
* REVISIT only boardinfo really needs static linking. the rest (device and
|
|
|
|
* driver registration) _could_ be dynamically linked (modular) ... costs
|
|
|
|
* include needing to have boardinfo data structures be much more public.
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
*/
|
2008-10-16 05:02:46 +00:00
|
|
|
postcore_initcall(spi_init);
|
[PATCH] spi: simple SPI framework
This is the core of a small SPI framework, implementing the model of a
queue of messages which complete asynchronously (with thin synchronous
wrappers on top).
- It's still less than 2KB of ".text" (ARM). If there's got to be a
mid-layer for something so simple, that's the right size budget. :)
- The guts use board-specific SPI device tables to build the driver
model tree. (Hardware probing is rarely an option.)
- This version of Kconfig includes no drivers. At this writing there
are two known master controller drivers (PXA/SSP, OMAP MicroWire)
and three protocol drivers (CS8415a, ADS7846, DataFlash) with LKML
mentions of other drivers in development.
- No userspace API. There are several implementations to compare.
Implement them like any other driver, and bind them with sysfs.
The changes from last version posted to LKML (on 11-Nov-2005) are minor,
and include:
- One bugfix (removes a FIXME), with the visible effect of making device
names be "spiB.C" where B is the bus number and C is the chipselect.
- The "caller provides DMA mappings" mechanism now has kerneldoc, for
DMA drivers that want to be fancy.
- Hey, the framework init can be subsys_init. Even though board init
logic fires earlier, at arch_init ... since the framework init is
for driver support, and the board init support uses static init.
- Various additional spec/doc clarifications based on discussions
with other folk. It adds a brief "thank you" at the end, for folk
who've helped nudge this framework into existence.
As I've said before, I think that "protocol tweaking" is the main support
that this driver framework will need to evolve.
From: Mark Underwood <basicmark@yahoo.com>
Update the SPI framework to remove a potential priority inversion case by
reverting to kmalloc if the pre-allocated DMA-safe buffer isn't available.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
2006-01-08 21:34:19 +00:00
|
|
|
|