linux/drivers/serial/mpc52xx_uart.c

840 lines
22 KiB
C
Raw Normal View History

/*
* drivers/serial/mpc52xx_uart.c
*
* Driver for the PSC of the Freescale MPC52xx PSCs configured as UARTs.
*
* FIXME According to the usermanual the status bits in the status register
* are only updated when the peripherals access the FIFO and not when the
* CPU access them. So since we use this bits to know when we stop writing
* and reading, they may not be updated in-time and a race condition may
* exists. But I haven't be able to prove this and I don't care. But if
* any problem arises, it might worth checking. The TX/RX FIFO Stats
* registers should be used in addition.
* Update: Actually, they seem updated ... At least the bits we use.
*
*
* Maintainer : Sylvain Munaut <tnt@246tNt.com>
*
* Some of the code has been inspired/copied from the 2.4 code written
* by Dale Farnsworth <dfarnsworth@mvista.com>.
*
* Copyright (C) 2004-2005 Sylvain Munaut <tnt@246tNt.com>
* Copyright (C) 2003 MontaVista, Software, Inc.
*
* This file is licensed under the terms of the GNU General Public License
* version 2. This program is licensed "as is" without any warranty of any
* kind, whether express or implied.
*/
/* Platform device Usage :
*
* Since PSCs can have multiple function, the correct driver for each one
* is selected by calling mpc52xx_match_psc_function(...). The function
* handled by this driver is "uart".
*
* The driver init all necessary registers to place the PSC in uart mode without
* DCD. However, the pin multiplexing aren't changed and should be set either
* by the bootloader or in the platform init code.
*
* The idx field must be equal to the PSC index ( e.g. 0 for PSC1, 1 for PSC2,
* and so on). So the PSC1 is mapped to /dev/ttyPSC0, PSC2 to /dev/ttyPSC1 and
* so on. But be warned, it's an ABSOLUTE REQUIREMENT ! This is needed mainly
* fpr the console code : without this 1:1 mapping, at early boot time, when we
* are parsing the kernel args console=ttyPSC?, we wouldn't know which PSC it
* will be mapped to.
*/
#include <linux/platform_device.h>
#include <linux/module.h>
#include <linux/tty.h>
#include <linux/serial.h>
#include <linux/sysrq.h>
#include <linux/console.h>
#include <asm/delay.h>
#include <asm/io.h>
#include <asm/mpc52xx.h>
#include <asm/mpc52xx_psc.h>
#if defined(CONFIG_SERIAL_MPC52xx_CONSOLE) && defined(CONFIG_MAGIC_SYSRQ)
#define SUPPORT_SYSRQ
#endif
#include <linux/serial_core.h>
/* We've been assigned a range on the "Low-density serial ports" major */
#define SERIAL_PSC_MAJOR 204
#define SERIAL_PSC_MINOR 148
#define ISR_PASS_LIMIT 256 /* Max number of iteration in the interrupt */
static struct uart_port mpc52xx_uart_ports[MPC52xx_PSC_MAXNUM];
/* Rem: - We use the read_status_mask as a shadow of
* psc->mpc52xx_psc_imr
* - It's important that is array is all zero on start as we
* use it to know if it's initialized or not ! If it's not sure
* it's cleared, then a memset(...,0,...) should be added to
* the console_init
*/
#define PSC(port) ((struct mpc52xx_psc __iomem *)((port)->membase))
/* Forward declaration of the interruption handling routine */
IRQ: Maintain regs pointer globally rather than passing to IRQ handlers Maintain a per-CPU global "struct pt_regs *" variable which can be used instead of passing regs around manually through all ~1800 interrupt handlers in the Linux kernel. The regs pointer is used in few places, but it potentially costs both stack space and code to pass it around. On the FRV arch, removing the regs parameter from all the genirq function results in a 20% speed up of the IRQ exit path (ie: from leaving timer_interrupt() to leaving do_IRQ()). Where appropriate, an arch may override the generic storage facility and do something different with the variable. On FRV, for instance, the address is maintained in GR28 at all times inside the kernel as part of general exception handling. Having looked over the code, it appears that the parameter may be handed down through up to twenty or so layers of functions. Consider a USB character device attached to a USB hub, attached to a USB controller that posts its interrupts through a cascaded auxiliary interrupt controller. A character device driver may want to pass regs to the sysrq handler through the input layer which adds another few layers of parameter passing. I've build this code with allyesconfig for x86_64 and i386. I've runtested the main part of the code on FRV and i386, though I can't test most of the drivers. I've also done partial conversion for powerpc and MIPS - these at least compile with minimal configurations. This will affect all archs. Mostly the changes should be relatively easy. Take do_IRQ(), store the regs pointer at the beginning, saving the old one: struct pt_regs *old_regs = set_irq_regs(regs); And put the old one back at the end: set_irq_regs(old_regs); Don't pass regs through to generic_handle_irq() or __do_IRQ(). In timer_interrupt(), this sort of change will be necessary: - update_process_times(user_mode(regs)); - profile_tick(CPU_PROFILING, regs); + update_process_times(user_mode(get_irq_regs())); + profile_tick(CPU_PROFILING); I'd like to move update_process_times()'s use of get_irq_regs() into itself, except that i386, alone of the archs, uses something other than user_mode(). Some notes on the interrupt handling in the drivers: (*) input_dev() is now gone entirely. The regs pointer is no longer stored in the input_dev struct. (*) finish_unlinks() in drivers/usb/host/ohci-q.c needs checking. It does something different depending on whether it's been supplied with a regs pointer or not. (*) Various IRQ handler function pointers have been moved to type irq_handler_t. Signed-Off-By: David Howells <dhowells@redhat.com> (cherry picked from 1b16e7ac850969f38b375e511e3fa2f474a33867 commit)
2006-10-05 13:55:46 +00:00
static irqreturn_t mpc52xx_uart_int(int irq,void *dev_id);
/* Simple macro to test if a port is console or not. This one is taken
* for serial_core.c and maybe should be moved to serial_core.h ? */
#ifdef CONFIG_SERIAL_CORE_CONSOLE
#define uart_console(port) ((port)->cons && (port)->cons->index == (port)->line)
#else
#define uart_console(port) (0)
#endif
/* ======================================================================== */
/* UART operations */
/* ======================================================================== */
static unsigned int
mpc52xx_uart_tx_empty(struct uart_port *port)
{
int status = in_be16(&PSC(port)->mpc52xx_psc_status);
return (status & MPC52xx_PSC_SR_TXEMP) ? TIOCSER_TEMT : 0;
}
static void
mpc52xx_uart_set_mctrl(struct uart_port *port, unsigned int mctrl)
{
/* Not implemented */
}
static unsigned int
mpc52xx_uart_get_mctrl(struct uart_port *port)
{
/* Not implemented */
return TIOCM_CTS | TIOCM_DSR | TIOCM_CAR;
}
static void
mpc52xx_uart_stop_tx(struct uart_port *port)
{
/* port->lock taken by caller */
port->read_status_mask &= ~MPC52xx_PSC_IMR_TXRDY;
out_be16(&PSC(port)->mpc52xx_psc_imr,port->read_status_mask);
}
static void
mpc52xx_uart_start_tx(struct uart_port *port)
{
/* port->lock taken by caller */
port->read_status_mask |= MPC52xx_PSC_IMR_TXRDY;
out_be16(&PSC(port)->mpc52xx_psc_imr,port->read_status_mask);
}
static void
mpc52xx_uart_send_xchar(struct uart_port *port, char ch)
{
unsigned long flags;
spin_lock_irqsave(&port->lock, flags);
port->x_char = ch;
if (ch) {
/* Make sure tx interrupts are on */
/* Truly necessary ??? They should be anyway */
port->read_status_mask |= MPC52xx_PSC_IMR_TXRDY;
out_be16(&PSC(port)->mpc52xx_psc_imr,port->read_status_mask);
}
spin_unlock_irqrestore(&port->lock, flags);
}
static void
mpc52xx_uart_stop_rx(struct uart_port *port)
{
/* port->lock taken by caller */
port->read_status_mask &= ~MPC52xx_PSC_IMR_RXRDY;
out_be16(&PSC(port)->mpc52xx_psc_imr,port->read_status_mask);
}
static void
mpc52xx_uart_enable_ms(struct uart_port *port)
{
/* Not implemented */
}
static void
mpc52xx_uart_break_ctl(struct uart_port *port, int ctl)
{
unsigned long flags;
spin_lock_irqsave(&port->lock, flags);
if ( ctl == -1 )
out_8(&PSC(port)->command,MPC52xx_PSC_START_BRK);
else
out_8(&PSC(port)->command,MPC52xx_PSC_STOP_BRK);
spin_unlock_irqrestore(&port->lock, flags);
}
static int
mpc52xx_uart_startup(struct uart_port *port)
{
struct mpc52xx_psc __iomem *psc = PSC(port);
int ret;
/* Request IRQ */
ret = request_irq(port->irq, mpc52xx_uart_int,
IRQF_DISABLED | IRQF_SAMPLE_RANDOM, "mpc52xx_psc_uart", port);
if (ret)
return ret;
/* Reset/activate the port, clear and enable interrupts */
out_8(&psc->command,MPC52xx_PSC_RST_RX);
out_8(&psc->command,MPC52xx_PSC_RST_TX);
out_be32(&psc->sicr,0); /* UART mode DCD ignored */
out_be16(&psc->mpc52xx_psc_clock_select, 0xdd00); /* /16 prescaler on */
out_8(&psc->rfcntl, 0x00);
out_be16(&psc->rfalarm, 0x1ff);
out_8(&psc->tfcntl, 0x07);
out_be16(&psc->tfalarm, 0x80);
port->read_status_mask |= MPC52xx_PSC_IMR_RXRDY | MPC52xx_PSC_IMR_TXRDY;
out_be16(&psc->mpc52xx_psc_imr,port->read_status_mask);
out_8(&psc->command,MPC52xx_PSC_TX_ENABLE);
out_8(&psc->command,MPC52xx_PSC_RX_ENABLE);
return 0;
}
static void
mpc52xx_uart_shutdown(struct uart_port *port)
{
struct mpc52xx_psc __iomem *psc = PSC(port);
/* Shut down the port, interrupt and all */
out_8(&psc->command,MPC52xx_PSC_RST_RX);
out_8(&psc->command,MPC52xx_PSC_RST_TX);
port->read_status_mask = 0;
out_be16(&psc->mpc52xx_psc_imr,port->read_status_mask);
/* Release interrupt */
free_irq(port->irq, port);
}
static void
mpc52xx_uart_set_termios(struct uart_port *port, struct termios *new,
struct termios *old)
{
struct mpc52xx_psc __iomem *psc = PSC(port);
unsigned long flags;
unsigned char mr1, mr2;
unsigned short ctr;
unsigned int j, baud, quot;
/* Prepare what we're gonna write */
mr1 = 0;
switch (new->c_cflag & CSIZE) {
case CS5: mr1 |= MPC52xx_PSC_MODE_5_BITS;
break;
case CS6: mr1 |= MPC52xx_PSC_MODE_6_BITS;
break;
case CS7: mr1 |= MPC52xx_PSC_MODE_7_BITS;
break;
case CS8:
default: mr1 |= MPC52xx_PSC_MODE_8_BITS;
}
if (new->c_cflag & PARENB) {
mr1 |= (new->c_cflag & PARODD) ?
MPC52xx_PSC_MODE_PARODD : MPC52xx_PSC_MODE_PAREVEN;
} else
mr1 |= MPC52xx_PSC_MODE_PARNONE;
mr2 = 0;
if (new->c_cflag & CSTOPB)
mr2 |= MPC52xx_PSC_MODE_TWO_STOP;
else
mr2 |= ((new->c_cflag & CSIZE) == CS5) ?
MPC52xx_PSC_MODE_ONE_STOP_5_BITS :
MPC52xx_PSC_MODE_ONE_STOP;
baud = uart_get_baud_rate(port, new, old, 0, port->uartclk/16);
quot = uart_get_divisor(port, baud);
ctr = quot & 0xffff;
/* Get the lock */
spin_lock_irqsave(&port->lock, flags);
/* Update the per-port timeout */
uart_update_timeout(port, new->c_cflag, baud);
/* Do our best to flush TX & RX, so we don't loose anything */
/* But we don't wait indefinitly ! */
j = 5000000; /* Maximum wait */
/* FIXME Can't receive chars since set_termios might be called at early
* boot for the console, all stuff is not yet ready to receive at that
* time and that just makes the kernel oops */
/* while (j-- && mpc52xx_uart_int_rx_chars(port)); */
while (!(in_be16(&psc->mpc52xx_psc_status) & MPC52xx_PSC_SR_TXEMP) &&
--j)
udelay(1);
if (!j)
printk( KERN_ERR "mpc52xx_uart.c: "
"Unable to flush RX & TX fifos in-time in set_termios."
"Some chars may have been lost.\n" );
/* Reset the TX & RX */
out_8(&psc->command,MPC52xx_PSC_RST_RX);
out_8(&psc->command,MPC52xx_PSC_RST_TX);
/* Send new mode settings */
out_8(&psc->command,MPC52xx_PSC_SEL_MODE_REG_1);
out_8(&psc->mode,mr1);
out_8(&psc->mode,mr2);
out_8(&psc->ctur,ctr >> 8);
out_8(&psc->ctlr,ctr & 0xff);
/* Reenable TX & RX */
out_8(&psc->command,MPC52xx_PSC_TX_ENABLE);
out_8(&psc->command,MPC52xx_PSC_RX_ENABLE);
/* We're all set, release the lock */
spin_unlock_irqrestore(&port->lock, flags);
}
static const char *
mpc52xx_uart_type(struct uart_port *port)
{
return port->type == PORT_MPC52xx ? "MPC52xx PSC" : NULL;
}
static void
mpc52xx_uart_release_port(struct uart_port *port)
{
if (port->flags & UPF_IOREMAP) { /* remapped by us ? */
iounmap(port->membase);
port->membase = NULL;
}
release_mem_region(port->mapbase, MPC52xx_PSC_SIZE);
}
static int
mpc52xx_uart_request_port(struct uart_port *port)
{
int err;
if (port->flags & UPF_IOREMAP) /* Need to remap ? */
port->membase = ioremap(port->mapbase, MPC52xx_PSC_SIZE);
if (!port->membase)
return -EINVAL;
err = request_mem_region(port->mapbase, MPC52xx_PSC_SIZE,
"mpc52xx_psc_uart") != NULL ? 0 : -EBUSY;
if (err && (port->flags & UPF_IOREMAP)) {
iounmap(port->membase);
port->membase = NULL;
}
return err;
}
static void
mpc52xx_uart_config_port(struct uart_port *port, int flags)
{
if ( (flags & UART_CONFIG_TYPE) &&
(mpc52xx_uart_request_port(port) == 0) )
port->type = PORT_MPC52xx;
}
static int
mpc52xx_uart_verify_port(struct uart_port *port, struct serial_struct *ser)
{
if ( ser->type != PORT_UNKNOWN && ser->type != PORT_MPC52xx )
return -EINVAL;
if ( (ser->irq != port->irq) ||
(ser->io_type != SERIAL_IO_MEM) ||
(ser->baud_base != port->uartclk) ||
(ser->iomem_base != (void*)port->mapbase) ||
(ser->hub6 != 0 ) )
return -EINVAL;
return 0;
}
static struct uart_ops mpc52xx_uart_ops = {
.tx_empty = mpc52xx_uart_tx_empty,
.set_mctrl = mpc52xx_uart_set_mctrl,
.get_mctrl = mpc52xx_uart_get_mctrl,
.stop_tx = mpc52xx_uart_stop_tx,
.start_tx = mpc52xx_uart_start_tx,
.send_xchar = mpc52xx_uart_send_xchar,
.stop_rx = mpc52xx_uart_stop_rx,
.enable_ms = mpc52xx_uart_enable_ms,
.break_ctl = mpc52xx_uart_break_ctl,
.startup = mpc52xx_uart_startup,
.shutdown = mpc52xx_uart_shutdown,
.set_termios = mpc52xx_uart_set_termios,
/* .pm = mpc52xx_uart_pm, Not supported yet */
/* .set_wake = mpc52xx_uart_set_wake, Not supported yet */
.type = mpc52xx_uart_type,
.release_port = mpc52xx_uart_release_port,
.request_port = mpc52xx_uart_request_port,
.config_port = mpc52xx_uart_config_port,
.verify_port = mpc52xx_uart_verify_port
};
/* ======================================================================== */
/* Interrupt handling */
/* ======================================================================== */
static inline int
IRQ: Maintain regs pointer globally rather than passing to IRQ handlers Maintain a per-CPU global "struct pt_regs *" variable which can be used instead of passing regs around manually through all ~1800 interrupt handlers in the Linux kernel. The regs pointer is used in few places, but it potentially costs both stack space and code to pass it around. On the FRV arch, removing the regs parameter from all the genirq function results in a 20% speed up of the IRQ exit path (ie: from leaving timer_interrupt() to leaving do_IRQ()). Where appropriate, an arch may override the generic storage facility and do something different with the variable. On FRV, for instance, the address is maintained in GR28 at all times inside the kernel as part of general exception handling. Having looked over the code, it appears that the parameter may be handed down through up to twenty or so layers of functions. Consider a USB character device attached to a USB hub, attached to a USB controller that posts its interrupts through a cascaded auxiliary interrupt controller. A character device driver may want to pass regs to the sysrq handler through the input layer which adds another few layers of parameter passing. I've build this code with allyesconfig for x86_64 and i386. I've runtested the main part of the code on FRV and i386, though I can't test most of the drivers. I've also done partial conversion for powerpc and MIPS - these at least compile with minimal configurations. This will affect all archs. Mostly the changes should be relatively easy. Take do_IRQ(), store the regs pointer at the beginning, saving the old one: struct pt_regs *old_regs = set_irq_regs(regs); And put the old one back at the end: set_irq_regs(old_regs); Don't pass regs through to generic_handle_irq() or __do_IRQ(). In timer_interrupt(), this sort of change will be necessary: - update_process_times(user_mode(regs)); - profile_tick(CPU_PROFILING, regs); + update_process_times(user_mode(get_irq_regs())); + profile_tick(CPU_PROFILING); I'd like to move update_process_times()'s use of get_irq_regs() into itself, except that i386, alone of the archs, uses something other than user_mode(). Some notes on the interrupt handling in the drivers: (*) input_dev() is now gone entirely. The regs pointer is no longer stored in the input_dev struct. (*) finish_unlinks() in drivers/usb/host/ohci-q.c needs checking. It does something different depending on whether it's been supplied with a regs pointer or not. (*) Various IRQ handler function pointers have been moved to type irq_handler_t. Signed-Off-By: David Howells <dhowells@redhat.com> (cherry picked from 1b16e7ac850969f38b375e511e3fa2f474a33867 commit)
2006-10-05 13:55:46 +00:00
mpc52xx_uart_int_rx_chars(struct uart_port *port)
{
struct tty_struct *tty = port->info->tty;
[PATCH] TTY layer buffering revamp The API and code have been through various bits of initial review by serial driver people but they definitely need to live somewhere for a while so the unconverted drivers can get knocked into shape, existing drivers that have been updated can be better tuned and bugs whacked out. This replaces the tty flip buffers with kmalloc objects in rings. In the normal situation for an IRQ driven serial port at typical speeds the behaviour is pretty much the same, two buffers end up allocated and the kernel cycles between them as before. When there are delays or at high speed we now behave far better as the buffer pool can grow a bit rather than lose characters. This also means that we can operate at higher speeds reliably. For drivers that receive characters in blocks (DMA based, USB and especially virtualisation) the layer allows a lot of driver specific code that works around the tty layer with private secondary queues to be removed. The IBM folks need this sort of layer, the smart serial port people do, the virtualisers do (because a virtualised tty typically operates at infinite speed rather than emulating 9600 baud). Finally many drivers had invalid and unsafe attempts to avoid buffer overflows by directly invoking tty methods extracted out of the innards of work queue structs. These are no longer needed and all go away. That fixes various random hangs with serial ports on overflow. The other change in here is to optimise the receive_room path that is used by some callers. It turns out that only one ldisc uses receive room except asa constant and it updates it far far less than the value is read. We thus make it a variable not a function call. I expect the code to contain bugs due to the size alone but I'll be watching and squashing them and feeding out new patches as it goes. Because the buffers now dynamically expand you should only run out of buffering when the kernel runs out of memory for real. That means a lot of the horrible hacks high performance drivers used to do just aren't needed any more. Description: tty_insert_flip_char is an old API and continues to work as before, as does tty_flip_buffer_push() [this is why many drivers dont need modification]. It does now also return the number of chars inserted There are also tty_buffer_request_room(tty, len) which asks for a buffer block of the length requested and returns the space found. This improves efficiency with hardware that knows how much to transfer. and tty_insert_flip_string_flags(tty, str, flags, len) to insert a string of characters and flags For a smart interface the usual code is len = tty_request_buffer_room(tty, amount_hardware_says); tty_insert_flip_string(tty, buffer_from_card, len); More description! At the moment tty buffers are attached directly to the tty. This is causing a lot of the problems related to tty layer locking, also problems at high speed and also with bursty data (such as occurs in virtualised environments) I'm working on ripping out the flip buffers and replacing them with a pool of dynamically allocated buffers. This allows both for old style "byte I/O" devices and also helps virtualisation and smart devices where large blocks of data suddenely materialise and need storing. So far so good. Lots of drivers reference tty->flip.*. Several of them also call directly and unsafely into function pointers it provides. This will all break. Most drivers can use tty_insert_flip_char which can be kept as an API but others need more. At the moment I've added the following interfaces, if people think more will be needed now is a good time to say int tty_buffer_request_room(tty, size) Try and ensure at least size bytes are available, returns actual room (may be zero). At the moment it just uses the flipbuf space but that will change. Repeated calls without characters being added are not cumulative. (ie if you call it with 1, 1, 1, and then 4 you'll have four characters of space. The other functions will also try and grow buffers in future but this will be a more efficient way when you know block sizes. int tty_insert_flip_char(tty, ch, flag) As before insert a character if there is room. Now returns 1 for success, 0 for failure. int tty_insert_flip_string(tty, str, len) Insert a block of non error characters. Returns the number inserted. int tty_prepare_flip_string(tty, strptr, len) Adjust the buffer to allow len characters to be added. Returns a buffer pointer in strptr and the length available. This allows for hardware that needs to use functions like insl or mencpy_fromio. Signed-off-by: Alan Cox <alan@redhat.com> Cc: Paul Fulghum <paulkf@microgate.com> Signed-off-by: Hirokazu Takata <takata@linux-m32r.org> Signed-off-by: Serge Hallyn <serue@us.ibm.com> Signed-off-by: Jeff Dike <jdike@addtoit.com> Signed-off-by: John Hawkes <hawkes@sgi.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by: Adrian Bunk <bunk@stusta.de> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-01-10 04:54:13 +00:00
unsigned char ch, flag;
unsigned short status;
/* While we can read, do so ! */
while ( (status = in_be16(&PSC(port)->mpc52xx_psc_status)) &
MPC52xx_PSC_SR_RXRDY) {
/* Get the char */
ch = in_8(&PSC(port)->mpc52xx_psc_buffer_8);
/* Handle sysreq char */
#ifdef SUPPORT_SYSRQ
IRQ: Maintain regs pointer globally rather than passing to IRQ handlers Maintain a per-CPU global "struct pt_regs *" variable which can be used instead of passing regs around manually through all ~1800 interrupt handlers in the Linux kernel. The regs pointer is used in few places, but it potentially costs both stack space and code to pass it around. On the FRV arch, removing the regs parameter from all the genirq function results in a 20% speed up of the IRQ exit path (ie: from leaving timer_interrupt() to leaving do_IRQ()). Where appropriate, an arch may override the generic storage facility and do something different with the variable. On FRV, for instance, the address is maintained in GR28 at all times inside the kernel as part of general exception handling. Having looked over the code, it appears that the parameter may be handed down through up to twenty or so layers of functions. Consider a USB character device attached to a USB hub, attached to a USB controller that posts its interrupts through a cascaded auxiliary interrupt controller. A character device driver may want to pass regs to the sysrq handler through the input layer which adds another few layers of parameter passing. I've build this code with allyesconfig for x86_64 and i386. I've runtested the main part of the code on FRV and i386, though I can't test most of the drivers. I've also done partial conversion for powerpc and MIPS - these at least compile with minimal configurations. This will affect all archs. Mostly the changes should be relatively easy. Take do_IRQ(), store the regs pointer at the beginning, saving the old one: struct pt_regs *old_regs = set_irq_regs(regs); And put the old one back at the end: set_irq_regs(old_regs); Don't pass regs through to generic_handle_irq() or __do_IRQ(). In timer_interrupt(), this sort of change will be necessary: - update_process_times(user_mode(regs)); - profile_tick(CPU_PROFILING, regs); + update_process_times(user_mode(get_irq_regs())); + profile_tick(CPU_PROFILING); I'd like to move update_process_times()'s use of get_irq_regs() into itself, except that i386, alone of the archs, uses something other than user_mode(). Some notes on the interrupt handling in the drivers: (*) input_dev() is now gone entirely. The regs pointer is no longer stored in the input_dev struct. (*) finish_unlinks() in drivers/usb/host/ohci-q.c needs checking. It does something different depending on whether it's been supplied with a regs pointer or not. (*) Various IRQ handler function pointers have been moved to type irq_handler_t. Signed-Off-By: David Howells <dhowells@redhat.com> (cherry picked from 1b16e7ac850969f38b375e511e3fa2f474a33867 commit)
2006-10-05 13:55:46 +00:00
if (uart_handle_sysrq_char(port, ch)) {
port->sysrq = 0;
continue;
}
#endif
/* Store it */
[PATCH] TTY layer buffering revamp The API and code have been through various bits of initial review by serial driver people but they definitely need to live somewhere for a while so the unconverted drivers can get knocked into shape, existing drivers that have been updated can be better tuned and bugs whacked out. This replaces the tty flip buffers with kmalloc objects in rings. In the normal situation for an IRQ driven serial port at typical speeds the behaviour is pretty much the same, two buffers end up allocated and the kernel cycles between them as before. When there are delays or at high speed we now behave far better as the buffer pool can grow a bit rather than lose characters. This also means that we can operate at higher speeds reliably. For drivers that receive characters in blocks (DMA based, USB and especially virtualisation) the layer allows a lot of driver specific code that works around the tty layer with private secondary queues to be removed. The IBM folks need this sort of layer, the smart serial port people do, the virtualisers do (because a virtualised tty typically operates at infinite speed rather than emulating 9600 baud). Finally many drivers had invalid and unsafe attempts to avoid buffer overflows by directly invoking tty methods extracted out of the innards of work queue structs. These are no longer needed and all go away. That fixes various random hangs with serial ports on overflow. The other change in here is to optimise the receive_room path that is used by some callers. It turns out that only one ldisc uses receive room except asa constant and it updates it far far less than the value is read. We thus make it a variable not a function call. I expect the code to contain bugs due to the size alone but I'll be watching and squashing them and feeding out new patches as it goes. Because the buffers now dynamically expand you should only run out of buffering when the kernel runs out of memory for real. That means a lot of the horrible hacks high performance drivers used to do just aren't needed any more. Description: tty_insert_flip_char is an old API and continues to work as before, as does tty_flip_buffer_push() [this is why many drivers dont need modification]. It does now also return the number of chars inserted There are also tty_buffer_request_room(tty, len) which asks for a buffer block of the length requested and returns the space found. This improves efficiency with hardware that knows how much to transfer. and tty_insert_flip_string_flags(tty, str, flags, len) to insert a string of characters and flags For a smart interface the usual code is len = tty_request_buffer_room(tty, amount_hardware_says); tty_insert_flip_string(tty, buffer_from_card, len); More description! At the moment tty buffers are attached directly to the tty. This is causing a lot of the problems related to tty layer locking, also problems at high speed and also with bursty data (such as occurs in virtualised environments) I'm working on ripping out the flip buffers and replacing them with a pool of dynamically allocated buffers. This allows both for old style "byte I/O" devices and also helps virtualisation and smart devices where large blocks of data suddenely materialise and need storing. So far so good. Lots of drivers reference tty->flip.*. Several of them also call directly and unsafely into function pointers it provides. This will all break. Most drivers can use tty_insert_flip_char which can be kept as an API but others need more. At the moment I've added the following interfaces, if people think more will be needed now is a good time to say int tty_buffer_request_room(tty, size) Try and ensure at least size bytes are available, returns actual room (may be zero). At the moment it just uses the flipbuf space but that will change. Repeated calls without characters being added are not cumulative. (ie if you call it with 1, 1, 1, and then 4 you'll have four characters of space. The other functions will also try and grow buffers in future but this will be a more efficient way when you know block sizes. int tty_insert_flip_char(tty, ch, flag) As before insert a character if there is room. Now returns 1 for success, 0 for failure. int tty_insert_flip_string(tty, str, len) Insert a block of non error characters. Returns the number inserted. int tty_prepare_flip_string(tty, strptr, len) Adjust the buffer to allow len characters to be added. Returns a buffer pointer in strptr and the length available. This allows for hardware that needs to use functions like insl or mencpy_fromio. Signed-off-by: Alan Cox <alan@redhat.com> Cc: Paul Fulghum <paulkf@microgate.com> Signed-off-by: Hirokazu Takata <takata@linux-m32r.org> Signed-off-by: Serge Hallyn <serue@us.ibm.com> Signed-off-by: Jeff Dike <jdike@addtoit.com> Signed-off-by: John Hawkes <hawkes@sgi.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by: Adrian Bunk <bunk@stusta.de> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-01-10 04:54:13 +00:00
flag = TTY_NORMAL;
port->icount.rx++;
if ( status & (MPC52xx_PSC_SR_PE |
MPC52xx_PSC_SR_FE |
[PATCH] TTY layer buffering revamp The API and code have been through various bits of initial review by serial driver people but they definitely need to live somewhere for a while so the unconverted drivers can get knocked into shape, existing drivers that have been updated can be better tuned and bugs whacked out. This replaces the tty flip buffers with kmalloc objects in rings. In the normal situation for an IRQ driven serial port at typical speeds the behaviour is pretty much the same, two buffers end up allocated and the kernel cycles between them as before. When there are delays or at high speed we now behave far better as the buffer pool can grow a bit rather than lose characters. This also means that we can operate at higher speeds reliably. For drivers that receive characters in blocks (DMA based, USB and especially virtualisation) the layer allows a lot of driver specific code that works around the tty layer with private secondary queues to be removed. The IBM folks need this sort of layer, the smart serial port people do, the virtualisers do (because a virtualised tty typically operates at infinite speed rather than emulating 9600 baud). Finally many drivers had invalid and unsafe attempts to avoid buffer overflows by directly invoking tty methods extracted out of the innards of work queue structs. These are no longer needed and all go away. That fixes various random hangs with serial ports on overflow. The other change in here is to optimise the receive_room path that is used by some callers. It turns out that only one ldisc uses receive room except asa constant and it updates it far far less than the value is read. We thus make it a variable not a function call. I expect the code to contain bugs due to the size alone but I'll be watching and squashing them and feeding out new patches as it goes. Because the buffers now dynamically expand you should only run out of buffering when the kernel runs out of memory for real. That means a lot of the horrible hacks high performance drivers used to do just aren't needed any more. Description: tty_insert_flip_char is an old API and continues to work as before, as does tty_flip_buffer_push() [this is why many drivers dont need modification]. It does now also return the number of chars inserted There are also tty_buffer_request_room(tty, len) which asks for a buffer block of the length requested and returns the space found. This improves efficiency with hardware that knows how much to transfer. and tty_insert_flip_string_flags(tty, str, flags, len) to insert a string of characters and flags For a smart interface the usual code is len = tty_request_buffer_room(tty, amount_hardware_says); tty_insert_flip_string(tty, buffer_from_card, len); More description! At the moment tty buffers are attached directly to the tty. This is causing a lot of the problems related to tty layer locking, also problems at high speed and also with bursty data (such as occurs in virtualised environments) I'm working on ripping out the flip buffers and replacing them with a pool of dynamically allocated buffers. This allows both for old style "byte I/O" devices and also helps virtualisation and smart devices where large blocks of data suddenely materialise and need storing. So far so good. Lots of drivers reference tty->flip.*. Several of them also call directly and unsafely into function pointers it provides. This will all break. Most drivers can use tty_insert_flip_char which can be kept as an API but others need more. At the moment I've added the following interfaces, if people think more will be needed now is a good time to say int tty_buffer_request_room(tty, size) Try and ensure at least size bytes are available, returns actual room (may be zero). At the moment it just uses the flipbuf space but that will change. Repeated calls without characters being added are not cumulative. (ie if you call it with 1, 1, 1, and then 4 you'll have four characters of space. The other functions will also try and grow buffers in future but this will be a more efficient way when you know block sizes. int tty_insert_flip_char(tty, ch, flag) As before insert a character if there is room. Now returns 1 for success, 0 for failure. int tty_insert_flip_string(tty, str, len) Insert a block of non error characters. Returns the number inserted. int tty_prepare_flip_string(tty, strptr, len) Adjust the buffer to allow len characters to be added. Returns a buffer pointer in strptr and the length available. This allows for hardware that needs to use functions like insl or mencpy_fromio. Signed-off-by: Alan Cox <alan@redhat.com> Cc: Paul Fulghum <paulkf@microgate.com> Signed-off-by: Hirokazu Takata <takata@linux-m32r.org> Signed-off-by: Serge Hallyn <serue@us.ibm.com> Signed-off-by: Jeff Dike <jdike@addtoit.com> Signed-off-by: John Hawkes <hawkes@sgi.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by: Adrian Bunk <bunk@stusta.de> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-01-10 04:54:13 +00:00
MPC52xx_PSC_SR_RB) ) {
if (status & MPC52xx_PSC_SR_RB) {
[PATCH] TTY layer buffering revamp The API and code have been through various bits of initial review by serial driver people but they definitely need to live somewhere for a while so the unconverted drivers can get knocked into shape, existing drivers that have been updated can be better tuned and bugs whacked out. This replaces the tty flip buffers with kmalloc objects in rings. In the normal situation for an IRQ driven serial port at typical speeds the behaviour is pretty much the same, two buffers end up allocated and the kernel cycles between them as before. When there are delays or at high speed we now behave far better as the buffer pool can grow a bit rather than lose characters. This also means that we can operate at higher speeds reliably. For drivers that receive characters in blocks (DMA based, USB and especially virtualisation) the layer allows a lot of driver specific code that works around the tty layer with private secondary queues to be removed. The IBM folks need this sort of layer, the smart serial port people do, the virtualisers do (because a virtualised tty typically operates at infinite speed rather than emulating 9600 baud). Finally many drivers had invalid and unsafe attempts to avoid buffer overflows by directly invoking tty methods extracted out of the innards of work queue structs. These are no longer needed and all go away. That fixes various random hangs with serial ports on overflow. The other change in here is to optimise the receive_room path that is used by some callers. It turns out that only one ldisc uses receive room except asa constant and it updates it far far less than the value is read. We thus make it a variable not a function call. I expect the code to contain bugs due to the size alone but I'll be watching and squashing them and feeding out new patches as it goes. Because the buffers now dynamically expand you should only run out of buffering when the kernel runs out of memory for real. That means a lot of the horrible hacks high performance drivers used to do just aren't needed any more. Description: tty_insert_flip_char is an old API and continues to work as before, as does tty_flip_buffer_push() [this is why many drivers dont need modification]. It does now also return the number of chars inserted There are also tty_buffer_request_room(tty, len) which asks for a buffer block of the length requested and returns the space found. This improves efficiency with hardware that knows how much to transfer. and tty_insert_flip_string_flags(tty, str, flags, len) to insert a string of characters and flags For a smart interface the usual code is len = tty_request_buffer_room(tty, amount_hardware_says); tty_insert_flip_string(tty, buffer_from_card, len); More description! At the moment tty buffers are attached directly to the tty. This is causing a lot of the problems related to tty layer locking, also problems at high speed and also with bursty data (such as occurs in virtualised environments) I'm working on ripping out the flip buffers and replacing them with a pool of dynamically allocated buffers. This allows both for old style "byte I/O" devices and also helps virtualisation and smart devices where large blocks of data suddenely materialise and need storing. So far so good. Lots of drivers reference tty->flip.*. Several of them also call directly and unsafely into function pointers it provides. This will all break. Most drivers can use tty_insert_flip_char which can be kept as an API but others need more. At the moment I've added the following interfaces, if people think more will be needed now is a good time to say int tty_buffer_request_room(tty, size) Try and ensure at least size bytes are available, returns actual room (may be zero). At the moment it just uses the flipbuf space but that will change. Repeated calls without characters being added are not cumulative. (ie if you call it with 1, 1, 1, and then 4 you'll have four characters of space. The other functions will also try and grow buffers in future but this will be a more efficient way when you know block sizes. int tty_insert_flip_char(tty, ch, flag) As before insert a character if there is room. Now returns 1 for success, 0 for failure. int tty_insert_flip_string(tty, str, len) Insert a block of non error characters. Returns the number inserted. int tty_prepare_flip_string(tty, strptr, len) Adjust the buffer to allow len characters to be added. Returns a buffer pointer in strptr and the length available. This allows for hardware that needs to use functions like insl or mencpy_fromio. Signed-off-by: Alan Cox <alan@redhat.com> Cc: Paul Fulghum <paulkf@microgate.com> Signed-off-by: Hirokazu Takata <takata@linux-m32r.org> Signed-off-by: Serge Hallyn <serue@us.ibm.com> Signed-off-by: Jeff Dike <jdike@addtoit.com> Signed-off-by: John Hawkes <hawkes@sgi.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by: Adrian Bunk <bunk@stusta.de> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-01-10 04:54:13 +00:00
flag = TTY_BREAK;
uart_handle_break(port);
} else if (status & MPC52xx_PSC_SR_PE)
[PATCH] TTY layer buffering revamp The API and code have been through various bits of initial review by serial driver people but they definitely need to live somewhere for a while so the unconverted drivers can get knocked into shape, existing drivers that have been updated can be better tuned and bugs whacked out. This replaces the tty flip buffers with kmalloc objects in rings. In the normal situation for an IRQ driven serial port at typical speeds the behaviour is pretty much the same, two buffers end up allocated and the kernel cycles between them as before. When there are delays or at high speed we now behave far better as the buffer pool can grow a bit rather than lose characters. This also means that we can operate at higher speeds reliably. For drivers that receive characters in blocks (DMA based, USB and especially virtualisation) the layer allows a lot of driver specific code that works around the tty layer with private secondary queues to be removed. The IBM folks need this sort of layer, the smart serial port people do, the virtualisers do (because a virtualised tty typically operates at infinite speed rather than emulating 9600 baud). Finally many drivers had invalid and unsafe attempts to avoid buffer overflows by directly invoking tty methods extracted out of the innards of work queue structs. These are no longer needed and all go away. That fixes various random hangs with serial ports on overflow. The other change in here is to optimise the receive_room path that is used by some callers. It turns out that only one ldisc uses receive room except asa constant and it updates it far far less than the value is read. We thus make it a variable not a function call. I expect the code to contain bugs due to the size alone but I'll be watching and squashing them and feeding out new patches as it goes. Because the buffers now dynamically expand you should only run out of buffering when the kernel runs out of memory for real. That means a lot of the horrible hacks high performance drivers used to do just aren't needed any more. Description: tty_insert_flip_char is an old API and continues to work as before, as does tty_flip_buffer_push() [this is why many drivers dont need modification]. It does now also return the number of chars inserted There are also tty_buffer_request_room(tty, len) which asks for a buffer block of the length requested and returns the space found. This improves efficiency with hardware that knows how much to transfer. and tty_insert_flip_string_flags(tty, str, flags, len) to insert a string of characters and flags For a smart interface the usual code is len = tty_request_buffer_room(tty, amount_hardware_says); tty_insert_flip_string(tty, buffer_from_card, len); More description! At the moment tty buffers are attached directly to the tty. This is causing a lot of the problems related to tty layer locking, also problems at high speed and also with bursty data (such as occurs in virtualised environments) I'm working on ripping out the flip buffers and replacing them with a pool of dynamically allocated buffers. This allows both for old style "byte I/O" devices and also helps virtualisation and smart devices where large blocks of data suddenely materialise and need storing. So far so good. Lots of drivers reference tty->flip.*. Several of them also call directly and unsafely into function pointers it provides. This will all break. Most drivers can use tty_insert_flip_char which can be kept as an API but others need more. At the moment I've added the following interfaces, if people think more will be needed now is a good time to say int tty_buffer_request_room(tty, size) Try and ensure at least size bytes are available, returns actual room (may be zero). At the moment it just uses the flipbuf space but that will change. Repeated calls without characters being added are not cumulative. (ie if you call it with 1, 1, 1, and then 4 you'll have four characters of space. The other functions will also try and grow buffers in future but this will be a more efficient way when you know block sizes. int tty_insert_flip_char(tty, ch, flag) As before insert a character if there is room. Now returns 1 for success, 0 for failure. int tty_insert_flip_string(tty, str, len) Insert a block of non error characters. Returns the number inserted. int tty_prepare_flip_string(tty, strptr, len) Adjust the buffer to allow len characters to be added. Returns a buffer pointer in strptr and the length available. This allows for hardware that needs to use functions like insl or mencpy_fromio. Signed-off-by: Alan Cox <alan@redhat.com> Cc: Paul Fulghum <paulkf@microgate.com> Signed-off-by: Hirokazu Takata <takata@linux-m32r.org> Signed-off-by: Serge Hallyn <serue@us.ibm.com> Signed-off-by: Jeff Dike <jdike@addtoit.com> Signed-off-by: John Hawkes <hawkes@sgi.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by: Adrian Bunk <bunk@stusta.de> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-01-10 04:54:13 +00:00
flag = TTY_PARITY;
else if (status & MPC52xx_PSC_SR_FE)
[PATCH] TTY layer buffering revamp The API and code have been through various bits of initial review by serial driver people but they definitely need to live somewhere for a while so the unconverted drivers can get knocked into shape, existing drivers that have been updated can be better tuned and bugs whacked out. This replaces the tty flip buffers with kmalloc objects in rings. In the normal situation for an IRQ driven serial port at typical speeds the behaviour is pretty much the same, two buffers end up allocated and the kernel cycles between them as before. When there are delays or at high speed we now behave far better as the buffer pool can grow a bit rather than lose characters. This also means that we can operate at higher speeds reliably. For drivers that receive characters in blocks (DMA based, USB and especially virtualisation) the layer allows a lot of driver specific code that works around the tty layer with private secondary queues to be removed. The IBM folks need this sort of layer, the smart serial port people do, the virtualisers do (because a virtualised tty typically operates at infinite speed rather than emulating 9600 baud). Finally many drivers had invalid and unsafe attempts to avoid buffer overflows by directly invoking tty methods extracted out of the innards of work queue structs. These are no longer needed and all go away. That fixes various random hangs with serial ports on overflow. The other change in here is to optimise the receive_room path that is used by some callers. It turns out that only one ldisc uses receive room except asa constant and it updates it far far less than the value is read. We thus make it a variable not a function call. I expect the code to contain bugs due to the size alone but I'll be watching and squashing them and feeding out new patches as it goes. Because the buffers now dynamically expand you should only run out of buffering when the kernel runs out of memory for real. That means a lot of the horrible hacks high performance drivers used to do just aren't needed any more. Description: tty_insert_flip_char is an old API and continues to work as before, as does tty_flip_buffer_push() [this is why many drivers dont need modification]. It does now also return the number of chars inserted There are also tty_buffer_request_room(tty, len) which asks for a buffer block of the length requested and returns the space found. This improves efficiency with hardware that knows how much to transfer. and tty_insert_flip_string_flags(tty, str, flags, len) to insert a string of characters and flags For a smart interface the usual code is len = tty_request_buffer_room(tty, amount_hardware_says); tty_insert_flip_string(tty, buffer_from_card, len); More description! At the moment tty buffers are attached directly to the tty. This is causing a lot of the problems related to tty layer locking, also problems at high speed and also with bursty data (such as occurs in virtualised environments) I'm working on ripping out the flip buffers and replacing them with a pool of dynamically allocated buffers. This allows both for old style "byte I/O" devices and also helps virtualisation and smart devices where large blocks of data suddenely materialise and need storing. So far so good. Lots of drivers reference tty->flip.*. Several of them also call directly and unsafely into function pointers it provides. This will all break. Most drivers can use tty_insert_flip_char which can be kept as an API but others need more. At the moment I've added the following interfaces, if people think more will be needed now is a good time to say int tty_buffer_request_room(tty, size) Try and ensure at least size bytes are available, returns actual room (may be zero). At the moment it just uses the flipbuf space but that will change. Repeated calls without characters being added are not cumulative. (ie if you call it with 1, 1, 1, and then 4 you'll have four characters of space. The other functions will also try and grow buffers in future but this will be a more efficient way when you know block sizes. int tty_insert_flip_char(tty, ch, flag) As before insert a character if there is room. Now returns 1 for success, 0 for failure. int tty_insert_flip_string(tty, str, len) Insert a block of non error characters. Returns the number inserted. int tty_prepare_flip_string(tty, strptr, len) Adjust the buffer to allow len characters to be added. Returns a buffer pointer in strptr and the length available. This allows for hardware that needs to use functions like insl or mencpy_fromio. Signed-off-by: Alan Cox <alan@redhat.com> Cc: Paul Fulghum <paulkf@microgate.com> Signed-off-by: Hirokazu Takata <takata@linux-m32r.org> Signed-off-by: Serge Hallyn <serue@us.ibm.com> Signed-off-by: Jeff Dike <jdike@addtoit.com> Signed-off-by: John Hawkes <hawkes@sgi.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by: Adrian Bunk <bunk@stusta.de> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-01-10 04:54:13 +00:00
flag = TTY_FRAME;
/* Clear error condition */
out_8(&PSC(port)->command,MPC52xx_PSC_RST_ERR_STAT);
}
[PATCH] TTY layer buffering revamp The API and code have been through various bits of initial review by serial driver people but they definitely need to live somewhere for a while so the unconverted drivers can get knocked into shape, existing drivers that have been updated can be better tuned and bugs whacked out. This replaces the tty flip buffers with kmalloc objects in rings. In the normal situation for an IRQ driven serial port at typical speeds the behaviour is pretty much the same, two buffers end up allocated and the kernel cycles between them as before. When there are delays or at high speed we now behave far better as the buffer pool can grow a bit rather than lose characters. This also means that we can operate at higher speeds reliably. For drivers that receive characters in blocks (DMA based, USB and especially virtualisation) the layer allows a lot of driver specific code that works around the tty layer with private secondary queues to be removed. The IBM folks need this sort of layer, the smart serial port people do, the virtualisers do (because a virtualised tty typically operates at infinite speed rather than emulating 9600 baud). Finally many drivers had invalid and unsafe attempts to avoid buffer overflows by directly invoking tty methods extracted out of the innards of work queue structs. These are no longer needed and all go away. That fixes various random hangs with serial ports on overflow. The other change in here is to optimise the receive_room path that is used by some callers. It turns out that only one ldisc uses receive room except asa constant and it updates it far far less than the value is read. We thus make it a variable not a function call. I expect the code to contain bugs due to the size alone but I'll be watching and squashing them and feeding out new patches as it goes. Because the buffers now dynamically expand you should only run out of buffering when the kernel runs out of memory for real. That means a lot of the horrible hacks high performance drivers used to do just aren't needed any more. Description: tty_insert_flip_char is an old API and continues to work as before, as does tty_flip_buffer_push() [this is why many drivers dont need modification]. It does now also return the number of chars inserted There are also tty_buffer_request_room(tty, len) which asks for a buffer block of the length requested and returns the space found. This improves efficiency with hardware that knows how much to transfer. and tty_insert_flip_string_flags(tty, str, flags, len) to insert a string of characters and flags For a smart interface the usual code is len = tty_request_buffer_room(tty, amount_hardware_says); tty_insert_flip_string(tty, buffer_from_card, len); More description! At the moment tty buffers are attached directly to the tty. This is causing a lot of the problems related to tty layer locking, also problems at high speed and also with bursty data (such as occurs in virtualised environments) I'm working on ripping out the flip buffers and replacing them with a pool of dynamically allocated buffers. This allows both for old style "byte I/O" devices and also helps virtualisation and smart devices where large blocks of data suddenely materialise and need storing. So far so good. Lots of drivers reference tty->flip.*. Several of them also call directly and unsafely into function pointers it provides. This will all break. Most drivers can use tty_insert_flip_char which can be kept as an API but others need more. At the moment I've added the following interfaces, if people think more will be needed now is a good time to say int tty_buffer_request_room(tty, size) Try and ensure at least size bytes are available, returns actual room (may be zero). At the moment it just uses the flipbuf space but that will change. Repeated calls without characters being added are not cumulative. (ie if you call it with 1, 1, 1, and then 4 you'll have four characters of space. The other functions will also try and grow buffers in future but this will be a more efficient way when you know block sizes. int tty_insert_flip_char(tty, ch, flag) As before insert a character if there is room. Now returns 1 for success, 0 for failure. int tty_insert_flip_string(tty, str, len) Insert a block of non error characters. Returns the number inserted. int tty_prepare_flip_string(tty, strptr, len) Adjust the buffer to allow len characters to be added. Returns a buffer pointer in strptr and the length available. This allows for hardware that needs to use functions like insl or mencpy_fromio. Signed-off-by: Alan Cox <alan@redhat.com> Cc: Paul Fulghum <paulkf@microgate.com> Signed-off-by: Hirokazu Takata <takata@linux-m32r.org> Signed-off-by: Serge Hallyn <serue@us.ibm.com> Signed-off-by: Jeff Dike <jdike@addtoit.com> Signed-off-by: John Hawkes <hawkes@sgi.com> Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by: Adrian Bunk <bunk@stusta.de> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-01-10 04:54:13 +00:00
tty_insert_flip_char(tty, ch, flag);
if (status & MPC52xx_PSC_SR_OE) {
/*
* Overrun is special, since it's
* reported immediately, and doesn't
* affect the current character
*/
tty_insert_flip_char(tty, 0, TTY_OVERRUN);
}
}
tty_flip_buffer_push(tty);
return in_be16(&PSC(port)->mpc52xx_psc_status) & MPC52xx_PSC_SR_RXRDY;
}
static inline int
mpc52xx_uart_int_tx_chars(struct uart_port *port)
{
struct circ_buf *xmit = &port->info->xmit;
/* Process out of band chars */
if (port->x_char) {
out_8(&PSC(port)->mpc52xx_psc_buffer_8, port->x_char);
port->icount.tx++;
port->x_char = 0;
return 1;
}
/* Nothing to do ? */
if (uart_circ_empty(xmit) || uart_tx_stopped(port)) {
mpc52xx_uart_stop_tx(port);
return 0;
}
/* Send chars */
while (in_be16(&PSC(port)->mpc52xx_psc_status) & MPC52xx_PSC_SR_TXRDY) {
out_8(&PSC(port)->mpc52xx_psc_buffer_8, xmit->buf[xmit->tail]);
xmit->tail = (xmit->tail + 1) & (UART_XMIT_SIZE - 1);
port->icount.tx++;
if (uart_circ_empty(xmit))
break;
}
/* Wake up */
if (uart_circ_chars_pending(xmit) < WAKEUP_CHARS)
uart_write_wakeup(port);
/* Maybe we're done after all */
if (uart_circ_empty(xmit)) {
mpc52xx_uart_stop_tx(port);
return 0;
}
return 1;
}
static irqreturn_t
IRQ: Maintain regs pointer globally rather than passing to IRQ handlers Maintain a per-CPU global "struct pt_regs *" variable which can be used instead of passing regs around manually through all ~1800 interrupt handlers in the Linux kernel. The regs pointer is used in few places, but it potentially costs both stack space and code to pass it around. On the FRV arch, removing the regs parameter from all the genirq function results in a 20% speed up of the IRQ exit path (ie: from leaving timer_interrupt() to leaving do_IRQ()). Where appropriate, an arch may override the generic storage facility and do something different with the variable. On FRV, for instance, the address is maintained in GR28 at all times inside the kernel as part of general exception handling. Having looked over the code, it appears that the parameter may be handed down through up to twenty or so layers of functions. Consider a USB character device attached to a USB hub, attached to a USB controller that posts its interrupts through a cascaded auxiliary interrupt controller. A character device driver may want to pass regs to the sysrq handler through the input layer which adds another few layers of parameter passing. I've build this code with allyesconfig for x86_64 and i386. I've runtested the main part of the code on FRV and i386, though I can't test most of the drivers. I've also done partial conversion for powerpc and MIPS - these at least compile with minimal configurations. This will affect all archs. Mostly the changes should be relatively easy. Take do_IRQ(), store the regs pointer at the beginning, saving the old one: struct pt_regs *old_regs = set_irq_regs(regs); And put the old one back at the end: set_irq_regs(old_regs); Don't pass regs through to generic_handle_irq() or __do_IRQ(). In timer_interrupt(), this sort of change will be necessary: - update_process_times(user_mode(regs)); - profile_tick(CPU_PROFILING, regs); + update_process_times(user_mode(get_irq_regs())); + profile_tick(CPU_PROFILING); I'd like to move update_process_times()'s use of get_irq_regs() into itself, except that i386, alone of the archs, uses something other than user_mode(). Some notes on the interrupt handling in the drivers: (*) input_dev() is now gone entirely. The regs pointer is no longer stored in the input_dev struct. (*) finish_unlinks() in drivers/usb/host/ohci-q.c needs checking. It does something different depending on whether it's been supplied with a regs pointer or not. (*) Various IRQ handler function pointers have been moved to type irq_handler_t. Signed-Off-By: David Howells <dhowells@redhat.com> (cherry picked from 1b16e7ac850969f38b375e511e3fa2f474a33867 commit)
2006-10-05 13:55:46 +00:00
mpc52xx_uart_int(int irq, void *dev_id)
{
struct uart_port *port = dev_id;
unsigned long pass = ISR_PASS_LIMIT;
unsigned int keepgoing;
unsigned short status;
spin_lock(&port->lock);
/* While we have stuff to do, we continue */
do {
/* If we don't find anything to do, we stop */
keepgoing = 0;
/* Read status */
status = in_be16(&PSC(port)->mpc52xx_psc_isr);
status &= port->read_status_mask;
/* Do we need to receive chars ? */
/* For this RX interrupts must be on and some chars waiting */
if ( status & MPC52xx_PSC_IMR_RXRDY )
IRQ: Maintain regs pointer globally rather than passing to IRQ handlers Maintain a per-CPU global "struct pt_regs *" variable which can be used instead of passing regs around manually through all ~1800 interrupt handlers in the Linux kernel. The regs pointer is used in few places, but it potentially costs both stack space and code to pass it around. On the FRV arch, removing the regs parameter from all the genirq function results in a 20% speed up of the IRQ exit path (ie: from leaving timer_interrupt() to leaving do_IRQ()). Where appropriate, an arch may override the generic storage facility and do something different with the variable. On FRV, for instance, the address is maintained in GR28 at all times inside the kernel as part of general exception handling. Having looked over the code, it appears that the parameter may be handed down through up to twenty or so layers of functions. Consider a USB character device attached to a USB hub, attached to a USB controller that posts its interrupts through a cascaded auxiliary interrupt controller. A character device driver may want to pass regs to the sysrq handler through the input layer which adds another few layers of parameter passing. I've build this code with allyesconfig for x86_64 and i386. I've runtested the main part of the code on FRV and i386, though I can't test most of the drivers. I've also done partial conversion for powerpc and MIPS - these at least compile with minimal configurations. This will affect all archs. Mostly the changes should be relatively easy. Take do_IRQ(), store the regs pointer at the beginning, saving the old one: struct pt_regs *old_regs = set_irq_regs(regs); And put the old one back at the end: set_irq_regs(old_regs); Don't pass regs through to generic_handle_irq() or __do_IRQ(). In timer_interrupt(), this sort of change will be necessary: - update_process_times(user_mode(regs)); - profile_tick(CPU_PROFILING, regs); + update_process_times(user_mode(get_irq_regs())); + profile_tick(CPU_PROFILING); I'd like to move update_process_times()'s use of get_irq_regs() into itself, except that i386, alone of the archs, uses something other than user_mode(). Some notes on the interrupt handling in the drivers: (*) input_dev() is now gone entirely. The regs pointer is no longer stored in the input_dev struct. (*) finish_unlinks() in drivers/usb/host/ohci-q.c needs checking. It does something different depending on whether it's been supplied with a regs pointer or not. (*) Various IRQ handler function pointers have been moved to type irq_handler_t. Signed-Off-By: David Howells <dhowells@redhat.com> (cherry picked from 1b16e7ac850969f38b375e511e3fa2f474a33867 commit)
2006-10-05 13:55:46 +00:00
keepgoing |= mpc52xx_uart_int_rx_chars(port);
/* Do we need to send chars ? */
/* For this, TX must be ready and TX interrupt enabled */
if ( status & MPC52xx_PSC_IMR_TXRDY )
keepgoing |= mpc52xx_uart_int_tx_chars(port);
/* Limit number of iteration */
if ( !(--pass) )
keepgoing = 0;
} while (keepgoing);
spin_unlock(&port->lock);
return IRQ_HANDLED;
}
/* ======================================================================== */
/* Console ( if applicable ) */
/* ======================================================================== */
#ifdef CONFIG_SERIAL_MPC52xx_CONSOLE
static void __init
mpc52xx_console_get_options(struct uart_port *port,
int *baud, int *parity, int *bits, int *flow)
{
struct mpc52xx_psc __iomem *psc = PSC(port);
unsigned char mr1;
/* Read the mode registers */
out_8(&psc->command,MPC52xx_PSC_SEL_MODE_REG_1);
mr1 = in_8(&psc->mode);
/* CT{U,L}R are write-only ! */
*baud = __res.bi_baudrate ?
__res.bi_baudrate : CONFIG_SERIAL_MPC52xx_CONSOLE_BAUD;
/* Parse them */
switch (mr1 & MPC52xx_PSC_MODE_BITS_MASK) {
case MPC52xx_PSC_MODE_5_BITS: *bits = 5; break;
case MPC52xx_PSC_MODE_6_BITS: *bits = 6; break;
case MPC52xx_PSC_MODE_7_BITS: *bits = 7; break;
case MPC52xx_PSC_MODE_8_BITS:
default: *bits = 8;
}
if (mr1 & MPC52xx_PSC_MODE_PARNONE)
*parity = 'n';
else
*parity = mr1 & MPC52xx_PSC_MODE_PARODD ? 'o' : 'e';
}
static void
mpc52xx_console_write(struct console *co, const char *s, unsigned int count)
{
struct uart_port *port = &mpc52xx_uart_ports[co->index];
struct mpc52xx_psc __iomem *psc = PSC(port);
unsigned int i, j;
/* Disable interrupts */
out_be16(&psc->mpc52xx_psc_imr, 0);
/* Wait the TX buffer to be empty */
j = 5000000; /* Maximum wait */
while (!(in_be16(&psc->mpc52xx_psc_status) & MPC52xx_PSC_SR_TXEMP) &&
--j)
udelay(1);
/* Write all the chars */
for (i = 0; i < count; i++, s++) {
/* Line return handling */
if (*s == '\n')
out_8(&psc->mpc52xx_psc_buffer_8, '\r');
/* Send the char */
out_8(&psc->mpc52xx_psc_buffer_8, *s);
/* Wait the TX buffer to be empty */
j = 20000; /* Maximum wait */
while (!(in_be16(&psc->mpc52xx_psc_status) &
MPC52xx_PSC_SR_TXEMP) && --j)
udelay(1);
}
/* Restore interrupt state */
out_be16(&psc->mpc52xx_psc_imr, port->read_status_mask);
}
static int __init
mpc52xx_console_setup(struct console *co, char *options)
{
struct uart_port *port = &mpc52xx_uart_ports[co->index];
int baud = CONFIG_SERIAL_MPC52xx_CONSOLE_BAUD;
int bits = 8;
int parity = 'n';
int flow = 'n';
if (co->index < 0 || co->index >= MPC52xx_PSC_MAXNUM)
return -EINVAL;
/* Basic port init. Needed since we use some uart_??? func before
* real init for early access */
spin_lock_init(&port->lock);
port->uartclk = __res.bi_ipbfreq / 2; /* Look at CTLR doc */
port->ops = &mpc52xx_uart_ops;
port->mapbase = MPC52xx_PA(MPC52xx_PSCx_OFFSET(co->index+1));
/* We ioremap ourself */
port->membase = ioremap(port->mapbase, MPC52xx_PSC_SIZE);
if (port->membase == NULL)
return -EINVAL;
/* Setup the port parameters accoding to options */
if (options)
uart_parse_options(options, &baud, &parity, &bits, &flow);
else
mpc52xx_console_get_options(port, &baud, &parity, &bits, &flow);
return uart_set_options(port, co, baud, parity, bits, flow);
}
static struct uart_driver mpc52xx_uart_driver;
static struct console mpc52xx_console = {
.name = "ttyPSC",
.write = mpc52xx_console_write,
.device = uart_console_device,
.setup = mpc52xx_console_setup,
.flags = CON_PRINTBUFFER,
.index = -1, /* Specified on the cmdline (e.g. console=ttyPSC0 ) */
.data = &mpc52xx_uart_driver,
};
static int __init
mpc52xx_console_init(void)
{
register_console(&mpc52xx_console);
return 0;
}
console_initcall(mpc52xx_console_init);
#define MPC52xx_PSC_CONSOLE &mpc52xx_console
#else
#define MPC52xx_PSC_CONSOLE NULL
#endif
/* ======================================================================== */
/* UART Driver */
/* ======================================================================== */
static struct uart_driver mpc52xx_uart_driver = {
.owner = THIS_MODULE,
.driver_name = "mpc52xx_psc_uart",
.dev_name = "ttyPSC",
.major = SERIAL_PSC_MAJOR,
.minor = SERIAL_PSC_MINOR,
.nr = MPC52xx_PSC_MAXNUM,
.cons = MPC52xx_PSC_CONSOLE,
};
/* ======================================================================== */
/* Platform Driver */
/* ======================================================================== */
static int __devinit
mpc52xx_uart_probe(struct platform_device *dev)
{
struct resource *res = dev->resource;
struct uart_port *port = NULL;
int i, idx, ret;
/* Check validity & presence */
idx = dev->id;
if (idx < 0 || idx >= MPC52xx_PSC_MAXNUM)
return -EINVAL;
if (!mpc52xx_match_psc_function(idx,"uart"))
return -ENODEV;
/* Init the port structure */
port = &mpc52xx_uart_ports[idx];
memset(port, 0x00, sizeof(struct uart_port));
spin_lock_init(&port->lock);
port->uartclk = __res.bi_ipbfreq / 2; /* Look at CTLR doc */
port->fifosize = 512;
port->iotype = UPIO_MEM;
port->flags = UPF_BOOT_AUTOCONF |
( uart_console(port) ? 0 : UPF_IOREMAP );
port->line = idx;
port->ops = &mpc52xx_uart_ops;
/* Search for IRQ and mapbase */
for (i=0 ; i<dev->num_resources ; i++, res++) {
if (res->flags & IORESOURCE_MEM)
port->mapbase = res->start;
else if (res->flags & IORESOURCE_IRQ)
port->irq = res->start;
}
if (!port->irq || !port->mapbase)
return -EINVAL;
/* Add the port to the uart sub-system */
ret = uart_add_one_port(&mpc52xx_uart_driver, port);
if (!ret)
platform_set_drvdata(dev, (void*)port);
return ret;
}
static int
mpc52xx_uart_remove(struct platform_device *dev)
{
struct uart_port *port = (struct uart_port *) platform_get_drvdata(dev);
platform_set_drvdata(dev, NULL);
if (port)
uart_remove_one_port(&mpc52xx_uart_driver, port);
return 0;
}
#ifdef CONFIG_PM
static int
mpc52xx_uart_suspend(struct platform_device *dev, pm_message_t state)
{
struct uart_port *port = (struct uart_port *) platform_get_drvdata(dev);
if (sport)
uart_suspend_port(&mpc52xx_uart_driver, port);
return 0;
}
static int
mpc52xx_uart_resume(struct platform_device *dev)
{
struct uart_port *port = (struct uart_port *) platform_get_drvdata(dev);
if (port)
uart_resume_port(&mpc52xx_uart_driver, port);
return 0;
}
#endif
static struct platform_driver mpc52xx_uart_platform_driver = {
.probe = mpc52xx_uart_probe,
.remove = mpc52xx_uart_remove,
#ifdef CONFIG_PM
.suspend = mpc52xx_uart_suspend,
.resume = mpc52xx_uart_resume,
#endif
.driver = {
.name = "mpc52xx-psc",
},
};
/* ======================================================================== */
/* Module */
/* ======================================================================== */
static int __init
mpc52xx_uart_init(void)
{
int ret;
printk(KERN_INFO "Serial: MPC52xx PSC driver\n");
ret = uart_register_driver(&mpc52xx_uart_driver);
if (ret == 0) {
ret = platform_driver_register(&mpc52xx_uart_platform_driver);
if (ret)
uart_unregister_driver(&mpc52xx_uart_driver);
}
return ret;
}
static void __exit
mpc52xx_uart_exit(void)
{
platform_driver_unregister(&mpc52xx_uart_platform_driver);
uart_unregister_driver(&mpc52xx_uart_driver);
}
module_init(mpc52xx_uart_init);
module_exit(mpc52xx_uart_exit);
MODULE_AUTHOR("Sylvain Munaut <tnt@246tNt.com>");
MODULE_DESCRIPTION("Freescale MPC52xx PSC UART");
MODULE_LICENSE("GPL");