With commit eaa7995c52 (regulator: core: avoid
regulator_resolve_supply() race condition) we started holding the rdev
lock while resolving supplies, an operation that requires holding the
regulator_list_mutex. This results in lockdep warnings since in other
places we take the list mutex then the mutex on an individual rdev.
Since the goal is to make sure that we don't call set_supply() twice
rather than a concern about the cost of resolution pull the rdev lock
and check for duplicate resolution down to immediately before we do the
set_supply() and drop it again once the allocation is done.
Fixes: eaa7995c52 (regulator: core: avoid regulator_resolve_supply() race condition)
Reported-by: Marek Szyprowski <m.szyprowski@samsung.com>
Tested-by: Marek Szyprowski <m.szyprowski@samsung.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
Link: https://lore.kernel.org/r/20210122132042.10306-1-broonie@kernel.org
Signed-off-by: Mark Brown <broonie@kernel.org>
The final step in regulator_register() is to call
regulator_resolve_supply() for each registered regulator
(including the one in the process of being registered). The
regulator_resolve_supply() function first checks if rdev->supply
is NULL, then it performs various steps to try to find the supply.
If successful, rdev->supply is set inside of set_supply().
This procedure can encounter a race condition if two concurrent
tasks call regulator_register() near to each other on separate CPUs
and one of the regulators has rdev->supply_name specified. There
is currently nothing guaranteeing atomicity between the rdev->supply
check and set steps. Thus, both tasks can observe rdev->supply==NULL
in their regulator_resolve_supply() calls. This then results in
both creating a struct regulator for the supply. One ends up
actually stored in rdev->supply and the other is lost (though still
present in the supply's consumer_list).
Here is a kernel log snippet showing the issue:
[ 12.421768] gpu_cc_gx_gdsc: supplied by pm8350_s5_level
[ 12.425854] gpu_cc_gx_gdsc: supplied by pm8350_s5_level
[ 12.429064] debugfs: Directory 'regulator.4-SUPPLY' with parent
'17a00000.rsc:rpmh-regulator-gfxlvl-pm8350_s5_level'
already present!
Avoid this race condition by holding the rdev->mutex lock inside
of regulator_resolve_supply() while checking and setting
rdev->supply.
Signed-off-by: David Collins <collinsd@codeaurora.org>
Link: https://lore.kernel.org/r/1610068562-4410-1-git-send-email-collinsd@codeaurora.org
Signed-off-by: Mark Brown <broonie@kernel.org>
A built-in regulator driver cannot link against a modular cmd_db driver:
qcom-rpmh-regulator.c:(.text+0x174): undefined reference to `cmd_db_read_addr'
There is already a dependency for RPMh, so add another one of this
type for cmd_db.
Fixes: 34c5aa2666 ("regulator: Kconfig: Fix REGULATOR_QCOM_RPMH dependencies to avoid build error")
Fixes: 46fc033eba ("regulator: add QCOM RPMh regulator driver")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Reviewed-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Link: https://lore.kernel.org/r/20201230145712.3133110-1-arnd@kernel.org
Signed-off-by: Mark Brown <broonie@kernel.org>
The pf8x00 driver supports three devices, the DT compatible strings
and I2C IDs should enumerate these specifically rather than using a
wildcard so that we don't collide with anything incompatible in the
same ID range in the future and so that we can handle any software
visible differences between the variants we find.
Signed-off-by: Mark Brown <broonie@kernel.org>
Link: https://lore.kernel.org/r/20201215132024.13356-1-broonie@kernel.org
Signed-off-by: Mark Brown <broonie@kernel.org>
This patch fixes a couple of bugs in the DA9121 driver.
One in an uninialised string I forgot to remove when changing to of_parse_cb()
The other is an index for an optional DT property which overflows
Adam Ward (2):
regulator: da9121: Remove uninitialised string variable
regulator: da9121: Fix index used for DT property
drivers/regulator/da9121-regulator.c | 5 ++---
1 file changed, 2 insertions(+), 3 deletions(-)
--
1.9.1
When I use the axp20x chip to power my SDIO device on the 5.4 kernel,
the output voltage of DLDO2 is wrong. After comparing the register
manual and source code of the chip, I found that the mask bit of the
driver register of the port was wrong. I fixed this error by modifying
the mask register of the source code. This error seems to be a copy
error of the macro when writing the code. Now the voltage output of
the DLDO2 port of axp20x is correct. My development environment is
Allwinner A40I of arm architecture, and the kernel version is 5.4.
Signed-off-by: DingHua Ma <dinghua.ma.sz@gmail.com>
Reviewed-by: Chen-Yu Tsai <wens@csie.org>
Cc: <stable@vger.kernel.org>
Fixes: db4a555f7c ("regulator: axp20x: use defines for masks")
Link: https://lore.kernel.org/r/20201201001000.22302-1-dinghua.ma.sz@gmail.com
Signed-off-by: Mark Brown <broonie@kernel.org>
When CONFIG_GPIOLIB is disabled, the declarations from linux/gpio/consumer.h
are not visible:
drivers/regulator/da9121-regulator.c:371:14: error: implicit declaration of function 'fwnode_gpiod_get_index' [-Werror,-Wimplicit-function-declaration]
ena_gpiod = fwnode_gpiod_get_index(of_fwnode_handle(np), "enable", 0,
^
drivers/regulator/da9121-regulator.c:372:7: error: use of undeclared identifier 'GPIOD_OUT_HIGH'
GPIOD_OUT_HIGH |
^
drivers/regulator/da9121-regulator.c:373:7: error: use of undeclared identifier 'GPIOD_FLAGS_BIT_NONEXCLUSIVE'
GPIOD_FLAGS_BIT_NONEXCLUSIVE,
Include this explicitly to help compile testing.
Fixes: 46c413d5bb ("regulator: da9121: Add support for device variants via devicetree")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Link: https://lore.kernel.org/r/20201204165229.3754763-1-arnd@kernel.org
Signed-off-by: Mark Brown <broonie@kernel.org>
Fix the following sparse warnings:
drivers/regulator/da9121-regulator.c:55:21: warning: symbol 'da9121_10A_2phase_current' was not declared. Should it be static?
drivers/regulator/da9121-regulator.c:63:21: warning: symbol 'da9121_6A_2phase_current' was not declared. Should it be static?
drivers/regulator/da9121-regulator.c:71:21: warning: symbol 'da9121_5A_1phase_current' was not declared. Should it be static?
drivers/regulator/da9121-regulator.c:79:21: warning: symbol 'da9121_3A_1phase_current' was not declared. Should it be static?
drivers/regulator/da9121-regulator.c:151:32: warning: symbol 'status_event_handling' was not declared. Should it be static?
Signed-off-by: Zou Wei <zou_wei@huawei.com>
Acked-by: Adam Ward <Adam.Ward.opensource@diasemi.com>
Link: https://lore.kernel.org/r/1606994795-36182-1-git-send-email-zou_wei@huawei.com
Signed-off-by: Mark Brown <broonie@kernel.org>
The kernel test robot reported the following build error:
All errors (new ones prefixed by >>):
xtensa-linux-ld: drivers/regulator/qcom-rpmh-regulator.o: in function `rpmh_regulator_vrm_get_voltage_sel':
qcom-rpmh-regulator.c:(.text+0x270): undefined reference to `rpmh_write'
xtensa-linux-ld: drivers/regulator/qcom-rpmh-regulator.o: in function `rpmh_regulator_send_request':
qcom-rpmh-regulator.c:(.text+0x2f2): undefined reference to `rpmh_write'
xtensa-linux-ld: drivers/regulator/qcom-rpmh-regulator.o: in function `rpmh_regulator_vrm_get_voltage_sel':
>> qcom-rpmh-regulator.c:(.text+0x274): undefined reference to `rpmh_write_async'
xtensa-linux-ld: drivers/regulator/qcom-rpmh-regulator.o: in function `rpmh_regulator_send_request':
qcom-rpmh-regulator.c:(.text+0x2fc): undefined reference to `rpmh_write_async'
Which is due to REGULATOR_QCOM_RPMH depending on
QCOM_RPMH || COMPILE_TEST. The problem is that QOM_RPMH can now
be a module, which in that case requires REGULATOR_QCOM_RPMH=m
to build.
However, if COMPILE_TEST is enabled, REGULATOR_QCOM_RPMH can be
set to =y while QCOM_RPMH=m which will cause build failures.
The fix here is to add (QCOM_RPMH=n && COMPILE_TEST) to the
dependency.
Feedback would be appreciated!
Cc: Todd Kjos <tkjos@google.com>
Cc: Saravana Kannan <saravanak@google.com>
Cc: Andy Gross <agross@kernel.org>
Cc: Bjorn Andersson <bjorn.andersson@linaro.org>
Cc: Rajendra Nayak <rnayak@codeaurora.org>
Cc: Maulik Shah <mkshah@codeaurora.org>
Cc: Stephen Boyd <swboyd@chromium.org>
Cc: Liam Girdwood <lgirdwood@gmail.com>
Cc: Mark Brown <broonie@kernel.org>
Cc: linux-arm-msm@vger.kernel.org
Reported-by: kernel test robot <lkp@intel.com>
Signed-off-by: John Stultz <john.stultz@linaro.org>
Reviewed-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Link: https://lore.kernel.org/r/20201123222359.103822-1-john.stultz@linaro.org
Signed-off-by: Mark Brown <broonie@kernel.org>
Add a simple regulator based on SCMI Voltage Domain Protocol.
Signed-off-by: Cristian Marussi <cristian.marussi@arm.com>
----
v6 --> v7
- add proper blank lines between semantic blocks
- fix return value on error path of scmi_reg_is_enabled()
- use generic Failure message on err path of info_get()
- fix comment containing apostrophe
v3 --> v4
- using of_match_full_name core regulator flag
- avoid coccinelle falde complaints about pointer-sized allocations
v2 --> v3
- remove multiple linear mappings support
- removed duplicated voltage name printout
- added a few comments
- simplified return path in scmi_reg_set_voltage_sel()
v1 --> v2
- removed duplicate regulator naming
- removed redundant .get/set_voltage ops: only _sel variants implemented
- removed condexpr on fail path to increase readability
v0 --> v1
- fixed init_data constraint parsing
- fixes for v5.8 (linear_range.h)
- fixed commit message content and subject line format
- factored out SCMI core specific changes to distinct patch
- reworked Kconfig and Makefile to keep proper alphabetic order
- fixed SPDX comment style
- removed unneeded inline functions
- reworked conditionals for legibility
- fixed some return paths to properly report SCMI original errors codes
- added some more descriptive error messages when fw returns invalid ranges
- removed unneeded explicit devm_regulator_unregister from .remove()
Link: https://lore.kernel.org/r/20201123202336.46701-4-cristian.marussi@arm.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Hi,
this series introduces the support for the new SCMI Voltage Domain Protocol
defined by the upcoming SCMIv3.0 specification, whose BETA release is
available at [1].
Afterwards, a new generic SCMI Regulator driver is developed on top of the
new SCMI VD Protocol.
In V4 Patch 3/5 introduced a needed fix in Regulator framework to cope with
generic named nodes.
The series is currently based on for-next/scmi [2] on top of:
commit b141fca08207 ("firmware: arm_scmi: Fix missing destroy_workqueue()")
Any feedback welcome,
Thanks,
Cristian
---
v5 --> v6
- reordered dt bindings patch
- removed single field struct
- reviewed args to scmi_init_voltage_levels()
- allocating scmi_voltage_info_array contiguously
v4 --> v5
- rebased
- VD Protocol
- removed inline
- moved segmented intervals defines
- fixed some macros complaints by checkpatch
v3 --> v4
- DT bindings
- using generic node names
- listing explicitly subset of supported regulators bindings
- SCMI Regulator
- using of_match_full_name core regulator flag
- avoid coccinelle false flag complaints
- VD Protocol
- avoid coccinelle false flag complaints
- avoiding fixed size typing
v2 --> v3
- DT bindings
- avoid awkard examples based on _cpu/_gpu regulators
- SCMI Regulator
- remove multiple linear mappings support
- removed duplicated voltage name printout
- added a few comments
- simplified return path in scmi_reg_set_voltage_sel()
- VD Protocol
- restrict segmented voltage domain descriptors to one triplet
- removed unneeded inline
- free allocated resources for invalid voltage domain
- added __must_check to info_get voltage operations
- added a few comments
- removed fixed size typing from struct voltage_info
v1 --> v2
- rebased on for-next/scmi v5.10
- DT bindings
- removed any reference to negative voltages
- SCMI Regulator
- removed duplicate regulator naming
- removed redundant .get/set_voltage ops: only _sel variants implemented
- removed condexpr on fail path to increase readability
- VD Protocol
- fix voltage levels query loop to reload full cmd description
between iterations as reported by Etienne Carriere
- ensure transport rx buffer is properly sized calli scmi_reset_rx_to_maxsz
between transfers
[1]:https://developer.arm.com/documentation/den0056/c/
[2]:https://git.kernel.org/pub/scm/linux/kernel/git/sudeep.holla/linux.git/log/?h=for-next/scmi
Cristian Marussi (5):
firmware: arm_scmi: Add Voltage Domain Support
firmware: arm_scmi: add SCMI Voltage Domain devname
regulator: core: add of_match_full_name boolean flag
dt-bindings: arm: add support for SCMI Regulators
regulator: add SCMI driver
.../devicetree/bindings/arm/arm,scmi.txt | 43 ++
drivers/firmware/arm_scmi/Makefile | 2 +-
drivers/firmware/arm_scmi/common.h | 1 +
drivers/firmware/arm_scmi/driver.c | 3 +
drivers/firmware/arm_scmi/voltage.c | 380 ++++++++++++++++
drivers/regulator/Kconfig | 9 +
drivers/regulator/Makefile | 1 +
drivers/regulator/of_regulator.c | 8 +-
drivers/regulator/scmi-regulator.c | 409 ++++++++++++++++++
include/linux/regulator/driver.h | 3 +
include/linux/scmi_protocol.h | 64 +++
11 files changed, 920 insertions(+), 3 deletions(-)
create mode 100644 drivers/firmware/arm_scmi/voltage.c
create mode 100644 drivers/regulator/scmi-regulator.c
--
2.17.1
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.orghttp://lists.infradead.org/mailman/listinfo/linux-arm-kernel
During regulators registration, if .of_match and .regulators_node are
defined as non-null strings in struct regulator_desc the core searches the
DT subtree rooted at .regulators_node trying to match, at first, .of_match
against the 'regulator-compatible' property and, then, falling back to use
the name of the node itself to determine a good match.
Property 'regulator-compatible', though, is now deprecated and falling back
to match against the node name, works fine only as long as the involved
nodes are named in an unique way across the searched subtree; if that's not
the case, like when using <common-name>@<unit> style naming for properties
indexed via 'reg' property (as advised by the standard), the above matching
mechanism based on the simple common name will lead to multiple matches and
the only viable alternative would be to properly define the now deprecated
'regulator-compatible' as the node full name, i.e. <common-name>@<unit>.
In order to address this case without using such deprecated binding, define
a new boolean flag .of_match_full_name in struct regulator_desc to force
the core to match against the node full-name instead of the plain name.
Signed-off-by: Cristian Marussi <cristian.marussi@arm.com>
Link: https://lore.kernel.org/r/20201119191051.46363-4-cristian.marussi@arm.com
Signed-off-by: Mark Brown <broonie@kernel.org>
At the start of driver initialization, we do not know what bias
setting the bootloader has configured the system for and we only know
for certain the very first time we do a transition.
However, since the initial value of the comparison index is -EINVAL,
this negative value results in an array out of bound access on the
very first transition.
Since we don't know what the setting is, we just set the bias
configuration as there is nothing to compare against. This prevents
the array out of bound access.
NOTE: Even though we could use a more relaxed check of "< 0" the only
valid values(ignoring cosmic ray induced bitflips) are -EINVAL, 0+.
Fixes: 40b1936efe ("regulator: Introduce TI Adaptive Body Bias(ABB) on-chip LDO driver")
Link: https://lore.kernel.org/linux-mm/CA+G9fYuk4imvhyCN7D7T6PMDH6oNp6HDCRiTUKMQ6QXXjBa4ag@mail.gmail.com/
Reported-by: Naresh Kamboju <naresh.kamboju@linaro.org>
Reviewed-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Nishanth Menon <nm@ti.com>
Link: https://lore.kernel.org/r/20201118145009.10492-1-nm@ti.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Hi,
This series adds support for ramp delay on mcp16502. It also adds
some cleanup on mcp16502.
Apart from that patches 1/6 fixes the selector validation in case
the regulator::desc::linear_min_sel is not zero.
Thank you,
Claudiu Beznea
Changes in v3:
- fix compilation error in patch 5/6
Reported-by: kernel test robot <lkp@intel.com>
Changes in v2:
- rebase on top of regulator/for-next
- checked 1/6 and 3/6 applies on top of regulator/for-5.10
Claudiu Beznea (6):
regulator: core: validate selector against linear_min_sel
regulator: core: do not continue if selector match
regulator: mcp16502: add linear_min_sel
regulator: mcp16502: adapt for get/set on other registers
regulator: mcp16502: add support for ramp delay
regulator: mcp16502: remove void documentation of struct mcp16502
drivers/regulator/core.c | 12 +++-
drivers/regulator/helpers.c | 3 +-
drivers/regulator/mcp16502.c | 135 ++++++++++++++++++++++++++++++++++++-------
3 files changed, 127 insertions(+), 23 deletions(-)
--
2.7.4
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.orghttp://lists.infradead.org/mailman/listinfo/linux-arm-kernel
