linux-stable.git/drivers/base/platform.c, branch v5.15.208

driver: platform: Add helper for safer setting of driver_override

2023-11-08T16:26:43+00:00

commit 6c2f421174273de8f83cde4286d1c076d43a2d35 upstream.

Several core drivers and buses expect that driver_override is a
dynamically allocated memory thus later they can kfree() it.

However such assumption is not documented, there were in the past and
there are already users setting it to a string literal. This leads to
kfree() of static memory during device release (e.g. in error paths or
during unbind):

    kernel BUG at ../mm/slub.c:3960!
    Internal error: Oops - BUG: 0 [#1] PREEMPT SMP ARM
    ...
    (kfree) from [] (platform_device_release+0x88/0xb4)
    (platform_device_release) from [] (device_release+0x2c/0x90)
    (device_release) from [] (kobject_put+0xec/0x20c)
    (kobject_put) from [] (exynos5_clk_probe+0x154/0x18c)
    (exynos5_clk_probe) from [] (platform_drv_probe+0x6c/0xa4)
    (platform_drv_probe) from [] (really_probe+0x280/0x414)
    (really_probe) from [] (driver_probe_device+0x78/0x1c4)
    (driver_probe_device) from [] (bus_for_each_drv+0x74/0xb8)
    (bus_for_each_drv) from [] (__device_attach+0xd4/0x16c)
    (__device_attach) from [] (bus_probe_device+0x88/0x90)
    (bus_probe_device) from [] (device_add+0x3dc/0x62c)
    (device_add) from [] (of_platform_device_create_pdata+0x94/0xbc)
    (of_platform_device_create_pdata) from [] (of_platform_bus_create+0x1a8/0x4fc)
    (of_platform_bus_create) from [] (of_platform_bus_create+0x20c/0x4fc)
    (of_platform_bus_create) from [] (of_platform_populate+0x84/0x118)
    (of_platform_populate) from [] (of_platform_default_populate_init+0xa0/0xb8)
    (of_platform_default_populate_init) from [] (do_one_initcall+0x8c/0x404)

Provide a helper which clearly documents the usage of driver_override.
This will allow later to reuse the helper and reduce the amount of
duplicated code.

Convert the platform driver to use a new helper and make the
driver_override field const char (it is not modified by the core).

Reviewed-by: Rafael J. Wysocki 
Acked-by: Rafael J. Wysocki 
Signed-off-by: Krzysztof Kozlowski 
Link: https://lore.kernel.org/r/20220419113435.246203-2-krzysztof.kozlowski@linaro.org
Signed-off-by: Greg Kroah-Hartman 
Signed-off-by: Lee Jones 
Signed-off-by: Greg Kroah-Hartman

platform: Provide a remove callback that returns no value

2023-05-11T14:00:24+00:00

[ Upstream commit 5c5a7680e67ba6fbbb5f4d79fa41485450c1985c ]

struct platform_driver::remove returning an integer made driver authors
expect that returning an error code was proper error handling. However
the driver core ignores the error and continues to remove the device
because there is nothing the core could do anyhow and reentering the
remove callback again is only calling for trouble.

So this is an source for errors typically yielding resource leaks in the
error path.

As there are too many platform drivers to neatly convert them all to
return void in a single go, do it in several steps after this patch:

 a) Convert all drivers to implement .remove_new() returning void instead
    of .remove() returning int;
 b) Change struct platform_driver::remove() to return void and so make
    it identical to .remove_new();
 c) Change all drivers back to .remove() now with the better prototype;
 d) drop struct platform_driver::remove_new().

While this touches all drivers eventually twice, steps a) and c) can be
done one driver after another and so reduces coordination efforts
immensely and simplifies review.

Signed-off-by: Uwe Kleine-König 
Link: https://lore.kernel.org/r/20221209150914.3557650-1-u.kleine-koenig@pengutronix.de
Signed-off-by: Greg Kroah-Hartman 
Stable-dep-of: c766c90faf93 ("media: rcar_fdp1: Fix refcount leak in probe and remove function")
Signed-off-by: Sasha Levin

driver core: platform: Remove platform_device_add_properties()

2021-08-18T14:08:11+00:00

There are no more users for it. The last place where it's
called is in platform_device_register_full(). Replacing that
call with device_create_managed_software_node() and
removing the function.

Reviewed-by: Andy Shevchenko 
Reviewed-by: Thierry Reding 
Signed-off-by: Heikki Krogerus 
Link: https://lore.kernel.org/r/20210817102449.39994-3-heikki.krogerus@linux.intel.com
Signed-off-by: Greg Kroah-Hartman

bus: Make remove callback return void

2021-07-21T09:53:42+00:00

The driver core ignores the return value of this callback because there
is only little it can do when a device disappears.

This is the final bit of a long lasting cleanup quest where several
buses were converted to also return void from their remove callback.
Additionally some resource leaks were fixed that were caused by drivers
returning an error code in the expectation that the driver won't go
away.

With struct bus_type::remove returning void it's prevented that newly
implemented buses return an ignored error code and so don't anticipate
wrong expectations for driver authors.

Reviewed-by: Tom Rix  (For fpga)
Reviewed-by: Mathieu Poirier 
Reviewed-by: Cornelia Huck  (For drivers/s390 and drivers/vfio)
Acked-by: Russell King (Oracle)  (For ARM, Amba and related parts)
Acked-by: Mark Brown 
Acked-by: Chen-Yu Tsai  (for sunxi-rsb)
Acked-by: Pali Rohár 
Acked-by: Mauro Carvalho Chehab  (for media)
Acked-by: Hans de Goede  (For drivers/platform)
Acked-by: Alexandre Belloni 
Acked-By: Vinod Koul 
Acked-by: Juergen Gross  (For xen)
Acked-by: Lee Jones  (For mfd)
Acked-by: Johannes Thumshirn  (For mcb)
Acked-by: Johan Hovold 
Acked-by: Srinivas Kandagatla  (For slimbus)
Acked-by: Kirti Wankhede  (For vfio)
Acked-by: Maximilian Luz 
Acked-by: Heikki Krogerus  (For ulpi and typec)
Acked-by: Samuel Iglesias Gonsálvez  (For ipack)
Acked-by: Geoff Levand  (For ps3)
Acked-by: Yehezkel Bernat  (For thunderbolt)
Acked-by: Alexander Shishkin  (For intel_th)
Acked-by: Dominik Brodowski  (For pcmcia)
Acked-by: Rafael J. Wysocki  (For ACPI)
Acked-by: Bjorn Andersson  (rpmsg and apr)
Acked-by: Srinivas Pandruvada  (For intel-ish-hid)
Acked-by: Dan Williams  (For CXL, DAX, and NVDIMM)
Acked-by: William Breathitt Gray  (For isa)
Acked-by: Stefan Richter  (For firewire)
Acked-by: Benjamin Tissoires  (For hid)
Acked-by: Thorsten Scherer  (For siox)
Acked-by: Sven Van Asbroeck  (For anybuss)
Acked-by: Ulf Hansson  (For MMC)
Acked-by: Wolfram Sang  # for I2C
Acked-by: Sudeep Holla 
Acked-by: Geert Uytterhoeven 
Acked-by: Dmitry Torokhov 
Acked-by: Finn Thain 
Signed-off-by: Uwe Kleine-König 
Link: https://lore.kernel.org/r/20210713193522.1770306-6-u.kleine-koenig@pengutronix.de
Signed-off-by: Greg Kroah-Hartman

drivers/base: Constify static attribute_group structs

2021-06-04T13:06:28+00:00

These are only used by putting their address in an array of pointers to
const struct attribute_group (either directly or via the
__ATTRIBUTE_GROUP macro). Make them const to allow the compiler to place
them in read-only memory.

Signed-off-by: Rikard Falkeborn 
Link: https://lore.kernel.org/r/20210528213408.20067-1-rikard.falkeborn@gmail.com
Signed-off-by: Greg Kroah-Hartman

driver core: Drop helper devm_platform_ioremap_resource_wc()

2021-05-27T13:51:33+00:00

Since the macro was introduced in 2019 (commit bb6243b4f73d ("drivers:
platform: provide devm_platform_ioremap_resource_wc()") there is only a
single user which hardly justifies the function for the small task it
provides.

So drop the helper and open-code it in the only user. Adapt the non-wc
case accordingly.

For a all-mod-config build on amd64 this change introduces the following
changes according to bloat-o-meter:

add/remove: 0/1 grow/shrink: 1/0 up/down: 20/-252 (-232)
Function                                     old     new   delta
devm_platform_ioremap_resource_wc            252       -    -252
sram_probe                                   796     816     +20

Signed-off-by: Uwe Kleine-König 
Link: https://lore.kernel.org/r/20210525103711.956438-1-u.kleine-koenig@pengutronix.de
Signed-off-by: Greg Kroah-Hartman

Revert "driver core: platform: Make platform_get_irq_optional() optional"

2021-04-07T09:47:56+00:00

This reverts commit ed7027fdf4ec41ed6df6814956dc11860232a9d5 as it
causes runtime issues:
	https://lore.kernel.org/r/20210406192514.GA34677@roeck-us.net

Link: https://lore.kernel.org/r/20210406192514.GA34677@roeck-us.net
Reported-by: Guenter Roeck 
Cc: Matthias Schiffer 
Cc: Andy Shevchenko 
Signed-off-by: Greg Kroah-Hartman

driver core: platform: Make platform_get_irq_optional() optional

2021-04-02T15:06:34+00:00

Currently the platform_get_irq_optional() returns an error code even
if IRQ resource sumply has not been found. It prevents caller to be
error code agnostic in their error handling.

Now:
	ret = platform_get_irq_optional(...);
	if (ret != -ENXIO)
		return ret; // respect deferred probe
	if (ret > 0)
		...we get an IRQ...

After proposed change:
	ret = platform_get_irq_optional(...);
	if (ret < 0)
		return ret;
	if (ret > 0)
		...we get an IRQ...

Reported-by: Matthias Schiffer 
Signed-off-by: Andy Shevchenko 
Link: https://lore.kernel.org/r/20210331144526.19439-1-andriy.shevchenko@linux.intel.com
Signed-off-by: Greg Kroah-Hartman

driver core: platform: Make clear error code used for missed IRQ

2021-04-02T15:02:22+00:00

We have few code paths where same error code is assigned and
returned for missed IRQ. Unify that under single error path.

Signed-off-by: Andy Shevchenko 
Link: https://lore.kernel.org/r/20210331145937.35980-1-andriy.shevchenko@linux.intel.com
Signed-off-by: Greg Kroah-Hartman

Merge tag 'driver-core-5.12-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/driver-core

2021-02-24T18:13:55+00:00

Pull driver core / debugfs update from Greg KH:
 "Here is the "big" driver core and debugfs update for 5.12-rc1

  This set of driver core patches caused a bunch of problems in
  linux-next for the past few weeks, when Saravana tried to set
  fw_devlink=on as the default functionality. This caused a number of
  systems to stop booting, and lots of bugs were fixed in this area for
  almost all of the reported systems, but this option is not ready to be
  turned on just yet for the default operation based on this testing, so
  I've reverted that change at the very end so we don't have to worry
  about regressions in 5.12

  We will try to turn this on for 5.13 if testing goes better over the
  next few months.

  Other than the fixes caused by the fw_devlink testing in here, there's
  not much more:

   - debugfs fixes for invalid input into debugfs_lookup()

   - kerneldoc cleanups

   - warn message if platform drivers return an error on their remove
     callback (a futile effort, but good to catch).

  All of these have been in linux-next for a while now, and the
  regressions have gone away with the revert of the fw_devlink change"

* tag 'driver-core-5.12-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/driver-core: (35 commits)
  Revert "driver core: Set fw_devlink=on by default"
  of: property: fw_devlink: Ignore interrupts property for some configs
  debugfs: do not attempt to create a new file before the filesystem is initalized
  debugfs: be more robust at handling improper input in debugfs_lookup()
  driver core: auxiliary bus: Fix calling stage for auxiliary bus init
  of: irq: Fix the return value for of_irq_parse_one() stub
  of: irq: make a stub for of_irq_parse_one()
  clk: Mark fwnodes when their clock provider is added/removed
  PM: domains: Mark fwnodes when their powerdomain is added/removed
  irqdomain: Mark fwnodes when their irqdomain is added/removed
  driver core: fw_devlink: Handle suppliers that don't use driver core
  of: property: Add fw_devlink support for optional properties
  driver core: Add fw_devlink.strict kernel param
  of: property: Don't add links to absent suppliers
  driver core: fw_devlink: Detect supplier devices that will never be added
  driver core: platform: Emit a warning if a remove callback returned non-zero
  of: property: Fix fw_devlink handling of interrupts/interrupts-extended
  gpiolib: Don't probe gpio_device if it's not the primary device
  device.h: Remove bogus "the" in kerneldoc
  gpiolib: Bind gpio_device to a driver to enable fw_devlink=on by default
  ...