public inbox for linux-arm-kernel@lists.infradead.org 
 help / color / mirror / Atom feed
* [PATCH v1 0/6] power: Use named initializers for platform_device_id arrays
@ 2026-05-29 10:18 Uwe Kleine-König (The Capable Hub)
  2026-05-29 10:18 ` [PATCH v1 1/6] power: Drop unused assignment of platform_device_id driver data Uwe Kleine-König (The Capable Hub)
                   ` (3 more replies)
  0 siblings, 4 replies; 9+ messages in thread
From: Uwe Kleine-König (The Capable Hub) @ 2026-05-29 10:18 UTC (permalink / raw)
  To: Sebastian Reichel
  Cc: Kuan-Wei Chiu, Benson Leung, Guenter Roeck, Thomas Weißschuh,
	Krzysztof Kozlowski, Matthias Brugger, AngeloGioacchino Del Regno,
	linux-pm, linux-kernel, chrome-platform, linux-arm-kernel,
	linux-mediatek, Hans de Goede, Marek Szyprowski,
	Sebastian Krzyszkowiak, Purism Kernel Team, Yixun Lan,
	Andreas Kemnade, Matti Vaittinen, Sven Peter, Janne Grunau,
	Neal Gompa, Amit Sunil Dhamne, Samuel Kayode, linux-riscv,
	spacemit, asahi, imx, Chen-Yu Tsai

Hello,

this series targets to use named initializers for platform_device_id
arrays. In general these are better readable for humans and more robust
to changes in the respective struct definition.

This robustness is needed as I want to do

	diff --git a/include/linux/mod_devicetable.h b/include/linux/mod_devicetable.h
	--- a/include/linux/mod_devicetable.h
	+++ b/include/linux/mod_devicetable.h
	@@ -610,4 +610,7 @@ struct dmi_system_id {
	 struct platform_device_id {
		char name[PLATFORM_NAME_SIZE];
	-	kernel_ulong_t driver_data;
	+	union {
	+		kernel_ulong_t driver_data;
	+		const void *driver_data_ptr;
	+	};
	 };

which allows dropping several casts and eases porting CHERI to mainline
linux. A possible follow-up change is the following example:

	diff --git a/drivers/gpio/gpio-pxa.c b/drivers/gpio/gpio-pxa.c
	index 5d61053e0596..03bc8e859d73 100644
	--- a/drivers/gpio/gpio-pxa.c
	+++ b/drivers/gpio/gpio-pxa.c
	@@ -534,7 +534,7 @@ static struct irq_chip pxa_muxed_gpio_chip = {
	 static int pxa_gpio_nums(struct platform_device *pdev)
	 {
		const struct platform_device_id *id = platform_get_device_id(pdev);
	-	struct pxa_gpio_id *pxa_id = (struct pxa_gpio_id *)id->driver_data;
	+	struct pxa_gpio_id *pxa_id = id->driver_data_ptr;
		int count = 0;
	 
		switch (pxa_id->type) {
	@@ -708,14 +708,14 @@ static int pxa_gpio_probe(struct platform_device *pdev)
	 }
	 
	 static const struct platform_device_id gpio_id_table[] = {
	-	{ .name = "pxa25x-gpio",	.driver_data = (unsigned long)&pxa25x_id },
	-	{ .name = "pxa26x-gpio",	.driver_data = (unsigned long)&pxa26x_id },
	-	{ .name = "pxa27x-gpio",	.driver_data = (unsigned long)&pxa27x_id },
	-	{ .name = "pxa3xx-gpio",	.driver_data = (unsigned long)&pxa3xx_id },
	-	{ .name = "pxa93x-gpio",	.driver_data = (unsigned long)&pxa93x_id },
	-	{ .name = "mmp-gpio",		.driver_data = (unsigned long)&mmp_id },
	-	{ .name = "mmp2-gpio",		.driver_data = (unsigned long)&mmp2_id },
	-	{ .name = "pxa1928-gpio",	.driver_data = (unsigned long)&pxa1928_id },
	+	{ .name = "pxa25x-gpio",	.driver_data_ptr = &pxa25x_id },
	+	{ .name = "pxa26x-gpio",	.driver_data_ptr = &pxa26x_id },
	+	{ .name = "pxa27x-gpio",	.driver_data_ptr = &pxa27x_id },
	+	{ .name = "pxa3xx-gpio",	.driver_data_ptr = &pxa3xx_id },
	+	{ .name = "pxa93x-gpio",	.driver_data_ptr = &pxa93x_id },
	+	{ .name = "mmp-gpio",		.driver_data_ptr = &mmp_id },
	+	{ .name = "mmp2-gpio",		.driver_data_ptr = &mmp2_id },
	+	{ .name = "pxa1928-gpio",	.driver_data_ptr = &pxa1928_id },
		{ }
	 };
 
increasing readability due to less casting. Also this results in the
compiler warning:

	drivers/gpio/gpio-pxa.c: In function ‘pxa_gpio_nums’:
	drivers/gpio/gpio-pxa.c:537:38: error: initialization discards ‘const’ qualifier from pointer target type [-Werror=discarded-qualifiers]
	  537 |         struct pxa_gpio_id *pxa_id = id->driver_data_ptr;
	      |                                      ^~

which is a good thing as adding the needed const to fix this warning
improves type safety. (Sorry, no driver below drivers/power benefits
here, so the example is from a different subsystem.)

Patch #5 is just an unrelated change where I spotted the patch
opportunity while preparing this series. It's included here just for my
convenience and could be applied independently.

If you consider the last patch mostly churn, just drop it.

There are no dependencies between the patches, the only file that is
touched in more than one patch is drivers/power/supply/mt6360_charger.c
and even if these (i.e. patches #1 and #5) are reordered, git should be
able to cope with that without conflicts.

Note that this series is only build tested, and there is the possibility
that I failed to spot driver data usage that render patches #2 and/or #3
wrong. So a deeper look on at least those two won't hurt.

Best regards
Uwe

Uwe Kleine-König (The Capable Hub) (6):
  power: Drop unused assignment of platform_device_id driver data
  power: supply: max14577: Drop driver data in of and platform device id
    arrays
  power: supply: max17042: Drop driver data in of and platform device id
    arrays
  power: Use named initializers for platform_device_id arrays
  power: supply: mt6360_charger: Use of match table unconditionally
  power: Unify code style for platform_device_id arrays

 drivers/power/reset/qemu-virt-ctrl.c           |  2 +-
 drivers/power/reset/spacemit-p1-reboot.c       |  4 ++--
 drivers/power/reset/tps65086-restart.c         |  2 +-
 drivers/power/supply/axp288_charger.c          |  2 +-
 drivers/power/supply/axp288_fuel_gauge.c       |  2 +-
 drivers/power/supply/bd71828-power.c           |  8 ++++----
 drivers/power/supply/charger-manager.c         |  4 ++--
 drivers/power/supply/cros_charge-control.c     |  4 ++--
 drivers/power/supply/cros_peripheral_charger.c |  4 ++--
 drivers/power/supply/cros_usbpd-charger.c      |  4 ++--
 drivers/power/supply/macsmc-power.c            |  2 +-
 drivers/power/supply/max14577_charger.c        | 12 +++++-------
 drivers/power/supply/max17042_battery.c        | 14 +++++++-------
 drivers/power/supply/max77693_charger.c        |  2 +-
 drivers/power/supply/max77759_charger.c        |  2 +-
 drivers/power/supply/max8997_charger.c         |  2 +-
 drivers/power/supply/max8998_charger.c         |  2 +-
 drivers/power/supply/mt6360_charger.c          |  6 +++---
 drivers/power/supply/pf1550-charger.c          |  2 +-
 drivers/power/supply/rt5033_charger.c          |  2 +-
 20 files changed, 40 insertions(+), 42 deletions(-)


base-commit: f7af91adc230aa99e23330ecf85bc9badd9780ad
-- 
2.47.3



^ permalink raw reply	[flat|nested] 9+ messages in thread
end of thread, other threads:[~2026-06-03 20:47 UTC | newest]

Thread overview: 9+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2026-05-29 10:18 [PATCH v1 0/6] power: Use named initializers for platform_device_id arrays Uwe Kleine-König (The Capable Hub)
2026-05-29 10:18 ` [PATCH v1 1/6] power: Drop unused assignment of platform_device_id driver data Uwe Kleine-König (The Capable Hub)
2026-05-29 22:02   ` Tzung-Bi Shih
2026-05-29 10:18 ` [PATCH v1 4/6] power: Use named initializers for platform_device_id arrays Uwe Kleine-König (The Capable Hub)
2026-05-29 11:29   ` Joshua Peisach
2026-05-29 18:49     ` Uwe Kleine-König (The Capable Hub)
2026-06-01  6:48   ` Matti Vaittinen
2026-05-29 10:18 ` [PATCH v1 5/6] power: supply: mt6360_charger: Use of match table unconditionally Uwe Kleine-König (The Capable Hub)
2026-06-03 20:46 ` (subset) [PATCH v1 0/6] power: Use named initializers for platform_device_id arrays Sebastian Reichel

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox