Skip to contentSkip to conversation

spi: bcm63xx: remove PM_SLEEP based conditional compilation

PATCH Applied Dhruva Gole 5 participants 9 messages Apr 20, 2023 – Apr 25, 2023

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 spi: bcm63xx: remove PM_SLEEP based conditional compilation

Get rid of conditional compilation based on CONFIG_PM_SLEEP because it may introduce build issues with certain configs where it maybe disabled This is because if above config is not enabled the suspend-resume functions are never part of the code but the bcm63xx_spi_pm_ops struct still inits them to non-existent suspend-resume functions.

Fixes b42dfed83d95 ("spi: add Broadcom BCM63xx SPI controller driver")
---
 drivers/spi/spi-bcm63xx.c | 2 --
 1 file changed, 2 deletions(-)
drivers/spi/spi-bcm63xx.c +0 -2
index 96633a0051b1..99395932074c 100644
@@ -617,7 +617,6 @@ static void bcm63xx_spi_remove(struct platform_device *pdev)
 	clk_disable_unprepare(bs->clk);
 }
 
-#ifdef CONFIG_PM_SLEEP
 static int bcm63xx_spi_suspend(struct device *dev)
 {
 	struct spi_master *master = dev_get_drvdata(dev);
@@ -644,7 +643,6 @@ static int bcm63xx_spi_resume(struct device *dev)
 
 	return 0;
 }
-#endif
 
 static const struct dev_pm_ops bcm63xx_spi_pm_ops = {
 	SET_SYSTEM_SLEEP_PM_OPS(bcm63xx_spi_suspend, bcm63xx_spi_resume)
Signature · 1 line
2.25.1
maintainer
#1 lore raw
Quoted 9 lines · Dhruva Gole

Get rid of conditional compilation based on CONFIG_PM_SLEEP because it may introduce build issues with certain configs where it maybe disabled This is because if above config is not enabled the suspend-resume functions are never part of the code but the bcm63xx_spi_pm_ops struct still inits them to non-existent suspend-resume functions.

Fixes b42dfed83d95 ("spi: add Broadcom BCM63xx SPI controller driver")

[...]

Applied to

https://git.kernel.org/pub/scm/linux/kernel/git/broonie/spi.git for-next

Thanks!

[1/1] spi: bcm63xx: remove PM_SLEEP based conditional compilation

      commit: 25f0617109496e1aff49594fbae5644286447a0f

All being well this means that it will be integrated into the linux-next tree (usually sometime in the next 24 hours) and sent to Linus during the next merge window (or sooner if it is a bug fix), however if problems are discovered then the patch may be dropped or reverted.

You may get further e-mails resulting from automated or manual testing and review of the tree, please engage with people reporting problems and send followup patches addressing any issues that are reported if needed.

If any updates are required or you are submitting further changes they should be sent as incremental updates against current git, existing patches will not be replaced.

Please add any relevant lists and maintainers to the CCs when replying to this mail.

Thanks,
Mark

#2 lore raw
In reply to Dhruva Gole
Quoted 18 lines · Dhruva Gole

Get rid of conditional compilation based on CONFIG_PM_SLEEP because it may introduce build issues with certain configs where it maybe disabled This is because if above config is not enabled the suspend-resume functions are never part of the code but the bcm63xx_spi_pm_ops struct still inits them to non-existent suspend-resume functions.

Fixes b42dfed83d95 ("spi: add Broadcom BCM63xx SPI controller driver")
---
  drivers/spi/spi-bcm63xx.c | 2 --
  1 file changed, 2 deletions(-)
drivers/spi/spi-bcm63xx.c +0 -0
index 96633a0051b1..99395932074c 100644

drivers/spi/spi-bcm63xx.c static void bcm63xx_spi_remove(struct platform_device *pdev) · line 620

Quoted 5 lines
drivers/spi/spi-bcm63xx.c +0 -1
@@ -617,7 +617,6 @@ static void bcm63xx_spi_remove(struct platform_device *pdev)
  	clk_disable_unprepare(bs->clk);
  }
 
-#ifdef CONFIG_PM_SLEEP
  static int bcm63xx_spi_suspend(struct device *dev)

Don't we need a __maybe_unused here?

Signature · 1 line
Florian
maintainer
#3 lore raw
Quoted 23 lines · Florian Fainelli
Quoted 21 lines · Dhruva Gole

Get rid of conditional compilation based on CONFIG_PM_SLEEP because it may introduce build issues with certain configs where it maybe disabled This is because if above config is not enabled the suspend-resume functions are never part of the code but the bcm63xx_spi_pm_ops struct still inits them to non-existent suspend-resume functions.

Fixes b42dfed83d95 ("spi: add Broadcom BCM63xx SPI controller driver")
---
  drivers/spi/spi-bcm63xx.c | 2 --
  1 file changed, 2 deletions(-)
drivers/spi/spi-bcm63xx.c +0 -0
index 96633a0051b1..99395932074c 100644
@@ -617,7 +617,6 @@ static void bcm63xx_spi_remove(struct platform_device *pdev)
      clk_disable_unprepare(bs->clk);
  }
 
Quoted 4 lines

drivers/spi/spi-bcm63xx.c static void bcm63xx_spi_remove(struct platform_device *pdev) · line 620

Quoted 2 lines
drivers/spi/spi-bcm63xx.c +0 -1
-#ifdef CONFIG_PM_SLEEP
  static int bcm63xx_spi_suspend(struct device *dev)

Don't we need a __maybe_unused here?

Actually the premise of this patch is wrong, and should rather be reverted.

The bcm63xx_spi_pm_ops struct is initialized with
SET_SYSTEM_SLEEP_PM_OPS(), which is defined as

#ifdef CONFIG_PM_SLEEP
#define SET_SYSTEM_SLEEP_PM_OPS(suspend_fn, resume_fn) \
SYSTEM_SLEEP_PM_OPS(suspend_fn, resume_fn) #else
#define SET_SYSTEM_SLEEP_PM_OPS(suspend_fn, resume_fn) #endif

so for !CONFIG_PM_SLEEP it won't initialize the struct at all (or reference non-existing functions), and therefore there will be no build issues.

Regards,
Jonas

#4 lore raw

Hi Jonas,

Quoted 34 lines · Jonas Gorski
Quoted 23 lines · Florian Fainelli
Quoted 21 lines · Dhruva Gole

Get rid of conditional compilation based on CONFIG_PM_SLEEP because it may introduce build issues with certain configs where it maybe disabled This is because if above config is not enabled the suspend-resume functions are never part of the code but the bcm63xx_spi_pm_ops struct still inits them to non-existent suspend-resume functions.

Fixes b42dfed83d95 ("spi: add Broadcom BCM63xx SPI controller driver")
---
   drivers/spi/spi-bcm63xx.c | 2 --
   1 file changed, 2 deletions(-)
drivers/spi/spi-bcm63xx.c +0 -0
index 96633a0051b1..99395932074c 100644
@@ -617,7 +617,6 @@ static void bcm63xx_spi_remove(struct platform_device *pdev)
       clk_disable_unprepare(bs->clk);
   }
 
Quoted 4 lines

drivers/spi/spi-bcm63xx.c static void bcm63xx_spi_remove(struct platform_device *pdev) · line 620

Quoted 2 lines
drivers/spi/spi-bcm63xx.c +0 -1
-#ifdef CONFIG_PM_SLEEP
   static int bcm63xx_spi_suspend(struct device *dev)

Don't we need a __maybe_unused here?

Actually the premise of this patch is wrong, and should rather be reverted.

The bcm63xx_spi_pm_ops struct is initialized with
SET_SYSTEM_SLEEP_PM_OPS(), which is defined as

Quoted 7 lines

#ifdef CONFIG_PM_SLEEP
#define SET_SYSTEM_SLEEP_PM_OPS(suspend_fn, resume_fn) \
SYSTEM_SLEEP_PM_OPS(suspend_fn, resume_fn) #else
#define SET_SYSTEM_SLEEP_PM_OPS(suspend_fn, resume_fn) #endif

Thanks for pointing this out, I must have missed this. Anyway, I have sent another patch to migrate to using DEFINE_SIMPLE_DEV_PM_OPS as per Mark's suggestion [0]. There I think it would be necessary to remove the CONFIG_PM_SLEEP checks in the driver. So no need to revert this patch.

Quoted 6 lines

so for !CONFIG_PM_SLEEP it won't initialize the struct at all (or reference non-existing functions), and therefore there will be no build issues.

Regards,
Jonas

[0] https://lore.kernel.org/all/[email protected]/

Signature · 7 lines
Thanks and Regards,
Dhruva Gole

_______________________________________________
linux-arm-kernel mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel
#5 lore raw
In reply to Dhruva Gole
Quoted 9 lines · Dhruva Gole

Get rid of conditional compilation based on CONFIG_PM_SLEEP because it may introduce build issues with certain configs where it maybe disabled This is because if above config is not enabled the suspend-resume functions are never part of the code but the bcm63xx_spi_pm_ops struct still inits them to non-existent suspend-resume functions.

Fixes b42dfed83d95 ("spi: add Broadcom BCM63xx SPI controller driver")

This patch results in:

drivers/spi/spi-bcm63xx.c:632:12: error: 'bcm63xx_spi_resume' defined but not used [-Werror=unused-function]
632 | static int bcm63xx_spi_resume(struct device *dev)

      |            ^~~~~~~~~~~~~~~~~~

drivers/spi/spi-bcm63xx.c:620:12: error: 'bcm63xx_spi_suspend' defined but not used [-Werror=unused-function]
620 | static int bcm63xx_spi_suspend(struct device *dev)

on architectures with no PM support (alpha, csky, m68k, openrisc, parisc, riscv, s390).

Guenter

#6 lore raw

Hi,

Quoted 12 lines · Guenter Roeck
Quoted 9 lines · Dhruva Gole

Get rid of conditional compilation based on CONFIG_PM_SLEEP because it may introduce build issues with certain configs where it maybe disabled This is because if above config is not enabled the suspend-resume functions are never part of the code but the bcm63xx_spi_pm_ops struct still inits them to non-existent suspend-resume functions.

Fixes b42dfed83d95 ("spi: add Broadcom BCM63xx SPI controller driver")

This patch results in:

Quoted 8 lines

drivers/spi/spi-bcm63xx.c:632:12: error: 'bcm63xx_spi_resume' defined but not used [-Werror=unused-function]
632 | static int bcm63xx_spi_resume(struct device *dev)

      |            ^~~~~~~~~~~~~~~~~~

drivers/spi/spi-bcm63xx.c:620:12: error: 'bcm63xx_spi_suspend' defined but not used [-Werror=unused-function]
620 | static int bcm63xx_spi_suspend(struct device *dev)

on architectures with no PM support (alpha, csky, m68k, openrisc, parisc, riscv, s390).

Thanks for pointing this out.

I could send a patch to address this as pointed here [0]

by adding a __maybe_unused.

However, do you think my other patch [1] could address this issue without the need for above?

I think it would because DEFINE_SIMPLE_DEV_PM_OPS doesn't seem to be under any conditional CONFIG_PM.

However, I may have missed something, please do let me know what's the best way to fix.

[0] https://lore.kernel.org/all/[email protected]/

[1] https://lore.kernel.org/all/[email protected]/

Quoted 2 lines

Guenter

Signature · 2 lines
Regards,
Dhruva Gole <[email protected]>
#7 lore raw
Quoted 24 lines · Dhruva

Hi,

Quoted 12 lines · Guenter Roeck
Quoted 9 lines · Dhruva Gole

Get rid of conditional compilation based on CONFIG_PM_SLEEP because it may introduce build issues with certain configs where it maybe disabled This is because if above config is not enabled the suspend-resume functions are never part of the code but the bcm63xx_spi_pm_ops struct still inits them to non-existent suspend-resume functions.

Fixes b42dfed83d95 ("spi: add Broadcom BCM63xx SPI controller driver")

This patch results in:

Quoted 8 lines

drivers/spi/spi-bcm63xx.c:632:12: error: 'bcm63xx_spi_resume' defined but not used [-Werror=unused-function]
632 | static int bcm63xx_spi_resume(struct device *dev)

       |            ^~~~~~~~~~~~~~~~~~

drivers/spi/spi-bcm63xx.c:620:12: error: 'bcm63xx_spi_suspend' defined but not used [-Werror=unused-function]
620 | static int bcm63xx_spi_suspend(struct device *dev)

on architectures with no PM support (alpha, csky, m68k, openrisc, parisc, riscv, s390).

Quoted 8 lines

Thanks for pointing this out.

I could send a patch to address this as pointed here [0]

by adding a __maybe_unused.

However, do you think my other patch [1] could address this issue without the need for above?

Personally I would go for [0] as the least invasive solution, but I really have no idea, sorry. I just hope that your (broken) patch doesn't make it as-is into the upstream kernel.

Guenter

Quoted 11 lines

I think it would because DEFINE_SIMPLE_DEV_PM_OPS doesn't seem to be under any conditional CONFIG_PM.

However, I may have missed something, please do let me know what's the best way to fix.

[0] https://lore.kernel.org/all/[email protected]/

[1] https://lore.kernel.org/all/[email protected]/

Quoted 2 lines

Guenter

maintainer Applied
#8 lore raw
Quoted 3 lines · Guenter Roeck

Personally I would go for [0] as the least invasive solution, but I really have no idea, sorry. I just hope that your (broken) patch doesn't make it as-is into the upstream kernel.

I've applied the SIMPLE_DEV_PM_OPS patch which seems to fix the issue for riscv.