| Message ID | 20240814121649.261693-1-rf@opensource.cirrus.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | i2c: Use IS_REACHABLE() for substituting empty ACPI functions | expand |
On Wed, 14 Aug 2024 14:16:49 +0200, Richard Fitzgerald wrote: > > Replace IS_ENABLED() with IS_REACHABLE() to substitute empty stubs for: > i2c_acpi_get_i2c_resource() > i2c_acpi_client_count() > i2c_acpi_find_bus_speed() > i2c_acpi_new_device_by_fwnode() > i2c_adapter *i2c_acpi_find_adapter_by_handle() > i2c_acpi_waive_d0_probe() > > commit f17c06c6608a ("i2c: Fix conditional for substituting empty ACPI > functions") partially fixed this conditional to depend on CONFIG_I2C, > but used IS_ENABLED(), which is wrong since CONFIG_I2C is tristate. > > CONFIG_ACPI is boolean but let's also change it to use IS_REACHABLE() > to future-proof it against becoming tristate. > > Somehow despite testing various combinations of CONFIG_I2C and CONFIG_ACPI > we missed the combination CONFIG_I2C=m, CONFIG_ACPI=y. > > Signed-off-by: Richard Fitzgerald <rf@opensource.cirrus.com> > Fixes: f17c06c6608a ("i2c: Fix conditional for substituting empty ACPI functions") > Reported-by: kernel test robot <lkp@intel.com> > Closes: https://lore.kernel.org/oe-kbuild-all/202408141333.gYnaitcV-lkp@intel.com/ I suppose Wolfram will take this? Reviewed-by: Takashi Iwai <tiwai@suse.de> thanks, Takashi
> > Signed-off-by: Richard Fitzgerald <rf@opensource.cirrus.com> > > Fixes: f17c06c6608a ("i2c: Fix conditional for substituting empty ACPI functions") > > Reported-by: kernel test robot <lkp@intel.com> > > Closes: https://lore.kernel.org/oe-kbuild-all/202408141333.gYnaitcV-lkp@intel.com/ > > I suppose Wolfram will take this? > > Reviewed-by: Takashi Iwai <tiwai@suse.de> Yes, for rc4. I wonder why buildbot did not spot it for the branch I submitted to rc3. Things happen, I guess...
On Wed, 14 Aug 2024 16:00:24 +0200, Wolfram Sang wrote: > > > > > Signed-off-by: Richard Fitzgerald <rf@opensource.cirrus.com> > > > Fixes: f17c06c6608a ("i2c: Fix conditional for substituting empty ACPI functions") > > > Reported-by: kernel test robot <lkp@intel.com> > > > Closes: https://lore.kernel.org/oe-kbuild-all/202408141333.gYnaitcV-lkp@intel.com/ > > > > I suppose Wolfram will take this? > > > > Reviewed-by: Takashi Iwai <tiwai@suse.de> > > Yes, for rc4. I wonder why buildbot did not spot it for the branch I > submitted to rc3. Things happen, I guess... Maybe because there was no user that can hit with randconfig. It was a preliminary fix for the change I recently put after the rc3 release, and now it surfaced. thanks, Takashi
> It was a preliminary fix for the change I recently put after the rc3 > release, and now it surfaced. Ah, that's an explanation. Thanks!
On Wed, Aug 14, 2024 at 01:16:49PM +0100, Richard Fitzgerald wrote: > Replace IS_ENABLED() with IS_REACHABLE() to substitute empty stubs for: > i2c_acpi_get_i2c_resource() > i2c_acpi_client_count() > i2c_acpi_find_bus_speed() > i2c_acpi_new_device_by_fwnode() > i2c_adapter *i2c_acpi_find_adapter_by_handle() > i2c_acpi_waive_d0_probe() > > commit f17c06c6608a ("i2c: Fix conditional for substituting empty ACPI > functions") partially fixed this conditional to depend on CONFIG_I2C, > but used IS_ENABLED(), which is wrong since CONFIG_I2C is tristate. > > CONFIG_ACPI is boolean but let's also change it to use IS_REACHABLE() > to future-proof it against becoming tristate. > > Somehow despite testing various combinations of CONFIG_I2C and CONFIG_ACPI > we missed the combination CONFIG_I2C=m, CONFIG_ACPI=y. > > Signed-off-by: Richard Fitzgerald <rf@opensource.cirrus.com> > Fixes: f17c06c6608a ("i2c: Fix conditional for substituting empty ACPI functions") > Reported-by: kernel test robot <lkp@intel.com> > Closes: https://lore.kernel.org/oe-kbuild-all/202408141333.gYnaitcV-lkp@intel.com/ Applied to for-current, thanks!
diff --git a/include/linux/i2c.h b/include/linux/i2c.h index 7eedd0c662da..377def497298 100644 --- a/include/linux/i2c.h +++ b/include/linux/i2c.h @@ -1066,7 +1066,7 @@ static inline int of_i2c_get_board_info(struct device *dev, struct acpi_resource; struct acpi_resource_i2c_serialbus; -#if IS_ENABLED(CONFIG_ACPI) && IS_ENABLED(CONFIG_I2C) +#if IS_REACHABLE(CONFIG_ACPI) && IS_REACHABLE(CONFIG_I2C) bool i2c_acpi_get_i2c_resource(struct acpi_resource *ares, struct acpi_resource_i2c_serialbus **i2c); int i2c_acpi_client_count(struct acpi_device *adev);
Replace IS_ENABLED() with IS_REACHABLE() to substitute empty stubs for: i2c_acpi_get_i2c_resource() i2c_acpi_client_count() i2c_acpi_find_bus_speed() i2c_acpi_new_device_by_fwnode() i2c_adapter *i2c_acpi_find_adapter_by_handle() i2c_acpi_waive_d0_probe() commit f17c06c6608a ("i2c: Fix conditional for substituting empty ACPI functions") partially fixed this conditional to depend on CONFIG_I2C, but used IS_ENABLED(), which is wrong since CONFIG_I2C is tristate. CONFIG_ACPI is boolean but let's also change it to use IS_REACHABLE() to future-proof it against becoming tristate. Somehow despite testing various combinations of CONFIG_I2C and CONFIG_ACPI we missed the combination CONFIG_I2C=m, CONFIG_ACPI=y. Signed-off-by: Richard Fitzgerald <rf@opensource.cirrus.com> Fixes: f17c06c6608a ("i2c: Fix conditional for substituting empty ACPI functions") Reported-by: kernel test robot <lkp@intel.com> Closes: https://lore.kernel.org/oe-kbuild-all/202408141333.gYnaitcV-lkp@intel.com/ --- include/linux/i2c.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)