On 2023-11-01 09:19, Jaehoon Chung wrote:
-----Original Message----- From: U-Boot u-boot-bounces@lists.denx.de On Behalf Of Jaehoon Chung Sent: Wednesday, November 1, 2023 5:11 PM To: 'Jonas Karlman' jonas@kwiboo.se Cc: sjg@chromium.org; patrice.chotard@foss.st.com; eugen.hristev@collabora.com; trini@konsulko.com; u- boot@lists.denx.de Subject: RE: [PATCH] power: regulator: Fix an handling error about EALREADY
-----Original Message----- From: U-Boot u-boot-bounces@lists.denx.de On Behalf Of Jaehoon Chung Sent: Wednesday, November 1, 2023 5:07 PM To: 'Jonas Karlman' jonas@kwiboo.se Cc: sjg@chromium.org; patrice.chotard@foss.st.com; eugen.hristev@collabora.com; trini@konsulko.com;
u-
boot@lists.denx.de Subject: RE: [PATCH] power: regulator: Fix an handling error about EALREADY
-----Original Message----- From: Jonas Karlman jonas@kwiboo.se Sent: Wednesday, November 1, 2023 4:47 PM To: Jaehoon Chung jh80.chung@samsung.com Cc: sjg@chromium.org; patrice.chotard@foss.st.com; eugen.hristev@collabora.com; trini@konsulko.com;
u-
boot@lists.denx.de Subject: Re: [PATCH] power: regulator: Fix an handling error about EALREADY
On 2023-11-01 08:23, Jaehoon Chung wrote:
If reegulator is already enabled, it will be return to EALREADY. But driver that its function is called can notice as error, even though it's working fine.
Fixes: 4fcba5d556b ("regulator: implement basic reference counter")
Signed-off-by: Jaehoon Chung jh80.chung@samsung.com
drivers/power/regulator/regulator-uclass.c | 4 ++++ 1 file changed, 4 insertions(+)
diff --git a/drivers/power/regulator/regulator-uclass.c b/drivers/power/regulator/regulator-
uclass.c
index 3a6ba69f6d5f..fc1c3eb93c9d 100644 --- a/drivers/power/regulator/regulator-uclass.c +++ b/drivers/power/regulator/regulator-uclass.c @@ -187,6 +187,10 @@ int regulator_set_enable(struct udevice *dev, bool enable) } }
- /* Regulator is already enabled */
- if (ret == -EALREADY)
return 0;Use of regulator_set_enable_if_allowed() will cover this error, and regulator_set_enable() should continue to return this error.
regulator_set_enable_if_allowed() can be covered. But regulator_set_enable is called in some drivers. You means that it needs to replace to regulator_set_enable_if_allowed() from regulator_set_enable() about all driver?
I think that was the consensus when basic reference counter was implemented for gpio/fixed regulators. It should return -EALREADY when a regulator is already enabled, and if a caller need a more relaxed enable/disable regulator_set_enable_if_allowed() could be used.
Tried to fix a few commons callers in following series, but did not cover hw I could not test. Can see that I missed the meson usb driver and was something I could have tested.
Keep fixed/gpio regulator enable count in balance https://patchwork.ozlabs.org/cover/1810049/
Regards, Jonas
Best Regards, Jaehoon Chung
Well.. I will recheck about your comment.
Best Regards, Jaehoon Chung
When I have checked on my target, It seems that it can't cover all cases.
On odroid-c4, USB doesn't work, even though its regulator is enabled.
=> ums 0 mmc 0 UMS: LUN 0, dev mmc 0, hwpart 0, sector 0x0, count 0x1dacc00 No USB device found Couldn't init USB controller.
Best Regards, Jaehoon Chung
Regards, Jonas
return ret; }