| Message ID | 20221130152148.2769768-1-u.kleine-koenig@pengutronix.de (mailing list archive) |
|---|---|
| Headers | show |
| Series | pwm: Allow .get_state to fail | expand |
Hey Uwe! On Wed, Nov 30, 2022 at 04:21:37PM +0100, Uwe Kleine-König wrote: > Hello, > > I forgot about this series and was remembered when I talked to Conor > Dooley about how .get_state() should behave in an error case. In the context of "my" driver, get_state() the proposal was to fail with -ETIMEDOUT rather than block a caller, potentially, for seconds or report a potentially "random" state. Specifically, values writen to the registers that control the PWM duty cycle are not visible to the cpu until the changes have propagated to the waveform at the start of a new period. The timeout would occur if the bit that signifies that the "shadow registers" contain a value which has not yet propagated. This bit is per PWM "controller" and not per PWM channel. Returning from apply() without waiting, possibly for seconds, for the writes to become visible could cause get_state() to see anything between the new and old states, inclusive! If anyone cares at all, the discussion is here: https://lore.kernel.org/linux-pwm/20221110093512.333881-1-conor.dooley@microchip.com/T/#m800eeabad29067940a5684e54106fd0bb7261944 > In v1 Thierry had the concern: > > | That raises the question about what to do in these cases. If we return > | an error, that could potentially throw off consumers. So perhaps the > | closest would be to return a disabled PWM? > | Or perhaps it'd be up to the > | consumer to provide some fallback configuration for invalidly configured > | or unconfigured PWMs. > > .get_state() is only called in pwm_device_request on a pwm_state that a > consumer might see. Before my series a consumer might have seen a > partial modified pwm_state (because .get_state() might have modified > .period, then stumbled and returned silently). The last patch ensures > that this partial modification isn't given out to the consumer. Instead > they now see the same as if .get_state wasn't implemented at all. Getting the same thing as if get_state() did not exist seems preferable to me in this context than "lying" and pretending that a PWM is disabled or potentially inconsistent reports from get_state() that I mentioned above. TL;DR, I quite like the ability to return an error and not mislead the caller. Thanks for sending a v2 of this so quickly :) Conor.
Hello Conor, On Thu, Dec 01, 2022 at 11:11:51AM +0000, Conor Dooley wrote: > TL;DR, I quite like the ability to return an error and not mislead the > caller. Is this an Ack? Best regards Uwe
On Thu, Dec 01, 2022 at 02:19:07PM +0100, Uwe Kleine-König wrote: > Hello Conor, > > On Thu, Dec 01, 2022 at 11:11:51AM +0000, Conor Dooley wrote: > > TL;DR, I quite like the ability to return an error and not mislead the > > caller. > > Is this an Ack? It is if you want it to be! I didn't really feel qualified to do so which is why I gave some context etc. I did check out the callsites for the non-void returning op, and it looked good to me, so sure, why not: Acked-by: Conor Dooley <conor.dooley@microchip.com> Thanks, Conor.
On Wed, Nov 30, 2022 at 04:21:37PM +0100, Uwe Kleine-König wrote: > Hello, > > I forgot about this series and was remembered when I talked to Conor > Dooley about how .get_state() should behave in an error case. > > Compared to (implicit) v1, sent with Message-Id: 20220916151506.298488-1-u.kleine-koenig@pengutronix.de > I changed: > > - Patch #1 which does the prototype change now just adds "return 0" to > all implementations and so gets simpler and doesn't change behaviour. > The adaptions to the different .get_state() implementations are split > out into individual patches to ease review. > - One minor inconsistency fixed in "pwm: Handle .get_state() failures" > that I noticed while looking into this patch. > - I skipped changing sun4i.c as I don't know how to handle the error > there. Someone might want to have a look. (That's not ideal, but it's > not worse than the same issue before this series.) > > In v1 Thierry had the concern: > > | That raises the question about what to do in these cases. If we return > | an error, that could potentially throw off consumers. So perhaps the > | closest would be to return a disabled PWM? Or perhaps it'd be up to the > | consumer to provide some fallback configuration for invalidly configured > | or unconfigured PWMs. > > .get_state() is only called in pwm_device_request on a pwm_state that a > consumer might see. Before my series a consumer might have seen a > partial modified pwm_state (because .get_state() might have modified > .period, then stumbled and returned silently). The last patch ensures > that this partial modification isn't given out to the consumer. Instead > they now see the same as if .get_state wasn't implemented at all. I'm wondering why we didn't see a compiler warning about mistyped function prototypes in some drivers. P.S. The series is good thing to do, thank you.
Hello Andy, On Fri, Dec 09, 2022 at 11:47:54PM +0200, Andy Shevchenko wrote: > On Wed, Nov 30, 2022 at 04:21:37PM +0100, Uwe Kleine-König wrote: > > In v1 Thierry had the concern: > > > > | That raises the question about what to do in these cases. If we return > > | an error, that could potentially throw off consumers. So perhaps the > > | closest would be to return a disabled PWM? Or perhaps it'd be up to the > > | consumer to provide some fallback configuration for invalidly configured > > | or unconfigured PWMs. > > > > .get_state() is only called in pwm_device_request on a pwm_state that a > > consumer might see. Before my series a consumer might have seen a > > partial modified pwm_state (because .get_state() might have modified > > .period, then stumbled and returned silently). The last patch ensures > > that this partial modification isn't given out to the consumer. Instead > > they now see the same as if .get_state wasn't implemented at all. > > I'm wondering why we didn't see a compiler warning about mistyped function > prototypes in some drivers. I don't understand where you expected a warning. Care to elaborate? > P.S. The series is good thing to do, thank you. It's already too late for an ack, the series is already in Thierry's tree. Best regards Uwe
On Sat, Dec 10, 2022 at 10:18:33AM +0100, Uwe Kleine-König wrote: > On Fri, Dec 09, 2022 at 11:47:54PM +0200, Andy Shevchenko wrote: > > On Wed, Nov 30, 2022 at 04:21:37PM +0100, Uwe Kleine-König wrote: ... > > I'm wondering why we didn't see a compiler warning about mistyped function > > prototypes in some drivers. > > I don't understand where you expected a warning. Care to elaborate? intel-lpss.c has the prototype that returns an int. IIRC it was like this before your patches. Now the above wondering passage...
Hello Andy, On Sat, Dec 10, 2022 at 10:57:16PM +0200, Andy Shevchenko wrote: > On Sat, Dec 10, 2022 at 10:18:33AM +0100, Uwe Kleine-König wrote: > > On Fri, Dec 09, 2022 at 11:47:54PM +0200, Andy Shevchenko wrote: > > > On Wed, Nov 30, 2022 at 04:21:37PM +0100, Uwe Kleine-König wrote: > > ... > > > > I'm wondering why we didn't see a compiler warning about mistyped function > > > prototypes in some drivers. > > > > I don't understand where you expected a warning. Care to elaborate? > > intel-lpss.c has the prototype that returns an int. IIRC it was like this Do you mean drivers/mfd/intel-lpss.c? That one doesn't implement a PWM?! And drivers/pwm/pwm-lpss.c is adapted by this series. One of us is confused ... Best regards Uwe
On Sat, Dec 10, 2022 at 11:41:54PM +0100, Uwe Kleine-König wrote: > On Sat, Dec 10, 2022 at 10:57:16PM +0200, Andy Shevchenko wrote: > > On Sat, Dec 10, 2022 at 10:18:33AM +0100, Uwe Kleine-König wrote: > > > On Fri, Dec 09, 2022 at 11:47:54PM +0200, Andy Shevchenko wrote: > > > > On Wed, Nov 30, 2022 at 04:21:37PM +0100, Uwe Kleine-König wrote: ... > > > > I'm wondering why we didn't see a compiler warning about mistyped function > > > > prototypes in some drivers. > > > > > > I don't understand where you expected a warning. Care to elaborate? > > > > intel-lpss.c has the prototype that returns an int. IIRC it was like this > > Do you mean drivers/mfd/intel-lpss.c? That one doesn't implement a PWM?! Nope, I meant pwm-lpss.c. > And drivers/pwm/pwm-lpss.c is adapted by this series. That's what I didn't see how. > One of us is confused ... Yes, because when I have checked the branch based on Linux Next I already saw that get_state() returns a code and I wasn't aware that the series is already there.
Hello, I forgot about this series and was remembered when I talked to Conor Dooley about how .get_state() should behave in an error case. Compared to (implicit) v1, sent with Message-Id: 20220916151506.298488-1-u.kleine-koenig@pengutronix.de I changed: - Patch #1 which does the prototype change now just adds "return 0" to all implementations and so gets simpler and doesn't change behaviour. The adaptions to the different .get_state() implementations are split out into individual patches to ease review. - One minor inconsistency fixed in "pwm: Handle .get_state() failures" that I noticed while looking into this patch. - I skipped changing sun4i.c as I don't know how to handle the error there. Someone might want to have a look. (That's not ideal, but it's not worse than the same issue before this series.) In v1 Thierry had the concern: | That raises the question about what to do in these cases. If we return | an error, that could potentially throw off consumers. So perhaps the | closest would be to return a disabled PWM? Or perhaps it'd be up to the | consumer to provide some fallback configuration for invalidly configured | or unconfigured PWMs. .get_state() is only called in pwm_device_request on a pwm_state that a consumer might see. Before my series a consumer might have seen a partial modified pwm_state (because .get_state() might have modified .period, then stumbled and returned silently). The last patch ensures that this partial modification isn't given out to the consumer. Instead they now see the same as if .get_state wasn't implemented at all. Best regards Uwe Uwe Kleine-König (11): pwm: Make .get_state() callback return an error code pwm/tracing: Also record trace events for failed API calls drm/bridge: ti-sn65dsi86: Propagate errors in .get_state() to the caller leds: qcom-lpg: Propagate errors in .get_state() to the caller pwm: crc: Propagate errors in .get_state() to the caller pwm: cros-ec: Propagate errors in .get_state() to the caller pwm: imx27: Propagate errors in .get_state() to the caller pwm: mtk-disp: Propagate errors in .get_state() to the caller pwm: rockchip: Propagate errors in .get_state() to the caller pwm: sprd: Propagate errors in .get_state() to the caller pwm: Handle .get_state() failures drivers/gpio/gpio-mvebu.c | 9 ++++++--- drivers/gpu/drm/bridge/ti-sn65dsi86.c | 14 ++++++++------ drivers/leds/rgb/leds-qcom-lpg.c | 14 ++++++++------ drivers/pwm/core.c | 28 +++++++++++++++++---------- drivers/pwm/pwm-atmel.c | 6 ++++-- drivers/pwm/pwm-bcm-iproc.c | 8 +++++--- drivers/pwm/pwm-crc.c | 10 ++++++---- drivers/pwm/pwm-cros-ec.c | 8 +++++--- drivers/pwm/pwm-dwc.c | 6 ++++-- drivers/pwm/pwm-hibvt.c | 6 ++++-- drivers/pwm/pwm-imx-tpm.c | 8 +++++--- drivers/pwm/pwm-imx27.c | 8 +++++--- drivers/pwm/pwm-intel-lgm.c | 6 ++++-- drivers/pwm/pwm-iqs620a.c | 6 ++++-- drivers/pwm/pwm-keembay.c | 6 ++++-- drivers/pwm/pwm-lpss.c | 6 ++++-- drivers/pwm/pwm-meson.c | 8 +++++--- drivers/pwm/pwm-mtk-disp.c | 12 +++++++----- drivers/pwm/pwm-pca9685.c | 8 +++++--- drivers/pwm/pwm-raspberrypi-poe.c | 8 +++++--- drivers/pwm/pwm-rockchip.c | 12 +++++++----- drivers/pwm/pwm-sifive.c | 6 ++++-- drivers/pwm/pwm-sl28cpld.c | 8 +++++--- drivers/pwm/pwm-sprd.c | 8 +++++--- drivers/pwm/pwm-stm32-lp.c | 8 +++++--- drivers/pwm/pwm-sun4i.c | 12 +++++++----- drivers/pwm/pwm-sunplus.c | 6 ++++-- drivers/pwm/pwm-visconti.c | 6 ++++-- drivers/pwm/pwm-xilinx.c | 8 +++++--- include/linux/pwm.h | 4 ++-- include/trace/events/pwm.h | 20 +++++++++---------- 31 files changed, 174 insertions(+), 109 deletions(-) base-commit: 50315945d178eebec4e8e2c50c265767ddb926eb