On 12/20/23 12:14, Tony Lindgren wrote: > * Thomas Richard <thomas.richard@xxxxxxxxxxx> [231220 10:50]: >> On 12/19/23 18:15, Thomas Richard wrote: >>> Hello, >> >> I add some people in this thread. >> >>> >>> I have a gpio expander (pca953x driver) connected to an i2c controller >>> managed by the omap-i2c driver. >>> And I have some issues with pm_runtime_force_suspend/resume during >>> suspend to ram. >>> For some reasons, related to hardware design, I need to access to this >>> gpio expander during suspend_noirq and resume_noirq. So I had to move >>> the suspend/resume of the pca953x to suspend_noirq/resume_noirq. > > Hmm at noirq level you need to do polling on the i2c controller? Hello Tony, Thanks for your reply. No, irq is still active in suspend_noirq for this i2c controller due to the flag IRQF_NO_SUSPEND [1]. If this flag is set, the interrupt is still enabled in suspend_noirq [2]. [1] https://elixir.bootlin.com/linux/v6.7-rc6/source/drivers/i2c/busses/i2c-omap.c#L1473 [2] https://www.kernel.org/doc/html/latest/power/suspend-and-interrupts.html#the-irqf-no-suspend-flag > >>> diff --git a/drivers/i2c/busses/i2c-omap.c b/drivers/i2c/busses/i2c-omap.c >>> index 42165ef57946..fe79b27b46fd 100644 >>> --- a/drivers/i2c/busses/i2c-omap.c >>> +++ b/drivers/i2c/busses/i2c-omap.c >>> @@ -1575,9 +1575,24 @@ static int __maybe_unused >>> omap_i2c_runtime_resume(struct device *dev) >>> return 0; >>> } >>> >>> +static int omap_i2c_suspend(struct device *dev) >>> +{ >>> + pm_runtime_get_sync(dev); >>> + pm_runtime_disable(dev); >>> + return 0; >>> +} > > If you want the i2c controller enabled during suspend, you can leave it > enabled above, and as we already have SET_NOIRQ_SYSTEM_SLEEP_PM_OPS > doing force_suspend() and force_resume(), you can runtime PM put on > resume. So something like below might do the trick: Ok I'll test it. Thanks > > static int omap_i2c_suspend(struct device *dev) > { > return pm_runtime_resume_and_get(dev); > } > > static int omap_i2c_resume(struct device *dev) > { > pm_runtime_mark_last_busy(dev); > pm_runtime_put_autosuspend(dev); > > return 0; > } > >>> static const struct dev_pm_ops omap_i2c_pm_ops = { >>> SET_NOIRQ_SYSTEM_SLEEP_PM_OPS(pm_runtime_force_suspend, >>> pm_runtime_force_resume) >>> + SET_SYSTEM_SLEEP_PM_OPS(omap_i2c_suspend, omap_i2c_resume) >>> SET_RUNTIME_PM_OPS(omap_i2c_runtime_suspend, >>> omap_i2c_runtime_resume, NULL) >>> }; > > And with the changes you did to omap_i2c_pm_ops naturally. This way > the controller should stay active until noirq ops. > > Of course it's possible I did not quite understand what you're trying > to do, but if so please let me know :) The context is: One gpio of this expander is the reset pin of PCIe endpoints. And this reset pin shall be managed during the suspend_noirq and resume_noirq of the PCIe controller. That's why I need to do some i2c accesses during suspend_noirq/resume_noirq. -- Thomas Richard, Bootlin Embedded Linux and Kernel engineering https://bootlin.com
