Commit 38ba835
pwm: sysfs: Do not apply state to already disabled PWMs
If the PWM is exported but not enabled, do not call pwm_class_apply_state().
First of all, in this case, period may still be unconfigured and this would
make pwm_class_apply_state() return -EINVAL, and then suspend would fail.
Second, it makes little sense to apply state onto PWM that is not enabled
before suspend.
Failing case:
"
$ echo 1 > /sys/class/pwm/pwmchip4/export
$ echo mem > /sys/power/state
...
pwm pwmchip4: PM: dpm_run_callback(): pwm_class_suspend+0x1/0xa8 returns -22
pwm pwmchip4: PM: failed to suspend: error -22
PM: Some devices failed to suspend, or early wake event detected
"
Working case:
"
$ echo 1 > /sys/class/pwm/pwmchip4/export
$ echo 100 > /sys/class/pwm/pwmchip4/pwm1/period
$ echo 10 > /sys/class/pwm/pwmchip4/pwm1/duty_cycle
$ echo mem > /sys/power/state
...
"
Do not call pwm_class_apply_state() in case the PWM is disabled
to fix this issue.
Fixes: 7fd4edc ("pwm: sysfs: Add suspend/resume support")
Signed-off-by: Marek Vasut <marex@denx.de>
Fixes: ef2bf49 ("pwm: Improve args checking in pwm_apply_state()")
Reviewed-by: Brian Norris <briannorris@chromium.org>
Reviewed-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Thierry Reding <thierry.reding@gmail.com>1 parent 661dfb7 commit 38ba835
1 file changed
Lines changed: 17 additions & 0 deletions
| Original file line number | Diff line number | Diff line change | |
|---|---|---|---|
| |||
424 | 424 | | |
425 | 425 | | |
426 | 426 | | |
| 427 | + | |
| 428 | + | |
| 429 | + | |
| 430 | + | |
| 431 | + | |
| 432 | + | |
| 433 | + | |
427 | 434 | | |
428 | 435 | | |
429 | 436 | | |
| |||
448 | 455 | | |
449 | 456 | | |
450 | 457 | | |
| 458 | + | |
| 459 | + | |
| 460 | + | |
| 461 | + | |
451 | 462 | | |
| 463 | + | |
| 464 | + | |
| 465 | + | |
| 466 | + | |
| 467 | + | |
| 468 | + | |
452 | 469 | | |
453 | 470 | | |
454 | 471 | | |
| |||
0 commit comments