On 6/28/21 12:01 AM, Can Guo wrote: > On 2021-06-25 07:42, Bart Van Assche wrote: >> On 6/23/21 12:35 AM, Can Guo wrote: >>> Rename pm_op_in_progress and is_sys_suspended to >>> wlu_pm_op_in_progress and >>> is_wlu_sys_suspended accordingly. >> >> Can the is_wlu_sys_suspended member variable be removed by checking >> dev->power.is_suspended where dev represents the WLUN? >> > > No, PM set dev->power.is_suspended to "false" even the device failed > resuming, > while is_wlu_sys_suspended can be used to tell that. (+Rafael) Hi Rafael, In drivers/base/power/main.c we found the following code: End: error = dpm_run_callback(callback, dev, state, info); dev->power.is_suspended = false; Is it a bug or a feature that dev->power.is_suspended is set to false if dpm_run_callback() fails? I'm asking this because only clearing dev->power.is_suspended if dpm_run_callback() returns 0 would allow to simplify the UFS driver. It can happen for UFS devices that runtime resume fails and if this fails we need to track this. Thanks, Bart.
