Re: [PATCH v4] scsi: ufs: Cleanup completed request without interrupt notification

[Bart Van Assche]

On 2020-07-31 01:00, Can Guo wrote:
> AFAIK, sychronization of scsi_done is not a problem here, because scsi
> layer
> use the atomic state, namely SCMD_STATE_COMPLETE, of a scsi cmd to prevent
> the concurrency of abort and real completion of it.
> 
> Check func scsi_times_out(), hope it helps.
> 
> enum blk_eh_timer_return scsi_times_out(struct request *req)
> {
> ...
> ÂÂÂÂÂÂÂ if (rtn == BLK_EH_DONE) {
> ÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ /*
> ÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ * Set the command to complete first in order to prevent
> a real
> ÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ * completion from releasing the command while error
> handling
> ÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ * is using it. If the command was already completed,
> then the
> ÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ * lower level driver beat the timeout handler, and it
> is safe
> ÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ * to return without escalating error recovery.
> ÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ *
> ÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ * If timeout handling lost the race to a real
> completion, the
> ÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ * block layer may ignore that due to a fake timeout
> injection,
> ÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ * so return RESET_TIMER to allow error handling another
> shot
> ÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ * at this command.
> ÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ */
> ÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ if (test_and_set_bit(SCMD_STATE_COMPLETE, &scmd->state))
> ÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ return BLK_EH_RESET_TIMER;
> ÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ if (scsi_abort_command(scmd) != SUCCESS) {
> ÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ set_host_byte(scmd, DID_TIME_OUT);
> ÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ scsi_eh_scmd_add(scmd);
> ÂÂÂÂÂÂÂÂÂÂÂÂÂÂÂ }
> ÂÂÂÂÂÂÂ }
> }

I am familiar with this mechanism. My concern is that both the regular
completion path and the abort handler must call scsi_dma_unmap() before
calling cmd->scsi_done(cmd). I don't see how
test_and_set_bit(SCMD_STATE_COMPLETE, &scmd->state) could prevent that
the regular completion path and the abort handler call scsi_dma_unmap()
concurrently since both calls happen before the SCMD_STATE_COMPLETE bit
is set?

Thanks,

Bart.