RE: [PATCH v4] scsi: ufs: core: handle PM commands timeout before SCSI EH

[Fang Hongjie(方洪杰)]

> From: Peter Wang (王信友) [mailto:peter.wang@xxxxxxxxxxxx]
> Sent: Wednesday, June 3, 2026 2:55 PM
> To: beanhuo@xxxxxxxxxx; James.Bottomley@xxxxxxxxxxxxxxxxxxxxx;
> Fang Hongjie(方洪杰) <hongjiefang@xxxxxxxxxxxx>;
> alim.akhtar@xxxxxxxxxxx; avri.altman@xxxxxxx;
> martin.petersen@xxxxxxxxxx; bvanassche@xxxxxxx
> Cc: linux-scsi@xxxxxxxxxxxxxxx; linux-kernel@xxxxxxxxxxxxxxx
> Subject: Re: [PATCH v4] scsi: ufs: core: handle PM commands timeout before
> SCSI EH
> 
> 
> On Tue, 2026-06-02 at 20:41 +0800, Hongjie Fang wrote:
> > +       /*
> > +        * ufshcd_link_recovery() may already have completed @scmd,
> > e.g. via
> > +        * the existing MCQ force-completion path.
> > +        */
> > +       if (!test_bit(SCMD_STATE_COMPLETE, &scmd->state)) {
> > +               if (!hba->mcq_enabled) {
> >
> 
> Hi Hongjie,
> 
> MCQ will complete all uncompleted scmd by setting force_compl.
> Hence, if there are still uncompleted scmd, it should be
> in legacy mode, and there is no need to check
> if (!hba->mcq_enabled), right?
> 

The !hba->mcq_enabled check is only for clearing outstanding_reqs,
which is the legacy single-doorbell software bitmap.
MCQ does not use outstanding_reqs for request tracking.


> 
> > +                       unsigned long flags;
> > +                       struct request *rq = scsi_cmd_to_rq(scmd);
> > +
> > +                       spin_lock_irqsave(&hba->outstanding_lock,
> > flags);
> > +                       __clear_bit(rq->tag, &hba->outstanding_reqs);
> > +                       spin_unlock_irqrestore(&hba-
> > >outstanding_lock, flags);
> > +               }
> > +
> > +               set_host_byte(scmd, DID_TIME_OUT);
> >
> 
> Why does MCQ mode set DID_REQUEUE, while legacy mode sets
> DID_TIME_OUT?
> 
Normal SCSI PM commands use DID_REQUEUE, matching the existing
MCQ force-completion behavior. Reserved internal device-management
commands continue to use DID_TIME_OUT, right ?



Best.