On Wed, Jan 22, 2020 at 08:23:41PM -0800, Bart Van Assche wrote: > Since the SCSI core does not reuse the tag of the SCSI command that is > being aborted by .eh_abort() before .eh_abort() has finished it is not > necessary to check from inside that callback whether or not the SCSI command > has already completed. Instead, rely on the firmware to return an error code > when attempting to abort a command that has already completed. Additionally, > rely on the firmware to return an error code when attempting to abort an > already aborted command. > > In qla2x00_abort_srb(), use blk_mq_request_started() instead of > sp->completed and sp->aborted. > > This patch eliminates several race conditions triggered by the removed member > variables. I can only guess here what the races are but I agree removing the logic here and relying on the SCSI layer to handle it correctly makes sense. > Acked-by: Himanshu Madhani <hmadhani@xxxxxxxxxxx> > Reviewed-by: Roman Bolshakov <r.bolshakov@xxxxxxxxx> > Cc: Quinn Tran <qutran@xxxxxxxxxxx> > Cc: Martin Wilck <mwilck@xxxxxxxx> > Cc: Daniel Wagner <dwagner@xxxxxxx> > Signed-off-by: Bart Van Assche <bvanassche@xxxxxxx> Reviewed-by: Daniel Wagner <dwagner@xxxxxxx> > +/* > + * The caller must ensure that no completion interrupts will happen > + * while this function is in progress. > + */ So could we add something like WARN_ON(irqs_disabled())?
