On Mon, 2023-06-12 at 08:35 -0700, Bart Van Assche wrote:
> On 6/12/23 08:03, mwilck@xxxxxxxx wrote:
> > -static int scsi_internal_device_block(struct scsi_device *sdev)
> > +static void scsi_device_block(struct scsi_device *sdev, void
> > *data)
> > {
> > int err;
> >
> > @@ -2805,7 +2804,8 @@ static int scsi_internal_device_block(struct
> > scsi_device *sdev)
> > scsi_stop_queue(sdev, false);
> > mutex_unlock(&sdev->state_mutex);
> >
> > - return err;
> > + WARN_ONCE(err, "__scsi_internal_device_block_nowait(%s)
> > failed: err = %d\n",
> > + dev_name(&sdev->sdev_gendev), err);
> > }
>
> Hmm ... shouldn't the WARN_ONCE() statement refer to
> scsi_device_block()
> instead of __scsi_internal_device_block_nowait()?
The message references the function that failed, and prints the return
value of that function. It used to be scsi_internal_device_block()
before my patch, and it's __scsi_internal_device_block_nowait() now.
But I agree that the message is ugly. I'll to come up with something
better.
Thanks,
Martin
