On Fri, 2009-10-16 at 11:29 +0200, Christof Schmitt wrote:
> On Thu, Oct 15, 2009 at 05:47:00PM -0700, Vasu Dev wrote:
> > diff --git a/drivers/scsi/scsi_error.c b/drivers/scsi/scsi_error.c
> > index 7b1e20f..3379da6 100644
> > --- a/drivers/scsi/scsi_error.c
> > +++ b/drivers/scsi/scsi_error.c
> > @@ -331,6 +331,42 @@ static int scsi_check_sense(struct scsi_cmnd *scmd)
> > }
> > }
> >
> > +static void scsi_handle_queue_ramp_up(struct scsi_device *sdev)
> > +{
> > + struct scsi_host_template *sht = sdev->host->hostt;
> > + struct scsi_device *tmp_sdev;
> > +
> > + if (!sht->change_queue_depth ||
> > + sdev->queue_depth == sdev->max_queue_depth)
> > + return;
> [...]
> > --- a/drivers/scsi/scsi_scan.c
> > +++ b/drivers/scsi/scsi_scan.c
> > @@ -251,6 +251,7 @@ static struct scsi_device *scsi_alloc_sdev(struct scsi_target *starget,
> > sdev->model = scsi_null_device_strs;
> > sdev->rev = scsi_null_device_strs;
> > sdev->host = shost;
> > + sdev->queue_ramp_up_period = SCSI_DEFAULT_RAMP_UP_PERIOD;
> > sdev->id = starget->id;
> > sdev->lun = lun;
> > sdev->channel = starget->channel;
> > @@ -312,6 +313,8 @@ static struct scsi_device *scsi_alloc_sdev(struct scsi_target *starget,
> > }
> > }
> >
> > + sdev->max_queue_depth = sdev->queue_depth;
> > +
> > return sdev;
>
> Running this patches series with the zfcp device driver increases the
> queue_depth beyond the maximum. The problem is that after slave_alloc,
> the queue_depth is 1 from cmd_per_lun and this is the value used for
> max_queue_depth.
>
I see, I missed this case to update max_queue_depth since libfc is not
using slave_configure.
> zfcp then adjust the queue_depth and tagging in slave_configure. Now,
> the queue_depth is 32 and the max_queue_depth is still 1. And the
> check for sdev->queue_depth == sdev->max_queue_depth is never true.
>
> What is the best way to solve this? Move the initial assignment of
> max_queue_depth to be called after the call to slave_configure? And/or
Move should work fine unless I'm missing any case where slave_configure
won't be called on new sdev alloc or max_queue_depth could change after
calling slave_configure also, apart from sysfs.
I've already added code to update this on qdepth change via sysfs. Am I
missing any other case ?
> adjust the check above to sdev->queue_depth >= sdev->max_queue_depth?
>
Adding check this way will work as safe guard in case any place left
updating max_queue_depth but that would be bug. I'll change check as
suggested above.
Thanks
Vasu
> --
> Christof
> --
> To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
> the body of a message to majordomo@xxxxxxxxxxxxxxx
> More majordomo info at http://vger.kernel.org/majordomo-info.html
--
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
