Re: [PATCH 2/3] scsi: disable VPD page check on error

2016-06-22 Thread Christoph Hellwig
On Mon, Jun 20, 2016 at 08:57:47AM +0200, Hannes Reinecke wrote: > If we encounter an error during initial VPD page scan we should be > setting the 'skip_vpd_pages' bit to avoid further accesses. > Any errors during rescan should be ignored, as we might encounter > temporary I/O errors during

Re: [PATCH 2/3] scsi: disable VPD page check on error

2016-06-21 Thread Martin K. Petersen
> "James" == James Bottomley writes: James> I also don't really like this. If the device has a failure like James> the QEMU one where it just hangs up, this won't help because the James> problem already happened. Conversely, if the device replies in

Re: [PATCH 2/3] scsi: disable VPD page check on error

2016-06-20 Thread Ewan D. Milne
On Mon, 2016-06-20 at 06:44 -0700, James Bottomley wrote: > On Mon, 2016-06-20 at 09:25 -0400, Ewan D. Milne wrote: > > On Mon, 2016-06-20 at 08:57 +0200, Hannes Reinecke wrote: > > > If we encounter an error during initial VPD page scan we should be > > > setting the 'skip_vpd_pages' bit to avoid

Re: [PATCH 2/3] scsi: disable VPD page check on error

2016-06-20 Thread James Bottomley
On Mon, 2016-06-20 at 09:25 -0400, Ewan D. Milne wrote: > On Mon, 2016-06-20 at 08:57 +0200, Hannes Reinecke wrote: > > If we encounter an error during initial VPD page scan we should be > > setting the 'skip_vpd_pages' bit to avoid further accesses. > > Any errors during rescan should be ignored,

[PATCH 2/3] scsi: disable VPD page check on error

2016-06-20 Thread Hannes Reinecke
If we encounter an error during initial VPD page scan we should be setting the 'skip_vpd_pages' bit to avoid further accesses. Any errors during rescan should be ignored, as we might encounter temporary I/O errors during rescan. Signed-off-by: Hannes Reinecke ---