Re: SG does not ignore dxferp (direct io + mmap)

[Laurence Oberman <loberman@redhat.com>]

----- Original Message -----
> From: "Eyal Ben David" <bdeyal@gmail.com>
> To: "Johannes Thumshirn" <jthumshirn@suse.de>
> Cc: "Ewan D. Milne" <emilne@redhat.com>, "Laurence Oberman" <loberman@redhat.com>, dgilbert@interlog.com,
> linux-scsi@vger.kernel.org
> Sent: Friday, November 25, 2016 7:36:34 AM
> Subject: Re: SG does not ignore dxferp (direct io + mmap)
> 
> On Fri, Nov 25, 2016 at 1:53 PM, Johannes Thumshirn <jthumshirn@suse.de>
> wrote:
> > On Fri, Nov 25, 2016 at 01:20:34PM +0200, Eyal Ben David wrote:
> >> Note that sg_mmap_read does not parse the SCSI sense, so the script
> >> might fail for other reasons (some SCSI error) and think its a zero
> >> byte corruption.
> >
> > But SCSI generic checks for errors and returns -EINVAL on CHECK_CONDITION
> > or
> > DRIVER_SENSE (and sets SG_INFO_CHECK in hdr.info).
> >
> Ah OK. We use async write/read instead of ioctl and forgot that ioctl
> checks the read.
> 
> > Anyways, can you test the patch Ewan found on one of your kernel's that are
> > known to fail?
> 
> All the examples I gave before were on physical hosts and storage at a
> testing lab.
> That would be difficult to do. Sorry.
> --
> To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> 

I applied just the patch myself and Ewan isolated to the 4.7.9 kernel and the issue is resolved for me.
I have 100000 loops of the test with no issue.

I repeated the test flushing cache each time as well to make sure

for i in `seq 1 100000`; do ./sg_mmap_read -d /dev/sg1 -l 0 -m -b | hexdump | grep 6300; done

for i in `seq 1 100000`; do ./sg_mmap_read -d /dev/sg1 -l 0 -m -b | hexdump | grep 6300; echo 3 > /proc/sys/vm/drop_caches;done

Johannes, you are reproducing another race in your test I think.