Re: [PATCH 03/23] sd: implement REQ_OP_WRITE_ZEROES

[Bart Van Assche <Bart.VanAssche@sandisk.com>]

On Wed, 2017-03-29 at 16:51 +0200, Paolo Bonzini wrote:
> On 28/03/2017 20:50, Bart Van Assche wrote:
> > This means that just like the start and end of a discard must be aligne=
d on a
> > discard_granularity boundary, WRITE SAME commands with the UNMAP bit se=
t must
> > also respect that granularity. I think this means that either
> > __blkdev_issue_zeroout() has to be modified such that it rejects unalig=
ned
> > REQ_OP_WRITE_ZEROES operations or that blk_bio_write_same_split() has t=
o be
> > modified such that it generates REQ_OP_WRITEs for the unaligned start a=
nd tail.
>=20
> I don't think this is the case.

Hello Paolo,

Can you cite the section(s) from the SCSI specs that support your view? I
reread the "5.49 WRITE SAME (10) command" and "4.7.3.4.4 WRITE SAME command
and unmap operations" sections but I have not found any explicit statement
that specifies the behavior for unaligned WRITE SAME commands with the UNMA=
P
bit set. It seems to me like=A0the OPTIMAL UNMAP GRANULARITY parameter was
overlooked when both sections were written. Should we ask the T10 committee
for a clarification?

Another question is, if the specification of WRITE SAME + UNMAP would be
made unambiguous in the SBC document, whether or not we should take the ris=
k
to trigger behavior that is not what we expect by sending unaligned WRITE
SAME + UNMAP commands to SCSI devices?

Thanks,

Bart.=