On Wed, 2014-07-16 at 14:45 -0400, Martin K. Petersen wrote:
> >>>>> "James" == James Bottomley <jbottomley@xxxxxxxxxxxxx> writes:
>
> >> I don't have a problem with a BLIST_PREFER_UNMAP flag or something
> >> like that. But BLIST_TRY_VPD_PAGES seems more generally useful and it
> >> does fix the problem at hand. That's why I went that route.
>
> James> Hang on ... unless we apply Christoph or my fix, we'll get the
> James> same issue with every raid driver (that's about 10 I think) that
> James> set no_write_same when they hit a >2TB RAID volume,
>
> no_write_same is set for all the RAID controller drivers (54b2b50c20a6).
>
> If a WRITE SAME(10/16) command fails and the UNMAP bit is not set we'll
> set no_write_same=1 and disable REQ_WRITE_SAME support.
>
> If a WRITE SAME(10/16) command fails and the UNMAP bit is set we'll
> disable REQ_DISCARD support.
>
> Not sure where the 10 vs. 16 byte 2TB limitation comes into play here?
It's the code we identified in sd.c:read_capacity_16():
if (buffer[14] & 0x80) { /* LBPME */
sdkp->lbpme = 1;
if (buffer[14] & 0x40) /* LBPRZ */
sdkp->lbprz = 1;
sd_config_discard(sdkp, SD_LBP_WS16);
}
If the inquiry shows LBPME set, we'll do write same regardless of the
no_write_same bit. That's why for SPC-2 devices it only shows up on
>2TB devices because they force RC16.
James
--
To unsubscribe from this list: send the line "unsubscribe stable" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html