On 22/11/2015 17:10, Christoph Hellwig wrote:
On Sun, Nov 22, 2015 at 04:55:49PM +0200, Sagi Grimberg wrote:
Also note that 4.4-rc prefer_fr=y register_always=n
!register_always still blows up badly with XFS and ext4 due to
data integrity errors.
So the register_always=N makes bad things happen? if we register
all the frags we're OK (i.e. register_always=Y)?
No. register_always=Y is already broken in 4.3, but register_always=N is
now also broken in 4.4.
OK, I'm confused so please let me understand slowly :)
Your patch "ib_srp: initialize dma_length in srp_map_idb" solves
the register_always=Y dma_length = 0 WARN_ON() on 4.4-rc, does it solve
the data integrity errors too?
The 4.3+ issue that only affects register_always=Y is the FR mapping
of the indirect buffer that is not accounted for.
So it fails because there are no free FRs to use. I see.
The 4.4+ issue that also affects register_always=N is something in your
MR patches when we hit a gappy SG list.
This is something I'd need to look at. I tested this code path using
Bart's discontigiuous_io utility and didn't see problems.
This code path is specific to srp because all other ULPs guarantee
no-gaps...
--
To unsubscribe from this list: send the line "unsubscribe target-devel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
