On Fri, Aug 26, 2016 at 12:12:07PM -0400, Doug Ledford wrote: > On 8/26/2016 10:44 AM, Leon Romanovsky wrote: > > On Thu, Aug 25, 2016 at 03:29:12PM -0400, Doug Ledford wrote: > > > > Hi Doug, > > > > These two patches were supposed to be carried by us [1] and I explicitly > > said that. Especially, the last patch in this series is wrong. Please > > revert it. The proper patch is [2] and it was supposed to be sent right > > after our shared code. > > > >> Yuval Shaia (2): > >> IB/mlx4: Make function use_tunnel_data return void > >> IB/mlx4: Return EAGAIN for any error in mlx4_ib_poll_one > > > > [1] https://www.spinics.net/lists/linux-rdma/msg38580.html > > [2] > > https://git.kernel.org/cgit/linux/kernel/git/leon/linux-rdma.git/commit/?h=topic/mlx-fixes&id=d5a1c56c3a49db6369b36aa04fbeabc7653ae107 > > > > I saw your request to carry it, but then never heard anything again from > you. The fix that the original patch implements is still valid and was > something I wanted to get into 4.8, where as your patch is being held > for 4.9 and goes down a different path to solving the issue. IMO, the > patch from Yuval is fine for 4.8, and if you want to apply yours to 4.9, > that's fine too. I don't see a need to revert the existing patch. From > what I can tell, the existing patch will work fine and do what Yuval > intended, it just won't do what you intend to do in 4.9. Please correct > me if I'm wrong. By our HW design and SW implementation poll_cq never fails and returns errors, so all these prints are to catch ULP bugs. In case of such bug, Yuval's patch will cause to reentry (EAGAIN) and kprints storm again and again. It is undesired and misleading behaviour. We targeted our patch to 4.9, because it is not actual fix, but help to ULP developers and there is no real need to hurry up. Will it be acceptable by you, if I revert Yuval's patch before sending our version for 4.9? Thanks. > > -- > Doug Ledford > GPG Key ID: 0E572FDD >