On Thu, Mar 15, 2018 at 02:12:29PM -0400, Laurence Oberman wrote:
> On Thu, 2018-03-15 at 12:34 -0400, Doug Ledford wrote:
> > On Thu, 2018-03-15 at 12:32 -0400, Doug Ledford wrote:
> > > On Wed, 2018-03-14 at 17:06 -0400, Laurence Oberman wrote:
> > > > Is this the fix for the issue I worked with Jason which we had to
> > > > pull
> > > > for RHEL 7.5
> > >
> > > No, they are unrelated I believe.
> > >
> >
> > I take that back. It could be. If this:
> >
> > 16c1975f1032 ("IB/mlx5: Create profile infrastructure to add and
> > remove
> > stages")
> >
> > is the upstream commit that got yanked, then yes.
> >
>
> Yep from my original patch we did not use
> My comment was:
>
> Changes to the mlx5_core call mlx5_try_fast_unload in the shutdown.
> This is causing error messages on shutdown and with the latest
> rdma tree panics due to list corruption.
> Remove the mlx5_try_fast_unload call so we go back to calling
> the original mlx5_unload_one call only.
Laurence, Doug,
This patch is unrelated to the issue you worked with Jason.
Thanks
Attachment:
signature.asc
Description: PGP signature
