On Tue, 24 Mar 2009, Alasdair G Kergon wrote: > On Mon, Feb 23, 2009 at 02:23:00PM -0500, Mikulas Patocka wrote: > > DM_ENDIO_REQUEUE doesn't work and I doubt that it can be made work with > > barriers. > > Didn't we discuss this before? > We *require* DM_ENDIO_REQUEUE functionality - I can't apply a patch that > breaks it. But it is not something that can be used arbitrarily - it is > only used in situations where core dm is firmly in control (and I/O cannot get > through to the device e.g. all paths down in multipath) and I don't see why it > can't continue to work with barriers. > > I hope it's just the comment that is wrong here - or does the patch need > updating? Oh, sorry, that is old comment for the old version of that patch. I fixed the code but forgot to remove the comment. The code in the patch should be correct w.r.t. DM_ENDIO_REQUEUE. Mikulas > Alasdair > -- > agk@xxxxxxxxxx > -- dm-devel mailing list dm-devel@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/dm-devel
