Messages in this thread | | | Subject | Re: tipc_udp_send_msg oops in 4.4 when setting link tolerance | From | Tommi Rantala <> | Date | Wed, 15 Nov 2017 11:42:15 +0200 |
| |
On 14.11.2017 13:35, Jon Maloy wrote: >> Found it, the missing patch is this one (9b3009604b8e does not help): >> >> commit d01332f1acacc0cb43a61f4244dd2b846d4cd585 >> Author: Richard Alpe <richard.alpe@ericsson.com> >> Date: Mon Feb 1 08:19:56 2016 +0100 >> >> tipc: fix link attribute propagation bug >> >> >> It does not apply as-is to 4.4, so backported it, see below. >> Does it look good? I can send it forward to Greg for inclusion in 4.4. > > Yes. I would be very grateful if you do.
OK, sent the patch to stable@vger.kernel.org
>> But with this patch included, I can easily reproduce the "BUG: Bad page state >> in process git" issue also in 4.4 like this: >> >> $ tipc link set tolerance 100 link $LINKNAME $ cd /tmp && git clone >> /path/to/linux-stable >> >> I can try to debug that a bit more to see if I can figure it out. > > I would appreciate that. If this turns out to also be an already fixed bug (most likely) you can send that directly to GKH too. > If you find a bug that requires a new patch send it to tipc-discussion for review first. > > Finally, I want to say that I very much appreciate users like you, who try solving the problems themselves and contribute to the code. > We are right now in an upgrade phase of TIPC, where we have been/are adding new features to TIPC (overlapping ring neighbor monitoring, group communication, new addressing model, performance improvements...) which inevitably entail some new bugs, but I expect the frequency of these changes/improvements to TIPC go down in a few months, and with that the number of fixes needed.
Thanks!
After more testing, I noticed that I could only reproduce the "BUG: Bad page state" in a single machine, so maybe something wrong with the machine, and not with tipc.
-Tommi
| |