lkml.org 
[lkml]   [2005]   [May]   [13]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
From
SubjectRe: x86-64 bad pmds in 2.6.11.6 II
Date
CW = Christopher Warner
CW>>>>> 2.6.11.5 kernel,
CW>>>>> Tyan S2882/dual AMD 246 opterons
CW>>>>> sh:18983: mm/memory.c:99: bad pmd
ffff810005974cc8(00007ffffffffe46).
CW>>>>> sh:18983: mm/memory.c:99: bad pmd
ffff810005974cd0(00007ffffffffe47).

DJ = Dave Jones
DJ>>>> That's the 3rd or 4th time I've seen this
DJ>>>> reported on this hardware.
DJ>>>> It's not exclusive to it, but it does seem more
DJ>>>> susceptible for some reason. Spooky.

AK = Andi Kleen wrote:
AK>>> It seems to be clear now that it is hardware
AK>>> independent.
AK>>>
AK>>> I actually got it once now too, but only after
AK>>> 24+h stress test :/
AK>>>
AK>>> I have a better debugging patch now that I will be
AK>>> testing soon, hopefully that turns something up.

DJ = Dave Jones
DJ>> Ok, I'm respinning the Fedora update kernel today
DJ>> for other reasons, if you have that patch in time,
DJ>> I'll toss it in too.
DJ>>
DJ>> Though as yet, no further reports from our users.

AK = Andi Kleen
AK> Here's the new patch. However it costs some memory
AK> bloat because I added a new field to struct page

I posted some information on the fedora-list concerning my experience
with this problem. I am using a Tyan S2885/dual 244 Opterons. For HW and
driver details see:
https://www.redhat.com/archives/fedora-list/2005-May/msg01690.html

I have been using Dave's FC3 test kernel (2.6.11-1.24_FC3smp) for a
little over a day and have been unable to generate the problem with the
computer under a larger than normal load.

Prior to May 12, I had been seeing the problem very regularly. It
started around April 14. I believe this is about the time I first
started using the 2.6.11-1.14_FC3smp kernel. I remember I had to get a
BIOS upgrade from Tyan (http://www.tyan.com/support/html/b_s2885.html
unfortunately a Beta release) to get my network back once I started
using the new kernel. After that the memory.c messages started showing
up. I though it might be BIOS related (see the note on the referenced
Tyan page), but when I saw Christopher's post I thought maybe it was the
kernel because his MOBO doesn't have a Beta BIOS release. I googled and
found this thread, and subscribed to this list.

I have "memory.c:97 bad pmd" entries in my /var/log/messages files going
back to April 14. The only days I don't have them are April 20, 22, 23,
24 and April 29 (22, 23, and 24 are a weekend with less activity). I
have had them every day in May until I installed Dave's test kernel.

I am very computer literate, but not a kernel developer. I hope I didn't
offend you guys by posting here. I would be willing to be your guinea
pig for testing. Currently I am unable to reproduce the problem. If I am
able to reproduce the problem, would you prefer I post here or on the
fedora list?

Thanks for all of your efforts,
Pete


-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

\
 
 \ /
  Last update: 2005-05-13 23:57    [W:0.092 / U:2.140 seconds]
©2003-2017 Jasper Spaans. hosted at Digital OceanAdvertise on this site