#505977 v86d: segfault

Package:
v86d
Source:
v86d
Description:
daemon to run x86 code in an emulated environment
Submitter:
"hugo vanwoerkom"
Date:
2010-06-10 15:21:04 UTC
Severity:
normal
#505977#5
Date:
2008-11-17 11:34:50 UTC
From:
To:
At times v86d segfaults, like so:
Nov 17 05:00:27 debian v86d: Trying to access an unsupported memory region
at 10ff99
Nov 17 05:00:37 debian last message repeated 2 times
Nov 17 05:00:41 debian kernel: [24114.544392] v86d[26680]: segfault at 0 ip
08049861 sp bf9faf60 error 4 in v86d
[8048000+16000]
Nov 17 05:00:41 debian kernel: [24117.992020] uvesafb: mode switch failed
(eax=0x4f02, err=1)
Nov 17 05:00:41 debian kernel: [24119.541427] v86d[27267]: segfault at 0 ip
08049861 sp bf8c0e20 error 4 in v86d
[8048000+16000]
Nov 17 05:00:41 debian kernel: [24122.992020] uvesafb: mode switch failed
(eax=0x4f02, err=1)
Nov 17 05:00:41 debian kernel: [24124.457259] v86d[27290]: segfault at 0 ip
08049861 sp bfde9350 error 4 in v86d
[8048000+16000]
Nov 17 05:00:41 debian kernel: [24128.000027] uvesafb: mode switch failed
(eax=0x4f02, err=1)
Nov 17 05:01:32 debian kernel: [24179.348262] 'OUT-unknown:'IN= OUT=vmnet1
SRC=172.16.178.1 DST=172.16.178.255 L
EN=242 TOS=0x00 PREC=0x00 TTL=64 ID=0 DF PROTO=UDP SPT=138 DPT=138 LEN=222
Nov 17 05:01:32 debian kernel: [24179.348397] 'OUT-unknown:'IN= OUT=vmnet8
SRC=172.16.113.1 DST=172.16.113.255 L
EN=242 TOS=0x00 PREC=0x00 TTL=64 ID=0 DF PROTO=UDP SPT=138 DPT=138 LEN=222
Nov 17 05:02:19 debian v86d: Trying to access an unsupported memory region
at 10ff99
Nov 17 05:02:22 debian kernel: [24226.147311] v86d[27899]: segfault at 0 ip
08049861 sp bf93fea0 error 4 in v86d
[8048000+16000]
Nov 17 05:02:22 debian kernel: [24229.616026] uvesafb: mode switch failed
(eax=0x4f02, err=1)
Nov 17 05:02:34 debian v86d: Trying to access an unsupported memory region
at 10ff99
Nov 17 05:02:37 debian kernel: [24241.006200] v86d[27984]: segfault at 0 ip
08049861 sp bfcda240 error 4 in v86d
[8048000+16000]
Nov 17 05:02:37 debian kernel: [24244.428055] uvesafb: mode switch failed
(eax=0x4f02, err=1)

I run a recompiled v86d w/o the patch and configured to:
        ./configure --with-x86emu

Is there a way to restart uvesafb? Because now to only way around it is a
reboot.

It doesn't always happen, but I can't tell why not.

I run the latest linux-image-2.6.26-1-686 on Sid.

Hugo Vanwoerkom

#505977#10
Date:
2008-11-17 12:26:54 UTC
From:
To:
The startup messages for uvesafb were:

Nov 16 22:19:05 debian kernel: [   23.476272] uvesafb: NVIDIA Corporation,
nv44 Board - p382h1  , Chip Rev   , O
EM: NVIDIA, VBE v3.0
Nov 16 22:19:05 debian kernel: [   23.508283] uvesafb: protected mode
interface info at c000:d3a0
Nov 16 22:19:05 debian kernel: [   23.508288] uvesafb: pmi: set display
start = c00cd3d6, set palette = c00cd440
Nov 16 22:19:05 debian kernel: [   23.508291] uvesafb: pmi: ports = 3b4 3b5
3ba 3c0 3c1 3c4 3c5 3c6 3c7 3c8 3c9
3cc 3ce 3cf 3d0 3d1 3d2 3d3 3d4 3d5 3da
Nov 16 22:19:05 debian kernel: [   23.511052] uvesafb: VBIOS/hardware
doesn't support DDC transfers
Nov 16 22:19:05 debian kernel: [   23.511056] uvesafb: no monitor limits
have been set, default refresh rate wil
l be used
Nov 16 22:19:05 debian kernel: [   23.516863] uvesafb: scrolling: ypan using
protected mode interface, yres_virt
ual=6144
Nov 16 22:19:05 debian kernel: [   23.743379] Console: switching to colour
frame buffer device 128x48
Nov 16 22:19:05 debian kernel: [   23.762168] uvesafb: framebuffer at
0xc0000000, mapped to 0xf8b80000, using 24
576k, total 262144k
Nov 16 22:19:05 debian kernel: [   23.762171] fb0: VESA VGA frame buffer
device

The only difference I see is that "no monitor limits" because I had the
monitors turned off.
I run a 2-seat system and normally gdm brings up the 2 monitors, but not
this time: it brought up only one, so I had to restart gdm.
Unfortunately I did not save the Xorg logs of that startup.


Hugo

#505977#15
Date:
2008-11-18 16:11:04 UTC
From:
To:
Hi Hugo,

Never seen that on my systems.
Can you try 0.1.9 from
http://die-welt.net/~evgeni/v86d/v86d_0.1.9-0_i386.deb ?
It's unpatched and compiled with x86emu, so you shouldn't need any
changes.

Regards
Evgeni

#505977#20
Date:
2008-11-18 16:19:26 UTC
From:
To:

Hi Evgeni,

Thanks for the quick response. This occurred several times on Oct. 3 2008,
but I did not submit a bug report. I'll try 0.1.9 and let you know, for
certain if it happens again.

Hugo

#505977#25
Date:
2008-11-19 16:47:53 UTC
From:
To:
Hi Evgeni,


It happened again:

Nov 19 10:20:19 debian v86d: Trying to access an unsupported memory region
at cffff
Nov 19 10:20:25 debian v86d: Trying to access an unsupported memory region
at cffff
Nov 19 10:20:28 debian kernel: [20426.182690] v86d[2307]: segfault at 0 ip
08049111 sp bfe52c00 error 4 in v86d[
8048000+18000]
Nov 19 10:20:28 debian kernel: [20430.116179] uvesafb: mode switch failed
(eax=0x4f02, err=1). Trying again with
 default timings.
Nov 19 10:20:28 debian kernel: [20431.481443] v86d[14642]: segfault at 0 ip
08049111 sp bfe443f0 error 4 in v86d
[8048000+18000]
Nov 19 10:20:28 debian kernel: [20435.128020] uvesafb: mode switch failed
(eax=0x4f02, err=1)
Nov 19 10:20:28 debian acpid: client connected from 3686[0:0]
Nov 19 10:20:35 debian v86d: Trying to access an unsupported memory region
at cffff
Nov 19 10:20:40 debian v86d: Trying to access an unsupported memory region
at cffff
Nov 19 10:20:43 debian do_modtemp_hdd[3305]: execute only:
/usr/bin/do_chrony2 (195)
Nov 19 10:20:43 debian do_modtemp_hdd[3305]: 18  24  32  15  31   (201)
Nov 19 10:20:45 debian v86d: Trying to access an unsupported memory region
at cffff
Nov 19 10:20:55 debian last message repeated 2 times

I had been on since 4:40, so almost 6 hours.
As far as I can tell it happened about the time one user logged off from gdm
and logged on again.

dpkg -l v86d:
ii  v86d                   0.1.9-0                daemon to run x86 code in
an emulated environment

Regards,
Hugo

#505977#30
Date:
2008-11-19 20:04:01 UTC
From:
To:
Hi Evgeni,

I don't remember if I told you why I use uvesafb: it's because since
linux-image-2.6.25-1-686 I cannot use 'vga=791':
http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=481063

But I *can* when I roll my own kernel with  the same version from kernel.org
.
It's something that is happening in the Debian kernel.

Don't know if it is related.

Regards,

Hugo

#505977#35
Date:
2008-11-21 11:46:49 UTC
From:
To:
Hi Evgeni,

I can now cause the problem at will by issuing:

startx -- :1 -layout X0 -dpi 110 -deferglyphs 16 -isolateDevice PCI:0:8:0
-sharevts

and I get:

Nov 21 05:24:09 debian kernel: [  391.380003] uvesafb: mode switch failed
(eax=0x14f, err=0). Trying again with
default timings.
Nov 21 05:24:09 debian kernel: [  391.380003] uvesafb: mode switch failed
(eax=0x14f, err=0)
Nov 21 05:24:10 debian kernel: [  392.124005] uvesafb: mode switch failed
(eax=0x14f, err=0). Trying again with
default timings.
Nov 21 05:24:10 debian kernel: [  392.124005] uvesafb: mode switch failed
(eax=0x14f, err=0)
Nov 21 05:24:37 debian kernel: [  418.725455] uvesafb: mode switch failed
(eax=0x14f, err=0). Trying again with
default timings.
Nov 21 05:24:37 debian kernel: [  418.726021] uvesafb: mode switch failed
(eax=0x14f, err=0)

I attach the applicable xorg.conf

Regards,

Hugo

#505977#40
Date:
2008-11-21 18:56:36 UTC
From:
To:
Mh, I can't reproduce that here with my ATI hardware :(
Everything is as stable as it should be.

Will mail upstream later, maybe he has an idea.

Regards

#505977#45
Date:
2008-11-24 23:38:24 UTC
From:
To:
Hi Hugo,

could you please test
http://die-welt.net/~evgeni/v86d/v86d_0.1.9-0+git1_i386.deb

And if that does not help, kernel 2.6.28-rc6?

Regards
Evgeni

#505977#50
Date:
2008-11-25 10:42:13 UTC
From:
To:

Will do. Will let you know. I went to my own rolled 2.6.26.7 + vesafb. Have
to switch systems.

Regards,

Hugo

#505977#55
Date:
2008-11-26 11:05:41 UTC
From:
To:
Hi Evgeni,

Booting Debian 2.6.26-10  with v86d 0.1.9-0+git1 gets:

Nov 25 13:33:00 debian kernel: [  114.182203] v86d[2327]: segfault at 0 ip
08049111 sp bfbcd980 error 4 in v86d$
Nov 25 13:33:00 debian kernel: [  118.508013] uvesafb: mode switch failed
(eax=0x4f02, err=1). Trying again wit$
Nov 25 13:33:00 debian kernel: [  119.177330] v86d[4000]: segfault at 0 ip
08049111 sp bfed2480 error 4 in v86d$
Nov 25 13:33:00 debian kernel: [  123.508012] uvesafb: mode switch failed
(eax=0x4f02, err=1)

Booting Linux version 2.6.28-rc6-hvw with it gets:

Nov 26 04:29:55 debian kernel: [  230.707808] uvesafb: mode switch failed
(eax=0x14f, err=0). Trying again with$
Nov 26 04:29:55 debian kernel: [  230.708371] uvesafb: mode switch failed
(eax=0x14f, err=0)
Nov 26 04:32:41 debian kernel: [  396.911367] uvesafb: mode switch failed
(eax=0x14f, err=0). Trying again with$
Nov 26 04:32:41 debian kernel: [  396.911950] uvesafb: mode switch failed
(eax=0x14f, err=0)

and apparently no segfault! But still an ugly hang, you can't do anything on
the VT's. Progress?
Also for uvesafb 'mode=' has been changed to 'mode_option='.

Regards,

Hugo

#505977#60
Date:
2008-11-27 10:49:53 UTC
From:
To:
Hi Evgeni,

But not segfaulting is not always the case. Yesterday afternoon I tried
2.6.28-rc6 again with that v86d and it *did* segfault:

Nov 26 16:59:41 debian kernel: [  109.489588] v86d[3759]: segfault at 0 ip
08049111 sp bf868820 error 4 in v86d[8048000+18000]
Nov 26 16:59:41 debian kernel: [  113.832773] uvesafb: mode switch failed
(eax=0x4f02, err=1). Trying again with default timings.
Nov 26 16:59:41 debian kernel: [  114.503727] v86d[5397]: segfault at 0 ip
08049111 sp bfd5eb00 error 4 in v86d[8048000+18000]
Nov 26 16:59:41 debian kernel: [  118.846089] uvesafb: mode switch failed
(eax=0x4f02, err=1)
Nov 26 17:00:23 debian kernel: [  151.726743] v86d[5605]: segfault at 0 ip
08049111 sp bfa2a7c0 error 4 in v86d[8048000+18000]
Nov 26 17:00:23 debian kernel: [  156.065180] uvesafb: mode switch failed
(eax=0x4f02, err=1). Trying again with default timings.
Nov 26 17:00:23 debian kernel: [  156.741595] v86d[5620]: segfault at 0 ip
08049111 sp bffba560 error 4 in v86d[8048000+18000]
Nov 26 17:00:23 debian kernel: [  161.077498] uvesafb: mode switch failed
(eax=0x4f02, err=1)
Nov 26 17:00:36 debian kernel: [  164.136780] v86d[5643]: segfault at 0 ip
08049111 sp bfa52ff0 error 4 in v86d[8048000+18000]
Nov 26 17:00:36 debian kernel: [  168.480537] uvesafb: mode switch failed
(eax=0x4f02, err=1). Trying again with default timings.
Nov 26 17:00:36 debian kernel: [  169.149999] v86d[5658]: segfault at 0 ip
08049111 sp bfdacb50 error 4 in v86d[8048000+18000]
Nov 26 17:00:36 debian kernel: [  173.493855] uvesafb: mode switch failed
(eax=0x4f02, err=1)
Nov 26 17:02:34 debian kernel: [  282.312115] v86d[6104]: segfault at 0 ip
08049111 sp bfd97b30 error 4 in v86d[8048000+18000]
Nov 26 17:02:34 debian kernel: [  286.659319] uvesafb: mode switch failed
(eax=0x4f02, err=1). Trying again with default timings.
Nov 26 17:02:34 debian kernel: [  287.324711] v86d[6119]: segfault at 0 ip
08049111 sp bfd90b30 error 4 in v86d[8048000+18000]
Nov 26 17:02:34 debian kernel: [  291.672635] uvesafb: mode switch failed
(eax=0x4f02, err=1)
Nov 26 17:07:39 debian kernel: [  587.901653] v86d[7403]: segfault at 0 ip
08049111 sp bfbf7990 error 4 in v86d[8048000+18000]
Nov 26 17:07:39 debian kernel: [  592.223852] uvesafb: mode switch failed
(eax=0x4f02, err=1). Trying again with default timings.
Nov 26 17:07:39 debian kernel: [  592.896713] v86d[7418]: segfault at 0 ip
08049111 sp bfe74c10 error 4 in v86d[8048000+18000]
Nov 26 17:07:39 debian kernel: [  597.236170] uvesafb: mode switch failed
(eax=0x4f02, err=1)

That's always the case with:

startx -- :1 -layout X0 -dpi 110 -deferglyphs 16 -isolateDevice PCI:0:8:0
-sharevts

starting X on the GeForce 6200 PCI card.

Regards,

Hugo

#505977#65
Date:
2009-06-06 10:45:49 UTC
From:
To:
Hi Hugo,

this bug is quite old, and many things happened since you reported it
(esp the kernel, Xorg etc). Thus I wanted to hear from you, if you
still can reproduce it with latest kernel (2.6.29 or 2.6.30-rcX),
uptodate Xorg from Sid and the appropriate drivers for your GeForce
(nv, nouveau, nvidia).

Regards
Evgeni

#505977#70
Date:
2009-06-07 16:17:09 UTC
From:
To:

Evgeni,

Since my discovery of edd=off as kernel parm to get framebuffer support
again I am using the Debian Sid kernels again. They don't use v86d.
I will try 2.6.29 and see what I get with v86d.
Regards

Hugo

#505977#75
Date:
2010-06-10 15:09:35 UTC
From:
To:
I also ran into this when I updated from 1.23-6 to 1.23-7, but if it
is missing RTC support, how come there is nothing in the changelog?
I presumed that it had to do with my kernel configuration. (But what?)

Hugo Vanwoerkom

#505977#80
Date:
2010-06-10 15:16:24 UTC
From:
To:
Sorry, I apologize, hit the send button too fast,  disregard the
previous message.

Hugo Vanwoerkom