1 |
> The current kernel that I am running is 2.6.26.5-rt9-gentoo2. |
2 |
> Sporadically this freezes, usually happening during the end of |
3 |
> emerging an ebuild or when stopping gpm. Usually, the fb only freezes |
4 |
> for a few seconds, but sometimes I have to <Alt><SysRq>O the machine |
5 |
> (it's unresponsive to <Alt><SysRq>{R,S,E,I,K,U}). |
6 |
> I have the feeling that this is a regression because the framebuffer |
7 |
> never froze on 2.6.25.4-rt5-gentoo nor on the kernel on the 2008.0 |
8 |
> install/minimal CD. |
9 |
> |
10 |
> Any suggestions? |
11 |
> |
12 |
> Note: I emailed about this on lkml, yet got no replies. |
13 |
> |
14 |
|
15 |
I think something similar happened to me; I am using vanilla-sources. |
16 |
As for version, see |
17 |
$ uname -a |
18 |
Linux jorge 2.6.26.6 #1 Fri Oct 10 00:52:35 BRT 2008 i686 AMD |
19 |
Athlon(tm) XP 2600 |
20 |
+ AuthenticAMD GNU/Linux |
21 |
(although, at the time of the problem, it might have been 2.6.26.5) |
22 |
|
23 |
The computer froze during an emerge of multiple ebuilds (I couldn't |
24 |
know if the emerge had finished or not, but I intensely hope it had - |
25 |
I hate system inconsistencies and filesystem corruptions). |
26 |
The system were completely unresponsive to <ctrl><alt>fn and the magic |
27 |
sysrq keys, except <Alt><SysRq>o. In fact, even the keyboard leds were |
28 |
unresponsive AFAIR (as far as I recall). |
29 |
|
30 |
Note: Despite the system *seeming* unresponsive to the other |
31 |
<alt><sysrq> keys, *maybe* they were indeed effective, and I did the |
32 |
whole "<Alt><SysRq>r, <Alt><SysRq>e, wait 5 seconds, <Alt><SysRq>i, |
33 |
<Alt><SysRq>s, wait 5 seconds, <Alt><SysRq>u, wait 5 seconds, |
34 |
<Alt><SysRq>o" hoping that this would cleanly shutdown my computer. |
35 |
AFAIR, the next boot did indicate that the filesystem had been cleanly |
36 |
umounted) |
37 |
|
38 |
|
39 |
On 2008-10-18 it happened again. Using vanilla sources 2.6.26.6. Again, it was |
40 |
after an emerge (but it wasn't immediately after, it was some minutes |
41 |
later). For the record, I am using a tmpfs on PORTAGE_TMPDIR (and it |
42 |
is 1536 MB, far more than enough for emerging the programs I was |
43 |
emerging - gimp-help, yasm, libpcre, subversion, giflib - and I have 1 |
44 |
GB RAM and 972 MB of swap - and I was not running any memory-intensive |
45 |
apps (just enlightenment, gkrellm, pidgin, and firefox viewing a |
46 |
single page on Ubuntu wiki) The emerge went fine with absolutely no |
47 |
errors. Some minutes later, when I was preparing to shutdown the |
48 |
system, the system became unresponsive just after I tried to close |
49 |
pidgin. |
50 |
|
51 |
It seems that the system becomes completely unresponsive to keyboard |
52 |
(except the magic sysrq keys); <ctrl><alt>fn , <ctrl><alt>backspace , |
53 |
<ctrl><alt><del>, nothing responds. The screen is completely frozen |
54 |
(so I don't know if the computer responds to the mouse). The system |
55 |
itself isn't frozen though - I have cron jobs that beep every 15 |
56 |
minutes (to keep me informed of the time), and I heard the expected |
57 |
beep when the time reached 7:00 (I woke up early, at 5:45). I then |
58 |
performed <alt><sysrq>r, but the system seemed still unresponsive to |
59 |
keyboard. I left the system idle for more than 30 minutes, to see if |
60 |
it would recover. It didn't. |
61 |
I then rebooted with the EISUB sysrq keys combination. The system |
62 |
indeed responded to theses keys - it rebooted and, on the next boot, |
63 |
the filesystem was reportedly clean. |
64 |
|
65 |
I was using e17. Since i changed to Xfce, I haven't had this problem. |
66 |
It is very possible, though, that this was just a coincidence, and the |
67 |
bug is not e17's fault |
68 |
|
69 |
-- |
70 |
Software is like sex: it is better when it is free - Linus Torvalds |