| Author |
Message |
Rikhard
|
|
Post subject: Kernel 3.4.0, a buggy kernel?
 Posted: 22.05.2012, 22:24
|
|
Joined: 2011-03-24
Posts: 6
Status: Offline
|
|
After dist-upgrade to kernel 3.4-0.slh.1-aptosid-amd64 my monitor shuts down during boot and I am forced to manually restart. The old kernel 3.3-6-slh2 is OK and I can log in.
Here some extracts from /var/log/kern.log
10.233775] Pid: 552, comm: modprobe Tainted: G D 3.4-0.slh.1-aptosid-amd64 #1
[ 10.234417] Call Trace:
[ 10.234763] [<ffffffff81319bde>] ? __schedule_bug+0x3e/0x45
[ 10.235090] [<ffffffff81320032>] ? __schedule+0x93/0x494
[ 10.235410] [<ffffffff8104fdc0>] ? __cond_resched+0x1d/0x26
[ 10.235735] [<ffffffff81320480>] ? _cond_resched+0x12/0x1c
[ 10.236087] [<ffffffff810b1850>] ? unmap_single_vma+0x56a/0x5d6
[ 10.236435] [<ffffffff810d9df1>] ? lookup_page_cgroup+0x23/0x2c
[ 10.236784] [<ffffffff810a313b>] ? add_page_to_lru_list+0x63/0x63
[ 10.237134] [<ffffffff810b1f17>] ? unmap_vmas+0x4e/0x79
[ 10.237484] [<ffffffff81185b13>] ? hweight_long+0x5/0x6
[ 10.237828] [<ffffffff810b6e25>] ? exit_mmap+0x97/0xe4
[ 10.238173] [<ffffffff81031097>] ? mmput+0x51/0xf8
[ 10.238514] [<ffffffff81036274>] ? exit_mm+0x10b/0x116
[ 10.238849] [<ffffffff81321045>] ? _raw_spin_lock_irq+0xb/0x25
[ 10.239185] [<ffffffff810364c4>] ? do_exit+0x245/0x799
[ 10.239529] [<ffffffff8103490a>] ? kmsg_dump+0x8e/0xec
[ 10.239860] [<ffffffff8100c017>] ? oops_end+0xa4/0xa9
[ 10.240196] [<ffffffff81318e54>] ? no_context+0x241/0x26c
[ 10.240549] [<ffffffff81029be2>] ? do_page_fault+0x138/0x2f2
[ 10.240906] [<ffffffffa046873b>] ? nv04_timer_read+0x33/0x46 [nouveau]
[ 10.241271] [<ffffffffa046873b>] ? nv04_timer_read+0x33/0x46 [nouveau]
[ 10.241619] [<ffffffffa0447ea9>] ? nouveau_wait_eq+0x3a/0x6a [nouveau]
[ 10.241972] [<ffffffffa046873b>] ? nv04_timer_read+0x33/0x46 [nouveau]
[ 10.242341] [<ffffffffa0447ea9>] ? nouveau_wait_eq+0x3a/0x6a [nouveau]
[ 10.242692] [<ffffffff813215a5>] ? page_fault+0x25/0x30
And it goes on in the same style, there are also several bugs, here are some:
BUG: scheduling while atomic: modprobe/552/0x10000003
BUG: unable to handle kernel NULL pointer dereference at (null)
[ 10.194166] IP: [<ffffffff8131f82c>] __mutex_lock_common.isra.7+0xd0/0x1cd
[ 10.194272] PGD 0
[ 10.194355] Oops: 0002 [#1] PREEMPT SMP
My boot parameters are:
BOOT_IMAGE=/boot/vmlinuz-3.4-0.slh.1-aptosid-amd64 root=UUID=80e05083-cba1-4f3d-9832-01354e4277dd ro single
and I use the nouveau driver for my Nvidia Geforce 8800 GTS 512 graphic card.
What I think is best to do in this situation is just to sit down and wait for the next kernel upgrade. According to http://www.kernel.org/ the latest stable kernel is 3.3.7 but we have this buggy 3.4.0.
But what I want to ask is:
What do all those question marks mean in kern.log. I suppose something is wrong then.
And if anyone else has experienced the same problem? |
|
|
| |
|
|
|
 |
|
slh
|
|
|
Post subject: RE: Kernel 3.4.0, a buggy kernel?
Posted: 22.05.2012, 23:03
|
|
Joined: 2010-08-25
Posts: 584
Status: Offline
|
|
"Tainted: D" means that there already was an OOPS or BUG before the one you quoted, getting to know that first one would be important, because subsequent ones may already have to deal with memory corruption and therefore aren't useful for debugging anymore.
That said, how do you define a "buggy kernel" in generic terms?
Given the size of the kernel, it's a statistic certainty that "something" will always be broken, the only question is if the breakage affects you or someone else. E.g. I know first hand that it works fine on several dozens of very diverse systems (on some of them for a couple of weeks already, among them our webserver) and I've also not gotten any bugreports within the team so far. Therefore there is no need for a big public health warning (on the contrary, the more bug reports we get, the easier it gets to find something in common between them), the big task is now to understand what fails on your specific system - ideally also why and how to fix it. Trying to sit it out and waiting for a fix to magically appear out of nothing, usually has a relatively slim chance of success.
While not necessarily representative for contemporary nouveau support, I know that it is working fine on an old system equipped with a GeForce 2MX, which is the only one I have available for testing nouveau. |
|
|
| |
|
|
|
|
|
Spaceferret
|
|
|
Post subject: RE: Kernel 3.4.0, a buggy kernel?
Posted: 23.05.2012, 05:20
|
|
Joined: 2010-09-13
Posts: 96
Status: Offline
|
|
| NVIDIA proprietary driver won't build for 3.4.0, I use .run package from NVIDIA website, worked perfectly with all previous kernels so far. The package builder displays error and exits. Also, shorewall exits with an error during boot time, went back to the old kernel, everything is back to normal |
|
|
| |
|
|
|
|
|
Rikhard
|
|
|
Post subject: RE: Kernel 3.4.0, a buggy kernel?
Posted: 23.05.2012, 10:19
|
|
Joined: 2011-03-24
Posts: 6
Status: Offline
|
|
|
Quote:
how do you define a "buggy kernel" in generic terms?
My headline was “Kernel 3.4.0, a buggy kernel?” formulated as a question, to find out if it is functional for others. I have several Linux distributions in my computer, and kern.log for this new kernel in Aptosid compared to kern.logs in the other distributions shows more bugs and Oops. It is in fact still an unstable kernel so I suppose that is normal.
Quote:
the only question is if the breakage affects you or someone else
Yes, that is what I want to know. It certainly affects me. I am glad the kernel works well for other users of Aptosid so perhaps the the problem I have is related to my hardware.
Quote:
Trying to sit it out and waiting for a fix to magically appear out of nothing, usually has a relatively slim chance of success.
The fault is obviously in the kernel. Trying to fix something to work in an unstable kernel is challenging and out of my knowledge. So better wait and see till next kernel upgrade. Interesting of course is to know what led to the monitor shutdown. If this had happened in a stable distribution I certainly had tried to pin down the problem, but in this case in Aptosid with massive upgrades every day I will wait for that “ slim chance of success” till I take any radical steps, because I can use my old kernel.
What I have read on Google the Nouveau driver should work on the kernel, and I want to keep that driver.
But still to my question:
What do all those question marks mean in kern.log? |
|
|
| |
|
|
|
|
|
Rikhard
|
|
|
Post subject: RE: Kernel 3.4.0, a buggy kernel?
Posted: 25.05.2012, 09:23
|
|
Joined: 2011-03-24
Posts: 6
Status: Offline
|
|
| After the kernel upgrade to 3.4.0-slh2 the monitor shuts down but not in recovery mode anymore. That is a progress. |
|
|
| |
|
|
|
|
|
Rikhard
|
|
|
Post subject: RE: Kernel 3.4.0, a buggy kernel?
Posted: 25.05.2012, 16:14
|
|
Joined: 2011-03-24
Posts: 6
Status: Offline
|
|
|
|
|
|
|
stefutz
|
|
|
Post subject: RE: Kernel 3.4.0, a buggy kernel?
Posted: 27.05.2012, 16:18
|
|
Joined: 2010-11-29
Posts: 19
Status: Offline
|
|
Hi,
I also have a problem with the 3.4.0 kernel. System responds very slowly and sometimes freezes, windows tear up when moving, looks like a bad video driver.
When dist-upgrade from 3.3.6, a message popped out saying that firmware files are missing.
I want to mention that both contrib and non-free are added to debian repositories.
Here is a link with a shot on the screen: http://en.zimagez.com/zimage/dscn43203.php
I also reverted to 3.3.6
Cheers,
Stefan |
|
|
| |
|
|
|
|
|
|
