google-chrome again, Re: kernel bug (?)

2010-09-06 Thread A Mennucc
hi list, hi Ben,

this is a followup to the message I posted in March, see
http://lists.debian.org/debian-kernel/2010/03/msg00667.html

With the latest update of the linux kernel and google-chrome, namely
linux-image-2.6.26-2-openvz-amd64   2.6.26-25 
google-chrome-stable   6.0.472.53-r57914 
the bug has changed.

Now when I try to start google-chrome, it immediatly dies, printing
Failed to move to new PID namespace: File exist

and the following backtrace appears in the kernel log,

[15268.091684] kobject_add_internal failed for lo with -EEXIST, don't 
try to register things with the same name in the same directory.
[15268.091693] Pid: 14873, comm: chrome-sandbox Tainted: P  
2.6.26-2-openvz-amd64 #1
[15268.091696] 
[15268.091698] Call Trace:
[15268.091715]  [8030fc29] kobject_add_internal+0x13f/0x17e
[15268.091721]  [80310066] kobject_add+0x74/0x7c
[15268.091733]  [8029f654] percpu_populate+0x77/0x9e
[15268.091738]  [8030e7e5] __next_cpu+0x19/0x26
[15268.091742]  [8030fa80] kobject_get+0x12/0x17
[15268.091748]  [80377e35] get_device+0x17/0x1f
[15268.091752]  [80378456] device_add+0x9b/0x53f
[15268.091757]  [8030f872] kobject_init_internal+0x12/0x2c
[15268.091764]  [803b1f3a] register_netdevice+0x282/0x358
[15268.091769]  [803b2049] register_netdev+0x39/0x46
[15268.091775]  [8037eb71] loopback_net_init+0x3e/0x75
[15268.091781]  [803ae61a] setup_net+0x85/0xe6
[15268.091787]  [803aee28] copy_net_ns+0x49/0xb5
[15268.091793]  [8024ae29] create_new_namespaces+0x125/0x1da
[15268.091800]  [8024b0d2] copy_namespaces+0x66/0x9c
[15268.091807]  [80233759] copy_process+0xa05/0x11fc
[15268.091818]  [8023404d] do_fork_pid+0xfd/0x2aa
[15268.091822]  [80221f52] do_page_fault+0x5c9/0x9ab
[15268.091828]  [8042350a] thread_return+0xc6/0x107
[15268.091836]  [8020c00a] system_call_after_swapgs+0x8a/0x8f
[15268.091840]  [8023420b] do_fork+0x11/0x13
[15268.091844]  [8020c397] ptregscall_common+0x67/0xb0

If I use the command line $ google-chrome -no-sandbox
then chrome starts fine (altough it complains about less security)

a.

ps: please CC me, I am not on the list


-- 
To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Archive: http://lists.debian.org/20100906102636.ga13...@tonelli.sns.it



Re: google-chrome again, Re: kernel bug (?)

2010-09-06 Thread Ben Hutchings
On Mon, 2010-09-06 at 12:26 +0200, A Mennucc wrote:
 hi list, hi Ben,
 
 this is a followup to the message I posted in March, see
 http://lists.debian.org/debian-kernel/2010/03/msg00667.html
 
 With the latest update of the linux kernel and google-chrome, namely
 linux-image-2.6.26-2-openvz-amd64   2.6.26-25 
 google-chrome-stable   6.0.472.53-r57914 
 the bug has changed.
 
 Now when I try to start google-chrome, it immediatly dies, printing
 Failed to move to new PID namespace: File exist
 
 and the following backtrace appears in the kernel log,
 
 [15268.091684] kobject_add_internal failed for lo with -EEXIST, don't 
 try to register things with the same name in the same directory.
 [15268.091693] Pid: 14873, comm: chrome-sandbox Tainted: P  
 2.6.26-2-openvz-amd64 #1
 [15268.091696] 
 [15268.091698] Call Trace:
 [15268.091715]  [8030fc29] kobject_add_internal+0x13f/0x17e
 [15268.091721]  [80310066] kobject_add+0x74/0x7c
 [15268.091733]  [8029f654] percpu_populate+0x77/0x9e
 [15268.091738]  [8030e7e5] __next_cpu+0x19/0x26
 [15268.091742]  [8030fa80] kobject_get+0x12/0x17
 [15268.091748]  [80377e35] get_device+0x17/0x1f
 [15268.091752]  [80378456] device_add+0x9b/0x53f
 [15268.091757]  [8030f872] kobject_init_internal+0x12/0x2c
 [15268.091764]  [803b1f3a] register_netdevice+0x282/0x358
 [15268.091769]  [803b2049] register_netdev+0x39/0x46
 [15268.091775]  [8037eb71] loopback_net_init+0x3e/0x75
 [15268.091781]  [803ae61a] setup_net+0x85/0xe6
[...]

This is OpenVZ-specific broken-ness.  NET_NS is not enabled in the
regular kernel, for good reasons.

Ben.

-- 
Ben Hutchings
Once a job is fouled up, anything done to improve it makes it worse.


signature.asc
Description: This is a digitally signed message part