| View previous topic :: View next topic |
| Author |
Message |
McTango
n00b
Joined: 11 Dec 2004
Posts: 66
Location: Formosa - Argentina
|
 Posted: Thu Jan 10, 2008 4:52 pm Post subject: Kernel unaligned access on Ultra5 |
 |
|
After upgraded Ultra5's kernel from 2.6.22-r5 to 2.6.23-r3, the system log shows this messages: | Code: | Jan 10 14:00:11 darwin [689098.362740] Kernel unaligned access at TPC[651f34] tcp_transmit_skb+0x240/0x7f0
Jan 10 14:00:11 darwin [689098.362825] Kernel unaligned access at TPC[651f3c] tcp_transmit_skb+0x248/0x7f0
Jan 10 14:00:11 darwin [689098.362867] Kernel unaligned access at TPC[641c2c] ip_queue_xmit+0x218/0x554
Jan 10 14:00:11 darwin [689098.362905] Kernel unaligned access at TPC[641c3c] ip_queue_xmit+0x228/0x554
Jan 10 14:00:11 darwin [689098.362940] Kernel unaligned access at TPC[57c0c4] ip_fast_csum+0xc/0x80 |
I don't know if they are serious or not, and have no idea about what are they.
Thanks.
_________________
Charly Tango
Formosa - Argentina
‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾‾
To emerge, or not to emerge: that is the question.
_________________ |
|
| Back to top |
|
 |
Ixar
n00b
Joined: 02 May 2007
Posts: 15
|
| Posted: Thu Jan 10, 2008 10:03 pm Post subject: |
|
|
It is not that bad 
It only means that an alignment constraint is not met, therefore the kernel catches the unaligned access and performs manually the fetch to overcome the alignment constraint. It is slower as 1 unaligned read requires 2 aligned reads, but it is acceptable as the kernel knows how to do it (it is a feature).
It is logged because on sparc architecture memory access alignement constraints should have been respected therefore it is unusual for the kernel to catch such accesses.
Best Regards,
Ixar |
|
| Back to top |
|
|
|
|
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
|
|
Gentoo on Sparc
