David Rusling wrote:
>
> Actually, you have to turn alignment traps on for the networking layer
> to work.
Actually there are very few such cases left. Most of my development
machines show zero alignment faults after several days of operation. On
busier machines, we do get a few here and there:
>From www.netwinder.org:
$ uname -a
Linux yuri 2.2.14-20000612 #1 Mon Jun 12 18:08:37 EDT 2000 armv4l
unknown
$ uptime
10:31am up 223 days, 22:17, 1 user, load average: 0.04, 0.04, 0.01
$ cat /proc/sys/debug/alignment
User: 7
System: 32
Skipped: 0
Half: 0
Word: 32
Multi: 0
>From ftp.netwinder.org:
$uname -a
Linux kei 2.2.14-20000612 #1 Mon Jun 12 18:08:37 EDT 2000 armv4l
unknown$ uptime
10:32am up 223 days, 22:29, 1 user, load average: 0.19, 0.08, 0.01
$ cat /proc/sys/debug/alignment
User: 506
System: 136
Skipped: 0
Half: 0
Word: 136
Multi: 0
As things stand right now, user-space faults are trapped, but no action
is taken - the application receives bad data. Kernel faults are
corrected.
-Ralph
_______________________________________________
http://lists.arm.linux.org.uk/mailman/listinfo/linux-arm
Please visit the above address for information on this list.
