On Wednesday 27 May 2009 18:35:42 Nathan Wharton wrote:
On Wed, May 27, 2009 at 11:21 AM, Nathan Wharton email@example.com wrote:
On Wed, May 27, 2009 at 11:16 AM, Sven Eckelmann firstname.lastname@example.org
Thanks. This is completely unexpecting for a kernel with alignment trap. The expected result is "Value: 12389ab 89abcdef". Can you check the value of /proc/cpu/alignment ? I will look further at it this evening.
User: 0 System: 0 Skipped: 0 Half: 0 Word: 0 DWord: 0 Multi: 0 User faults: 0 (ignored)
I do get the correct result after echo 2 > /proc/cpu/alignment
Ok, so 100% sure that it is a alignment problem. The kernel fixup is always activated and shouldn't create such problems as the page faults you mentioned in your first post. Maybe some communication between kernelland and userland has gone wrong. Can you test batmand with batgat and activated alignment fixup?
Best regards, Sven
(Thanks for the link - it is a real nice overview).