At 13:40 10.03.2002 +0000, Chris Clayton wrote:
EIP; c012d67f
<===== Trace; c01c5d46 <__make_request+86/5b0> Trace; c01c6308 Trace; d089c93f <[pktcdvd]pkt_read_bh+3b/50> I don't have access to any proper Linux workstation right now, but this looks like it could a problem with highmem support.
Do you have support for more than 1GB of physical RAM compiled in your kernel and does changing the option make a difference? Arnd <><
On Sun, 10 Mar 2002, Arnd Bergmann wrote:
At 13:40 10.03.2002 +0000, Chris Clayton wrote:
EIP; c012d67f
<===== Trace; c01c5d46 <__make_request+86/5b0> Trace; c01c6308 Trace; d089c93f <[pktcdvd]pkt_read_bh+3b/50> I don't have access to any proper Linux workstation right now, but this looks like it could a problem with highmem support. Do you have support for more than 1GB of physical RAM compiled in your kernel and does changing the option make a difference?
Yes, this is indeed a highmem problem. This patch should fix it. --- linux/drivers/block/pktcdvd.c.old Sun Mar 10 16:10:00 2002 +++ linux/drivers/block/pktcdvd.c Sun Mar 10 16:01:45 2002 @@ -264,6 +264,7 @@ kmem_cache_free(bh_cachep, bh); break; } + bh->b_page = virt_to_page(bh->b_data); spin_lock_irq(&pd->lock); bh->b_pprev = &cdrw->bhlist; -- Peter Osterlund - petero2@telia.com http://w1.894.telia.com/~u89404340
Peter, Thanks for the patch - it does indeed stop the oops when highmem is configured on. Chris --- This message has been virus checked by Grisoft AVG during dispatch. Checked by AVG anti-virus system (http://www.grisoft.com). Version: 6.0.336 / Virus Database: 188 - Release Date: 11/03/02
participants (3)
-
Arnd Bergmann
-
Chris Clayton
-
Peter Osterlund