On Tue, 12 Jun 2001, D. Stimits wrote:
> At one point I had the NULL pointer dereference at the same address
> during boot of a new kernel, which did not have SGI patches. It appeared
> to be aic7xxx failure, but was not. I keep mentioning this one patch
> which is part of Alan Cox's ac 2.4.5 series, but which is not yet in the
> main kernel source. Even if you cannot use this patch at all times, can
> you test the following? In linux source, fs/block_dev.c, near line 596
> (depending on kernel version), there is a function "ioctl_by_bdev". In
> that function, add the line I have below that starts with "+" (don't use
> the "+"):
I have only just setup NFS here at work but I will see if I can provoke
some odd behaviour by putting the presure on the NFS server.
The server is currently 2.4.6-pre1 based on my testing machine.
I have run Bonnie with different sizes on this NFS mount and with 2
clients concurrently. I can't make it go boom yet. The 3Com 100Mbit switch
at works is having a harder time then the actual NFS server which is a
desktop machine. The clients are SMP and the server desktop machine is UP.
The current CVS 2.4.6-pre2 tree at least contains the patch/fix mentioned
above. The 2.4.6-pre1 tree does not contain this fix?
The machines at work currently use this tree. And they did not crash yet.
You seem to be using a ramdisk. Can you compile the kernel to get rid of
that? (it's just that I don't like ramdisks).
I personally compile all the important hardware drivers like network scsi
and fs into the kernel.
Seth
|