Mailinglist Archive: packet-writing (124 mails)

[Jens Axboe <axboe@xxxxxxx>]

On Mon, Apr 23 2001, Adam Huffman wrote:
> On Mon, 23 Apr 2001, Jens Axboe wrote:
> 
> > > >>EIP; 00000000 Before first symbol
> > > Trace; c017e9ea <generic_make_request+14a/160>
> > > Trace; c017ea5d <submit_bh+5d/90>
> > > Trace; c017ece7 <ll_rw_block+207/250>
> > 
> > Queue problem again, looks like blk_get_queue returned queue that wasn't
> > setup (NULL q->make_request_fn). Most of these look pretty odd -- it
> > could be leftover from the very first oops you posted I suppose, if they
> > happened without a reboot in between?
> > 
> 
> No, there was no reboot.  There was a flurry of Oopsen (about a dozen),
> which stopped, then started again when I tried to change virtual
> console, some Alt-SysRq combinations and Ctrl-Alt-Del.  Eventually I had
> to use the reset button.

Ok, that's what I suspected. Then the following oopsen are most likely
due to the first panic, although they really shouldn't be in this
case... Oh well.

> > If not, then... What compiler did you use?
> > 
> 
> Latest RedHat gcc-2.96-81

Hmm, I'll keep this in mind.

> > > Hardware is a Yamaha 8424 internal SCSI on an Adaptec 2904.
> > 
> > Ok, should be good.
> > 
> 
> A few kernels ago I was having some problems with the new Adaptec
> driver, which I referred to Justin Gibbs;  they seem to have been solved
> with the 6.1.11 driver, but it's possible a remaining problem there is
> interacting with your packet stuff.

Shouldn't be, I'm using the very same driver for SCSI testing here. The
only requirements of the low level scsi driver (and card) for pktcdvd is
that it can handle at least 32 segments in a request.

-- 
Jens Axboe