Re: Possible bug or flaw detected.
On Wed, 26 Nov 1997, Wintermute wrote:
> From: Wintermute <wntrmute@tampabay.rr.com>
> Subject: Possible bug or flaw detected.
...
> Attached to this email are applicable files from my Debian Linux 2.0.29
> installation.
> I have compiled 3c509.c with #define EL3_DEBUG 9 to ensure that I get
> enough error output to show what's going on.
..
> The 3c509 card in my present system is identical to those we use at
> work. The boxes at work are Compaq Deskpro P100's running Debian Linux
> as well. The 3c509 driver has no problem working on those. My home
> machine (and current problem child), is running Debian 2.0.29 (as I
> mentioned) and is an AMD 5x86 133 w/ True Green 486 main board (pretty
> generic board.. has the ability to run most any 486 class chip.. which
> the 5x86 is technically).
>
> Now I've tried compiling the driver into the kernel and as a module.
> I've even downloaded the 1.14 version driver and put that in place of
> the 1.07 driver. I've remembered to use the ether= and reserve= on the
> lilo boot prompt and the io= and irq= (except in the case of the newest
> version where io is not an option).
...
> I have stripped the machine down to bare essentials in my adventure so
> far. The last tested configuration was thus:
Good. The usual conflict problem has been with AWE32 boards at the
activation address of 0x100. The updated 3c509 driver new checks for a free
activation address.
Another common problem with any board is not actually running the updated
kernel. Checking /proc/version shows the build date of the running kernel.
Testing with driver modules avoids this problem.
> I'm really at wits end here, and the only thing I can think of after
> watching the diagnostic output and stepping through the source of the
> driver is that the driver is not reading the values for the card from
> the right address.. the 6b50 value that your driver is testing is in the
> output from el3-diag right where it should be, however if you notice the
> dmesg output, when it tries to get that value it find some random word
> lying around, nothing even close to 6b50.
You didn't include the 'dmesg' output in the attachments. What value is
being detected?
> I am thinking that perhaps whatever method is being used to determine
> the settings in el3-diag could be applied to the 3c509 driver, and
> perhaps things would work correctly. However, seeing as the main code
> in el3.c and 3c509.c is almost identical I don't know why 3c509 wouldn't
> work in the first place.
The two activation sequences are deliberately nearly identical.
A difference is the timing routines: the user-level diagnostic might be
doing something different for udelay().
Are you are running your ISA bus at a very high speed?
Donald Becker becker@cesdis.gsfc.nasa.gov
USRA-CESDIS, Center of Excellence in Space Data and Information Sciences.
Code 930.5, Goddard Space Flight Center, Greenbelt, MD. 20771
301-286-0882 http://cesdis.gsfc.nasa.gov/pub/people/becker/whoiam.html
--
TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to
debian-user-request@lists.debian.org .
Trouble? e-mail to templin@bucknell.edu .
Reply to: