The Answer Gang

re: the answer guy

From hazmouz on Thu, 21 Sep 2000

can you please help me to find the driver for linux or linuxlike of the SIS 6326 graphic card
thanks a lot

PS: excuse me for this sudden intrusion in your life

...and...

Vorrei ricevere i driver d'installazione per SIS 6326

From Cataldo Pellegrino on Mon, 28 Aug 2000

Now as it turns out, none of the usual Gang answered, but it had been discussed aboard my Star Trek shuttle mailing list a few months ago...

Beware the SiS 6326!

Answered by: The Armadillo with the Mask

Politely:

It's a real bother to get working under XFree86.

Frankly:

[ Hey, we can't print that sort of punctuated language here! not without entities anyway. ]... #$@%!!!!!!!

In fact, if you have this or a derivative card, don't even try. Go out immediately and get Xfree86 4.0. This is not a matter of compatibility it is a matter that it just doesn't work at all under anything but. You need at least 3.3.6 to get even faltering support let alone anything that actually works.

Seeing as 3.3.6 is still the current version, it's probably still the case. Feel free to build it yourself from the source repository if you feel up to it, but, then you may as well try the new version instead and get real support. -- Heather

So what is it?---A very popular chipset on the low-end $30-50 SVGA card market. I've got one in komodo for preciselyt that reason. When I first put the machine together, I just needed a card that would allow me to stand the box until I could put it on a TTY and decide what I wanted to do for a real card/monitor combo. As such I never tried to bring X up on it. They are in a number of the testbed boxen at [my work] for similar reasons.

I kept running into the same problem. I could run 'startx' by hand, but whenever I ran /etc/init.d/xdm start, 'parse_xf86config' would complain that there was an error in the XF86Config in /etc/X11 and refuse to start xdm. First, I noticed that the variable PROBLEM in /etc/init.d/xdm is automatically set to "yes" with the intent that if the parse_xf86config went well, PROBLEM would get set to null and all would be well.

It turns out this never happens and even when the output of parse_xf86config is clean, PROBLEM doesn't get reset. I was having other problems too. The screen interlaced horribly and occasionally blanked out and then came back. The cursor was a big huge blob. I was really just about to toss the card across the room. A final search on RedHat under "all Linux sites" turned up the answer. If you add the following under XF86Config, it works.

Option no_bitblt
Option no_imageblt
Option no_accel
Option sw_cursor

This also fixed the issue I was having with PROBLEM in /etc/init.d/xdm which is the REALLY bizarre part. From all appearances, it looks like parse_xf86config is not only doing a syntactical check on the config file but that it is also checking to see if the options that you are specifying actually WORK on the card or if it's going to cause trouble.

I have the exact same /etc/init.d/xdm script on chameleon, which is also running debian and it runs without a hitch(it is however, using the vastly more ubiquitous C&T 65550 chipset)

Moral of the story:

Avoid cheapy video cards and when you can't or REALLY don't want to, avoid the SiS 6326. Dropping an extra $50 is far more worth your time.

Also, startup scripts aren't infallible. I'm still wondering if there are maybe some bugs in the Debian /bin/sh that was causing the evaluation problems with /etc/init.d/xdm I was seeing.

Note: Probably resolved if you stay on Potato; he was looking at Debian/frozen in mid July, and it didn't release until mid August. -- Heather

"That's my story and I'm stickin' to it!"

'dillo

Copyright © 2000, James T. Dennis
Published in the Linux Gazette Issue 58 October 2000

HTML transformation by of Tuxtops, Inc., http://www.tuxtops.com/

1   2   3   4   5   6   7