On Sun, 07 Mar 1999 at 00:46, Matthias Runge wrote:
> Hi,
> I've got a question/bug-report for the matrox-fb-driver:
> booting with enabled framebuffer via lilo and command line
> video=matrox:vesa:0x114:fh:85kHz,fv=85Hz
^ ","
> works very well. starting an X is possible, but quitting X
> leaves the screen blank. My monitor says signal error, horintal 27 kHz,
> vertical 42 Hz. This is too less to say, whats going on.
Maybe you should indicate, which X server are you running. There
should be no problems with XFree-SVGA 3.3.2 or 3.3.3... Because of
PLL takes some time to lock (.5 sec on Millennium (I/II)), driver reprograms
PLL only when they changes. That means, if you have same pixclock on all
consoles, PLL is programmed only on boot-up. Could you try to
`fbset -pixclock 21212' on any console (say VT3) and then, after exiting
from X, switch from VT1 to VT3 (and back) ? It should fix problem.
Also, are you using XFree in interlaced videomode ?
[side note: works here with XFree 3.3.3, Mystique; even switching from/to
wine directx fallout2 works ok; switching fbtv <-> X too]
> I use kernel version 2.2.2-ac-7 and a matrox mystique
> graphics adaptor, but I experienced this in earlier releases, too.
Could you send me some more info (/etc/X11/XF86Config...) privately,
maybe together with lspci -vx (or, if you do not have NCR SCSI, lspci -vxxx).
and On Sun, 7 Mar 1999, 16:11, Riccardo Facchetti wrote:
> Another report about matroxfb:
> I use matroxfb (dual headed but only 1 monitor attached to a matrox
> mystique 220 ... the second card is a mystique 170).
> I use X86_FBDev.
> I use an MSS sound card.
> When starting X FBDev server _while_ sound card is playing sounds crash my
> machine ... solid lockup ...
Maybe you should check, whether MSS and videocards do not try to share
IRQ. I do not play with IRQ state in device, so it should be disabled.
But maybe it is not and MSS code gets sick from spurious frame IRQs
(but I never saw IRQ enabled; did you run, by any chance, Windows before
Linux without hardware reset?)
Also, could you check whether machine dies when switching videomode by
fbset? There is 5sec timeout for PLL to lock, if PLL does not lock in 5 secs,
with current driver machine dies (because of
`printk(KERN_ERR "PLL not locked after 5 secs, all is lost");'... this
printk() inside videodriver kills computer completely (last time it
happened to me, in 2.1.128, there was no SAK, no num-lock, even no reaction
to ATX poweroff button, fortunately `reset' button is hardwired to whole
computer reset)).
> Sometimes the machine crash when exiting X FBDev too but here I can SAK
> and then ... problems over problems ... one time I have seen SCSI timeouts
> ... another time I have seen GPF's another time I have seen _nothing_ ...
> another more time I have seen nothing (screen blank) ...
> I would really like to help debugging this crashes (I ever thought that X
> FBDev+kernel fb is _much_ more kernel-friendly than the X SVGA ... leading
> to less kernel problems [read crashes] due to the lack of direct-io).
It looks to me like that you have big hardware troubles... Could you
send me output of lspci -vxxx + isapnptools config + soundcard settings
+ boot options + output from fbset ?
Thanks,
Petr Vandrovec
vandrove@vc.cvut.cz
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/