Bug#206907: xserver-xfree86: hangs on screensaver
On Tue, Aug 26, 2003 at 02:55:21PM -0500, Branden Robinson wrote:
> On Tue, Aug 26, 2003 at 07:45:13PM +0200, David N. Welton wrote:
> > Enabled PciRetry.
> >
> > It bailed again, this time in another place (I guess):
> >
> > 0x080833b5 in Permedia2SubsequentHorVertLine (pScrn=0x8942ef0, x=864, y=7,
> > len=5, dir=3) at pm2_accel.c:583
> > 583 pm2_accel.c: No such file or directory.
> > in pm2_accel.c
> > (gdb) bt
> > #0 0x080833b5 in Permedia2SubsequentHorVertLine (pScrn=0x8942ef0, x=864, y=7,
> > len=5, dir=3) at pm2_accel.c:583
>
> Here's the source code of that function:
>
> 578 static void
> 579 Permedia2SubsequentHorVertLine(ScrnInfoPtr pScrn,int x,int y,int len,int dir)
> 580 {
> 581 GLINTPtr pGlint = GLINTPTR(pScrn);
> 582
> 583 GLINT_WAIT(6);
> 584 GLINT_WRITE_REG(x<<16, StartXDom);
> 585 GLINT_WRITE_REG(y<<16, StartY);
> 586 if (dir == DEGREES_0) {
> 587 GLINT_WRITE_REG(1<<16, dXDom);
> 588 GLINT_WRITE_REG(0<<16, dY);
> 589 } else {
> 590 GLINT_WRITE_REG(0<<16, dXDom);
> 591 GLINT_WRITE_REG(1<<16, dY);
> 592 }
> 593
> 594 GLINT_WRITE_REG(len, GLINTCount);
> 595 GLINT_WRITE_REG(PrimitiveLine, Render);
> 596 }
This don't bring us much, it is hanging in glint_wait, which is a macro
waiting for space in the fifo, and since the space never comes, it
waits forever. Same diagnostic as before. The pipeline is deadlocked,
but i don't know why. Strong suspisions of a hardware problems, the
hardware cannot be all that new after all. But maybe there is some
behavior that makes this appear more often or something such, i don't
know. Overheating of the graphic chip or failing powersupply (either on
the board or the agp power lines) might be potential culprits.
I will try to build the full 4.3.0-0pre1v1 (BTW, why both pre1 and v1,
is this not doing double usage) with the glint debug option enabled, so
you can run the non-loadable server in gdb too. I need to free space
again on my disks (or buy a new disk) though.
Friendly,
Sven Luther
>
> --
> G. Branden Robinson | Damnit, we're all going to die;
> Debian GNU/Linux | let's die doing something *useful*!
> branden@debian.org | -- Hal Clement, on comments that
> http://people.debian.org/~branden/ | space exploration is dangerous
Reply to: