Bug#236187: xserver-xfree86: [r128] GL lockups

[Daniel Jacobowitz <dan@debian.org>]

Could something based on this patch, or one of the other suggestions in
the bug log, be applied?

[Overquoting for history, since this was a while back.]

On Sun, Mar 07, 2004 at 05:52:53PM -0500, Daniel Jacobowitz wrote:
> On Sun, Mar 07, 2004 at 02:45:59PM -0500, Daniel Jacobowitz wrote:
> > On Sun, Mar 07, 2004 at 06:21:11PM +0100, Michel Dänzer wrote:
> > > On Sun, 2004-03-07 at 17:38, Daniel Jacobowitz wrote:
> > > > 
> > > > Downgrading to xserver-xfree86 4.2.1-16 has eliminated the problem so
> > > > my money is on the 4.3.0 r128 driver.
> > > 
> > > If you agree that your problem seems related to
> > > http://bugs.xfree86.org/show_bug.cgi?id=271, can you try the workarounds
> > > discussed there?
> > 
> > I'll give it a try.  Looking at the code in 4.2.1 and 4.3.0 (you
> > probably know all this already, I'm just educating myself):
> 
> > And yes, looking the source to the DRM module I'm running, it does
> > honor the timeout.  Which defaults to 10_000, so the old working
> > timeout was 10K * 2M * 32, and the new not working timeout is 10K * 32.
> > And the max settable timeout is 100K * 32.  So it seems pretty clear
> > that, however long it takes to return to idle, it's longer than this
> > vastly reduced timeout.
> > 
> > I'm going to try your suggested changes now, along with some
> > instrumentation to see how long resets are really taking.
> 
> I've verified that some xscreensaver modules take up to fifty (as
> opposed to 32) calls to the ioctl before they quiesce.  Probably it
> gets even higher.  So, this is the same problem.
> 
> The code I'm testing is more patient, so I don't know if your fix to
> call R128CCE_STOP helps with the actual resets; the other fellow in the
> bug report said that it did though.  Here's what I was using for
> testing.
> 
> --- xc/programs/Xserver/hw/xfree86/drivers/ati/r128_accel.c.orig	2004-03-07 14:34:13.000000000 -0500
> +++ xc/programs/Xserver/hw/xfree86/drivers/ati/r128_accel.c	2004-03-07 14:39:29.000000000 -0500
> @@ -229,25 +229,35 @@
>  void R128CCEWaitForIdle(ScrnInfoPtr pScrn)
>  {
>      R128InfoPtr info = R128PTR(pScrn);
> -    int         ret, i;
> +    int         ret, i, j;
>  
>      FLUSH_RING();
>  
>      for (;;) {
> +        j = 0;
> +
> +        do {
>          i = 0;
>          do {
>              ret = drmCommandNone(info->drmFD, DRM_R128_CCE_IDLE);
>          } while ( ret && errno == EBUSY && i++ < R128_IDLE_RETRY );
> +        } while ( ret && errno == EBUSY && j++ < R128_TIMEOUT );
>  
>  	if (ret && ret != -EBUSY) {
>  	    xf86DrvMsg(pScrn->scrnIndex, X_ERROR,
>  		       "%s: CCE idle %d\n", __FUNCTION__, ret);
>  	}
>  
> +	if (i != 0 || j != 0) {
> +	    xf86DrvMsg(pScrn->scrnIndex, X_ERROR,
> +		       "%s: CCE idle took j == %d\n", __FUNCTION__, j);
> +	}
> +
>  	if (ret == 0) return;
>  
>  	xf86DrvMsg(pScrn->scrnIndex, X_ERROR,
>  		   "Idle timed out, resetting engine...\n");
> +	R128CCE_STOP(pScrn, info);
>  	R128EngineReset(pScrn);
>  
>  	/* Always restart the engine when doing CCE 2D acceleration */

-- 
Daniel Jacobowitz