6.9RC1 MergedFB on radeon 7500 breaks

[Benjamin Herrenschmidt]

On Fri, 2005-10-28 at 00:12 -0400, Chao Cao wrote:

That's interesting as your backtrace seem to not involve the radeon driver at
all... I don't know how much of those things can actually be trusted on x86 but
there are a few weird values there, like the pSrc pointer to fbCopyAreammx,
which is wrong. Maybe a bug in fbCompositeCopyAreammx() ? However, the xDst
argument passed to  fbCompositeCopyAreammx() itself looks fairly bad too,
so I'm not too sure at this point what's going on.

Ben.

>     f=0x81f5768 "Caught signal %d.  Server aborting\n") at log.c:408
> #8  0x0808bc9a in xf86SigHandler (signo=11) at xf86Events.c:1412
> #9  <signal handler called>
> #10 fbCopyAreammx (pSrc=0x11c0, pDst=0x82d7340, src_x=0, src_y=0, dst_x=0,
>     dst_y=0, width=2304, height=1024) at fbmmx.c:2218
> #11 0xb7876490 in fbCompositeCopyAreammx (op=1 '\001', pSrc=0xb4991260,
>     pMask=0x0, pDst=0xb4991260, xSrc=0, ySrc=0, xMask=0, yMask=0, xDst=20896,
>     yDst=96, width=0, height=4704) at fbmmx.c:2265
> #12 0xb786eb75 in fbComposite (op=1 '\001', pSrc=0x84720f8, pMask=0x0,
>     pDst=0x84751a0, xSrc=0, ySrc=0, xMask=0, yMask=0, xDst=1, yDst=0,
>     width=2304, height=1024) at fbpict.c:1297
> #13 0xb783b0ad in exaComposite (op=1 '\001', pSrc=0x84720f8, pMask=0x82a9670,
>     pDst=0x84751a0, xSrc=0, ySrc=0, xMask=0, yMask=0, xDst=0, yDst=0,
>     width=2304, height=1024) at exapict.c:471
> #14 0x08192576 in damageComposite (op=1 '\001', pSrc=0xb4991260,
> ---Type <return> to continue, or q <return> to quit---
>     pMask=0xb4991260, pDst=0x84751a0, xSrc=0, ySrc=0, xMask=0, yMask=0,
>     xDst=0, yDst=0, width=2304, height=1024) at damage.c:539
> #15 0x081acabb in ProcRenderComposite (client=0x0) at render.c:755
> #16 0x080d7f81 in Dispatch () at dispatch.c:459
> #17 0x080e80fd in main (argc=4, argv=0xbfaaa3f4, envp=0xb4991260) at main.c:450