Re: Possible bug in xaaFillPoly.c
Mark Vojkovich wrote: Can you try the attached patch? X work stable now, core is not dumped, and verification test (xsuit, chapter 06) passed too. I guess, you may commit the patch. ___ Devel mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/devel
Re: Possible bug in xaaFillPoly.c
Thanks for verifying. The fix has been committed to CVS. Mark. On Mon, 24 Mar 2003, Alexandr Andreev wrote: Mark Vojkovich wrote: Can you try the attached patch? X work stable now, core is not dumped, and verification test (xsuit, chapter 06) passed too. I guess, you may commit the patch. ___ Devel mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/devel
Re: Possible bug in xaaFillPoly.c
This is not the correct solution. Mono8x8PatternRectHelper needs
a pattern, yet the API doesn't pass one. It was intended that
it be passed in pCache-pat, yet that wasn't set up. There
was a scratch cache structure set aside for that type of thing.
Can you try the attached patch?
MArk.
On Wed, 19 Mar 2003, Mark Vojkovich wrote:
There's definitely a bug there. I'm not sure if that's the
correct solution, or there is a bigger problem. I will review
this code tonight. Thanks for pointing this out.
Mark.
On Wed, 19 Mar 2003, Alexandr Andreev wrote:
Hi,
Possible, there is a bug in the Mono8x8PatternRectHelper() function.
Lets look at XAAFillPolygonStippled():
Assume:
1. type == DO_MONO_8x8,
2. infoRec-Mono8x8PatternFillFlags:
HARDWARE_PATTERN_SCREEN_ORIGIN is NOT set,
HARDWARE_PATTERN_PROGRAMMED_BITS is set,
HARDWARE_PATTERN_PROGRAMMED_ORIGIN is NOT set.
In this case, pCache == NULL, and RectFunc = Mono8x8PatternRectHelper
Later, in Mono8x8PatternRectHelper() we will crash when trying to access
pCache-pat0.
I think that this patch may help, but i'm not sure.
*** xaaFillPoly.c.old Thu Mar 20 10:46:51 2003
--- xaaFillPoly.c Thu Mar 20 10:54:21 2003
***
*** 770,776
pCache = (*infoRec-CacheMono8x8Pattern)(
infoRec-pScrn, patx, paty);
patx = pCache-x; paty = pCache-y;
! }
RectFunc = Mono8x8PatternRectHelper;
}
--- 770,780
pCache = (*infoRec-CacheMono8x8Pattern)(
infoRec-pScrn, patx, paty);
patx = pCache-x; paty = pCache-y;
! } else {
! pCache = (infoRec-ScratchCacheInfoRec);
! pCache-pat0 = patx;
! pCache-pat1 = paty;
! }
RectFunc = Mono8x8PatternRectHelper;
}
Possible bug in xaaFillPoly.c
Hi,
Possible, there is a bug in the Mono8x8PatternRectHelper() function.
Lets look at XAAFillPolygonStippled():
Assume:
1. type == DO_MONO_8x8,
2. infoRec-Mono8x8PatternFillFlags:
HARDWARE_PATTERN_SCREEN_ORIGIN is NOT set,
HARDWARE_PATTERN_PROGRAMMED_BITS is set,
HARDWARE_PATTERN_PROGRAMMED_ORIGIN is NOT set.
In this case, pCache == NULL, and RectFunc = Mono8x8PatternRectHelper
Later, in Mono8x8PatternRectHelper() we will crash when trying to access
pCache-pat0.
I think that this patch may help, but i'm not sure.
--- xc.orig/programs/Xserver/hw/xfree86/xaa/xaaFillPoly.c Wed Mar 19 20:27:42
2003
+++ xc/programs/Xserver/hw/xfree86/xaa/xaaFillPoly.cWed Mar 19 20:29:26 2003
@@ -535,9 +535,9 @@
yorg = (y - yorg) 0x07;
if(!(infoRec-Mono8x8PatternFillFlags
- HARDWARE_PATTERN_PROGRAMMED_ORIGIN)){
+ HARDWARE_PATTERN_PROGRAMMED_BITS)){
if(infoRec-Mono8x8PatternFillFlags
- HARDWARE_PATTERN_PROGRAMMED_BITS) {
+ HARDWARE_PATTERN_PROGRAMMED_ORIGIN) {
int patx = pCache-pat0;
int paty = pCache-pat1;
XAARotateMonoPattern(patx, paty, xorg, yorg,
Re: Possible bug in xaaFillPoly.c
There's definitely a bug there. I'm not sure if that's the correct solution, or there is a bigger problem. I will review this code tonight. Thanks for pointing this out. Mark. On Wed, 19 Mar 2003, Alexandr Andreev wrote: Hi, Possible, there is a bug in the Mono8x8PatternRectHelper() function. Lets look at XAAFillPolygonStippled(): Assume: 1. type == DO_MONO_8x8, 2. infoRec-Mono8x8PatternFillFlags: HARDWARE_PATTERN_SCREEN_ORIGIN is NOT set, HARDWARE_PATTERN_PROGRAMMED_BITS is set, HARDWARE_PATTERN_PROGRAMMED_ORIGIN is NOT set. In this case, pCache == NULL, and RectFunc = Mono8x8PatternRectHelper Later, in Mono8x8PatternRectHelper() we will crash when trying to access pCache-pat0. I think that this patch may help, but i'm not sure. ___ Devel mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/devel
