On Thu, Mar 3, 2022 at 9:43 AM Michael <macallan%netbsd.org@localhost> wrote: > > Hello, > > On Wed, 23 Feb 2022 14:40:27 -0800 > foo bar <tokenalt%gmail.com@localhost> wrote: > > > The bad news is that 24bit mode now has some image corruption on the > > bottom of the screen. Roughly a 110 pixel tall strip along the bottom > > gets garbled whenever any window uses acceleration. > > Please try this: > Index: cg14_accel.c > =================================================================== > RCS file: /cvsroot/xsrc/external/mit/xf86-video-suncg14/dist/src/cg14_accel.c,v > retrieving revision 1.28 > diff -u -w -r1.28 cg14_accel.c > --- cg14_accel.c 24 Dec 2021 05:22:54 -0000 1.28 > +++ cg14_accel.c 3 Mar 2022 17:39:24 -0000 > @@ -1495,7 +1495,7 @@ > > pExa->memoryBase = p->fb; > pExa->memorySize = p->memsize; > - pExa->offScreenBase = p->width * p->height * (pScrn->depth >> 3); > + pExa->offScreenBase = p->width * p->height * (pScrn->bitsPerPixel >> 3); > > /* > * SX memory instructions are written to 64bit aligned > addresses with > Using this patch X correctly identifies the amount of off screen memory and x11perf will no longer corrupt the screen. However in the process of testing this patch I found a few programs will cause the X server to seg fault. The easiest one to trigger it is netsurf where all you have to do is move the cursor to the refresh button next to the address bar and it will instantly seg fault the server. I rebuilt the xserver without the patch and it has this issue as well so it does not look like a new problem. X log below.
Description: Binary data