On Mon, 2015-09-28 at 22:49 +0200, Thomas Klausner wrote: > PCI_REGION_BASE returns a 64-bit value, which needs to be converted > down to 32-bit before it can be put into a 32-bit pointer.
Ugh, DGA sure is awful. The problem here is if the fb bar really is above 4G this is going to return garbage to the client; but that's not really different from what happens anyway (since the dispatch code never fills in mem2 on 32-bit-pointer machines), so I guess the patch isn't wrong so much as not a real improvement. The OpenFramebuffer interface here is just wrong, the driver ought to return a u64 and the server should copy that into the reply properly. I'm in no hurry to "fix" DGA so I've gone ahead and applied these two patches. remote: I: patch #60569 updated using rev 9c8a08c9c7a0f7ec591042abd5b4a0fb50c028f1. remote: I: patch #60568 updated using rev 47f43694645f3e6347b536d5e3fb2fee56e715ab. remote: I: 2 patch(es) updated to state Accepted. To ssh://git.freedesktop.org/git/xorg/driver/xf86-video-s3virge 8a54ff3..47f4369 master -> master - ajax _______________________________________________ [email protected]: X.Org development Archives: http://lists.x.org/archives/xorg-devel Info: http://lists.x.org/mailman/listinfo/xorg-devel
