The following checkin to XFree86 CVS broke the savage driver for some Savage MX and Savage IX users. Tim Roberts put up a binary driver 1.1.27t, which users have tested and said it worked for them. Tim just released the sources for 1.1.27t a couple of days ago, and I've gotten people to test his stock 1.1.27t driver, as well as 1.1.27t patched with differences between 1.1.26t and 1.1.26 in XFree86 CVS. I called this driver 1.1.27mh.
Testing of 1.1.27h caused the same failures, implying that some difference between Tim's stock drivers and what is in XFree86 CVS had caused the breakage. A user who has this hardware diffed the log files from a good working driver and the non-working driver and the only differences were that VBE was not being enabled and used. I looked through the 1.1.26t->1.1.26 (current CVS) patch for VBE related changes, and the only change that looked potentially problematic was on Dec 16th below. I have reverted this change, and had multiple users now confirm that the driver works for them for about a week. In addition the 1.1.27t driver fixes a few other problems for people, but it is too late for 1.1.27t to get into 4.3.0. Please apply this one patch to the 1.1.26 driver in CVS however to revert the change below and fix the Savage driver for Savage MX/IX. -- Mike A. Harris ftp://people.redhat.com/mharris OS Systems Engineer - XFree86 maintainer - Red Hat ---------- Forwarded message ---------- Date: Mon, 16 Dec 2002 01:38:54 -0800 (PST) From: Alan Hourihane <[EMAIL PROTECTED]> To: [EMAIL PROTECTED] List-Id: CVS commit messages from the XFree86 repository <cvs-commit.XFree86.Org> Subject: CVS Update: xc (branch: trunk) CVSROOT: /home/x-cvs Module name: xc Changes by: [EMAIL PROTECTED] 02/12/16 01:38:27 Log message: no need to softboot twice, reduces startup time Modified files: xc/programs/Xserver/hw/xfree86/drivers/savage/: savage_driver.c Revision Changes Path 1.30 +2 -3 xc/programs/Xserver/hw/xfree86/drivers/savage/savage_driver.c _______________________________________________ Cvs-commit mailing list [EMAIL PROTECTED] http://XFree86.Org/mailman/listinfo/cvs-commit
Patch by Mike A. Harris <[EMAIL PROTECTED]> Revert patch done in XFree86 CVS which I believe is causing the VESA VBE related problems on Savage MX. Here is the CVS commit message from the change: Date: Mon, 16 Dec 2002 01:38:54 -0800 (PST) From: Alan Hourihane <[EMAIL PROTECTED]> To: [EMAIL PROTECTED] List-Id: CVS commit messages from the XFree86 repository <cvs-commit.XFree86.Org> Subject: CVS Update: xc (branch: trunk) CVSROOT: /home/x-cvs Module name: xc Changes by: [EMAIL PROTECTED] 02/12/16 01:38:27 Log message: no need to softboot twice, reduces startup time Modified files: xc/programs/Xserver/hw/xfree86/drivers/savage/: savage_driver.c Revision Changes Path 1.30 +2 -3 xc/programs/Xserver/hw/xfree86/drivers/savage/savage_driver.c --- xc/programs/Xserver/hw/xfree86/drivers/savage-1.1.26t/savage_driver.c 2002-10-07 15:17:33.000000000 -0400 +++ xc/programs/Xserver/hw/xfree86/drivers/savage/savage_driver.c 2003-02-07 06:21:08.000000000 -0500 @@ -954,6 +945,7 @@ if (xf86LoadSubModule(pScrn, "vbe")) { xf86LoaderReqSymLists(vbeSymbols, NULL); + psav->pVbe = VBEInit(psav->pInt10, pEnt->index); } @@ -2228,7 +2221,7 @@ psav = SAVPTR(pScrn); pEnt = xf86GetEntityInfo(pScrn->entityList[0]); - psav->pVbe = VBEInit(psav->pInt10, pEnt->index); + psav->pVbe = VBEInit(NULL, pEnt->index); SavageEnableMMIO(pScrn);