[Openchrome-users] [openChrome] #287: Problem with CX700 + VBEModes=TRUE on Ubuntu Jaunty

[OpenChrome Trac]

#287: Problem with CX700 + VBEModes=TRUE on Ubuntu Jaunty
--------------------+-------------------------------------------------------
Reporter:  klaussm  |       Owner:  somebody
    Type:  defect   |      Status:  new     
Priority:  major    |   Component:  vbe     
 Version:  trunk    |    Keywords:          
Blocking:           |   Blockedby:          
--------------------+-------------------------------------------------------
 Hi,

 I'm having a problem using the openchrome driver after I upgraded to
 Ubuntu Jaunty, the upcomming 9.04 release. It worked before the upgrade
 where I was running Ubuntu Gutsy (7.10). I have tried both with the
 openchrome driver included with Ubuntu, and with the trunk from SVN.

 I want to use the DVI output, and therefore I have set the VBEModes option
 to TRUE. However this results in the X process running in an infinite
 loop, with the last line in the log file being:

 {{{
 [  288.638223] (II) CHROME(0): Attempting to use 60Hz refresh for mode
 "1600x1200" (124)
 }}}

 I have discovered the the new Ubuntu X server uses an X86 emulator instead
 of the vm86 call, so I've tried switching libint10 back to using vm86 to
 see if that made a difference.

 With a libint10 that uses vm86, I instead of an infinite loop, get a
 SIGTRAP. The last line of the log file is the same "Attempting to use"
 line as before. The backtrace when the SIGTRAP is received is:

 {{{
 #0  0xb7636b68 in xf86ExecX86int10 (pInt=0x8213d28) at
 ../../../../hw/xfree86/int10/../os-
 support/linux/int10/vm86/linux_vm86.c:277
 #1  0xb762d2cb in VBEGetPixelClock (pVbe=0x8213cd0, mode=292,
 clock=162000000) at ../../../../hw/xfree86/vbe/vbe.c:1036
 #2  0xb762f0db in VBESetModeParameters (pScrn=0x820bee0, pVbe=0x8213cd0)
 at ../../../../hw/xfree86/vbe/vbeModes.c:416
 #3  0xb7861a2c in ViaVbeModePreInit (pScrn=0x820bee0) at via_vbe.c:390
 #4  0xb7854333 in VIAPreInit (pScrn=0x820bee0, flags=<value optimized
 out>) at via_driver.c:1689
 #5  0x080b0733 in InitOutput (pScreenInfo=0x81f6c20, argc=1,
 argv=0xbf9a2274) at ../../../../hw/xfree86/common/xf86Init.c:1008
 #6  0x08072141 in main (argc=1, argv=0xbf9a2274, envp=0xbf9a227c) at
 ../../dix/main.c:315
 }}}

 And the contents of the pInt structure when entering the xf86ExecX86int10
 function is:

 {{{
 *pInt = {entityIndex = 0, scrnIndex = 0, cpuRegs = 0xb763a1a0, BIOSseg =
 49152, inb40time = 43499, BIOSScratch = 0x8213ca8 "\003P", Flags = 3,
 private = 0x8213268,
   mem = 0xb763a154, num = 16, ax = 20235, bx = 0, cx = 162000000, dx =
 292, si = 0, di = 8192, es = 0, bp = 0, flags = 12802, stackseg = 4096,
 dev = 0x8206bf8, ioBase = 0}
 }}}
 I'm currently running kernel version 2.6.29-rc8, but the problem was also
 showing on 2.6.27. The 2.6.27 kernel is the exact same kernel it worked on
 when I was running Gutsy, so it doesn't seem to be kernel releated.

-- 
Ticket URL: <http://www.openchrome.org/trac/ticket/287>
openChrome <http://www.openchrome.org/>
The openChrome project