Re: xsrc/48344: X sesrver crashes with "lockup" on certain i810e-based systems

["John D. Baker" <jdbaker%mylinuxisp.com@localhost>]

The following reply was made to PR xsrc/48344; it has been noted by GNATS.

From: "John D. Baker" <jdbaker%mylinuxisp.com@localhost>
To: gnats-bugs%NetBSD.org@localhost
Cc: 
Subject: Re: xsrc/48344: X sesrver crashes with "lockup" on certain i810e-based
 systems
Date: Sun, 20 Jul 2014 23:27:45 -0500 (CDT)

 I've found the change that causes the X-server to fail on Dell Optiplex
 GX-110 systems (82810E graphics).
 
 5.99.44 (25-Jan 2011 10:52 UTC) works  (agp_i810.c r1.69)
 5.99.44 (25-Jan 2011 10:53 UTC) fails  (agp_i810.c r1.70)
 
 I also have an HP Pavilion something-or-other ("LOMITA") that is also
 built around the 82810E and related chips.  It does NOT suffer from the
 problem brought about by the above change.  Hardware-wise, the HP is
 practically identical to the DELLs (output of 'pcictl pci0 list').
 
 Perhaps their BIOSes differ in how completely they intialize the hardware
 and the change exposes this in the DELLs?
 
 I built a kernel with "options AGP_DEBUG" and extracted the results from
 "/var/log/messages" for the version that works and the version that fails.
 The two logs are essentially identical until the very end of each.  The
 version that succeeds ends with:
 
 $ tail i810-success.log                                    
 Jul 20 20:11:28 plexor /netbsd: agp: searching for memory block 2
 Jul 20 20:11:28 plexor /netbsd: agp: considering memory block 0
 Jul 20 20:11:29 plexor /netbsd: agp: considering memory block 1
 Jul 20 20:11:29 plexor /netbsd: agp: considering memory block 2
 Jul 20 20:11:29 plexor /netbsd: agp: searching for memory block 3
 Jul 20 20:11:29 plexor /netbsd: agp: considering memory block 0
 Jul 20 20:11:29 plexor /netbsd: agp: considering memory block 1
 Jul 20 20:11:29 plexor /netbsd: agp: considering memory block 2
 Jul 20 20:11:29 plexor /netbsd: agp: considering memory block 3
 Jul 20 20:11:30 plexor sysop: X11 finished
 
 
 The version that fails ends with:
 
 $ tail i810-fail.log     
 Jul 20 22:25:42 plexor /netbsd: agp: searching for memory block 3
 Jul 20 22:25:42 plexor /netbsd: agp: considering memory block 0
 Jul 20 22:25:42 plexor /netbsd: agp: considering memory block 1
 Jul 20 22:25:42 plexor /netbsd: agp: considering memory block 2
 Jul 20 22:25:42 plexor /netbsd: agp: considering memory block 3
 Jul 20 22:26:04 plexor /netbsd: agpclose: mem 0 is bound
 Jul 20 22:26:04 plexor /netbsd: agpclose: mem 1 is bound
 Jul 20 22:26:04 plexor /netbsd: agpclose: mem 2 is bound
 Jul 20 22:26:04 plexor /netbsd: agpclose: mem 3 is bound
 Jul 20 22:26:04 plexor sysop: X11 finished
 
 
 If the full logs of each are more helpful, I can send them, but they
 are about 56KB each.
 
 I am in the process of updating my test source tree to netbsd-6 after
 which I will revert to r1.69 of "sys/dev/pci/agp_i810.c" to see if it
 builds and runs (the diffs versus r1.73 in netbsd-6 lead me to believe
 it will).
 
 
 (The HP is having issues starting up, but once I've got it sorted out
 I hope to collect similar data from it.)
 
 -- 
 |/"\ John D. Baker, KN5UKS               NetBSD     Darwin/MacOS X
 |\ / jdbaker[snail]mylinuxisp[flyspeck]com    OpenBSD            FreeBSD
 | X  No HTML/proprietary data in email.   BSD just sits there and works!
 |/ \ GPGkeyID:  D703 4A7E 479F 63F8 D3F4  BD99 9572 8F23 E4AD 1645