On 08/02/2010 14:07, Robert Milkowski wrote:
On 05/02/2010 15:25, Jürgen Keil wrote:
Robert Milkowski wrote:
On 03/02/2010 12:25, Jürgen Keil wrote:
Robert Milkowski wrote:
Toshiba R600, up-to snv_129 X is working fine.
snv_131 and the moment Xorg starts the notebook does hard-reset
(quick
power-off) - no crash dump, no nothing, When I booted under kernel
debugger the box just hangs and I can't go into the debugger.
I tried to boot into single user mode, console is fine, then manually
start Xorg binary and laptop does a hard-reset almost right-away.
What happens when you boot into single user mode
under kernel debugger, and open the /dev/agpgart device,
e.g. like this
true< /dev/agpgart
Does that panic the kernel?
yes, it does in hat_devload()
?
That looks a lot like 6914386
"X freeze (and reboot) a build 130 system"
http://bugs.opensolaris.org/bugdatabase/view_bug.do?bug_id=6914386
Which got a "commit to fix" for b134.
I've just installed b132 on a Tecra A10, like this:
- boot and install b132 using the vesa install option
unfortunately there was a new problem with vesa,
compiz and a white screen (defect 14387,
http://defect.opensolaris.org/bz/show_bug.cgi?id=14387 )
but with metacity instead of compiz the gui installer
worked ok
- boot b132 from hdd into single user mode
- disable the broken chipset flush code in the agptarget driver,
using the following patch:
echo "intel_chipset_flush_setup?w c3" | mdb -w /kernel/drv/agptarget
echo "intel_chipset_flush_setup?w c3" | mdb -w
/kernel/drv/amd64/agptarget
halt
(My Tecra A10 seems to work just fine with the chipset flush code
disabled; YMMV)
- reboot into b132
- remove the /etc/X11/xorg.conf file
(it forces the use of the vesa driver, after the vesa install)
and logout from Xorg
- Xorg should no be using the intel video driver
A similar patch should work when you upgrade to b132 using
the dev pkg. After the upgrade, mount the new b132 BE and
apply the above patch to the b132 kernel/drv/agptarget and
kernel/drv/amd64/agptarget binary.
I upgraded to snv_132 then patched a copy of agptarget binaries.
I confirmed first that snv_132 does crashes with original binariesm,
then I put in place a patched binaries as you described above, updated
boot archives, rebooted and got a working X11! I double checked if
reading from /dev/agpgart doeasn't crash OS - and it doesn't.
Thank you.
I only hope it will be fixed for 2010.03
btw: looks like it is going to be fixed in snv_134
--
Robert Milkowski
http://milek.blogspot.com
_______________________________________________
opensolaris-discuss mailing list
[email protected]
