[Suspend2-devel] Re: Is this a known issue with SMP

[Sunil Kumar]

Its seems to be mysteriously fixed with the latest kernel 2.6.17.3 and
suspend2 2.2.7 combo. I have no idea, I can't reproduce it anymore.

On 7/3/06, Sunil Kumar < devsku at gmail.com> wrote:
>
> I see this on resume in the /var/log/messages and my system freezes if put
> into high cpu load after resume (I didn't notice this problem till now
> because I didn't do much high cpu intensive work after resume):
>
> Jul  3 08:36:05 localhost Thawing cpus ...
> Jul  3 08:36:06 localhost SMP alternatives: switching to SMP code
> Jul  3 08:36:06 localhost Booting processor 1/1 eip 3000
> Jul  3 08:36:06 localhost Initializing CPU#1
> Jul  3 08:36:06 localhost Calibrating delay using timer specific routine..
> 4812.35 BogoMIPS (lpj=24061761)
> Jul  3 08:36:06 localhost CPU: Vendor unknown, using generic init.
> Jul  3 08:36:06 localhost CPU: Your system may be unstable.
> Jul  3 08:36:06 localhost CPU: After generic identify, caps: 178bfbff
> e3d3fbff 00000000 00000000 00000001 00000000 00000003
> Jul  3 08:36:06 localhost CPU: After all inits, caps: 178bfbff e3d3fbff
> 00000000 00000000 00000001 00000000 00000003
> Jul  3 08:36:06 localhost CPU1: AuthenticAMD AMD Athlon(tm) 64 X2 Dual
> Core Processor 3800+ stepping 02
> Jul  3 08:36:06 localhost CPU1 is up
> Jul  3 08:36:06 localhost fbsplash: switched splash state to 'off' on
> console 15
> Jul  3 08:36:06 localhost login(pam_unix)[11750]: session closed for user
> root
> Jul  3 08:36:06 localhost forcedeth.c: Reverse Engineered nForce ethernet
> driver. Version 0.54.
> Jul  3 08:36:06 localhost ACPI: PCI Interrupt 0000:00:0a.0[A] -> Link
> [APCH] -> GSI 23 (level, low) -> IRQ 217
> Jul  3 08:36:06 localhost PCI: Setting latency timer of device 0000:00:
> 0a.0 to 64
> Jul  3 08:36:06 localhost forcedeth: using HIGHDMA
> Jul  3 08:36:07 localhost BUG: warning at
> drivers/char/vt.c:2940/set_palette()
> Jul  3 08:36:07 localhost <c02b8f4f> set_palette+0x5f/0x70  <c02b4f09>
> redraw_screen+0xc9/0x200
> Jul  3 08:36:07 localhost <c0109362> convert_fxsr_to_user+0x122/0x190
> <c02affce> complete_change_console+0x2e/0xf0
> Jul  3 08:36:07 localhost <c02af812> vt_ioctl+0x1612/0x1c80  <c0109579>
> save_i387_fxsave+0xa9/0xd0
> Jul  3 08:36:07 localhost <c010295f> setup_sigcontext+0xdf/0x130
> <c0102fce> handle_signal+0x11e/0x180
> Jul  3 08:36:07 localhost <c01030d2> do_signal+0xa2/0x160  <c02ae200>
> vt_ioctl+0x0/0x1c80
> Jul  3 08:36:07 localhost <c02a938d> tty_ioctl+0x1dd/0x5d0  <c018e6b0>
> do_ioctl+0x70/0xb0
> Jul  3 08:36:07 localhost <c018e8a5> vfs_ioctl+0x65/0x200  <c018ea85>
> sys_ioctl+0x45/0x70
> Jul  3 08:36:07 localhost <c010326f> sysenter_past_esp+0x54/0x75
> Jul  3 08:36:07 localhost BUG: warning at
> drivers/char/vt.c:445/complement_pos()
>
> There are two things: One about CPU identification and the other with BUG
> at vt.c. It seems to me that  "CPU: Vendor unknown, using generic init."
> means that after this reinit of CPU1, CPU0 and CPU1 are not inited in the
> same way because during boot system intialized both CPUs with something
> which was not "generic".
>
> After resume when I start something which is CPU intensive, one CPU
> becomes 100% (red in gkrellm, meaning kernel bound) and other completely
> free with occasional usage shown. My keyboard locks up. Mouse responds but
> no new programs can be started. KDE logout menu works but 'restart' doesn't
> do anything for half an hour. And I just have to do the sysrq thing to
> reboot.
>
> I am using 2.2.6 with 2.6.17.1 gentoo sources kernel patched with
> suspend2.
>
> Is this a known SMP issue? What about that BUG warning at vt.c?
>
> Thanks,
> Sunil
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.tuxonice.net/pipermail/tuxonice-devel/attachments/20060703/855baac6/attachment.html>