This looks like that the BIOS did not load the same microcode patch in any individual core. So some cores have SVM disabled which triggers the #GP. I guess this is a BIOS bug.
On Fri, Jan 02, 2009 at 11:45:12PM -0500, [email protected] wrote: > This is resolved, but since other people might have the problem I'll still > post what I > figured out. It seems like a strange corner case of memory allocation. > > I've upgraded to kernel 2.6.28 with kvm-82, but the problem kept persisting. > I've decided to > remove the evga 9800gt and suddenly it kvm-amd would load without any issue. > > I've decided to check the Asus website and sure enough there was a bios > update, I've > installed it then installed the evga card back it and since then everything > works (well > except rebooting but anyways). > > The bios update is 0603, it lists fix ups for some memory modules as an > improvement, so that > might be it. > > On Wed, 31 Dec 2008, Avi Kivity wrote: > > >Date: Wed, 31 Dec 2008 12:17:15 +0200 > >From: Avi Kivity <[email protected]> > >To: [email protected] > >Cc: [email protected], Joerg Roedel <[email protected]> > >Subject: Re: GPF on modprobe kvm-amd > >(adding cc) > > > >Joerg, all I can make of it is that svm is enabled on one cpu but not on the > >other. Can > >you help here? > > > >[email protected] wrote: > >> This is with kvm-81 > >> I'm getting a kernel panic when I modprobe kvm-amd > >> > >> It used to work until I had to use the CMOS jumper to boot. > >> The motherboard is an Asus M3N78 PRO > >> The CPU is an AMD Phenom 9950 Black Edition > >> > >> Here are the steps I've already tried: > >> - Checking that virtualization is enabled in the BIOS > >> - Updating the BIOS > >> - Restoring the defaults of the BIOS > >> - Downgrading the BIOS > >> - Trying various versions of kvm > >> - Trying various linux kernel versions > >> - Trying various vcore settings > >> - Enabling/Disabling Cool & Quiet and AMD C1E > >> - Tried a known good kernel and modules combo > >> - Interrupting the boot process to only have udev as a process, and > >> killing it. > >> - Not having any other modules loaded > >> > >> What I plan to try: > >> - Updating to 2.6.28 and kvm-82 > >> - Tranfering the nvram from an identical motherboard I own to the > >> problematic one, using /dev/nvram > >> > >> Here are the traces > >> [ 36.870419] general protection fault: 0000 [1] PREEMPT SMP > >> [ 36.870579] CPU 0 > >> [ 36.870667] Modules linked in: kvm_amd(+) kvm nvidia(P) snd_hda_intel > >> [ 36.870893] Pid: 0, comm: swapper Tainted: P 2.6.27-gentoo-r7 > >> #4 > >> [ 36.870955] RIP: 0010:[<ffffffffa07e7f47>] [<ffffffffa07e7f47>] > >> svm_hardware_enable+0x87/0xf0 [kvm_amd] > >> [ 36.871070] RSP: 0018:ffffffff808d9f28 EFLAGS: 00010006 > >> [ 36.871130] RAX: 0000000000001d01 RBX: 0000000000000040 RCX: > >> 00000000c0000080 > >> [ 36.871193] RDX: 0000000000000000 RSI: ffff88021c41c9c0 RDI: > >> 0000000000000000 > >> [ 36.871255] RBP: ffffffff808d9f48 R08: 0000000000000000 R09: > >> 0000000000000000 > >> [ 36.871317] R10: 0000000000000000 R11: ffffffff80869e88 R12: > >> ffff88021e843cc0 > >> [ 36.871379] R13: ffff88021e843ce8 R14: 0000000000000000 R15: > >> 0000000000000000 > >> [ 36.871413] FS: 00007fbabbbec6f0(0000) GS:ffffffff80863340(0000) > >> knlGS:0000000000000000 > >> [ 36.871413] CS: 0010 DS: 0018 ES: 0018 CR0: 000000008005003b > >> [ 36.871413] CR2: 00007fbabbc22000 CR3: 0000000000201000 CR4: > >> 00000000000006e0 > >> [ 36.871413] DR0: 0000000000000000 DR1: 0000000000000000 DR2: > >> 0000000000000000 > >> [ 36.871413] DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: > >> 0000000000000400 > >> [ 36.871413] Process swapper (pid: 0, threadinfo ffffffff80868000, task > >> ffffffff80808340) > >> [ 36.871413] Stack: 880028036000007f ffffffff8025ffff ffff880028038d20 > >> 0000000000000011 > >> [ 36.871413] ffffffff808d9f58 ffffffffa07c002e ffffffff808d9f68 > >> ffffffffa07bceaf > >> [ 36.871413] ffffffff808d9f98 ffffffff8026515b 0000000000000011 > >> ffffffff808ae500 > >> [ 36.871413] Call Trace: > >> [ 36.871413] <IRQ> [<ffffffff8025ffff>] ? > >> tick_broadcast_oneshot_control+0xff/0x130 > >> [ 36.871413] [<ffffffffa07c002e>] kvm_arch_hardware_enable+0xe/0x10 > >> [kvm] > >> [ 36.871413] [<ffffffffa07bceaf>] hardware_enable+0x2f/0x40 [kvm] > >> [ 36.871413] [<ffffffff8026515b>] > >> generic_smp_call_function_interrupt+0x6b/0x160 > >> [ 36.871413] [<ffffffff8021e3f9>] smp_call_function_interrupt+0x19/0x30 > >> [ 36.871413] [<ffffffff8020d0d6>] call_function_interrupt+0x66/0x70 > >> [ 36.871413] <EOI> [<ffffffff80214232>] ? default_idle+0x42/0x50 > >> [ 36.871413] [<ffffffff80214348>] ? c1e_idle+0x38/0x100 > >> [ 36.871413] [<ffffffff8025ae51>] ? > >> atomic_notifier_call_chain+0x11/0x20 > >> [ 36.871413] [<ffffffff8020b496>] ? cpu_idle+0x56/0xc0 > >> [ 36.871413] [<ffffffff80642856>] ? rest_init+0x86/0x90 > >> [ 36.871413] > >> [ 36.871413] > >> [ 36.871413] Code: 46 10 0f 01 45 e0 48 8b 45 e2 b9 80 00 00 c0 48 83 c0 > >> 40 48 89 46 18 0f 32 48 c1 e2 20 89 c0 48 09 c2 89 d0 48 c1 ea 20 80 cc 10 > >> <0f> 30 48 ba 00 00 00 00 00 1e 00 00 48 03 56 20 48 b8 b7 6d db > >> [ 36.871413] RIP [<ffffffffa07e7f47>] svm_hardware_enable+0x87/0xf0 > >> [kvm_amd] > >> [ 36.871413] RSP <ffffffff808d9f28> > >> [ 36.871413] ---[ end trace 91fceceaf959d326 ]--- > >> [ 36.871413] Kernel panic - not syncing: Aiee, killing interrupt > >> handler! > >> [ 36.871413] ------------[ cut here ]------------ > >> [ 36.871413] WARNING: at kernel/smp.c:332 > >> smp_call_function_mask+0x21c/0x230() > >> [ 36.871413] Modules linked in: kvm_amd(+) kvm nvidia(P) snd_hda_intel > >> [ 36.871413] Pid: 0, comm: swapper Tainted: P D 2.6.27-gentoo-r7 > >> #4 > >> [ 36.871413] > >> [ 36.871413] Call Trace: > >> [ 36.871413] <IRQ> [<ffffffff8023fb6f>] warn_on_slowpath+0x5f/0x90 > >> [ 36.871413] [<ffffffff8023feae>] ? __call_console_drivers+0x6e/0x90 > >> [ 36.871413] [<ffffffff8025a5e4>] ? up+0x34/0x50 > >> [ 36.871413] [<ffffffff8024049c>] ? release_console_sem+0x1dc/0x1f0 > >> [ 36.871413] [<ffffffff80264f6c>] smp_call_function_mask+0x21c/0x230 > >> [ 36.871413] [<ffffffff8021e380>] ? stop_this_cpu+0x0/0x30 > >> [ 36.871413] [<ffffffff806578c6>] ? printk+0x67/0x69 > >> [ 36.871413] [<ffffffff8026a8e4>] ? crash_kexec+0x74/0x100 > >> [ 36.871413] [<ffffffff8023feae>] ? __call_console_drivers+0x6e/0x90 > >> [ 36.871413] [<ffffffff80264faf>] smp_call_function+0x2f/0x70 > >> [ 36.871413] [<ffffffff8021e363>] native_smp_send_stop+0x23/0x40 > >> [ 36.871413] [<ffffffff806577ae>] panic+0xbe/0x16f > >> [ 36.871413] [<ffffffff806578c6>] ? printk+0x67/0x69 > >> [ 36.871413] [<ffffffff806578c6>] ? printk+0x67/0x69 > >> [ 36.871413] [<ffffffff80243986>] do_exit+0x786/0x900 > >> [ 36.871413] [<ffffffff8020e3a5>] ? show_registers+0x95/0x2c0 > >> [ 36.871413] [<ffffffff8040ff7f>] ? do_unblank_screen+0xf/0x140 > >> [ 36.871413] [<ffffffff8020dc18>] oops_end+0x88/0x90 > >> [ 36.871413] [<ffffffff8020eade>] die+0x5e/0x90 > >> [ 36.871413] [<ffffffff8020f6d0>] do_general_protection+0x150/0x160 > >> [ 36.871413] [<ffffffff8065ab79>] error_exit+0x0/0x51 > >> [ 36.871413] [<ffffffffa07e7f47>] ? svm_hardware_enable+0x87/0xf0 > >> [kvm_amd] > >> [ 36.871413] [<ffffffffa07e7ed6>] ? svm_hardware_enable+0x16/0xf0 > >> [kvm_amd] > >> [ 36.871413] [<ffffffff8025ffff>] ? > >> tick_broadcast_oneshot_control+0xff/0x130 > >> [ 36.871413] [<ffffffffa07c002e>] kvm_arch_hardware_enable+0xe/0x10 > >> [kvm] > >> [ 36.871413] [<ffffffffa07bceaf>] hardware_enable+0x2f/0x40 [kvm] > >> [ 36.871413] [<ffffffff8026515b>] > >> generic_smp_call_function_interrupt+0x6b/0x160 > >> [ 36.871413] [<ffffffff8021e3f9>] smp_call_function_interrupt+0x19/0x30 > >> [ 36.871413] [<ffffffff8020d0d6>] call_function_interrupt+0x66/0x70 > >> [ 36.871413] <EOI> [<ffffffff80214232>] ? default_idle+0x42/0x50 > >> [ 36.871413] [<ffffffff80214348>] ? c1e_idle+0x38/0x100 > >> [ 36.871413] [<ffffffff8025ae51>] ? > >> atomic_notifier_call_chain+0x11/0x20 > >> [ 36.871413] [<ffffffff8020b496>] ? cpu_idle+0x56/0xc0 > >> [ 36.871413] [<ffffffff80642856>] ? rest_init+0x86/0x90 > >> [ 36.871413] > >> [ 36.871413] ---[ end trace 91fceceaf959d326 ]--- > >> > >> Here is the hexdump of /dev/nvram for the mobo that works: > >> diskgrinder ~ # hexdump /dev/nvram > >> 0000000 0000 ec00 0200 8007 c002 2fff f62f d7ef > >> 0000010 0000 0000 0000 0000 0000 0010 a400 ff95 > >> 0000020 9808 ffc0 ff20 fee7 ecf7 ff57 077d d1bf > >> 0000030 13f0 8a7b 2877 d9cd af11 cf5f 5d7f fffc > >> 0000040 f799 6b29 0bfc 0e8f e25e bc65 faf7 ff9f > >> 0000050 81b2 fcfe 40fd 0122 f488 dc01 3f05 d2ff > >> 0000060 ffff 7fff 25ff edde fffe f9ff 22ab 0083 > >> 0000070 2000 > >> 0000072 > >> > >> and here is the one that doesn't work: > >> gandalf ~ # hexdump /dev/nvram > >> 0000000 0000 f840 0200 800b c002 2fff 472f bfcf > >> 0000010 0000 0000 0000 0000 0000 0000 0000 0000 > >> 0000020 b905 ffc0 dd20 3be3 ffac fbff 07e6 d1bb > >> 0000030 13f0 88e9 08f7 d9ed 7f11 ef73 fd3b 4bb4 > >> 0000040 7f9f ea7f 0ffe 0e8e e77a bc35 ef7f 5eef > >> 0000050 03ba ff3e 60df 1026 f482 dc01 f705 ffcb > >> 0000060 df7d fdfc f7df cff2 ef5e fbfd 2097 00b6 > >> 0000070 2000 > >> 0000072 > >> > >> -- > >> To unsubscribe from this list: send the line "unsubscribe kvm" in > >> the body of a message to [email protected] > >> More majordomo info at http://vger.kernel.org/majordomo-info.html > > > > > > > -- | AMD Saxony Limited Liability Company & Co. KG Operating | Wilschdorfer Landstr. 101, 01109 Dresden, Germany System | Register Court Dresden: HRA 4896 Research | General Partner authorized to represent: Center | AMD Saxony LLC (Wilmington, Delaware, US) | General Manager of AMD Saxony LLC: Dr. Hans-R. Deppe, Thomas McCoy -- To unsubscribe from this list: send the line "unsubscribe kvm" in the body of a message to [email protected] More majordomo info at http://vger.kernel.org/majordomo-info.html
