Xen devel - Jan 2006 - crash on VM creation with changeset 8503 (3eeabf448f91)

I'm running stock Xen:

$ hg tip
changeset:   8503:3eeabf448f91
tag:         tip
user:        kaf24@firebug.cl.cam.ac.uk
date:        Fri Jan  6 16:47:25 2006 +0100
summary:     Clean up domain shutdown and modification of vcpu

The crash happens when I try to create a VM. Curiously, the crash
happens ONLY when running with the flags nosmp and noht. With SMP
enabled, things work just fine.

The dump on console when the crash happens:

(XEN) CPU:    0
(XEN) EIP:    e008:[<ff10d8d7>] sedf_wake+0x2d7/0x5b0
(XEN) EFLAGS: 00010002   CONTEXT: hypervisor
(XEN) eax: 00000000   ebx: ff1b3e80   ecx: 00000011   edx: 00000020
(XEN) esi: 00000001   edi: ff1b3080   ebp: ff1b3e80   esp: ff16ddb4
(XEN) cr0: 80050033   cr3: 39930000
(XEN) ds: e010   es: e010   fs: 0000   gs: 0033   ss: e010   cs: e008
(XEN) Xen stack trace from esp=ff16ddb4:
(XEN)    00000000 00000000 00000000 00000000 000020b6 80000006 ffbf9080 ffbf9080
(XEN)    05f5e100 00000000 eea7ce90 00000020 0000bc00 ff1b3080 00000002 ff1b3080
(XEN)    00000292 ff16dfb4 ffbe7080 ff10efec ff1b3080 00000555 00000000 ff16de84
(XEN)    0000e4e1 00000004 00000002 00000000 c4b40000 ff1b3080 00000001 ff104a58
(XEN)    ff1b3080 ff16dfb4 ffbe2280 ff103a7a ffbe2280 b6e78c70 00000088 00000000
(XEN)    ffbe7e80 00000000 00000004 ff10d204 c4b40000 00000004 002550ef 00000000
(XEN)    00000000 00000000 ffffffea 00000000 00004c4b ffbfac00 00000004 ece16a1f
(XEN)    00000020 ffbfac08 00000000 22ff0000 00000000 c4b40000 ffbfac08 00000000
(XEN)    ff108fcf ffbfac08 00000004 00e4e1c0 ffbfac08 00000000 ffbe7080 ffbfac08
(XEN)    ffbfac00 00000000 ffbfac10 ffbfac18 ffbe7080 0007a120 00000000 00000092
(XEN)    00000000 0007a120 00000000 00000292 0000000b 03000000 b7880001 4fb2e42c
(XEN)    00000000 0810b230 0000001f 00000000 a5dba1ee a5dba1ee b7bce200 4fb2e42c
(XEN)    b7bce200 b788dd74 b6e78cc8 4faaecc2 b788dd74 b7bce200 a5dba1ee 4fb2e42c
(XEN)    b7bce200 b7bce200 b6e78cf8 4fa919b7 b788dd74 b7bce200 4fab1c33 4fb2e42c
(XEN)    b7f69f98 b788a8ac 4faf5700 b7c32220 b7fb3120 b788a84c 00000000 00000000
(XEN)    eea22c6a 00000020 ffbe7080 ffbe7080 4fb2e42c b78898c4 cc954000 ff137423
(XEN)    b6e78c70 b788fa44 4fa8e5de 4fb2e42c b78898c4 cc954000 00000007 00070000
(XEN)    c02dbfe6 00000061 00000246 cc955d38 00000069 0000007b 0000007b 00000000
(XEN)    00000033 00000000 ffbe7080
(XEN) Xen call trace:
(XEN)    [<ff10d8d7>] sedf_wake+0x2d7/0x5b0
(XEN)    [<ff10efec>] vcpu_wake+0x8c/0xd0
(XEN)    [<ff104a58>] domain_unpause_by_systemcontroller+0x28/0x40
(XEN)    [<ff103a7a>] do_dom0_op+0x8aa/0xc20
(XEN)    [<ff10d204>] sedf_do_schedule+0x924/0xc60
(XEN)    [<ff108fcf>] __divdi3+0x4f/0xa0
(XEN)    [<ff137423>] hypercall+0x53/0x57
(XEN)
(XEN) Faulting linear address: 00000018
(XEN) Pagetable walk from 00000018:
(XEN)   L2 = 00000000 55555555
(XEN) ************************************
(XEN) CPU0 FATAL TRAP 14 (page fault), ERROR_CODE 0000, IN INTERRUPT CONTEXT.
(XEN) System shutting down -- need manual reset.
(XEN) ************************************
--
Web/Blog/Gallery: http://floatingsun.net

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel

On 7 Jan 2006, at 01:06, Diwaker Gupta wrote:
> $ hg tip
> changeset:   8503:3eeabf448f91
> tag:         tip
> user:        kaf24@firebug.cl.cam.ac.uk
> date:        Fri Jan  6 16:47:25 2006 +0100
> summary:     Clean up domain shutdown and modification of vcpu
>
> The crash happens when I try to create a VM. Curiously, the crash
> happens ONLY when running with the flags nosmp and noht. With SMP
> enabled, things work just fine.
When did this start happening? Will a tip changeset from Thursday cause 
the problem?

  -- Keir


_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel

> >> When did this start happening? Will a tip changeset from Thursday
> >> cause
> >> the problem?
> >
> > I'm not sure. I pulled in changes after almost a week, so its
> > difficult to isolate the problematic changeset.
>
> I couldn't repro the bug myself, but I was screwing around with the
> scheduler interfaces yesterday so it's unlikely that your problem is a
> coincidence.
The affinity stuff?
> I added some more fixes today, so perhaps the problem will have gone
> away now.
I'll test and let you know.

Diwaker
--
Web/Blog/Gallery: http://floatingsun.net

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xensource.com
http://lists.xensource.com/xen-devel