[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Xen-devel] crash in csched_load_balance after xl vcpu-pin

Il Mer 11 Apr 2018, 22:48 Olaf Hering <olaf@xxxxxxxxx> ha scritto:
On Wed, Apr 11, Dario Faggioli wrote:

> It will crash, again, possibly with the same stack trace, but I think
> it's worth a try.


(XEN) Xen BUG at sched_credit.c:876
(XEN) ----[ Xen-4.11.20180410T125709.50f8ba84a5-7.bug1087289_411  x86_64  debug=y   Not tainted ]----
(XEN) CPU:    108
(XEN) RIP:    e008:[<ffff82d080229ab4>] sched_credit.c#csched_vcpu_migrate+0x27/0x54
(XEN) RFLAGS: 0000000000010006   CONTEXT: hypervisor
(XEN) Xen call trace:
(XEN)    [<ffff82d080229ab4>] sched_credit.c#csched_vcpu_migrate+0x27/0x54
(XEN)    [<ffff82d080236348>] schedule.c#vcpu_move_locked+0xbb/0xc2
(XEN)    [<ffff82d08023764c>] schedule.c#vcpu_migrate+0x226/0x25b
(XEN)    [<ffff82d080239367>] context_saved+0x95/0x9c
(XEN)    [<ffff82d08027797d>] context_switch+0xe66/0xeb0
(XEN)    [<ffff82d080236943>] schedule.c#schedule+0x5f4/0x627
(XEN)    [<ffff82d080239f15>] softirq.c#__do_softirq+0x85/0x90
(XEN)    [<ffff82d080239f6a>] do_softirq+0x13/0x15
(XEN)    [<ffff82d08031f5db>] vmx_asm_do_vmentry+0x2b/0x30

So, really *exactly* the same. Ok, thanks.

I think that from "CONTEXT: hypervisor", we can tell that the current vcpu is the idle one, and I'm starting to wonder whether the lazy context switch logic may play a role in all this.

But, for now, it's just a gut feeling. I'll investigate tomorrow.

Another thing we could do, would be to try George's migratiin refactoring series. I haven't reviewed it in details yet, but it seemed reasonable at a first glance.

Not that that could be the solution (backportabiliry, ecc), but, if it works, it might give us ideas on where to look, and on how to produce a stepgap patch, "just" solving the issue.

Thanks and regards,
Xen-devel mailing list



Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.