Re: [Xen-devel] HVM guests hanging in hpet_rtc_interrupt

["Jan Beulich" <JBeulich@xxxxxxxx>]

>>> On 31.01.13 at 19:46, Olaf Hering <olaf@xxxxxxxxx> wrote:
> On Thu, Jan 31, Olaf Hering wrote:
> 
>> On Thu, Jan 31, Ian Campbell wrote:
>> 
>> > On Thu, 2013-01-31 at 17:07 +0000, Olaf Hering wrote:
>> > > Current xen-unstable has appearently issues with HVM guest.
>> > 
>> > staging or tested? which changeset?
>> 
>> staging, 26493:6727070b4129
>> Its the first time since many weeks that I run xen-unstable again, so I
>> cant tell when it started.
> 
> It works with xend for some reason, pvops dom0+domU and changeset
> 26502:d1bf3b21f783. What I see in guest dmesg is shown below. With xl the 
> irq#8
> message does not appear.

Odd - I didn't observe any problems specifically with SLE guests
(which is what I used for testing). In fact I also, subsequent to
the series that's already in staging, figured the problem with the
remaining changes left from the submission about half a year
ago, and was hoping to post that one soon too (but of course
won't until what's in is really sane).

Looking over the changes again (and without having looked at
the Linux side yet again), I think the most likely candidates for
reverting are

26457:aa82638d58b0 (x86/HVM: consolidate toggling of RTC IRQ)
26461:78e91e9e4d61 (x86/HVM: generalize IRQ raising on RTC_REG_B writes)

Especially the latter I can't really see to be responsible for
apparently superfluous IRQ8 instances, as my understanding
is that this would imply repeated REG_B writes (which I specifically
didn't observe with extra debugging code added).

Considering that I didn't see any problems with the patches
myself - any chance you could try the individual reverts above
as a first step? (Obviously reverting 26461 should be strait
forward, but reverting 26457 alone might not be as simple.)

Thanks, Jan

> ...
> [    0.832275] udev: starting version 147
> [    1.312097] tsc: Refined TSC clocksource calibration: 2926.333 MHz
> [   14.378118] udev: starting version 147
> [   14.469299] input: Xen Virtual Keyboard as /devices/virtual/input/input5
> [   14.469532] input: Xen Virtual Pointer as /devices/virtual/input/input6
> [   14.480053] microcode: CPU0 sig=0x206c2, pf=0x1, revision=0xc
> [   14.480626] piix4_smbus 0000:00:01.3: SMBus base address uninitialized - 
> upgrade BIOS or use force_addr=0xaddr
> [   14.546814] microcode: CPU1 sig=0x206c2, pf=0x1, revision=0xc
> [   14.601594] NET: Registered protocol family 10
> [   14.725108] microcode: Microcode Update Driver: v2.00 
> <tigran@xxxxxxxxxxxxxxxxxxxx>, Peter Oruba
> [   15.084864] Adding 331772k swap on /dev/xvda2.  Priority:-1 extents:1 
> across:331772k SS
> [   18.810217] irq 8: nobody cared (try booting with the "irqpoll" option)
> [   18.812058] Pid: 0, comm: swapper/0 Not tainted 
> 3.7.5-9.home_olh_kernel_sles11sp1.1-kernel-linux-3_7 #1
> [   18.812058] Call Trace:
> [   18.812058]  <IRQ>  [<ffffffff8109ce78>] __report_bad_irq+0x38/0xe0
> [   18.812058]  [<ffffffff8109d0ef>] note_interrupt+0x1cf/0x210
> [   18.812058]  [<ffffffff8109aa77>] handle_irq_event_percpu+0x97/0x170
> [   18.812058]  [<ffffffff8109ab8c>] handle_irq_event+0x3c/0x60
> [   18.812058]  [<ffffffff8109d79d>] handle_edge_irq+0x6d/0x120
> [   18.812058]  [<ffffffff8142b474>] __xen_evtchn_do_upcall+0x1a4/0x280
> [   18.812058]  [<ffffffff8142ca6a>] xen_evtchn_do_upcall+0x2a/0x40
> [   18.812058]  [<ffffffff81660aad>] xen_hvm_callback_vector+0x6d/0x80
> [   18.812058]  <EOI>  [<ffffffff810668f3>] ? hrtimer_start+0x13/0x20
> [   18.812058]  [<ffffffff810360d6>] ? native_safe_halt+0x6/0x10
> [   18.812058]  [<ffffffff81569e8f>] ? cpuidle_idle_call+0x1f/0xc0
> [   18.812058]  [<ffffffff81014735>] default_idle+0x45/0x50
> [   18.812058]  [<ffffffff81014a68>] cpu_idle+0x88/0xd0
> [   18.812058]  [<ffffffff81642428>] rest_init+0x68/0x70
> [   18.812058]  [<ffffffff81c7bc8f>] start_kernel+0x2f3/0x3b4
> [   18.812058]  [<ffffffff81c7b795>] ? repair_env_string+0x58/0x58
> [   18.812058]  [<ffffffff81cb8c36>] ? memblock_reserve+0x2e/0x4e
> [   18.812058]  [<ffffffff81c7b2d5>] x86_64_start_reservations+0xa5/0xc2
> [   18.812058]  [<ffffffff81c7b401>] x86_64_start_kernel+0x10f/0x12a
> [   18.812058]  [<ffffffff81c7b120>] ? early_idt_handlers+0x120/0x120
> [   18.812058] handlers:
> [   18.812058] [<ffffffff81550740>] cmos_interrupt
> [   18.812058] Disabling IRQ #8
> [   19.420384] loop: module loaded
> [   19.548950] kjournald starting.  Commit interval 5 seconds
> [   19.548970] EXT3-fs (xvda1): mounted filesystem with ordered data mode
> [   59.305223] PM: freeze of devices complete after 2.291 msecs
> [   59.305227] suspending xenstore...
> [   59.305290] PM: late freeze of devices complete after 0.058 msecs
> [   59.309521] PM: noirq freeze of devices complete after 4.227 msecs
> [   59.312080] Xen HVM callback vector for event delivery is enabled
> [   59.312080] Xen Platform PCI: I/O protocol version 1
> [   59.312080] Grant tables using version 1 layout.
> [   59.312080] xen: --> irq=8, pirq=16
> [   59.312080] xen: --> irq=12, pirq=17
> [   59.312080] xen: --> irq=1, pirq=18
> [   59.312080] xen: --> irq=6, pirq=19
> [   59.312080] xen: --> irq=4, pirq=20
> [   59.312080] xen: --> irq=7, pirq=21
> [   59.312080] xen: --> irq=28, pirq=22
> [   59.313545] ata_piix 0000:00:01.1: restoring config space at offset 0x4 
> (was 0x2800001, writing 0x2800005)
> [   59.313545] PM: noirq restore of devices complete after 7.838 msecs
> [   59.313545] PM: early restore of devices complete after 0.046 msecs
> [   59.319592] pci 0000:00:00.0: calling quirk_passive_release+0x0/0xa0
> [   59.479464] PM: restore of devices complete after 159.921 msecs
> [   59.488309] Setting capacity to 2097152
> [   59.498494] Setting capacity to 2097152
> 
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@xxxxxxxxxxxxx 
> http://lists.xen.org/xen-devel 



_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
http://lists.xen.org/xen-devel