Re: [KVM PATCH v9 0/5] irqfd fixes and enhancements

From: Gregory Haskins <ghaskins@novell.com>
To: Avi Kivity <avi@redhat.com>
Cc: kvm@vger.kernel.org, linux-kernel@vger.kernel.org,
	mst@redhat.com, davidel@xmailserver.org
Subject: Re: [KVM PATCH v9 0/5] irqfd fixes and enhancements
Date: Sun, 05 Jul 2009 17:21:17 -0400	[thread overview]
Message-ID: <4A51194D.8000403@novell.com> (raw)
In-Reply-To: <4A50723E.6030305@redhat.com>

[-- Attachment #1: Type: text/plain, Size: 3915 bytes --]

Avi Kivity wrote:
> On 07/02/2009 06:50 PM, Avi Kivity wrote:
>> On 07/02/2009 06:37 PM, Gregory Haskins wrote:
>>> (Applies to kvm.git/master:1f9050fd)
>>>
>>> The following is the latest attempt to fix the races in
>>> irqfd/eventfd, as
>>> well as restore DEASSIGN support.  For more details, please read the
>>> patch
>>> headers.
>>>
>>> As always, this series has been tested against the kvm-eventfd unit
>>> test
>>> and everything appears to be functioning properly. You can download
>>> this
>>> test here:
>>
>> Applied, thanks.
>>
>
> ... and unapplied.  There's a refcounting mismatch in irqfd_cleanup: a
> reference is taken for each irqfd, but dropped for each guest.

Doh!

Note that the kvm->irqfds.init boolean was supposed to make it both
acquire and release the reference on a per-vm basis.  Obviously,
something is wrong. :)

But I see later in the thread you guys have already figured out that
only the last patch needs to be reverted.  Ill take a look at the last
patch and fix the issue once I am back in the office tomorrow.  Sorry
for the trouble.

Kind Regards,
-Greg

>   This causes an oops if a guest with no irqfds is created and destroyed:
>
> IP: [<ffffffff8105254a>] flush_workqueue+0x23/0x6e
> PGD 0
> Oops: 0000 [4] SMP
> CPU 1
> Modules linked in: kvm_intel kvm nfs lockd nfs_acl sco bridge stp bnep
> l2cap bluetooth autofs4 sunrpc ipv6 dm_multipath uinput i5000_edac
> e1000e edac_core iTCO_wdt
>  iTCO_vendor_support i2c_i801 i2c_core e100 mii floppy pcspkr shpchp
> serio_raw ata_generic pata_acpi [last unloaded: kvm]
> Pid: 2088, comm: qemu Tainted: G      D  
> 2.6.27.19-170.2.35.fc10.x86_64 #1 TYAN Transport GT20-B5372
> RIP: 0010:[<ffffffff8105254a>]  [<ffffffff8105254a>]
> flush_workqueue+0x23/0x6e
> RSP: 0018:ffff8801077d1b08  EFLAGS: 00010292
> RAX: ffffffff8156de08 RBX: ffff8801097e8a50 RCX: 0000000000000000
> RDX: 0000000000000000 RSI: 000000000000019f RDI: 0000000000000000
> RBP: ffff8801077d1b28 R08: 0000000000000000 R09: ffffffff81140027
> R10: ffff88012f402340 R11: ffff880125daf820 R12: ffffffff8156de10
> R13: 0000000000000000 R14: ffff88012f449cd8 R15: ffff88012f002900
> FS:  00007f9b67295950(0000) GS:ffff88012fc04980(0000)
> knlGS:0000000000000000
> CS:  0010 DS: 002b ES: 002b CR0: 000000008005003b
> CR2: 0000000000000020 CR3: 0000000000201000 CR4: 00000000000026e0
> DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
> DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400
> Process qemu (pid: 2088, threadinfo ffff8801077d0000, task
> ffff88012984c530)
> Stack:  ffff8801077d1b28 ffff8801097e8a50 ffff8801097e8000
> ffff8801097e8a68
>  ffff8801077d1b58 ffffffffa01c4eec 0000000800000000 ffff8801097e8000
>  ffff88012f449cd8 ffff88012c80c600 ffff8801077d1b78 ffffffffa01af34b
> Call Trace:
>  [<ffffffffa01c4eec>] kvm_irqfd_release+0x7a/0xcc [kvm]
>  [<ffffffffa01af34b>] kvm_vm_release+0x18/0x27 [kvm]
>  [<ffffffff810c14c7>] __fput+0xca/0x16d
>  [<ffffffff810c157f>] fput+0x15/0x17
>  [<ffffffff810bea29>] filp_close+0x67/0x72
>  [<ffffffff810433ec>] put_files_struct+0x74/0xc8
>  [<ffffffff81043488>] exit_files+0x48/0x51
>  [<ffffffff81044de9>] do_exit+0x26a/0x8a0
>  [<ffffffffa01da330>] ? vmx_vcpu_put+0x9/0xb [kvm_intel]
>  [<ffffffff810454a1>] do_group_exit+0x82/0xaf
>  [<ffffffff8104eabb>] get_signal_to_deliver+0x2b0/0x2dc
>  [<ffffffff81010379>] ? sysret_signal+0x42/0x71
>  [<ffffffff8100f45f>] do_notify_resume+0x90/0x93f
>  [<ffffffff81060cca>] ? do_futex+0x90/0x973
>  [<ffffffffa01ad956>] ? kvm_vcpu_ioctl+0x470/0x485 [kvm]
>  [<ffffffff81333801>] ? trace_hardirqs_on_thunk+0x3a/0x3c
>  [<ffffffff810616a2>] ? sys_futex+0xf5/0x113
>  [<ffffffff81010379>] ? sysret_signal+0x42/0x71
>  [<ffffffff81010737>] ptregscall_common+0x67/0xb0
>
> irqfd_cleanup.wq has never been initialized, but is destroyed.
>

[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 266 bytes --]