Re: [Xen-devel] Failure to boot HVM guest with more than 32 VCPUS

[Vitaly Kuznetsov <vkuznets@xxxxxxxxxx>]

"Hao, Xudong" <xudong.hao@xxxxxxxxx> writes:

>
> Attach the Fedora22 log, a different issue.
>

Here is the crash:

[ 6399.814801] xen_netfront: Initialising Xen virtual ethernet driver
[ 6399.814838] [drm] Initialized drm 1.1.0 20060810
[ 6399.838500] xen_netfront: can't alloc rx grant refs
[ 6399.838504] net eth0: only created 31 queues
[ 6399.839930] BUG: unable to handle kernel NULL pointer dereference at 
0000000000000018
[ 6399.839938] IP: [<ffffffffa01a00cb>] netback_changed+0x8eb/0xef0 
[xen_netfront]
[ 6399.839940] PGD 0
[ 6399.839943] Oops: 0000 [#1] SMP
[ 6399.839948] Modules linked in: drm xen_netfront(+) xen_blkfront(+) 
crc32c_intel ata_generic pata_acpi
[ 6399.839955] CPU: 0 PID: 241 Comm: xenwatch Not tainted 4.0.4-301.fc22.x86_64 
#1
[ 6399.839957] Hardware name: Xen HVM domU, BIOS 4.6.0-rc 08/13/2015
[ 6399.839959] task: ffff880107ffc590 ti: ffff880107920000 task.ti: 
ffff880107920000
[ 6399.839963] RIP: 0010:[<ffffffffa01a00cb>]  [<ffffffffa01a00cb>] 
netback_changed+0x8eb/0xef0 [xen_netfront]
[ 6399.839965] RSP: 0000:ffff880107923d68  EFLAGS: 00010202
[ 6399.839966] RAX: 0000000000000000 RBX: ffff8800e6568000 RCX: 0000000000000001
[ 6399.839968] RDX: 00000000000e6568 RSI: ffff8800e65c40f8 RDI: 0000000000003f41
[ 6399.839969] RBP: ffff880107923df8 R08: ffffc90000940000 R09: 0000000000000000
[ 6399.839971] R10: ffffea0003995a00 R11: 0000000000100000 R12: ffff8800e65c4000
[ 6399.839972] R13: ffff8800e65c40f8 R14: ffff8800e6567000 R15: 0000000000044000
[ 6399.839974] FS:  0000000000000000(0000) GS:ffff880109600000(0000) 
knlGS:0000000000000000
[ 6399.839975] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[ 6399.839977] CR2: 0000000000000018 CR3: 0000000001c0b000 CR4: 00000000001407f0
[ 6399.839980] Stack:
[ 6399.839983]  ffff880107923d78 ffff8800e65c1e04 ffff8800e65c1f21 
ffff8800eb8b1400
[ 6399.839985]  ffff8800e73d8000 ffff8800eb8b1400 0000002107923dec 
ffff880000000020
[ 6399.839988]  0000000107ffc590 0000004000000001 ffff8800ebb261f1 
0000000087ee3bd9
[ 6399.839989] Call Trace:
[ 6399.839998]  [<ffffffff8147161d>] xenbus_otherend_changed+0xad/0x120
[ 6399.840014]  [<ffffffff810decd7>] ? prepare_to_wait_event+0x87/0x100
[ 6399.840017]  [<ffffffff814709c0>] ? unregister_xenbus_watch+0x1d0/0x1d0
[ 6399.840021]  [<ffffffff81473b63>] backend_changed+0x13/0x20
[ 6399.840023]  [<ffffffff81470a4f>] xenwatch_thread+0x8f/0x150
[ 6399.840026]  [<ffffffff810dedc0>] ? wake_atomic_t_function+0x70/0x70
[ 6399.840031]  [<ffffffff810bb678>] kthread+0xd8/0xf0
[ 6399.840035]  [<ffffffff810bb5a0>] ? kthread_worker_fn+0x180/0x180
[ 6399.840041]  [<ffffffff81788bd8>] ret_from_fork+0x58/0x90
[ 6399.840044]  [<ffffffff810bb5a0>] ? kthread_worker_fn+0x180/0x180
[ 6399.840056] Code: 48 8b 04 f0 48 83 f8 ff 0f 84 e2 04 00 00 48 89 c6 48 b8 
ff ff ff ff ff ff ff 3f 48 21 c6 e9 ad fd ff ff 49 8b 44 24 20 4c 89 ee <48> 8b 
78 18 e8 fc ee 2c e1 85 c0 0f 88 be fd ff ff 49 8b 44 24
[ 6399.840059] RIP  [<ffffffffa01a00cb>] netback_changed+0x8eb/0xef0 
[xen_netfront]
[ 6399.840060]  RSP <ffff880107923d68>
[ 6399.840061] CR2: 0000000000000018
[ 6399.840064] ---[ end trace 790f9d91e3f3059b ]---

The main issue (I suppose) is:
[ 6399.838500] xen_netfront: can't alloc rx grant refs

it happens when gnttab_alloc_grant_references() fails. As far as I can
see it can only happen when gnttab_expand() fails. Do you have anything
in your Xen dmesg ('xl dmegs' output on your host)? In case not, can you
try increasing you guest's memory to see if the issue goes away?

-- 
  Vitaly

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