RE: [Xen-devel] Xen unstable crash

["Jiang, Yunhong" <yunhong.jiang@xxxxxxxxx>]

I'm not sure if we can temporaryly print some information (or BUG) in 
put_page() when the count_info is 0 already. (I think that should be bug even 
in the long run, since the get_page will fail is count_info is overflow).

As I can't reproduce this issue locally, I have to guess the possibility root 
cause. If there is a spurious put_page(), it may works well before patch 
19285/19286 (Especially when the windows in the two put_page is quite small), 
since originally free_heap_pages() will not check count_info. However, it will 
trigger the ASSERT() in free_heap_pages() now for sure. And it can explain the 
BUG() in alloc page, although I can't explain why it works before.

Thanks
Yunhong Jiang



>-----Original Message-----
>From: Keir Fraser [mailto:keir.fraser@xxxxxxxxxxxxx] 
>Sent: 2009年3月13日 22:19
>To: Jiang, Yunhong; Andrew Lyon; Xen-devel
>Subject: Re: [Xen-devel] Xen unstable crash
>
>If you can reproduce this bug, it's worth trying to revert c/s 
>19285 and try
>again:
> hg export 19285 | patch -Rp1
>To put the tree back into clean state afterwards:
> hg diff | patch -Rp1
>
>If the bug still reproduces another possible culprit is c/s 19317.
>
> -- Keir
>
>On 13/03/2009 12:55, "Jiang, Yunhong" <yunhong.jiang@xxxxxxxxx> wrote:
>
>> Thanks for the log, seems the count info  is -1UL in such 
>situation, I think
>> it may because some change to count_info, and I will try to check it.
>> 
>> Thanks
>> Yunhong Jiang
>> 
>>> -----Original Message-----
>>> From: Keir Fraser [mailto:keir.fraser@xxxxxxxxxxxxx]
>>> Sent: 2009年3月13日 18:39
>>> To: Andrew Lyon; Xen-devel
>>> Cc: Jiang, Yunhong
>>> Subject: Re: [Xen-devel] Xen unstable crash
>>> 
>>> Thanks. Our testing has showed this up too. The cause hasn't
>>> been tracked
>>> down yet unfortunately.
>>> 
>>> -- Keir
>>> 
>>> On 13/03/2009 10:33, "Andrew Lyon" <andrew.lyon@xxxxxxxxx> wrote:
>>> 
>>>> Hi,
>>>> 
>>>> Running Xen unstable on a Dell Optiplex 755, after starting and
>>>> shutting down a few hvm's the system crashes with the following
>>>> message:
>>>> 
>>>> This is with Xensource 2.6.18.8 kernel:
>>>> 
>>>> (XEN) ** page_alloc.c:407 -- 449/512 ffffffffffffffff
>>>> (XEN) Xen BUG at page_alloc.c:409
>>>> (XEN) ----[ Xen-3.4-unstable  x86_64  debug=y  Not tainted ]----
>>>> (XEN) CPU:    1
>>>> (XEN) RIP:    e008:[<ffff828c8011206f>] 
>alloc_heap_pages+0x35a/0x486
>>>> (XEN) RFLAGS: 0000000000010286   CONTEXT: hypervisor
>>>> (XEN) rax: 0000000000000000   rbx: ffff82840199b820   rcx:
>>> 0000000000000001
>>>> (XEN) rdx: 000000000000000a   rsi: 000000000000000a   rdi:
>>> ffff828c801fedec
>>>> (XEN) rbp: ffff830127fdfcb8   rsp: ffff830127fdfc58   r8:
>>> 0000000000000004
>>>> (XEN) r9:  0000000000000004   r10: 0000000000000010   r11:
>>> 0000000000000010
>>>> (XEN) r12: ffff828401998000   r13: 00000000000001c1   r14:
>>> 0000000000000200
>>>> (XEN) r15: 0000000000000000   cr0: 0000000080050033   cr4:
>>> 00000000000026f0
>>>> (XEN) cr3: 000000011f46e000   cr2: 000000000133a000
>>>> (XEN) ds: 0000   es: 0000   fs: 0063   gs: 0000   ss: e010 
>  cs: e008
>>>> (XEN) Xen stack trace from rsp=ffff830127fdfc58:
>>>> (XEN)    0000000900000001 0000000000000098 0000000000000200
>>> 0000000100000001
>>>> (XEN)    ffff830127fdfcf8 ffff828c801a5c68 ffff830127fdfccc
>>> ffff830127fdff28
>>>> (XEN)    0000000000000027 0000000000000000 ffff83011d9c4000
>>> 0000000000000000
>>>> (XEN)    ffff830127fdfcf8 ffff828c8011383b 0100000400000009
>>> ffff830127fdff28
>>>> (XEN)    0000000000000006 0000000044803760 00000000448037c0
>>> 0000000000000000
>>>> (XEN)    ffff830127fdff08 ffff828c80110591 0000000000000000
>>> 0000000000000000
>>>> (XEN)    0000000000000000 0000000000000000 0000000000000000
>>> 0000000000000000
>>>> (XEN)    0000000000000200 0000000000000001 0000000000000000
>>> 0000000000000001
>>>> (XEN)    ffff828c80214d00 ffff830127fdfda8 ffff830127fdfde8
>>> ffff828c80115811
>>>> (XEN)    0000000000000001 ffff828c80151777 ffff830127fdfda8
>>> ffff828c8013c559
>>>> (XEN)    0000000000000004 0000020000000001 ffff830127fdfdb8
>>> ffff828c8013c5f6
>>>> (XEN)    ffff830127fdfde8 ffff828c80107247 ffff830127fdfde8
>>> ffff828c8011d73e
>>>> (XEN)    0000000000000001 0000000000000000 ffff830127fdfe28
>>> ffff83011d9c4000
>>>> (XEN)    0000000000000282 0000000400000009 0000000044803750
>>> ffff8300cfdfc030
>>>> (XEN)    ffff830127ff1f28 0000000000000002 ffff830127fdfe58
>>> ffff828c80119cb7
>>>> (XEN)    00007cfed8020197 ffff828c80239180 0000000000000002
>>> ffff830127fdfe68
>>>> (XEN)    ffff828c8011c0c8 ffff828c80239180 ffff830127fdfe78
>>> ffff828c8014c173
>>>> (XEN)    ffff830127fdfe98 ffff828c801d166d ffff830127fdfe98
>>> 00000000000cce00
>>>> (XEN)    000000000001f600 ffff828c801d2063 0000000044803750
>>> 0000000000000004
>>>> (XEN)    0000000000000009 000000000000000a 00002b820b4a5eb7
>>> ffff83011d9c4000
>>>> (XEN) Xen call trace:
>>>> (XEN)    [<ffff828c8011206f>] alloc_heap_pages+0x35a/0x486
>>>> (XEN)    [<ffff828c8011383b>] alloc_domheap_pages+0x128/0x17b
>>>> (XEN)    [<ffff828c80110591>] do_memory_op+0x988/0x17a7
>>>> (XEN)    [<ffff828c801cf1bf>] syscall_enter+0xef/0x149
>>>> (XEN)
>>>> (XEN)
>>>> (XEN) ****************************************
>>>> (XEN) Panic on CPU 1:
>>>> (XEN) Xen BUG at page_alloc.c:409
>>>> (XEN) ****************************************
>>>> (XEN)
>>>> 
>>>> And here again running opensuse 2.6.27 kernel:
>>>> 
>>>> (XEN) ** page_alloc.c:534 -- 0/1 ffffffffffffffff
>>>> (XEN) Xen BUG at page_alloc.c:536
>>>> (XEN) ----[ Xen-3.4-unstable  x86_64  debug=y  Not tainted ]----
>>>> (XEN) CPU:    1
>>>> (XEN) RIP:    e008:[<ffff828c80112d77>] free_heap_pages+0x12f/0x4b8
>>>> (XEN) RFLAGS: 0000000000010206   CONTEXT: hypervisor
>>>> (XEN) rax: 0000000000000000   rbx: ffff82840236e3e0   rcx:
>>> 0000000000000001
>>>> (XEN) rdx: ffffffffffffffff   rsi: 000000000000000a   rdi:
>>> ffff828c801fedec
>>>> (XEN) rbp: ffff830127fdfe90   rsp: ffff830127fdfe40   r8:
>>> 0000000000000004
>>>> (XEN) r9:  0000000000000004   r10: 0000000000000010   r11:
>>> 0000000000000010
>>>> (XEN) r12: ffff82840236e3e0   r13: 0000000000000000   r14:
>>> 0000000000000000
>>>> (XEN) r15: 0080000000000000   cr0: 000000008005003b   cr4:
>>> 00000000000026f0
>>>> (XEN) cr3: 00000000bf1b2000   cr2: 00000000004878d5
>>>> (XEN) ds: 0000   es: 0000   fs: 0000   gs: 0000   ss: e010 
>  cs: e008
>>>> (XEN) Xen stack trace from rsp=ffff830127fdfe40:
>>>> (XEN)    0000000000000001 ffff82840236e3e0 0000000127fdfe90
>>> 0000000000000001
>>>> (XEN)    0000000000000000 0000000000000200 c2c2c2c2c2c2c2c2
>>> ffff82840236e3c0
>>>> (XEN)    ffff82840236e2e0 ffff830000000000 ffff830127fdfed0
>>> ffff828c8011329a
>>>> (XEN)    000000985dfe4f8c 0000000000000001 ffff830127fdff28
>>> ffff828c80297880
>>>> (XEN)    0000000000000002 ffff828c80239100 ffff830127fdff00
>>> ffff828c8011ba21
>>>> (XEN)    ffff8800e9be5d80 ffff830127fdff28 ffff828c802375b0
>>> ffff8300cee8a000
>>>> (XEN)    ffff830127fdff20 ffff828c8013ca78 0000000000000001
>>> ffff8300cfaee000
>>>> (XEN)    ffff830127fdfda8 ffff8800e9be5d80 ffff8800ea1006c0
>>> ffffffff8070f1c0
>>>> (XEN)    000000000000008f ffff8800c4df3c98 0000000000000184
>>> 0000000000000246
>>>> (XEN)    ffff8800c4df3d68 ffff8800eab76b00 0000000000000000
>>> 0000000000000000
>>>> (XEN)    ffffffff802073aa 0000000000000009 00000000deadbeef
>>> 00000000deadbeef
>>>> (XEN)    0000010000000000 ffffffff802073aa 000000000000e033
>>> 0000000000000246
>>>> (XEN)    ffff8800c4df3c60 000000000000e02b 7f766dfbff79beef
>>> fddffff4f3b9beef
>>>> (XEN)    008488008022beef 0001000a0a03beef f7f5ff7b00000001
>>> ffff8300cfaee000
>>>> (XEN) Xen call trace:
>>>> (XEN)    [<ffff828c80112d77>] free_heap_pages+0x12f/0x4b8
>>>> (XEN)    [<ffff828c8011329a>] page_scrub_softirq+0x19a/0x23c
>>>> (XEN)    [<ffff828c8011ba21>] do_softirq+0x6a/0x77
>>>> (XEN)    [<ffff828c8013ca78>] idle_loop+0x9d/0x9f
>>>> (XEN)
>>>> (XEN)
>>>> (XEN) ****************************************
>>>> (XEN) Panic on CPU 1:
>>>> (XEN) Xen BUG at page_alloc.c:536
>>>> (XEN) ****************************************
>>>> (XEN)
>>>> (XEN) Reboot in five seconds...
>>>> 
>>>> 
>>>> Andy
>>>> 
>>>> _______________________________________________
>>>> Xen-devel mailing list
>>>> Xen-devel@xxxxxxxxxxxxxxxxxxx
>>>> http://lists.xensource.com/xen-devel
>>> 
>>> 
>>> 
>
>
>

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