Re: [Qemu-devel] [RFC] dump memory when host pci device is used by guest

[Jan Kiszka <jan.kiszka@web.de>]

[-- Attachment #1: Type: text/plain, Size: 9528 bytes --]

On 2011-11-16 14:29, Dave Anderson wrote:
> 
> 
> ----- Original Message -----
>> Hi, all
>>
>> 'virsh dump' can not work when host pci device is used by guest. We have
>> discussed this issue here:
>> http://lists.nongnu.org/archive/html/qemu-devel/2011-10/msg00736.html
>>
>> We have determined to introduce a new command dump to dump memory.
>> The core file's format can be elf.
>>
>> I created a kdump-elf vmcore, and found that it can be used by both
>> crash and gdb:
>>
>> # gdb /usr/lib/debug/lib/modules/2.6.32-71.el6.x86_64/vmlinux
>> /work/core/vmcore
>> GNU gdb (GDB) Red Hat Enterprise Linux (7.2-48.el6)
>> Copyright (C) 2010 Free Software Foundation, Inc.
>> License GPLv3+: GNU GPL version 3 or later
>> <http://gnu.org/licenses/gpl.html>
>> This is free software: you are free to change and redistribute it.
>> There is NO WARRANTY, to the extent permitted by law.  Type "show
>> copying"
>> and "show warranty" for details.
>> This GDB was configured as "x86_64-redhat-linux-gnu".
>> For bug reporting instructions, please see:
>> <http://www.gnu.org/software/gdb/bugs/>...
>> Reading symbols from
>> /usr/lib/debug/lib/modules/2.6.32-71.el6.x86_64/vmlinux...done.
>> [New Thread 1691]
>> [New <main task>]
>> #0  sysrq_handle_crash (key=99, tty=0x0) at drivers/char/sysrq.c:130
>> 130	drivers/char/sysrq.c: No such file or directory.
>> 	in drivers/char/sysrq.c
>> (gdb) bt
>> #0  sysrq_handle_crash (key=99, tty=0x0) at drivers/char/sysrq.c:130
>> #1  0xffffffff8130d822 in __handle_sysrq (key=99, tty=0x0,
>> check_mask=<value optimized out>) at drivers/char/sysrq.c:521
>> #2  0xffffffff8130d8de in write_sysrq_trigger (file=<value optimized
>> out>, buf=<value optimized out>, count=2, ppos=<value optimized
>> out>) at drivers/char/sysrq.c:599
>> #3  0xffffffff811cf31e in proc_reg_write (file=<value optimized out>,
>> buf=0x7fdabafea000 <Address 0x7fdabafea000 out of bounds>, count=2,
>> ppos=<value optimized out>)
>>     at fs/proc/inode.c:207
>> #4  0xffffffff8116c818 in vfs_write (file=0xffff88003c7bb740,
>> buf=0x7fdabafea000 <Address 0x7fdabafea000 out of bounds>,
>> count=<value optimized out>, pos=0xffff88003767ff48)
>>     at fs/read_write.c:347
>> #5  0xffffffff8116d251 in sys_write (fd=<value optimized out>,
>> buf=0x7fdabafea000 <Address 0x7fdabafea000 out of bounds>, count=2)
>> at fs/read_write.c:399
>> #6  0xffffffff81013172 in ?? () at arch/x86/kernel/entry_64.S:487
>> #7  0x0000000000000246 in ?? ()
>> #8  0x00000000ffffffff in ?? ()
>> #9  0x00007fdabafde700 in ?? ()
>> #10 0x000000000000000a in ?? ()
>> #11 0x0000000000000001 in ?? ()
>> #12 0x0000000000000002 in ?? ()
>> #13 0x0000000000000001 in ?? ()
>> #14 0x00000030f80d4230 in ?? ()
>> #15 0x0000000000000033 in ?? ()
>> #16 0x0000000000010206 in ?? ()
>> #17 0x00007fff8a126470 in ?? ()
>> #18 0x000000000000002b in ?? ()
>> #19 0xffff8800374f5000 in ?? ()
>> #20 0xffff88003c6f9000 in ?? ()
>> #21 0x0000000000000080 in ?? ()
>> #22 0xffff880037680080 in ?? ()
>> #23 0xffffffff00000014 in ?? ()
>> #24 0x0000000000000000 in ?? ()
>> (gdb) q
>> # crash -s /usr/lib/debug/lib/modules/2.6.32-71.el6.x86_64/vmlinux /work/core/vmcore
>> crash> bt
>> PID: 1691   TASK: ffff88003711d520  CPU: 0   COMMAND: "bash"
>>  #0 [ffff88003767fae0] machine_kexec at ffffffff8103695b
>>  #1 [ffff88003767fb40] crash_kexec at ffffffff810b8f08
>>  #2 [ffff88003767fc10] oops_end at ffffffff814cbbd0
>>  #3 [ffff88003767fc40] no_context at ffffffff8104651b
>>  #4 [ffff88003767fc90] __bad_area_nosemaphore at ffffffff810467a5
>>  #5 [ffff88003767fce0] bad_area at ffffffff810468ce
>>  #6 [ffff88003767fd10] do_page_fault at ffffffff814cd740
>>  #7 [ffff88003767fd60] page_fault at ffffffff814caf45
>>     [exception RIP: sysrq_handle_crash+22]
>>     RIP: ffffffff8130d566  RSP: ffff88003767fe18  RFLAGS: 00010096
>>     RAX: 0000000000000010  RBX: 0000000000000063  RCX: 0000000000000000
>>     RDX: 0000000000000000  RSI: 0000000000000000  RDI: 0000000000000063
>>     RBP: ffff88003767fe18   R8: 0000000000000000   R9: ffffffff815106c0
>>     R10: 0000000000000001  R11: 0000000000000000  R12: 0000000000000000
>>     R13: ffffffff8179e6c0  R14: 0000000000000286  R15: 0000000000000007
>>     ORIG_RAX: ffffffffffffffff  CS: 0010  SS: 0018
>>  #8 [ffff88003767fe20] __handle_sysrq at ffffffff8130d822
>>  #9 [ffff88003767fe70] write_sysrq_trigger at ffffffff8130d8de
>> #10 [ffff88003767fea0] proc_reg_write at ffffffff811cf31e
>> #11 [ffff88003767fef0] vfs_write at ffffffff8116c818
>> #12 [ffff88003767ff30] sys_write at ffffffff8116d251
>> #13 [ffff88003767ff80] system_call_fastpath at ffffffff81013172
>>     RIP: 00000030f80d4230  RSP: 00007fff8a126470  RFLAGS: 00010206
>>     RAX: 0000000000000001  RBX: ffffffff81013172  RCX: 0000000000000400
>>     RDX: 0000000000000002  RSI: 00007fdabafea000  RDI: 0000000000000001
>>     RBP: 00007fdabafea000   R8: 000000000000000a   R9: 00007fdabafde700
>>     R10: 00000000ffffffff  R11: 0000000000000246  R12: 0000000000000002
>>     R13: 00000030f8379780  R14: 0000000000000002  R15: 00000030f8379780
>>     ORIG_RAX: 0000000000000001  CS: 0033  SS: 002b
>> crash>
>>
>> I wrote a sample(not finished). It only can works on x86_64(both host and guest)
>> I use it to create a core file:
>> # readelf -h /tmp/vm2.save
>> ELF Header:
>>   Magic:   7f 45 4c 46 02 01 01 00 00 00 00 00 00 00 00 00
>>   Class:                             ELF64
>>   Data:                              2's complement, little endian
>>   Version:                           1 (current)
>>   OS/ABI:                            UNIX - System V
>>   ABI Version:                       0
>>   Type:                              CORE (Core file)
>>   Machine:                           Advanced Micro Devices X86-64
>>   Version:                           0x1
>>   Entry point address:               0x0
>>   Start of program headers:          64 (bytes into file)
>>   Start of section headers:          0 (bytes into file)
>>   Flags:                             0x0
>>   Size of this header:               64 (bytes)
>>   Size of program headers:           56 (bytes)
>>   Number of program headers:         9
>>   Size of section headers:           0 (bytes)
>>   Number of section headers:         0
>>   Section header string table index: 0
>> # readelf -l /tmp/vm2.save
>>
>> Elf file type is CORE (Core file)
>> Entry point 0x0
>> There are 9 program headers, starting at offset 64
>>
>> Program Headers:
>>   Type           Offset             VirtAddr           PhysAddr
>>                  FileSiz            MemSiz              Flags  Align
>>   NOTE           0x0000000000000238 0x0000000000000000 0x0000000000000000
>>                  0x00000000000002c8 0x00000000000002c8         0
>>   LOAD           0x0000000000000500 0xffffffff81000000 0x0000000001000000
>>                  0x000000001f000000 0x000000001f000000         0
>>   LOAD           0x000000001f000500 0x0000000000000000 0x0000000000000000
>>                  0x0000000001000000 0x0000000001000000         0
>>   LOAD           0x0000000020000500 0x0000000000000000 0x0000000020000000
>>                  0x0000000000020000 0x0000000000020000         0
>>   LOAD           0x0000000020020500 0x0000000000000000 0x0000000020870000
>>                  0x0000000000010000 0x0000000000010000         0
>>   LOAD           0x0000000020030500 0x0000000000000000 0x0000000020850000
>>                  0x0000000000020000 0x0000000000020000         0
>>   LOAD           0x0000000020050500 0x0000000000000000 0x0000000020840000
>>                  0x0000000000010000 0x0000000000010000         0
>>   LOAD           0x0000000020060500 0x0000000000000000 0x0000000020040000
>>                  0x0000000000800000 0x0000000000800000         0
>>   LOAD           0x0000000020860500 0x0000000000000000 0x0000000020020000
>>                  0x0000000000020000 0x0000000000020000         0
>>
>> I can use crash to anaylze the file, but I can not use gdb to anaylze it.
>> # gdb /usr/lib/debug/lib/modules/2.6.32-71.el6.x86_64/vmlinux /tmp/vm2.save
>> GNU gdb (GDB) Red Hat Enterprise Linux (7.2-48.el6)
>> Copyright (C) 2010 Free Software Foundation, Inc.
>> License GPLv3+: GNU GPL version 3 or later
>> <http://gnu.org/licenses/gpl.html>
>> This is free software: you are free to change and redistribute it.
>> There is NO WARRANTY, to the extent permitted by law.  Type "show
>> copying"
>> and "show warranty" for details.
>> This GDB was configured as "x86_64-redhat-linux-gnu".
>> For bug reporting instructions, please see:
>> <http://www.gnu.org/software/gdb/bugs/>...
>> Reading symbols from
>> /usr/lib/debug/lib/modules/2.6.32-71.el6.x86_64/vmlinux...done.
>> [New <main task>]
>> [New <main task>]
>> #0  0x8103be8b00000000 in ?? ()
>> (gdb) bt
>> #0  0x8103be8b00000000 in ?? ()
>> Cannot access memory at address 0x8170dec800000000
>> (gdb) q
>>
>> My first and the most important question is that: Is there necessary
>> to continue this work?
>>
>> The attachment is the sample patch.
>>
>> Thanks
>> Wen Congyang
> 
> From an enterprise/support point of view, the wholesale replacement
> of the current use of the savevm dumpfile format by "virsh dump" with
> this ELF style format would be a *huge* improvement. 

Yes, fully agree. Would be cool if that could actually work for both
crash and gdb. Looking forward!

Jan


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 262 bytes --]