[CRIU] Restoring failed on ARMv7

Pavel Emelyanov xemul at parallels.com
Mon Nov 2 23:12:19 PST 2015


On 11/03/2015 08:02 AM, Artem Kuzmitskiy wrote:
> Kernel version is 3.16, criu 1.6&1.7, libc 2.19, gcc 4.8.3.
> I asked in the start email about kernel options bz for criu 1.7 check has failed.

OK, would you pick two patches from the mailing list:

[PATCH 1/2] pie: Warn about unknown for log-simple format
[PATCH 2/2] pie: Support %p printing

and try again?

>>     >     > Result of criu check --ms for criu 1.7(4c5b23e52c1dc4e3fbbc7472b92e7b1ce9d22f02):
>>     >     >
>>     >     > Warn  (cr-check.c:189): Skipping unssuported PR_SET_MM_MAP
>>     >     >
>>     >     > Warn  (cr-check.c:588): Skipping PTRACE_O_SUSPEND_SECCOMP check
>>     >     >
>>     >     > Warn  (cr-check.c:615): Dirty tracking is OFF. Memory snapshot will not work.
>>     >     >
>>     >     > Error (timerfd.c:55): timerfd: No timerfd support for c/r: Inappropriate ioctl for device
>>     >     >
>>     >     > Warn  (cr-check.c:702): Skipping unsupported AIO remap
>>     >     >
>>     >     > Warn  (cr-check.c:718): fdinfo doesn't contain the lock field
> 
> 2015-11-02 19:32 GMT+03:00 Christopher Covington <cov at codeaurora.org <mailto:cov at codeaurora.org>>:
> 
>     On 11/02/2015 10:59 AM, Pavel Emelyanov wrote:
>     > On 10/30/2015 05:48 PM, Artem Kuzmitskiy wrote:
>     >>
>     >>
>     >> 2015-10-30 16:33 GMT+03:00 Pavel Emelyanov <xemul at parallels.com <mailto:xemul at parallels.com> <mailto:xemul at parallels.com <mailto:xemul at parallels.com>>>:
>     >>
>     >>     On 10/30/2015 04:17 PM, Artem Kuzmitskiy wrote:
>     >>     >
>     >>     >
>     >>     > 2015-10-30 12:52 GMT+03:00 Pavel Emelyanov <xemul at parallels.com <mailto:xemul at parallels.com> <mailto:xemul at parallels.com <mailto:xemul at parallels.com>> <mailto:xemul at parallels.com <mailto:xemul at parallels.com> <mailto:xemul at parallels.com <mailto:xemul at parallels.com>>>>:
>     >>     >
>     >>     >     On 10/30/2015 12:38 PM, Artem Kuzmitskiy wrote:
>     >>     >     > Hi all,
>     >>     >     >
>     >>     >     > I checked criu on armv7 with 3.16 kernel and got next error when restoring:
>     >>     >     > ....
>     >>     >     > (00.087176)   6034: task_args: 0xd6000
>     >>     >     > task_args->pid: 6034
>     >>     >     > task_args->nr_threads: 1
>     >>     >     > task_args->clone_restore_fn: 0xd1c90
>     >>     >     > task_args->thread_args: 0xd62c0
>     >>     >     > pie: Switched to the restorer 6034
>     >>     >     > pie: Error (pie/restorer.c:772): Unable to unmap (-): 1992687616
>     >>     >
>     >>     >     Ouch, the pie logging is ... not complete :) Can you strace the restoring
>     >>     >     to see what's going on with this syscall?
>     >>     >
>     >>     >
>     >>     > Strace log (command -strace criu restore -d -D images -o restore.log -v4 --shell-job -t 11758)
>     >>
>     >>     Plz, add -f option to strace, the subtasks syscalls are of the main interest.
>     >
>     > I see no failed munmap-s in the strace below. And this part
>     >
>     >> clone(Process 4820 attached
>     >> child_stack=0x7e894b00, flags=SIGCHLD) = 4820
>     >> [pid  5387] gettimeofday({1446215777, 101923}, NULL) = 0
>     >> [pid  5387] write(1023, "(00.031539) PID: real 4820 virt "..., 37) = 37
>     >> [pid  5387] flock(3, LOCK_UN)           = 0
>     >> [pid  5387] close(3)                    = 0
>     >> [pid  5387] gettimeofday({1446215777, 103520}, NULL) = 0
>     >> [pid  5387] write(1023, "(00.033136) Wait until namespace"..., 46) = 46
>     >> [pid  5387] futex(0x76f7a010, FUTEX_WAIT, 1, NULL
> 
>     >> syscall: unknown syscall trap 0x0f000000
> 
>     I overlooked this before. Is that a syscall number? What syscall is that
>     supposed to be? Could there be a criu vs kernel or libc vs kernel feature
>     mismatch?
> 
>     Christopher Covington
> 
>     --
>     Qualcomm Innovation Center, Inc.
>     The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum,
>     a Linux Foundation Collaborative Project
> 
> 
> 
> 
> -- 
> Best regards,
> Artem Kuzmitskiy
> 



More information about the CRIU mailing list