[CRIU] [PATCH 0/8] [v3] Prepare the support for AArch64

[Alexander Kartashov]

On 11/05/2013 02:48 PM, Andrew Vagin wrote:
> Looks like we allocate not enough memory for parasite arguments.
> A following patch helps me.
>
> diff --git a/parasite-syscall.c b/parasite-syscall.c
> index e67b1c3..1d14b48 100644
> --- a/parasite-syscall.c
> +++ b/parasite-syscall.c
> @@ -1152,7 +1152,7 @@ struct parasite_ctl *parasite_infect_seized(pid_t
> pid, struct pstree_item *item,
>           * without using ptrace at all.
>           */
>   
> -       ctl->args_size = parasite_args_size(vma_area_list, dfds, timer_n);
> +       ctl->args_size = parasite_args_size(vma_area_list, dfds, timer_n) + 40960;
>          map_exchange_size = parasite_size + ctl->args_size;
>          map_exchange_size += RESTORE_STACK_SIGFRAME + PARASITE_STACK_SIZE;
>          if (item->nr_threads > 1)
OK for now. However, adding the code dumping the parasite argument area
helps too. I suspect the problem has something to do with the size of 
the parasite
blob. Moreover, the same offset in the area is corrupted on a particular
build of the parasite blob.

-- 
Sincerely yours,
Alexander Kartashov

Intern
Core team

www.parallels.com

Skype: aleksandr.kartashov
Email: alekskartashov at parallels.com