[CRIU] [PATCH 11/13] restore: Add restoration of the blocked threads signals from the image

[Pavel Emelyanov]

On 11/07/2012 04:32 PM, Cyrill Gorcunov wrote:
> On Wed, Nov 07, 2012 at 02:07:36PM +0400, Pavel Emelyanov wrote:
>> On 11/06/2012 08:53 PM, Cyrill Gorcunov wrote:
>>>
>>> Signed-off-by: Cyrill Gorcunov <gorcunov at openvz.org>
>>> ---
>>>  cr-restore.c       |    2 ++
>>>  include/restorer.h |    3 +++
>>>  restorer.c         |    3 +++
>>>  3 files changed, 8 insertions(+), 0 deletions(-)
>>>
>>
>>
>>> @@ -199,6 +199,9 @@ static int restore_thread_common(struct rt_sigframe *sigframe,
>>>  		}
>>>  	}
>>>  
>>> +	if (args->has_blk_sigset)
>>> +		sigframe->uc.uc_sigmask.sig[0] = args->blk_sigset;
>>> +
>>>  	restore_sched_info(&args->sp);
>>>  
>>>  	return restore_gpregs(sigframe, &args->gpregs);
>>
>> This should either be in thread-specific code, or the blk_sigset should be accepted
>> as function argument. The latter looks more reasonable to me.
> 
> OK, i must admit I don't get it. The restore_thread_common is already is thread
> specific code, 

It's not. It's in the routine which gets called by both -- threads and task (thread leader).

> and @blk_sigset is sitting in thread_restore_args structure, what
> the point to introduce another level of abstraction and pass it as function
> argument? Frankly, it's restorer code which is supposed to be as small as
> possible.
> .
>