Eric W. Biederman wrote:
> Nadia.Derbey@bull.net writes:
> 
> 
>>This patchset is a part of an effort to change some syscalls behavior for
>>checkpoint restart.
> 
> 
> Thanks for doing this.
> 
> Unfortunately this makes a very good case of why we don't want to go down
> this route.  Adding magic parameters to syscalls that are only useful
> in one very specific restart case.
> 
> We need good clean interfaces with well defined semantics.
> 
> Something as narrow focused on this is not really useful and it takes
> a lot of code to do something very few people will want to actively
> do.

All this seems reasonable.
Ok, so since we are taking the "new syscalls" direction, I'll try to 
make a list of the potentially duplicated syscalls.

Regards,
Nadia

> 
> 
>>The syntax is:
>># echo "LONG1 XX" > /proc/self/task/<my_tid>/next_syscall_data
>>     next object to be created will have an id set to XX
> 
> 
> Which his horrible in another way because it is hugely race prone.
> 
> Eric
> 
> 
> 



_______________________________________________
Containers mailing list
Containers@lists.linux-foundation.org
https://lists.linux-foundation.org/mailman/listinfo/containers