Re: Stop signals and interruption of system calls on Linux

[<yyyyyy@xxxxxxx>]

Michael T Kerrisk <yyyyyyyyy@xxxxxxx> said:
> I remembered that Geoff Clare also pointed out to me the following text 
> from Section 2.4.4:
> 
>     If the action of the signal is to stop the process, the 
>     process shall stop until continued or terminated. Generation 
>     of a SIGCONT signal for the process shall cause the process 
>     to be continued, and the original function shall continue at 
>     the point the process was stopped.
> 
> I think that saves my argument: Linux should not be returning 
> EINTR at this point.  (And of course I'm still curious if any other 
> implementation behaves like Linux.)

I imagine that if you showed this quote to an implementor’s lawyer, he might 
argue that it’s not entirely clear that in the case of a call that was 
stopped while waiting for something, "continue at the point the process was 
stopped" means "continue waiting" rather than "continue from the point where 
it was waiting", and that the standard doesn’t say that it must continue on 
exactly the same code path as it would if the process hadn’t been stopped.  
But I agree that without the above quote, his case would be much stronger...