Re: [PATCH 4/4] signal: Only handle on NSIG signals on signal functions (BZ #25657)

[Adhemerval Zanella <adhemerval.zanella@linaro.org>]

On 21/04/2020 10:14, Florian Weimer wrote:
> * Adhemerval Zanella:
> 
>> On 21/04/2020 09:53, Florian Weimer wrote:
>>> * Adhemerval Zanella via Libc-alpha:
>>>
>>>> The upper bits of the sigset_t s not fully initialized in the signal
>>>> mask calls that return information from kernel (sigprocmask,
>>>> sigpending, and pthread_sigmask), since the exported sigset_t size
>>>> (1024 bits) is larger than Linux support one (64 or 128 bits).
>>>> It might make sigisemptyset/sigorset/sigandset fail if the mask
>>>> is filled prior the call.
>>>>
>>>> This patch changes the internal signal function to handle up to
>>>> supported Linux signal number (_NSIG), the remaining bits are
>>>> untouched.
>>>
>>> Sorry, I have trouble what the intended outcome of this change is.
>>>
>>> What is the expected behavior here?  We have an overly large sigset_t,
>>> but only signal numbers from 0 to NSIG - 1 (inclusive) are valid in
>>> all function calls?
>>>
>>
>> The bug report link to a stackoverflow example, which simplified is:
>>
>>   #define _GNU_SOURCE
>>   #include <stdio.h>
>>   #include <signal.h>
>>   int
>>   main (int argc, char *argv[])
>>   { 
>>     sigset_t set;
>>     sigfillset (&set);
>>     sigprocmask (SIG_BLOCK, NULL, &set);
>>     printf ("%d\n", sigisemptyset (&set));
>>    return 0;
>>   }
>>
>> The main issue is GNU extensions sigisemptyset/sigorset/sigandset
>> operates on full internal signal bits where kernel only updates
>> up to _NSIG signals.
> 
> I get that, but what do you want to do to fix it?
> 
> It's not entirely clear from the changes in the patch.
> 

Only handle __NSIG_WORDS on linux sigsetopts.h instead of _SIGSET_NWORDS
when manipulating sigset_t masks.