>>>       int rseq(struct rseq * rseq, uint32_t rseq_len, int flags, uint32_t 
>>>sig);
>> 
>> Really dumb question -- and one I'm sorry to bring up at the last minute. 
>> Should
>> we consider making the syscall name something more generic 
>> "register_tls_abi"?
> I proposed that approach back in 2016 ("tls abi" system call), and the 
> feedback
> I received back then is that it was preferred to have a dedicated "rseq" 
> system
> call than an "open ended" and generic "tls abi" system call.

Ultimately I'm fine either way. I do think that in the past few months of 
review it has become clear that creating this tls abi requires a fair bit of 
work. It'd be a shame to see a future attempt to use such an ABI made difficult 
by forcing the author to figure out the registration process yet again. I 
assume the maintainers of glibc would also like to avoid the need to register 
multiple ABIs.

-b

Reply via email to