Re: [musl] Why is setrlimit() considered to have per-thread effect?

[Alexey Izbyshev <izbyshev@ispras.ru>]

On 2020-10-15 11:50, Szabolcs Nagy wrote:
> note that prlimit does not have synccall in
> musl: the kernel implemented the per process
> rlimit setting when prlimit was added.
> (i think this is linux commit
>  1c1e618ddd15f69fd87ccea596769f78c8065504 )
> 
> but older kernels don't have that.
> 
Ah, thank you for checking that, though the transition appear to have 
happened much earlier than the commit you referenced (which is not 
relevant), in pre-git epoch between 2.6.9 and 2.6.10[1, 2]. I was 
confused because Linux man pages never mention that and explicitly say 
"Resource limits are per-process attributes that are shared by all of 
the threads in a process."[3], but I should have checked old sources.

>> Tangentially, setgroups() is not called via __synccall(), though it 
>> does
>> have per-thread effect. Is this intentional?
> 
> that may be a bug, but it's not a posix api
> so not a conformance issue, but a linux issue:
> if other linux libcs don't do synccall then
> that's the defacto interface contract.
> 
FWIW, glibc does synccall since 2011: 
https://sourceware.org/git/?p=glibc.git;a=commit;h=70181fddf14

[1] https://elixir.bootlin.com/linux/v2.6.9/source/kernel/sys.c#L1537
[2] https://elixir.bootlin.com/linux/v2.6.10/source/kernel/sys.c#L1487
[3] https://man7.org/linux/man-pages/man2/setrlimit.2.html