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

[Rich Felker <dalias@libc.org>]

On Thu, Oct 15, 2020 at 08:01:00AM +0300, Alexey Izbyshev wrote:
> Hello,
> 
> Commit 544ee752cd[1] claims that setrlimit() is per-thread on Linux,
> similarly to setxid() calls, so it should be called via
> __synccall(). But this appears to be wrong: the kernel code operates
> on tsk->signal[2], which is a per-thread-group structure. Glibc
> doesn't call setrlimit() for each thread either. Am I missing
> something?

POSIX specifies that it sets the limits for the process. If the kernel
doesn't do that, we have to implement in userspace.

> Tangentially, setgroups() is not called via __synccall(), though it
> does have per-thread effect. Is this intentional?

POSIX doesn't define setgroups, so it's up to the implementation.
Conceptually since POSIX has supplemental groups they probably
*should* be forced process-global, so maybe we should change this.

For an example that's more clear-cut, setfs[ug]id is explicitly not
process-global because it's not a security boundary and the whole
purpose is to be able to do local id changes then revert them for the
sake of performing access as a different user/group.

Rich