Re: [musl] vfork()-based posix_spawn() has more failure modes than fork()-based one

[Alexey Izbyshev <izbyshev@ispras.ru>]

On 2022-05-03 00:31, Rich Felker wrote:
> On Mon, May 02, 2022 at 11:25:07PM +0200, Florian Weimer wrote:
>> * Rich Felker:
>> 
>> > I'm trying to understand how this comes to be. The child should
>> > inherit the namespaces of the parent and thus should not be in a
>> > different namespace that precludes spawn. I'm guessing this is some
>> > oddity where unshare doesn't affect the process itself, only its
>> > children? If so, it seems like a bug that it doesn't affect the
>> > process itself after execve (after unshare(1) runs your test program),
>> > but that probably can't be fixed now on the Linux side for stability
>> > reasons. :(
>> 
>> It's about fundamentally conflicting requirements.
>> 
>> The vDSO data mapping needs to store the time offset, so it has to be
>> distinct from the original namespace.  vfork preserves the VM sharing.
>> It's not possible to do both things at the same time.
>> 
>> unshare(CLONE_NEWTIME) should have been specified to only take effect
>> after execve, when the vDSO is remapped anyway.
> 
> Yes, exactly. The bug is that someone confused processes and process
> images (fork and exec) when coming up with the constraint and now
> we're probably stuck with it. *sigh*
> 
Status update: the recently released kernel 6.2 contains the fix[1], so 
we're not stuck with this bug after all!

[1] 
https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=2b5f9dad32ed19e8db3b0f10a84aa824a219803b

Alexey