Re: [PATCH] revert zap_other_threads breakage, disallow CLONE_THREAD without CLONE_DETACHED

[Jamie Lokier]

Linus Torvalds wrote:
> > Don't forget to mention that software that may be run on 2.5 kernels
> > needs to set both bits, else won't work as expected.
> 
> Well, the CLONE_DETACHED without CLONE_THREAD case was never legal, and my
> current patch will actually warn about the newly disallowed case too. I've
> not gotten any warnings with RH-9, and I don't think anybody else has a 
> new enough glibc to even use CLONE_THREAD at all.

I'm thinking of programs that don't use Glibc, but do use these features.
Perhaps I'm the only person who writes such code :)

> But yes, there will be a warning, at least for a time (and eventually 
> we'll just return -EINVAL silently - ie the program will _fail_ the 
> clone(), it won't just act strangely).

-EINVAL would be great.

-- Jamie
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/