Re: -mregparm=3

Linus Torvalds (torvalds@transmeta.com)
8 Jul 1998 18:55:29 GMT


In article <r27m1omika.fsf@happy.cygnus.com>,
Ulrich Drepper <drepper@cygnus.com> wrote:
>"Benjamin C.R. LaHaise" <blah@kvack.org> writes:
>
>> <sigh> It's a wonderful idea that really cuts down on the cost of function
>> calls quite a bit, but alas, neither gcc nor egcs work even vaguely
>> reliably with the -mregparm option.
>
>I'm using regparm (using attributes, though) in glibc 2.1 on Intel and
>it does work. Don't spread these wrong information.

Uli, he's right. It's you who are spreading mis-information.

regparm _does_ work on a small scale, and I've been using it in 2.1.x
series. However, claiming that it is reliable in all cases is a pipe
dream - both gcc and egcs _do_ mess it up quite badly in some cases.
They may be rare, but when they happen the code is complete crap.

I've been using it very carefully on a function-by-function basis for
cases where I really care, but it really doesn't work in the general
case.

Linus

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu