[lkml]   [2019]   [Sep]   [19]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
Messages in this thread
SubjectRe: [RFC] Improve memset
On Tue, Sep 17, 2019 at 01:45:20PM -0700, Linus Torvalds wrote:
> That sounds better, but I'm a bit nervous about the whole thing
> because who knows when the alternatives code itself internally uses
> memset() and then we have a nasty little chicken-and-egg problem.

You mean memcpy()...?

> Also, for it to make sense to inline rep stosb, I think we also need
> to just make the calling conventions for the alternative calls be that
> they _don't_ clobber other registers than the usual rep ones
> (cx/di/si). Otherwise one big code generation advantage of inlining
> the thing just goes away.

Yah, that is tricky and I have no smart idea how. The ABI puts the
operands in rdi,rsi,rdx, ... while REP; STOSB wants them in rax,rcx,rdi.
And if it were only that, then we could probably accept the 2 movs and
a push but then the old functions clobber three more: "rdx", "r8", "r9".

I could try to rewrite the old functions to see if I can save some regs...

> On the whole I get the feeling that this is all painful complexity and
> we shouldn't do it. At least not without some hard performance numbers
> for some huge improvement, which I don't think we've seen.

Yap, it is starting to become hairy.

> Because I find the thing fascinating conceptually, but am not at all
> convinced I want to deal with the pain in practice ;)

I hear ya.



ECO tip #101: Trim your mails when you reply.

 \ /
  Last update: 2019-09-19 14:56    [W:0.087 / U:0.116 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site