prefetch

From: Linus Torvalds <[e-mail]>Newsgroups: fa.linux.kernelSubject: Re: [patch] epoll use a single inode ...Date: Wed, 07 Mar 2007 22:58:57 UTCMessage-ID: <fa.xioENCdH8/[e-mail]>On Wed, 7 Mar 2007, Anton Blanchard wrote:>> Funny you mention this. We found some noticeable ppc64 regressions when> moving the dcache to standard list macros and had to do this to fix it> up:>> static inline void prefetch(const void *x)> {>         if (unlikely(!x))>                 return;>>         __asm__ __volatile__ ("dcbt 0,%0" : : "r" (x));> }>> Urgh :)Yeah, I'm not at all surprised. Any implementation of "prefetch" thatdoesn't just turn into a no-op if the TLB entry doesn't exist (which makesthem weaker for *actual* prefetching) will generally have a hard time witha NULL pointer. Exactly because it will try to do a totally unnecessaryTLB fill - and since most CPU's will not cache negative TLB entries, thatunnecessary TLB fill will be done over and over and over again..In general, using software prefetching is just a stupid idea, unless - the prefetch really is very strict (ie for a linked list you do exactly   the above kinds of things to make sure that you don't try to prefetch   the non-existent end entry)AND - the CPU is stupid (in-order in particular).I think Intel even suggests in their optimization manuals to *not* dosoftware prefetching, because hw can usually simply do better without it.Linus