Re: [PATCH] mm/slab.c : prefetchw the start of new allocated objects

2005-07-30 Thread David S. Miller
From: Eric Dumazet <[EMAIL PROTECTED]> Date: Fri, 29 Jul 2005 11:05:09 +0200 > Some CPU lacks a prefetchw() and currently do nothing, so I ask this > question : Should'nt make prefetchw() do at least a prefetch() ? A > read hint is better than nothing. This is not true, especially on SMP. If the

Re: [PATCH] mm/slab.c : prefetchw the start of new allocated objects

2005-07-29 Thread Ingo Molnar
* Eric Dumazet <[EMAIL PROTECTED]> wrote: > local_irq_restore(save_flags); > objp = cache_alloc_debugcheck_after(cachep, flags, objp, > __builtin_return_address(0)); > + prefetchw(objp); > return objp; the idea is good, but i'd suggest to do the prefetchw a bit earlier, r

Re: [PATCH] mm/slab.c : prefetchw the start of new allocated objects

2005-07-29 Thread Andrew Morton
Eric Dumazet <[EMAIL PROTECTED]> wrote: > > Most of objects returned by __cache_alloc() will be written by the caller, > (but not all callers want to write all the object, but just at the begining) > prefetchw() tells the modern CPU to think about the future writes, ie start > some memory transa

[PATCH] mm/slab.c : prefetchw the start of new allocated objects

2005-07-29 Thread Eric Dumazet
[MM] slab.c : prefetchw the start of new allocated objects Most of objects returned by __cache_alloc() will be written by the caller, (but not all callers want to write all the object, but just at the begining) prefetchw() tells the modern CPU to think about the future writes, ie start some memor