![]() | |||||||||||||
Messages in this thread |
Andrew Morton wrote: > > > Is prefetching 4k at a time optimal? Should the prefetch be embedded > in copy_*_user? > For the Pentium III, that's the Intel recommended sequence. > > The code makes no attempt to align the prefetch address to anything. > Should it? Not needed. > What happens if a prefetch spills over the end of the page and > faults? > Doesn't hurt, prefetch instruction never cause page faults. BUT: Prefetch doesn't preload the TLB. If the TLB entry for the kmap is not in the TLB, all prefetch instructions are treated as nops.(on pIII). -- Manfred - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/ | ||||||||||||
| Last update: 2005-03-22 12:14 [from the cache] ©2003-2008 | |||||||||||||