Re: [uml-devel] Re: [PATCH 4/6] x86_64: fix L1_CACHE_SHIFT_MAX for Intel EM64T [for 2.6.14?]

From: Blaisorblade (blaisorblade_at_yahoo.it)
Date: 10/26/05

  • Next message: John Benes: "Re: Call for PIIX4 chipset testers"
    To: user-mode-linux-devel@lists.sourceforge.net
    Date:	Wed, 26 Oct 2005 02:00:15 +0200
    
    

    On Wednesday 26 October 2005 01:33, Andi Kleen wrote:
    > On Wednesday 26 October 2005 00:44, Blaisorblade wrote:
    > > For what I see, that's based on the tradeoff between space and contention
    > > - for instance there are few zones only, so there's no big waste.

    > If space is precious it shouldn't be padded at all.
    Ah, when the structure has many instances padding should be reduced, you mean.
    But well, I think slab itself adds some padding to separate instances, for
    this effect and for cache colouring (I guess that means to prevent too many
    different things from being allocated on addresses matching on the low-order
    bits, so that inserting one in the cache means evicting another of the two -
    it's something I guessed when I studied set-associative caches).

    > > In practice, interpreting !X86_GENERIC as "I will run this kernel on
    > > _this_ processor" could also be done.
    >
    > That is what it always meant yes.
    >
    > > However, in case you didn't note, max_align is never enough on EM64T
    > > currently, right?
    >
    > I will prepare patches for .15 to remove it completely, that should fix
    > that problem.
    Making L1_CACHE_SHIFT_MAX and L1_CACHE_SHIFT match, right? Or forcing all
    architectures to support something like X86_GENERIC?

    Also,

    #define L1_CACHE_BYTES (1 << L1_CACHE_SHIFT)

    should probably be made general, while at it.

    -- 
    Inform me of my mistakes, so I can keep imitating Homer Simpson's "Doh!".
    Paolo Giarrusso, aka Blaisorblade (Skype ID "PaoloGiarrusso", ICQ 215621894)
    http://www.user-mode-linux.org/~blaisorblade
    		
    ___________________________________ 
    Yahoo! Messenger: chiamate gratuite in tutto il mondo 
    http://it.messenger.yahoo.com
    -
    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/
    

  • Next message: John Benes: "Re: Call for PIIX4 chipset testers"

    Relevant Pages

    • RE: why swap at all?
      ... But this doesn't help in this case as the image-file is up to 4,4GB in ... whole which means that it ALONE can fill up the whole cache. ... Real Programmers consider "what you see is what you get" to be just as ... send the line "unsubscribe linux-kernel" in ...
      (Linux-Kernel)
    • Re: statfs() / statvfs() syscall ballsup...
      ... regular writes and read stuff off the disk that was a potential security ... So right now we have extra code and extra complexity (which implies not ... through the page cache to make sure that it's safe. ... send the line "unsubscribe linux-kernel" in ...
      (Linux-Kernel)
    • fastboot, diskstat
      ... that the total time for prefetching + actual boot was only 10% shorter, ... actually cache the pages I touched, ... Also, regarding the directory entries, are they accessed via the buffer ... send the line "unsubscribe linux-kernel" in ...
      (Linux-Kernel)
    • Re: statfs() / statvfs() syscall ballsup...
      ... > touching the page will obviously map in and synchronize the result. ... > populated the cache for a directory structure really quickly using it. ... uptodate, so perhaps calling sys_readaheadon a range always causes ... send the line "unsubscribe linux-kernel" in ...
      (Linux-Kernel)
    • Re: 2.6.0, cdrom still showing directories after being erased
      ... There's no command to invalidate read cache, ... of the SYNC_CACHE command to flush dirty data to media (which is what LG ... send the line "unsubscribe linux-kernel" in ...
      (Linux-Kernel)