Re: [RFC PATCH] sched_domains: Make SD_NODE_INIT per-arch

From: Andi Kleen (ak_at_suse.de)
Date: 09/30/04

  • Next message: Jody McIntyre: "Re: [PATCH] Realtime LSM" 
    Date:	Thu, 30 Sep 2004 23:12:28 +0200
To: Matthew Dobson <colpatch@us.ibm.com>

    > Well, you can certainly base the x86_64 CMP values on the current
    > SD_SIBLING_INIT values. Those are well publicized, see
    > include/linux/sched.h! ;)

    Current BK has it in kernel/sched.c.

    And it also broke NUMA kernels on UP, but that's a different issue.

    > I suppose it would be pretty trivial to define defaults in
    > include/asm-generic/topology.h, and allow arches that care to define
    > their own SD_*_INITs without disrupting anyone else. Actually, that's
    > far better than what I've got now. I'll run that patch up after the
    > meeting I'm currently late for and post it in a couple hours.

    Full override isn't good imho because it could lead to bit rot,
    better is to have defaults that can be used as a base, but tweaked.

    -Andi
    -
    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: Jody McIntyre: "Re: [PATCH] Realtime LSM"