2.6.0-test6-mjb1

From: Martin J. Bligh (mbligh_at_aracnet.com)
Date: 10/03/03

  • Next message: Jamie Lokier: "Re: idt change in a running kernel? what locking?"
    Date:	Fri, 03 Oct 2003 09:54:47 -0700
    To: linux-kernel <linux-kernel@vger.kernel.org>
    
    

    The patchset contains mainly scalability and NUMA stuff, and anything
    else that stops things from irritating me. It's meant to be pretty stable,
    not so much a testing ground for new stuff.

    I'd be very interested in feedback from anyone willing to test on any
    platform, however large or small.

    ftp://ftp.kernel.org/pub/linux/kernel/people/mbligh/2.6.0-test6/patch-2.6.0-test6-mjb1.bz2

    additional patches that can be applied if desired:

    Since 2.6.0-test2-mjb1 (~ = changed, + = added, - = dropped)

    Notes:
            Mostly a merge forwards.

    Now in Linus' tree:

    - config_debug Dave Hansen
            Make '-g' for the kernel a config option
            (equivalent patch merged)

    - 32bit_dev_t Andries Brouwer
            Make dev_t 32 bit

    - dynamic_hd_struct Badari Pulavarty
            Allocate hd_structs dynamically

    - 16way_x440 Matt Dobson
            Fix problem with booting on 16x x440

    - iosched_hashes Badari Pulavarty
            Twiddle with the iosched hash tables for fun & profit

    Dropped:

    - irq_affinity Martin J. Bligh
            Workaround for irq_affinity on clustered apic mode systems (eg x440)
            (fixed differently)

    - kgdb Andrew Morton
            The older version of kgdb, synched with 2.5.54-mm1
            (need to pick up the new version of kgdb)

    - thread_under_page William Lee Irwin
            Fix THREAD_SIZE < PAGE_SIZE case
            (not needed for now)

    - reiserfs_dio Mingming Cao
            DIO for Reiserfs
            (doesn't compile)

    - sched_interactive Ingo Molnar
            Bugfix for interactive scheduler
            (not needed anymore)

    - kgdb_cleanup Martin J. Bligh
            Stop kgdb renaming schedule to do_schedule when it's not even enabled
            (need to pick up the new version of kgdb)

    - lotsa_sds Badari Pulavarty
            Create some insane number of sds
            (not easily mergable. Badari to give me a new version)

    - config_numasched Dave Hansen
            Turn NUMA scheduler into a config option
            (ummm. I lost it. Will try to find it again)

    - separate_pmd Dave Hansen
            Separate kernel pmd per task.
            (4/4 split will conflict heavily with this)

    - banana_split Dave Hansen
            Make PAGE_OFFSET play twister and limbo.
            (4/4 split will conflict heavily with this)

    - proc_pid_readdir Manfred Spraul
            Make proc_pid_readdir more efficent. Allegedly.
            (didn't merge easily, and I've never seen it do anything)

    - lockmeter_tytso Ted Tso
            Fix lockmeter
            (merged into main lockmeter patch)

    - early_printk_fix Dave Hansen
            Fix epk conflict with ppc64
            (merged into main epk patch)

    New:

    + page_lock William Lee Irwin
            Conditionally convert mapping->page_lock back to an rwlock

    Pending:
    lotsa_sds
    config_numasched
    4/4 split
    new kgdb
    list_of_lists
    Hyperthreaded scheduler (Ingo Molnar)
    scheduler callers profiling (Anton or Bill Hartner)
    Child runs first (akpm)
    Kexec
    e1000 fixes
    Update the lost timer ticks code
    pidmaps_nodepages (Dave Hansen)

    Present in this patch:

    early_printk Dave Hansen / Keith Mannthey
            Allow printk before console_init

    confighz Andrew Morton / Dave Hansen
            Make HZ a config option of 100 Hz or 1000 Hz

    config_page_offset Dave Hansen / Andrea
            Make PAGE_OFFSET a config option

    numameminfo Martin Bligh / Keith Mannthey
            Expose NUMA meminfo information under /proc/meminfo.numa

    sched_tunables Robert Love
            Provide tunable parameters for the scheduler (+ NUMA scheduler)

    partial_objrmap Dave McCracken
            Object based rmap for filebacked pages.

    spinlock_inlining Andrew Morton & Martin J. Bligh
            Inline spinlocks for profiling. Made into a ugly config option by me.

    lockmeter John Hawkes / Hanna Linder
            Locking stats.

    sched_interactive Ingo Molnar
            Bugfix for interactive scheduler

    local_balance_exec Martin J. Bligh
            Modify balance_exec to use node-local queues when idle

    tcp_speedup Martin J. Bligh
            Speedup TCP (avoid double copy) as suggested by Linus

    disable preempt Martin J. Bligh
            I broke preempt somehow, temporarily disable it to stop accidents

    ppc64 pci fix Anton Blanchard
            Fix some ppc64 pci thing or other.

    per_node_idt Zwane Mwaikambo
            Per node IDT so we can do silly numbers of IO-APICs on NUMA-Q

    aiofix2 Mingming Cao
            fixed a bug in ioctx_alloc()

    config_irqbal Keith Mannthey
            Make irqbalance a config option

    percpu_real_loadavg Dave Hansen / Martin J. Bligh
            Tell me what the real load average is, and tell me per cpu.

    nolock Dave McCracken
            Nah, we don't like locks.

    mbind_part1 Matt Dobson
            Bind some memory for NUMA.

    mbind_part2 Matt Dobson
            Bind some more memory for NUMA.

    per_node_rss Matt Dobson
            Track which nodes tasks mem is on, so sched can be sensible.

    pfn_to_nid Martin J. Bligh
            Dance around the twisted rats nest of crap in i386 include.

    gfp_node_strict Dave Hansen
            Add a node strict binding as a gfp mask option

    page_lock William Lee Irwin
            Conditionally convert mapping->page_lock back to an rwlock

    -mjb Martin J. Bligh
            Add a tag to the makefile

    -
    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: Jamie Lokier: "Re: idt change in a running kernel? what locking?"

    Relevant Pages

    • 2.6.0-test3-mjb1
      ... else that stops things from irritating me. ... scheduler callers profiling ... send the line "unsubscribe linux-kernel" in ...
      (Linux-Kernel)
    • 2.6.0-test8-mjb1
      ... else that stops things from irritating me. ... scheduler callers profiling ... send the line "unsubscribe linux-kernel" in ...
      (Linux-Kernel)
    • 2.6.0-test9-mjb1
      ... else that stops things from irritating me. ... scheduler callers profiling ... send the line "unsubscribe linux-kernel" in ...
      (Linux-Kernel)
    • 2.6.0-test9-mjb3
      ... and anything else that stops things from irritating me. ... scheduler callers profiling ... send the line "unsubscribe linux-kernel" in ...
      (Linux-Kernel)
    • 2.6.0-test11-mjb1
      ... and anything else that stops things from irritating me. ... scheduler callers profiling ... send the line "unsubscribe linux-kernel" in ...
      (Linux-Kernel)