Re: 1352 NUL bytes at the end of a page? (was Re: Assertion `s && s->tree' failed: The saga continues.)

From: Steven Cole (elenstev_at_mesatop.com)
Date: 05/16/04

  • Next message: Jose Fernandez: "10 cargas por el precio de 1" 
    To: Andrew Morton <akpm@osdl.org>
Date:	Sun, 16 May 2004 07:49:24 -0600

    On Sunday 16 May 2004 04:01 am, Andrew Morton wrote:
    > Steven Cole <elenstev@mesatop.com> wrote:
    > >
    > > The script only reported one
    > > iteration finished, while I got it to do 36 iterations over several hours earlier
    > > today (with a 2.6.3-4mdk vendor kernel), so I'm going to add some timing
    > > tests to the script to see if things are really slowing down with current kernels,
    > > or if it's just my worried imaginings.
    >
    > I did a bit of testing on a 256MB laptop with a fairly slow disk, ext3.
    > Three iterations of the test took:
    >
    > 2.6.6: 1055.53s user 327.14s system 32% cpu 1:10:06.71 total
    >
    > 2.4.27-pre2: 1042.03s user 307.21s system 32% cpu 1:09:46.00 total
    >
    > 2.6.3: 1053.23s user 326.16s system 27% cpu 1:22:07.24 total
    >
    >
    > So there's nothing particularly wild there. It's possible I guess that the
    > 2.6 VM is very sucky but something else made up for it - possibly the
    > anticipatory scheduler but more likely the Orlov allocator.
    >
    > You're using reiserfs, yes?

    Yes, but I also have a BK repo on an ext3 fs which saw the exact
    same initial problem. Well, almost in that I got the Assertion `s && s->tree' failed
    message, but didn't preserve the possibly null-containing s.ChangeSet file.

    >
    > Are you sure the IDE disks are in DMA mode?
    >
    >
    For the 2.6.3 tests, yes.
    For the overnight 2.6.6 tests, apparently not, but it is now.
    Without DMA, the clone took about 22 minutes, but even with
    DMA back on, 2.6.6-current is much slower than 2.6.3-4mdk.

    It's hard to make your wheels fall off when putting around the
    race track at 2/3 speed. If I get time, I'll try the patch from Nick
    and/or a kernel from around 4/15 when I first saw this problem,
    but the weather's nice outside, so it may be later this evening
    before I return to testing.

    /dev/hda:
     multcount = 0 (off)
     IO_support = 0 (default 16-bit)
     unmaskirq = 0 (off)
     using_dma = 1 (on)
     keepsettings = 0 (off)
     readonly = 0 (off)
     readahead = 256 (on)
     geometry = 65535/16/63, sectors = 78125000, start = 0

    ----------------------------
    2.6.3-4mdk:

    time bk clone -qlr40514130hBbvgP4CvwEVEu27oxm46w testing-2.6 foo
    287.93user 59.33system 11:14.03elapsed 51%CPU (0avgtext+0avgdata 0maxresident)k
    0inputs+0outputs (41264major+523595minor)pagefaults 0swaps

    (cd foo; time bk pull -q)
    397.05user 184.05system 16:38.35elapsed 58%CPU (0avgtext+0avgdata 0maxresident)k
    0inputs+0outputs (30287major+2115127minor)pagefaults 0swaps
    -----------------------------

    -----------------------------
    2.6.6-current:

    time bk clone -qlr40514130hBbvgP4CvwEVEu27oxm46w testing-2.6 foo
    290.48user 96.76system 15:00.85elapsed 42%CPU (0avgtext+0avgdata 0maxresident)k
    0inputs+0outputs (50893major+515118minor)pagefaults 0swaps

    (cd foo; time bk pull -q)
    402.74user 254.98system 23:25.43elapsed 46%CPU (0avgtext+0avgdata 0maxresident)k
    0inputs+0outputs (60130major+2089908minor)pagefaults 0swaps

    ------------------------------

    Steven
    -
    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: Jose Fernandez: "10 cargas por el precio de 1"

    Relevant Pages

    • Re: super.clone() puzzlement
      ... If your design can make do with just copy-ctors and forgo clone(), ... private Foo { ... Even then you could just dump the generics and leave clone to return Object and callers to cast. ... The lack of multiple inheritance in Java makes it tricky but you can make a class that contains an ArrayListand has getInputStream and getOutputStream methods that furnish inner class instances, one which appends to the ArrayList at one end, and the other that consumes from the other end and blocks if need be. ...
      (comp.lang.java.programmer)
    • Re: [PHP] Passing variables between pages
      ... You always have the option to write one huge script for any application. ... we use sessions to pass variables between scripts -- but we could ... So just defining $foo saying include 'steptwo.php' ... I do use this technique inside of my templating class though. ...
      (php.general)
    • Re: Redirect file output into variables
      ... Hmm, i would use a safe interpreter for that..., as you say limited ... interp alias foo puts buffered_puts ... set script "script you want to execute in a safe environment" ...
      (comp.lang.tcl)
    • Re: A subst-antial question
      ... Suppose I have a script: ... I know I have to substitute variables locally before I can set them ... stuff with $foo in it ... An alternative, in Tcl 8.5 (or 8.4 with the dict package), would be to not perform the substitutions locally, but instead to snapshot the environment of the script as a dict, and then send that along with the code to be evaluated remotely using: ...
      (comp.lang.tcl)
    • Re: Restricting users to their own home directories / not letting users view other users files...?
      ... Write a script to copy the user's files over on a schedule...? ... $ cat foo ... That way they can edit via the browser or their favorite blog posting ...
      (freebsd-questions)