HFS Patch 16 for GNU Parted 1.6.15

From: Guillaume Knispel (k_guillaume_at_libertysurf.fr)
Date: 09/30/04

  • Next message: Jim Hart: "weirdness compiling firefox 1.0" 
    Date: 30 Sep 2004 11:54:04 -0700

    Hi,

    I've released HFS Patch 16 for Parted 1.6.15 :
    http://xilun.nerim.net/Projet/Parted/parted-1.6.15-hfs-16.patch.gz
    ( patched tbz2 : http://xilun.nerim.net/Projet/Parted/parted-1.6.15-hfs-16.tar.bz2
      staticaly linked for ppc :
    http://xilun.nerim.net/Projet/Parted/parted-hfs16-ppc.static )
    Use the '-T' option of patch to apply it, otherwise you will uselessly
    rebuild the whole autoconf system of Parted.

    I've identified the bottleneck that induces poor performance. Because
    resolving it means modifying portion of Parted that aren't related to
    HFS at all, this version of the HFS Patch doesn't solve the issue at
    all. But I've written a quick hack that desactivated the paranoid
    block devices cache flush (bottleneck), to be used until Parted
    provide some official non paranoid sync :) I think this is still
    safe(*) for HFS resizing (I've no idea weither it's safe or not for
    other FS resizing). The patch that does that is here :
    http://xilun.nerim.net/Projet/Parted/Experimental/parted-1.6.15-sync.patch
    The static binary version I builded include this patch.
    The performance improvement I get on my test system is 29x...

    Bugfixes in this version :
    - The catalog file was written instead of the attributes files, in
    case the attributes files exists (this is very unlikely to happen)
    (this bug exists only in patch 15, but in that version I also
    introduced a warning about the experimental state of the code in case
    an attributes file exists, so at least people were warned :)
    - An error condition is now correctly detected when reading the bad
    blocks list (this can only happen on incorrect filesystem anyway)
    - The return value of every memory allocation is now correctly checked
    ( I forgot 4 of them...)

    Remember that Linux 2.4 is still incompatible with resized volumes
    that are also used under Os X.

    Enjoy,
    Guillaume Knispel 

    ---
(*) Safe means you can't loose data in case of interruption of Parted
during the resize process, like SIGKILL or power failure. (though I
DON'T provide any warranty about that).
  • Next message: Jim Hart: "weirdness compiling firefox 1.0"

    Relevant Pages

    • Re: coretemp - take3
      ... CONFIG_CPU_HOTPLUG y/n ... versions of the functions (safe and unsafe). ... I attached the patch here. ... #ifdef CONFIG_PARAVIRT ...
      (Debian-User)
    • Re: [PATCH] allow root to modify raw scsi command permissions list
      ... > My patch leaves the defaults as what are currently in the kernel. ... Yes but what I wanted to say the filter currently in the kernel is not ... purpose of giving read/write permissions in this case. ... The starting point must be safe and it can be relaxed. ...
      (Linux-Kernel)
    • Re: fuel tank leak - epoxy question
      ... Since YOU choose to effect such a feeble, potentially dangerous patch ... repair, and continue to drive the vehicle, I could only assume that YOU had ... A car with a leaking/patched/weak gas tank is neither incredibly reliable ... nor is it incredibly safe. ...
      (rec.autos.tech)
    • Re: CONFIG_VFAT_FS_DUALNAMES regressions
      ... patch used a space followed by a nul). ... the probability is much easier to calculate as ... only writes the short-names it's also safe. ...
      (Linux-Kernel)
    • HFS(+) resize patch for GNU Parted.
      ... My 9th patch for HFS and HFS+ shrink support is ... To test this patch: download the parted-1.6.6 sources, ...
      (comp.os.linux.powerpc)