lkml.org 
[lkml]   [2009]   [Apr]   [23]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
    Patch in this message
    /
    Date
    From
    SubjectRe: [PATCH 02/22] Do not sanity check order in the fast path
    On Wed, 22 Apr 2009 18:11:51 +0100
    Mel Gorman <mel@csn.ul.ie> wrote:

    > > I depend on the allocator to tell me when I've fed it too high of an
    > > order. If we really need this, perhaps we should do an audit and then
    > > add a WARN_ON() for a few releases to catch the stragglers.
    > >
    >
    > I consider it buggy to ask for something so large that you always end up
    > with the worst option - vmalloc().

    Nevertheless, it's a pretty common pattern for initialisation code all
    over the kernel to do

    while (allocate(huge_amount) == NULL)
    huge_amount /= 2;

    and the proposed change will convert that from "works" to "either goes
    BUG or mysteriously overindexes zone->free_area[] in
    __rmqueue_smallest()". The latter of which is really nasty.

    > How about leaving it as a VM_BUG_ON
    > to get as many reports as possible on who is depending on this odd
    > behaviour?

    That would be quite disruptive. Even emitting a trace for each call
    would be irritating. How's about this:

    --- a/mm/page_alloc.c~page-allocator-do-not-sanity-check-order-in-the-fast-path-fix
    +++ a/mm/page_alloc.c
    @@ -1405,7 +1405,8 @@ get_page_from_freelist(gfp_t gfp_mask, n

    classzone_idx = zone_idx(preferred_zone);

    - VM_BUG_ON(order >= MAX_ORDER);
    + if (WARN_ON_ONCE(order >= MAX_ORDER))
    + return NULL;

    zonelist_scan:
    /*
    _

    and then we revisit later?


    \
     
     \ /
      Last update: 2009-04-24 00:53    [W:0.046 / U:30.792 seconds]
    ©2003-2016 Jasper Spaans. hosted at Digital OceanAdvertise on this site