[lkml]   [2009]   [Sep]   [9]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
Messages in this thread
SubjectRe: [patch 3/4 -mm] flex_array: poison free elements
On Wed, 9 Sep 2009, Li Zefan wrote:

> > Gack, ftrace_page would probably be better off simply using the flex list
> > implementation I suggested which constructs a list from page->lru since it
> > doesn't appear to require indexing into the array, it simply requires an
> > iteration in all cases (or am I wrong?). Flex lists could easily be made
> > to support seqfiles.
> >
> Right, iteration is needed but indexing is not. I thought
> about flex list too, flex array has some limitations which
> make it hard to find a real user.

I agree, the cgroup pidarray could be converted into a flex list
implementation as well as ftrace_page, it appears. All the metadata can
be stored by overloading struct page and linking them in a dynamically
expanding list on page->lru.

I'm thinking Andrew will want to see some user of this library in 2.6.32
that cannot be served by a flex list type of implementation if we're to
keep it merged in the kernel.

 \ /
  Last update: 2009-09-09 06:19    [W:0.045 / U:48.220 seconds]
©2003-2018 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site