[lkml]   [2006]   [Apr]   [12]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
Messages in this thread
SubjectRe: CSCAN I/O scheduler for 2.6.10 kernel

I am attaching a final CSCAN scheduler patch for the kernel.
The earlier patch that I had posted had a bug in the
"cscan_merged_requests" function. This has been taken care of in the
attached patch. I would really appreciate if some one could help me
in conducting performance tests for the attached patch.

Many thanks for to all of you all for your inputs on this.

- Vishal

On 4/11/06, Jan Engelhardt <> wrote:
> >> >I am attaching the CSCAN scheduler patch for kernel.
> >>
> >> Thanks, I will try this.
> >>
> >> I have a question, why does not it use the kernel's rbtree implementation?
> >
> >It does, I dunno why you think it doesn't?
> My bad. I thought because a function is named
> static struct cscan_request *__rb_insert_request
> led me to believe this is the main insert function (when in fact it
> rb_link_node is). Maybe it should be just
> called "insert_request".

Every passing minute is another chance to turn it all around.
[unhandled content-type:application/octet-stream]
 \ /
  Last update: 2006-04-13 01:58    [from the cache]
©2003-2011 Jasper Spaans