httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Ryan Bloom <>
Subject Re: Concatenation of small buckets in core_output_filter Re: cvs commit: httpd-2.0/server core.c
Date Mon, 26 Nov 2001 17:55:46 GMT
On Monday 26 November 2001 09:46 am, Brian Pane wrote:
> Ryan Bloom wrote:
> >On Monday 26 November 2001 09:39 am, Brian Pane wrote:
> >>Ryan Bloom wrote:
> >>>On Monday 26 November 2001 09:28 am, Brian Pane wrote:
> >>
> >>[...]
> >>
> >>>>* If we end up in the concatenation again during the foreach loop
> >>>> through the brigade, add the small buckets to the end of the
> >>>> previously allocated heap bucket.  If the heap bucket size is
> >>>> about to exceed 8KB, stop.
> >>>
> >>>This should be done by moving the first bucket into a separate brigade
> >>>and using the brigade_write functions again.
> >>
> >>Thanks for the pointer; using apr_brigade_write for this will make the
> >>code a lot cleaner.  I'll implement this later today if nobody beats
> >>me to it.
> >
> >That is what is being done now.
> But in the current implemention, we end up allocating new space
> each time the buckets are compacted, because we start with an
> empty brigade, right?  So even if the first bucket in the list
> to be compacted happens to have been created by a previous
> apr_brigade_write (because we end up compacting twice on the
> same request, due to a long stream of small buckets produced by
> a CGI or mod_include), we can't take advantage of the extra space
> within that bucket.  Or am I missing something?

You've got it right, but I want to be sure that you know we are currently using
those functions. We just need to split the brigade correctly so that we have
a brigade that ends with the previously allocated bucket.


Ryan Bloom
Covalent Technologies

View raw message