The patch looks decent, but can be done better.
It would be better if there was a flag set on the filedescriptor, than
to call the pending_method all the time. This part of the code is
accessed very frequently and is already a major bottleneck...
Also, you still need to poll/select to keep a good balance in the
processing power given to the filedescriptors. If not, a single SSL
session could theoretically monopolize the usage. Skipping the
poll/select only because there is some pending filedescriptor is no good
as filedescriptors having data in their kernel buffers is as pending as
the ones having data in their SSL buffers.. (only the notification
capabilities differ).
I'll work on a new patch. Should be available within an hour.
Regards
Henrik
Noel Burton-Krahn wrote:
>
> Oops! Here's the real patch (with both pending_method() and X-Forwarded-URI and
> pending_menthod()).
>
> I tried to make the pending_method() test fit in with the logic in
> comm_select(). pending_method() is tested only if the files read
> status would be checked by poll() or select(). It shouldn't be
> checked if the file is disabled.
Received on Fri Oct 19 2001 - 06:19:42 MDT
This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:14:33 MST