You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@httpd.apache.org by "MATHIHALLI,MADHUSUDAN (HP-Cupertino,ex1)" <ma...@hp.com> on 2001/11/07 06:49:13 UTC
[PATCH]1. SHMEM (repost)
Hi,
As you might be aware, the current code in apr_shmem.c doesn't do
any memory management. I'd posted this patch sometime back, but I didn't get
any feedback regarding whether it's accepted / rejected (but for the code
structure comments from Justin - thanks)..
Here's a repost of the SHMEM patch. It'd be great if somebody could
pl. review and let me know you comments / suggestions.. This would be
required for the SHMHT / SHMCB patch that follows..
-Madhu
Re: [PATCH]1. SHMEM (repost)
Posted by Ryan Bloom <rb...@covalent.net>.
On Tuesday 06 November 2001 09:57 pm, Justin Erenkrantz wrote:
> On Wed, Nov 07, 2001 at 12:49:13AM -0500, MATHIHALLI,MADHUSUDAN (HP-Cupertino,ex1) wrote:
> > Hi,
> > As you might be aware, the current code in apr_shmem.c doesn't do
> > any memory management. I'd posted this patch sometime back, but I didn't
> > get any feedback regarding whether it's accepted / rejected (but for the
> > code structure comments from Justin - thanks)..
> > Here's a repost of the SHMEM patch. It'd be great if somebody could
> > pl. review and let me know you comments / suggestions.. This would be
> > required for the SHMHT / SHMCB patch that follows..
>
> Can you please post the SHMHT / SHMCB patches? I'm still not sold on
> why we need the shared memory management at all. If you can provide a
> good use for this (i.e. whatever SHMHT is), that may provide some
> impetus for adding this.
>
> I guess I'd also like to know what value this shared memory cache adds
> for mod_ssl. I'm not at all clear on how shared memory would make
> SSL any better - as SSL is only a connection-oriented protocol and
> shouldn't use shared memory. None of the MPMs in httpd migrate
> connections between processes (I don't see this changing anytime soon).
> Since SSL works on the connection level, I'm not sure what this is.
> So, I guess some explanation here would do wonders. -- justin
Two problems with this paragraph. The cache is a session cache, so
that SSL doesn't need to redo everything everytime somebody connects
with an SSL request.
And, the Perchild MPM does migrate connections between processes.
Ryan
______________________________________________________________
Ryan Bloom rbb@apache.org
Covalent Technologies rbb@covalent.net
--------------------------------------------------------------
Re: [PATCH]1. SHMEM (repost)
Posted by Justin Erenkrantz <je...@ebuilt.com>.
On Wed, Nov 07, 2001 at 12:49:13AM -0500, MATHIHALLI,MADHUSUDAN (HP-Cupertino,ex1) wrote:
> Hi,
> As you might be aware, the current code in apr_shmem.c doesn't do
> any memory management. I'd posted this patch sometime back, but I didn't get
> any feedback regarding whether it's accepted / rejected (but for the code
> structure comments from Justin - thanks)..
> Here's a repost of the SHMEM patch. It'd be great if somebody could
> pl. review and let me know you comments / suggestions.. This would be
> required for the SHMHT / SHMCB patch that follows..
Can you please post the SHMHT / SHMCB patches? I'm still not sold on
why we need the shared memory management at all. If you can provide a
good use for this (i.e. whatever SHMHT is), that may provide some
impetus for adding this.
I guess I'd also like to know what value this shared memory cache adds
for mod_ssl. I'm not at all clear on how shared memory would make
SSL any better - as SSL is only a connection-oriented protocol and
shouldn't use shared memory. None of the MPMs in httpd migrate
connections between processes (I don't see this changing anytime soon).
Since SSL works on the connection level, I'm not sure what this is.
So, I guess some explanation here would do wonders. -- justin