Hi Bill,
        Thanks for picking up that patch.. Another update that I had from
the HPUX TCP team was that they considering changing the return code from
ENOBUFS to ECONNRESET (which is what it is on most platforms) - but I'm not
sure when are they going to do that.. So, for the time being, the patch
holds good.

-Madhu

-----Original Message-----
From: Bill Stoddard [mailto:[EMAIL PROTECTED]]
Sent: Thursday, January 17, 2002 1:14 PM
To: [EMAIL PROTECTED]
Subject: Re: [PATCH - Apache 1.3] ENOBUFS error on HPUX


+1

My opinion is that we should not make the HPUX11 specific check but both
Jeff and Doug
believe otherwise, so I will commit the patch as is.

Thanks,
Bill

----- Original Message -----
From: "MATHIHALLI,MADHUSUDAN (HP-Cupertino,ex1)"
<[EMAIL PROTECTED]>
To: <[EMAIL PROTECTED]>
Sent: Friday, December 14, 2001 7:35 PM
Subject: [PATCH - Apache 1.3] ENOBUFS error on HPUX


> Hi,
> This was a small fix that we had done for Apache 1.3 on HPUX some
> time back.. (Thanks Doug, for pointing it out). Can somebody pl. review
and
> commit the foll. piece of code.
>
> -Madhu
>
> Index: http_main.c
> ===================================================================
> RCS file: /home/cvspublic/apache-1.3/src/main/http_main.c,v
> retrieving revision 1.562
> diff -u -r1.562 http_main.c
> --- http_main.c 2001/11/26 17:26:54     1.562
> +++ http_main.c 2001/12/15 00:20:41
> @@ -4344,6 +4344,19 @@
>                  * to just exit in most cases.
>                  */
>                  switch (errno) {
> +
> +#if defined(HPUX11) && defined(ENOBUFS)
> +                    /* On HPUX 11.x, the 'ENOBUFS, No buffer space
> available'
> +                     * error occures because the accept() cannot
complete.
> +                     * You will not see ENOBUFS at 10.20 because the
kernel
> +                     * hides any occurrence from being returned from user
> space.
> +                     * BUt, ENOBUFS at 11.0 TCP/IP is quite possible, and
> could
> +                     * occur intermittently. As a work-around, we are
going
> to
> +                     * ingnore ENOBUFS.
> +                     */
> +                case ENOBUFS:
> +#endif
> +
>  #ifdef EPROTO
>                     /* EPROTO on certain older kernels really means
>                      * ECONNABORTED, so we need to ignore it for them.
>

Reply via email to