httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Bill Stoddard <b...@wstoddard.com>
Subject Re: cvs commit: httpd-2.0/server/mpm/winnt child.c mpm_winnt.c mpm_winnt.h
Date Thu, 13 Mar 2003 19:44:37 GMT
William A. Rowe, Jr. wrote:
> At 12:57 PM 3/13/2003, Bill Stoddard wrote:
> 
>>ake@apache.org wrote:
>>
>>>ake         2003/03/04 14:15:52
>>> Modified:    .        CHANGES
>>>              server/mpm/winnt child.c mpm_winnt.c mpm_winnt.h
>>> Log:
>>> Added the WindowsSocketsWorkaroud directive for Windows NT/2000/XP
>>> to work around problems with certain VPN and Firewall products that
>>> have buggy AcceptEx implementations
>>> 
>>> Revision  Changes    Path
>>> 1.1103    +5 -0      httpd-2.0/CHANGES
>>> 
>>
>><snip>
>>
>>>  
>>>  static const command_rec winnt_cmds[] = {
>>>  LISTEN_COMMANDS,
>>> @@ -224,6 +243,9 @@
>>>    "Number of threads each child creates" ),
>>>  AP_INIT_TAKE1("ThreadLimit", set_thread_limit, NULL, RSRC_CONF,
>>>    "Maximum worker threads in a server for this run of Apache"),
>>> +AP_INIT_TAKE1("WindowsSocketsWorkaround", set_sockets_workaround, NULL, RSRC_CONF,
>>> +  "Set \"on\" to work around buggy Winsock provider implementations of certain
VPN or Firewall software"),
>>> +
>>>  { NULL }
>>>  };
>>
>>Rather than WindowsSocketsWorkaround, why not WinUseWinsock1 or ??. It would be better
I think if the directive somehow indicated exactly what it was doing (causing the winnt mpm
to use the select/accept winsock1 calls rather than AcceptEx, a winsock2 call).
> 
> 
> That would be a misnomer - since our handle inheritance requires winsock2.
> 
> What about WindowsUseAcceptEx on|off?  That's really descriptive of what 
> the workaround does.  Or, we could call it WindowsFastSockets on|off,
> which is the effect of the workaround.
> 
> I was also looking at this entire patch - it seems silly to retest for both the
> version of windows and this flag throughout the code.  Why not simply
> initialize it in the register_hooks() call based on the OS version?  Then let
> the directive override its default value.  

That might be cool. Post a patch and I'll review it.

> We should prevent Win9x users
> from enabling the flag, however :-)
> 
> Bill 
> 

WindowsUseAcceptEx is much better I think. Since it is on by default on 
systems which support it. Another (better?) suggestion: 
WindowsDisableAcceptEx (with no arguments)?

Bill



Mime
View raw message