This is the mail archive of the glibc-bugs@sourceware.org mailing list for the glibc project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

[Bug libc/11687] Allow customizing of __FD_SETSIZE


------- Additional Comments From mtk dot manpages at gmail dot com  2010-06-21 08:52 -------
(In reply to comment #1)
> You cannot simply change FD_SETSIZE like this and expect anything to work. The
> FD_SETSIZE value is part of the ABI and depends mainly on the kernel - if you
> use larger fdset size, the syscall will not work anyway. 

Petr, Did you test this? I've happily raised FD_SETSIZE by modifying glibc
headers, and used set sizes up to 16k. No kernel mod required.


> You would have to hack
> your kernel too, and then you are completely on your own anyway; other strange
> problems can also pop up due to different structure sizes in already compiled
> programs.
> 
> In short, FD_SETSIZE is not remotely a variable that would be
> runtime-adjustable, and it's not defined as a fixed value just for the sake of
> it. If you need to handle more fds, you will need different methods like poll()
> or epoll().

Elsewhere, the reporter explained why that's not possible for him. People
porting legacy applications do run into this problem. There seems no obvious
reason for glibc not to allow FD_SETSIZE to conditionally raised in the headers.
I posted exactly the same request years ago, and got the same response to my bug
report from the maintainer. But, AFAICS, there is no problem.

-- 


http://sourceware.org/bugzilla/show_bug.cgi?id=11687

------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]