Re: HOW TO INCREASE FIFO SIZE LIMIT?

From: Michael Kerrisk (michael-dot-kerrisk-at-gmx-dot-net_at_NOSPAM.COM)
Date: 01/20/04 

Date: Tue, 20 Jan 2004 09:25:42 +0100

On 19 Jan 2004 16:16:35 -0800, sandeep_onsite@yahoo.com (Sandeep)
wrote:

>Hi,
>
>Does anybody know how to increase FIFO Max. size limit on RedHat 9.0?
>My current limit is 8*512 bytes, which is as shown by "ulimit -p" command.
>I tried to increase it to 20*512 bytes using command: "ulimit -p 20".
>Got this error:
>
>ulimit: pipe size: cannot modify limit: Invalid argument
>
>Let me know is there is any other way to change the Max. size of Fifo/Pipe?

Hello Sandeep,

There is no way to do this with a standard kernel (the bash man page
explicitly notes that you can't set this limit). There are patches
around that allow you to rebuild the kernel with a different value for
this limit. However, doing this is almost certainly a bad idea: an
application built with such a kernel dependency won't be portable.
Furthermore, looking for this facility is usually an indicator of bad
application design: In general, an application should never need to be
concerned with exactly what the capacity of a pipe is -- if we want to
prevent the writer process from blocking, the process reading from the
pipe should be designed to read data as soon as it available.

It might be a good idea if you explain what you want to do and why.
Then people might be able to make suggestions about better ways of
solving the problem.

Cheers,

Michael

Relevant Pages

  • Re: pppd crashes, was: kde-freebsd
    ... User PPP is very easy to use, Kernel PPP is not. ... It appears to me that PPP is the more normal way on FreeBSD, whereas, in my own experience Linux, prefer PPPD. ... Over time FreeBSD and Linux drifted apart on this design issue, and it became something of a characteristic of BSD, perhaps that is why Kernel PPP became less well maintained ... Regarding the various comments by Michael Nottebrock, Firstly: The bug you mentioned I have not experienced. ...
    (freebsd-stable)
  • Re: A Layered Kernel: Proposal
    ... (I can't believe that a kernel programmer will not ... So give us a design. ... layer, the SCSI layer, the network layer, etc.), but it is always ...
    (Linux-Kernel)
  • Re: [PROPOSAL/PATCH] Fortuna PRNG in /dev/random
    ... > in random.c with the Fortuna PRNG designed by Ferguson and Schneier (Practical ... The kernel will break if CONFIG_CRYPTO is false ... don't want crypto, then you don't want secure random numbers." ... design a system that is closer to "true randomness" as possible. ...
    (Linux-Kernel)
  • Re: OpenGL-based framebuffer concepts
    ... How is a printk generated in the kernel going to find this ... User space is messed up ... Your display process has ... Why don't you write up a comprehensive design for your system and post ...
    (Linux-Kernel)
  • Re: Allocating kernel memory
    ... > design of the kernel you are using. ... memory available within the kernel does not imply I do not understand ... > You don't understand the architecture you are ...
    (comp.os.linux.development.system)
Loading