linux-mips-fnet
[Top] [All Lists]

Re: pause()

To: Florian Lohoff <flo@rfc822.org>
Subject: Re: pause()
From: Ralf Baechle <ralf@oss.sgi.com>
Date: Sun, 2 Apr 2000 02:34:25 +0200
Cc: linux@engr.sgi.com, linux-mips@fnet.fr, linux-mips@vger.rutgers.edu, aj@suse.de
In-reply-to: <20000331195303.B20241@paradigm.rfc822.org>; from flo@rfc822.org on Fri, Mar 31, 2000 at 07:53:03PM +0200
References: <20000331151555.A5911@uni-koblenz.de> <20000331195303.B20241@paradigm.rfc822.org>
On Fri, Mar 31, 2000 at 07:53:03PM +0200, Florian Lohoff wrote:

> On Fri, Mar 31, 2000 at 03:15:55PM +0200, Ralf Baechle wrote:
> > I just found another brainfart in the libc / kernel interface.  In the
> > believe libc wouldn't use the pause(2) syscall any longer I removed it.
> > This makes a number of programs like screen burn all CPU they can get.
> > I'll provide two fixes, one to libc and a second for the kernel and
> > either one will be sufficient.
> 
> Could this also be the cause of "top" refreshing the screen as fast
> as it can ? I noticed that when i rebuild the debian procps package
> and tried top ... It then consumes the cpu it gets ...

I haven't tried myself but sounds very probable.  The sympthon is that
an strace will show large numbers of pause(2) syscalls which all return
with -ENOSYS as error.

  Ralf

<Prev in Thread] Current Thread [Next in Thread>