Coda File System

Re: mounting venus process not waited for

From: Jan Harkes <>
Date: Thu, 5 Jun 2003 11:00:40 -0400
On Thu, Jun 05, 2003 at 08:32:32AM -0400, Greg Troxel wrote:
>   Linux is using signal(SIGCHLD, SIG_IGN) which allows us to do nothing
>   at all. This does not seem to be a POSIX standard. Does BSD happen to
>   have this feature?
> The default behavior for SIGCHLD is to discard the signal (says 'man 7
> signal' on NetBSD 1.6.1).  So that call looks redundant.  Does
> explicitly asking to ignore SIGCHLD mean that the process fully exits
> without the parent wait()ing, and avoids the zombie state?  That
> sounds, at first blush, like a POSIX violation.

It seems to be a SysV-ism that might have been taken up in the Single
Unix Standard.

But most references that google found seem to indicate that it is not
very portable.

Received on 2003-06-05 11:03:35