Skip site navigation (1)Skip section navigation (2)
Date:      Sat, 23 Feb 2002 12:46:10 -0800
From:      Alfred Perlstein <bright@mu.org>
To:        dillon@FreeBSD.org
Cc:        Seigo Tanimura <tanimura@FreeBSD.org>, cvs-committers@FreeBSD.org, cvs-all@FreeBSD.org
Subject:   Re: cvs commit: src/sys/coda coda_venus.c src/sys/compat/linprocfs
Message-ID:  <20020223204610.GG80761@elvis.mu.org>
In-Reply-To: <200202232030.g1NKU4m87533@freefall.freebsd.org>
References:  <200202232030.g1NKU4m87533@freefall.freebsd.org>

next in thread | previous in thread | raw e-mail | index | archive | help
* dillon@FreeBSD.org <dillon@FreeBSD.org> [020223 12:30] wrote:
> 
> :  New locks are:
> :  
> :  - pgrpsess_lock which locks the whole pgrps and sessions,
> :  - pg_mtx which protects the pgrp members, and
> :  - s_mtx which protects the session members.
> :...
> 
>     I see you are using mtx_lock_giant().  Thanks!  That is precisely the
>     use it is designed for!
> 
>     --
> 
>     Seigo, it looks like s_mtx can be turned into a pool mutex.  Do you
>     mind if I make this change?
> 
>     -
> 
>     I would also like to document some of the routines.  For example,
>     funsetownlst() has some rather complex locking requirements which
>     are not documented at all at the base of the routine.  It would also
>     probably be a good idea for the two callers of this routine to assert
>     the sign of sio_pgid before calling funsetownlst().  pgfind() is also
>     only partially documented.  And so forth.
> 
>     I also noticed that you are leaving PGRPSESS_SLOCK()'d while calling
>     ktrsetchildren() and ktrops().  This is the only place where 
>     PGRPSESS_SLOCK() is used outside of its domain and if it is at all
>     possible to remove it I would remove it.  If that isn't possible, then
>     both ktr*() routines *AND* VREF() and vrele() need to be documented in
>     regards to the process group session potentially being locked on call
>     (which I would be happy to do as well).

Matt, any chance you can co-ordinate with jhb (or not) to add
functionality to witness such that it can detect when a mutex
is aquired while holding a pool mutex?

-Alfred

To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe cvs-all" in the body of the message




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20020223204610.GG80761>