Skip site navigation (1)Skip section navigation (2)
Date:      Tue, 03 Apr 2001 12:50:24 -0700 (PDT)
From:      John Baldwin <jhb@FreeBSD.org>
To:        Alfred Perlstein <alfred@FreeBSD.org>
Cc:        cvs-committers@FreeBSD.org, cvs-all@FreeBSD.org
Subject:   Re: cvs commit: src/sys/sys mbuf.h src/sys/kern uipc_mbuf.c
Message-ID:  <XFMail.010403125024.jhb@FreeBSD.org>
In-Reply-To: <20010403095120.E813@fw.wintelcom.net>

next in thread | previous in thread | raw e-mail | index | archive | help

On 03-Apr-01 Alfred Perlstein wrote:
> * John Baldwin <jhb@FreeBSD.org> [010403 09:19] wrote:
>> 
>> On 03-Apr-01 Alfred Perlstein wrote:
>> > alfred      2001/04/02 20:15:12 PDT
>> > 
>> >   Modified files:
>> >     sys/sys              mbuf.h 
>> >     sys/kern             uipc_mbuf.c 
>> >   Log:
> 
> ...
> 
>> >   There's also some if (1) code that should check if the "how"
>> >   operation specifies blocking/non-blocking behavior, we _could_ make
>> >   it so that we hold onto the mutex through calls into kmem_alloc
>> >   when non-blocking requests are made, but for safety reasons we
>> >   currently drop and reaquire the mutex around the calls.
>> >   
>> >   Also, note that calling kmem_alloc is rare and only happens during
>> >   a shortage so drop/re-getting the mutex will not be a common
>> >   occurance.
>> 
>> It is perfectly safe to drop the lock so long as the mbuf subsystem is in a
>> consistent state before calling kmem_alloc().
> 
> Yes, I know that, not dropping the mutex would happen when how == M_NOWAIT
> but the code to do that is commented out:
> 
> if (1 /* XXX: how == M_NOWAIT */)
>   mtx_unlock(&mbuf_mtx);
> kmem_alloc(...)
> if (1 /* XXX: how == M_NOWAIT */)
>   mtx_lock(&mbuf_mtx);
> 
> Basically, if kmem_alloc isn't going to sleep then there's no reason
> I can see to drop the mutex except for the lock order revesal problem
> with Giant being required by kmem_alloc.  (which you have patches for)

I have patches to acquire Giant inside of kmem_alloc, I do not have patches to
deal with the lock order problem.

>> >   Remove some #define's that seemed to obfuscate the code to me.
>> >   
>> >   Remove an extranious comment.
>> >   
>> >   Remove an XXX, including mutex.h isn't a crime.
>> 
>> Yes it is, and is going away soon. :)  It will also be leaving other headers
>> such as sys/ucred.h, sys/resourcevar.h, etc.
> 
> Yeah, Bruce slapped me around for this, are you planning on fixing it?
> Or should I proceed to remove the includes?

It's on the SMP todo list.  Bruce has sent me a script to do lots of header
include checks and it apparently spits out 9000 lines of errors right now. :) 
I'll see if I can't get some of thise paired down at some point in time.

-- 

John Baldwin <jhb@FreeBSD.org> -- http://www.FreeBSD.org/~jhb/
PGP Key: http://www.Baldwin.cx/~john/pgpkey.asc
"Power Users Use the Power to Serve!"  -  http://www.FreeBSD.org/

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?XFMail.010403125024.jhb>