Date: Tue, 22 Mar 2005 09:08:00 +0100
From: Ingo Molnar <mingo@elte.hu>
To: Kyle Moffett <mrmacman_g4@mac.com>
Cc: Herbert Xu <herbert@gondor.apana.org.au>, rusty@au1.ibm.com,
       bhuey@lnxw.com, gh@us.ibm.com, tgall@us.ibm.com,
       jim.houston@comcast.net, linux-kernel@vger.kernel.org,
       torvalds@osdl.org, akpm@osdl.org, paulmck@us.ibm.com,
       manfred@colorfullife.com, shemminger@osdl.org, dipankar@in.ibm.com
Subject: Re: Real-Time Preemption and RCU
Message-ID: <20050322080800.GA9497@elte.hu>
References: <20050318160229.GC25485@elte.hu> <E1DCPut-0005XI-00@gondolin.me.apana.org.au> <20050319163128.GB28958@elte.hu> <24e1c77a699dcab771a0280df69d44eb@mac.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <24e1c77a699dcab771a0280df69d44eb@mac.com>
User-Agent: Mutt/1.4.1i
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 1031
Lines: 25


* Kyle Moffett <mrmacman_g4@mac.com> wrote:

> One solution I can think of, although it bloats memory usage for
> many-way boxen, is to just have a table in the rwlock with one entry
> per cpu.  Each CPU would get one concurrent reader, others would need
> to sleep

yes, it bloats memory usage, and complicates PI quite significantly. 
We've been there and have done something similar to that in earlier -RT
patches - 64 'owner' entries in the lock already caused problems on 8K
stacks. 32 seemed to work but caused quite some bloat in data structure
sizes.

i'd rather live with some scalability bottleneck for now, in exchange of
a wastly easier to handle design. We can still complicate things later
on for better scalability, once all the other issues have been sorted
out.

	Ingo
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/