From mboxrd@z Thu Jan  1 00:00:00 1970
From: erik quanstrom <quanstro@quanstro.net>
Date: Fri, 25 Feb 2011 01:12:30 -0500
To: 9fans@9fans.net
Message-ID: <d1633b1172672456010aff5bbbfaf94f@ladd.quanstro.net>
In-Reply-To: <AANLkTim7W5rkgP00_M0AprNGpFhYgP=UUtWtOytJDKHO@mail.gmail.com>
References: <11da45046fa8267e7445128ed00724cd@ladd.quanstro.net>
	<AANLkTinuyGgidt_Lz9rQ8VyVtTVdnU2TABmbZG7NWt8g@mail.gmail.com>
	<24bb48f61c5eab87a133b82a9ef32474@coraid.com>
	<AANLkTim7W5rkgP00_M0AprNGpFhYgP=UUtWtOytJDKHO@mail.gmail.com>
MIME-Version: 1.0
Content-Type: text/plain; charset="US-ASCII"
Content-Transfer-Encoding: 7bit
Subject: Re: [9fans] sleep/wakeup bug?
Topicbox-Message-UUID: b46f71cc-ead6-11e9-9d60-3106f5b1d025

On Fri Feb 25 01:02:16 EST 2011, rsc@swtch.com wrote:
> you are fixating on the case you fixed and not
> the case i am describing.  in the case i am describing
> sleep returns before wakeup even starts.

what i'm saying is that with the code i posted
a) if sleep does not sleep, wakeup is not holding the rendez lock when sleep returns
b) if sleep does sleep, the same assertion holds

- erik