9fans archive / 2006 / 09 / 356 prev next
From: Russ Cox <rsc@swt...>
Subject: drawterm cpunote
Date: Mon, 25 Sep 2006 07:12:48 -0400
> this code in kern/devcons.c
>
> case Qcpunote:
> sleep(&up->sleep, return0, nil);
>
> does it hang a blockingslave in exportfs, pending a read?
> for drawterm, the net effect is that rmtnoteproc in
> cpu.c gets an error when drawterm goes away which
> then kills rc. correct?
sounds right to me, though i'm not 100%
sure that the failed read on cpunote is what
kills everything off.
in plan 9 the cpu client relays notes via that
file, but since drawterm doesn't have notes
to relay, it just makes the reads block until it exits.
russ