netdev
[Top] [All Lists]

Re: tx_timeout and timer serialisation

To: "David S. Miller" <davem@xxxxxxxxxx>
Subject: Re: tx_timeout and timer serialisation
From: Donald Becker <becker@xxxxxxxxx>
Date: Sat, 20 May 2000 04:31:33 -0400 (EDT)
Cc: netdev@xxxxxxxxxxx
In-reply-to: <200005200731.AAA19020@xxxxxxxxxxxxxxx>
Sender: owner-netdev@xxxxxxxxxxx
On Sat, 20 May 2000, David S. Miller wrote:

> del_timer deschedules a timer, no more, no less.

That's a little too trite of an answer to be useful.

Immediately after the return of del_timer(&timer), I expect
 - kfree(timer.data) is safe
 - the timer.function() code may be immediately be 'rmmod'ed

Specifically, I expect that if the timer.function() will not be
running on another processor when del_timer(&timer) returns.  If this
doesn't hold true, the timer.function() might still be in the function entry
phase and entire timer.function() could be run after del_timer() has
nominally finished.  And if that's acceptable behavior, I have a new version
of del_timer() that's very small and fast ;->.


Donald Becker                           becker@xxxxxxxxx
Scyld Computing Corporation
410 Severn Ave. Suite 210
Annapolis MD 21403



<Prev in Thread] Current Thread [Next in Thread>