Date: Wed, 2 Jun 2004 01:57:40 +0400
From: Andrew Zabolotny <zap@homelink.ru>
To: Jeff Garzik <jgarzik@pobox.com>
Cc: greg@kroah.com, linux-kernel@vger.kernel.org
Subject: Re: two patches - request for comments
Message-Id: <20040602015740.3cedd197.zap@homelink.ru>
In-Reply-To: <40BCF3BF.3020202@pobox.com>
References: <20040529012030.795ad27e.zap@homelink.ru>
	<20040528221006.GB13851@kroah.com>
	<20040529124421.28c776cc.zap@homelink.ru>
	<40BCF3BF.3020202@pobox.com>
Organization: home
Mime-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 1170
Lines: 38

On Tue, 01 Jun 2004 17:23:11 -0400
Jeff Garzik <jgarzik@pobox.com> wrote:

> Typical Linux usage to an item being registered is
> 	ptr = alloc_foo()
> 	register_foo(ptr)
> 	unregister_foo(ptr)
> 	free_foo()
In this case it is:

register_lcd_device("foo", ...);
...
unregister_lcd_device("foo");

The name is guaranteed to be unique by sysfs design during the whole
device lifetime, and calling unregister_xxx() outside the lifetime brackets
is clearly an error.

> It is quite unusual to unregister based on name.  Pointers are far more 
> likely to be unique, and the programmer is far less likely to screw up 
> the unregister operation.
I understand this, I see why it looks unusual. I'll fix this if it matters.
It'll be something like:

lcd_device = register_lcd_device ("foo", ...);
...
unregister_lcd_device (lcd_device);

However, this extra pointer will take some memory; not much of course.

--
Greetings,
   Andrew
-
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/