From mboxrd@z Thu Jan  1 00:00:00 1970
From: Timur Tabi <timur@freescale.com>
Date: Tue, 08 Jun 2010 09:55:43 -0500
Subject: [U-Boot] [PATCH] [v2] tsec: fix the return value for
 tsec_eth_init() and tsec_standard_init()
In-Reply-To: <4C0E56E3.6000007@emk-elektronik.de>
References: <1275935487-2031-1-git-send-email-timur@freescale.com>	<201006071849.29511.vapier@gentoo.org>	<4C0D78D6.2010707@freescale.com>	<201006080004.05839.vapier@gentoo.org>
	<4C0E54F4.8000006@freescale.com>
	<4C0E56E3.6000007@emk-elektronik.de>
Message-ID: <4C0E59EF.5030909@freescale.com>
List-Id: <u-boot.lists.denx.de>
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: 7bit
To: u-boot@lists.denx.de

Reinhard Meyer (-VC) wrote:

> That poses the general question what a function that initializes several
> devices should do if one of the devices should return an error and what
> to return if ALL devices return an error.

I believe the consensus is that any device initialization function that
returns an error is simply ignored and skipped, which is what my patch does:

	for (i = 0; i < num; i++) {
		ret = tsec_initialize(bis, &tsecs[i]);
		if (ret > 0)
			count += ret;
	}

I could have done "if (ret >= 0)", but that's less efficient.

> At least in the last case I would assume to return the error code of one
> of the devices.

But which error code?  What if the first device returns -1, and the second
one returns -2?

> In the cases where not all, but at least one of the devices get initialized
> without error, the number of successful devices might be returned.

That's what we do today.

> However, in most contexts that might not be helpful to the rest of the
> system anyway.
> 
>