Date: Wed, 16 Dec 2009 00:02:58 +0000
From: Alan Cox <alan@lxorguk.ukuu.org.uk>
To: Keith Mannthey <kmannth@us.ibm.com>
Cc: lkml <linux-kernel@vger.kernel.org>, John Stultz <johnstul@us.ibm.com>
Subject: Re: [RFC][Patch] IBM Real-Time "SMI Free" mode drive -v2
Message-ID: <20091216000258.4c294d90@lxorguk.ukuu.org.uk>
In-Reply-To: <1260907788.6521.10.camel@keith-laptop>
References: <1260907788.6521.10.camel@keith-laptop>
Mime-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 1393
Lines: 35

> --- linux-2.6.32/drivers/misc/ibmrtl/ibm_rtl.c	1969-12-31 16:00:00.000000000 -0800
> +++ linux-2.6.32-rtl/drivers/misc/ibmrtl/ibm_rtl.c	2009-12-14 16:37:19.000000000 -0800

Probably belongs in drivers/firmware as its not a misc_* interface user.

> +	/* Get the address for the Extende Biso Data Area */
> +	ebda_addr = *(u16 *) phys_to_virt(EDBA_ADDR);
> +	ebda_addr <<= 4;
> +	edba_map = ioremap(ebda_addr, 4);
> +	if (!edba_map)
> +		return -ENOMEM;

This is wrong. I wish we had a single proper EBDA handling function
because this keeps coming up.

The rules for the EBDA on PC class hardware as that the EBDA address
provides a real segment offset for the EBDA *if one is present*. Lots of
machines including quite a few suprisingly modern boxes (eg AMD76x
systems without a PS/2 mouse plugged in) don't have an EBDA so you need
to check if the word you read is zero before doing an ioremap.

> +	for (table = (u32 *) edba_map ; \
> +			table < (u32 *) edba_map + ebda_size/4; table++)

You need to check the table is valid and has a correct signature. Not
doing so is completely unsafe.

The rest of the code looks pretty sane to me.

Alan
--
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/