From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1751886Ab0J3Gcp (ORCPT ); Sat, 30 Oct 2010 02:32:45 -0400 Received: from mail-wy0-f174.google.com ([74.125.82.174]:37706 "EHLO mail-wy0-f174.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751544Ab0J3Gcm (ORCPT ); Sat, 30 Oct 2010 02:32:42 -0400 Date: Sat, 30 Oct 2010 07:32:37 +0100 From: Grant Likely To: David Daney Cc: linux-mips@linux-mips.org, ralf@linux-mips.org, devicetree-discuss@lists.ozlabs.org, linux-kernel@vger.kernel.org, Jeremy Kerr , Benjamin Herrenschmidt , Dan Carpenter , Greg Kroah-Hartman Subject: Re: [PATCH] of: of_mdio: Fix some endianness problems. Message-ID: <20101030063237.GC2456@angua.secretlab.ca> References: <1288227827-5447-1-git-send-email-ddaney@caviumnetworks.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <1288227827-5447-1-git-send-email-ddaney@caviumnetworks.com> User-Agent: Mutt/1.5.20 (2009-06-14) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Wed, Oct 27, 2010 at 06:03:47PM -0700, David Daney wrote: > In of_mdiobus_register(), the __be32 *addr variable is dereferenced. > This will not work on little-endian targets. Also since it is > unsigned, checking for less than zero is redundant. > > Fix these two issues. > > Signed-off-by: David Daney > Cc: Grant Likely > Cc: Jeremy Kerr > Cc: Benjamin Herrenschmidt > Cc: Dan Carpenter > Cc: Greg Kroah-Hartman > --- > drivers/of/of_mdio.c | 23 ++++++++++++++--------- > 1 files changed, 14 insertions(+), 9 deletions(-) > > diff --git a/drivers/of/of_mdio.c b/drivers/of/of_mdio.c > index 1fce00e..b370306 100644 > --- a/drivers/of/of_mdio.c > +++ b/drivers/of/of_mdio.c > @@ -52,27 +52,32 @@ int of_mdiobus_register(struct mii_bus *mdio, struct device_node *np) > > /* Loop over the child nodes and register a phy_device for each one */ > for_each_child_of_node(np, child) { > - const __be32 *addr; > + const __be32 *paddr; > + u32 addr; > int len; > > /* A PHY must have a reg property in the range [0-31] */ > - addr = of_get_property(child, "reg", &len); > - if (!addr || len < sizeof(*addr) || *addr >= 32 || *addr < 0) { > + paddr = of_get_property(child, "reg", &len); > + if (!paddr || len < sizeof(*paddr)) { > +addr_err: > dev_err(&mdio->dev, "%s has invalid PHY address\n", > child->full_name); > continue; > } > + addr = be32_to_cpup(paddr); > + if (addr >= 32) > + goto addr_err; goto's are fine for jumping to the end of a function to unwind allocations, but please don't use it in this manner. The original structure will actually work just fine if you do it thusly: if (!paddr || len < sizeof(*paddr) || *(addr = be32_to_cpup(paddr)) >= 32) { dev_err(&mdio->dev, "%s has invalid PHY address\n", child->full_name); continue; } Otherwise this patch looks good. After you've reworked and retested I'll pick it up for 2.6.37 (or dave will). g. > > if (mdio->irq) { > - mdio->irq[*addr] = irq_of_parse_and_map(child, 0); > - if (!mdio->irq[*addr]) > - mdio->irq[*addr] = PHY_POLL; > + mdio->irq[addr] = irq_of_parse_and_map(child, 0); > + if (!mdio->irq[addr]) > + mdio->irq[addr] = PHY_POLL; > } > > - phy = get_phy_device(mdio, be32_to_cpup(addr)); > + phy = get_phy_device(mdio, addr); > if (!phy || IS_ERR(phy)) { > dev_err(&mdio->dev, "error probing PHY at address %i\n", > - *addr); > + addr); > continue; > } > phy_scan_fixups(phy); > @@ -91,7 +96,7 @@ int of_mdiobus_register(struct mii_bus *mdio, struct device_node *np) > } > > dev_dbg(&mdio->dev, "registered phy %s at address %i\n", > - child->name, *addr); > + child->name, addr); > } > > return 0; > -- > 1.7.2.3 >