Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1752809Ab1DZSvE (ORCPT ); Tue, 26 Apr 2011 14:51:04 -0400 Received: from iolanthe.rowland.org ([192.131.102.54]:36899 "HELO iolanthe.rowland.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with SMTP id S1751504Ab1DZSvC (ORCPT ); Tue, 26 Apr 2011 14:51:02 -0400 Date: Tue, 26 Apr 2011 14:51:00 -0400 (EDT) From: Alan Stern X-X-Sender: stern@iolanthe.rowland.org To: Rabin Vincent cc: Arnd Bergmann , , , Ulrich Weigand , , , Peter Maydell Subject: Re: ARM unaligned MMIO access with attribute((packed)) In-Reply-To: Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=UTF-8 Content-Transfer-Encoding: 8BIT Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 1741 Lines: 43 On Tue, 26 Apr 2011, Rabin Vincent wrote: > On Wed, Feb 2, 2011 at 21:30, Arnd Bergmann wrote: > > As noticed by Peter Maydell, the EHCI device driver in Linux gets > > miscompiled by some versions of arm-gcc (still need to find out which) > > due to a combination of problems: > > > > 1. In include/linux/usb/ehci_def.h, struct ehci_caps is defined > > with __attribute__((packed)), for no good reason. This is clearly > > a bug and needs to get fixed, but other drivers have the same bug > > Was a patch submitted for this? I couldn't find it in the archives. > U-Boot seems to be fixing this by adding an "aligned(4)" instead > of removing the packed: > > http://www.mail-archive.com/u-boot@lists.denx.de/msg51418.html ISTR a patch was submitted, but apparently it never got picked up. > > and it used to work. The attribute forces byte access on all members > > accessed through pointer dereference, which is not allowed on > > MMIO accesses in general. The specific code triggering the problem > > in Peter's case is in ehci-omap.c: > > ? ? ? ?omap->ehci->regs = hcd->regs > > ? ? ? ? ? ? ? ?+ HC_LENGTH(readl(&omap->ehci->caps->hc_capbase)); > > In my case it's this writel() in ehci-hub.c that gets chopped into > strbs: > > /* force reset to complete */ > ehci_writel(ehci, temp & ~(PORT_RWC_BITS | PORT_RESET), > status_reg); Why would that get messed up? The status_reg variable doesn't have any __atribute__((packed)) associated with it. Alan Stern -- 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/