From: Ralf Baechle Date: Wed, 15 Mar 2006 11:36:31 +0000 (+0000) Subject: [MIPS] Work around bad code generation for . X-Git-Tag: v2.6.16~2^2~5 X-Git-Url: https://err.no/cgi-bin/gitweb.cgi?a=commitdiff_plain;h=966f4406d903a4214fdc74bec54710c6232a95b8;p=linux-2.6 [MIPS] Work around bad code generation for . If a call to set_io_port_base() was being followed by usage of mips_io_port_base in the same function gcc was possibly using the old value due to some clever abuse of const. Adding a barrier will keep the optimization and result in correct code with latest gcc. Signed-off-by: Ralf Baechle --- diff --git a/include/asm-mips/io.h b/include/asm-mips/io.h index 8c011aa61a..ba1d7bbc15 100644 --- a/include/asm-mips/io.h +++ b/include/asm-mips/io.h @@ -4,7 +4,7 @@ * for more details. * * Copyright (C) 1994, 1995 Waldorf GmbH - * Copyright (C) 1994 - 2000 Ralf Baechle + * Copyright (C) 1994 - 2000, 06 Ralf Baechle * Copyright (C) 1999, 2000 Silicon Graphics, Inc. * Copyright (C) 2004, 2005 MIPS Technologies, Inc. All rights reserved. * Author: Maciej W. Rozycki @@ -103,8 +103,20 @@ */ extern const unsigned long mips_io_port_base; -#define set_io_port_base(base) \ - do { * (unsigned long *) &mips_io_port_base = (base); } while (0) +/* + * Gcc will generate code to load the value of mips_io_port_base after each + * function call which may be fairly wasteful in some cases. So we don't + * play quite by the book. We tell gcc mips_io_port_base is a long variable + * which solves the code generation issue. Now we need to violate the + * aliasing rules a little to make initialization possible and finally we + * will need the barrier() to fight side effects of the aliasing chat. + * This trickery will eventually collapse under gcc's optimizer. Oh well. + */ +static inline void set_io_port_base(unsigned long base) +{ + * (unsigned long *) &mips_io_port_base = base; + barrier(); +} /* * Thanks to James van Artsdalen for a better timing-fix than