Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1754025AbYFSO7s (ORCPT ); Thu, 19 Jun 2008 10:59:48 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1751188AbYFSO7k (ORCPT ); Thu, 19 Jun 2008 10:59:40 -0400 Received: from g5t0007.atlanta.hp.com ([15.192.0.44]:37535 "EHLO g5t0007.atlanta.hp.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750858AbYFSO7k convert rfc822-to-8bit (ORCPT ); Thu, 19 Jun 2008 10:59:40 -0400 From: "Mingarelli, Thomas" To: Linus Torvalds , Wim Van Sebroeck CC: Andrew Morton , LKML Date: Thu, 19 Jun 2008 14:58:40 +0000 Subject: RE: [WATCHDOG] v2.6.26 hpwdt.c fixes Thread-Topic: [WATCHDOG] v2.6.26 hpwdt.c fixes Thread-Index: AcjRgjQ95fouv7IlQY6ynNieA4ujfQAmTgqA Message-ID: <183C1D5A376DE343AA8F94FC2A1EC1491BFC2E931E@GVW1091EXB.americas.hpqcorp.net> References: <20080618194917.GC2741@infomag.infomag.iguana.be> In-Reply-To: Accept-Language: en-US Content-Language: en-US X-MS-Has-Attach: X-MS-TNEF-Correlator: acceptlanguage: en-US Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 8BIT MIME-Version: 1.0 X-Brightmail-Tracker: AAAAAQAAAAI= X-Whitelist: TRUE Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 4337 Lines: 108 I agree that that portion of the driver can be improved upon immensely. It was a pain getting something to work. Your 32 bit changes have been tested by two separate parties and proven to work without the -O flag. I am in the process of testing the 64 bit changes and will post the results asap. Thanks, Tom -----Original Message----- From: Linus Torvalds [mailto:torvalds@linux-foundation.org] Sent: Wednesday, June 18, 2008 3:30 PM To: Wim Van Sebroeck Cc: Andrew Morton; LKML; Mingarelli, Thomas Subject: Re: [WATCHDOG] v2.6.26 hpwdt.c fixes On Wed, 18 Jun 2008, Linus Torvalds wrote: > > Quite frankly, this is *not* the right thing to do either: the proper > thing to do is to just move the low-level call into the "asm" statement, > and leave all the function prologue/epilogue entirely to the compiler. Something like the following. AGAIN! This is totally untested. It's meant as an *example* of how to use inline asm properly, not meant to necessarily work or do the right thing. It may be buggy as hell, for all I know. And there are probably better ways to do this. (This one does just the 64-bit version, because it's a bit easier: the 32-bit version needs to probably do some of the loading and storing of registers manually in the inline asm just to avoid running out of them, since the register pressure is worse). Linus --- drivers/watchdog/hpwdt.c | 50 +++++++++++++++------------------------------ 1 files changed, 17 insertions(+), 33 deletions(-) diff --git a/drivers/watchdog/hpwdt.c b/drivers/watchdog/hpwdt.c index 2686f3e..028c957 100644 --- a/drivers/watchdog/hpwdt.c +++ b/drivers/watchdog/hpwdt.c @@ -336,39 +336,23 @@ static int __devinit detect_cru_service(void) asmlinkage void asminline_call(struct cmn_registers *pi86Regs, unsigned long *pRomEntry) { - asm("pushq %rbp \n\t" - "movq %rsp, %rbp \n\t" - "pushq %rax \n\t" - "pushq %rbx \n\t" - "pushq %rdx \n\t" - "pushq %r12 \n\t" - "pushq %r9 \n\t" - "movq %rsi, %r12 \n\t" - "movq %rdi, %r9 \n\t" - "movl 4(%r9),%ebx \n\t" - "movl 8(%r9),%ecx \n\t" - "movl 12(%r9),%edx \n\t" - "movl 16(%r9),%esi \n\t" - "movl 20(%r9),%edi \n\t" - "movl (%r9),%eax \n\t" - "call *%r12 \n\t" - "pushfq \n\t" - "popq %r12 \n\t" - "popfq \n\t" - "movl %eax, (%r9) \n\t" - "movl %ebx, 4(%r9) \n\t" - "movl %ecx, 8(%r9) \n\t" - "movl %edx, 12(%r9) \n\t" - "movl %esi, 16(%r9) \n\t" - "movl %edi, 20(%r9) \n\t" - "movq %r12, %rax \n\t" - "movl %eax, 28(%r9) \n\t" - "popq %r9 \n\t" - "popq %r12 \n\t" - "popq %rdx \n\t" - "popq %rbx \n\t" - "popq %rax \n\t" - "leave \n\t" "ret"); + asm("pushq %0 ; popfq ; call *%6; pushfq; popfq %0" + :"=r" (pi86Regs->reflags), + "=a" (pi86Regs->u1.reax), + "=b" (pi86Regs->u2.rebx), + "=c" (pi86Regs->u3.recx), + "=d" (pi86Regs->u4.redx), + "=S" (pi86Regs->resi), + "=D" (pi86Regs->redi) + :"r" (pRomEntry), + "0" (pi86Regs->reflags), + "1" (pi86Regs->u1.reax), + "2" (pi86Regs->u2.rebx), + "3" (pi86Regs->u3.recx), + "4" (pi86Regs->u4.redx), + "5" (pi86Regs->resi), + "6" (pi86Regs->redi) + :"cc", "memory"); } /* -- 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/