Return-path: Received: from mail-bw0-f46.google.com ([209.85.214.46]:48958 "EHLO mail-bw0-f46.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1756570Ab1CSCDP (ORCPT ); Fri, 18 Mar 2011 22:03:15 -0400 Received: by bwz15 with SMTP id 15so3815481bwz.19 for ; Fri, 18 Mar 2011 19:03:13 -0700 (PDT) From: Christian Lamparter To: Ignacy Gawedzki Subject: Re: Dividing by a non-static value in carl9170-fw? Date: Sat, 19 Mar 2011 03:03:06 +0100 Cc: linux-wireless@vger.kernel.org References: <20110116031600.GA21973@zenon.in.qult.net> <201103190143.27042.chunkeey@googlemail.com> <20110319012244.GA4157@zenon.in.qult.net> In-Reply-To: <20110319012244.GA4157@zenon.in.qult.net> MIME-Version: 1.0 Content-Type: Multipart/Mixed; boundary="Boundary-00=_b7AhNqTUE5y4bJZ" Message-Id: <201103190303.07310.chunkeey@googlemail.com> Sender: linux-wireless-owner@vger.kernel.org List-ID: --Boundary-00=_b7AhNqTUE5y4bJZ Content-Type: Text/Plain; charset="iso-8859-15" Content-Transfer-Encoding: 7bit On Saturday 19 March 2011 02:22:44 Ignacy Gawedzki wrote: > On Sat, Mar 19, 2011 at 01:43:26AM +0100, thus spake Christian Lamparter: > > I gave it a try and It worked?! > > > > [54670.551750] ieee80211 phy14: FW: a:0x4537 b:0x2c a/b:0x192 (0x192 = 402). > > > > (It was nothing else than just define "a" and "b" (as global variables) and > > add the INFO("a:0x%x b:0x%x a/b:0x%x", a, b, a/b);) > > I suppose in your case the values are known at compile time and the compiler > does the calculation statically. sometimes its just down to "luck" ;) > If you define the global variables as volatile, then the optimization is > forbidden and you should get the bogus results. ??? No, "global" is enough. As long as you don't select "const". (In fact there's more of a story behind "const volatile" than you might think, or have you ever wondered why carl9170 still ships with gcc 4.4 instead of 4.5? Anyway that's OT) > BTW, I used udivsi3_i4i.S from the Linux kernel, modified it as you indicated. > Please find my version attached. You are right, I can reproduce your problem with the attached udiv. The crux is seems to be "hidden" inside the kernel's Makefile: >udivsi3-y := udivsi3_i4i-Os.o > >ifneq ($(CONFIG_CC_OPTIMIZE_FOR_SIZE),y) >udivsi3-$(CONFIG_CPU_SH3) := udivsi3_i4i.o >udivsi3-$(CONFIG_CPU_SH4) := udivsi3_i4i.o >endif It looks like the version you are using is "made" for SH3+, however the AR9170 has a SH2. (Jup, I can't believe it either :-D) Best Regards, Chr (didn't really bother with the "signed" div.) --Boundary-00=_b7AhNqTUE5y4bJZ Content-Type: text/x-csrc; charset="UTF-8"; name="udivsi3_i4i-Os.S" Content-Transfer-Encoding: 7bit Content-Disposition: attachment; filename="udivsi3_i4i-Os.S" /* Copyright (C) 2006 Free Software Foundation, Inc. This file is free software; you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation; either version 2, or (at your option) any later version. In addition to the permissions in the GNU General Public License, the Free Software Foundation gives you unlimited permission to link the compiled version of this file into combinations with other programs, and to distribute those combinations without any restriction coming from the use of this file. (The General Public License restrictions do apply in other respects; for example, they cover modification of the file, and distribution when not linked into a combine executable.) This file is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details. You should have received a copy of the GNU General Public License along with this program; see the file COPYING. If not, write to the Free Software Foundation, 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA. */ /* Moderately Space-optimized libgcc routines for the Renesas SH / STMicroelectronics ST40 CPUs. Contributed by J"orn Rennecke joern.rennecke@st.com. */ /* Size: 186 bytes jointly for udivsi3_i4i and sdivsi3_i4i sh4-200 run times: udiv small divisor: 55 cycles udiv large divisor: 52 cycles sdiv small divisor, positive result: 59 cycles sdiv large divisor, positive result: 56 cycles sdiv small divisor, negative result: 65 cycles (*) sdiv large divisor, negative result: 62 cycles (*) (*): r2 is restored in the rts delay slot and has a lingering latency of two more cycles. */ .balign 4 .global ___udivsi3_i4i .global ___sdivsi3_i4i .global ___udivsi3_i4 .set ___udivsi3_i4, ___udivsi3_i4i .type ___udivsi3_i4i, @function .type ___sdivsi3_i4i, @function ___udivsi3_i4i: ___sdivsi3_i4i: sts pr,r1 mov.l r4,@-r15 extu.w r5,r0 cmp/eq r5,r0 swap.w r4,r0 shlr16 r4 bf/s large_divisor div0u mov.l r5,@-r15 shll16 r5 sdiv_small_divisor: div1 r5,r4 bsr div6 div1 r5,r4 div1 r5,r4 bsr div6 div1 r5,r4 xtrct r4,r0 xtrct r0,r4 bsr div7 swap.w r4,r4 div1 r5,r4 bsr div7 div1 r5,r4 xtrct r4,r0 mov.l @r15+,r5 swap.w r0,r0 mov.l @r15+,r4 jmp @r1 rotcl r0 div7: div1 r5,r4 div6: div1 r5,r4; div1 r5,r4; div1 r5,r4 div1 r5,r4; div1 r5,r4; rts; div1 r5,r4 divx3: rotcl r0 div1 r5,r4 rotcl r0 div1 r5,r4 rotcl r0 rts div1 r5,r4 large_divisor: mov.l r5,@-r15 sdiv_large_divisor: xor r4,r0 .rept 4 rotcl r0 bsr divx3 div1 r5,r4 .endr mov.l @r15+,r5 mov.l @r15+,r4 jmp @r1 rotcl r0 .global __sdivsi3_i4i .global __sdivsi3_i4 .global __sdivsi3 .set __sdivsi3_i4, __sdivsi3_i4i .set __sdivsi3, __sdivsi3_i4i __sdivsi3_i4i: mov.l r4,@-r15 cmp/pz r5 mov.l r5,@-r15 bt/s pos_divisor cmp/pz r4 neg r5,r5 extu.w r5,r0 bt/s neg_result cmp/eq r5,r0 neg r4,r4 pos_result: swap.w r4,r0 bra sdiv_check_divisor sts pr,r1 pos_divisor: extu.w r5,r0 bt/s pos_result cmp/eq r5,r0 neg r4,r4 neg_result: mova negate_result,r0 ; mov r0,r1 swap.w r4,r0 lds r2,macl sts pr,r2 sdiv_check_divisor: shlr16 r4 bf/s sdiv_large_divisor div0u bra sdiv_small_divisor shll16 r5 .balign 4 negate_result: neg r0,r0 jmp @r2 sts macl,r2 --Boundary-00=_b7AhNqTUE5y4bJZ--