libvorbis patch for armel
Of the flags that -ffast-math sets, turning -ffinite-math-only off again avoids the erroneous optimization. With the attached patch, libvorbis's examples/encoder_example produces the same (correct) output as on arm-oldabi and it makes oggenc work on armel too. M On armel, oggenc creates short output files that decode to the correct amount of silence. This is caused by an optimization bug present in gcc 4.[123] that miscompiles the MAX(x,y) macro, optimizing it away completely. Fixes: http://bugs.debian.org/515949 Analysis: https://trac.xiph.org/ticket/1526 Martin Guy martinw...@yahoo.it March 2009 --- libvorbis-1.2.0.dfsg/configure.in.old 2007-07-25 17:27:00.0 +0100 +++ libvorbis-1.2.0.dfsg/configure.in 2009-03-19 07:44:38.0 + @@ -168,6 +168,12 @@ CFLAGS=-O20 -D__NO_MATH_INLINES -fsigned-char PROFILE=-O20 -g -pg -D__NO_MATH_INLINES -fsigned-char ;; esac + + # Avoid an optimization bug in gcc-4.[123] + case $host in + arm*-*-linux-gnueabi) + CFLAGS+= -fno-finite-math-only ;; + esac fi CFLAGS=$CFLAGS $cflags_save --- libvorbis-1.2.0.dfsg/configure.old 2007-07-25 17:46:37.0 +0100 +++ libvorbis-1.2.0.dfsg/configure 2009-03-19 07:45:54.0 + @@ -19484,6 +19484,12 @@ CFLAGS=-O20 -D__NO_MATH_INLINES -fsigned-char PROFILE=-O20 -g -pg -D__NO_MATH_INLINES -fsigned-char ;; esac + + # Avoid an optimization bug in gcc-4.[123] + case $host in + arm*-*-linux-gnueabi) + CFLAGS+= -fno-finite-math-only ;; + esac fi CFLAGS=$CFLAGS $cflags_save
Re: GCC bug when -ffast-math is set on armel
On Thu, Mar 19, 2009 at 09:40:47AM +, Martin Guy wrote:
I thought I should mention this here too in case anyone else gets
bitten by this armel-specific bug.
A GCC bug has just turned up that affects the arm-*-gnueabi
architecture in gcc-4.[123]. While debugging libvorbisenc (which
produces silent output files on armel), it turns out that when -O
-ffast-math are set, GCC can produce incorrect code for the max(x,y)
macro applied to floating point values.
Thanks for finding this out. Has a bug been filed to the GCC bugtracker?
Thanks to Erik de Castro Lopo for some incisive debugging and for
producing a minimal example (should print 0 0 but doesn't)
/*
**This file is in the Public Domain.
**
**This program demonstrates a bug in the -ffast-math option of the gcc
**armel compiler : gcc version 4.3.2 (Debian 4.3.2-1.1)
**
**This works as expected:
**
** gcc -Wall -O3 gcc-test.c -o gcc-test ./gcc-test
**min : 0.max : 0.
**
**Compile with -ffast-math and things goes screwy.
**
** gcc -Wall -O3 -ffast-math gcc-test.c -o gcc-test ./gcc-test
**min : 9.max : 0.
*/
#include stdio.h
#define COUNT 10
#define test_max(x,y) ((x) (y) ? (y) : (x))
#define test_min(x,y) ((x) (y) ? (y) : (x))
int
main (void)
{ /* C Standard says static data gets initialized to zero. */
static float data [COUNT] ;
float max = -9.0, min = 9.0 ;
int k ;
for (k = 0 ; k COUNT ; k++)
{ max = test_max (max, data [k]) ;
min = test_min (min, data [k]) ;
} ;
printf (min : %12.4fmax : %12.4f\n, min, max) ;
return 0 ;
}
Full details at http://bugs.debian.org/515949
M
--
To UNSUBSCRIBE, email to debian-arm-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
--
rm -rf only sounds scary if you don't have backups
--
To UNSUBSCRIBE, email to debian-arm-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: GCC bug when -ffast-math is set on armel
Thanks for finding this out. Has a bug been filed to the GCC bugtracker? You're welcome, and yes: http://gcc.gnu.org/bugzilla/show_bug.cgi?id=39501 M -- To UNSUBSCRIBE, email to debian-arm-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
RE: GCC bug when -ffast-math is set on armel
We're working on a fix David A Rusling, ARM Fellow Processor Division Tools ARM Limited 110 Fulbourn Road, Cambridge, CB1 9NJ Tel: +44 1223 400424 -Original Message- From: martinw...@gmail.com [mailto:martinw...@gmail.com] On Behalf Of Martin Guy Sent: Thursday, March 19, 2009 4:35 PM To: Riku Voipio Cc: debian-arm@lists.debian.org Subject: Re: GCC bug when -ffast-math is set on armel Thanks for finding this out. Has a bug been filed to the GCC bugtracker? You're welcome, and yes: http://gcc.gnu.org/bugzilla/show_bug.cgi?id=39501 M -- To UNSUBSCRIBE, email to debian-arm-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org -- IMPORTANT NOTICE: The contents of this email and any attachments are confidential and may also be privileged. If you are not the intended recipient, please notify the sender immediately and do not disclose the contents to any other person, use it for any purpose, or store or copy the information in any medium. Thank you. -- To UNSUBSCRIBE, email to debian-arm-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Re: GCC bug when -ffast-math is set on armel
In case anyone's in a hurry for a working oggenc (libvorbis) or lame on armel, I've dumped fixed packages under http://martinwguy.co.uk/martin/debian/no-finite-math-only I haven't changed the version numbers on the packages, so you need to download and dpkg -i them and they will get replaced automatically by the fixed mainline versions when they are ready and you next upgrade. M -- To UNSUBSCRIBE, email to debian-arm-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
NSLU2 Lenny install problem
Hi, I have tried to install Debian Lenny on my second NSLU2 using a USB stick. I followed the instructions here: http://www.cyrius.com/debian/nslu2/ During boot the following error is seen on the serial console: [42949405.56] sd 0:0:0:0: [sda] Result: hostbyte=0x01 driverbyte=0x00 [42949405.57] end_request: I/O error, dev sda, sector 1246395 [42949405.57] sd 0:0:0:0: [sda] Result: hostbyte=0x01 driverbyte=0x00 [42949405.58] end_request: I/O error, dev sda, sector 1225939 [42949405.59] EXT3-fs error (device sda2): ext3_find_entry: reading directory #24327 offset 0 [42949405.61] EXT3-fs error (device sda2): ext3_find_entry: reading directory #24825 offset 0 [42949405.63] EXT3-fs error (device sda2): ext3_find_entry: reading directory #24001 offset 0 [42949405.64] EXT3-fs error (device sda2): ext3_find_entry: reading directory #24325 offset 0 [42949405.65] EXT3-fs error (device sda2): ext3_find_entry: reading directory #8001 offset 0 [42949405.67] EXT3-fs error (device sda2): ext3_find_entry: reading directory #24690 offset 0 [42949405.69] EXT3-fs error (device sda2): ext3_find_entry: reading directory #24690 offset 0 [42949405.70] ehci_hcd :00:01.2: USB 2.0 started, EHCI 1.00, driver 10 Dec 2004 [42949405.70] usb usb3: configuration #1 chosen from 1 choice I have tried using a second USB stick with the same result. Any ideas are welcome! /Magnus See full console log here: http://www.nabble.com/file/p22608464/capture.TXT capture.TXT -- View this message in context: http://www.nabble.com/NSLU2-Lenny-install-problem-tp22608464p22608464.html Sent from the debian-arm mailing list archive at Nabble.com. -- To UNSUBSCRIBE, email to debian-arm-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
