[gentoo-amd64] Compiling mozilla/firefox with -O2

[gentoo-amd64] Compiling mozilla/firefox with -O2

[Simon Strandman]

The mozconfig eclass strips flags down to -O1 on amd64 with the 
explanation "more than -O1 causes segfaults on 64-bit (bug 33767)". That 
bug is two years old and we have a much improved amd64 toolchain since. 
I've been compiling firefox and and mozilla with -O2 and -Os for quite 
some time and never had any problems at all. And this is also what 
binary distros does. I opened a bug about it but nobody responded:
http://bugs.gentoo.org/show_bug.cgi?id=86571

So please test compiling firefox and other mozilla apps with -O2 and 
respond to that bug so this can be fixed!

-- 
Simon Strandman <simon.strandman@telia.com>

-- 
gentoo-amd64@gentoo.org mailing list

Re: [gentoo-amd64] Compiling mozilla/firefox with -O2

[David Edge]

Simon Strandman wrote:

> The mozconfig eclass strips flags down to -O1 on amd64 with the 
> explanation "more than -O1 causes segfaults on 64-bit (bug 33767)". 
> That bug is two years old and we have a much improved amd64 toolchain 
> since. I've been compiling firefox and and mozilla with -O2 and -Os 
> for quite some time and never had any problems at all. And this is 
> also what binary distros does. I opened a bug about it but nobody 
> responded:
> http://bugs.gentoo.org/show_bug.cgi?id=86571
>
> So please test compiling firefox and other mozilla apps with -O2 and 
> respond to that bug so this can be fixed!
>
Not sure about this - I have built both Mozilla and Firefox with both O1 
and O2, and get a problem. At seemingly random intervals, on no site in 
particular, when drawing a new page, I get rubbish on the screen - both 
inside the browser window and in other windows and on the root window - 
I have to shut down and restart X to have a usable system again.
This only seems to happen when I use the browser, so I'm sure it's a 
Mozilla/Firefox problem. I haven't really had time to investigate and am 
new to Gentoo, so didn't really want to raise a bug without more 
investigation.
Cheers, Dave

Portage 2.0.51.22-r3 (default-linux/amd64/2005.1, gcc-3.4.4, 
glibc-2.3.5-r2, 2.6.13-gentoo-r5 x86_64)
=================================================================
System uname: 2.6.13-gentoo-r5 x86_64 AMD Opteron(tm) Processor 248
Gentoo Base System version 1.6.13
ccache version 2.3 [enabled]
dev-lang/python:     2.3.5-r2, 2.4.2
sys-apps/sandbox:    1.2.12
sys-devel/autoconf:  2.13, 2.59-r6
sys-devel/automake:  1.4_p6, 1.5, 1.6.3, 1.7.9-r1, 1.8.5-r3, 1.9.6-r1
sys-devel/binutils:  2.15.92.0.2-r10
sys-devel/libtool:   1.5.20
virtual/os-headers:  2.6.11-r2
ACCEPT_KEYWORDS="amd64"
AUTOCLEAN="yes"
CBUILD="x86_64-pc-linux-gnu"
CFLAGS="-march=k8 -pipe -O2"
CHOST="x86_64-pc-linux-gnu"
CONFIG_PROTECT="/etc /usr/kde/2/share/config /usr/kde/3/share/config 
/usr/share/config /var/qmail/control"
CONFIG_PROTECT_MASK="/etc/gconf /etc/env.d"
CXXFLAGS="-march=k8 -pipe -O2"
DISTDIR="/usr/portage/distfiles"
FEATURES="autoconfig ccache distlocks sandbox sfperms strict"
GENTOO_MIRRORS="http://ftp.belnet.be/mirror/rsync.gentoo.org/gentoo/ 
http://ftp.heanet.ie/pub/gentoo/ 
http://ftp.snt.utwente.nl/pub/os/linux/gentoo 
http://gentoo.blueyonder.co.uk"
MAKEOPTS="-j3"
PKGDIR="/usr/portage/packages"
PORTAGE_TMPDIR="/var/tmp"
PORTDIR="/usr/portage"
SYNC="rsync://rsync.europe.gentoo.org/gentoo-portage"
USE="amd64 X a52 aac acl acpi alsa avi berkdb bitmap-fonts bzip2 cdr 
crypt cups curl dvd dvdr dvdread eds emboss emul-linux-x86 encode esd 
exif fam ffmpeg flac foomaticdb fortran ftp gif glut gnome gphoto2 gpm 
gstreamer gtk gtk2 hardened imagemagick imlib ipv6 java jpeg lzw 
lzw-tiff mozilla mp3 mpeg ncurses nls nsplugin offensive ogg opengl pam 
pdflib perl png python quicktime readline sdl spell ssl tcpd tiff 
truetype-fonts type1-fonts udev usb userlocales vorbis xml2 xmms xpm xv 
zlib userland_GNU kernel_linux elibc_glibc"
Unset:  ASFLAGS, CTARGET, LANG, LC_ALL, LDFLAGS, LINGUAS, PORTDIR_OVERLAY


-- 
gentoo-amd64@gentoo.org mailing list

[gentoo-amd64] Re: Compiling mozilla/firefox with -O2

[Duncan]

Simon Strandman posted <436F70C0.1040303@telia.com>, excerpted below,  on
Mon, 07 Nov 2005 16:20:32 +0100:

> The mozconfig eclass strips flags down to -O1 on amd64 with the 
> explanation "more than -O1 causes segfaults on 64-bit (bug 33767)". That 
> bug is two years old and we have a much improved amd64 toolchain since. 
> I've been compiling firefox and and mozilla with -O2 and -Os for quite 
> some time and never had any problems at all. And this is also what 
> binary distros does. I opened a bug about it but nobody responded:
> http://bugs.gentoo.org/show_bug.cgi?id=86571
> 
> So please test compiling firefox and other mozilla apps with -O2 and 
> respond to that bug so this can be fixed!

For something of this nature (definitely 64-bit/amd64 related, and where 
the amd64 arch testers (ATs) can be put to work confirming the results,
especially since the package owners may not be able to test...), it
generally works much better to either assign it to amd64@, or assign it to
the package owner (as you did) but CC amd64@ (as you did /not/). 

Apparently, many of the herds, and I'd guess mozilla is one of them, are
really stretched far too thin, and for something like this that's going to
require a lot of testing they don't have the direct ability to do, and
where it's working as is, what resources there are will likely get
prioritized to security bugs and the like, since they can't do much with
this directly anyway.  Get amd64@ in the CC (or assign it to them
originally if appropriate, as I think it could have been here), and it's
likely to get action rather sooner than it might otherwise.

(Me... I find konqueror, which I have on my system anyway since KDE's my
default desktop, to be quite sufficient.  Merging Firefox/Mozilla would
just mean one more package to continually compile from source to keep
updated -- and a security critical package at that.)

-- 
Duncan - List replies preferred.   No HTML msgs.
"Every nonfree program has a lord, a master --
and if you use the program, he is your master."  Richard Stallman in
http://www.linuxdevcenter.com/pub/a/linux/2004/12/22/rms_interview.html


-- 
gentoo-amd64@gentoo.org mailing list

Re: [gentoo-amd64] Compiling mozilla/firefox with -O2

[Simon Strandman]

David Edge skrev:

> Simon Strandman wrote:

> Not sure about this - I have built both Mozilla and Firefox with both 
> O1 and O2, and get a problem. At seemingly random intervals, on no 
> site in particular, when drawing a new page, I get rubbish on the 
> screen - both inside the browser window and in other windows and on 
> the root window - I have to shut down and restart X to have a usable 
> system again.

Does it make a difference if you compile it with -O1 or -O2? It sounds 
more like a X problem to me. You aren't running with composite on perhaps?

-- 
Simon Strandman <simon.strandman@telia.com>

-- 
gentoo-amd64@gentoo.org mailing list

Re: [gentoo-amd64] Re: Compiling mozilla/firefox with -O2

[Simon Strandman]

Duncan skrev:

>For something of this nature (definitely 64-bit/amd64 related, and where 
>the amd64 arch testers (ATs) can be put to work confirming the results,
>especially since the package owners may not be able to test...), it
>generally works much better to either assign it to amd64@, or assign it to
>the package owner (as you did) but CC amd64@ (as you did /not/). 
>  
>
Ok, will do! Actually I didn't assign it to anyone initially.

-- 
Simon Strandman <simon.strandman@telia.com>

-- 
gentoo-amd64@gentoo.org mailing list

Re: [gentoo-amd64] Compiling mozilla/firefox with -O2

[Richard Freeman]

[-- Attachment #1: Type: text/plain, Size: 967 bytes --]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Simon Strandman wrote:
> 
> So please test compiling firefox and other mozilla apps with -O2 and
> respond to that bug so this can be fixed!
> 

I edited the eclass to hard code in -Os and firefox built fine, and so
far appears to run fine.  I confirmed that -Os was used in the actual
emerge log so it didn't get filtered out.  So, that is a good sign.
Perhaps we can be a little more liberal with the flags in the future.
I'll have to try it out with thunderbird.  Too bad I don't have memory
usage stats on firefox before the upgrade.  I'll have to generate those
for thunderbird.  I love -Os since RAM tends to be more limiting than
CPU for most applications.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org

iD8DBQFDb6nog2bN8aFizRkRAuf+AJsGr/nPSTTOqEgw2+klgfZOMqxtegCeJJl/
gxWGTcZXl3z/cbWXOXeFNr4=
=x/Iz
-----END PGP SIGNATURE-----

[-- Attachment #2: S/MIME Cryptographic Signature --]
[-- Type: application/x-pkcs7-signature, Size: 4275 bytes --]

Re: [gentoo-amd64] Compiling mozilla/firefox with -O2

[David Edge]

Simon Strandman wrote:

> David Edge skrev:
>
>> Simon Strandman wrote:
>
>
>> Not sure about this - I have built both Mozilla and Firefox with both 
>> O1 and O2, and get a problem. At seemingly random intervals, on no 
>> site in particular, when drawing a new page, I get rubbish on the 
>> screen - both inside the browser window and in other windows and on 
>> the root window - I have to shut down and restart X to have a usable 
>> system again.
>
>
> Does it make a difference if you compile it with -O1 or -O2? It sounds 
> more like a X problem to me. You aren't running with composite on 
> perhaps?
>
Sorry, what's composite? May be an X problem, but it only ever happens 
when the browser displays a new page. I don't see anything in any error 
logs in /var/log.
Dave
-- 
gentoo-amd64@gentoo.org mailing list

Re: [gentoo-amd64] Compiling mozilla/firefox with -O2

[Simon Strandman]

Richard Freeman skrev:

>-----BEGIN PGP SIGNED MESSAGE-----
>Hash: SHA1
>
>  
>
>I edited the eclass to hard code in -Os and firefox built fine, and so
>far appears to run fine.  I confirmed that -Os was used in the actual
>emerge log so it didn't get filtered out.  So, that is a good sign.
>Perhaps we can be a little more liberal with the flags in the future.
>I'll have to try it out with thunderbird.  Too bad I don't have memory
>usage stats on firefox before the upgrade.  I'll have to generate those
>for thunderbird.  I love -Os since RAM tends to be more limiting than
>CPU for most applications.
>  
>
Yes, -Os seems indeed to be a good flag to compile firefox with. I 
remember reading a pdf created by some redhat guys about various 
optimizations on mozilla and they came to the conclusion that -Os what 
pretty optimal. So the mozilla/firefox/thunderbird rpms in RHEL/fedora 
are built with -Os, and so are the official mozilla.org builds. Actually 
I opened a bug report about that to:
https://bugs.gentoo.org/show_bug.cgi?id=107584

-- 
Simon Strandman <simon.strandman@telia.com>

-- 
gentoo-amd64@gentoo.org mailing list

Re: [gentoo-amd64] Compiling mozilla/firefox with -O2

[Simon Strandman]

David Edge skrev:

> Simon Strandman wrote:
>
> Sorry, what's composite? May be an X problem, but it only ever happens 
> when the browser displays a new page. I don't see anything in any 
> error logs in /var/log.
> Dave

Then you probably don't have it on. :) It can cause corruption like this.

http://gentoo-wiki.com/TIP_Xorg_X11_and_Transparency

-- 
Simon Strandman <simon.strandman@telia.com>

-- 
gentoo-amd64@gentoo.org mailing list

Re: [gentoo-amd64] Compiling mozilla/firefox with -O2

[David Edge]

Simon,
no I don't have composite on! I have emerged firefox with debug in 
/etc/portage/package.use and am running "firefox -debug" and so far I 
haven't seen the problem. Due to the apparently random nature of the 
problem I can't be sure about it yet.

If you are interested I have a screenshot of my display just after the 
error happened - it's about 365kB. Let me know what to do with it if you 
want to see it.
Cheers, Dave
-- 
gentoo-amd64@gentoo.org mailing list