Bug#342105: certain JavaScript sites make firefox consume 100% CPU
On Mon, 2007-01-29 at 02:19 -0500, Eric Dorland wrote: > * Ben Hutchings ([EMAIL PROTECTED]) wrote: > > I've also experienced this when scrolling pages on > > http://www.worldchanging.com which doesn't seem like it would involve > > any JS. Scrolling it a page at a time, or dragging the scrollbar > > quickly, seems to keep the X server busy for about half a second. This > > is long-enough that a page-down key can auto-repeat! Eric Wong's > > suggestion to use EXA fixes the problem for me on an "ATI Technologies > > Inc RV350 [Mobility Radeon 9600 M10]". > > > > I haven't tried using XAA and Firefox/Iceweasel 2.0, but could do if > > necessary. In any case this seems more like an X server (or driver) > > bug, even if the browser is particularly prone to provoking it. > > Does running it with MOZ_DISABLE_PANGO make a difference? I just tested all four combinations of {EXA,XAA} and MOZ_DISABLE_PANGO {unset,=1} and the latter variable makes no difference. Ben. -- Ben Hutchings I say we take off; nuke the site from orbit. It's the only way to be sure. signature.asc Description: This is a digitally signed message part
Bug#342105: certain JavaScript sites make firefox consume 100% CPU
* Ben Hutchings ([EMAIL PROTECTED]) wrote: > I've also experienced this when scrolling pages on > http://www.worldchanging.com which doesn't seem like it would involve > any JS. Scrolling it a page at a time, or dragging the scrollbar > quickly, seems to keep the X server busy for about half a second. This > is long-enough that a page-down key can auto-repeat! Eric Wong's > suggestion to use EXA fixes the problem for me on an "ATI Technologies > Inc RV350 [Mobility Radeon 9600 M10]". > > I haven't tried using XAA and Firefox/Iceweasel 2.0, but could do if > necessary. In any case this seems more like an X server (or driver) > bug, even if the browser is particularly prone to provoking it. Does running it with MOZ_DISABLE_PANGO make a difference? -- Eric Dorland <[EMAIL PROTECTED]> ICQ: #61138586, Jabber: [EMAIL PROTECTED] 1024D/16D970C6 097C 4861 9934 27A0 8E1C 2B0A 61E9 8ECF 16D9 70C6 signature.asc Description: Digital signature
Bug#342105: certain JavaScript sites make firefox consume 100% CPU
I've also experienced this when scrolling pages on http://www.worldchanging.com which doesn't seem like it would involve any JS. Scrolling it a page at a time, or dragging the scrollbar quickly, seems to keep the X server busy for about half a second. This is long-enough that a page-down key can auto-repeat! Eric Wong's suggestion to use EXA fixes the problem for me on an "ATI Technologies Inc RV350 [Mobility Radeon 9600 M10]". I haven't tried using XAA and Firefox/Iceweasel 2.0, but could do if necessary. In any case this seems more like an X server (or driver) bug, even if the browser is particularly prone to provoking it. Ben. -- Ben Hutchings For every complex problem there is a solution that is simple, neat, and wrong. signature.asc Description: This is a digitally signed message part
Bug#342105: certain JavaScript sites make firefox consume 100% CPU
* Eric Wong ([EMAIL PROTECTED]) [20060828 17:39]: > Ah, I forgot to note that I'm using xserver-xorg-video-ati and > xserver-xorg-core from experimental. I see. FTM I'd rather not wander to experimental-land... > You can use --no-vmlinux (--vmlinux is for profiling the kernel > itself, you don't need it for userspace). That's correct. However, the sluggishness was so severe that I suspected kernel-related issues; moving a window fast and dropping it really stalled the machine for seconds (2.4 GHz P4, Xorg nice 0). I try to get a test box and see if experimental makes a difference. Thanks, Peter -- .+'''+. .+'''+. .+'''+. .+'''+. .+'' Kelemen Péter / \ / \ [EMAIL PROTECTED] .+' `+...+' `+...+' `+...+' `+...+' -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#342105: certain JavaScript sites make firefox consume 100% CPU
KELEMEN Peter <[EMAIL PROTECTED]> wrote: > * Eric Wong ([EMAIL PROTECTED]) [20060827 20:23]: > > > Option "AccelMethod" "exa" > > Hm. I tried EXA long time ago but switched back to XAA while > hunting a DRI-related hang and I forgot to switch back to EXA. > Now I did and it is dog slow (only moving windows around spikes > to 100% CPU), interestingly I have the same RV280 9200SE chip > than you do. No oprofile because I didn't find the vmlinux image > matching my Debian kernel. Any hints? Ah, I forgot to note that I'm using xserver-xorg-video-ati and xserver-xorg-core from experimental. You can use --no-vmlinux (--vmlinux is for profiling the kernel itself, you don't need it for userspace). Moving an opaque window in FVWM 2.5.17 brings Xorg up to 60% CPU (2500 AthlonXP) or so on my 9200SE system, but it remains responsive. Moving a window frame (no animation) doesn't seem to use much CPU at all :) I'm using 24-bit color on a 1600x1200 display here without problems, Otherwise, my xorg config is fairly standard, I'm not using the Composite extension, or even backingstore (never noticed a difference with it). I'm also using the following settings (set for stability), but they don't seem to affect 2D performance Option "AGPFastWrite""off" Option "EnablePageFlip" "false" I'm also using a stock 2.6.16.16 kernel on my 9200SE with a fairly conservative config (no preempt, 100Hz). -- Eric Wong -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#342105: certain JavaScript sites make firefox consume 100% CPU
* Eric Wong ([EMAIL PROTECTED]) [20060827 20:23]: > Option "AccelMethod" "exa" Hm. I tried EXA long time ago but switched back to XAA while hunting a DRI-related hang and I forgot to switch back to EXA. Now I did and it is dog slow (only moving windows around spikes to 100% CPU), interestingly I have the same RV280 9200SE chip than you do. No oprofile because I didn't find the vmlinux image matching my Debian kernel. Any hints? Peter -- .+'''+. .+'''+. .+'''+. .+'''+. .+'' Kelemen Péter / \ / \ [EMAIL PROTECTED] .+' `+...+' `+...+' `+...+' `+...+' -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#342105: certain JavaScript sites make firefox consume 100% CPU
Mike Hommey <[EMAIL PROTECTED]> wrote: > On Fri, Aug 04, 2006 at 02:44:51PM +0200, KELEMEN Peter <[EMAIL PROTECTED]> > wrote: > > * Mike Hommey ([EMAIL PROTECTED]) [20060206 07:00]: > > > > > More likely to be something like a javascript making some text > > > moving fast... > > > > The problem is back, although differently; even with JavaScript > > turned off, blogger.com is a CPU hog in Firefox. It turns out > > if I switch CSS off (View->Page Style->No Style), the problem > > vanishes. Font subpixel rendering is off, DRI is active. > > There might be a fixed background or something similar. This has been bugging me for a long time now, and my solution had been to stick to the 1.0.x versions which weren't affected by this (at least not for me). Today I finally got around to looking into it and started oprofiling to look for a bottleneck. Oprofile pointed me to the fbBlt() function in libfb.so; and Googling for these terms brought me to this thread: http://lists.freedesktop.org/archives/xorg/2006-March/thread.html#13602 and particularly this message: http://lists.freedesktop.org/archives/xorg/2006-March/013608.html So I just added this line to the "Device" section of my /etc/X11/xorg.conf: Option "AccelMethod" "exa" And restarted X... Firefox is suddenly faster than it's ever been for me :D Unfortunately (as far as I understand), EXA doesn't work for everyone yet, but with my Radeon machines (Mobility M6 LY on my Thinkpad, and RV280 9200 SE on my desktop) seem to be working fine so far with EXA enabled. -- Eric Wong -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#342105: certain JavaScript sites make firefox consume 100% CPU
On Fri, Aug 04, 2006 at 02:44:51PM +0200, KELEMEN Peter <[EMAIL PROTECTED]> wrote: > * Mike Hommey ([EMAIL PROTECTED]) [20060206 07:00]: > > > More likely to be something like a javascript making some text > > moving fast... > > The problem is back, although differently; even with JavaScript > turned off, blogger.com is a CPU hog in Firefox. It turns out > if I switch CSS off (View->Page Style->No Style), the problem > vanishes. Font subpixel rendering is off, DRI is active. There might be a fixed background or something similar. Mike -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#342105: certain JavaScript sites make firefox consume 100% CPU
* Mike Hommey ([EMAIL PROTECTED]) [20060206 07:00]: > More likely to be something like a javascript making some text > moving fast... The problem is back, although differently; even with JavaScript turned off, blogger.com is a CPU hog in Firefox. It turns out if I switch CSS off (View->Page Style->No Style), the problem vanishes. Font subpixel rendering is off, DRI is active. Peter -- .+'''+. .+'''+. .+'''+. .+'''+. .+'' Kelemen Péter / \ / \ /[EMAIL PROTECTED] .+' `+...+' `+...+' `+...+' `+...+' -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#342105: certain JavaScript sites make firefox consume 100% CPU
On Sun, Feb 05, 2006 at 04:35:47PM -0500, Eric Dorland <[EMAIL PROTECTED]> wrote: > * KELEMEN Peter ([EMAIL PROTECTED]) wrote: > > * Eric Dorland ([EMAIL PROTECTED]) [20051224 02:48]: > > > > > blogger.com seems to make the X server consume a ton of CPU (but > > > i could be wrong), not firefox, and I had no problem with > > > qlogic.com. Any extensions installed? > > > > None. It is something that Firefox does that makes X burn CPU > > cycles, and it turns out, switching off subpixel font rendering > > makes the problem go away. > > Please keep the bug address when replying. > > That's interesting, maybe it's a bug in libxft? More likely to be something like a javascript making some text moving fast... Mike -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#342105: certain JavaScript sites make firefox consume 100% CPU
* KELEMEN Peter ([EMAIL PROTECTED]) wrote: > * Eric Dorland ([EMAIL PROTECTED]) [20051224 02:48]: > > > blogger.com seems to make the X server consume a ton of CPU (but > > i could be wrong), not firefox, and I had no problem with > > qlogic.com. Any extensions installed? > > None. It is something that Firefox does that makes X burn CPU > cycles, and it turns out, switching off subpixel font rendering > makes the problem go away. Please keep the bug address when replying. That's interesting, maybe it's a bug in libxft? -- Eric Dorland <[EMAIL PROTECTED]> ICQ: #61138586, Jabber: [EMAIL PROTECTED] 1024D/16D970C6 097C 4861 9934 27A0 8E1C 2B0A 61E9 8ECF 16D9 70C6 -BEGIN GEEK CODE BLOCK- Version: 3.12 GCS d- s++: a-- C+++ UL+++ P++ L++ E++ W++ N+ o K- w+ O? M++ V-- PS+ PE Y+ PGP++ t++ 5++ X+ R tv++ b+++ DI+ D+ G e h! r- y+ --END GEEK CODE BLOCK-- signature.asc Description: Digital signature
Bug#342105: certain JavaScript sites make firefox consume 100% CPU
* KELEMEN Peter ([EMAIL PROTECTED]) wrote: > Package: firefox > Version: 1.4.99+1.5rc3.dfsg-2 > Severity: important > > SSIA, notably http://blogger.com/ and http://qlogic.com/ are such sites. blogger.com seems to make the X server consume a ton of CPU (but i could be wrong), not firefox, and I had no problem with qlogic.com. Any extensions installed? -- Eric Dorland <[EMAIL PROTECTED]> ICQ: #61138586, Jabber: [EMAIL PROTECTED] 1024D/16D970C6 097C 4861 9934 27A0 8E1C 2B0A 61E9 8ECF 16D9 70C6 -BEGIN GEEK CODE BLOCK- Version: 3.12 GCS d- s++: a-- C+++ UL+++ P++ L++ E++ W++ N+ o K- w+ O? M++ V-- PS+ PE Y+ PGP++ t++ 5++ X+ R tv++ b+++ DI+ D+ G e h! r- y+ --END GEEK CODE BLOCK-- signature.asc Description: Digital signature
Bug#342105: certain JavaScript sites make firefox consume 100% CPU
Package: firefox Version: 1.4.99+1.5rc3.dfsg-2 Severity: important SSIA, notably http://blogger.com/ and http://qlogic.com/ are such sites. -- System Information: Debian Release: testing/unstable APT prefers unstable APT policy: (500, 'unstable'), (500, 'stable'), (1, 'experimental') Architecture: i386 (i686) Shell: /bin/sh linked to /bin/dash Kernel: Linux 2.6.13-atapass-chihiro.1 Locale: LANG=C, LC_CTYPE=en_US.UTF-8 (charmap=UTF-8) Versions of packages firefox depends on: ii debianutils 2.15.1 Miscellaneous utilities specific t ii fontconfig 2.3.2-1.1 generic font configuration library ii libatk1.0-0 1.10.3-1The ATK accessibility toolkit ii libc62.3.5-8.1 GNU C Library: Shared libraries an ii libcairo21.0.2-3 The Cairo 2D vector graphics libra ii libfontconfig1 2.3.2-1.1 generic font configuration library ii libfreetype6 2.1.10-1FreeType 2 font engine, shared lib ii libgcc1 1:4.0.2-5 GCC support library ii libglib2.0-0 2.8.4-2 The GLib library of C routines ii libgtk2.0-0 2.6.10-2The GTK+ graphical user interface ii libidl0 0.8.5-1 library for parsing CORBA IDL file ii libjpeg626b-10 The Independent JPEG Group's JPEG ii libpango1.0-01.8.2-3 Layout and rendering of internatio ii libpng12-0 1.2.8rel-5 PNG library - runtime ii libstdc++6 4.0.2-5 The GNU Standard C++ Library v3 ii libx11-6 6.8.2.dfsg.1-11 X Window System protocol client li ii libxext6 6.8.2.dfsg.1-11 X Window System miscellaneous exte ii libxft2 2.1.7-1 FreeType-based font drawing librar ii libxinerama1 6.8.2.dfsg.1-11 X Window System multi-head display ii libxp6 6.8.2.dfsg.1-11 X Window System printing extension ii libxt6 6.8.2.dfsg.1-11 X Toolkit Intrinsics ii psmisc 21.8-1 Utilities that use the proc filesy ii xlibs6.8.2.dfsg.1-11 X Window System client libraries m ii zlib1g 1:1.2.3-8 compression library - runtime firefox recommends no packages. -- no debconf information -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]