When the acrobat reader is running as a plugin in Mozilla is makes the "X" process use >80% CPU time. I have only noticed this in Gentoo. I used Debian in the past and never had this problem. Either it is a new version of Adobe Acrobat that is doing this, or it could be an ebuild problem.
This only happens with the plugin, not running acrobat separately.
I cannot quite reproduce the problem (it doesn't consume CPU time so much). Please post output of "emerge info", "emerge -pv mozilla" and "emerge -pv acroread". Does upgrading mozilla (or mozilla-firebird) or/and acroread solve the problem?
Before opening a PDF in mozilla: PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ Command 3665 root 15 0 300m 43m 2756 S 3.2 8.7 152:28.84 X 18964 david 16 0 996 996 780 R 2.3 0.2 0:00.23 top 3853 david 15 0 19760 19m 15m S 1.3 3.8 90:17.44 kdeinit After opening a PDF in mozilla: PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ Command 3665 root 25 0 302m 45m 2756 R 71.9 9.1 152:10.09 X 19221 david 16 0 84052 81m 30m S 21.9 16.1 8:09.63 mozilla-bin 18963 david 16 0 996 996 780 R 2.3 0.2 0:00.22 top 3853 david 15 0 19760 19m 15m S 1.6 3.8 90:16.68 kdeinit root@sidicpc22 david # emerge info Portage 2.0.48-r5 (default-x86-1.4, gcc-3.2.3, glibc-2.3.2-r1) ================================================================= System uname: 2.4.20-gentoo-r5-fb i686 AMD Athlon(tm) Processor GENTOO_MIRRORS="http://gentoo.oregonstate.edu http://distro.ibiblio.org/pub/Linux/distributions/gentoo" CONFIG_PROTECT="/etc /var/qmail/control /usr/kde/2/share/config /usr/kde/3/share/config /usr/X11R6/lib/X11/xkb /usr/kde/3.1/share/config /usr/share/config" CONFIG_PROTECT_MASK="/etc/gconf /etc/env.d" PORTDIR="/usr/portage" DISTDIR="/usr/portage/distfiles" PKGDIR="/usr/portage/packages" PORTAGE_TMPDIR="/var/tmp" PORTDIR_OVERLAY="/usr/local/portage" USE="x86 oss 3dnow apm avi crypt cups encode foomaticdb gif jpeg libg++ mad mmx mpeg ncurses nls pdflib png quicktime spell truetype xml2 xmms xv zlib directfb alsa gdbm berkdb slang readline arts tetex aalib nas svga tcltk java mysql postgres X sdl gpm tcpd pam libwww ssl perl python imlib oggvorbis qt kde motif opengl mozilla cdr artswrappersuid atlas curl emacs fbcon gd gtk2 guile joystick mcal mozcalendar mozp3p mozsvg mpi offensive pda pic plotutils ppds samba tiff type1 usb v4l wxwindows xvid -mikmod -esd -gnome -gtk" COMPILER="gcc3" CHOST="i686-pc-linux-gnu" CFLAGS="-march=athlon-tbird -O3 -pipe -fomit-frame-pointer" CXXFLAGS="-march=athlon-tbird -O3 -pipe -fomit-frame-pointer" ACCEPT_KEYWORDS="x86" MAKEOPTS="-j2" AUTOCLEAN="yes" SYNC="rsync://rsync.gentoo.org/gentoo-portage" FEATURES="sandbox ccache" root@sidicpc22 david # emerge -vp mozilla [ebuild R ] net-www/mozilla-1.4-r3 +java +crypt -ipv6 +gtk2 +ssl -ldap -gnome -debug +mozsvg +mozcalendar -mozaccess +mozp3p -mozxmlterm -moznoirc -moznomail -moznocompose -moznoxft root@sidicpc22 david # emerge -vp acroread [ebuild R ] app-text/acroread-5.07
Hum, optimize flag may cause the problem, so please remove "-march=athlon-tbird" from CFLAGS and CXXFLAGS, re-emerge mozilla (you might need to re-emerge X as well) and try again? Debian doesn't optimize their packages to specific architecture (they usually only set "-O" and no -march nor -mcpu flag), they may not encounter the problem.
I haven't tried fixing it yet, although I think your advice may be right. Is it possible that we can force certain flags to not be used in this ebuild if they do in fact cause this problem? I'll let you know if I get around to re-building acroread.
Yes, we can disable certain flags using strip-flags, filter-flags, replace-flags and so on from flag-o-matic.eclass. You may want to have a look at flag-o-matic man page or /usr/portage/eclass/flag-o-matic.eclass directly. If you look into mozilla ebuild you will see most of your CFLAGS/CXXFLAGS are removed or replaced. If some of your CFLAGS/CXXFLAGS turn out to cause this problem we can add them to the list of flags to be removed and disable them.
I can confirm this problem. Additional after closing the tab that includes the acrobat reader window, the subprocess won't be killed.
Carlo, can you post your "emerge info" please?
Sure. Portage 2.0.48-r5 (default-x86-1.4, gcc-3.2.3, glibc-2.2.5-r2,2.3.2-r1) ================================================================= System uname: 2.4.20-gentoo-r6 i686 AMD Duron(tm) Processor GENTOO_MIRRORS="ftp://sunsite.informatik.rwth-aachen.de/pub/Linux/gentoo ftp://ftp.snt.utwente.nl/pub/os/linux/gentoo ftp://ftp.berlios.de/pub/gentoo-deutsch http://www.ibiblio.org/pub/Linux/distributions/gentoo" CONFIG_PROTECT="/etc /var/qmail/control /usr/kde/2/share/config /usr/kde/3/share/config /var/bind /usr/X11R6/lib/X11/xkb /usr/kde/3.1/share/config /usr/share/config" CONFIG_PROTECT_MASK="/etc/gconf /etc/env.d" PORTDIR="/usr/portage" DISTDIR="/usr/portage/distfiles" PKGDIR="/usr/portage/packages" PORTAGE_TMPDIR="/var/tmp" PORTDIR_OVERLAY="/usr/local/portage" USE="x86 oss 3dnow apm avi crypt cups encode foomaticdb gif jpeg gnome libg++ mad mikmod mmx mpeg ncurses nls pdflib png spell truetype xml2 xmms xv zlib directfb gtkhtml alsa gdbm berkdb slang readline arts tetex aalib nas bonobo svga ggi tcltk java ruby mysql postgres X sdl gpm tcpd pam libwww ssl perl python imlib oggvorbis gtk qt kde motif opengl mozilla gphoto2 ldap snmp cdr scanner acl acpi acpi4linux artswrappersuid curl dga dnd doc dvd emacs faad fbcon firebird gd gtk2 hbci imap innodb jack jack-tmpfs jikes junit ladcca lcms libgda lirc maildir mbox mng moznocompose moznoirc moznomail mozsvg mozxmlterm oav openal plotutils pnp samba sasl slp sse tiff usb vim-with-x wmf xml xvid zeo -quicktime -guile -esd" COMPILER="gcc3" CHOST="i686-pc-linux-gnu" CFLAGS="-O3 -march=athlon-xp -pipe -foptimize-sibling-calls -fomit-frame-pointer -funroll-loops -fforce-addr -msse -mmmx -m3dnow " CXXFLAGS="-O2 -mcpu=i686 -pipe" ACCEPT_KEYWORDS="x86" MAKEOPTS="-j2" AUTOCLEAN="yes" SYNC="rsync://rsync.europe.gentoo.org/gentoo-portage" FEATURES="sandbox ccache keeptemp keepwork"
Have a look at http://forums.gentoo.org/viewtopic.php?p=470249#470249. I think, it may be the same problem. I never had a version of Mozilla installed that was 100% stable, btw..
Ok, I re-emerged mozilla using CFLAGS="" and I'm still getting 80%+ CPU usage by the X process. Next, I'll try re-emerging X using CFLAGS="" and see if that fixes it.
I have re-emerged X and Mozilla now with CFLAGS="" and I'm still getting this 80%+ CPU usage by the "X" process. Any advice would be greatly appeciated.
Thanks for testing hard. I really appreciate your effort.I have no idea about the problem but I'm going to have access toDuron box next week, so I will give it a try.
I'm sorry to say that I didn't succeed to reproduce the problem. I attach `emerge info` at the Duron box I tested. I pretty appreciate all of your help and assistance. If new info will be available, please reopen the bug. ginger% emerge -pv acroread These are the packages that I would merge, in order: Calculating dependencies ...done! [ebuild R ] app-text/acroread-5.08 +cjk ginger% emerge -pv mozilla These are the packages that I would merge, in order: Calculating dependencies ...done! [ebuild R ] net-www/mozilla-1.4-r3 +java +crypt -ipv6 -gtk2 +ssl -ldap +gnome -debug -mozsvg -mozcalendar -mozaccess -mozp3p -mozxmlterm -moznoirc -moznomail -moznocompose -moznoxft ginger% emerge -pv xfree These are the packages that I would merge, in order: Calculating dependencies ...done! [ebuild R ] x11-base/xfree-4.3.0-r3 -3dfx -sse -mmx -3dnow +xml2 +truetype +nls +cjk -doc -ipv6 -debug -static +pam -sdk -gatos ginger% emerge info Portage 2.0.49-r13 (default-x86-1.4, gcc-3.2.3, glibc-2.3.2-r1, 2.4.21) ================================================================= System uname: 2.4.21 i686 AMD Duron(tm) Processor Gentoo Base System version 1.4.3.10p1 distcc 2.9 i686-pc-linux-gnu (protocols 1 and 2) (default port 3632) [disabled] ccache version 2.3 [enabled] ACCEPT_KEYWORDS="x86" AUTOCLEAN="yes" CFLAGS="-mcpu=athlon -march=i586 -O -pipe" CHOST="i686-pc-linux-gnu" COMPILER="gcc3" CONFIG_PROTECT="/etc /var/qmail/control /usr/kde/2/share/config /usr/kde/3/share /config /usr/X11R6/lib/X11/xkb /usr/kde/3.1/share/config /usr/share/texmf/tex/generic/config/ /usr/share/texmf/tex/platex/config/ /usr/share/texmf/dvips/config/ /usr/share/texmf/dvipdfm/config/ /usr/share/texmf/xdvi/ /usr/share/config" CONFIG_PROTECT_MASK="/etc/gconf /etc/env.d" CXXFLAGS="-mcpu=athlon -march=i586 -O -pipe" DISTDIR="/home/distfiles" FEATURES="sandbox ccache autoaddcvs digest buildpkg" GENTOO_MIRRORS="ftp://ftp.ecc.u-tokyo.ac.jp/GENTOO http://mh.sodan.ecc.u-tokyo.ac.jp/pub/linux/gentoo http://ftp.gentoo.or.kr/ http://gentoo.gnukorea.org/ http://www.zentek-international.com/mirrors/gentoo/ http://www.zentek-international.com/mirrors/gentoo/ http://212.219.56.146/sites/www.ibiblio.org/gentoo/ http://gentoo.oregonstate.edu http://www.ibiblio.org/pub/Linux/distributions/gentoo" MAKEOPTS="-j2" PKGDIR="/home/packages/i586" PORTAGE_TMPDIR="/var/tmp" PORTDIR="/usr/portage" PORTDIR_OVERLAY="/home/gentoo-x86" SYNC="rsync://sb.itc.u-tokyo.ac.jp/gentoo-portage" USE="x86 oss apm avi crypt cups encode foomaticdb gif jpeg libg++ mad mikmod mpeg ncurses pdflib png quicktime spell xml2 xmms xv zlib gdbm berkdb slang readline arts aalib bonobo svga tcltk java guile ruby sdl gpm tcpd pam libwww perl python esd oggvorbis qt kde motif opengl cdr X gtk gnome cjk nls imlib ssl canna emacs freewnn mozilla truetype linguas_ja"
I just tried Opera today, and acroread has 0% CPU usage when a PDF file is open inside Opera. So it is definitely not an acroread problem... Also, do you think this can be narrowed down to a mozilla problem, as opposed to an X problem? This is my CFLAGS: CFLAGS="-march=athlon-tbird -O3 -pipe -fomit-frame-pointer" when I tried to emerge with no optimizations, I just did: CFLAGS="" emerge mozilla xfree Is this correct? Did the optimizations actually get removed? I'm not sure how to change flags the fancy way, using flag-o-matic. It seems that these require me to edit the ebuild, but that seems like a less robust way than say, changing make.conf, since it will get stomped on when I do emerge sync right? I also noticed that your flags were: CFLAGS="-mcpu=athlon -march=i586 -O -pipe" I didn't use the -mcpu option, should I? I also used the march=athlon-tbird option. Is this bad? Should I ust i586, i686, or just athlon instead?
Right. If you want to make sure you get the desired CFLAGS/CXXFLAGS,you can have a look at /var/db/pkg/[category]/[pkgname]/CFLAGS (and CXXFLAGS). As for flag-o-matic, you need to modify the ebuild to employ CFLAGS sanitisation, so using enviroment variable as you did is the easiest way to configure CFLAGS/CXXFLAGS. I don't want to use too strong -march CFLAGS because it will create "illegal instruction" error (I use created tbz2 files on different machines) but on the other hand I want to have a benefit of CPU optimisation (that's why I add -mcpu=athlon flag though -march=i586 automatically adds -mcpu=i586). -mcpu does optimise binaries but the binaries will run on other machines. If Opera works fine with Acrobat Reader, can you try net-www/mozilla-firebird-bin? It is a pre-built binary so that your CFLAGS won't affect it.
I already had mozilla-firebird-bin installed (ver 6.1) and it worked fine. acroread was at 0% and X was at 6% (normal level). I guess this is to be expected, I used binaries in Debian (from Debian unstable as well as from mozilla.org) and I never had this problem. I'll keep working at it. One thing to try is to compile firebird and work with that instead of mozilla. That will allow me to continue to use mozilla today with no monkeying around, while I compile and test mozilla-firebird in the background.
It is happening in mozilla-firebird, compiled today by me using: CFLAGS="-march=athlon-tbird -O3 -pipe -fomit-frame-pointer" which is what has always been in /etc/make.conf file I will do some more testing now, by re-compiling.
I edited my make.conf file directly, to change CFLAGS, this was the simplest way for me without messing around too much. I first removed the arch flag completely, then I removed O3, each with no success. Now my CFLAGS according to /var/db/pkg/net-www/mozilla-firebird/CFLAGS is the following: -pipe -Wno-return-type -w -Wno-return-type -w -Wno-return-type -w -s -fforce-addr It looks like -fforce-addr is the only optimization left in there! That is added by the mozilla-firebird ebuild for whatever reason, along with -s (which is harmless). However configure is using --enable-optimize=-O2 and from looking at the output while compiling, it looks like -O2 is being used. I'll try removing this manually in the ebuild, then recompiling, and see what happens.
Do anybody know what gcc mozilla.org/adobe uses to compile their binary packages? It may just be another ABI issue ... ?
From http://www.mozilla.org/releases/mozilla1.4/README.html#new-issues "The Linux binaries distributed by mozilla.org are now compiled with GCC 3.2" Not sure of the subversion. I have a question. What does ABI issue mean? reply by email if you like.
It must be terrible code, that they were not able to fix these three problems from 1.4 to 1.5: - netstat zombie process - acrobat sucking cpu time - loosing focus instead they're trying to improve these included email/composer components... :(
I see some bugs on Mozilla.org about this. I think it might be a gtk issue... I'm compiling with USE="-gtk2" right now. I don't know why I never thought of checking mozilla's bugzilla before! I guess I assumed if mozilla.org's binaries worked then it must not be a mozilla.org problem. silly me...
I can confirm I have this problem too...it struck me as odd, because I just reinstalled my beautifully stable Gentoo 1.2 installation to put on Gentoo 1.4, and this problem jumped out straight away. Emerge info out put: Portage 2.0.49-r15 (default-x86-1.4, gcc-3.2.3, glibc-2.3.2-r1, 2.4.20-gentoo-r7) ================================================================= System uname: 2.4.20-gentoo-r7 i686 Intel(R) Pentium(R) III Mobile CPU 1000MHz Gentoo Base System version 1.4.3.10p1 ACCEPT_KEYWORDS="x86" AUTOCLEAN="yes" CFLAGS="-O3 -march=pentium3 -fprefetch-loop-arrays -funroll-loops -pipe" CHOST="i686-pc-linux-gnu" COMPILER="gcc3" CONFIG_PROTECT="/etc /var/qmail/control /usr/share/config /usr/kde/2/share/config /usr/kde/3/share/config /usr/X11R6/lib/X11/xkb:/usr/kde/3.1/share/config:/usr/share/config" CONFIG_PROTECT_MASK="/etc/gconf /etc/env.d" CXXFLAGS="-O3 -march=pentium3 -fprefetch-loop-arrays -funroll-loops -pipe" DISTDIR="/usr/portage/distfiles" FEATURES="sandbox ccache autoaddcvs" GENTOO_MIRRORS="http://csociety-ftp.ecn.purdue.edu/pub/gentoo/ ftp://csociety-ftp.ecn.purdue.edu/pub/gentoo/" MAKEOPTS="-j2" PKGDIR="/usr/portage/packages" PORTAGE_TMPDIR="/var/tmp" PORTDIR="/usr/portage" PORTDIR_OVERLAY="" SYNC="rsync://rsync.gentoo.org/gentoo-portage" USE="x86 oss apm avi crypt cups encode foomaticdb gif jpeg gnome libg++ libwww mad mikmod mpeg ncurses nls pdflib png quicktime spell truetype xml2 xmms xv zlib gdbm berkdb slang readline arts bonobo svga java ruby mysql postgres X sdl gpm tcpd pam python imlib oggvorbis gtk qt kde motif opengl aalib acl alsa cdr directfb dvd esd ethereal evo fbcon gb gd ggi gstreamer gtk2 gtkhtml imap lcms ldap mozilla perl plotutils samba scanner slp ssl tcltk tiff usb wmf Xaw3d xinerama xosd xml"
I think I've solved the bug. It just had to do with the gtk2 USE flag. Get rid of that and the problem goes away. I think that Gentoo should force the gtk2 crap to NOT be used in their distribution until the problem is fixed upstream.
mozilla also seems a bit snappier now without gtk2. esp. startup seemed faster
What a great job! Thanks for tracking down the problem. I can reproduce the same problem on my machine. It occurs when I use mozilla based brouwsers (I tried mozilla, epiphany, and kazehakase) built with GTK+2 and CPU usage of X sticks to 65%. I don't have the problem with mozilla and firebird built with GTK+1. This is a bug for sure, but there are a number of people who want mozilla with GTK+2, so warning about this bug in pkg_postinst of mozilla (and other GTK+2 mozilla embedded browsers) might be good enough for it? (cf. GNOME 2.4 requires epiphany and epiphany depends on GTK+2 enabled mozilla)
Ok, I'm fine with putting it in postinst. But we could put it in pkg_preinst to save people the compile time. Actually pkg_preinst will run after compile before before install, so maybe we can put a little einfo message in src_unpack? BTW, I have noticed that I lost anti-aliasing in mozilla menus and toolbars and tab headings. It's not that bad though. But I still have in in web pages, thanks to mozilla xft I guess.
Good point. We could use pkg_setup for that purpose.(Personally I don't compile mozilla with gtk2 so I haven't comeacross this problem)
Man, mozilla is MUCH snappier now. I'm so much happier now. Ooo, that rhymes. I use mozilla more than any other program I think. Actually this is a funny story, this is the reason I switched over to Gentoo from Debian... I found that the Debian binaries ran much slower than the binaries from moz.org. It was twice as slow to start up, and marking a pile of messages as Junk took twice as long. I filed a bug about this to the debian mozilla guy, I think it's Tatuo-somthing, and the problem was never solved. I compiled my own Mozilla a few times from CVS but that was a pain, so I switched to Gentoo so that a machine could automate the compiling for me. Then I got hooked on it. But now I realize, it must have been because Debian's package was compiled with gtk2 support. What a relief this is solved.... The slowness went way beyond acrobat problems...
Then you want the message rather in pkg_setup() if you want it before its even unpacking stuff ....
First cut at it: in pkg_setup() or in moz_setup() (called by src_unpack) if [ -n "`use gtk2`"] then einfo "You have the gtk2 flag enabled. This can cause a slowdown in Mozilla's performance especially when using the acroread plugin to view a PDF file. Consider disabling this flag unless you know that you require it." fi please revise as you see fit
Mozilla 1.5 just came out recently. I'd like to see something added to the mozilla ebuild about this USE="gtk2" issue. Actually, this is a bug in gtk2, so I think it is a serious issue and we need a workaround, if not a serious warning to the user in the form of some einfo statements.
Created attachment 23267 [details] Patch to fix this bug. Please commit this
While the PDF viewer consistently triggers the CPU suck problem, it goes beyond PDF viewing. I just had epiphany (current stable) with two windows, one window being gentoo.org, the other window linuxtoday.com, and cpu usage was pegged. I had had epiphany up just a few minutes. I was trying Epiphany to see if that worked better than Mozilla, so problem is in the underlying common library. Email me if you want my USE flags, etc. and I'll post them. By box is AMD 2200 XP UP, 1.5G RAM. This is frustrating. I find myself killing and restarting mozilla several times a day, and when cpu usage is pegged just killing mozilla can take a minute. I also find myself running 'emerge sync emerge -u sys/wor' seq often hoping that something in gentoo will straighten this out soon. But that thrashes both your servers and my box and so far hasn't fixed Mozilla. What's the current recommended work around? Re-emerging with gtk2 off? Is there a fix upstream that will percolate down to stable soon? Thanks
re-emergeing with gtk2 flag off is the only solution. But I heard that epiphany has to be gtk2, and gnome 2.4 requires epiphany. Is that right? I'm not sure if it's a mozilla problem or a gtk problem. I knew this before but I forgot, and I can't find anything on google about it right now.
Same problem, same solution, if gtk2 is off acrobat works as expected.
*** Bug 39689 has been marked as a duplicate of this bug. ***
Why not just change ( gtk2? >=x11-libs/gtk+-2.2.0 : =x11-libs/gtk+-1.2* ) ( gtk2? >=dev-libs/glib-2.2.0 : =dev-libs/glib-1.2* ) to be just: =dev-libs/gtk+-1.2* =dev-libs/glib-1.2*
Ok, I just spoke with some people on #gentoo and it seems the warning route would be better as in the patch that is already attached.
I agree, so can someone commit this change please? I don't know why there is a 1 month delay to get a simple patch in, with 2 @gentoo.org addresses on the CC list and one assignee (mozilla team).
Mozilla ebuild has no maintainer and belongs to mozilla herd, so mozilla herd takes first precedence in changing the ebuild. If mozilla herd allows me to commit the change, I'll be glad to do that. azarah and brad, what do you guys think about this bug?
http://bugzilla.mozilla.org/show_bug.cgi?id=198954 The GTK+2 port of mozilla is important for a few reasons, not the least of which: GNOME now depends on Mozilla being built with GTK+2, as does galeon GTK+2 has far more support for i18n and other standards that will make Linux on the desktop sucessful while GTK+1 will not GTK+1 is deprecated and no longer developed, and will eventually create more bugs than it solves A suggested workaround is to use 'plugger' to embed the adobe acrobat program until the issue is resolved upstream - but since it's just one plugin wouldn't it make more sense to place an advisory in the adobe acrobat ebuild? Telling people to build Mozilla against GTK+1 or it will eat their babies would cause a lot of confusion.
Telling someone in the Acrobat ebuild may cause someone to have to rebuild mozilla. The wording of the patch can be changed: --- mozilla-1.6_beta.ebuild 2003-12-20 13:37:49.000000000 -0500 +++ mozilla-1.6_beta.ebuild.mod 2004-01-06 18:49:13.000000000 -0500 @@ -97,6 +97,15 @@ dev-lang/perl java? ( >=dev-java/java-config-0.2.0 )" +pkg_setup() { + if [ -n "`use gtk2`" ] + then + einfo "You have the gtk2 flag enabled. This can cause a slowdown" + einfo "in Mozilla's performance when using the acroread" + einfo "plugin to view a PDF file. Consider disabling this flag" + einfo "if you use the the Adobe Acrobat plugin with Mozilla." + fi +} moz_setup() {
Hi, I used to have similar speed problems with acrobat plugin in mozila. The problem was DRI enabled in X windows. Disable the "Load "dri"" line in XF86Config-4 file, restart xdm and try again. It should go away. I use radeon from dri.sourceforge.net (aka xfree-drm) module.
This is isn't just a "speed" program. I don't have dri enabled in my XFree86 config file, I use the nvidia module/driver.
No comments so far from azarah or brad. This is needs to be implemented or another alternative found. I used Gentoo for a LONG time with this problem and it is really critical IMO.
Brad wrote his comment in Comment #43. I'm also inclined to add an advisory to acrobat reader's ebuild. David, are you ok with it?
Yes, putting the warning with the plugin's ebuild makes more sense. As long as it gets out there.
Added an advisory to acroread-5.08.ebuild. Thanks.