Home | Docs | Forums | Lists | Bugs | Planet | Store | GMN | Get Gentoo!
Not eligible to see or edit group visibility for this bug.
View Bug Activity | Format For Printing | XML | Clone This Bug
As per summary: 1) emerge =rtorrent-0.8.2 (which is a C++ application) 2) Watch in horror as one's available RAM and swap are saturated within mere seconds and the OOM starts shooting down everything in sight. Confirmed reproducible using gcc-3.4.6-r2 (with any available specs) on x86 and amd64. gcc-4.1.2 confirmed as working in a standard 2008.0_beta2 chroot on the same system. Strictly speaking, this probably isn't a hardened bug at all but I think it belongs with the herd as, let's face it, no-one else cares much about gcc-3* anymore.
Same here: Portage 2.1.4.4 (hardened/amd64/multilib, gcc-3.4.6, glibc-2.6.1-r0, 2.6.23-hardened-r12 x86_64) ================================================================= System uname: 2.6.23-hardened-r12 x86_64 AMD Athlon(tm) 64 X2 Dual Core Processor 5600+ Timestamp of tree: Wed, 28 May 2008 23:45:01 +0000 app-shells/bash: 3.2_p33 dev-java/java-config: 1.3.7, 2.1.6 dev-lang/python: 2.4.4-r13 dev-python/pycrypto: 2.0.1-r6 sys-apps/baselayout: 1.12.11.1 sys-apps/sandbox: 1.2.18.1-r2 sys-devel/autoconf: 2.61-r1 sys-devel/automake: 1.5, 1.7.9-r1, 1.9.6-r2, 1.10.1 sys-devel/binutils: 2.18-r1 sys-devel/gcc-config: 1.4.0-r4 sys-devel/libtool: 1.5.26 virtual/os-headers: 2.6.23-r3 ACCEPT_KEYWORDS="amd64" CBUILD="x86_64-pc-linux-gnu" CFLAGS="-march=athlon64 -O2 -pipe" CHOST="x86_64-pc-linux-gnu" CONFIG_PROTECT="/etc" CONFIG_PROTECT_MASK="/etc/env.d /etc/env.d/java/ /etc/fonts/fonts.conf /etc/gconf /etc/php/apache2-php5/ext-active/ /etc/php/cgi-php5/ext-active/ /etc/php/cli-php5/ext-active/ /etc/revdep-rebuild /etc/terminfo /etc/texmf/web2c /etc/udev/rules.d" CXXFLAGS="-march=athlon64 -O2 -pipe" DISTDIR="/usr/portage/distfiles" FEATURES="autoconfig distlocks metadata-transfer noinfo parallel-fetch sandbox sfperms strict unmerge-orphans userfetch" GENTOO_MIRRORS="ftp://mirror.q-mex.net/gentoo/ http://linux.rz.ruhr-uni-bochum.de/download/gentoo-mirror/ http://ftp-stud.fht-esslingen.de/pub/Mirrors/gentoo/" LC_ALL="en_US.UTF-8" MAKEOPTS="-j3" PKGDIR="/usr/portage/packages" PORTAGE_RSYNC_OPTS="--recursive --links --safe-links --perms --times --compress --force --whole-file --delete --stats --timeout=180 --exclude=/distfiles --exclude=/local --exclude=/packages" PORTAGE_TMPDIR="/var/tmp" PORTDIR="/usr/portage" PORTDIR_OVERLAY="/usr/portage/local/layman/webapps-experimental /usr/portage/local/local" SYNC="rsync://rsync.europe.gentoo.org/gentoo-portage" USE="7zip aim amd64 authdaemond bash-completion berkdb bitmap-fonts bzip2 caps cgi chroot cli cracklib crypt cscope ctype curl curlwrappers dynamicplugin encode expat fam fastcgi fasttrack flatfile ftp gd gdbm geoip gif gnutella gnutls gpgme hardened iconv idn imagemagick imap iproute2 ithreads jpeg jpeg2k kqemu libg++ libwww lighttpd logrotate maildir mailwrapper mime mng mudflap multilib mysql ncurses nls nptl nptlonly ntlm offensive ogg pam pcre pdf perl php pic png pop python readline reflection rrdtool ruby sasl session slang smime smtp snmp socks5 spell spl sse sse2 ssl svg sysfs tcpd theora threads tidy tiff tordns truetype truetype-fonts type1-fonts unicode userlocales vhosts vim x264 xml xvid zip zlib" ALSA_CARDS="ali5451 als4000 atiixp atiixp-modem bt87x ca0106 cmipci emu10k1x ens1370 ens1371 es1938 es1968 fm801 hda-intel intel8x0 intel8x0m maestro3 trident usb-audio via82xx via82xx-modem ymfpci" ALSA_PCM_PLUGINS="adpcm alaw asym copy dmix dshare dsnoop empty extplug file hooks iec958 ioplug ladspa lfloat linear meter mulaw multi null plug rate route share shm softvol" APACHE2_MODULES="actions alias auth_basic authn_alias authn_anon authn_dbm authn_default authn_file authz_dbm authz_default authz_groupfile authz_host authz_owner authz_user autoindex cache dav dav_fs dav_lock deflate dir disk_cache env expires ext_filter file_cache filter headers include info log_config logio mem_cache mime mime_magic negotiation rewrite setenvif speling status unique_id userdir usertrack vhost_alias" ELIBC="glibc" INPUT_DEVICES="mouse keyboard" KERNEL="linux" LCD_DEVICES="bayrad cfontz cfontz633 glk hd44780 lb216 lcdm001 mtxorb ncurses text" USERLAND="GNU" VIDEO_CARDS="apm ark chips cirrus cyrix dummy fbdev glint i128 i810 mach64 mga neomagic nv r128 radeon rendition s3 s3virge savage siliconmotion sis sisusb tdfx tga trident tseng v4l vesa vga via vmware voodoo" Unset: CPPFLAGS, CTARGET, EMERGE_DEFAULT_OPTS, INSTALL_MASK, LANG, LDFLAGS, LINGUAS, PORTAGE_COMPRESS, PORTAGE_COMPRESS_FLAGS, PORTAGE_RSYNC_EXTRA_OPTS
CCing Timothy Redaelli here as he is listed as a maintainer in metadata.xml. I think package.masking net-p2p/rtorrent-0.8.2 at least on hardened profiles would be appropriate in the mean time. Will save users from this unhappy surprise.
Yes, either that or check gcc-major-version in the ebuild itself until we figure out just what to do. I don't believe that the problem is connected with the hardened toolchain; I'll seek to confirm one way or the other at some point today.
*** This bug has been marked as a duplicate of bug 216118 ***
This is not a duplicate of bug 216118 (which is really solved), re-opening.
Ok, my mistake. Using MAKEOPTS="-j1" The build does appear to be failing on dht_manager.cc as in bug 216118. However, that bug is against an SVN version, not against a version in portage as this bug is. Also compiling the SVN version stated there results in immediate segfault without the memory saturation problem, so the failure is different. For these reasons that bug should be marked a dupe of this one IMO. make[3]: Entering directory `/var/tmp/portage/net-p2p/rtorrent-0.8.2/work/rtorrent-0.8.2/src/core' i686-pc-linux-gnu-g++ -DHAVE_CONFIG_H -I. -I../.. -I. -I./.. -I../.. -march=athlon64 -msse3 -O2 -fweb -fno-ident -pipe -fno-strict-aliasing -DNDEBUG -I/usr/include/sigc++-2.0 -I/usr/lib/sigc++-2.0/include -I/usr/include -c -o curl_get.o curl_get.cc i686-pc-linux-gnu-g++ -DHAVE_CONFIG_H -I. -I../.. -I. -I./.. -I../.. -march=athlon64 -msse3 -O2 -fweb -fno-ident -pipe -fno-strict-aliasing -DNDEBUG -I/usr/include/sigc++-2.0 -I/usr/lib/sigc++-2.0/include -I/usr/include -c -o curl_stack.o curl_stack.cc i686-pc-linux-gnu-g++ -DHAVE_CONFIG_H -I. -I../.. -I. -I./.. -I../.. -march=athlon64 -msse3 -O2 -fweb -fno-ident -pipe -fno-strict-aliasing -DNDEBUG -I/usr/include/sigc++-2.0 -I/usr/lib/sigc++-2.0/include -I/usr/include -c -o dht_manager.o dht_manager.cc In file included from ./../rpc/command_map.h:45, from ./../rpc/parse_commands.h:43, from dht_manager.cc:46: ./../rpc/command.h: In function `rpc::target_type rpc::get_target_left(const rpc::target_type&)': ./../rpc/command.h:153: internal compiler error: in type_unification_real, at cp/pt.c:9134 Please submit a full bug report, ...
*** Bug 216118 has been marked as a duplicate of this bug. ***
(In reply to comment #6) > [...] Also compiling the SVN version stated there results in immediate > segfault without the memory saturation problem, so the failure is different. > For these reasons that bug should be marked a dupe of this one IMO. once again: when using the custom ebuild or the make command (even on an svn source), it does not throw segfault, it starts eating up all the memory and the swapfile, until all of it is gone [...] so i still think, it's the same bug, and has nothing to do where do i download the source from (svn, tar.gz from the official libtorrent page, or a gentoo mirror)
Well in that case I apologize. That was not the behavior I observed when I tried it so perhaps my observations were wrong. Regardless this bug has more useful build output, more complete CC list, possible temporary courses of action and since it is actually IN portage seems to be now getting attention where as the last one received almost none outside myself who reproduced it shortly after you reported it.
This is p.masked in the hardened profile for now.
libtorrent 0.12.0 should also be masked on hardened, since the good old rtorrent 0.7.9 won't compile if this one is installed.
Was rtorrent-0.8.0 removed for this reason as well? I went to rebuild rtorrent, and after masking 0.8.2*, it's going back to 0.7.9. I had 0.8.0 working fine, but I have to go back to 0.7.9 now because it's not available anymore, and I'm still getting 0.8.2-r2 in my list, which doesn't build still.
Re Comment 12 - that rtorrent-0.8.0 was removed has nothing to do with this bug. If it works for you then your best bet is to retrieve the ebuild from here: http://sources.gentoo.org/viewcvs.py/gentoo-x86/net-p2p/rtorrent/?hideattic=0 Then place it an overlay and use it until such time as we (may) have a toolchain that's actually supported both by gentoo and upstream.
Brave option to resolve: https://hardened.gentooexperimental.org/secure/ Regards, Dw.
With today's sync updates brings libtorrent-0.12.2-r3, curl-7.18.2 and rtorrent-0.8.2-r3. An emerge --update --deep world results in rtorrent's 0.8.2-r3 a failed build. emerge --info: Portage 2.1.4.4 (hardened/linux/x86, gcc-3.4.6, glibc-2.6.1-r0, 2.6.24-hardened-r3 i686) ================================================================= System uname: 2.6.24-hardened-r3 i686 Intel(R) Pentium(R) M processor 1300MHz Timestamp of tree: Wed, 30 Jul 2008 07:45:03 +0000 app-shells/bash: 3.2_p33 dev-lang/python: 2.5.2-r5 sys-apps/baselayout: 1.12.11.1 sys-apps/sandbox: 1.2.18.1-r2 sys-devel/autoconf: 2.61-r2 sys-devel/automake: 1.10.1 sys-devel/binutils: 2.18-r3 sys-devel/gcc-config: 1.4.0-r4 sys-devel/libtool: 1.5.26 virtual/os-headers: 2.6.23-r3 ACCEPT_KEYWORDS="x86" CBUILD="i686-pc-linux-gnu" CFLAGS="-O2 -march=pentium-m -pipe -fomit-frame-pointer" CHOST="i686-pc-linux-gnu" CONFIG_PROTECT="/etc" CONFIG_PROTECT_MASK="/etc/env.d /etc/gconf /etc/revdep-rebuild /etc/terminfo /etc/udev/rules.d" CXXFLAGS="-O2 -march=pentium-m -pipe -fomit-frame-pointer" DISTDIR="/usr/portage/distfiles" FEATURES="ccache collision-protect distlocks metadata-transfer parallel-fetch sandbox sfperms strict unmerge-orphans userfetch" GENTOO_MIRRORS="http://cesium.di.uminho.pt/pub/gentoo/ ftp://cesium.di.uminho.pt/pub/gentoo/ http://ftp.dei.uc.pt/pub/linux/gentoo/ ftp://ftp.dei.uc.pt/pub/linux/gentoo/ " LC_ALL="en_US.UTF-8" LDFLAGS="-Wl,-O1" MAKEOPTS="-j2" PKGDIR="/usr/portage/packages" PORTAGE_RSYNC_OPTS="--recursive --links --safe-links --perms --times --compress --force --whole-file --delete --stats --timeout=180 --exclude=/distfiles --exclude=/local --exclude=/packages" PORTAGE_TMPDIR="/var/tmp" PORTDIR="/usr/portage" PORTDIR_OVERLAY="/usr/portage/local/layman/sunrise" SYNC="rsync://rsync.europe.gentoo.org/gentoo-portage" USE="acpi bash-completion berkdb cli cracklib crypt dbus dri gpm hal hardened iconv midi mudflap ncurses nptl nptlonly openmp pam pcre perl pic pppd python readline reflection samba session spl ssl symlink tcpd thread threads truetype type1 unicode urandom x86 xorg zlib" ALSA_CARDS="intel8x0 intel8x0m" ALSA_PCM_PLUGINS="adpcm alaw asym copy dmix dshare dsnoop empty extplug file hooks iec958 ioplug ladspa lfloat linear meter mulaw multi null plug rate route share shm softvol" APACHE2_MODULES="actions alias auth_basic authn_alias authn_anon authn_dbm authn_default authn_file authz_dbm authz_default authz_groupfile authz_host authz_owner authz_user autoindex cache dav dav_fs dav_lock deflate dir disk_cache env expires ext_filter file_cache filter headers include info log_config logio mem_cache mime mime_magic negotiation rewrite setenvif speling status unique_id userdir usertrack vhost_alias" ELIBC="glibc" INPUT_DEVICES="keyboard mouse evdev synaptics" KERNEL="linux" LCD_DEVICES="bayrad cfontz cfontz633 glk hd44780 lb216 lcdm001 mtxorb ncurses text" USERLAND="GNU" VIDEO_CARDS="radeon" Unset: CPPFLAGS, CTARGET, EMERGE_DEFAULT_OPTS, INSTALL_MASK, LANG, LINGUAS, PORTAGE_COMPRESS, PORTAGE_COMPRESS_FLAGS, PORTAGE_RSYNC_EXTRA_OPTS
>=net-p2p/rtorrent-0.8.2 is package.masked on hardened profiles until someone adds a check for >=gcc-4 such as that in media-video/cinelerra. If you unmask it, as it appears you have, that is NOP.
My system is _not_ hardened (amd64, gcc 3.4.6) and I get this saturation issue. Daniel
Thanks Daniel. We are aware that this is a bug in the affected version of gcc and that it doesn't specifically apply to the hardened toolchain. However, as I alluded to in my original report, it is unrealistic to expect anyone else outside of the hardened herd to be at all concerned. The silence from the toolchain herd and the package maintainer is unlikely to be broken. Therefore, unless a better solution can be found, and as gcc-3.4.6-r2 is still the "supported" compiler in Hardened Gentoo, it seems that the only way to get this bug truly fixed is for the herd to either fix the existing compiler or provide a supported gcc-4.x release (whichever comes first).
(In reply to comment #16) > >=net-p2p/rtorrent-0.8.2 is package.masked on hardened profiles until someone adds a check for >=gcc-4 such as that in media-video/cinelerra. If you unmask it, as it appears you have, that is NOP. > The package is not masked on hardened profiles. That is why I had to mask it alongside with libtorrent-0.12.2-r3 to able to emerge rtorrent.
(In reply to comment #19) > (In reply to comment #16) > > >=net-p2p/rtorrent-0.8.2 is package.masked on hardened profiles until someone adds a check for >=gcc-4 such as that in media-video/cinelerra. If you unmask it, as it appears you have, that is NOP. > > > > The package is not masked on hardened profiles. That is why I had to mask it > alongside with libtorrent-0.12.2-r3 to able to emerge rtorrent. > I should clarify. # eselect profile show Current make.profile symlink: /usr/portage/profiles/hardened/linux/x86 # cat /etc/portage/package.mask >=net-p2p/rtorrent-0.8.2-r3 >=net-libs/libtorrent-0.12.2-r3 >=net-misc/curl-7.18.2 By masking as above I compile the packages without problems. Here's the default eix output for rtorrent: # eix rtorrent [U] net-p2p/rtorrent Available versions: 0.7.9 0.8.2-r3 {debug ipv6 xmlrpc} Installed versions: 0.7.9(01:36:46 PM 07/30/2008)(-debug -xmlrpc) Homepage: http://libtorrent.rakshasa.no/ Description: BitTorrent Client using libtorrent Hope that helps.
Created an attachment (id=161905) [edit] rt_triple patch The attached patch was lifted from an upstream bug report: http://libtorrent.rakshasa.no/ticket/1271 I can confirm that this leads to a clean compile with gcc-3.4.6
(In reply to comment #19) > The package is not masked on hardened profiles. ... Yes, it is. http://sources.gentoo.org/viewcvs.py/gentoo-x86/profiles/hardened/package.mask?rev=1.41&view=log (In reply to comment #20) > # eselect profile show > Current make.profile symlink: > /usr/portage/profiles/hardened/linux/x86 Therein lies the problem. http://archives.gentoo.org/gentoo-hardened/msg_ce5a9e2f10213b51b91d04326509338e.xml (In reply to comment #21) > Created an attachment (id=161905) [edit] > rt_triple patch > > The attached patch was lifted from an upstream bug report: > http://libtorrent.rakshasa.no/ticket/1271 > > I can confirm that this leads to a clean compile with gcc-3.4.6 > Nice find! Doesn't look like upstream rtorrent cares about gcc 3.4.6 and will not be taking this patch. Hopefully one of the maintainers of the package are willing to carry it. Otherwise please add the requested check for gcc version. We p.masked this on Hardened to protect our users, but obviously that doesn't cover anyone using gcc 3.4 in "regular" Gentoo. CCing loki_val <at> g.o since he is now listed as a maintainer as well.
Thanks Gordon. :) Wasn't aware of the profile issue.
+ 01 Aug 2008; Peter Alfredsen <loki_val@gentoo.org> + +files/rtorrent-0.8.2-gcc34.patch, rtorrent-0.8.2-r3.ebuild: + Add patch for gcc-3.4, bug #224031. Thanks goes to Herbie Hopkins + <herbie@hopkins.net>. Upstream http://libtorrent.rakshasa.no/ticket/1271 . + Verified with Josef Drexler via IRC that this is the correct solution. + It should work for you guys now too. It works here on ~x86 w/ gcc-4.3.1-r1. Please verify and unmask as appropriate. Leaving status as it is for the same purpose.
To both Peter and Herbie, many thanks indeed. The patch in Comment 21 does indeed solve the problem (tested on an Opteron system here). I just wish I knew how this could be addressed in the compiler itself.
It sounds like this is fixed then? Reopen if not.
Works for me on hardened. Somebody needs to unmask it (I had to add it to package.unmask today).
(In reply to comment #27) > Works for me on hardened. Somebody needs to unmask it.. Done.
