I had this issue already with gcc 3.4.4 and now with 3.4.5. When ever I compile Postfix with the VDA patch, then I get the following error while Postfix delivers to a virtual user: Jan 5 06:58:56 mail postfix/qmgr[14713]: warning: premature end-of-input on private/virtual socket while reading input attribute name Jan 5 06:58:56 mail postfix/qmgr[14713]: warning: private/virtual socket: malformed response Jan 5 06:58:56 mail postfix/qmgr[14713]: warning: transport virtual failure -- see a previous warning/fatal/panic logfile record for the problem description Jan 5 06:58:56 mail postfix/master[14707]: warning: process /usr/lib/postfix/virtual pid 14783 killed by signal 6 Jan 5 06:58:56 mail postfix/master[14707]: warning: /usr/lib/postfix/virtual: bad command startup -- throttling Compiling with GCC 3.3.6 with exactly the same USE flags and CFLAGS/CXXFLAGS works without any problems. Compiling Postfix without the VDA patch / vda USE flag does work without any problem (using GCC 3.3.6, 3.4.4, 3.4.5, 4.0.2, 4.1.0-beta20051230)
you neglected to post `emerge info`
(In reply to comment #1) > you neglected to post `emerge info` > Sorry! Portage 2.0.53 (hardened/x86/2.6, gcc-3.4.5, glibc-2.3.5-r2, 2.6.15 i686) ================================================================= System uname: 2.6.15 i686 AMD Athlon (TM) Gentoo Base System version 1.6.14 distcc 2.18.3 i686-pc-linux-gnu (protocols 1 and 2) (default port 3632) [disabled] ccache version 2.3 [disabled] 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: 2.16.1 sys-devel/libtool: 1.5.20 virtual/os-headers: 2.6.11-r3 ACCEPT_KEYWORDS="x86" AUTOCLEAN="yes" CBUILD="i686-pc-linux-gnu" CFLAGS="-march=athlon-tbird -Os -pipe -mmmx -m3dnow -fforce-addr -fomit-frame-pointer -funroll-loops -falign-functions=4 -maccumulate-outgoing-args" CHOST="i686-pc-linux-gnu" CONFIG_PROTECT="/etc /etc/mail/dspam /etc/postfix /usr/kde/2/share/config /usr/kde/3/share/config /usr/share/config /var/bind /var/qmail/control /var/run/dspam /var/spool/vacation" CONFIG_PROTECT_MASK="/etc/gconf /etc/terminfo /etc/env.d" CXXFLAGS="-march=athlon-tbird -Os -pipe -mmmx -m3dnow -fforce-addr -fomit-frame-pointer -funroll-loops -falign-functions=4 -maccumulate-outgoing-args" DISTDIR="/usr/portage/distfiles" FEATURES="autoconfig buildpkg distlocks sandbox sfperms strict userpriv" GENTOO_MIRRORS="ftp://sunsite.cnlab-switch.ch/mirror/gentoo http://www.ibiblio.org/pub/Linux/distributions/gentoo http://gentoo.oregonstate.edu/" MAKEOPTS="-j2" PKGDIR="/usr/portage/packages" PORTAGE_TMPDIR="/var/tmp" PORTDIR="/usr/portage" PORTDIR_OVERLAY="/mnt/gentoo.overlay" SYNC="rsync://rsync.gentoo.org/gentoo-portage" USE="x86 GAPING_SECURITY_HOLE X509 acl acpi apache2 authdaemond autofs berkdb big-tables bzip2 cdb cgi crypt curl dlloader enscript erandom exif expat extraengine fam fastcgi ffmpeg fla freetds gd gd-external gdbm geoip gif gmp graphviz hardened icc icc-pgo idn imagemagick imap innodb inode ithreads j2ee java javacomm javamail jbig jboss jce jikes jms jmx jpeg jta kerberos krb4 large-domain lcms ldap libwww logrotate lzw lzw-tiff maildir mbox mcal mhash ming mmx mpeg mpm-worker mysql nagios-dns nagios-game nagios-ntp nagios-ping nagios-ssh ncurses netpbm neural nls nptl oav odbc pam pam-mysql pcre pdflib pear-db perl php pic png python readline rrdtool ruby samba sasl session silvercity skey snmp spell sqlite srp sse ssl svg svga svgz symlink tcpd threads tiff truetype udev urandom userlocales vda vhosts virtual-users wmf xml xml2 yaz zlib userland_GNU kernel_linux elibc_glibc" Unset: ASFLAGS, CTARGET, LANG, LC_ALL, LDFLAGS, LINGUAS
with CFLAGS like that i cant imagine why :P does CFLAGS='-O2 -pipe' + gcc-3.4.5 yield a working system ?
(In reply to comment #3) > with CFLAGS like that i cant imagine why :P > > does CFLAGS='-O2 -pipe' + gcc-3.4.5 yield a working system ? > YES! What now? Closing this bug? Or should I try to find out wich flag was/is the problem and then enhance the postfix ebuild to filter that flag out?
If you could isolate the issue and find out what flag causes the problem, that would be helpful. We could filter it, or we could try and actually get a fix for it upstream.
(In reply to comment #5) > If you could isolate the issue and find out what flag causes the problem, that > would be helpful. We could filter it, or we could try and actually get a fix > for it upstream. > Okay. Give me some time to check out serval combinations of diefferent flags. Will keep you informed about my progress.
Works: CFLAGS="-O2 -pipe" CXXFLAGS="-O2 -pipe" emerge -v postfix CFLAGS="-march=athlon-tbird -Os -pipe" CXXFLAGS="-march=athlon-tbird -Os -pipe" emerge -v postfix CFLAGS="-march=athlon-tbird -Os -pipe -mmmx -m3dnow" CXXFLAGS="-march=athlon-tbird -Os -pipe -mmmx -m3dnow" emerge -v postfix CFLAGS="-march=athlon-tbird -Os -pipe -mmmx -m3dnow -fforce-addr -fomit-frame-pointer" CXXFLAGS="-march=athlon-tbird -Os -pipe -mmmx -m3dnow -fforce-addr -fomit-frame-pointer" 571 CFLAGS="-march=athlon-tbird -Os -pipe -mmmx -m3dnow -fforce-addr -fomit-frame-pointer -funroll-loops" CXXFLAGS="-march=athlon-tbird -Os -pipe -mmmx -m3dnow -fforce-addr -fomit-frame-pointer -funroll-loops" emerge -v postfix CFLAGS="-march=athlon-tbird -Os -pipe -mmmx -m3dnow -fforce-addr -fomit-frame-pointer -funroll-loops -falign-functions=4" CXXFLAGS="-march=athlon-tbird -Os -pipe -mmmx -m3dnow -fforce-addr -fomit-frame-pointer -funroll-loops -falign-functions=4" emerge -v postfix CFLAGS="-march=athlon-tbird -Os -pipe -mmmx -m3dnow -fforce-addr -fomit-frame-pointer -funroll-loops -falign-functions=4" CXXFLAGS="-march=athlon-tbird -Os -pipe -mmmx -m3dnow -fforce-addr -fomit-frame-pointer -funroll-loops -falign-functions=4" emerge -v postfix Does not work: CFLAGS="-march=athlon-tbird -Os -pipe -mmmx -m3dnow -fforce-addr -fomit-frame-pointer -funroll-loops -falign-functions=4 -maccumulate-outgoing-args" CXXFLAGS="-march=athlon-tbird -Os -pipe -mmmx -m3dnow -fforce-addr -fomit-frame-pointer -funroll-loops -falign-functions=4 -maccumulate-outgoing-args" emerge -v postfix CFLAGS="-march=athlon-tbird -Os -pipe -mmmx -m3dnow -fforce-addr -fomit-frame-pointer -falign-functions=4 -maccumulate-outgoing-args" CXXFLAGS="-march=athlon-tbird -Os -pipe -mmmx -m3dnow -fforce-addr -fomit-frame-pointer -falign-functions=4 -maccumulate-outgoing-args" emerge -v postfix Looks like the "-maccumulate-outgoing-args" flag is the guilty one. If you have VDA on your setup, could you try to compile with "-maccumulate-outgoing-args" and see if it is failing for you as well? cheers SteveB
This does not work as well: CFLAGS="-Os -pipe -maccumulate-outgoing-args" CXXFLAGS="Os -pipe -maccumulate-outgoing-args" emerge -v postfix It looks that realy the "-maccumulate-outgoing-args" flag is the problem.
Is this issue only reproducible with 3.4.5? Or does 4.x suffer from the same issue? Also, thanks for reducing the flags for me. Makes my life easier :)
(In reply to comment #9) > Is this issue only reproducible with 3.4.5? Or does 4.x suffer from the same > issue? > > Also, thanks for reducing the flags for me. Makes my life easier :) > I don't know. On the systems where I have 4.0.2 and 4.1.0 installed, Postifix does not have any virutal delivery enabled. I would need to change that and enable VDA in order to verify this issue with 4.x. Do you want me to do that? Or do you have by accident a system with GCC 4.x and Postfix and VDA around where you could quickly test that? cheers SteveB
If you could provide me with a minimal setup/set of steps to reproduce this, it would help a lot. (Testing and letting me know on top of that would help as well) :)
Perhaps you want to just filter this flag for the time being.
(In reply to comment #12) > Perhaps you want to just filter this flag for the time being. > Probably this ist the faster way of solving this issue? Currently I am on my way to upgrade a (not so much important) system to gcc 4.0.2 and gcc 4.1.0-beta. Looks like I can't compile Cyrus-IMAPD with any of the 4-series of GCC. This is actualy holding me back from testing the issue with gcc 4.x. However... I did not forgett that I need to do this test. Will keep you informed about my progress.
I tried this with CFLAGS="-march=pentium4 -Os -pipe -mmmx -fforce-addr -fomit-frame-pointer -funroll-loops -falign-functions=4 -maccumulate-outgoing-args" as well as with my usual CFLAGS="-march=pentium4 -O2 -pipe -ggdb", but delivering to virtual users works all the time. Perhaps you can post your postfix setup relevant to virtual domains. I have used a simple setup: main.cf: virtual_alias_domains = foo.sk virtual_alias_maps = hash:/etc/postfix/virtual virtual: ticho@foo.sk ticho @foo.sk root BTW, I'm also using gcc-3.4.5 with the same profile as you. Which gcc profile do you use? (gcc-config -l)
(In reply to comment #14) > I tried this with CFLAGS="-march=pentium4 -Os -pipe -mmmx -fforce-addr > -fomit-frame-pointer -funroll-loops -falign-functions=4 > -maccumulate-outgoing-args" as well as with my usual CFLAGS="-march=pentium4 > -O2 -pipe -ggdb", but delivering to virtual users works all the time. > > Perhaps you can post your postfix setup relevant to virtual domains. I have > used a simple setup: > > main.cf: > virtual_alias_domains = foo.sk > virtual_alias_maps = hash:/etc/postfix/virtual > > virtual: > ticho@foo.sk ticho > @foo.sk root > > BTW, I'm also using gcc-3.4.5 with the same profile as you. Which gcc profile > do you use? (gcc-config -l) > Try adding quota to the ticho@foo.sk mailbox and then try again. More info here: http://web.onda.com.br/nadal/
Ok, I tried this: main.cf: virtual_mailbox_limit_maps = hash:/etc/postfix/vquota vquota: ticho@foo.sk 100000000 Still, I get the mail okay: Jan 23 23:15:17 src@thelair postfix/postfix-script: stopping the Postfix mail system Jan 23 23:15:17 src@thelair postfix/master[20220]: terminating on signal 15 Jan 23 23:15:18 src@thelair postfix/postfix-script: starting the Postfix mail system Jan 23 23:15:18 src@thelair postfix/master[27982]: daemon started -- version 2.2.5, conf iguration /etc/postfix Jan 23 23:15:21 src@thelair postfix/smtpd[2999]: connect from localhost[127.0.0.1] Jan 23 23:15:47 src@thelair postfix/smtpd[2999]: 28A583472F: client=localhost[127.0.0.1] Jan 23 23:15:57 src@thelair postfix/cleanup[24601]: 28A583472F: message-id=<200601232215 29.28A583472F@thelair.ynet.sk> Jan 23 23:15:57 src@thelair postfix/qmgr[3228]: 28A583472F: from=<asd@foo.cz>, size=368, nrcpt=1 (queue active) Jan 23 23:15:57 src@thelair postfix/local[31993]: 28A583472F: to=<ticho@thelair.ynet.sk> , orig_to=<ticho@foo.sk>, relay=local, delay=28, status=sent (delivered to command: /usr /bin/procmail) Jan 23 23:15:57 src@thelair postfix/qmgr[3228]: 28A583472F: removed Jan 23 23:15:59 src@thelair postfix/smtpd[2999]: disconnect from localhost[127.0.0.1] This all with ricey CFLAGS as mentioned above. Perhaps try removing your -march (or -m3dnow) ? Anyway, this is best raised with postfix upstream developers, as they are the ones who can fix the code.
This doesn't look like a toolchain issue since Ticho isn't having the same problems.
I think we should close this entry. It is now 1 year old. Can I close it?
(In reply to comment #18) > I think we should close this entry. It is now 1 year old. Can I close it? > That depends on whether you're still having the issue or not.
(In reply to comment #19) > (In reply to comment #18) > > I think we should close this entry. It is now 1 year old. Can I close it? > > > > That depends on whether you're still having the issue or not. > I moved ahead with GCC and with Postfix and with VDA. So to be honest: I have not anymore this issue. But I as well don't have anymore the setup described in this bug report. So from my viewpoint we can close this bug.
Fair enough. I'm glad it's working for you now.
