On a RELENG_4 system (sources as of Jan 13 @16:45 UT), a "make all" in /usr/ports/emulators/linux_base-8 produces: ======================================================================== scatest2# make all ===> Vulnerability check disabled, database not found ===> Extracting for linux_base-8-8.0_4 => Checksum OK for rpm/i386/8.0/redhat-release-8.0-8.noarch.rpm. => Checksum OK for rpm/i386/8.0/glibc-common-2.3.2-4.80.8.i386.rpm. => Checksum OK for rpm/i386/8.0/glibc-2.3.2-4.80.8.i386.rpm. => Checksum OK for rpm/i386/8.0/setup-2.5.20-1.noarch.rpm. => Checksum OK for rpm/i386/8.0/filesystem-2.1.6-5.noarch.rpm. => Checksum OK for rpm/i386/8.0/basesystem-8.0-1.noarch.rpm. => Checksum OK for rpm/i386/8.0/libattr-2.0.8-3.i386.rpm. => Checksum OK for rpm/i386/8.0/libacl-2.0.11-2.i386.rpm. => Checksum OK for rpm/i386/8.0/libelf-0.8.2-2.i386.rpm. => Checksum OK for rpm/i386/8.0/bzip2-libs-1.0.2-5.i386.rpm. => Checksum OK for rpm/i386/8.0/termcap-11.0.1-13.noarch.rpm. => Checksum OK for rpm/i386/8.0/compat-db-3.3.11-2.i386.rpm. => Checksum OK for rpm/i386/8.0/db4-4.0.14-14.i386.rpm. => Checksum OK for rpm/i386/8.0/gdbm-1.8.0-18.i386.rpm. => Checksum OK for rpm/i386/8.0/glib-1.2.10-8.i386.rpm. => Checksum OK for rpm/i386/8.0/libtermcap-2.0.8-31.i386.rpm. => Checksum OK for rpm/i386/8.0/bash-2.05b-5.i386.rpm. => Checksum OK for rpm/i386/8.0/bzip2-1.0.2-5.i386.rpm. => Checksum OK for rpm/i386/8.0/compat-libstdc++-7.3-2.96.110.i386.rpm. => Checksum OK for rpm/i386/8.0/ncurses-5.2-28.i386.rpm. => Checksum OK for rpm/i386/8.0/info-4.2-5.i386.rpm. => Checksum OK for rpm/i386/8.0/grep-2.5.1-4.i386.rpm. => Checksum OK for rpm/i386/8.0/fileutils-4.1.9-11.i386.rpm. => Checksum OK for rpm/i386/8.0/popt-1.7-1.06.i386.rpm. => Checksum OK for rpm/i386/8.0/readline-4.3-3.i386.rpm. => Checksum OK for rpm/i386/8.0/setserial-2.17-9.i386.rpm. => Checksum OK for rpm/i386/8.0/libstdc++-3.2-7.i386.rpm. => Checksum OK for rpm/i386/8.0/slang-1.4.5-11.i386.rpm. => Checksum OK for rpm/i386/8.0/sh-utils-2.0.12-3.i386.rpm. => Checksum OK for rpm/i386/8.0/rpm-4.1-1.06.i386.rpm. => Checksum OK for rpm/i386/8.0/libgcc-3.2-7.i386.rpm. => Checksum OK for rpm/i386/8.0/freetype-2.1.5-0.ximian.5.1.i386.rpm. => Checksum OK for rpm/i386/8.0/zlib-1.1.4-8.8x.i386.rpm. ===> linux_base-8-8.0_4 depends on executable: rpm - found kern.fallback_elf_brand: 3 -> 3 redhat-release-8.0-8.noarch.rpm glibc-common-2.3.2-4.80.8.i386.rpm execution of glibc-common-2.3.2-4.80.8 script failed, exit status 0 *** Error code 1 Stop in /usr/ports/emulators/linux_base-8. scatest2# ======================================================================= The following message is logged in /var/log/messages and a core file is left in /usr/ports/emulators/linux_base-8/work/linux_base-8-8.0: Jan 14 13:27:15 scatest2 /kernel: pid 1975 (build-locale-arc), uid 0: exited on signal 11 The problem is repeatable. I have installed this port before on recent RELENG_4 systems, so I suspect that it has something to do with the recent changes to the linux emulator. Since linux_base-8 appears to be the default for 4.11-RELEASE, this may affect the release bits. Fix: N/A How-To-Repeat: "make all" in /usr/ports/emulators/linux_base-8 on a very recent RELENG_4
Responsible Changed From-To: freebsd-ports-bugs->emulation Over to maintainers
I observe exactly the same problem here: FreeBSD plexi 4.11-STABLE FreeBSD 4.11-STABLE #0: Thu Jan 13 18:39:22 JST 2005 devin@plexi.pun-pun.prv:/usr/obj/usr/src/4-stable/src/sys/PLEXI i386 but not here: FreeBSD kanemochi 4.10-STABLE FreeBSD 4.10-STABLE #0: Fri Nov 19 17:30:57 JST 2004 devin@kanemochi.tky.swissbank.com:/usr/obj/usr/src/sys/KANEMOCHI i386 (dates correspond to cvs checkout date) So it may also have something to do with recent changes in RELENG_4 -- Tod McQuillin
I MFCed a bit more of the Linux kernel emulation into RELENG_4 yesterday. That seems to fix this problem. I'll probably move it into RELENG_4_11 later today. This is the patch, if anyone has some time to give it a try on their machines (or if you're tracking RELENG_4 just update): Index: sys/i386/linux/linux_sysvec.c =================================================================== RCS file: /home/ncvs/src/sys/i386/linux/linux_sysvec.c,v retrieving revision 1.55.2.9 retrieving revision 1.55.2.10 diff -u -r1.55.2.9 -r1.55.2.10 --- sys/i386/linux/linux_sysvec.c 12 Jan 2002 11:03:30 -0000 1.55.2.9+++ sys/i386/linux/linux_sysvec.c 18 Jan 2005 22:11:12 -0000 1.55.2.10 @@ -722,6 +722,7 @@ args[2] = tf->tf_edx; args[3] = tf->tf_esi; args[4] = tf->tf_edi; + args[5] = tf->tf_ebp; /* Unconfirmed */ *params = NULL; /* no copyin */ }
State Changed From-To: open->closed Another piece of the -current Linux kernel emulation was MFCed to RELENG_4 yesterday that seems to fix this. Scott has confirmed it fixed the problem for him. I'll merge it into RELENG_4_11 before the release.