| Summary: | sys-kernel/vanilla-sources-3.9.4-r1 with PaX - Oops when try to load openvswitch module again - IP: [<ffffffffa0027015>] ovs_init_net+0x15/0x3c [openvswitch] | ||
|---|---|---|---|
| Product: | Gentoo Linux | Reporter: | Marcin Mirosław <bug> |
| Component: | [OLD] Core system | Assignee: | Gentoo Kernel Bug Wranglers and Kernel Maintainers <kernel> |
| Status: | RESOLVED TEST-REQUEST | ||
| Severity: | normal | CC: | hardened-kernel+disabled, pageexec, spender |
| Priority: | Normal | ||
| Version: | unspecified | ||
| Hardware: | All | ||
| OS: | Linux | ||
| Whiteboard: | |||
| Package list: | Runtime testing required: | --- | |
| Attachments: | kernel config | ||
|
Description
Marcin Mirosław
2013-06-01 21:16:34 UTC
# emerge --info Portage 2.1.11.62 (hardened/linux/amd64, gcc-4.7.3, glibc-2.15-r3, 3.9.4-hardened-r1 x86_64) ================================================================= System uname: Linux-3.9.4-hardened-r1-x86_64-Intel-R-_Xeon-R-_CPU_E3-1230_V2_@_3.30GHz-with-gentoo-2.2 KiB Mem: 16450984 total, 10936604 free KiB Swap: 4193264 total, 4193264 free Timestamp of tree: Sat, 01 Jun 2013 04:15:01 +0000 ld GNU gold (GNU Binutils 2.22) 1.11 ccache version 3.1.9 [enabled] app-shells/bash: 4.2_p45 dev-lang/python: 2.7.3-r3, 3.2.3-r2 dev-util/ccache: 3.1.9 dev-util/cmake: 2.8.10.2-r2 dev-util/pkgconfig: 0.28 sys-apps/baselayout: 2.2 sys-apps/openrc: 0.11.8 sys-apps/sandbox: 2.5 sys-devel/autoconf: 2.69 sys-devel/automake: 1.11.6, 1.12.6 sys-devel/binutils: 2.22-r1 sys-devel/gcc: 4.7.3, 4.8.0 sys-devel/gcc-config: 1.7.3 sys-devel/libtool: 2.4-r1 sys-devel/make: 3.82-r4 sys-kernel/linux-headers: 3.7 (virtual/os-headers) sys-libs/glibc: 2.15-r3 Repositories: gentoo qemu-init ACCEPT_KEYWORDS="amd64" ACCEPT_LICENSE="* -@EULA" CBUILD="x86_64-pc-linux-gnu" CFLAGS="-O2 -pipe -march=native -frecord-gcc-switches -fno-unwind-tables -fno-asynchronous-unwind-tables -fexpensive-optimiza tions" CHOST="x86_64-pc-linux-gnu" CONFIG_PROTECT="/etc" CONFIG_PROTECT_MASK="/etc/ca-certificates.conf /etc/env.d /etc/gconf /etc/gentoo-release /etc/revdep-rebuild /etc/sandbox.d /etc/terminfo" CXXFLAGS="-O2 -pipe -march=native -frecord-gcc-switches -fno-unwind-tables -fno-asynchronous-unwind-tables -fexpensive-optimi zations" DISTDIR="/usr/portage/distfiles" FCFLAGS="-O2 -pipe" FEATURES="assume-digests binpkg-logs ccache collision-protect config-protect-if-modified distlocks ebuild-locks fixlafiles merge-sync news p arallel-fetch preserve-libs protect-owned sandbox sfperms strict unknown-features-warn unmerge-logs unmerge-orphans userfetch userpriv users andbox xattr" FFLAGS="-O2 -pipe" GENTOO_MIRRORS="http://distfiles.gentoo.org" LANG="pl_PL.utf8" LDFLAGS="-Wl,-O1 -Wl,--as-needed -Wl,--sort-common" MAKEOPTS="-j8" PKGDIR="/usr/portage/packages" PORTAGE_CONFIGROOT="/" PORTAGE_RSYNC_EXTRA_OPTS="-O" PORTAGE_RSYNC_OPTS="--recursive --links --safe-links --perms --times --compress --force --whole-file --delete --stats --human-readable --tim eout=180 --exclude=/distfiles --exclude=/local --exclude=/packages" PORTAGE_TMPDIR="/var/tmp" PORTDIR="/usr/portage" PORTDIR_OVERLAY="/var/lib/layman/qemu-init" SYNC="rsync://gentoo-mirror/gentoo-portage" USE="acl acpi amd64 avx bash-completion caps custom-cflags cxx hardened hwdb iconv mmxext multilib nls openmp sse2 sse3 sse4 sse41 sse4_1 ss se3 threads udev unicode vim-syntax xattr" ABI_X86="64" CURL_SSL="openssl" ELIBC="glibc" GRUB_PLATFORMS="multiboot pc" KERNEL="linux" LINGUA S="en" PYTHON_TARGETS="python2_7 python3_2" QEMU_SOFTMMU_TARGETS="i386 x86_64" QEMU_USER_TARGETS="i386 x86_64" USERLAND="GNU" Unset: CPPFLAGS, CTARGET, EMERGE_DEFAULT_OPTS, INSTALL_MASK, LC_ALL, PORTAGE_BUNZIP2_COMMAND, PORTAGE_COMPRESS, PORTAGE_COMPRESS_FLAGS, USE _PYTHON Created attachment 349874 [details]
kernel config
Certainly I can't reproduce it... (In reply to Marcin Mirosław from comment #0) > I'm not sure if it can be somehow related to bug #469500. I didn't recompile > net-misc/openvswitch with new kernel. how about you recompile the module for the current kernel and try that instead? mismatched versions are never a good idea, CONFIG_MODVERSIONS exists for a reason. I'm using module from kernel not from net-misc/openvswitch - I just noticed it. So it was used module from running kernel. I don't have good enviroment to try reproduce problem because it's production machine. (In reply to Marcin Mirosław from comment #5) > I'm using module from kernel not from net-misc/openvswitch - I just noticed > it. So it was used module from running kernel. in that case it'd be nice if you could eventually test a vanilla kernel as well, right now i don't see what in grsec could cause this. In second try I couldn't reproduce it. If I find way to reproduce it on hardened then I'll try vanilla kernel. Now I have hard lock of host when I'm trying to do `rmmmod openvswitch`. Hard lock appears on vanilla-sources-3.9.7 also. it looks grsec changes nothing in kernel behaviour. (In reply to Marcin Mirosław from comment #8) > Now I have hard lock of host when I'm trying to do `rmmmod openvswitch`. > Hard lock appears on vanilla-sources-3.9.7 also. it looks grsec changes > nothing in kernel behaviour. Okay this is purely a vanilla issue, so I'll send it their way. 1. Is something written to log (eg. /var/log/messages) before the hard lock? 2. Are you still able to us Magic SysRq? [1] 3. Is this sys-apps/kmod or something else? Which version? 3. Which kernel version was the last working version? 4. Can you try a later version (3.9.7) and a development version (3.10-rc7)? It would be nice to know if it work(ed|s) at some point so we can inspect what changes have been done that have caused this. Also, if there is a chance that there is some kind of output being written before hard lock then that would be very useful debugging information. [1]: http://en.wikipedia.org/wiki/Magic_SysRq_key Ad.1. No Ad.2. Rather no. Firstly I have access to console using IPMI - it isn't very reliable. Pressing any key doesn't wake up console from blanking, I didn't test magic Sysrq. I'll try to do it. Ad.3. [ebuild R ] sys-apps/kmod-12-r1 USE="tools -debug -doc -lzma -static-libs -zlib" Ad.4. I didn't try. I'm using openvswitch newly Ad.5. I'll try. It's production machine, I can't reboot it too frequently:) I can't promise I'll do it in this week. I'll try to reproduce problem in VM, it would be easier to make tests. Have you since been able to reproduce this? Did you try a more recent kernel?
As for the original call trace, I just noted this line:
> [36079.368177] [<ffffffff81031056>] ? native_pax_close_kernel+0x26/0x40
That call is not present in a vanilla kernel, that makes inspecting this harder.
Could you obtain us a new call trace for vanilla sources without PaX?
I've tried to reproduce problem on host and inside vm (with gentoo-sources and with hardened-sources). Without success. I've also tried on 3.9.4-hardened-r1, again without success. (In reply to Marcin Mirosław from comment #13) > I've tried to reproduce problem on host and inside vm (with gentoo-sources > and with hardened-sources). Without success. I've also tried on > 3.9.4-hardened-r1, again without success. I'm confused. Is the bug in comment 0 reproduceable or not? Anthony, when I did tests two days ago I couldn't reproduce it. Maybe because I changed a little configuration of openvswitch, dunno. Maybe it was a some kind of glitch. (In reply to Marcin Mirosław from comment #15) > Anthony, when I did tests two days ago I couldn't reproduce it. Maybe > because I changed a little configuration of openvswitch, dunno. Maybe it was > a some kind of glitch. Please reopen this bug if you get this again in the future, thank you in advance. |
