Report forwarded
to debian-bugs-dist@lists.debian.org, GNU/kFreeBSD Maintainers <debian-bsd@lists.debian.org>: Bug#743141; Package kfreebsd-9.
(Sun, 30 Mar 2014 21:51:07 GMT) (full text, mbox, link).
Acknowledgement sent
to Ivo De Decker <ivo.dedecker@ugent.be>:
New Bug report received and forwarded. Copy sent to GNU/kFreeBSD Maintainers <debian-bsd@lists.debian.org>.
(Sun, 30 Mar 2014 21:51:07 GMT) (full text, mbox, link).
package: kfreebsd-9
severity: serious
version: 9.2-2
Hi,
The kfreebsd-amd64 buildds seem to reboot when building gcc-4.[6789]. I was
able to reproduce this issue in a sid pbuilder chroot on a fresh jessie
install. With kfreebsd-9, I get a crash (screenshot in attach). I also tried
with kfreebsd-10 (from jessie) and kfreebsd-11 (from experimental), and both
of them reboot (I don't have a screenshot for those).
The buildds probably run the kernels for stable. In that case, the found
version for this bug should be updated.
If there is any specific test I can run to try to pinpoint the issue, feel
free to let me know.
Cheers,
Ivo
Added indication that 743141 affects gcc-4.6
Request was from Julien Cristau <julien.cristau@logilab.fr>
to control@bugs.debian.org.
(Wed, 23 Apr 2014 08:39:07 GMT) (full text, mbox, link).
Added indication that 743141 affects gcc-4.7
Request was from Julien Cristau <julien.cristau@logilab.fr>
to control@bugs.debian.org.
(Wed, 23 Apr 2014 08:39:08 GMT) (full text, mbox, link).
Added indication that 743141 affects gcc-4.8
Request was from Julien Cristau <julien.cristau@logilab.fr>
to control@bugs.debian.org.
(Wed, 23 Apr 2014 08:39:09 GMT) (full text, mbox, link).
Added indication that 743141 affects gcc-4.9
Request was from Julien Cristau <julien.cristau@logilab.fr>
to control@bugs.debian.org.
(Wed, 23 Apr 2014 08:39:10 GMT) (full text, mbox, link).
Changed Bug title to 'system hangs when building gcc on kfreebsd-amd64' from 'system hangs when building gcc on kreebsd-amd64'
Request was from Niels Thykier <niels@thykier.net>
to control@bugs.debian.org.
(Wed, 23 Apr 2014 17:21:15 GMT) (full text, mbox, link).
Information forwarded
to debian-bugs-dist@lists.debian.org, GNU/kFreeBSD Maintainers <debian-bsd@lists.debian.org>: Bug#743141; Package kfreebsd-9.
(Wed, 23 Apr 2014 17:27:14 GMT) (full text, mbox, link).
Acknowledgement sent
to Niels Thykier <niels@thykier.net>:
Extra info received and forwarded to list. Copy sent to GNU/kFreeBSD Maintainers <debian-bsd@lists.debian.org>.
(Wed, 23 Apr 2014 17:27:14 GMT) (full text, mbox, link).
Cc: Debian Release Team <debian-release@lists.debian.org>,
"debian-bsd@lists.debian.org" <debian-bsd@lists.debian.org>
Subject: system hangs when building gcc on kfreebsd-amd64
Date: Wed, 23 Apr 2014 19:22:20 +0200
Dear BSD porters,
Do you have any news on this bug? It is severely affecting *all* GCC
versions and prevents new versions of them from migrating to testing.
~Niels
Added tag(s) help.
Request was from Steven Chamberlain <steven@pyro.eu.org>
to control@bugs.debian.org.
(Wed, 23 Apr 2014 18:00:05 GMT) (full text, mbox, link).
Information forwarded
to debian-bugs-dist@lists.debian.org, GNU/kFreeBSD Maintainers <debian-bsd@lists.debian.org>: Bug#743141; Package kfreebsd-9.
(Fri, 25 Apr 2014 18:12:15 GMT) (full text, mbox, link).
Acknowledgement sent
to Petr Salinger <Petr.Salinger@seznam.cz>:
Extra info received and forwarded to list. Copy sent to GNU/kFreeBSD Maintainers <debian-bsd@lists.debian.org>.
(Fri, 25 Apr 2014 18:12:15 GMT) (full text, mbox, link).
Subject: Re: system hangs when building gcc on kfreebsd-amd64
Date: Fri, 25 Apr 2014 19:38:10 +0200 (CEST)
> Do you have any news on this bug? It is severely affecting *all* GCC
> versions and prevents new versions of them from migrating to testing.
I tested it:
system up-to-date jessie, with kfreebsd-image-10.0-1-amd64/10.0-4
chroot up-to-date sid, building of gcc-4.6 4.6.4-6 rebooted the system.
It seems to be during test of "gcc unix/-m32".
Petr
Information forwarded
to debian-bugs-dist@lists.debian.org, GNU/kFreeBSD Maintainers <debian-bsd@lists.debian.org>: Bug#743141; Package kfreebsd-9.
(Mon, 28 Apr 2014 20:21:05 GMT) (full text, mbox, link).
Acknowledgement sent
to Steven Chamberlain <steven@pyro.eu.org>:
Extra info received and forwarded to list. Copy sent to GNU/kFreeBSD Maintainers <debian-bsd@lists.debian.org>.
(Mon, 28 Apr 2014 20:21:05 GMT) (full text, mbox, link).
Subject: Re: Bug#743141: system hangs when building gcc on kfreebsd-amd64
Date: Mon, 28 Apr 2014 21:18:59 +0100
On 25/04/14 18:38, Petr Salinger wrote:
> I tested it:
> system up-to-date jessie, with kfreebsd-image-10.0-1-amd64/10.0-4
Same here. I used a plain ufs filesystem and no swap.
> It seems to be during test of "gcc unix/-m32".
Yes. I reproduced this with kernel messages going to a serial console
(booted with -D). Manually triggering a panic (sysctl
debug.kdb.panic=1) normally outputs a stack trace from KDB to serial,
but this crash did not. I think it rebooted instantly. Obviously
nothing is logged to syslog either.
I was building gcc-4.7 with -j1 as a non-root user from a remote SSH
session, so I had a record of the very last console messages:
> Running target unix/-m32
> Using /usr/share/dejagnu/baseboards/unix.exp as board description file for target.
> Using /usr/share/dejagnu/config/unix.exp as generic interface file for target.
> Using /home/steven/gcc-4.7-4.7.3/src/gcc/testsuite/config/default.exp as tool-and-target-specific interface file.
> Running /home/steven/gcc-4.7-4.7.3/src/gcc/testsuite/gcc.c-torture/compile/compile.exp ...
This is the third part of the gcc testsuite. Previous sections are
called "unix" and "unix/-fstack-protector"; both of those had finished
before this.
It seems possible to get much more output during the testsuite by
creating a ~/.dejagnurc containing "set verbose 2". I've started
another run of this now but it may take several hours.
Regards,
--
Steven Chamberlain
steven@pyro.eu.org
Information forwarded
to debian-bugs-dist@lists.debian.org, GNU/kFreeBSD Maintainers <debian-bsd@lists.debian.org>: Bug#743141; Package kfreebsd-9.
(Wed, 30 Apr 2014 23:09:04 GMT) (full text, mbox, link).
Acknowledgement sent
to Steven Chamberlain <steven@pyro.eu.org>:
Extra info received and forwarded to list. Copy sent to GNU/kFreeBSD Maintainers <debian-bsd@lists.debian.org>.
(Wed, 30 Apr 2014 23:09:04 GMT) (full text, mbox, link).
Subject: Re: Bug#743141: system hangs when building gcc on kfreebsd-amd64
Date: Thu, 01 May 2014 00:04:48 +0100
To reproduce this, the unix/-m32 test suite can be invoked with:
$ cd ~/gcc-4.7-4.7.3/src/gcc/testsuite
$ make check RUNTESTFLAGS='--target_board "unix/-m32" -v -v'
First, many compiler regression tests are run successfully.
Later, some helper executable tls_runtime2029.exe is built. I think the
number part varies on each run.
When it is executed (with more stuff coming from LD_LIBRARY_PATH), as
non-root user, the system reboots before it can show a panic message or
drop to KDB.
I can't find a tls_runtime2029.c source file after rebooting. The crash
means tls_runtime2029.exe contains only junk after a reboot. I might
try again with the sync mount option.
Last console messages before the crash:
> Checking pattern "sparc-*-sunos*" with x86_64-pc-kfreebsd-gnu
> Checking pattern "alpha*-*-*" with x86_64-pc-kfreebsd-gnu
> Checking pattern "hppa*-*-hpux*" with x86_64-pc-kfreebsd-gnu
> Checking pattern "sparc-*-sunos*" with x86_64-pc-kfreebsd-gnu
> Checking pattern "alpha*-*-*" with x86_64-pc-kfreebsd-gnu
> Checking pattern "hppa*-*-hpux*" with x86_64-pc-kfreebsd-gnu
> Testing compile/pr44246.c, -O2 -flto -fuse-linker-plugin -fno-fat-lto-objects
> doing compile
> Invoking the compiler as ~/gcc-4.7-4.7.3/build/gcc/xgcc -B~/gcc-4.7-4.7.3/build/gcc/ -O2 -flto -fuse-linker-plugin -fno-fat-lto-objects -w -c -m32 -o pr44246.o ~/gcc-4.7-4.7.3/src/gcc/testsuite/gcc.c-torture/compile/pr44246.c
> Setting timeout to 300
> Executing on host: ~/gcc-4.7-4.7.3/build/gcc/xgcc -B~/gcc-4.7-4.7.3/build/gcc/ -O2 -flto -fuse-linker-plugin -fno-fat-lto-objects -w -c -m32 -o pr44246.o ~/gcc-4.7-4.7.3/src/gcc/testsuite/gcc.c-torture/compile/pr44246.c (timeout = 300)
> pid is 76488 -76488
> waitres is 76488 exp8 0 0
> output is status 0
> Checking pattern "sparc-*-sunos*" with x86_64-pc-kfreebsd-gnu
> Checking pattern "alpha*-*-*" with x86_64-pc-kfreebsd-gnu
> Checking pattern "hppa*-*-hpux*" with x86_64-pc-kfreebsd-gnu
> Checking pattern "sparc-*-sunos*" with x86_64-pc-kfreebsd-gnu
> Checking pattern "alpha*-*-*" with x86_64-pc-kfreebsd-gnu
> Checking pattern "hppa*-*-hpux*" with x86_64-pc-kfreebsd-gnu
> Testing compile/pr44686.c, -O0
> Profiling argument is <-fprofile-generate>
> Checking "alpha*-dec-osf*" against "x86_64-pc-kfreebsd-gnu"
> Checking "*-*-solaris2.[89]*" against "x86_64-pc-kfreebsd-gnu"
> check_cached_effective_target tls_runtime: checking unix/-m32
> check_compile tool: gcc for tls_runtime
> doing compile
> Invoking the compiler as ~/gcc-4.7-4.7.3/build/gcc/xgcc -B~/gcc-4.7-4.7.3/build/gcc/ tls_runtime2029.c -lm -m32 -o tls_runtime2029.exe
> Setting timeout to 300
> Executing on host: ~/gcc-4.7-4.7.3/build/gcc/xgcc -B~/gcc-4.7-4.7.3/build/gcc/ tls_runtime2029.c -lm -m32 -o tls_runtime2029.exe (timeout = 300)
> pid is 76496 -76496
> waitres is 76496 exp8 0 0
> output is status 0
> Checking pattern "sparc-*-sunos*" with x86_64-pc-kfreebsd-gnu
> Checking pattern "alpha*-*-*" with x86_64-pc-kfreebsd-gnu
> Checking pattern "hppa*-*-hpux*" with x86_64-pc-kfreebsd-gnu
> loading to unix/-m32
> Setting LD_LIBRARY_PATH to :~/gcc-4.7-4.7.3/build/gcc:~/gcc-4.7-4.7.3/build/gcc/32::~/gcc-4.7-4.7.3/build/gcc:~/gcc-4.7-4.7.3/build/gcc/32
> spawning command ./tls_runtime2029.exe
Regards,
--
Steven Chamberlain
steven@pyro.eu.org
Information forwarded
to debian-bugs-dist@lists.debian.org, GNU/kFreeBSD Maintainers <debian-bsd@lists.debian.org>: Bug#743141; Package kfreebsd-9.
(Thu, 01 May 2014 00:12:04 GMT) (full text, mbox, link).
Acknowledgement sent
to Steven Chamberlain <steven@pyro.eu.org>:
Extra info received and forwarded to list. Copy sent to GNU/kFreeBSD Maintainers <debian-bsd@lists.debian.org>.
(Thu, 01 May 2014 00:12:04 GMT) (full text, mbox, link).
Subject: Re: Bug#743141: system hangs when building gcc on kfreebsd-amd64
Date: Thu, 01 May 2014 01:08:52 +0100
Everything can be deleted from
~/gcc-4.7-4.7.3/src/gcc/testsuite/gcc.c-torture/compile/ except for
pr44686.c, which necessitates building tls_runtime, which triggers the
bug immediately.
The tls_runtime exe survives a reboot if building on ufs mounted with -o
sync. But executing it manually (in the same cwd, with the same
LD_LIBRARY_PATH) doesn't trigger a crash, it just exits with status 0.
Running the testsuite under ktrace doesn't help because it does not seem
to flush any data to ktrace.out soon enough to survive the crash.
Running the testsuite under gdb it still crashes the whole system, and I
don't know how/where I could set a breakpoint in order to get more details.
Regards,
--
Steven Chamberlain
steven@pyro.eu.org
Information forwarded
to debian-bugs-dist@lists.debian.org, GNU/kFreeBSD Maintainers <debian-bsd@lists.debian.org>: Bug#743141; Package kfreebsd-9.
(Sat, 10 May 2014 14:51:07 GMT) (full text, mbox, link).
Acknowledgement sent
to Christoph Egger <christoph@debian.org>:
Extra info received and forwarded to list. Copy sent to GNU/kFreeBSD Maintainers <debian-bsd@lists.debian.org>.
(Sat, 10 May 2014 14:51:08 GMT) (full text, mbox, link).
Hi!
Steven Chamberlain <steven@pyro.eu.org> writes:
> Everything can be deleted from
> ~/gcc-4.7-4.7.3/src/gcc/testsuite/gcc.c-torture/compile/ except for
> pr44686.c, which necessitates building tls_runtime, which triggers the
> bug immediately.
>
> The tls_runtime exe survives a reboot if building on ufs mounted with -o
> sync. But executing it manually (in the same cwd, with the same
> LD_LIBRARY_PATH) doesn't trigger a crash, it just exits with status 0.
>
> Running the testsuite under ktrace doesn't help because it does not seem
> to flush any data to ktrace.out soon enough to survive the crash.
>
> Running the testsuite under gdb it still crashes the whole system, and I
> don't know how/where I could set a breakpoint in order to get more details.
Did you manage to figure something additional out? Do we know if this
also kills plain freebsd?
Thanks
Christoph
--
9FED 5C6C E206 B70A 5857 70CA 9655 22B9 D49A E731
Debian Developer | Lisp Hacker | CaCert Assurer
Information forwarded
to debian-bugs-dist@lists.debian.org, GNU/kFreeBSD Maintainers <debian-bsd@lists.debian.org>: Bug#743141; Package kfreebsd-9.
(Sun, 11 May 2014 16:39:08 GMT) (full text, mbox, link).
Acknowledgement sent
to Steven Chamberlain <steven@pyro.eu.org>:
Extra info received and forwarded to list. Copy sent to GNU/kFreeBSD Maintainers <debian-bsd@lists.debian.org>.
(Sun, 11 May 2014 16:39:08 GMT) (full text, mbox, link).
Hi,
On 10/05/14 15:52, Christoph Egger wrote:
> Steven Chamberlain <steven@pyro.eu.org> writes:
>> Everything can be deleted from
>> ~/gcc-4.7-4.7.3/src/gcc/testsuite/gcc.c-torture/compile/ except for
>> pr44686.c, which necessitates building tls_runtime, which triggers the
>> bug immediately.
> Did you manage to figure something additional out? Do we know if this
> also kills plain freebsd?
I'm afraid I'm stuck with this.
I narrowed it down to that particular test (or prerequisites of the
test) when executed by the GCC testsuite. A large chroot is needed,
having all the GCC build-deps; once GCC is finished building it's
fairly easy to trigger this on demand.
Even if it might trigger the bug on plain FreeBSD, I think all this is
too unwieldy to provide as a testcase yet (0.5GiB for the GNU/kFreeBSD
chroot and 4GiB for the GCC build tree).
I'm still not sure exactly how the test is invoked by the testsuite's
expect scripts; which executable, with what arguments, libraries, and
how everything was compiled.
My next idea is to try wheezy on kfreebsd-9 (perhaps built with
LOCK_DEBUG etc.?), because Ivo showed a screenshot with some kernel
output and a hang. Whereas kfreebsd-10 seems to reboot instantly and
not output anything; even the older builds for experimental which had
some extra debug options enabled.
Regards,
--
Steven Chamberlain
steven@pyro.eu.org
Information forwarded
to debian-bugs-dist@lists.debian.org, GNU/kFreeBSD Maintainers <debian-bsd@lists.debian.org>: Bug#743141; Package kfreebsd-9.
(Sun, 11 May 2014 17:21:04 GMT) (full text, mbox, link).
Acknowledgement sent
to Petr Salinger <Petr.Salinger@seznam.cz>:
Extra info received and forwarded to list. Copy sent to GNU/kFreeBSD Maintainers <debian-bsd@lists.debian.org>.
(Sun, 11 May 2014 17:21:04 GMT) (full text, mbox, link).
To: Steven Chamberlain <steven@pyro.eu.org>, 743141@bugs.debian.org
Cc: Christoph Egger <christoph@debian.org>
Subject: Re: Bug#743141: system hangs when building gcc on kfreebsd-amd64
Date: Sun, 11 May 2014 19:18:56 +0200 (CEST)
Hi.
>> Did you manage to figure something additional out? Do we know if this
>> also kills plain freebsd?
>
> I'm afraid I'm stuck with this.
>
> I narrowed it down to that particular test (or prerequisites of the
> test) when executed by the GCC testsuite. A large chroot is needed,
> having all the GCC build-deps; once GCC is finished building it's
> fairly easy to trigger this on demand.
>
> Even if it might trigger the bug on plain FreeBSD, I think all this is
> too unwieldy to provide as a testcase yet (0.5GiB for the GNU/kFreeBSD
> chroot and 4GiB for the GCC build tree).
I tried under plain FreeBSD kernel (via kfreebsd-downloader-10)
and it does not reboot for me.
It reboots for me under current sid 10 kernel and current experimental 11
kernel.
Petr
Information forwarded
to debian-bugs-dist@lists.debian.org, GNU/kFreeBSD Maintainers <debian-bsd@lists.debian.org>: Bug#743141; Package kfreebsd-9.
(Sun, 11 May 2014 23:30:09 GMT) (full text, mbox, link).
Acknowledgement sent
to Steven Chamberlain <steven@pyro.eu.org>:
Extra info received and forwarded to list. Copy sent to GNU/kFreeBSD Maintainers <debian-bsd@lists.debian.org>.
(Sun, 11 May 2014 23:30:09 GMT) (full text, mbox, link).
To: Petr Salinger <Petr.Salinger@seznam.cz>, 743141@bugs.debian.org
Cc: Christoph Egger <christoph@debian.org>
Subject: Re: Bug#743141: system hangs when building gcc on kfreebsd-amd64
Date: Mon, 12 May 2014 00:27:10 +0100
On 11/05/14 18:18, Petr Salinger wrote:
> I tried under plain FreeBSD kernel (via kfreebsd-downloader-10)
> and it does not reboot for me.
It still crashes reproducibly for me, on upstream's build of:
GNU/kFreeBSD debian 10.0-RELEASE FreeBSD 10.0-RELEASE #0 r260789: Thu
Jan 16 22:34:59 UTC 2014
root@snap.freebsd.org:/usr/obj/usr/src/sys/GENERIC x86_64
I deleted all tests except for
src/gcc/testsuite/gcc.c-torture/compile/pr44686.c
and then:
> steven@debian:~/gcc-4.7-4.7.3/build/gcc$ make check RUNTESTFLAGS='--target_board "unix/-m32" -v -v'
...
> Testing gcc.c-torture/execute/builtins/20010124-1.c, -O0
> doing compile
> Invoking the compiler as /home/steven/gcc-4.7-4.7.3/build/gcc/xgcc -B/home/steven/gcc-4.7-4.7.3/build/gcc/ /home/steven/gcc-4.7-4.7.3/src/gcc/testsuite/gcc.c-torture/execute/builtins/20010124-1.c /home/steven/gcc-4.7-4.7.3/src/gcc/testsuite/gcc.c-torture/execute/builtins/20010124-1-lib.c /home/steven/gcc-4.7-4.7.3/src/gcc/testsuite/gcc.c-torture/execute/builtins/lib/main.c -w -O0 -lm -m32 -o /home/steven/gcc-4.7-4.7.3/build/gcc/testsuite/gcc/20010124-1.x0
> Setting timeout to 300
> Executing on host: /home/steven/gcc-4.7-4.7.3/build/gcc/xgcc -B/home/steven/gcc-4.7-4.7.3/build/gcc/ /home/steven/gcc-4.7-4.7.3/src/gcc/testsuite/gcc.c-torture/execute/builtins/20010124-1.c /home/steven/gcc-4.7-4.7.3/src/gcc/testsuite/gcc.c-torture/execute/builtins/20010124-1-lib.c /home/steven/gcc-4.7-4.7.3/src/gcc/testsuite/gcc.c-torture/execute/builtins/lib/main.c -w -O0 -lm -m32 -o /home/steven/gcc-4.7-4.7.3/build/gcc/testsuite/gcc/20010124-1.x0 (timeout = 300)
> pid is 2166 -2166
> waitres is 2166 exp8 0 0
> output is status 0
> Checking pattern "sparc-*-sunos*" with x86_64-pc-kfreebsd-gnu
> Checking pattern "alpha*-*-*" with x86_64-pc-kfreebsd-gnu
> Checking pattern "hppa*-*-hpux*" with x86_64-pc-kfreebsd-gnu
> Checking pattern "sparc-*-sunos*" with x86_64-pc-kfreebsd-gnu
> Checking pattern "alpha*-*-*" with x86_64-pc-kfreebsd-gnu
> Checking pattern "hppa*-*-hpux*" with x86_64-pc-kfreebsd-gnu
> Checking "spu-*-*" against "x86_64-pc-kfreebsd-gnu"
> Checking x86_64-pc-kfreebsd-gnu against x86_64-pc-kfreebsd-gnu
> loading to unix/-m32
> Setting LD_LIBRARY_PATH to :/home/steven/gcc-4.7-4.7.3/build/gcc:/home/steven/gcc-4.7-4.7.3/build/gcc/32::/home/steven/gcc-4.7-4.7.3/build/gcc:/home/steven/gcc-4.7-4.7.3/build/gcc/32
> spawning command /home/steven/gcc-4.7-4.7.3/build/gcc/testsuite/gcc/20010124-1.x0
<reboots immediately>
This even works inside of a BSD jail.
After a reboot, 20010124-1.x0 was missing. Even though the filesystem
is UFS, mounted with the sync option.
I can compile it by hand, then run it with the LD_LIBRARY_PATH as above,
but it doesn't trigger the crash that way.
I'd really like to go to upstream with this, but the testcase is
currently "a full Debian GNU/kFreeBSD sid chroot, a fully built GCC-4.7
source tree and all build dependencies" and I think we need to find
something smaller.
Regards,
--
Steven Chamberlain
steven@pyro.eu.org
Information forwarded
to debian-bugs-dist@lists.debian.org, GNU/kFreeBSD Maintainers <debian-bsd@lists.debian.org>: Bug#743141; Package kfreebsd-9.
(Mon, 12 May 2014 00:09:04 GMT) (full text, mbox, link).
Acknowledgement sent
to Steven Chamberlain <steven@pyro.eu.org>, 743141@bugs.debian.org:
Extra info received and forwarded to list. Copy sent to GNU/kFreeBSD Maintainers <debian-bsd@lists.debian.org>.
(Mon, 12 May 2014 00:09:04 GMT) (full text, mbox, link).
To: Petr Salinger <Petr.Salinger@seznam.cz>, 743141@bugs.debian.org
Cc: Christoph Egger <christoph@debian.org>
Subject: Re: Bug#743141: system hangs when building gcc on kfreebsd-amd64
Date: Mon, 12 May 2014 01:06:47 +0100
Finally getting somewhere!
testcase.c can be just:
> int main() { }
Use the system GCC, don't need to compile it from source.
> $ gcc -o testcase testcase.c
This is on kfreebsd-amd64, the executable is 64-bit.
Use expect to spawn it like this (same way that dejagnu does in the GCC
testsuite) :
> $ expect
> expect1.1> open "| ./testcase |& cat" "r"
> file6
This is okay.
Now need gcc-multilib installed for this:
> $ gcc -m32 -o testcase testcase.c
> $ expect
> expect1.1> open "| ./testcase |& cat" "r"
The 32-bit executable, spawned this way by expect, triggers an immediate
reboot.
Regards,
--
Steven Chamberlain
steven@pyro.eu.org
Information forwarded
to debian-bugs-dist@lists.debian.org, GNU/kFreeBSD Maintainers <debian-bsd@lists.debian.org>: Bug#743141; Package kfreebsd-9.
(Mon, 12 May 2014 00:57:07 GMT) (full text, mbox, link).
Acknowledgement sent
to Steven Chamberlain <steven@pyro.eu.org>, 743141@bugs.debian.org:
Extra info received and forwarded to list. Copy sent to GNU/kFreeBSD Maintainers <debian-bsd@lists.debian.org>.
(Mon, 12 May 2014 00:57:07 GMT) (full text, mbox, link).
Attached is a ktrace of expect spawning a 64-bit executable.
I can't do this for the 32-bit testcase because the system crashes
before any ktrace output is written to disk. Ivo's screenshot
implicated a 'spin lock held too long'.
I tried replacing the testcase with something that does "sleep(10)"
instead. The crash still happens immediately (upon program start, not
program exit).
In Ivo's screenshot, "sched lock 4" implicates the ULE scheduler,
something to do with thread queues. kern/kern_mutex.c suggests that
kfreebsd 9.2 compiled with WITNESS might give a little more info.
Regards,
--
Steven Chamberlain
steven@pyro.eu.org
Information forwarded
to debian-bugs-dist@lists.debian.org, GNU/kFreeBSD Maintainers <debian-bsd@lists.debian.org>: Bug#743141; Package kfreebsd-9.
(Mon, 12 May 2014 02:12:04 GMT) (full text, mbox, link).
Acknowledgement sent
to Steven Chamberlain <steven@pyro.eu.org>, 743141@bugs.debian.org:
Extra info received and forwarded to list. Copy sent to GNU/kFreeBSD Maintainers <debian-bsd@lists.debian.org>.
(Mon, 12 May 2014 02:12:04 GMT) (full text, mbox, link).
To: Petr Salinger <Petr.Salinger@seznam.cz>, 743141@bugs.debian.org
Subject: Re: Bug#743141: system hangs when building gcc on kfreebsd-amd64
Date: Mon, 12 May 2014 03:09:08 +0100
An even simpler way, don't need the expect program. Create a thread in
a 64-bit process, and try to execve() a 32-bit executable:
> #include <unistd.h>
> #include <pthread.h>
>
> void *thread_main() {
> char *cmdline[] = { "./testcase32", NULL };
> execve(cmdline[0], cmdline, NULL);
> }
>
> int main() {
> pthread_t thread;
> pthread_create(&thread, NULL, thread_main, NULL);
> pthread_join(thread, NULL);
> return 0;
> }
Regards,
--
Steven Chamberlain
steven@pyro.eu.org
Severity set to 'grave' from 'serious'
Request was from Steven Chamberlain <steven@pyro.eu.org>
to control@bugs.debian.org.
(Mon, 12 May 2014 13:42:04 GMT) (full text, mbox, link).
Added tag(s) upstream and security.
Request was from Steven Chamberlain <steven@pyro.eu.org>
to control@bugs.debian.org.
(Mon, 12 May 2014 13:42:05 GMT) (full text, mbox, link).
Set Bug forwarded-to-address to 'secteam@FreeBSD.org'.
Request was from Steven Chamberlain <steven@pyro.eu.org>
to control@bugs.debian.org.
(Mon, 12 May 2014 13:42:06 GMT) (full text, mbox, link).
Information forwarded
to debian-bugs-dist@lists.debian.org, GNU/kFreeBSD Maintainers <debian-bsd@lists.debian.org>: Bug#743141; Package kfreebsd-9.
(Tue, 20 May 2014 11:09:05 GMT) (full text, mbox, link).
Acknowledgement sent
to Steven Chamberlain <steven@pyro.eu.org>:
Extra info received and forwarded to list. Copy sent to GNU/kFreeBSD Maintainers <debian-bsd@lists.debian.org>.
(Tue, 20 May 2014 11:09:05 GMT) (full text, mbox, link).
Added tag(s) fixed-upstream.
Request was from Steven Chamberlain <steven@pyro.eu.org>
to control@bugs.debian.org.
(Tue, 20 May 2014 11:09:11 GMT) (full text, mbox, link).
Bug 743141 cloned as bugs 748743, 748744, 748745
Request was from Steven Chamberlain <steven@pyro.eu.org>
to control@bugs.debian.org.
(Tue, 20 May 2014 11:45:04 GMT) (full text, mbox, link).
Changed Bug title to 'kfreebsd-9: triple fault on execve from 64-bit thread to 32-bit process' from 'system hangs when building gcc on kfreebsd-amd64'
Request was from Steven Chamberlain <steven@pyro.eu.org>
to control@bugs.debian.org.
(Tue, 20 May 2014 11:45:05 GMT) (full text, mbox, link).
Marked as found in versions kfreebsd-9/9.0-10+deb70.6.
Request was from Steven Chamberlain <steven@pyro.eu.org>
to control@bugs.debian.org.
(Tue, 20 May 2014 11:45:06 GMT) (full text, mbox, link).
Removed tag(s) help.
Request was from Steven Chamberlain <steven@pyro.eu.org>
to control@bugs.debian.org.
(Tue, 20 May 2014 11:57:08 GMT) (full text, mbox, link).
Information forwarded
to debian-bugs-dist@lists.debian.org, GNU/kFreeBSD Maintainers <debian-bsd@lists.debian.org>: Bug#743141; Package kfreebsd-9.
(Thu, 22 May 2014 13:51:08 GMT) (full text, mbox, link).
Acknowledgement sent
to Robert Millan <rmh@debian.org>:
Extra info received and forwarded to list. Copy sent to GNU/kFreeBSD Maintainers <debian-bsd@lists.debian.org>.
(Thu, 22 May 2014 13:51:08 GMT) (full text, mbox, link).
Subject: Re: Processed: Re: Bug#743141: system hangs when building gcc on
kfreebsd-amd64
Date: Thu, 22 May 2014 15:49:48 +0200
Control: tag 743141 wontfix
On 20/05/14 13:45, Debian Bug Tracking System wrote:
>> retitle 743141 kfreebsd-9: triple fault on execve from 64-bit thread to 32-bit process
Tentatively tagging 743141 wontfix as kfreebsd-9 is planned for removal.
But, AFAIK, it can't be removed yet as latest D-I release still uses it? (CCing -boot)
--
Robert Millan
Added tag(s) wontfix.
Request was from Robert Millan <rmh@debian.org>
to 743141-submit@bugs.debian.org.
(Thu, 22 May 2014 13:51:08 GMT) (full text, mbox, link).
Information forwarded
to debian-bugs-dist@lists.debian.org, GNU/kFreeBSD Maintainers <debian-bsd@lists.debian.org>: Bug#743141; Package kfreebsd-9.
(Thu, 22 May 2014 14:12:08 GMT) (full text, mbox, link).
Acknowledgement sent
to Cyril Brulebois <kibi@debian.org>:
Extra info received and forwarded to list. Copy sent to GNU/kFreeBSD Maintainers <debian-bsd@lists.debian.org>.
(Thu, 22 May 2014 14:12:08 GMT) (full text, mbox, link).
Robert Millan <rmh@debian.org> (2014-05-22):
> Control: tag 743141 wontfix
>
> On 20/05/14 13:45, Debian Bug Tracking System wrote:
> >>retitle 743141 kfreebsd-9: triple fault on execve from 64-bit thread to 32-bit process
>
> Tentatively tagging 743141 wontfix as kfreebsd-9 is planned for removal.
>
> But, AFAIK, it can't be removed yet as latest D-I release still uses it? (CCing -boot)
You can run dak rm to check. Maybe the source can stay for compliance
(Built-Using, d-i, etc.) and binaries go away anyway.
See http://wiki.debian.org/ftpmaster_Removals
Mraw,
KiBi.
Debbugs is free software and licensed under the terms of the GNU General
Public License version 2. The current version can be obtained
from https://bugs.debian.org/debbugs-source/.