Apache / Passenger segfault at startup (Armel architecture) -- help?

[Chris P.]

Hi there. I'm running into an issue and would be grateful for
advice. On an Armel-based SheevaPlug, I've compiled Passenger with
success. (The only change I made was editing line 58 of
sp_counted_base.hpp, as suggested elsewhere online.)

However, Apache refuses to start when Passenger-enabled. The error is:

[ pid=9079 file=ext/common/ApplicationPoolServerExecutable.cpp:732
time=2009-08-28 22:32:37.713 ]:
*** Fatal error: Cannot send file descriptor with sendmsg(): Invalid
argument (22)
in 'int Server::start()' (ApplicationPoolServerExecutable.cpp:
703)

I can't seem to find other people in my boat, which is worrisome.
Does anyone have any ideas or suggestions?

Linux 2.6.30.5, Ubuntu Jaunty.

Regards,
Chris

[Hongli Lai]

sendmsg() probably needs to be called differently on ARM. Could you
run 'cpp -dM /dev/null' and paste the output here?

--
Phusion | The Computer Science Company

Web: http://www.phusion.nl/
E-mail: in...@phusion.nl
Chamber of commerce no: 08173483 (The Netherlands)

[Chris P.]

On Aug 29, 1:08 am, Hongli Lai <hon...@phusion.nl> wrote:
> sendmsg() probably needs to be called differently on ARM. Could you
> run 'cpp -dM /dev/null' and paste the output here?

Certainly (and thanks). Output follows.

Regards,
Chris

$ cpp -dM /dev/null
#define __DBL_MIN_EXP__ (-1021)
#define __FLT_MIN__ 1.17549435e-38F
#define __DEC64_DEN__ 0.000000000000001E-383DD
#define __CHAR_BIT__ 8
#define __WCHAR_MAX__ 4294967295U
#define __DBL_DENORM_MIN__ 4.9406564584124654e-324
#define __FLT_EVAL_METHOD__ 0
#define __unix__ 1
#define __DBL_MIN_10_EXP__ (-307)
#define __FINITE_MATH_ONLY__ 0
#define __ARMEL__ 1
#define __GNUC_PATCHLEVEL__ 3
#define __DEC64_MAX_EXP__ 384
#define __SHRT_MAX__ 32767
#define __LDBL_MAX__ 1.7976931348623157e+308L
#define __GXX_TYPEINFO_EQUALITY_INLINE 0
#define __UINTMAX_TYPE__ long long unsigned int
#define __linux 1
#define __DEC32_EPSILON__ 1E-6DF
#define __CHAR_UNSIGNED__ 1
#define __LDBL_MAX_EXP__ 1024
#define __linux__ 1
#define __SCHAR_MAX__ 127
#define __DBL_DIG__ 15
#define _FORTIFY_SOURCE 2
#define __SIZEOF_INT__ 4
#define __SIZEOF_POINTER__ 4
#define __USER_LABEL_PREFIX__
#define __STDC_HOSTED__ 1
#define __LDBL_HAS_INFINITY__ 1
#define __FLT_EPSILON__ 1.19209290e-7F
#define __APCS_32__ 1
#define __LDBL_MIN__ 2.2250738585072014e-308L
#define __DEC32_MAX__ 9.999999E96DF
#define __SIZEOF_LONG__ 4
#define __DECIMAL_DIG__ 17
#define __gnu_linux__ 1
#define __LDBL_HAS_QUIET_NAN__ 1
#define __GNUC__ 4
#define __FLT_HAS_DENORM__ 1
#define __SIZEOF_LONG_DOUBLE__ 8
#define __ARM_ARCH_5T__ 1
#define __DBL_MAX__ 1.7976931348623157e+308
#define __DBL_HAS_INFINITY__ 1
#define __DEC32_MIN_EXP__ (-95)
#define __THUMB_INTERWORK__ 1
#define __LDBL_HAS_DENORM__ 1
#define __DEC128_MAX__ 9.999999999999999999999999999999999E6144DL
#define __DEC32_MIN__ 1E-95DF
#define __DBL_MAX_EXP__ 1024
#define __DEC128_EPSILON__ 1E-33DL
#define __LONG_LONG_MAX__ 9223372036854775807LL
#define __SIZEOF_SIZE_T__ 4
#define __SIZEOF_WINT_T__ 4
#define __GXX_ABI_VERSION 1002
#define __SOFTFP__ 1
#define __FLT_MIN_EXP__ (-125)
#define __DBL_MIN__ 2.2250738585072014e-308
#define __DEC128_MIN__ 1E-6143DL
#define __REGISTER_PREFIX__
#define __DBL_HAS_DENORM__ 1
#define __NO_INLINE__ 1
#define __FLT_MANT_DIG__ 24
#define __VERSION__ "4.3.3"
#define __DEC64_EPSILON__ 1E-15DD
#define __DEC128_MIN_EXP__ (-6143)
#define unix 1
#define __SIZE_TYPE__ unsigned int
#define __DEC32_DEN__ 0.000001E-95DF
#define __ELF__ 1
#define __FLT_RADIX__ 2
#define __LDBL_EPSILON__ 2.2204460492503131e-16L
#define __VFP_FP__ 1
#define __SIZEOF_PTRDIFF_T__ 4
#define __FLT_HAS_QUIET_NAN__ 1
#define __FLT_MAX_10_EXP__ 38
#define __LONG_MAX__ 2147483647L
#define __FLT_HAS_INFINITY__ 1
#define __unix 1
#define __DEC64_MAX__ 9.999999999999999E384DD
#define __DEC64_MANT_DIG__ 16
#define __DEC32_MAX_EXP__ 96
#define linux 1
#define __DEC128_DEN__ 0.000000000000000000000000000000001E-6143DL
#define __LDBL_MANT_DIG__ 53
#define __DBL_HAS_QUIET_NAN__ 1
#define __WCHAR_TYPE__ unsigned int
#define __SIZEOF_FLOAT__ 4
#define __DEC64_MIN_EXP__ (-383)
#define __FLT_DIG__ 6
#define __INT_MAX__ 2147483647
#define __FLT_MAX_EXP__ 128
#define __DBL_MANT_DIG__ 53
#define __DEC64_MIN__ 1E-383DD
#define __WINT_TYPE__ unsigned int
#define __SIZEOF_SHORT__ 2
#define __LDBL_MIN_EXP__ (-1021)
#define __arm__ 1
#define __LDBL_MAX_10_EXP__ 308
#define __DBL_EPSILON__ 2.2204460492503131e-16
#define __SIZEOF_WCHAR_T__ 4
#define __DEC_EVAL_METHOD__ 2
#define __INTMAX_MAX__ 9223372036854775807LL
#define __FLT_DENORM_MIN__ 1.40129846e-45F
#define __FLT_MAX__ 3.40282347e+38F
#define __SIZEOF_DOUBLE__ 8
#define __FLT_MIN_10_EXP__ (-37)
#define __INTMAX_TYPE__ long long int
#define __DEC128_MAX_EXP__ 6144
#define __GNUC_MINOR__ 3
#define __DEC32_MANT_DIG__ 7
#define __DBL_MAX_10_EXP__ 308
#define __LDBL_DENORM_MIN__ 4.9406564584124654e-324L
#define __STDC__ 1
#define __PTRDIFF_TYPE__ int
#define __ARM_EABI__ 1
#define __DEC128_MANT_DIG__ 34
#define __LDBL_MIN_10_EXP__ (-307)
#define __SIZEOF_LONG_LONG__ 8
#define __LDBL_DIG__ 15
#define __GNUC_GNU_INLINE__ 1

[Hongli Lai]

On Sat, Aug 29, 2009 at 9:25 PM, Chris P.<christophe...@gmail.com> wrote:
> Certainly (and thanks).  Output follows.

Try this:

Edit ext/common/MessageChannel.h.
Find:
#if defined(__APPLE__) || defined(__SOLARIS__)
There are two such occurrences. Replace both with:
#if defined(__APPLE__) || defined(__SOLARIS__) || defined(__arm__)

Next, edit ext/phusion_passenger/native_support.c.
Find:
#if defined(__APPLE__) || defined(__SOLARIS__)
There are two such occurrences. Replace both with:
#if defined(__APPLE__) || defined(__SOLARIS__) || defined(__arm__)

Then recompile Phusion Passenger with 'rake apache2' and restart
Apache. Does this work?

[Chris P.]

On Aug 29, 4:36 pm, Hongli Lai <hon...@phusion.nl> wrote:
> Try this:
<snip>

OK, Passenger compiles and loads successfully with your proposed
alterations, and initial testing indicates that it's working. This is
fantastic -- thanks very much.

One note: whereas you say "two such occurrences", there are actually
*four* occurrences in each file. Lacking any info to the contrary I
simply shrugged and changed all eight.

So in summary I am now running with your suggested edits above, plus
the "line 58 of sp_counted_base.hpp" change. I'll bang on it more,
but I'm feeling very optimistic. I'm grateful for your assistance.

Regards,
Chris

[Hongli Lai]

On Sun, Aug 30, 2009 at 2:23 AM, Chris P.<christophe...@gmail.com> wrote:
> OK, Passenger compiles and loads successfully with your proposed
> alterations, and initial testing indicates that it's working.  This is
> fantastic -- thanks very much.
>
> One note: whereas you say "two such occurrences", there are actually
> *four* occurrences in each file.  Lacking any info to the contrary I
> simply shrugged and changed all eight.

Yes you're right, there are four.

> So in summary I am now running with your suggested edits above, plus
> the "line 58 of sp_counted_base.hpp" change.  I'll bang on it more,
> but I'm feeling very optimistic.  I'm grateful for your assistance.

Good to hear.

I'd like to make a few more changes so that this all works out of the
box in Armel. Could you run 'ruby -e "puts RUBY_PLATFORM"' and tell me
what you see?

[Chris P.]

On Aug 30, 3:10 am, Hongli Lai <hon...@phusion.nl> wrote:
> I'd like to make a few more changes so that this all works out of the
> box in Armel. Could you run 'ruby -e "puts RUBY_PLATFORM"' and tell me
> what you see?

Sure, here you go:

$ ruby -e "puts RUBY_PLATFORM"
arm-linux-eabi

Please let me know how I can provide additional help; I'm happy to.

Regards,
Chris

[Hongli Lai]

Could you try the latest code in the Phusion Passenger git repository
and check whether that works out of the box? Installation instructions
are in the README.

[Chris P.]

On Aug 30, 11:58 am, Hongli Lai <hon...@phusion.nl> wrote:
> Could you try the latest code in the Phusion Passenger git repository
> and check whether that works out of the box? Installation instructions
> are in the README.

I just checked it out with Git, compiled and installed it, and loaded
a Rails app page without trouble. Looks fantastic to me.

Question: I see that the Boost workaround for "undefined symbol:
__sync_val_compare_and_swap_4" (mentioned here
http://groups.google.com/group/phusion-passenger/browse_thread/thread/72637ecd4f87596a
) is not present in the code I gitted. How did you end up fixing it?

Regards,
Chris

[Hongli Lai]

I did it by defining BOOST_SP_USE_PTHREADS (or something similar to
that) on ARM systems. This has the same effect as making that code
change.

Regards,
Hongli Lai

[Chris P.]

On Aug 31, 11:17 am, Hongli Lai <hon...@phusion.nl> wrote:

> On Mon, Aug 31, 2009 at 6:28 PM, Chris P.<christopher.b.pow...@gmail.com> wrote:
> > Question: I see that the Boost workaround for "undefined symbol:
> > __sync_val_compare_and_swap_4" (mentioned here

> >http://groups.google.com/group/phusion-passenger/browse_thread/thread...

> > ) is not present in the code I gitted.  How did you end up fixing it?
>
> I did it by defining BOOST_SP_USE_PTHREADS (or something similar to
> that) on ARM systems. This has the same effect as making that code
> change.

Fair enough!

Thanks again for your quick and courteous assistance with the ARM
platform.

Regards,
Chris