A successful native build using clang at last---just!

[Paul Mather]

After quite a bit of trying, I finally was able to complete a native build and upgrade using clang on FreeBSD/arm.  I used r255420 as the build platform and was able to upgrade to r255586 on my Raspberry Pi.

Here are some observations:

1) For some reason this same upgrade failed on my BeagleBone Black (also running r255420).  It fails early on with this error and doesn't get past it:

[[...]]
c++  -O -pipe -I/usr/src/lib/clang/libllvmtablegen/../../../contrib/llvm/include -I/usr/src/lib/clang/libllvmtablegen/../../../contrib/llvm/tools/clang/include -I/usr/src/lib/clang/libllvmtablegen/../../../contrib/llvm/lib/TableGen -I. -I/usr/src/lib/clang/libllvmtablegen/../../../contrib/llvm/../../lib/clang/include -DLLVM_ON_UNIX -DLLVM_ON_FREEBSD -D__STDC_LIMIT_MACROS -D__STDC_CONSTANT_MACROS -fno-strict-aliasing -DLLVM_DEFAULT_TARGET_TRIPLE=\"armv6-gnueabi-freebsd10.0\" -DLLVM_HOST_TRIPLE=\"armv6-unknown-freebsd10.0\" -DDEFAULT_SYSROOT=\"\" -I/usr/obj/usr/src/tmp/legacy/usr/include -Wno-c++11-extensions -fno-exceptions -fno-rtti -c /usr/src/lib/clang/libllvmtablegen/../../../contrib/llvm/lib/TableGen/TGParser.cpp -o TGParser.o
Assertion failed: (!FoundVal && "Key already in new map?"), function moveFromOldBuckets, file /usr/src/lib/clang/libclangast/../../../contrib/llvm/include/llvm/ADT/DenseMap.h, line 289.
Stack dump:
0.	Program arguments: /usr/bin/c++ -cc1 -triple armv6--freebsd10.0-gnueabi -S -disable-free -main-file-name TGParser.cpp -mrelocation-model static -mdisable-fp-elim -relaxed-aliasing -mconstructor-aliases -target-abi aapcs-linux -target-cpu arm1136jf-s -msoft-float -mfloat-abi soft -target-feature +soft-float -target-feature +soft-float-abi -target-feature -neon -coverage-file /tmp/TGParser-KtbaOq.s -resource-dir /usr/bin/../lib/clang/3.3 -D LLVM_ON_UNIX -D LLVM_ON_FREEBSD -D __STDC_LIMIT_MACROS -D __STDC_CONSTANT_MACROS -D LLVM_DEFAULT_TARGET_TRIPLE="armv6-gnueabi-freebsd10.0" -D LLVM_HOST_TRIPLE="armv6-unknown-freebsd10.0" -D DEFAULT_SYSROOT="" -I /usr/src/lib/clang/libllvmtablegen/../../../contrib/llvm/include -I /usr/src/lib/clang/libllvmtablegen/../../../contrib/llvm/tools/clang/include -I /usr/src/lib/clang/libllvmtablegen/../../../contrib/llvm/lib/TableGen -I . -I /usr/src/lib/clang/libllvmtablegen/../../../contrib/llvm/../../lib/clang/include -I /usr/obj/usr/src/tmp/legacy/usr/include -internal-isystem /usr/include/c++/v1 -O2 -Wno-c++11-extensions -fdeprecated-macro -fno-dwarf-directory-asm -fdebug-compilation-dir /usr/obj/usr/src/tmp/usr/src/lib/clang/libllvmtablegen -ferror-limit 19 -fmessage-length 0 -mstackrealign -fno-rtti -fno-signed-char -fobjc-runtime=gnustep -fobjc-default-synthesize-properties -fdiagnostics-show-option -backend-option -vectorize-loops -o /tmp/TGParser-KtbaOq.s -x c++ /usr/src/lib/clang/libllvmtablegen/../../../contrib/llvm/lib/TableGen/TGParser.cpp 
1.	/usr/include/c++/v1/locale:3072:48: current parser token 'init'
2.	/usr/include/c++/v1/locale:205:1 <Spelling=/usr/include/c++/v1/__config:321:37>: parsing namespace 'std'
3.	/usr/include/c++/v1/locale:205:1 <Spelling=/usr/include/c++/v1/__config:321:59>: parsing namespace '__1'
c++: error: unable to execute command: Abort trap (core dumped)
c++: error: clang frontend command failed due to signal (use -v to see invocation)
FreeBSD clang version 3.3 (tags/RELEASE_33/final 183502) 20130610
Target: armv6--freebsd10.0-gnueabi
Thread model: posix
c++: note: diagnostic msg: PLEASE submit a bug report to http://llvm.org/bugs/ and include the crash backtrace, preprocessed source, and associated run script.
c++: note: diagnostic msg: Error generating preprocessed source(s).
*** Error code 254

Stop.
make[3]: stopped in /usr/src/lib/clang/libllvmtablegen
*** Error code 1

Stop.
make[2]: stopped in /usr/src
*** Error code 1

Stop.
make[1]: stopped in /usr/src
*** Error code 1

Stop.
make: stopped in /usr/src

2) The successful upgrade on Raspberry Pi had to be restarted a couple of times (using -DNO_CLEAN) due to hangs.  I have noticed problems with the system when swap is used.  Paging out to swap seems okay, but processes appear to die when they are paged back in from swap.  Sometimes, this hangs the system.

3) FWIW, I am running with super pages enabled: vm.pmap.sp_enabled=1.

Cheers,

Paul.