[ruby-core:101927] [Ruby master Bug#14480] miniruby crashing when compiled with -O2 or -O1 on aarch64

[mame@ruby-lang.org]

Issue #14480 has been updated by mame (Yusuke Endoh).


Looks like the status of `__builtin_setjmp` is very complicated:

[The documentation of GCC](https://gcc.gnu.org/onlinedocs/gcc/Nonlocal-Gotos.html) says:

> GCC provides the built-in functions __builtin_setjmp and __builtin_longjmp which are similar to, but not interchangeable with, the C library functions setjmp and longjmp. The built-in versions are used internally by GCC’s libraries to implement exception handling on some targets. You should use the standard C library functions declared in <setjmp.h> in user code instead of the builtins.

> An important caveat is that GCC arranges to save and restore only those registers known to the specific architecture variant being compiled for. This can make __builtin_setjmp and __builtin_longjmp more efficient than their library counterparts in some cases, but it can also cause incorrect and mysterious behavior when mixing with code that uses the full register set.

So, in general, it looks good to avoid using `__builtin_setjmp`/`__builtin_longjmp` in Ruby.

It was introduced for a performance reason on Mac OS X in 2008.

http://blade.nagaokaut.ac.jp/cgi-bin/vframe.rb/ruby/ruby-core/16023?15980-16545+split-mode-vertical

I'm unsure if this performance improvement on macOS is still significant. Anyone is interested in measuring the peformance?

Actually, using `__builtin_setjmp` has been very troublesome. It is explicitly disabled on some architectures/conditions:

* AIX: d612c44dad8a5d4c373b0c4bc08a99e3dc06498e
* ppc64* LInux: 67fcbf9328614379fee28bed83ae6749ce7a3dda
* when -fcf-protection is used: #15307
* GCC 4 or before: #10272

And looks like we need to disable it on Android too. #17511

But interestingly, it is explicitly enabled on mingw64 because it was apparently needed to fix an issue: #15348.

----------------------------------------
Bug #14480: miniruby crashing when compiled with -O2 or -O1 on aarch64
https://bugs.ruby-lang.org/issues/14480#change-89774

* Author: vo.x (Vit Ondruch)
* Status: Open
* Priority: Normal
* ruby -v: ruby 2.5.0p0 (2017-12-25 revision 61468) [aarch64-linux]
* Backport: 2.3: UNKNOWN, 2.4: UNKNOWN, 2.5: UNKNOWN
----------------------------------------
Recently, it is not possible to build Ruby 2.5.0 on aarch64 on Fedora Rawhide, because miniruby fails during build:

~~~
... snip ...

./miniruby -I./lib -I. -I.ext/common  -n \
-e 'BEGIN{version=ARGV.shift;mis=ARGV.dup}' \
-e 'END{abort "UNICODE version mismatch: #{mis}" unless mis.empty?}' \
-e '(mis.delete(ARGF.path); ARGF.close) if /ONIG_UNICODE_VERSION_STRING +"#{Regexp.quote(version)}"/o' \
10.0.0 ./enc/unicode/10.0.0/casefold.h ./enc/unicode/10.0.0/name2ctype.h 
generating encdb.h
./miniruby -I./lib -I. -I.ext/common  ./tool/generic_erb.rb -c -o encdb.h ./template/encdb.h.tmpl ./enc enc
generating prelude.c
./miniruby -I./lib -I. -I.ext/common  ./tool/generic_erb.rb -I. -c -o prelude.c \
	./template/prelude.c.tmpl ./prelude.rb ./gem_prelude.rb ./abrt_prelude.rb
*** stack smashing detected ***: <unknown> terminated
encdb.h updated

... snip ...
~~~

This might by Ruby or gcc issue. Not sure yet. However, there is already lengthy analysis available in Fedora's Bugzilla [1]. Would be anybody able to help to resolve this issue?


[1]: https://bugzilla.redhat.com/show_bug.cgi?id=1545239

---Files--------------------------------
Dockerfile (573 Bytes)


-- 
https://bugs.ruby-lang.org/