Bug #15335: Ruby 2.6.0 is not properly fortified

• Status: Closed
• Assignee: ioquatix (Samuel Williams)
• Target version: -
• ruby -v: ruby 2.6.0dev (2018-11-22 trunk 65928) [x86_64-linux]
• Backport: 2.3: UNKNOWN, 2.4: UNKNOWN, 2.5: UNKNOWN

[vo.x (Vit Ondruch)]

For some while, we carry this test in Fedora package 1:

checksec -f libruby.so.%{ruby_version} | \	
  grep "Full RELRO.*Canary found.*NX enabled.*DSO.*No RPATH.*No RUNPATH.*Yes.*\d*.*\d*.*libruby.so.%{ruby_version}"

This should ensure, that the library is properly fortified 2. This test was passing with preview3, but it started to fail, testing with r65928:

$ checksec -f libruby.so.2.6.0
WARNING: 'openssl' not found! It's required for most checks.

WARNING: Not all necessary commands found. Some tests might not work!

RELRO           STACK CANARY      NX            PIE             RPATH      RUNPATH	FORTIFY	Fortified Fortifiable  FILE
Full RELRO      Canary found      NX disabled   DSO             No RPATH   No RUNPATH   Yes	16		42	libruby.so.2.6.0

The NX disabled is the difference. Looking at the log, it is definitely not about configuration options. So if I should point finger at something, it seems to me that this must be it:

... snip ...

assembling coroutine/amd64/Context.s
gcc -I. -I.ext/include/x86_64-linux -I./include -I. -I./enc/unicode/10.0.0 -o coroutine/amd64/Context.o -c coroutine/amd64/Context.s

... snip ...

gcc -shared -Wl,-z,relro -Wl,--as-needed  -Wl,-z,now -specs=/usr/lib/rpm/redhat/redhat-hardened-ld -Wl,-soname,libruby.so.2.6 -fstack-protector-strong  -m64 dln.o localeinit.o loadpath.o array.o ast.o bignum.o class.o compar.o compile.o complex.o cont.o debug.o debug_counter.o dir.o dln_find.o encoding.o enum.o enumerator.o error.o eval.o file.o gc.o hash.o inits.o io.o iseq.o load.o marshal.o math.o mjit.o mjit_compile.o node.o numeric.o object.o pack.o parse.o proc.o process.o random.o range.o rational.o re.o regcomp.o regenc.o regerror.o regexec.o regparse.o regsyntax.o ruby.o safe.o signal.o sprintf.o st.o strftime.o string.o struct.o symbol.o thread.o time.o transcode.o transient_heap.o util.o variable.o version.o vm.o vm_backtrace.o vm_dump.o vm_trace.o coroutine/amd64/Context.o probes.o enc/ascii.o enc/us_ascii.o enc/unicode.o enc/utf_8.o enc/trans/newline.o setproctitle.o strlcat.o strlcpy.o addr2line.o  prelude.o dmyext.o dmyenc.o -lpthread -lrt -lrt -lgmp -ldl -lcrypt -lm   -o libruby.so.2.6.0

... snip ...

I.e. the coroutines assembly. Not sure how to prove it nor fix it.

[mame (Yusuke Endoh)]

• Status changed from Open to Assigned
• Assignee set to ioquatix (Samuel Williams)

[ioquatix (Samuel Williams)]

How can we solve this problem?

[normalperson (Eric Wong)]

https://bugs.ruby-lang.org/issues/15335#change-75132

[e-mail] wrote:

How can we solve this problem?

Fedora can configure with --disable-fiber-coroutine option
as a stopgap...

I.e. the coroutines assembly. Not sure how to prove it nor fix it.

Maybe this can help?
https://wiki.gentoo.org/wiki/Hardened/GNU_stack_quickstart

[ioquatix (Samuel Williams)]

Thanks Eric. Those two suggestions are really helpful. I'll investigate it now.

[ioquatix (Samuel Williams)]

I'm attempting to fix this issue here: https://github.com/ruby/ruby/pull/2027

@vo.x (Vit Ondruch) do you mind trying to build that PR?

[ioquatix (Samuel Williams)]

I've merged the first set of changes into trunk, for amd64 - @vo.x (Vit Ondruch) if you confirm this has fixed the issue I will extend to other architectures.

[vo.x (Vit Ondruch)]

Thx. I hope I'll be able to check it on Monday.

[vo.x (Vit Ondruch)]

Testing with r65990 on x86_64 and the test passed. I guess I should not try to build on other arches yet, right?

[vo.x (Vit Ondruch)]

Just FTR trying to build r66092 on Fedora Rawhide 1, x86_64, ppc64le, aarch64, s390x, and armv7hl were properly fortified while only i686 failed to pass 2 the fortification test.

• Status changed from Assigned to Closed

Applied in changeset trunk|r66341.

---

Ensure x86 stack is fortified, fixed #15335.

[ioquatix (Samuel Williams)]

I've fixed x86 implementation too, now that it was confirmed the previous effort to fix x64 worked as expected. This issue should be completely resolved now (dc6908ab44c3a3fc78319422410b57d3b7fb6c0c / r66341).

[vo.x (Vit Ondruch)]

ioquatix (Samuel Williams) wrote:

I've fixed x86 implementation too.

Thx, I can build Ruby 2.6.0 on all platforms just fine.

[vo.x (Vit Ondruch)]

• Related to Bug #16762: Ruby is not properly fortified on armv7hl added