Bug 863047 – sci-mathematics/glpk-5.0-r1 fails to compile (lto): minisat.c:139:8: error: dereferencing type-punned pointer will break strict-aliasing rules [-Werror=strict-aliasing]

• Product: Gentoo Linux
• Status: RESOLVED FIXED

[Description Agostino Sarubbo 2022-08-02 14:08:08 UTC]

https://blogs.gentoo.org/ago/2020/07/04/gentoo-tinderbox/

Issue: sci-mathematics/glpk-5.0-r1 fails to compile (lto).
Discovered on: amd64 (internal ref: lto_tinderbox)

NOTE:
This machine uses lto with CFLAGS=-flto -Werror=odr -Werror=lto-type-mismatch -Werror=strict-aliasing

Here is a bit of explanation:

-Werror=lto-type-mismatch:
User to find possible runtime issues in packages. It likely means the package is unsafe to build & use with LTO.
For projects using the same identifier but with different types across different files, they must be fixed to be consistent across the codebase.

-Werror=odr:
Used to find possible runtime issues in packages. These bugs are a problem anyway but may be even worse when combined with LTO. C++ code must comply with the One Definition Rule (ODR) - see https://en.cppreference.com/w/cpp/language/definition#One_Definition_Rule.

-Werror=strict-aliasing:
Used to find possible runtime issues in packages. These bugs are a problem anyway but may be even worse when combined with LTO.

Workarounds:
- If upstream is friendly and still active, file a bug upstream. For emulators, codecs, games, or multimedia packages, it may be worth just applying a workaround instead, as upstreams sometimes aren't receptive to these bugs (VALID FOR ALL).
- Use the new 'filter-lto' from flag-o-matic.eclass as it's likely to be unsafe with LTO (VALID FOR lto-type-mismatch - odr).
- Fix it yourself if interested, of course (VALID FOR ALL).
- Append-flags -fno-strict-aliasing (VALID FOR strict-aliasing).
- Use memcpy() but a union is sometimes suitable too (VALID FOR strict-aliasing).
- -fstrict-aliasing is implied by -O2, so this must be addressed in some form (VALID FOR strict-aliasing).

See also: https://marc.info/?l=gentoo-dev&m=165639574126280&w=2

[Comment 1 Agostino Sarubbo 2022-08-02 14:08:11 UTC]

Created attachment 796894 [details]
build.log

build log and emerge --info

[Comment 2 François Bissey 2022-08-02 21:06:57 UTC]

I just inspected the code in question and I am not sure if it should count as a false positive or not. I am guessing it may break some compiler optimizations. All the cases reported in minisat.c are `#define` statements, so I would count it as casting rather than aliasing. But it could cause problem when inlining.

[Comment 3 Michael Orlitzky 2022-08-06 01:02:39 UTC]

Created attachment 798121 [details, diff]
0001-src-minisat-minisat.c-update-types-to-comply-with-al.patch

(In reply to François Bissey from comment #2)
> I just inspected the code in question and I am not sure if it should count
> as a false positive or not. I am guessing it may break some compiler
> optimizations. All the cases reported in minisat.c are `#define` statements,
> so I would count it as casting rather than aliasing. But it could cause
> problem when inlining.

I think the problem is that it's casting a pointer to one type (lit) to a pointer to another type (float), and then dereferencing. If so, that's undefined behavior in C even if it happens to work out correctly. I hacked the types to agree and it all worked on the first try, so that gives me some confidence even though I haven't tested it beyond "make check". Patch attached.

[Comment 4 François Bissey 2022-08-06 01:07:04 UTC]

Looks good to me.

[Comment 5 Michael Orlitzky 2022-08-06 01:17:30 UTC]

Reported it upstream...

[Comment 6 Larry the Git Cow 2024-03-05 04:48:41 UTC]

The bug has been closed via the following commit(s):

https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=dbda7d259111a90728d62fc56e6d664d5a745f0e

commit dbda7d259111a90728d62fc56e6d664d5a745f0e
Author:     Eli Schwartz <[e-mail]>
AuthorDate: 2024-03-05 01:37:21 +0000
Commit:     Sam James <[e-mail]>
CommitDate: 2024-03-05 04:47:50 +0000

    sci-mathematics/glpk: mark as LTO-unsafe, strict-aliasing unsafe
    
    It has been reported upstream but no response.
    
    Closes: https://bugs.gentoo.org/863047
    Signed-off-by: Eli Schwartz <[e-mail]>
    Signed-off-by: Sam James <[e-mail]>

 sci-mathematics/glpk/glpk-5.0-r1.ebuild | 11 ++++++++++-
 sci-mathematics/glpk/glpk-5.0-r2.ebuild |  9 +++++++++
 2 files changed, 19 insertions(+), 1 deletion(-)

[Comment 7 Larry the Git Cow 2024-10-24 14:36:44 UTC]

The bug has been referenced in the following commit(s):

https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=e9cc5b98df6ccbfc116cd7340f1aa63e56a233cb

commit e9cc5b98df6ccbfc116cd7340f1aa63e56a233cb
Author:     Michael Orlitzky <[e-mail]>
AuthorDate: 2024-10-21 17:04:12 +0000
Commit:     Michael Orlitzky <[e-mail]>
CommitDate: 2024-10-24 14:36:14 +0000

    sci-mathematics/glpk: steal strict-aliasing patch from Fedora
    
    This lets us bring LTO back without having to be responsible for a
    custom patch.
    
    Bug: https://bugs.gentoo.org/863047
    Signed-off-by: Michael Orlitzky <[e-mail]>

 sci-mathematics/glpk/files/glpk-5.0-aliasing.patch | 89 ++++++++++++++++++++++
 .../{glpk-5.0-r2.ebuild => glpk-5.0-r3.ebuild}     | 10 +--
 2 files changed, 90 insertions(+), 9 deletions(-)