https://blogs.gentoo.org/ago/2020/07/04/gentoo-tinderbox/ Issue: media-libs/libvisual-0.4.0-r3 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
Created attachment 793010 [details] build.log build log and emerge --info
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
The issue with strict aliasing here seems to be this: There is a struct "_color16" containing a bit field defined as: >> typedef struct { >> uint16_t b:5, g:6, r:5; >> } _color16; Now function "visual_color_to_uint16" uses a pointer cast to be able to access the combined uint16_t value in the return statement: >> uint16_t visual_color_to_uint16 (VisColor *color) >> { >> _color16 colors; >> >> visual_log_return_val_if_fail (color != NULL, 0); >> >> colors.r = color->r >> 2; >> colors.g = color->g >> 3; >> colors.b = color->b >> 2; >> >> return *((uint16_t *) &colors); >> } One way to workaround the strict aliasing violation would be to define a union that has two members — one of type _color16, one of type uint16_t — and access the underlaying bytes using that union (and I now see you did mention use of a union as an option). While it's good to know about the violation of strict aliasing, I'm not sure this is a practical issue rather than just an anomaly. What I'm trying to say is: Does this even need a fix in Gentoo and/or upstream? What do you suggest for moving forward?
In Fedora they are disabling strict-aliasing as a workaround https://src.fedoraproject.org/rpms/libvisual/blob/rawhide/f/libvisual.spec
(In reply to Sebastian Pipping from comment #3) > The issue with strict aliasing here seems to be this: > > There is a struct "_color16" containing a bit field defined as: > > >> typedef struct { > >> uint16_t b:5, g:6, r:5; > >> } _color16; > > Now function "visual_color_to_uint16" uses a pointer cast to be able to > access the combined uint16_t value in the return statement: > > >> uint16_t visual_color_to_uint16 (VisColor *color) > >> { > >> _color16 colors; > >> > >> visual_log_return_val_if_fail (color != NULL, 0); > >> > >> colors.r = color->r >> 2; > >> colors.g = color->g >> 3; > >> colors.b = color->b >> 2; > >> > >> return *((uint16_t *) &colors); > >> } > > One way to workaround the strict aliasing violation would be to define a > union that has two members — one of type _color16, one of type uint16_t — > and access the underlaying bytes using that union (and I now see you did > mention use of a union as an option). > > While it's good to know about the violation of strict aliasing, I'm not sure > this is a practical issue rather than just an anomaly. What I'm trying to > say is: Does this even need a fix in Gentoo and/or upstream? What do you > suggest for moving forward? It _will_ break at some point, so yes, please fix it, or we might have to mask it again, given how obvious the transgression is.
Made a pull request upstream for review now…
The bug has been closed via the following commit(s): https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=577e8bde7cef90e98d5fcdca120abd09318fc9da commit 577e8bde7cef90e98d5fcdca120abd09318fc9da Author: Sebastian Pipping <sping@gentoo.org> AuthorDate: 2022-12-05 23:57:42 +0000 Commit: Sebastian Pipping <sping@gentoo.org> CommitDate: 2022-12-06 00:15:44 +0000 media-libs/libvisual: 0.4.1 + EAPI 8 Closes: https://bugs.gentoo.org/859922 Closes: https://bugs.gentoo.org/871015 Closes: https://bugs.gentoo.org/882811 Signed-off-by: Sebastian Pipping <sping@gentoo.org> media-libs/libvisual/Manifest | 1 + media-libs/libvisual/libvisual-0.4.1.ebuild | 50 +++++++++++++++++++++++++++++ 2 files changed, 51 insertions(+)
