bsd.compiler.mk: Fix cross-building from non-FreeBSD
cef65082ae12
Actions

Description

bsd.compiler.mk: Fix cross-building from non-FreeBSD

On non-FreeBSD, the various MACHINE variables for the host when
bootstrapping can be missing or not match FreeBSD's naming, causing
bsd.endian.mk to be unable to infer the endianness. Work around this by
assuming it's unsupported.

Note that we can't check BOOTSTRAPPING here as Makefile.inc1 includes
bsd.compiler.mk before that is set, and so we are unable to catch errors
during buildworld itself when cross-building and bsd.endian.mk failed,
but such errors should also show up when building on FreeBSD.

Fixes: 47363e99d3d3 ("Enable compressed debug on little-endian targets")

Details

Provenance

jrtc27

Authored on Aug 12 2021, 10:45 PM

Parents

rGc5bf58add0d5: Revert "ar: diff reduction against ELF Tool Chain"

Branches

Unknown

Tags

Unknown

Event Timeline

jrtc27 committed rGcef65082ae12: bsd.compiler.mk: Fix cross-building from non-FreeBSD (authored by jrtc27).Aug 12 2021, 10:45 PM

arichardson added a subscriber: arichardson.Aug 16 2021, 8:40 AM

arichardson added inline comments.

/share/mk/bsd.compiler.mk
248	Wouldn't it make more sense to add something like .else .error "Cannot infer endianess for MACHINE=${MACHINE} MACHINE_ARCH=${MACHINE_ARCH} MACHINE_CPUARCH=${MACHINE_CPUARCH} " to bsd.endian.mk?

jrtc27 added inline comments.Aug 16 2021, 9:20 AM

/share/mk/bsd.compiler.mk
248	Then you'd have to add a complex guard around `.include <bsd.endian.mk>` to avoid triggering the error, or have a simple "not FreeBSD" one that then doesn't infer endianness even when the machine is known. This way it does the best it can. You still get an error if you blindly use TARGET_ENDIANNESS (what this commit is fixing, the old malformed conditional due to an undefined variable).

arichardson added inline comments.Aug 16 2021, 9:24 AM

/share/mk/bsd.compiler.mk
248	I just had a look at bsd.endian.mk and I see also sets other variables that will be empty for unknown MACHINE/etc. variables. I believe those are only needed for the TARGET_* values so yeah I guess the host value doesn't really matter right now. Maybe a .warning would be appropriate though (in addition to this commit)?

jrtc27 added inline comments.Aug 16 2021, 9:32 AM

/share/mk/bsd.compiler.mk
248	That just seems like it'll be noisy, giving out warnings users can't do anything about and we know get hit. It's going to fail anyway if the include site isn't prepared for them to be undefined.