LGTM. This makes much more sense.

The concept seems fine with me and IIRC no one objected at the last core meeting.

`

In D22147#486635, @jhb wrote:

In D22147#486621, @imp wrote:

Nothing caught my eye, though the PLIST is getting to be rather long and a bit of a PITA to deal with :(

What I would really like is to have a shared pkg-plist and separate pkg-plist.${TARGETARCH} files. There doesn't seem to be an easy way to do this currently.

link command:

cc -target x86_64-unknown-freebsd13.0 --sysroot=/home/bed22/obj/home/bed22/git/freebsd/amd64.amd64/tmp -B/home/bed22/obj/home/bed22/git/freebsd/amd64.amd64/tmp/usr/bin -O2 -pipe -DCOMPAT_32BIT -march=i686 -mmmx -msse -msse2 -target x86_64-unknown-freebsd13.0 -m32 -L/home/bed22/obj/home/bed22/git/freebsd/amd64.amd64/obj-lib32/tmp/usr/lib32 --sysroot=/home/bed22/obj/home/bed22/git/freebsd/amd64.amd64/obj-lib32/tmp -B/home/bed22/obj/home/bed22/git/freebsd/amd64.amd64/obj-lib32/tmp/usr/lib32 -Wall -DFREEBSD_ELF -DIN_RTLD -ffreestanding -I/home/bed22/git/freebsd/lib/csu/common -I/home/bed22/git/freebsd/libexec/rtld-elf/i386 -I/home/bed22/git/freebsd/libexec/rtld-elf -fpic -DPIC -fvisibility=hidden -I/home/bed22/git/freebsd/libexec/rtld-elf/rtld-libc -mno-mmx -mno-sse -mno-avx -mno-avx2 -msoft-float -g -std=gnu99 -Wno-format-zero-length -Wsystem-headers -Werror -Wall -Wno-format-y2k -W -Wno-unused-parameter -Wstrict-prototypes -Wmissing-prototypes -Wpointer-arith -Wreturn-type -Wcast-qual -Wwrite-strings -Wswitch -Wshadow -Wunused-parameter -Wchar-subscripts -Winline -Wnested-externs -Wredundant-decls -Wold-style-definition -Wno-pointer-sign -Wformat=2 -Wno-format-extra-args -Werror -Wmissing-variable-declarations -Wthread-safety -Wno-empty-body -Wno-string-plus-int -Wno-unused-const-variable -Qunused-arguments  -nostdlib -e .rtld_start -shared -Wl,-Bsymbolic -Wl,-z,defs -Wl,--version-script=Version.map   -o ld-elf32.so.1.full rtld_start.o reloc.o rtld.o rtld_lock.o rtld_malloc.o rtld_printf.o map_object.o xmalloc.o debug.o libmap.o opendir.o closedir.o readdir.o telldir.o rtld_libc.o reallocf.o realpath.o getenv.o merge.o reallocarray.o errlst.o getcwd.o getprogname.o raise.o sigsetops.o sysctlnametomib.o __xuname.o  /home/bed22/obj/home/bed22/git/freebsd/amd64.amd64/libexec/rtld-elf32/rtld_libc.a

file output:

~/git/freebsd/libexec/rtld-elf32 buildenv> file `make -V .OBJDIR`/ld-elf32.so.1
/home/bed22/obj/home/bed22/git/freebsd/amd64.amd64/libexec/rtld-elf32/ld-elf32.so.1: ELF 32-bit LSB shared object, Intel 80386, version 1 (FreeBSD), dynamically linked, not stripped

• Default LIB32 to on where supported and mark broken elsewhere.

In D22147#486624, @jhb wrote:

@brooks suggested he would consider merging the llvm-xtoolchain ports into the llvm ports as well. I think it is nice that you can 'pkg install <foo>' to get the bits for 'make CROSS_TOOLCHAIN=<foo>'

In D22251#486697, @jhb wrote:

If this permits easily doing something like 'make buildenv; cd rtld-elf32; make' then this really will be nicer than our current approach.

In D22251#486698, @kib wrote:

I do not quite understand how even ld-elf32.so.1 can be built during non-compat build, not to mention normal binaries. For rtld, we need some object files from libc build. For binaries, we need libc.so.7 to link with. So how this stuff would work now ?

I've now testing this in an amd64 VM with i386 binaries so it's ready for review.

For the records, the CheriBSD commit is: https://github.com/CTSRD-CHERI/cheribsd/commit/805aa75408d105dd2c069c04e3b551caae24c558

This is extracted from CheriBSD. I'm not 100% it's something we want, but figure we should either upstream it (so I don't have compile errors every time a call to get_struct() is added) or remove it.

ARG, this was forked form the wrong branch. Will redo

This is extracted from CheriBSD. I'm not 100% it's something we want, but figure we should either upstream it (so I don't have compile errors every time a call to get_struct() is added) or remove it.

• Allow bsd.compat.mk to be reliably included outside Makefile.inc1.
• Get ARCH and CPUTYPE from contextually appropriate sources.
• Require that bsd.compat.mk be explicitly included and update docs.

I'm assuming this is something the powerpc people are going to want to land since it's a flag day change.

• Apply patch that arc applied, but didn't submit?!?!

• libstats: Fix ABI assertion.
• Require that bsd.compat.mk be explicitly included and update docs.

• Add a metion that "^" may need to be escaped.

Any further comments on this version?

A least AQR105_phy.c is still missing license headers. Please add headers to all source and header files.

A number of files are missing license headers including at least AQR105_phy.c and CX4.c. Please add headers to all source and header files.

LGTM

This looks good. I would like to commit it myself along with an update of the port to limit needless rebuilds (options changes require PORTREVISION bumps).

I believe the correct way to handle this is to update the vendor tree and then merge. I'm happy to do that.

• Get ARCH and CPUTYPE from contextually appropriate sources.

In D22059#482035, @imp wrote:

The question is, do we need them to be TARGET_* in this makefile.inc1 usage?
I kinda think it does. Checking into *that* detail...

Preparatory commits done in rS353592 rS353593 rS353594, rS353595.