Backout previous (making sbin/efi arch dependent), will be redone
differently

sbin/efi is only available on architectures that provide efi runtime
services

Revert the duplication of compat set lists; do it another way.

1. Restore the compatdir and compatfile tags in base and debug.

=> compatdir marks directories that should have per-ABI compat
subdirectories, like /usr/lib having /usr/lib/o32 and
/usr/lib/64 on mips64.

=> compatfile marks files that have per-ABI versions in the
nearest compatdir, like /usr/lib/o32/libc.so.12.221.

This is the mechanism we use to avoid duplicating every library
that has a compat version in multiple set lists.

2. Teach sets.subr to split the compatdir and compatfile entries out
into separate sets for base and debug.

The mechanism is to internally use:

- a new tag `omitcompat' when processing base and debug, so we
don't generate the per-ABI compat directories or per-ABI compat
files under them;

- a new tag `takecompat=${ABI}' when processing base32/64 and
debug32/64, so we _do_ generate the per-ABI compat directories
for ${ABI}.

=> On, e.g., mips64, base32 uses `takecompat=o32' and base64
uses `takecompat=64'.

=> On, e.g., aarch64, base32 uses
`takecompat=eabi,takecompat=eabihf', and there is no base64.

3. Record in sets.subr the list of `32' and `64' architectures for
each port.

This is maybe not the right place to put it, but it's also not
clear that we want to keep exactly this division -- maybe it would
be more sensible to just have per-ABI , e.g.:

(aarch64) base-eabi, base-eabihf
(amd64) base-i386
(mips64) base-o32, base-64
(mipsn64) base-o32, base-n32

But I'll leave cleaning that up for another day.

4. Adjust some sets. Mostly this commit shouldn't change what file
go in what sets, but, e.g., on mips64, the following move from
base32 to base64 as part of this change, which I think is
appropriate:

- /lib/64
- /libexec/ld.elf_so-64
- /usr/lib/64
- /usr/libdata/debug/lib/64
- /usr/libdata/debug/usr/lib/64
- /usr/libexec/ld.elf_so-64

TBD: Remove ./lib and ./usr/lib from debug/mi; they're already in
base/mi. Maybe move ./usr/libdata/debug to from base/mi to debug/mi.

This reverts most of the set list changes in:

https://mail-index.netbsd.org/source-changes/2024/04/09/msg150763.html

PR misc/58365

branches: 1.88.2;
Add new sets: base32, debug32, manhtml

- base32 contains (when MKCOMPAT=yes) shared libraries for 32-bit
compatibility, previously included in base

- debug32 contains (when MKCOMPAT=yes) debug symbols and static libraries
containing debug symbols for 32-bit compatiblity, previously included
in debug

- manhtml contains (when MKHTML=yes) the HTML files previously included
in 'man', which are of limited utility without third-party software.

The motivation for this change is to be able to easily exclude sets
from CD-ROM images that go over the size limit without xz compression
(which many NetBSD platforms struggle to extract at acceptable speeds).

lists: Remove more bogus shlib obsolete entries.

Too much trouble to have to remember about libuv.so, libgmp.so, &c.
Just apply the rule uniformly.

All that's left is test libraries and loadable modules.

Searched with:

git grep 'lib.*\.so\.[0-9][0-9]*.*obsolete' distrib/sets/lists \
| grep -v '/modules/' \
| grep -v '/locale/lib/' \
| grep -v '/lib/npf/' \
| grep -v '/lib/librump' \
| grep -v '/lib/netbsd/libclang_rt' \
| grep -v '/runemodule/'

branches: 1.86.2;
Good bye oabi! Tested building earmv7hf.

Add HAVE_ACPI and HAVE_UEFI flags to bsd.own.mk and use them to
conditionally include ACPI and UEFI bits instead of relying on adhoc
MACHINE_ARCH tests.

Limit bootarm.efi expectation further.

Should fix iyonix, and zaurus builds. Also probably fixes all oarm AKA
'arm' builds too.

fix earmv7hfeb set list for bootarm.efi

Do this by explicitly marking all the non-earmv7hf* arm archs as not wanted,
because the set lists processor doesn't have a better way to be told.

Build EFI bootloader on big endian armv7 and aarch64 targets.

Add bootarm.efi

include the boot32 and nbfs boot programs for eabi, as acorn32
with eabi still wants them.

not tested. someone, please test. this fixes the acorn32 build
since it was switched to eabi.

Retire libpmc. It uses the legacy PMC interface in the kernel, which has
support for only one ARM CPU. It used to have x86 support, but it was broken
and I removed it. The legacy PMC interface will be removed from the kernel
too. Sent on tech-kern@ yesterday, thorpej was fine.

branches: 1.78.2; 1.78.4;
Remove port-acorn26

OK core@

Remove entries made redundant by compatdir/compatfile

new libssh

Fix set lists for NetBSD/ARM after the OABI addition.

bump libintl

bump liblzma

bump libssh

bump libpcap.

stuff added by blacklist

bump libamu.

bump libc for strtoi and strtou

Update of OpenSSL to 1.0.1k part 3:
- bump the shared library versions
- adjust set lists
- regenerate man pages
- note the import in doc/3RDPARTY and doc/CHANGES

there were no changes in asm parts

elftosb depends upon MKCXX.

bump libmagic

remove obsolete entries for libisccfg which are not.

bump for bind update

Consistently tag libasan with gcc=48

Add libasan

bump libssh

bump libc and add man pages for open_{,w}memstream

bump libstdc++ minor due to new complex etc functions enabled.

bump libm for long double complex functions

add execvpe,execlpe bump.

add strnstr from FreeBSD

branches: 1.54.2;
merge openssl 1.0.1i

Punt - backout sysinst message catlogs

Sigh - move sysinst catalog files (which differ per arch, ouch!)
to md lists.

Update set lists for libc bump and new man pages.

introduce a new pci_drvname(3) to libpci. implement it in the backend
via PCI_IOC_DRVNAME.

update manual and set lists (and remove a couple of doubled entries.)o

this will be used in libpciaccess() to implement the has_kernel_driver()
method.

fixed removal of Lua 5.1 entries from distrib/sets/lists

lua: updated from 5.1 to 5.3 work3

* lua(1):
- changed lua_Integer to intmax_t
- updated distrib/sets/lists and etc/mtree
- updated bsd.lua.mk
- fixed bozohttpd (lua-bozo.c)
- compatibilized bindings: gpio, sqlite
* lua(4):
- removed floating-point and libc dependencies using '#ifndef _KERNEL'
- fixed division by zero and exponentiation
- libkern: added isalnum(), iscntrl(), isgraph(), isprint() and ispunct()
- acpica: removed isprint() from acnetbsd.h
- libc: moved strcspn.c, strpbrk.c and strspn.c to common
- removed stub headers
- compatibilized bindings: luapmf, luasystm
* reorganized luaconf.h
* updated doc/CHANGES and doc/RESPONSIBLE

infrastructure for bind-9.10.0-P2:
adding the tools delv and dnssec-importkey
adding libirs to support delv

bump libc.

Add asysctl(3) and asysctlbyname(3) wrappers for the common idiom of
fetching dynamically sized data via sysctl.

Update for OpenLDAP 2.4.39.

libossaudio minor version bump to signify new ioctls.

sort

branches: 1.41.2;
bump libssl

fix libelf.so sets

Fix typo in library version.

bump libelf and libdwarf

Replace all the libbind9.so.7 with libbind9.so.8.
Hmm... why isn't the list of 32bit compat libs based on the normal list?

bump bind libraries

bump sqlite3

Update file lists for atf 0.19.

This has been tested with both MKKYUA=no (the default) and MKKYUA=yes, as
well as with a non-clean tree and a clean build. Hopefully I got the
details right... but they were tricky!

bump libc for ___lwp_park60

Finish updating set lists for libc minor bump.

bump openssl libraries

bump libc for ptsname_r

bump libisc for new bind

Add the syslog(3) Lua binding to the build.

bump libssh

more sets fixes (libgomp, libobjc)

add support for GCC 4.8 sets.

sort

bump libc

Add liblutok.so.2.

Also mark liblutok.so.1 as obsolete. While this should not be done in the
general case, I think it's fine here because MKKYUA still defaults to NO
and therefore lutok is not built by default.

bump for binutils

Update set lists after libc minor bump.

Sort entries in strcmp order.

Update sets lists for new bind9 and friends.

lists for new bind 9.9.3-P2

from the failed builds, it seems we always build ofctl and eeprom

This change arm, armeb, earm, earmeb, earmhf, earmhfeb so all builds that
share a MACHINE_ARCH for userland so that except for etc, will produce
identical sets.

usr/include/machine now points to usr/include/arm

Add eabi to compat dirs.
Add missing compat ld.elf_so-oabi

Now that sets.subr understands endian=1234 just use ad.arm for everything
since it can support {,e}arm{,hf}{,eb}

branches: 1.12.6; 1.12.8; 1.12.12; 1.12.14; 1.12.20;
Build fdformat on a MACHINE_ARCH not MACHINE basis

branches: 1.11.2;
Catch up with shlib major bump for all arm platforms.

make MKPIC=no builds work on i368, shark, vax, pmax, sgimips, sparc,
maybe all mips and all arm as well.

branches: 1.9.30; 1.9.38;
Move the lib<foo>.so entry (link) from comp to base, so that
applications on systems without the `comp' set installed can
dlopen() a library without supplying the major number.

reformat whitespace

* Clean up the handling of rescue.* files makeflist.
* Move all md.* and ad.* /rescue/... files into their
own list files.

tweak the ./rescue entries

Add libpmc.

* Deal with libarm32 -> libarm name change.
* Tidy up some of the ARM port set lists in preparation for
the complete switch to ELF.

Move files common to all ARM distribs to ad.arm

Now that LKMs work all over, bring the VFS LKMs and if_ipl LKM to mi list from
arch ones. Also make all VFS LKMs 'base-vfs-lkm'.

Fixup ARM distrib lists. Add evbarm distrib lists.