Update amd in base from 6.1.5 to 6.2.0.

[cy]

Summary

This revision will MFV r305100 to update amd in base to 6.2.0.

The patch will fail to apply to some man pages due to compare failures of $FreeBSD$ in the diff.

Test Plan

Already running on my infrastructure of four machines at home.

make tinderbox succeeded on universe12b.

[cy]

Added kib to review.

[kib]

What exactly do you want me to review ? I am not familiar with the amd code. I used it, but that was many years ago.

If you have specific questions which you think I am competent to answer, please state them explicitly.

[cy]

Hi kib@,

Since amd is upstream, I'm primarily interested in the changes to Makefiles in usr.sbin/amd. Even though it tinderbox builds and I'm running it here, I'd appreciate a quick look through the Makefiles to ensure I haven't missed anything.

~cy@

[kib]

In D8405#175185, @cy wrote:

Hi kib@,

Since amd is upstream, I'm primarily interested in the changes to Makefiles in usr.sbin/amd. Even though it tinderbox builds and I'm running it here, I'd appreciate a quick look through the Makefiles to ensure I haven't missed anything.

I am not too familiar with our build system, you might ask e.g. bdrewery, emaste and brooks about it.

I took some look over the config.h, and I find it strange that e.g. HAVE_FFS is not defined. We also have MFS and TMPFS. I am not sure whether what we have is similar to what amd wants from things with that names.

[cy]

I've updated config.h to enable MFS.

Though "HAVE_FFS" exists in the upline config.h.in, the source makes no reference to it:

slippy$ ag -sc HAVE_FFS
slippy$ ag -sc HAVE_UFS
amd/ops_ufs.c:8
amd/ops_udf.c:1
conf/nfs_prot/nfs_prot_freebsd3.h:4
include/am_defs.h:5
include/mount_headers1.h:4
slippy$

I've added bdrewery, emaste and brooks to look over the Makefiles. I'm especially interested in a review of the yacc and lex invocations.

[brooks]

I don't think it would be very hard to fix the infrastructure to support multiple yacc and lex source files, I've got some other work to procrastinate on so I'l take a stab at that. It would clean these Makefile changes up quite a bit.

[cy]

usr.sbin/acpi/iasl has the same issue. What if I commit the update to amd now and we address acpi and amd later? I haven't looked but I'm willing to bet there are others as well.

If you don't mind, I can put together a patch and post it here on phabricator, and you and others can review/approve it?

~cy

[brooks]

I've added the share/mk changes that should remove the need for most of the Makefile hackery to D8438.

[cy]

D8438 doesn't address custom symbol prefixes (-p for yacc and -P for lex), use here in amd and in acpi/iasl. Let's work through D8438 first before we continue here.

[brooks]

I've abandoned the other revision, it was wrong in several ways and doesn't appear to be easily fixed.

usr.sbin/amd/amd/Makefile

12 ↗

(On Diff #21962)

I'm a bit surprised this part is necessary. It seems like make should find things in .OBJDIR on it's own.

[cy]

usr.sbin/amd/amd/Makefile

12 ↗

(On Diff #21962)

It does. I inserted that while working on the multiple yacc and lex sources at some point and failed to remove it. It's been removed locally.

[pfg]

Sorry, I have been AFK for a while.
Looks good to me.