Avoid racy access to the p_pid member of possibly reclaimed struct proc.
Other review' comments are handled.

In D54785#1252692, @jtl wrote:

In D54785#1252666, @kib wrote:

I do not think that trying to schedule a task is very robust with a LOR detected. We are already in the situation risking the deadlock.

Fair point!

If you refuse to make this a default, I am asking to add a simple knob which would make this additional info to be printed in all situations.

I don't "refuse"; I was just giving my thinking.

Ok, refuse was probably too strong word, sorry about this.

I do not think that trying to schedule a task is very robust with a LOR detected. We are already in the situation risking the deadlock.
Also I do not think that the doubts of the stack overflow is very grounded, since WITNESS typically has spinlock recording disabled II believe there are not solved issues with console spinlock which in reality cause hard hang). So there is definitely the space for interrupt frame, and mode, when WITNESS running.

Correct macro names.

Grammar and wording of the man page update.

Add note about the pid reuse.

In D54592#1252227, @asomers wrote:

This looks great, @kib! If you commit it, I'll follow up with the test suite. A few questions:

• Are you planning to write the man page yourself, or would you like help with that?

Add documentation.
Fix pdwait() prototype.

This should be extremely useful, I already anticipate it. But, why not make this information available by default? I am not even sure that the knob is needed to hide it.

Renumber AUE_PDRFORK.
Drop contrib/openbsm patch, similar to how other new AUE_ constants are not imported there.

In D54323#1252001, @cy wrote:

DSO bump.

Use lapic_maxlvt() everywhere.

Save file before diffing.

Add comment in vm_init.c

In D54768#1251863, @markj wrote:

To be clear, the point is to have an explicit panic rather than a boot hang because we end up spinning in an uninitialized trap handler?

Handle review' comments.

Add helper

Make this MI.

After a second thought, the pthread_tryjoin_np(3) function must be non-cancelling.

Convert test to the proper atf, instead of using err()

You might consider EXTERROR-ise kern_ktimer_create(), there are too many EINVALs there. Specific diagnostic would be useful.

Use array member for freebsd32_uint64_t on amd64.

Drop the patch to sort exterr_cat_filenames.h

In D54663#1250789, @brooks wrote:

In D54663#1250782, @kib wrote:

In D54663#1250781, @brooks wrote:

In D54663#1250766, @kib wrote:

After we agree on the approach and this change lands, I plan to do the pass over uses of val1/val2 in compat32, where int64_t val is used by the host ABI, converting them to freebsd32_uint64_t.

Please consider limiting this sweep to cases where the value will be mis-aligned relative to the struct.

Can you elaborate please, why? I understand that it formally changes nothing.

I'd rather we only have messy MD things where they are needed. That being said, the benefit of doing it everywhere is you do it once and don't have to think about it again so perhaps you are right that we should do it everywhere.

Switch to memcpy

In D54663#1250781, @brooks wrote:

In D54663#1250766, @kib wrote:

After we agree on the approach and this change lands, I plan to do the pass over uses of val1/val2 in compat32, where int64_t val is used by the host ABI, converting them to freebsd32_uint64_t.

Please consider limiting this sweep to cases where the value will be mis-aligned relative to the struct.

After we agree on the approach and this change lands, I plan to do the pass over uses of val1/val2 in compat32, where int64_t val is used by the host ABI, converting them to freebsd32_uint64_t.

freebsd32_uint64_t

In D54713#1250637, @markj wrote:

I wonder why we bother with bogus page replacement in sendfile_swapin(). Why not pass valid pages to the pager and let it decide how to handle it?

Make free_pid automatic, remote the proc_reap(free_pid) arg

Fix pdwait() userspace prototype.
Drop not useful operations on curproc procdesc.
Postpone freeing the zombie' pid until procdesc is freed.

In D54592#1249200, @asomers wrote:

I had to build an aarch64 VM just to test the RFSPAWN change, but it works now. It probably works on amd64 too, but I can't test it without writing assembly code.

I am fine with jrtc27' suggested name.

Did you tested this with lib/libc/tests/stdlib/tsearch_test.c?
If it works, I am fine with it.

Ok

In D54710#1249915, @jhb wrote:

In D54710#1249909, @kib wrote:

I think this is fine, but for completeness we probably should check that ehdr->e_phentsize == sizeof(Elf_Phdr), and return if not.

Yes. I do think we already check AT_PHENT, but might not be checking the size in the ehdr. Granted, that's a bug in the current code, and I might fix it in a separate commit.

I also noted it, but used find | sort in D54592.

You need to explain what you put into the summary, in the man page itself. Otherwise the reference to stack(9) looks out of context.

No, this is not going to work. This brings stdio + locale into rtld.

I think this is fine, but for completeness we probably should check that ehdr->e_phentsize == sizeof(Elf_Phdr), and return if not.