I do not have any opinion on the change, perhaps the only feedback you need is from jkim@ who does the acpica imports.
I only have style/cosmetics notes.

In D55494#1269805, @kib wrote:

Sorry for spamming.

More, I suggest to implement it by adding the new value to the exterr_verbose_state enum. The value of the variable would be checked only once then.

In D55230#1269768, @asomers wrote:

@kib is there anything else you'd like me to do for this PR, short of rewriting everything so the vnode lock isn't held while waiting for the daemon?

IB core does pass the DMA_ATTR_WRITE_BARRIER. What is the semantic for the flag?

In D55480#1269558, @jrtc27 wrote:

In D55480#1269551, @kib wrote:

First, there are two unrelated changes. One for the in-kernel ELF image activator, another for the kernel linker.

Then, for the image activator, why the is_cheri checks needs to be done in MI code? We have brand matchers (header_supported/brand_supported), which get the pointer to the mapped ELF header. Why cannot it be done there, in <arch>/elf_machdep.c?

You could make the same argument for e_machine/EI_CLASS/EI_DATA. Like those, this is an MI concept, so it's in MI code rather than having MD boilerplate (that, I will note, also has some odd semantics when it comes to the bool-returning functions, though I don't remember off the top of my head what those were). The MD hooks are for MD things like "v1 or v2 of this architecture's psABI". Whereas this is part of ILP32-vs-LP64-vs-"L64PC128" (or whatever name gets given to it, if any), and corresponds to a specific instantiation of the imgact_elf.c template (Elf32 vs Elf64 vs "Elf64C")

Sorry for spamming.

Thanks

First, there are two unrelated changes. One for the in-kernel ELF image activator, another for the kernel linker.

I also wanted to suggest that the reasons should be explained not only in the commit message, but also in the comment in the file itself. I hesitated, since this is tools/build and of limited scope.

IMO other bits should be also moved to specialreg.h

Perhaps the defines should go into x86/include/specialreg.h together with all other CPUID bits definitions.

Also use SIZE_TYPE instead of size_t since it's not guaranteed to be defined.

I would add a local like const struct asmc_model *m;, and set it to the found 'model'. Then goto to the end of the function, which would do

free(model);
return (m);

This way you have free() only in one place, and all returns would be guaranteed to free the memory.

The would it be the time to also change the macros itself to not pass 0 but the last arg?
(I do not have objections to either this patch or updated patch).

In D55378#1266700, @markj wrote:

In D55378#1266686, @kib wrote:

You could divide by 128 if this is so important.

Could you please explain the motivation for the changes?

I was looking at sources of frequent priv_check() calls from unprivileged processes, trying to see if any are avoidable. These calls might do some non-trivial work if MAC hooks are installed.

You could divide by 128 if this is so important.

In D55365#1266528, @bapt wrote:

https://reviews.freebsd.org/D55364 now has your tests and they pass

In D55365#1266416, @bapt wrote:

In D55365#1266404, @kib wrote:

In D55365#1266392, @bapt wrote:

I may be missing some idea on what to test

Might be add some tests for e.g. ru_RU.UTF-8? What do you need for this?

In D55365#1266404, @kib wrote:

In D55365#1266392, @bapt wrote:

I may be missing some idea on what to test

Might be add some tests for e.g. ru_RU.UTF-8? What do you need for this?

juste provide me a set of command you would expect to work in ru_RU.UTF-8 and and expected output and I will add it to the test suite.

In D55365#1266392, @bapt wrote:

I may be missing some idea on what to test

Do regexps addresses and commands like 's' correctly operate on mb strings?

Add comment.

Return on any error, not excluding ESRCH.

Fix grammar.