Seems correct at a glance, but maybe let @imp weight in before committing.

Does pig-state eventually lose the -N argument entirely?

Seems reasonable. Sorry I missed that MCLBYTES is something people adjust upward.

In D50126#1144298, @ziaee wrote:

A level 0 dump is like dd. It gets absolutely everything, unlike a boot environment. I'll work on that after I merge this one.

To fix an a separate commit: the foot note ordering is quite jarring.

If this were i386 or amd64 then yes, but armv7 is a tier-2 architecture so we can break the ABI contract in ways that are harmless. IMO there's lots of value in making lower tier platforms less weird.

There's an argument there should be a DECLARE_SFF_8024_ID macro that consumers add to their .c file, but is easy and has precedent.

Use Nm to render libutil in HISTORY

getmntopts.c -> mntopts.c to match header and manpage.

• actually install the manpage
• add guards to header

Sorry for the churn, but I think @kib is right and this should be declared in libc_private.h and not namespaced.

Sorry, I should have added you to D49951 and D49952 which this duplicates. I don't think a new library is warranted and if it is, it should probably be libmntopts or the like not something as generic as libmount.

This came in with D17587 for compatibility with old object files. It was roughly contemporaneous with GCC 6.5, the last release with GCJ support.

This seems alright.

This should be less confusing.

You still need to add a an #undef execvpe line to un-namespace.h and then you shouldn't need the libc_private.h entry. The point of namespace.h/un-namespace.h is that the declarations are impacted by the macro so the execvpe declaration in unistd.h becomes an __libc_execvpe declaration which you can then use.

I think you need to bring the __weak_reference line back from the original change.

In D49666#1132660, @emaste wrote:

Is f3a73a7055a8 the right hash -- maybe e6e5cd297ab4cae4995895b99c713df9cf6e8c0f? Change lgtm.

Seems fine conceptually, a couple comments.

fix typo pointed out by @emaste

• s/answers answers/values/ as pointed out by @jbo