refactor: deprecate *COUNT consts fitting #3131

[dybucc]

Description

Constants matching the *COUNT naming scheme, and whose purpose fits that of a "delimitter" value
have been deprecated as part of the deprecation efforts stemming from #3131.

This patch continues deprecation of symbols matching one of the most common suffixes in identifiers
for constants whose value fits that of a "limit." The changes here follow from those in
#3620, with the recommended fixes by Amanieu (not pinging for now.)

This is a follow up PR to #5120.

Notes

A few of the constants I couldn't decide on under mach/host_info.h and sys/attr.h in Apple's
distribution of the XNU kernel have been deprecated as per prior work by Lilit0x and comments by
Amanieu in #3620.

I couldn't pinpoint where was the DLFO_STRUCT_HAS_EH_COUNT constant declared and used in the GNU
Hurd upstream repo, so I haven't been able to make any decisions there.

Sources

• Haiku sources showing how the B_MEDIA_STALE_CHANGE_COUNT is not the type of constant we
  would want to deprecate. The code upstream itself does not seem to use it as a delimitter value,
  as can be seen in the second search result of the below reference.

  https://github.com/search?q=repo%3Ahaiku%2Fhaiku%20B_MEDIA_STALE_CHANGE_COUNT&type=code

• DragonFly BSD sources showing how the CPUCTL_CPUID_COUNT constant is not one we would want
  to deprecate, even though it may seem like it could be deprecated (because it appears as a
  *COUNT symbol after a bunch of other symbols.) This constant is actually used for literal CPU
  counts in
  https://github.com/DragonFlyBSD/DragonFlyBSD/blob/master/sys/dev/misc/cpuctl/cpuctl.c#L145.

  https://github.com/DragonFlyBSD/DragonFlyBSD/blob/master/sys/sys/cpuctl.h#L62

• OpenBSD sources showing how the KERN_TTYCOUNT is actually used as a counter for TTY devices and
  not for other constants (just another enum-like "variant".) This does not qualify it for
  deprecation.

  https://github.com/search?q=repo%3Aopenbsd%2Fsrc%20KERN_TTYCOUNT&type=code

• OpenBSD sources showing how the KERN_EVCOUNT is used as an event counter (with use akin to that
  of the above list item.) This does not qualify it for deprecation.

  https://github.com/search?q=repo%3Aopenbsd%2Fsrc%20KERN_EVCOUNT&type=code

• NetBSD sources showing how the DCCP_OPT_NDP_COUNT constant is used as just a regular constant
  (not meant for deprecation.)

  https://github.com/NetBSD/src/blob/d28519920c966c62fedd65ccc0dbd8b628401da5/sys/netinet/dccp.h#L177

• Linux sources showing how the TIOCGICOUNT constant seems to be used not as a counter for other
  symbols but as (possibly runtime) interrumpt counter (not meant to be deprecated.)

  https://github.com/torvalds/linux/blob/eb3f4b7426cfd2b79d65b7d37155480b32259a11/include/uapi/asm-generic/ioctls.h#L99

• FreeBSD sources showing how the IFMIB_IFCOUNT constant is promptly liable to upstream changes as
  it denotes the number of configured interfaces.

  https://github.com/freebsd/freebsd-src/blob/f9f46294d6af2a937afa74938bd4bb6826cbb921/sys/net/if_mib.h#L60

• XNU kernel sources showing how the RTV_HOPCOUNT is not the type of constant we would want to
  deprecate. It seems to be more of an enum "variant" among other similar constants.

  https://github.com/apple-oss-distributions/xnu/blob/f6217f891ac0bb64f3d375211650a4c1ff8ca1ea/bsd/net/route.h#L228

• Illumos sources showing that the IPV6_MINHOPCOUNT is not dependent of other source code chagnes
  (and thus does not require deprecation.)

  https://github.com/illumos/illumos-gate/blob/7124a0cccb6ebf2bcffdab36f9bdf8fe4c23d77f/usr/src/uts/common/netinet/in.h#L1309

• L4RE sources showing how the SIOCGIFCOUNT constant is more of an enum "variant" among other
  values, and thus not a targe of these deprecation efforts.

  https://github.com/search?q=repo%3Akernkonzept%2Fl4re-core%20SIOCGIFCOUNT&type=code

• Illumos sources showing how the LGRP_RSRC_COUNT constant is meant to count the number of
  resource types, which for the time being are the two constants below it.

  https://github.com/search?q=repo%3Aillumos%2Fillumos-gate%20LGRP_RSRC_COUNT&type=code

• Linux sources mentioning that the constants declared prior to the NFNL_SUBSYS_COUNT symbol could
  have been in an enumeration where the last value is the current "limit."

  https://github.com/search?q=repo%3Atorvalds%2Flinux%20NFNL_SUBSYS_COUNT&type=code

• Android sources replicating similar behavior as in Linux.

  https://cs.android.com/search?q=NFNL_SUBSYS_COUNT&sq=&ss=android%2Fplatform%2Fsuperproject

Checklist

• Relevant tests in libc-test/semver have been updated
• No placeholder or unstable values like *LAST or *MAX are
  included (see #3131)
• Tested locally (cd libc-test && cargo test --target mytarget);
  especially relevant for platforms that may not be checked in CI

[rustbot]

Some changes occurred in an Android module

cc @maurer

Some changes occurred in a solarish module

cc @jclulow, @pfmooney