release-22.2: storage: avoid EncodeMVCCValue struct comparison

[blathers-crl]

Backport 1/1 commits from #88989 on behalf of @erikgrinaker.

/cc @cockroachdb/release

---

In Go 1.19, struct comparisons saw a significant performance regression, apparently due to the use of memeqbody. This patch changes EncodeMVCCValue to use field comparisons instead of struct comparisons, which yields a significant performance gain.

Unfortunately, this prevents mid-stack inlining. However, the struct comparison regression is significantly larger than the inlining gains. We should reconsider this once Go 1.20 lands, where the regression has been fixed.

name                                                              old time/op  new time/op  delta
EncodeMVCCValue/header=empty/value=tombstone-24                   5.96ns ± 1%  4.66ns ± 0%  -21.85%  (p=0.000 n=9+9)
EncodeMVCCValue/header=empty/value=short-24                       5.93ns ± 0%  4.66ns ± 0%  -21.40%  (p=0.000 n=9+9)
EncodeMVCCValue/header=empty/value=long-24                        5.92ns ± 0%  4.66ns ± 0%  -21.31%  (p=0.000 n=10+10)
EncodeMVCCValue/header=local_walltime/value=tombstone-24          51.9ns ± 1%  49.5ns ± 1%   -4.81%  (p=0.000 n=9+10)
EncodeMVCCValue/header=local_walltime/value=short-24              54.2ns ± 1%  52.5ns ± 1%   -3.25%  (p=0.000 n=10+10)
EncodeMVCCValue/header=local_walltime/value=long-24               1.34µs ± 2%  1.36µs ± 1%   +1.69%  (p=0.001 n=10+9)
EncodeMVCCValue/header=local_walltime+logical/value=tombstone-24  56.3ns ± 0%  53.3ns ± 1%   -5.40%  (p=0.000 n=10+10)
EncodeMVCCValue/header=local_walltime+logical/value=short-24      58.8ns ± 0%  56.3ns ± 2%   -4.18%  (p=0.000 n=10+9)
EncodeMVCCValue/header=local_walltime+logical/value=long-24       1.36µs ± 3%  1.36µs ± 1%     ~     (p=0.269 n=10+9)

Resolves #88818.

Release note: None

---

Release justification: fixes Go 1.19 performance regression.

[blathers-crl]

Thanks for opening a backport.

Please check the backport criteria before merging:

• Patches should only be created for serious issues or test-only changes.
• Patches should not break backwards-compatibility.
• Patches should change as little code as possible.
• Patches should not change on-disk formats or node communication protocols.
• Patches should not add new functionality.
• Patches must not add, edit, or otherwise modify cluster versions; or add version gates.

If some of the basic criteria cannot be satisfied, ensure that the exceptional criteria are satisfied within.

• There is a high priority need for the functionality that cannot wait until the next release and is difficult to address in another way.
• The new functionality is additive-only and only runs for clusters which have specifically “opted in” to it (e.g. by a cluster setting).
• New code is protected by a conditional check that is trivial to verify and ensures that it only runs for opt-in clusters.
• The PM and TL on the team that owns the changed code have signed off that the change obeys the above rules.

Add a brief release justification to the body of your PR to justify this backport.

Some other things to consider:

• What did we do to ensure that a user that doesn’t know & care about this backport, has no idea that it happened?
• Will this work in a cluster of mixed patch versions? Did we test that?
• If a user upgrades a patch version, uses this feature, and then downgrades, what happens?

[cockroach-teamcity]

This change is 

[jbowens]

lgtm