storage: avoid expensive logging in MVCCAcquireLock

[nvb]

This commit puts expensive logging in MVCCAcquireLock behind a log.ExpensiveLogEnabled condition. This logging was bad in two ways:

1. it was boxing structs inside of interfaces, which can cause them to allocate and risks them escaping to the heap.
2. to avoid this risk of escaping, it was prematurely calling .String() on them. This was expensive for the key value, which was being pretty-printed even when tracing was disabled.

This showed up as taking 0.17% of CPU time in a run of TPC-C with Read Committed isolation:

This also showed up in the new MVCCAcquireLock_Pebble benchmark, which this commit improves:

name                                                                                          old time/op    new time/op    delta
MVCCAcquireLock_Pebble/batch=false/heldOtherTxn=false/heldSameTxn=true/strength=Shared-10       4.01µs ± 3%    2.42µs ± 1%  -39.70%  (p=0.000 n=10+9)
MVCCAcquireLock_Pebble/batch=false/heldOtherTxn=false/heldSameTxn=true/strength=Exclusive-10    4.02µs ± 2%    2.42µs ± 1%  -39.75%  (p=0.000 n=9+9)
MVCCAcquireLock_Pebble/batch=true/heldOtherTxn=false/heldSameTxn=true/strength=Shared-10        2.92µs ± 3%    1.60µs ± 2%  -45.18%  (p=0.000 n=9+10)
MVCCAcquireLock_Pebble/batch=true/heldOtherTxn=false/heldSameTxn=true/strength=Exclusive-10     2.91µs ± 2%    1.60µs ± 1%  -45.13%  (p=0.000 n=10+9)

name                                                                                          old alloc/op   new alloc/op   delta
MVCCAcquireLock_Pebble/batch=false/heldOtherTxn=false/heldSameTxn=true/strength=Shared-10         834B ± 0%      372B ± 0%  -55.38%  (p=0.000 n=10+10)
MVCCAcquireLock_Pebble/batch=false/heldOtherTxn=false/heldSameTxn=true/strength=Exclusive-10      834B ± 0%      372B ± 0%  -55.40%  (p=0.000 n=9+10)
MVCCAcquireLock_Pebble/batch=true/heldOtherTxn=false/heldSameTxn=true/strength=Shared-10          811B ± 0%      352B ± 0%  -56.58%  (p=0.000 n=10+10)
MVCCAcquireLock_Pebble/batch=true/heldOtherTxn=false/heldSameTxn=true/strength=Exclusive-10       810B ± 0%      352B ± 0%  -56.56%  (p=0.000 n=10+10)

name                                                                                          old allocs/op  new allocs/op  delta
MVCCAcquireLock_Pebble/batch=false/heldOtherTxn=false/heldSameTxn=true/strength=Shared-10         23.0 ± 0%       8.0 ± 0%  -65.22%  (p=0.000 n=10+10)
MVCCAcquireLock_Pebble/batch=false/heldOtherTxn=false/heldSameTxn=true/strength=Exclusive-10      23.0 ± 0%       8.0 ± 0%  -65.22%  (p=0.000 n=10+10)
MVCCAcquireLock_Pebble/batch=true/heldOtherTxn=false/heldSameTxn=true/strength=Shared-10          22.0 ± 0%       7.0 ± 0%  -68.18%  (p=0.000 n=10+10)
MVCCAcquireLock_Pebble/batch=true/heldOtherTxn=false/heldSameTxn=true/strength=Exclusive-10       22.0 ± 0%       7.0 ± 0%  -68.18%  (p=0.000 n=10+10)

Epic: None
Release note: None

[cockroach-teamcity]

This change is 

[arulajmani]

Reviewed 3 of 3 files at r1, 1 of 1 files at r2, all commit messages.
Reviewable status: complete! 1 of 0 LGTMs obtained (waiting on @sumeerbhola)

[sumeerbhola]

Reviewable status: complete! 2 of 0 LGTMs obtained (waiting on @nvanbenschoten)

[blathers-crl]

It looks like your PR touches production code but doesn't add or edit any test code. Did you consider adding tests to your PR?

_{🦉 Hoot! I am a Blathers, a bot for CockroachDB. My owner is dev-inf.}

[nvb]

TFTRs!

bors r+

[craig]

Build succeeded:

• Bazel Essential CI (Cockroach)