core/db: Revert Patch Added In Attempt To Improve I/O Performance

Reverting the patch since it was causing a regression and after fixing the
regression, there were no performance gains.  At least based on my method
for measurement.

ASTERISK-24050

Review: https://reviewboard.asterisk.org/r/3841/


git-svn-id: https://origsvn.digium.com/svn/asterisk/trunk@419504 65c4cc65-6c06-0410-ace0-fbb531ad65f3