Refresh db_buf pointer after lock and conditional wait.

[IvanVolosyuk]

Motivation and Context

Try to solve the encrypted zfs send issue. This is purely theoretical change, I wasn't been able to reproduce the issue myself, but looking at the code this have come to mind.

Description

I'm not very familiar with the zfs code base, but I think there is a logic issue in the code.
The db_buf reference can be changed by the code holding dndb->db_mtx mutex and waiting on mutex can take a while, so the reference can become stale.

How Has This Been Tested?

Show pass the test suite test without an issue I would imagine.

Types of changes

• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Performance enhancement (non-breaking change which improves efficiency)
• Code cleanup (non-breaking change which makes code smaller or more readable)
• Breaking change (fix or feature that would cause existing functionality to change)
• Library ABI change (libzfs, libzfs_core, libnvpair, libuutil and libzfsbootenv)
• Documentation (a change to man pages or other documentation)

Checklist:

• My code follows the OpenZFS code style requirements.
• I have updated the documentation accordingly.
• I have read the contributing document.
• I have added tests to cover my changes.
• I have run the ZFS Test Suite with this change applied.
• All commit messages are properly formatted and contain Signed-off-by.

[amotin]

Dnode buffer should never change, since it is accessed from both open and syncing contexts. Even decryption of it is a special case, done in place.