Fix dmu_recv_stream test for resumable

[PaulZ-98]

Use dsl_dataset_has_resume_receive_state()
not dsl_dataset_is_zapified() to check if
stream is resumable.

Signed-off-by: Paul Zuchowski [email protected]

There is code at the top of dmu_recv_stream() to check if the stream is resumable. The code as written seems to imply that if dsl_dataset_is_zapified(), then a DS_FIELD_RESUME_BYTES zap entry will be found. This is inconsistent with other places in ZFS where the more complete dsl_dataset_has_resume_receive_state() is used for this.

Because of this test, the zap_lookup for DS_FIELD_RESUME_BYTES does fail from time to time with ENOENT, and an uninitialized value can end up in drc_bytes_read, sometimes causing a panic at places where bytes_read is VERIFY'd.

Motivation and Context

This caused a panic in #9372 automated testing, under FreeBSD 13.

Description

This fix uses dsl_dataset_has_resume_receive_state() as the test, and initializes the bytes variable to 0 in case the zap lookup fails.

How Has This Been Tested?

Ran rsend and redacted_send portions of zfs tests.

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.

[ghost]

I don't think changing the check helps in practice, but it illustrates the intent at least. Nice find.

[PaulZ-98]

I just investigated this, and noted that the change to dsl_dataset_has_resume_receive_state eliminates the occurrences of zap_lookup / DS_FIELD_RESUME_BYTES returning ENOENT while running the redacted_send zfs tests. If the check is dsl_dataset_is_zapified, you get many instances of ENOENT while running the redacted_send zfs tests.