Restrict short block cloning requests

[amotin]

If we are copying only one block and it is smaller than recordsize property, do not allow destination to grow beyond one block if it is not there yet. Otherwise the destination will get stuck with that block size forever, that can be as small as 512 bytes, no matter how big the destination grow later.

I've noticed this while working on #15251.

How Has This Been Tested?

Before the patch:

# cd /optane
# dd if=/dev/random of=f1 bs=512 count=1
# dd if=/dev/random of=f2 bs=512 count=1
# zpool sync optane
# cat f1 f2 > f3
# zpool sync optane
# zdb -O optane f3

    Object  lvl   iblk   dblk  dsize  dnsize  lsize   %full  type
       640    2    32K    512     2K     512     1K  100.00  ZFS plain file

After the patch:

# zdb -O optane f3

    Object  lvl   iblk   dblk  dsize  dnsize  lsize   %full  type
      1025    1    32K     1K     4K     512     1K  100.00  ZFS plain file

For the same test with bs=128K cloning is used in both cases.

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.

[behlendorf]

@robn @pjd please take a look.

[robn]

Good find!

(My gosh, these edge cases are hard work).