Let's look at the end of the commit message of commit fb57ea13e04d82866cbc8e86c83261148bb3e231:

    4. This patch causes new failures in the test suite:
    - recording-surface*: possibly an effect of improved quality.
    - surface-pattern-scale-down*, surface-pattern-big-scale-down: the
    reference images should be updated.
    - pthread-same-source: I have no idea why this is failing, since this
    test shouldn't even trigger the new code.
    - large-source-roi: this test attempts to downscale an image which is
    30000 pixels wide down to 7 pixels. The filter parameters seem to be
    created correctly, but they might trigger an overflow somewhere in the
    convolution code; the output rectangle is white instead of red, as if
    nothing was drawn.
    - device-offset-scale: there are subtle differences which look like
    convolution-related smoothing; I'm not sure whether this is OK or not

These changes should be reflected in the reference images for these tests.

Just for completeness, here is the full list of tests which started failing due to that commit (btw the above commit was found by git bisect... so much time wasted pointlessly...):

device-offset-scale
large-source-roi
pthread-same-source
recording-surface-extend-none
recording-surface-extend-pad
recording-surface-extend-reflect
recording-surface-extend-repeat
recording-surface-over
recording-surface-source
surface-pattern-big-scale-down
surface-pattern-scale-down
surface-pattern-scale-down-extend-none
surface-pattern-scale-down-extend-pad
surface-pattern-scale-down-extend-reflect
surface-pattern-scale-down-extend-repeat

Of these tests, as mentioned in the commit message, large-source-roi is the only real new failure, the rest looks like changes due to the new scaling code. For some weird reason this is also the case for pthread-same-source, although this code only scales some source up, not down. Sounds like there is still some bug hiding in the original commit...?