44 lines
1.4 KiB
Text
44 lines
1.4 KiB
Text
|
Hallo Thomas,
|
||
|
|
|
||
|
|
Dank je voor de nieuwe versie.
|
||
|
|
|
||
|
|
ov13b10-of/0001-media-i2c-ov13b10.patch
|
||
|
|
|
||
|
|
De eerste patch ziet er goed uit; zie punt 1 hieronder.
|
||
|
|
|
||
|
|
ov13b10-of/0000-cover-letter.patch
|
||
|
|
|
||
|
|
De cover noemt de test op 2026-09-10; dat klopt met wat ik zag.
|
||
|
|
|
||
|
|
Dates: every "still", "unmerged", "in review" and "latest version" below is
|
||
|
|
as of 2026-09-21 (work/linux at v7.1-rc2, work/milos-linux at v7.1.2-milos).
|
||
|
|
|
||
|
|
Reversals of the 2026-08-18 review (two points where that review was wrong):
|
||
|
|
|
||
|
|
R1. The active-array origin (ov13b10 3/3). The 08-18 review said the origin
|
||
|
|
was (0, 0); the datasheet [1] gives (8, 8), so the earlier ask is withdrawn.
|
||
|
|
|
||
|
|
R2. The link-frequencies constraint (ov13b10 1/3) depends on R1 and on the
|
||
|
|
binding example in [2]; it is still open as of 2026-09-21.
|
||
|
|
|
||
|
|
camss-milos
|
||
|
|
|
||
|
|
1. num_clients (camss 2/5). The field is read once; the v3 series on the
|
||
|
|
list [1] does the same.
|
||
|
|
a. The error path leaves the clock enabled.
|
||
|
|
b. The comment above it names the wrong register.
|
||
|
|
|
||
|
|
2. Required supplies (camss 1/5). Matches the binding in [2].
|
||
|
|
|
||
|
|
ov13b10-of
|
||
|
|
|
||
|
|
1. The crop default. A plain paragraph item without a citation at all.
|
||
|
|
|
||
|
|
across the three series
|
||
|
|
|
||
|
|
- Every series cites the same datasheet; the three cover letters should
|
||
|
|
agree on its revision.
|
||
|
|
- The Cc lists differ between the series.
|
||
|
|
|
||
|
|
[1] https://lore.kernel.org/r/7927bdab-9c3b-460d-a2ba-2e5b06ee8804@kernel.org/
|
||
|
|
[2] https://lore.kernel.org/all/20260815.4567@example.org/
|