Differences

This shows you the differences between two versions of the page.

--- fourier:upstream_submissions [2026/04/28 22:06] – fourier: upstream submission drafts index (Qt, KDE, linux-media) markus_fritsche
+++ fourier:upstream_submissions [2026/04/30 08:56] (current) – kwin MR 9157 closed; vb2 RFC v1 sent + reviewed; v2 sketch backlogged; bugzilla + Kwiboo channels markus_fritsche
@@ Line 1: / Line 1: @@
 ====== fourier — upstream submission drafts ======
-Three submission drafts, written and parked, awaiting validation gates.
+Three submission drafts. Live in ''[[https://git.reauktion.de/marfrit/marfrit-packages|gitea/marfrit-packages]]'' under ''upstream-submissions/''.
-All live in &#039;&#039;[[https://git.reauktion.de/marfrit/marfrit-packages|gitea/marfrit-packages]]&#039;&#039; under
-&#039;&#039;upstream-submissions/&#039;&#039;.
 Index: [[fourier:start]].
@@ Line 9: / Line 7: @@
 ===== qt6-base-fourier → bugreports.qt.io + Gerrit =====
-  * **File:** &#039;&#039;upstream-submissions/qt6-base-fourier/qt-bug-report.md&#039;&#039;
+  * **File:** ''upstream-submissions/qt6-base-fourier/qt-bug-report.md''
   * **Targets:** bugreports.qt.io issue + Qt Gerrit MR
-  * **Why now-ish:** pure spec-correctness GLES3 fix, validated on two Mali generations (G52 ohm, T860 fresnel), zero regressions visible. Cleanly upstream-targetable.
+  * **Status:** parked. Pure spec-correctness GLES3 fix, validated on G52 ohm + T860 fresnel, zero regressions visible.
-===== kwin-fourier 0002 → invent.kde.org MR =====
+===== kwin-fourier → invent.kde.org MR =====
-  * **File:** &#039;&#039;upstream-submissions/kwin-fourier/kde-mr-body.md&#039;&#039;
+  * **Status (2026-04-30):** **MR 9157 closed.** Reviewer showed the upstream-shape patch (0002) was technically wrong — sync_file is a fence snapshot at export, dmabuf-fd polling tracks later-added fences, Mutter explicitly moved off that on amdgpu. No semantic difference, no syscall-count win, kernel blocking would be a kernel bug to fix.
-  * **Targets:** KDE GitLab (invent.kde.org) merge request against kwin
+  * **Currently shipped:** ''kwin-fourier 1:6.6.4-3'' with the **0001-skip-wait** patch (the pragmatic bypass). Fixes ohm V4L2 stateless playback stall. Will become a no-op once vb2-dma-resv RFC lands.
-  * **Patch:** &#039;&#039;0002-transaction-poll-dmabuf-fd-directly-upstream-shape.patch&#039;&#039;
+  * **0002 status:** retained on disk for reference; not for re-submission.
-  * **Why staged:** the diagnostic patch (0001) is a bypass not suitable for upstream; 0002 is the upstream-bound shape (poll dmabuf fd directly instead of going through &#039;&#039;EXPORT_SYNC_FILE&#039;&#039; + &#039;&#039;QSocketNotifier&#039;&#039;).
+  * **File:** ''upstream-submissions/kwin-fourier/kde-mr-body.md'' (post-mortem retained).
 ===== vb2-dma-resv RFC → linux-media =====
-  * **File:** &#039;&#039;upstream-submissions/vb2-dma-resv/lkml-submission-notes.md&#039;&#039; — full &#039;&#039;git send-email&#039;&#039; recipe with CC list
+  * **Status (2026-04-30):** **v1 sent 2026-04-29.** Two reviewer replies received same day from Nicolas Dufresne and Christian König. Not a NACK; directed redirect with permission slip for the fourier use case. **v2 sketch backlogged.**
-  * **Source:** &#039;&#039;kernel/vb2-dma-resv-rfc/&#039;&#039;
+  * **Cover Msg-Id:** ''<20260429195306.239666-1-mfritsche@reauktion.de>''
-  * **Target:** linux-media mailing list (RFC)
+  * **Lore thread:** [[https://lore.kernel.org/linux-media/20260429195306.239666-1-mfritsche@reauktion.de/]]
-  * **Patches:** vb2 helper API + hantro opt-in + rockchip-rga opt-in
+  * **Source patches:** ''kernel/vb2-dma-resv-rfc/''
-  * **APIs verified:** all against &#039;&#039;include/linux/dma-{fence,resv}.h&#039;&#039;
+  * **v1 send recipe:** ''upstream-submissions/vb2-dma-resv/lkml-submission-notes.md''
+  * **v2 sketch:** ''upstream-submissions/vb2-dma-resv/v2-cover-letter-sketch.md''
+  * **v2 prior-art reference:** ''upstream-submissions/vb2-dma-resv/v2-prior-art-references.md''
+  * **Full thread state:** [[fourier:vb2_dma_resv_rfc|fourier:vb2_dma_resv_rfc]] — reviewer points, decisions taken, lore-watch caveat.
-See [[fourier:vb2_dma_resv_rfc|fourier:vb2_dma_resv_rfc]] for why this
+===== Other submissions made this campaign =====
-exists and how it relates to the kwin patch.
+  * **Bugzilla 1969297 (Mozilla):** comment draft at ''upstream-submissions/firefox-fourier/bugzilla-1969297-comment.md''. User emailed David Turner directly with the gist 2026-04-29.
+  * **Kwiboo/FFmpeg libudev-bypass patch:** mailed to ''jonas@kwiboo.se'' (PRs disabled on the Kwiboo FFmpeg fork). Branch also at ''[[https://github.com/marfrit/FFmpeg/tree/libudev-bypass-fallback|github.com/marfrit/FFmpeg/tree/libudev-bypass-fallback]]''.