dri-devel Archive on lore.kernel.org
 help / color / mirror / Atom feed
* [PATCH v4 00/20] drm: Increase COMPILE_TEST=y coverage
@ 2026-11-18 21:44 Ville Syrjala
  2026-11-18 21:44 ` [PATCH v4 01/20] drm/tilcdc: Allow build without __iowmb() Ville Syrjala
                   ` (19 more replies)
  0 siblings, 20 replies; 36+ messages in thread
From: Ville Syrjala @ 2026-11-18 21:44 UTC (permalink / raw)
  To: dri-devel
  Cc: Alessio Belle, Alison Wang, asahi, Danilo Krummrich,
	Geert Uytterhoeven, imagination, Janne Grunau, Jyri Sarha,
	Kieran Bingham, Kory Maincent, Laurent Pinchart, linux-mips,
	linux-renesas-soc, Luigi Santivetti, Lyude Paul, Magnus Damm,
	nouveau, Paul Cercueil, Russell King, Sasha Finkelstein,
	Stefan Agner, Tomi Valkeinen, Tomi Valkeinen

From: Ville Syrjälä <ville.syrjala@linux.intel.com>

This series contains a few of my earlier COMPILE_TEST=y coverage
patches that are still not applied. I've also included whole new
set of patches based a more recent excursion with Claude. It
managed to identify some additional drivers we can target.

Compile tested on x86-32, x86-64, arm, and arm64, and I've also 
tried to silence all the resulting new compiler and sparse warnings.
Also a few kernel-doc warnings crept up in pvr so also tried to
clear those out.

The tilcdc DT change is the only one that I think might have
any functional impact. And since I have zero idea about DT
treat it with suspicion!

Assisted by Claude Sonnet 5.5

Cc: Alessio Belle <alessio.belle@imgtec.com>
Cc: Alison Wang <alison.wang@nxp.com>
Cc: asahi@lists.linux.dev
Cc: Danilo Krummrich <dakr@kernel.org>
Cc: Geert Uytterhoeven <geert+renesas@glider.be>
Cc: imagination@lists.freedesktop.org
Cc: Janne Grunau <j@jannau.net>
Cc: Jyri Sarha <jyri.sarha@iki.fi>
Cc: Kieran Bingham <kieran.bingham+renesas@ideasonboard.com>
Cc: Kory Maincent <kory.maincent@bootlin.com>
Cc: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Cc: linux-mips@vger.kernel.org
Cc: linux-renesas-soc@vger.kernel.org
Cc: Luigi Santivetti <luigi.santivetti@imgtec.com>
Cc: Lyude Paul <lyude@redhat.com>
Cc: Magnus Damm <magnus.damm@gmail.com>
Cc: nouveau@lists.freedesktop.org
Cc: Paul Cercueil <paul@crapouillou.net>
Cc: Russell King <linux@armlinux.org.uk>
Cc: Sasha Finkelstein <k@chaosmail.tech>
Cc: Stefan Agner <stefan@agner.ch>
Cc: Tomi Valkeinen <tomi.valkeinen@ideasonboard.com>
Cc: Tomi Valkeinen <tomi.valkeinen+renesas@ideasonboard.com>

Ville Syrjälä (20):
  drm/tilcdc: Allow build without __iowmb()
  drm/tilcdc: Drop volatile from the mmio pointer in tilcdc_write64()
  drm/tilcdc: Use __le64 for the fallback 64bit write
  drm/tilcdc: Fix endpoint node names in the legacy panel overlay
  drm/tilcdc: Allow build with COMPILE_TEST=y
  drm/armada: Fix printk arguments
  drm/armada: Fix armada_debugfs_crtc_reg_write() return type
  drm/armada: Fix sparse __iomem address space warnings
  drm/armada: Allow build with COMPILE_TEST=y
  drm/fsl-dcu: Allow build with COMPILE_TEST=y
  drm/adp: Include linux/bitfield.h
  drm/adp: Allow build with COMPILE_TEST=y
  drm/imagination: Include io-64-nonatomic-lo-hi.h
  drm/imagination: Use rcu_access_pointer() for dma_fence::ops
  drm/imagination: Allow build with COMPILE_TEST=y
  drm/imagination: Drop stale kernel-doc for
    pvr_mmu_backing_page.pvr_dev
  drm/imagination: Fix kernel-doc for pvr_mmu_op_context anonymous
    structs
  drm/ingenic: Allow ingenic-dw-hdmi build with COMPILE_TEST=y
  drm/nouveau: Allow platform driver build with COMPILE_TEST=y
  drm/rcar-du: Allow VSP and writeback build with COMPILE_TEST=y

 drivers/gpu/drm/adp/Kconfig                     |  2 +-
 drivers/gpu/drm/adp/adp_drv.c                   |  1 +
 drivers/gpu/drm/armada/Kconfig                  |  2 +-
 drivers/gpu/drm/armada/armada_debugfs.c         |  2 +-
 drivers/gpu/drm/armada/armada_fbdev.c           |  2 +-
 drivers/gpu/drm/armada/armada_gem.c             |  7 ++++---
 drivers/gpu/drm/fsl-dcu/Kconfig                 |  2 +-
 drivers/gpu/drm/imagination/Kconfig             |  3 ++-
 drivers/gpu/drm/imagination/pvr_ccb.c           |  5 +++--
 drivers/gpu/drm/imagination/pvr_device.h        |  1 +
 drivers/gpu/drm/imagination/pvr_mmu.c           | 14 ++++++--------
 drivers/gpu/drm/imagination/pvr_queue.c         | 17 +++++++++--------
 drivers/gpu/drm/ingenic/Kconfig                 |  2 +-
 drivers/gpu/drm/nouveau/Kconfig                 |  2 +-
 drivers/gpu/drm/renesas/rcar-du/Kconfig         |  4 ++--
 drivers/gpu/drm/tilcdc/Kconfig                  |  2 +-
 drivers/gpu/drm/tilcdc/tilcdc_panel_legacy.dtso |  4 ++--
 drivers/gpu/drm/tilcdc/tilcdc_regs.h            |  7 +++++--
 18 files changed, 43 insertions(+), 36 deletions(-)

-- 
2.56.0

^ permalink raw reply	[flat|nested] 36+ messages in thread

* [PATCH v4 01/20] drm/tilcdc: Allow build without __iowmb()
  2026-11-18 21:44 [PATCH v4 00/20] drm: Increase COMPILE_TEST=y coverage Ville Syrjala
@ 2026-11-18 21:44 ` Ville Syrjala
  2026-11-18 21:44 ` [PATCH v4 02/20] drm/tilcdc: Drop volatile from the mmio pointer in tilcdc_write64() Ville Syrjala
                   ` (18 subsequent siblings)
  19 siblings, 0 replies; 36+ messages in thread
From: Ville Syrjala @ 2026-11-18 21:44 UTC (permalink / raw)
  To: dri-devel; +Cc: Tomi Valkeinen, Jyri Sarha

From: Ville Syrjälä <ville.syrjala@linux.intel.com>

__iowmb() isn't available on most architectures. Make
its use optional so that the driver can be built on
other architectures with COMPILE_TEST=y.

v2: Add a comment stating this is for COMPILE_TEST (Tomi)

Cc: Tomi Valkeinen <tomi.valkeinen@ideasonboard.com>
Acked-by: Jyri Sarha <jyri.sarha@iki.fi>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
---
 drivers/gpu/drm/tilcdc/tilcdc_regs.h | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/drivers/gpu/drm/tilcdc/tilcdc_regs.h b/drivers/gpu/drm/tilcdc/tilcdc_regs.h
index 26ebaf1e0f70..55fe598fe770 100644
--- a/drivers/gpu/drm/tilcdc/tilcdc_regs.h
+++ b/drivers/gpu/drm/tilcdc/tilcdc_regs.h
@@ -125,7 +125,10 @@ static inline void tilcdc_write64(struct drm_device *dev, u32 reg, u64 data)
 #if defined(iowrite64) && !defined(iowrite64_is_nonatomic)
 	iowrite64(data, addr);
 #else
+	/* allow compilation without __iowmb() for COMPILE_TEST */
+#ifdef __iowmb
 	__iowmb();
+#endif
 	/* This compiles to strd (=64-bit write) on ARM7 */
 	*(volatile u64 __force *)addr = __cpu_to_le64(data);
 #endif
-- 
2.56.0


^ permalink raw reply related	[flat|nested] 36+ messages in thread

* [PATCH v4 02/20] drm/tilcdc: Drop volatile from the mmio pointer in tilcdc_write64()
  2026-11-18 21:44 [PATCH v4 00/20] drm: Increase COMPILE_TEST=y coverage Ville Syrjala
  2026-11-18 21:44 ` [PATCH v4 01/20] drm/tilcdc: Allow build without __iowmb() Ville Syrjala
@ 2026-11-18 21:44 ` Ville Syrjala
  2026-11-18 21:45 ` [PATCH v4 03/20] drm/tilcdc: Use __le64 for the fallback 64bit write Ville Syrjala
                   ` (17 subsequent siblings)
  19 siblings, 0 replies; 36+ messages in thread
From: Ville Syrjala @ 2026-11-18 21:44 UTC (permalink / raw)
  To: dri-devel; +Cc: Jyri Sarha, Tomi Valkeinen

From: Ville Syrjälä <ville.syrjala@linux.intel.com>

The address is passed to iowrite64() when the architecture provides
a native one. On some architectures (eg. alpha and parisc) that
takes a plain void __iomem *, so passing a volatile qualified
pointer discards the qualifier. This triggers -Wdiscarded-qualifiers
from gcc and a "different modifiers" warning from sparse. It is not
visible on ARM or arm64, where iowrite64() takes a volatile pointer.

The volatile on the pointer variable is not needed. The fallback
path already uses an explicit volatile cast for the actual access.

Cc: Jyri Sarha <jyri.sarha@iki.fi>
Cc: Tomi Valkeinen <tomi.valkeinen@ideasonboard.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
---
 drivers/gpu/drm/tilcdc/tilcdc_regs.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/tilcdc/tilcdc_regs.h b/drivers/gpu/drm/tilcdc/tilcdc_regs.h
index 55fe598fe770..23746b134c42 100644
--- a/drivers/gpu/drm/tilcdc/tilcdc_regs.h
+++ b/drivers/gpu/drm/tilcdc/tilcdc_regs.h
@@ -120,7 +120,7 @@ static inline void tilcdc_write(struct drm_device *dev, u32 reg, u32 data)
 static inline void tilcdc_write64(struct drm_device *dev, u32 reg, u64 data)
 {
 	struct tilcdc_drm_private *priv = ddev_to_tilcdc_priv(dev);
-	volatile void __iomem *addr = priv->mmio + reg;
+	void __iomem *addr = priv->mmio + reg;
 
 #if defined(iowrite64) && !defined(iowrite64_is_nonatomic)
 	iowrite64(data, addr);
-- 
2.56.0


^ permalink raw reply related	[flat|nested] 36+ messages in thread

* [PATCH v4 03/20] drm/tilcdc: Use __le64 for the fallback 64bit write
  2026-11-18 21:44 [PATCH v4 00/20] drm: Increase COMPILE_TEST=y coverage Ville Syrjala
  2026-11-18 21:44 ` [PATCH v4 01/20] drm/tilcdc: Allow build without __iowmb() Ville Syrjala
  2026-11-18 21:44 ` [PATCH v4 02/20] drm/tilcdc: Drop volatile from the mmio pointer in tilcdc_write64() Ville Syrjala
@ 2026-11-18 21:45 ` Ville Syrjala
  2026-11-18 21:45 ` [PATCH v4 04/20] drm/tilcdc: Fix endpoint node names in the legacy panel overlay Ville Syrjala
                   ` (16 subsequent siblings)
  19 siblings, 0 replies; 36+ messages in thread
From: Ville Syrjala @ 2026-11-18 21:45 UTC (permalink / raw)
  To: dri-devel; +Cc: Jyri Sarha, Tomi Valkeinen

From: Ville Syrjälä <ville.syrjala@linux.intel.com>

The fallback path in tilcdc_write64() stores the result of
__cpu_to_le64() through a volatile u64 pointer. That assigns a
__le64 to a plain u64 and makes sparse complain about the differing
base types. Use a volatile __le64 pointer for the store instead, so
the types match.

Cc: Jyri Sarha <jyri.sarha@iki.fi>
Cc: Tomi Valkeinen <tomi.valkeinen@ideasonboard.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
---
 drivers/gpu/drm/tilcdc/tilcdc_regs.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/tilcdc/tilcdc_regs.h b/drivers/gpu/drm/tilcdc/tilcdc_regs.h
index 23746b134c42..9dc3e1b9e20e 100644
--- a/drivers/gpu/drm/tilcdc/tilcdc_regs.h
+++ b/drivers/gpu/drm/tilcdc/tilcdc_regs.h
@@ -130,7 +130,7 @@ static inline void tilcdc_write64(struct drm_device *dev, u32 reg, u64 data)
 	__iowmb();
 #endif
 	/* This compiles to strd (=64-bit write) on ARM7 */
-	*(volatile u64 __force *)addr = __cpu_to_le64(data);
+	*(volatile __le64 __force *)addr = __cpu_to_le64(data);
 #endif
 }
 
-- 
2.56.0


^ permalink raw reply related	[flat|nested] 36+ messages in thread

* [PATCH v4 04/20] drm/tilcdc: Fix endpoint node names in the legacy panel overlay
  2026-11-18 21:44 [PATCH v4 00/20] drm: Increase COMPILE_TEST=y coverage Ville Syrjala
                   ` (2 preceding siblings ...)
  2026-11-18 21:45 ` [PATCH v4 03/20] drm/tilcdc: Use __le64 for the fallback 64bit write Ville Syrjala
@ 2026-11-18 21:45 ` Ville Syrjala
  2026-11-19  9:32   ` Kory Maincent
  2026-11-18 21:45 ` [PATCH v4 05/20] drm/tilcdc: Allow build with COMPILE_TEST=y Ville Syrjala
                   ` (15 subsequent siblings)
  19 siblings, 1 reply; 36+ messages in thread
From: Ville Syrjala @ 2026-11-18 21:45 UTC (permalink / raw)
  To: dri-devel; +Cc: Kory Maincent, Jyri Sarha, Tomi Valkeinen

From: Ville Syrjälä <ville.syrjala@linux.intel.com>

The overlay names its two graph endpoints "endpoint@0" but gives them
no reg property, so dtc complains with W=1:

  Warning (unit_address_vs_reg): .../endpoint@0: node has a unit name,
  but no reg or ranges property

There is only one endpoint per port, and the overlay refers to the
nodes through the panel_in and lcd_0 labels, not by name. Use plain
"endpoint" for the node names, which is the usual form for a single
endpoint, and drop the bogus unit address. The remote-endpoint
phandle links between the two nodes are unchanged.

Fixes: 0ff223d99147 ("drm/tilcdc: Convert legacy panel binding via DT overlay at boot time")
Cc: Kory Maincent <kory.maincent@bootlin.com>
Cc: Jyri Sarha <jyri.sarha@iki.fi>
Cc: Tomi Valkeinen <tomi.valkeinen@ideasonboard.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
---
Caveat! I really have no clue about DT, so this was pretty
much all Claude's idea how to fix the warning. I'd buy what
it's selling, but can't really judge it one way or the other.
Nor can I actually test this.

 drivers/gpu/drm/tilcdc/tilcdc_panel_legacy.dtso | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/gpu/drm/tilcdc/tilcdc_panel_legacy.dtso b/drivers/gpu/drm/tilcdc/tilcdc_panel_legacy.dtso
index ae71d10f5ec1..ca82f3bed15c 100644
--- a/drivers/gpu/drm/tilcdc/tilcdc_panel_legacy.dtso
+++ b/drivers/gpu/drm/tilcdc/tilcdc_panel_legacy.dtso
@@ -13,7 +13,7 @@
 	tilcdc-panel-dpi {
 		compatible = "panel-dpi";
 		port {
-			panel_in: endpoint@0 {
+			panel_in: endpoint {
 				remote-endpoint = <&lcd_0>;
 			};
 		};
@@ -22,7 +22,7 @@ panel_in: endpoint@0 {
 
 &lcdc {
 	port {
-		lcd_0: endpoint@0 {
+		lcd_0: endpoint {
 			remote-endpoint = <&panel_in>;
 		};
 	};
-- 
2.56.0


^ permalink raw reply related	[flat|nested] 36+ messages in thread

* [PATCH v4 05/20] drm/tilcdc: Allow build with COMPILE_TEST=y
  2026-11-18 21:44 [PATCH v4 00/20] drm: Increase COMPILE_TEST=y coverage Ville Syrjala
                   ` (3 preceding siblings ...)
  2026-11-18 21:45 ` [PATCH v4 04/20] drm/tilcdc: Fix endpoint node names in the legacy panel overlay Ville Syrjala
@ 2026-11-18 21:45 ` Ville Syrjala
  2026-11-18 21:45 ` [PATCH v4 06/20] drm/armada: Fix printk arguments Ville Syrjala
                   ` (14 subsequent siblings)
  19 siblings, 0 replies; 36+ messages in thread
From: Ville Syrjala @ 2026-11-18 21:45 UTC (permalink / raw)
  To: dri-devel; +Cc: Jyri Sarha, Tomi Valkeinen

From: Ville Syrjälä <ville.syrjala@linux.intel.com>

Allow tilcdc to be built with COMPILE_TEST=y for greater coverage.
Builds fine on x86, x86_64, arm and arm64 at least.

Acked-by: Jyri Sarha <jyri.sarha@iki.fi>
Reviewed-by: Tomi Valkeinen <tomi.valkeinen@ideasonboard.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
---
 drivers/gpu/drm/tilcdc/Kconfig | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/tilcdc/Kconfig b/drivers/gpu/drm/tilcdc/Kconfig
index 89df7528756c..dfe780475973 100644
--- a/drivers/gpu/drm/tilcdc/Kconfig
+++ b/drivers/gpu/drm/tilcdc/Kconfig
@@ -1,7 +1,7 @@
 # SPDX-License-Identifier: GPL-2.0-only
 config DRM_TILCDC
 	tristate "DRM Support for TI LCDC Display Controller"
-	depends on DRM && OF && ARM
+	depends on DRM && OF && (ARM || COMPILE_TEST)
 	select DRM_CLIENT_SELECTION
 	select DRM_KMS_HELPER
 	select DRM_GEM_DMA_HELPER
-- 
2.56.0


^ permalink raw reply related	[flat|nested] 36+ messages in thread

* [PATCH v4 06/20] drm/armada: Fix printk arguments
  2026-11-18 21:44 [PATCH v4 00/20] drm: Increase COMPILE_TEST=y coverage Ville Syrjala
                   ` (4 preceding siblings ...)
  2026-11-18 21:45 ` [PATCH v4 05/20] drm/tilcdc: Allow build with COMPILE_TEST=y Ville Syrjala
@ 2026-11-18 21:45 ` Ville Syrjala
  2026-11-18 21:45 ` [PATCH v4 07/20] drm/armada: Fix armada_debugfs_crtc_reg_write() return type Ville Syrjala
                   ` (13 subsequent siblings)
  19 siblings, 0 replies; 36+ messages in thread
From: Ville Syrjala @ 2026-11-18 21:45 UTC (permalink / raw)
  To: dri-devel; +Cc: Russell King

From: Ville Syrjälä <ville.syrjala@linux.intel.com>

../drivers/gpu/drm/armada/armada_gem.c: In function ‘armada_gem_pwrite_ioctl’:
../drivers/gpu/drm/armada/armada_gem.c:367:27: warning: format ‘%u’ expects argument of type ‘unsigned int’, but argument 2 has type ‘size_t’ {aka ‘long unsigned int’} [-Wformat=]
  367 |                 DRM_ERROR("invalid size: object size %u\n", dobj->obj.size);
      |                           ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~  ~~~~~~~~~~~~~~
      |                                                                      |
      |                                                                      size_t {aka long unsigned int}

Cc: Russell King <linux@armlinux.org.uk>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
---
 drivers/gpu/drm/armada/armada_gem.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/armada/armada_gem.c b/drivers/gpu/drm/armada/armada_gem.c
index da7335cbe82d..8a8f4da6f44f 100644
--- a/drivers/gpu/drm/armada/armada_gem.c
+++ b/drivers/gpu/drm/armada/armada_gem.c
@@ -365,7 +365,7 @@ int armada_gem_pwrite_ioctl(struct drm_device *dev, void *data,
 
 	if (args->offset > dobj->obj.size ||
 	    args->size > dobj->obj.size - args->offset) {
-		DRM_ERROR("invalid size: object size %u\n", dobj->obj.size);
+		DRM_ERROR("invalid size: object size %zu\n", dobj->obj.size);
 		ret = -EINVAL;
 		goto unref;
 	}
-- 
2.56.0


^ permalink raw reply related	[flat|nested] 36+ messages in thread

* [PATCH v4 07/20] drm/armada: Fix armada_debugfs_crtc_reg_write() return type
  2026-11-18 21:44 [PATCH v4 00/20] drm: Increase COMPILE_TEST=y coverage Ville Syrjala
                   ` (5 preceding siblings ...)
  2026-11-18 21:45 ` [PATCH v4 06/20] drm/armada: Fix printk arguments Ville Syrjala
@ 2026-11-18 21:45 ` Ville Syrjala
  2026-11-18 21:45 ` [PATCH v4 08/20] drm/armada: Fix sparse __iomem address space warnings Ville Syrjala
                   ` (12 subsequent siblings)
  19 siblings, 0 replies; 36+ messages in thread
From: Ville Syrjala @ 2026-11-18 21:45 UTC (permalink / raw)
  To: dri-devel; +Cc: Russell King

From: Ville Syrjälä <ville.syrjala@linux.intel.com>

Change the armada_debugfs_crtc_reg_write() return type to
the correct ssize_t. This makes the code actually build on
certain architectures.

Cc: Russell King <linux@armlinux.org.uk>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
---
 drivers/gpu/drm/armada/armada_debugfs.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/armada/armada_debugfs.c b/drivers/gpu/drm/armada/armada_debugfs.c
index 2445365c823f..5d2af08b9397 100644
--- a/drivers/gpu/drm/armada/armada_debugfs.c
+++ b/drivers/gpu/drm/armada/armada_debugfs.c
@@ -50,7 +50,7 @@ static int armada_debugfs_crtc_reg_open(struct inode *inode, struct file *file)
 			   inode->i_private);
 }
 
-static int armada_debugfs_crtc_reg_write(struct file *file,
+static ssize_t armada_debugfs_crtc_reg_write(struct file *file,
 	const char __user *ptr, size_t len, loff_t *off)
 {
 	struct armada_crtc *dcrtc;
-- 
2.56.0


^ permalink raw reply related	[flat|nested] 36+ messages in thread

* [PATCH v4 08/20] drm/armada: Fix sparse __iomem address space warnings
  2026-11-18 21:44 [PATCH v4 00/20] drm: Increase COMPILE_TEST=y coverage Ville Syrjala
                   ` (6 preceding siblings ...)
  2026-11-18 21:45 ` [PATCH v4 07/20] drm/armada: Fix armada_debugfs_crtc_reg_write() return type Ville Syrjala
@ 2026-11-18 21:45 ` Ville Syrjala
  2026-11-18 21:45 ` [PATCH v4 09/20] drm/armada: Allow build with COMPILE_TEST=y Ville Syrjala
                   ` (11 subsequent siblings)
  19 siblings, 0 replies; 36+ messages in thread
From: Ville Syrjala @ 2026-11-18 21:45 UTC (permalink / raw)
  To: dri-devel; +Cc: Russell King

From: Ville Syrjälä <ville.syrjala@linux.intel.com>

The gem object keeps a single void *addr that holds either a
page_address() pointer for page backed objects, or an ioremap_wc()
mapping for linear ones. Sparse complains about the address space
mismatches when the ioremap_wc() result is stored in it, when it is
passed to iounmap(), and when it is assigned to screen_base.

Splitting the pointer in two would be a lot of churn for little gain,
so just cast around the __iomem annotation at those three places.
The casts are consistent with how the pointer is used: iounmap() is
only called for linear objects, and the fbdev framebuffer is always
a linear object mapped through armada_gem_map_object().

No functional change.

Cc: Russell King <linux@armlinux.org.uk>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
---
 drivers/gpu/drm/armada/armada_fbdev.c | 2 +-
 drivers/gpu/drm/armada/armada_gem.c   | 5 +++--
 2 files changed, 4 insertions(+), 3 deletions(-)

diff --git a/drivers/gpu/drm/armada/armada_fbdev.c b/drivers/gpu/drm/armada/armada_fbdev.c
index f35afec2125f..073d7883d988 100644
--- a/drivers/gpu/drm/armada/armada_fbdev.c
+++ b/drivers/gpu/drm/armada/armada_fbdev.c
@@ -95,7 +95,7 @@ int armada_fbdev_driver_fbdev_probe(struct drm_fb_helper *fbh,
 	info->fix.smem_start = obj->phys_addr;
 	info->fix.smem_len = obj->obj.size;
 	info->screen_size = obj->obj.size;
-	info->screen_base = ptr;
+	info->screen_base = (void __iomem __force *)ptr;
 	fbh->funcs = &armada_fbdev_helper_funcs;
 	fbh->fb = &dfb->fb;
 
diff --git a/drivers/gpu/drm/armada/armada_gem.c b/drivers/gpu/drm/armada/armada_gem.c
index 8a8f4da6f44f..3503f633a8b2 100644
--- a/drivers/gpu/drm/armada/armada_gem.c
+++ b/drivers/gpu/drm/armada/armada_gem.c
@@ -61,7 +61,7 @@ void armada_gem_free_object(struct drm_gem_object *obj)
 		mutex_unlock(&priv->linear_lock);
 		kfree(dobj->linear);
 		if (dobj->addr)
-			iounmap(dobj->addr);
+			iounmap((void __iomem __force *)dobj->addr);
 	}
 
 	if (dobj->obj.import_attach) {
@@ -183,7 +183,8 @@ armada_gem_map_object(struct drm_device *dev, struct armada_gem_object *dobj)
 {
 	/* only linear objects need to be ioremap'd */
 	if (!dobj->addr && dobj->linear)
-		dobj->addr = ioremap_wc(dobj->phys_addr, dobj->obj.size);
+		dobj->addr = (void __force *)ioremap_wc(dobj->phys_addr,
+							dobj->obj.size);
 	return dobj->addr;
 }
 
-- 
2.56.0


^ permalink raw reply related	[flat|nested] 36+ messages in thread

* [PATCH v4 09/20] drm/armada: Allow build with COMPILE_TEST=y
  2026-11-18 21:44 [PATCH v4 00/20] drm: Increase COMPILE_TEST=y coverage Ville Syrjala
                   ` (7 preceding siblings ...)
  2026-11-18 21:45 ` [PATCH v4 08/20] drm/armada: Fix sparse __iomem address space warnings Ville Syrjala
@ 2026-11-18 21:45 ` Ville Syrjala
  2026-11-18 21:45 ` [PATCH v4 10/20] drm/fsl-dcu: " Ville Syrjala
                   ` (10 subsequent siblings)
  19 siblings, 0 replies; 36+ messages in thread
From: Ville Syrjala @ 2026-11-18 21:45 UTC (permalink / raw)
  To: dri-devel; +Cc: Russell King

From: Ville Syrjälä <ville.syrjala@linux.intel.com>

Allow armada to be built with COMPILE_TEST=y for greater coverage.
Builds fine on x86, x86_64, arm and arm64 at least.

Cc: Russell King <linux@armlinux.org.uk>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
---
 drivers/gpu/drm/armada/Kconfig | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/armada/Kconfig b/drivers/gpu/drm/armada/Kconfig
index b22c891a670b..1376337548da 100644
--- a/drivers/gpu/drm/armada/Kconfig
+++ b/drivers/gpu/drm/armada/Kconfig
@@ -1,7 +1,7 @@
 # SPDX-License-Identifier: GPL-2.0-only
 config DRM_ARMADA
 	tristate "DRM support for Marvell Armada SoCs"
-	depends on DRM && HAVE_CLK && ARM && MMU
+	depends on DRM && HAVE_CLK && MMU && (ARM || COMPILE_TEST)
 	select DRM_CLIENT_SELECTION
 	select DRM_KMS_HELPER
 	select FB_IOMEM_HELPERS if DRM_FBDEV_EMULATION
-- 
2.56.0


^ permalink raw reply related	[flat|nested] 36+ messages in thread

* [PATCH v4 10/20] drm/fsl-dcu: Allow build with COMPILE_TEST=y
  2026-11-18 21:44 [PATCH v4 00/20] drm: Increase COMPILE_TEST=y coverage Ville Syrjala
                   ` (8 preceding siblings ...)
  2026-11-18 21:45 ` [PATCH v4 09/20] drm/armada: Allow build with COMPILE_TEST=y Ville Syrjala
@ 2026-11-18 21:45 ` Ville Syrjala
  2026-11-18 21:45 ` [PATCH v4 11/20] drm/adp: Include linux/bitfield.h Ville Syrjala
                   ` (9 subsequent siblings)
  19 siblings, 0 replies; 36+ messages in thread
From: Ville Syrjala @ 2026-11-18 21:45 UTC (permalink / raw)
  To: dri-devel; +Cc: Stefan Agner, Alison Wang

From: Ville Syrjälä <ville.syrjala@linux.intel.com>

Allow fsl-dcu to be built with COMPILE_TEST=y for greater coverage.
Builds fine on x86, x86_64, arm and arm64 at least.

Cc: Stefan Agner <stefan@agner.ch>
Cc: Alison Wang <alison.wang@nxp.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
---
 drivers/gpu/drm/fsl-dcu/Kconfig | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/fsl-dcu/Kconfig b/drivers/gpu/drm/fsl-dcu/Kconfig
index 0e0f910ceb9f..ef7fd7fc9fd6 100644
--- a/drivers/gpu/drm/fsl-dcu/Kconfig
+++ b/drivers/gpu/drm/fsl-dcu/Kconfig
@@ -1,7 +1,7 @@
 # SPDX-License-Identifier: GPL-2.0-only
 config DRM_FSL_DCU
 	tristate "DRM Support for Freescale DCU"
-	depends on DRM && OF && ARM && COMMON_CLK
+	depends on DRM && OF && COMMON_CLK && (ARM || COMPILE_TEST)
 	select BACKLIGHT_CLASS_DEVICE
 	select DRM_CLIENT_SELECTION
 	select DRM_GEM_DMA_HELPER
-- 
2.56.0


^ permalink raw reply related	[flat|nested] 36+ messages in thread

* [PATCH v4 11/20] drm/adp: Include linux/bitfield.h
  2026-11-18 21:44 [PATCH v4 00/20] drm: Increase COMPILE_TEST=y coverage Ville Syrjala
                   ` (9 preceding siblings ...)
  2026-11-18 21:45 ` [PATCH v4 10/20] drm/fsl-dcu: " Ville Syrjala
@ 2026-11-18 21:45 ` Ville Syrjala
  2026-11-18 21:45 ` [PATCH v4 12/20] drm/adp: Allow build with COMPILE_TEST=y Ville Syrjala
                   ` (8 subsequent siblings)
  19 siblings, 0 replies; 36+ messages in thread
From: Ville Syrjala @ 2026-11-18 21:45 UTC (permalink / raw)
  To: dri-devel; +Cc: Sasha Finkelstein, Janne Grunau, asahi

From: Ville Syrjälä <ville.syrjala@linux.intel.com>

adp_drv.c uses FIELD_GET() but doesn't include linux/bitfield.h,
relying on an indirect include that isn't present on all
configurations. Include it explicitly so the driver can be built
with COMPILE_TEST=y on other architectures.

Cc: Sasha Finkelstein <k@chaosmail.tech>
Cc: Janne Grunau <j@jannau.net>
Cc: asahi@lists.linux.dev
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
---
 drivers/gpu/drm/adp/adp_drv.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/drivers/gpu/drm/adp/adp_drv.c b/drivers/gpu/drm/adp/adp_drv.c
index fdf0e6316175..21257612f5d0 100644
--- a/drivers/gpu/drm/adp/adp_drv.c
+++ b/drivers/gpu/drm/adp/adp_drv.c
@@ -1,5 +1,6 @@
 // SPDX-License-Identifier: GPL-2.0-only
 
+#include <linux/bitfield.h>
 #include <linux/component.h>
 #include <linux/iopoll.h>
 #include <linux/of.h>
-- 
2.56.0


^ permalink raw reply related	[flat|nested] 36+ messages in thread

* [PATCH v4 12/20] drm/adp: Allow build with COMPILE_TEST=y
  2026-11-18 21:44 [PATCH v4 00/20] drm: Increase COMPILE_TEST=y coverage Ville Syrjala
                   ` (10 preceding siblings ...)
  2026-11-18 21:45 ` [PATCH v4 11/20] drm/adp: Include linux/bitfield.h Ville Syrjala
@ 2026-11-18 21:45 ` Ville Syrjala
  2026-11-18 21:45 ` [PATCH v4 13/20] drm/imagination: Include io-64-nonatomic-lo-hi.h Ville Syrjala
                   ` (7 subsequent siblings)
  19 siblings, 0 replies; 36+ messages in thread
From: Ville Syrjala @ 2026-11-18 21:45 UTC (permalink / raw)
  To: dri-devel; +Cc: Sasha Finkelstein, Janne Grunau, asahi

From: Ville Syrjälä <ville.syrjala@linux.intel.com>

Allow adp to be built with COMPILE_TEST=y for greater coverage. Builds
fine on x86, x86_64, arm and arm64 at least.

Cc: Sasha Finkelstein <k@chaosmail.tech>
Cc: Janne Grunau <j@jannau.net>
Cc: asahi@lists.linux.dev
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
---
 drivers/gpu/drm/adp/Kconfig | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/adp/Kconfig b/drivers/gpu/drm/adp/Kconfig
index acfa21ee06d2..6073aa6b46a4 100644
--- a/drivers/gpu/drm/adp/Kconfig
+++ b/drivers/gpu/drm/adp/Kconfig
@@ -1,7 +1,7 @@
 # SPDX-License-Identifier: GPL-2.0-only OR MIT
 config DRM_ADP
 	tristate "DRM Support for pre-DCP Apple display controllers"
-	depends on DRM && OF && ARM64
+	depends on DRM && OF && (ARM64 || COMPILE_TEST)
 	depends on ARCH_APPLE || COMPILE_TEST
 	select DRM_KMS_HELPER
 	select DRM_BRIDGE_CONNECTOR
-- 
2.56.0


^ permalink raw reply related	[flat|nested] 36+ messages in thread

* [PATCH v4 13/20] drm/imagination: Include io-64-nonatomic-lo-hi.h
  2026-11-18 21:44 [PATCH v4 00/20] drm: Increase COMPILE_TEST=y coverage Ville Syrjala
                   ` (11 preceding siblings ...)
  2026-11-18 21:45 ` [PATCH v4 12/20] drm/adp: Allow build with COMPILE_TEST=y Ville Syrjala
@ 2026-11-18 21:45 ` Ville Syrjala
  2026-11-18 21:45 ` [PATCH v4 14/20] drm/imagination: Use rcu_access_pointer() for dma_fence::ops Ville Syrjala
                   ` (6 subsequent siblings)
  19 siblings, 0 replies; 36+ messages in thread
From: Ville Syrjala @ 2026-11-18 21:45 UTC (permalink / raw)
  To: dri-devel; +Cc: Alessio Belle, Luigi Santivetti, imagination

From: Ville Syrjälä <ville.syrjala@linux.intel.com>

ioread64() and iowrite64() are provided by asm-generic/io.h on 64bit
architectures that use it, such as arm64, but not on all architectures.
Include linux/io-64-nonatomic-lo-hi.h, which only supplies them when the
architecture doesn't, so the driver can be built with COMPILE_TEST=y
elsewhere. There is no functional change on architectures that provide
native 64-bit accessors.

Cc: Alessio Belle <alessio.belle@imgtec.com>
Cc: Luigi Santivetti <luigi.santivetti@imgtec.com>
Cc: imagination@lists.freedesktop.org
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
---
 drivers/gpu/drm/imagination/pvr_device.h | 1 +
 1 file changed, 1 insertion(+)

diff --git a/drivers/gpu/drm/imagination/pvr_device.h b/drivers/gpu/drm/imagination/pvr_device.h
index c246122b4caf..975b4e671418 100644
--- a/drivers/gpu/drm/imagination/pvr_device.h
+++ b/drivers/gpu/drm/imagination/pvr_device.h
@@ -19,6 +19,7 @@
 #include <linux/compiler_types.h>
 #include <linux/device.h>
 #include <linux/io.h>
+#include <linux/io-64-nonatomic-lo-hi.h>
 #include <linux/iopoll.h>
 #include <linux/kernel.h>
 #include <linux/math.h>
-- 
2.56.0


^ permalink raw reply related	[flat|nested] 36+ messages in thread

* [PATCH v4 14/20] drm/imagination: Use rcu_access_pointer() for dma_fence::ops
  2026-11-18 21:44 [PATCH v4 00/20] drm: Increase COMPILE_TEST=y coverage Ville Syrjala
                   ` (12 preceding siblings ...)
  2026-11-18 21:45 ` [PATCH v4 13/20] drm/imagination: Include io-64-nonatomic-lo-hi.h Ville Syrjala
@ 2026-11-18 21:45 ` Ville Syrjala
  2026-11-18 21:45 ` [PATCH v4 15/20] drm/imagination: Allow build with COMPILE_TEST=y Ville Syrjala
                   ` (5 subsequent siblings)
  19 siblings, 0 replies; 36+ messages in thread
From: Ville Syrjala @ 2026-11-18 21:45 UTC (permalink / raw)
  To: dri-devel; +Cc: Alessio Belle, Luigi Santivetti, imagination

From: Ville Syrjälä <ville.syrjala@linux.intel.com>

dma_fence::ops is annotated __rcu, but the driver compares and
tests it directly. That makes sparse complain about comparing
pointers from different address spaces. The driver only uses the
pointer to identify the kind of fence and never dereferences it,
so use rcu_access_pointer() for those checks, as dma-fence.h does
for its own identity checks.

No functional change.

Cc: Alessio Belle <alessio.belle@imgtec.com>
Cc: Luigi Santivetti <luigi.santivetti@imgtec.com>
Cc: imagination@lists.freedesktop.org
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
---
 drivers/gpu/drm/imagination/pvr_ccb.c   |  5 +++--
 drivers/gpu/drm/imagination/pvr_queue.c | 17 +++++++++--------
 2 files changed, 12 insertions(+), 10 deletions(-)

diff --git a/drivers/gpu/drm/imagination/pvr_ccb.c b/drivers/gpu/drm/imagination/pvr_ccb.c
index e408df74853d..92f041b4208a 100644
--- a/drivers/gpu/drm/imagination/pvr_ccb.c
+++ b/drivers/gpu/drm/imagination/pvr_ccb.c
@@ -18,6 +18,7 @@
 #include <linux/kernel.h>
 #include <linux/mutex.h>
 #include <linux/overflow.h>
+#include <linux/rcupdate.h>
 #include <linux/types.h>
 #include <linux/workqueue.h>
 
@@ -608,10 +609,10 @@ void pvr_kccb_fence_put(struct dma_fence *fence)
 	if (!fence)
 		return;
 
-	if (!fence->ops) {
+	if (!rcu_access_pointer(fence->ops)) {
 		dma_fence_free(fence);
 	} else {
-		WARN_ON(fence->ops != &pvr_kccb_fence_ops);
+		WARN_ON(rcu_access_pointer(fence->ops) != &pvr_kccb_fence_ops);
 		dma_fence_put(fence);
 	}
 }
diff --git a/drivers/gpu/drm/imagination/pvr_queue.c b/drivers/gpu/drm/imagination/pvr_queue.c
index 24a8645aa913..f86e39fcb57f 100644
--- a/drivers/gpu/drm/imagination/pvr_queue.c
+++ b/drivers/gpu/drm/imagination/pvr_queue.c
@@ -4,6 +4,7 @@
 #include <drm/drm_managed.h>
 #include <drm/gpu_scheduler.h>
 #include <linux/overflow.h>
+#include <linux/rcupdate.h>
 
 #include "pvr_cccb.h"
 #include "pvr_context.h"
@@ -193,7 +194,7 @@ pvr_queue_fence_is_ufo_backed(struct dma_fence *f)
 	 * Currently the only dma_fence backed by a UFO object is the job fence,
 	 * e.g. pvr_job::done_fence, wrapped by a pvr_queue_fence object.
 	 */
-	return f && f->ops == &pvr_queue_job_fence_ops;
+	return f && rcu_access_pointer(f->ops) == &pvr_queue_job_fence_ops;
 }
 
 /**
@@ -246,13 +247,13 @@ static void pvr_queue_fence_put(struct dma_fence *f)
 	if (!f)
 		return;
 
-	if (WARN_ON(f->ops &&
-		    f->ops != &pvr_queue_cccb_fence_ops &&
-		    f->ops != &pvr_queue_job_fence_ops))
+	if (WARN_ON(rcu_access_pointer(f->ops) &&
+		    rcu_access_pointer(f->ops) != &pvr_queue_cccb_fence_ops &&
+		    rcu_access_pointer(f->ops) != &pvr_queue_job_fence_ops))
 		return;
 
 	/* If the fence hasn't been initialized yet, free the object directly. */
-	if (f->ops)
+	if (rcu_access_pointer(f->ops))
 		dma_fence_put(f);
 	else
 		dma_fence_free(f);
@@ -339,7 +340,7 @@ pvr_queue_cccb_fence_init(struct dma_fence *fence, struct pvr_queue *queue)
 static void
 pvr_queue_job_fence_init(struct dma_fence *fence, struct pvr_queue *queue)
 {
-	if (!fence->ops)
+	if (!rcu_access_pointer(fence->ops))
 		pvr_queue_fence_init(fence, queue, &pvr_queue_job_fence_ops,
 				     &queue->job_fence_ctx);
 }
@@ -508,7 +509,7 @@ pvr_queue_get_job_kccb_fence(struct pvr_queue *queue, struct pvr_job *job)
 	if (!job->kccb_fence)
 		return NULL;
 
-	if (!WARN_ON(job->kccb_fence->ops)) {
+	if (!WARN_ON(rcu_access_pointer(job->kccb_fence->ops))) {
 		kccb_fence = pvr_kccb_reserve_slot(pvr_dev, job->kccb_fence);
 		job->kccb_fence = NULL;
 	}
@@ -755,7 +756,7 @@ static struct dma_fence *pvr_queue_run_job(struct drm_sched_job *sched_job)
 	 * done_fence that's been initialized earlier.
 	 */
 	if (job->paired_job && job->type == DRM_PVR_JOB_TYPE_FRAGMENT &&
-	    job->done_fence->ops) {
+	    rcu_access_pointer(job->done_fence->ops)) {
 		return dma_fence_get(job->done_fence);
 	}
 
-- 
2.56.0


^ permalink raw reply related	[flat|nested] 36+ messages in thread

* [PATCH v4 15/20] drm/imagination: Allow build with COMPILE_TEST=y
  2026-11-18 21:44 [PATCH v4 00/20] drm: Increase COMPILE_TEST=y coverage Ville Syrjala
                   ` (13 preceding siblings ...)
  2026-11-18 21:45 ` [PATCH v4 14/20] drm/imagination: Use rcu_access_pointer() for dma_fence::ops Ville Syrjala
@ 2026-11-18 21:45 ` Ville Syrjala
  2026-11-18 21:58   ` sashiko-bot
  2026-11-18 21:45 ` [PATCH v4 16/20] drm/imagination: Drop stale kernel-doc for pvr_mmu_backing_page.pvr_dev Ville Syrjala
                   ` (4 subsequent siblings)
  19 siblings, 1 reply; 36+ messages in thread
From: Ville Syrjala @ 2026-11-18 21:45 UTC (permalink / raw)
  To: dri-devel; +Cc: Alessio Belle, Luigi Santivetti, imagination

From: Ville Syrjälä <ville.syrjala@linux.intel.com>

Allow powervr to be built with COMPILE_TEST=y for greater
coverage. The driver only supports 64bit, so keep requiring
64BIT explicitly; previously that was implied by the
(ARM64 || RISCV && 64BIT) dependency. Builds fine on
x86_64 and arm64 at least.

Cc: Alessio Belle <alessio.belle@imgtec.com>
Cc: Luigi Santivetti <luigi.santivetti@imgtec.com>
Cc: imagination@lists.freedesktop.org
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
---
 drivers/gpu/drm/imagination/Kconfig | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/imagination/Kconfig b/drivers/gpu/drm/imagination/Kconfig
index 1fd4c635c2c9..a048d3c85938 100644
--- a/drivers/gpu/drm/imagination/Kconfig
+++ b/drivers/gpu/drm/imagination/Kconfig
@@ -3,7 +3,8 @@
 
 config DRM_POWERVR
 	tristate "Imagination Technologies PowerVR (Series 6 and later) & IMG Graphics"
-	depends on (ARM64 || RISCV && 64BIT)
+	depends on ARM64 || RISCV || COMPILE_TEST
+	depends on 64BIT
 	depends on DRM
 	depends on MMU
 	depends on PM
-- 
2.56.0


^ permalink raw reply related	[flat|nested] 36+ messages in thread

* [PATCH v4 16/20] drm/imagination: Drop stale kernel-doc for pvr_mmu_backing_page.pvr_dev
  2026-11-18 21:44 [PATCH v4 00/20] drm: Increase COMPILE_TEST=y coverage Ville Syrjala
                   ` (14 preceding siblings ...)
  2026-11-18 21:45 ` [PATCH v4 15/20] drm/imagination: Allow build with COMPILE_TEST=y Ville Syrjala
@ 2026-11-18 21:45 ` Ville Syrjala
  2026-11-18 21:45 ` [PATCH v4 17/20] drm/imagination: Fix kernel-doc for pvr_mmu_op_context anonymous structs Ville Syrjala
                   ` (3 subsequent siblings)
  19 siblings, 0 replies; 36+ messages in thread
From: Ville Syrjala @ 2026-11-18 21:45 UTC (permalink / raw)
  To: dri-devel; +Cc: Alessio Belle, Luigi Santivetti, imagination

From: Ville Syrjälä <ville.syrjala@linux.intel.com>

W=1 builds warn about the documented @pvr_dev member of struct
pvr_mmu_backing_page, which is below the "private:" marker and so is
not part of the documented interface. Drop the stale entry.

Cc: Alessio Belle <alessio.belle@imgtec.com>
Cc: Luigi Santivetti <luigi.santivetti@imgtec.com>
Cc: imagination@lists.freedesktop.org
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
---
 drivers/gpu/drm/imagination/pvr_mmu.c | 2 --
 1 file changed, 2 deletions(-)

diff --git a/drivers/gpu/drm/imagination/pvr_mmu.c b/drivers/gpu/drm/imagination/pvr_mmu.c
index d1442515553b..494436e673bf 100644
--- a/drivers/gpu/drm/imagination/pvr_mmu.c
+++ b/drivers/gpu/drm/imagination/pvr_mmu.c
@@ -219,8 +219,6 @@ static_assert(PAGE_SIZE >= PVR_MMU_BACKING_PAGE_SIZE);
  *                              table of any level.
  * @dma_addr: DMA address of this page.
  * @host_ptr: CPU address of this page.
- * @pvr_dev: The PowerVR device to which this page is associated. **For
- *           internal use only.**
  */
 struct pvr_mmu_backing_page {
 	dma_addr_t dma_addr;
-- 
2.56.0


^ permalink raw reply related	[flat|nested] 36+ messages in thread

* [PATCH v4 17/20] drm/imagination: Fix kernel-doc for pvr_mmu_op_context anonymous structs
  2026-11-18 21:44 [PATCH v4 00/20] drm: Increase COMPILE_TEST=y coverage Ville Syrjala
                   ` (15 preceding siblings ...)
  2026-11-18 21:45 ` [PATCH v4 16/20] drm/imagination: Drop stale kernel-doc for pvr_mmu_backing_page.pvr_dev Ville Syrjala
@ 2026-11-18 21:45 ` Ville Syrjala
  2026-11-18 21:45 ` [PATCH v4 18/20] drm/ingenic: Allow ingenic-dw-hdmi build with COMPILE_TEST=y Ville Syrjala
                   ` (2 subsequent siblings)
  19 siblings, 0 replies; 36+ messages in thread
From: Ville Syrjala @ 2026-11-18 21:45 UTC (permalink / raw)
  To: dri-devel; +Cc: Alessio Belle, Luigi Santivetti, imagination

From: Ville Syrjälä <ville.syrjala@linux.intel.com>

W=1 builds warn because struct pvr_mmu_op_context documents the
members of its anonymous map and unmap structs without the struct
name prefix. Use @map.<member> and @unmap.<member> so kernel-doc can
match them.

Cc: Alessio Belle <alessio.belle@imgtec.com>
Cc: Luigi Santivetti <luigi.santivetti@imgtec.com>
Cc: imagination@lists.freedesktop.org
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
---
 drivers/gpu/drm/imagination/pvr_mmu.c | 12 ++++++------
 1 file changed, 6 insertions(+), 6 deletions(-)

diff --git a/drivers/gpu/drm/imagination/pvr_mmu.c b/drivers/gpu/drm/imagination/pvr_mmu.c
index 494436e673bf..c2e1bffa014c 100644
--- a/drivers/gpu/drm/imagination/pvr_mmu.c
+++ b/drivers/gpu/drm/imagination/pvr_mmu.c
@@ -1388,24 +1388,24 @@ struct pvr_mmu_op_context {
 	/** @map: Data specifically for map operations. */
 	struct {
 		/**
-		 * @sgt: Scatter gather table containing pages pinned for use by
+		 * @map.sgt: Scatter gather table containing pages pinned for use by
 		 * this context - these are currently pinned when initialising
 		 * the VM bind operation.
 		 */
 		struct sg_table *sgt;
 
-		/** @sgt_offset: Offset into @sgt of the start of the mapping. */
+		/** @map.sgt_offset: Offset into @map.sgt of the start of the mapping. */
 		u64 sgt_offset;
 
 		/**
-		 * @l1_prealloc_tables: Preallocated l1 page table objects
+		 * @map.l1_prealloc_tables: Preallocated l1 page table objects
 		 * use by this context when creating a page mapping. Linked list
 		 * fully created during initialisation.
 		 */
 		struct pvr_page_table_l1 *l1_prealloc_tables;
 
 		/**
-		 * @l0_prealloc_tables: Preallocated l0 page table objects
+		 * @map.l0_prealloc_tables: Preallocated l0 page table objects
 		 * use by this context when creating a page mapping. Linked list
 		 * fully created during initialisation.
 		 */
@@ -1415,13 +1415,13 @@ struct pvr_mmu_op_context {
 	/** @unmap: Data specifically for unmap operations. */
 	struct {
 		/**
-		 * @l1_free_tables: Collects page table objects freed by unmap
+		 * @unmap.l1_free_tables: Collects page table objects freed by unmap
 		 * ops. Linked list empty at creation.
 		 */
 		struct pvr_page_table_l1 *l1_free_tables;
 
 		/**
-		 * @l0_free_tables: Collects page table objects freed by unmap
+		 * @unmap.l0_free_tables: Collects page table objects freed by unmap
 		 * ops. Linked list empty at creation.
 		 */
 		struct pvr_page_table_l0 *l0_free_tables;
-- 
2.56.0


^ permalink raw reply related	[flat|nested] 36+ messages in thread

* [PATCH v4 18/20] drm/ingenic: Allow ingenic-dw-hdmi build with COMPILE_TEST=y
  2026-11-18 21:44 [PATCH v4 00/20] drm: Increase COMPILE_TEST=y coverage Ville Syrjala
                   ` (16 preceding siblings ...)
  2026-11-18 21:45 ` [PATCH v4 17/20] drm/imagination: Fix kernel-doc for pvr_mmu_op_context anonymous structs Ville Syrjala
@ 2026-11-18 21:45 ` Ville Syrjala
  2026-11-19  9:31   ` Paul Cercueil
  2026-11-18 21:45 ` [PATCH v4 19/20] drm/nouveau: Allow platform driver " Ville Syrjala
  2026-11-18 21:45 ` [PATCH v4 20/20] drm/rcar-du: Allow VSP and writeback " Ville Syrjala
  19 siblings, 1 reply; 36+ messages in thread
From: Ville Syrjala @ 2026-11-18 21:45 UTC (permalink / raw)
  To: dri-devel; +Cc: Paul Cercueil, linux-mips

From: Ville Syrjälä <ville.syrjala@linux.intel.com>

Allow DRM_INGENIC_DW_HDMI to be built with COMPILE_TEST=y for greater
coverage. The rest of the ingenic driver already allows this. Builds
fine on x86, x86_64, arm and arm64 at least.

Cc: Paul Cercueil <paul@crapouillou.net>
Cc: linux-mips@vger.kernel.org
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
---
 drivers/gpu/drm/ingenic/Kconfig | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/ingenic/Kconfig b/drivers/gpu/drm/ingenic/Kconfig
index 987ed42f6f7a..60bff9ebee5d 100644
--- a/drivers/gpu/drm/ingenic/Kconfig
+++ b/drivers/gpu/drm/ingenic/Kconfig
@@ -31,7 +31,7 @@ config DRM_INGENIC_IPU
 
 config DRM_INGENIC_DW_HDMI
 	tristate "Ingenic specific support for Synopsys DW HDMI"
-	depends on MACH_JZ4780
+	depends on MACH_JZ4780 || COMPILE_TEST
 	select DRM_DW_HDMI
 	help
 	  Choose this option to enable Synopsys DesignWare HDMI based driver.
-- 
2.56.0


^ permalink raw reply related	[flat|nested] 36+ messages in thread

* [PATCH v4 19/20] drm/nouveau: Allow platform driver build with COMPILE_TEST=y
  2026-11-18 21:44 [PATCH v4 00/20] drm: Increase COMPILE_TEST=y coverage Ville Syrjala
                   ` (17 preceding siblings ...)
  2026-11-18 21:45 ` [PATCH v4 18/20] drm/ingenic: Allow ingenic-dw-hdmi build with COMPILE_TEST=y Ville Syrjala
@ 2026-11-18 21:45 ` Ville Syrjala
  2026-11-18 21:57   ` sashiko-bot
  2026-11-18 21:45 ` [PATCH v4 20/20] drm/rcar-du: Allow VSP and writeback " Ville Syrjala
  19 siblings, 1 reply; 36+ messages in thread
From: Ville Syrjala @ 2026-11-18 21:45 UTC (permalink / raw)
  To: dri-devel; +Cc: Lyude Paul, Danilo Krummrich, nouveau

From: Ville Syrjälä <ville.syrjala@linux.intel.com>

Allow NOUVEAU_PLATFORM_DRIVER to be built with COMPILE_TEST=y for
greater coverage. Builds fine on x86, x86_64, arm and arm64 at least.

Cc: Lyude Paul <lyude@redhat.com>
Cc: Danilo Krummrich <dakr@kernel.org>
Cc: nouveau@lists.freedesktop.org
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
---
 drivers/gpu/drm/nouveau/Kconfig | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/nouveau/Kconfig b/drivers/gpu/drm/nouveau/Kconfig
index 3b5757aed9c8..d7b432c839e5 100644
--- a/drivers/gpu/drm/nouveau/Kconfig
+++ b/drivers/gpu/drm/nouveau/Kconfig
@@ -34,7 +34,7 @@ config DRM_NOUVEAU
 
 config NOUVEAU_PLATFORM_DRIVER
 	bool "Nouveau (NVIDIA) SoC GPUs"
-	depends on DRM_NOUVEAU && ARCH_TEGRA
+	depends on DRM_NOUVEAU && (ARCH_TEGRA || COMPILE_TEST)
 	default y
 	help
 	  Support for Nouveau platform driver, used for SoC GPUs as found
-- 
2.56.0


^ permalink raw reply related	[flat|nested] 36+ messages in thread

* [PATCH v4 20/20] drm/rcar-du: Allow VSP and writeback build with COMPILE_TEST=y
  2026-11-18 21:44 [PATCH v4 00/20] drm: Increase COMPILE_TEST=y coverage Ville Syrjala
                   ` (18 preceding siblings ...)
  2026-11-18 21:45 ` [PATCH v4 19/20] drm/nouveau: Allow platform driver " Ville Syrjala
@ 2026-11-18 21:45 ` Ville Syrjala
  2026-11-18 22:11   ` Laurent Pinchart
  2026-11-19  7:33   ` Geert Uytterhoeven
  19 siblings, 2 replies; 36+ messages in thread
From: Ville Syrjala @ 2026-11-18 21:45 UTC (permalink / raw)
  To: dri-devel
  Cc: Laurent Pinchart, Tomi Valkeinen, Kieran Bingham,
	Geert Uytterhoeven, Magnus Damm, linux-renesas-soc

From: Ville Syrjälä <ville.syrjala@linux.intel.com>

Make DRM_RCAR_VSP selectable with COMPILE_TEST=y, and enable the hidden
DRM_RCAR_WRITEBACK option by default, so the VSP and writeback code gets
built for greater coverage. Builds fine on x86, x86_64, arm and arm64 at
least.

Cc: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
Cc: Tomi Valkeinen <tomi.valkeinen+renesas@ideasonboard.com>
Cc: Kieran Bingham <kieran.bingham+renesas@ideasonboard.com>
Cc: Geert Uytterhoeven <geert+renesas@glider.be>
Cc: Magnus Damm <magnus.damm@gmail.com>
Cc: linux-renesas-soc@vger.kernel.org
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
---
 drivers/gpu/drm/renesas/rcar-du/Kconfig | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/gpu/drm/renesas/rcar-du/Kconfig b/drivers/gpu/drm/renesas/rcar-du/Kconfig
index 80bd770ae0f6..5394f131c58f 100644
--- a/drivers/gpu/drm/renesas/rcar-du/Kconfig
+++ b/drivers/gpu/drm/renesas/rcar-du/Kconfig
@@ -76,7 +76,7 @@ config DRM_RCAR_MIPI_DSI
 	select RESET_CONTROLLER
 
 config DRM_RCAR_VSP
-	bool "R-Car DU VSP Compositor Support" if ARM
+	bool "R-Car DU VSP Compositor Support" if ARM || COMPILE_TEST
 	default y if ARM64
 	depends on DRM_RCAR_DU
 	depends on VIDEO_RENESAS_VSP1=y || (VIDEO_RENESAS_VSP1 && DRM_RCAR_DU=m)
@@ -85,5 +85,5 @@ config DRM_RCAR_VSP
 
 config DRM_RCAR_WRITEBACK
 	bool
-	default y if ARM64
+	default y if ARM64 || COMPILE_TEST
 	depends on DRM_RCAR_DU
-- 
2.56.0


^ permalink raw reply related	[flat|nested] 36+ messages in thread

* Re: [PATCH v4 19/20] drm/nouveau: Allow platform driver build with COMPILE_TEST=y
  2026-11-18 21:45 ` [PATCH v4 19/20] drm/nouveau: Allow platform driver " Ville Syrjala
@ 2026-11-18 21:57   ` sashiko-bot
  2026-11-18 22:28     ` Ville Syrjälä
  0 siblings, 1 reply; 36+ messages in thread
From: sashiko-bot @ 2026-11-18 21:57 UTC (permalink / raw)
  To: Ville Syrjala, dri-devel

Thank you for your contribution! Sashiko AI review found 1 potential
issue to consider:
- [Medium] drm/nouveau: build failure in nouveau_platform_resume() when
  PM_DEVFREQ=n
--

commit 7f3a768f8ef913fb90efb9c01241f3d8714fc297
Author: Ville Syrjälä <ville.syrjala@linux.intel.com>

drm/nouveau: Allow platform driver build with COMPILE_TEST=y

This commit allows NOUVEAU_PLATFORM_DRIVER to be enabled when COMPILE_TEST=y
even on non-Tegra architectures.

> diff --git a/drivers/gpu/drm/nouveau/Kconfig b/drivers/gpu/drm/nouveau/Kconfig
> index 3b5757aed9c84..d7b432c839e5f 100644
> --- a/drivers/gpu/drm/nouveau/Kconfig
> +++ b/drivers/gpu/drm/nouveau/Kconfig
> @@ -34,7 +34,7 @@ config DRM_NOUVEAU
>  
>  config NOUVEAU_PLATFORM_DRIVER
>  	bool "Nouveau (NVIDIA) SoC GPUs"
> -	depends on DRM_NOUVEAU && ARCH_TEGRA
> +	depends on DRM_NOUVEAU && (ARCH_TEGRA || COMPILE_TEST)
>  	default y
>  	help
>  	  Support for Nouveau platform driver, used for SoC GPUs as found

[Severity: Medium]
Does this break the build in nouveau_platform_resume() when
CONFIG_PM_SLEEP=y and CONFIG_PM_DEVFREQ=n?

DRM_NOUVEAU only has "select PM_DEVFREQ if ARCH_TEGRA", so when
ARCH_TEGRA=n and COMPILE_TEST=y, CONFIG_PM_DEVFREQ can remain disabled.
The fallback stub for gk20a_devfreq_resume() in
drivers/gpu/drm/nouveau/nvkm/subdev/clk/gk20a_devfreq.h is missing the
pointer asterisk on its parameter:

drivers/gpu/drm/nouveau/nvkm/subdev/clk/gk20a_devfreq.h:
    #if defined(CONFIG_PM_DEVFREQ)
    int gk20a_devfreq_init(struct nvkm_clk *base, struct gk20a_devfreq **devfreq);
    int gk20a_devfreq_resume(struct device *dev);
    int gk20a_devfreq_suspend(struct device *dev);
    #else
    static inline int gk20a_devfreq_init(struct nvkm_clk *base, struct gk20a_devfreq **devfreq)
    {
    	return 0;
    }

    static inline int gk20a_devfreq_resume(struct device dev) { return 0; }
    static inline int gk20a_devfreq_suspend(struct device *dev) { return 0; }
    #endif

Passing struct device *dev from nouveau_platform_resume() to
gk20a_devfreq_resume() then fails to compile with an incompatible type
error.

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20261118214518.29317-1-ville.syrjala@linux.intel.com?part=19

^ permalink raw reply	[flat|nested] 36+ messages in thread

* Re: [PATCH v4 15/20] drm/imagination: Allow build with COMPILE_TEST=y
  2026-11-18 21:45 ` [PATCH v4 15/20] drm/imagination: Allow build with COMPILE_TEST=y Ville Syrjala
@ 2026-11-18 21:58   ` sashiko-bot
  2026-11-18 22:30     ` Ville Syrjälä
  0 siblings, 1 reply; 36+ messages in thread
From: sashiko-bot @ 2026-11-18 21:58 UTC (permalink / raw)
  To: Ville Syrjala, dri-devel

Thank you for your contribution! Sashiko AI review found 1 potential
issue to consider:
- [Medium] drm/imagination: COMPILE_TEST build failure on 64-bit
  8KB/32KB page configs
--

commit 3a0a5d98906b6758538e2c13acfa780422a1a949
Author: Ville Syrjälä <ville.syrjala@linux.intel.com>

drm/imagination: Allow build with COMPILE_TEST=y

This commit allows DRM_POWERVR to be built on any 64-bit architecture when
COMPILE_TEST is enabled.

> diff --git a/drivers/gpu/drm/imagination/Kconfig b/drivers/gpu/drm/imagination/Kconfig
> index 1fd4c635c2c96..a048d3c859385 100644
> --- a/drivers/gpu/drm/imagination/Kconfig
> +++ b/drivers/gpu/drm/imagination/Kconfig
> @@ -3,7 +3,8 @@
>  
>  config DRM_POWERVR
>  	tristate "Imagination Technologies PowerVR (Series 6 and later) & IMG Graphics"
> -	depends on (ARM64 || RISCV && 64BIT)
> +	depends on ARM64 || RISCV || COMPILE_TEST
> +	depends on 64BIT
>  	depends on DRM
>  	depends on MMU
>  	depends on PM

[Severity: Medium]
Does enabling DRM_POWERVR under COMPILE_TEST && 64BIT break the build on
64-bit targets with 8KiB or 32KiB pages (such as sparc64, alpha, or 64-bit
mips with PAGE_SIZE_8KB/PAGE_SIZE_32KB)?

drivers/gpu/drm/imagination/pvr_mmu.h defines PVR_DEVICE_PAGE_SIZE as
PAGE_SIZE, and drivers/gpu/drm/imagination/pvr_mmu.c only supports 4KiB,
16KiB, 64KiB, 256KiB, 1MiB, and 2MiB pages, failing with an #error on any
other host page size.  Should DRM_POWERVR also depend on !PAGE_SIZE_8KB &&
!PAGE_SIZE_32KB (or restrict to supported page sizes)?

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20261118214518.29317-1-ville.syrjala@linux.intel.com?part=15

^ permalink raw reply	[flat|nested] 36+ messages in thread

* Re: [PATCH v4 20/20] drm/rcar-du: Allow VSP and writeback build with COMPILE_TEST=y
  2026-11-18 21:45 ` [PATCH v4 20/20] drm/rcar-du: Allow VSP and writeback " Ville Syrjala
@ 2026-11-18 22:11   ` Laurent Pinchart
  2026-11-18 22:32     ` Ville Syrjälä
  2026-11-19  7:33   ` Geert Uytterhoeven
  1 sibling, 1 reply; 36+ messages in thread
From: Laurent Pinchart @ 2026-11-18 22:11 UTC (permalink / raw)
  To: Ville Syrjala
  Cc: dri-devel, Tomi Valkeinen, Kieran Bingham, Geert Uytterhoeven,
	Magnus Damm, linux-renesas-soc

On Wed, Nov 18, 2026 at 11:45:17PM +0200, Ville Syrjala wrote:
> From: Ville Syrjälä <ville.syrjala@linux.intel.com>
> 
> Make DRM_RCAR_VSP selectable with COMPILE_TEST=y, and enable the hidden
> DRM_RCAR_WRITEBACK option by default, so the VSP and writeback code gets
> built for greater coverage. Builds fine on x86, x86_64, arm and arm64 at
> least.

Let's see what build bots will report on more exotic architectures. I
may volunteer you to send a fix if I'm too busy when the builders start
yelling :-)

> Cc: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
> Cc: Tomi Valkeinen <tomi.valkeinen+renesas@ideasonboard.com>
> Cc: Kieran Bingham <kieran.bingham+renesas@ideasonboard.com>
> Cc: Geert Uytterhoeven <geert+renesas@glider.be>
> Cc: Magnus Damm <magnus.damm@gmail.com>
> Cc: linux-renesas-soc@vger.kernel.org
> Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
> ---
>  drivers/gpu/drm/renesas/rcar-du/Kconfig | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/gpu/drm/renesas/rcar-du/Kconfig b/drivers/gpu/drm/renesas/rcar-du/Kconfig
> index 80bd770ae0f6..5394f131c58f 100644
> --- a/drivers/gpu/drm/renesas/rcar-du/Kconfig
> +++ b/drivers/gpu/drm/renesas/rcar-du/Kconfig
> @@ -76,7 +76,7 @@ config DRM_RCAR_MIPI_DSI
>  	select RESET_CONTROLLER
>  
>  config DRM_RCAR_VSP
> -	bool "R-Car DU VSP Compositor Support" if ARM
> +	bool "R-Car DU VSP Compositor Support" if ARM || COMPILE_TEST

This will make it possible to disable the option when ARM64 &&
COMPILE_TEST, which will break operation on Gen3 and up. It would be
best to avoid that.

>  	default y if ARM64
>  	depends on DRM_RCAR_DU
>  	depends on VIDEO_RENESAS_VSP1=y || (VIDEO_RENESAS_VSP1 && DRM_RCAR_DU=m)
> @@ -85,5 +85,5 @@ config DRM_RCAR_VSP
>  
>  config DRM_RCAR_WRITEBACK
>  	bool
> -	default y if ARM64
> +	default y if ARM64 || COMPILE_TEST
>  	depends on DRM_RCAR_DU

-- 
Regards,

Laurent Pinchart

^ permalink raw reply	[flat|nested] 36+ messages in thread

* Re: [PATCH v4 19/20] drm/nouveau: Allow platform driver build with COMPILE_TEST=y
  2026-11-18 21:57   ` sashiko-bot
@ 2026-11-18 22:28     ` Ville Syrjälä
  0 siblings, 0 replies; 36+ messages in thread
From: Ville Syrjälä @ 2026-11-18 22:28 UTC (permalink / raw)
  To: sashiko-reviews; +Cc: dri-devel

On Wed, Nov 18, 2026 at 09:57:31PM +0000, sashiko-bot@kernel.org wrote:
> Thank you for your contribution! Sashiko AI review found 1 potential
> issue to consider:
> - [Medium] drm/nouveau: build failure in nouveau_platform_resume() when
>   PM_DEVFREQ=n
> --
> 
> commit 7f3a768f8ef913fb90efb9c01241f3d8714fc297
> Author: Ville Syrjälä <ville.syrjala@linux.intel.com>
> 
> drm/nouveau: Allow platform driver build with COMPILE_TEST=y
> 
> This commit allows NOUVEAU_PLATFORM_DRIVER to be enabled when COMPILE_TEST=y
> even on non-Tegra architectures.
> 
> > diff --git a/drivers/gpu/drm/nouveau/Kconfig b/drivers/gpu/drm/nouveau/Kconfig
> > index 3b5757aed9c84..d7b432c839e5f 100644
> > --- a/drivers/gpu/drm/nouveau/Kconfig
> > +++ b/drivers/gpu/drm/nouveau/Kconfig
> > @@ -34,7 +34,7 @@ config DRM_NOUVEAU
> >  
> >  config NOUVEAU_PLATFORM_DRIVER
> >  	bool "Nouveau (NVIDIA) SoC GPUs"
> > -	depends on DRM_NOUVEAU && ARCH_TEGRA
> > +	depends on DRM_NOUVEAU && (ARCH_TEGRA || COMPILE_TEST)
> >  	default y
> >  	help
> >  	  Support for Nouveau platform driver, used for SoC GPUs as found
> 
> [Severity: Medium]
> Does this break the build in nouveau_platform_resume() when
> CONFIG_PM_SLEEP=y and CONFIG_PM_DEVFREQ=n?
> 
> DRM_NOUVEAU only has "select PM_DEVFREQ if ARCH_TEGRA", so when
> ARCH_TEGRA=n and COMPILE_TEST=y, CONFIG_PM_DEVFREQ can remain disabled.
> The fallback stub for gk20a_devfreq_resume() in
> drivers/gpu/drm/nouveau/nvkm/subdev/clk/gk20a_devfreq.h is missing the
> pointer asterisk on its parameter:
> 
> drivers/gpu/drm/nouveau/nvkm/subdev/clk/gk20a_devfreq.h:
>     #if defined(CONFIG_PM_DEVFREQ)
>     int gk20a_devfreq_init(struct nvkm_clk *base, struct gk20a_devfreq **devfreq);
>     int gk20a_devfreq_resume(struct device *dev);
>     int gk20a_devfreq_suspend(struct device *dev);
>     #else
>     static inline int gk20a_devfreq_init(struct nvkm_clk *base, struct gk20a_devfreq **devfreq)
>     {
>     	return 0;
>     }
> 
>     static inline int gk20a_devfreq_resume(struct device dev) { return 0; }
>     static inline int gk20a_devfreq_suspend(struct device *dev) { return 0; }
>     #endif
> 
> Passing struct device *dev from nouveau_platform_resume() to
> gk20a_devfreq_resume() then fails to compile with an incompatible type
> error.

Ack. Fix is now on my local branch. I'll include it if/when I resend.

-- 
Ville Syrjälä
Intel

^ permalink raw reply	[flat|nested] 36+ messages in thread

* Re: [PATCH v4 15/20] drm/imagination: Allow build with COMPILE_TEST=y
  2026-11-18 21:58   ` sashiko-bot
@ 2026-11-18 22:30     ` Ville Syrjälä
  0 siblings, 0 replies; 36+ messages in thread
From: Ville Syrjälä @ 2026-11-18 22:30 UTC (permalink / raw)
  To: sashiko-reviews; +Cc: dri-devel

On Wed, Nov 18, 2026 at 09:58:47PM +0000, sashiko-bot@kernel.org wrote:
> Thank you for your contribution! Sashiko AI review found 1 potential
> issue to consider:
> - [Medium] drm/imagination: COMPILE_TEST build failure on 64-bit
>   8KB/32KB page configs
> --
> 
> commit 3a0a5d98906b6758538e2c13acfa780422a1a949
> Author: Ville Syrjälä <ville.syrjala@linux.intel.com>
> 
> drm/imagination: Allow build with COMPILE_TEST=y
> 
> This commit allows DRM_POWERVR to be built on any 64-bit architecture when
> COMPILE_TEST is enabled.
> 
> > diff --git a/drivers/gpu/drm/imagination/Kconfig b/drivers/gpu/drm/imagination/Kconfig
> > index 1fd4c635c2c96..a048d3c859385 100644
> > --- a/drivers/gpu/drm/imagination/Kconfig
> > +++ b/drivers/gpu/drm/imagination/Kconfig
> > @@ -3,7 +3,8 @@
> >  
> >  config DRM_POWERVR
> >  	tristate "Imagination Technologies PowerVR (Series 6 and later) & IMG Graphics"
> > -	depends on (ARM64 || RISCV && 64BIT)
> > +	depends on ARM64 || RISCV || COMPILE_TEST
> > +	depends on 64BIT
> >  	depends on DRM
> >  	depends on MMU
> >  	depends on PM
> 
> [Severity: Medium]
> Does enabling DRM_POWERVR under COMPILE_TEST && 64BIT break the build on
> 64-bit targets with 8KiB or 32KiB pages (such as sparc64, alpha, or 64-bit
> mips with PAGE_SIZE_8KB/PAGE_SIZE_32KB)?
> 
> drivers/gpu/drm/imagination/pvr_mmu.h defines PVR_DEVICE_PAGE_SIZE as
> PAGE_SIZE, and drivers/gpu/drm/imagination/pvr_mmu.c only supports 4KiB,
> 16KiB, 64KiB, 256KiB, 1MiB, and 2MiB pages, failing with an #error on any
> other host page size.  Should DRM_POWERVR also depend on !PAGE_SIZE_8KB &&
> !PAGE_SIZE_32KB (or restrict to supported page sizes)?

I suppose. I might have to install a few more cross compilers and set
Claude to work with them.

-- 
Ville Syrjälä
Intel

^ permalink raw reply	[flat|nested] 36+ messages in thread

* Re: [PATCH v4 20/20] drm/rcar-du: Allow VSP and writeback build with COMPILE_TEST=y
  2026-11-18 22:11   ` Laurent Pinchart
@ 2026-11-18 22:32     ` Ville Syrjälä
  2026-11-18 22:39       ` Laurent Pinchart
  0 siblings, 1 reply; 36+ messages in thread
From: Ville Syrjälä @ 2026-11-18 22:32 UTC (permalink / raw)
  To: Laurent Pinchart
  Cc: dri-devel, Tomi Valkeinen, Kieran Bingham, Geert Uytterhoeven,
	Magnus Damm, linux-renesas-soc

On Thu, Nov 19, 2026 at 12:11:14AM +0200, Laurent Pinchart wrote:
> On Wed, Nov 18, 2026 at 11:45:17PM +0200, Ville Syrjala wrote:
> > From: Ville Syrjälä <ville.syrjala@linux.intel.com>
> > 
> > Make DRM_RCAR_VSP selectable with COMPILE_TEST=y, and enable the hidden
> > DRM_RCAR_WRITEBACK option by default, so the VSP and writeback code gets
> > built for greater coverage. Builds fine on x86, x86_64, arm and arm64 at
> > least.
> 
> Let's see what build bots will report on more exotic architectures. I
> may volunteer you to send a fix if I'm too busy when the builders start
> yelling :-)
> 
> > Cc: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
> > Cc: Tomi Valkeinen <tomi.valkeinen+renesas@ideasonboard.com>
> > Cc: Kieran Bingham <kieran.bingham+renesas@ideasonboard.com>
> > Cc: Geert Uytterhoeven <geert+renesas@glider.be>
> > Cc: Magnus Damm <magnus.damm@gmail.com>
> > Cc: linux-renesas-soc@vger.kernel.org
> > Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
> > ---
> >  drivers/gpu/drm/renesas/rcar-du/Kconfig | 4 ++--
> >  1 file changed, 2 insertions(+), 2 deletions(-)
> > 
> > diff --git a/drivers/gpu/drm/renesas/rcar-du/Kconfig b/drivers/gpu/drm/renesas/rcar-du/Kconfig
> > index 80bd770ae0f6..5394f131c58f 100644
> > --- a/drivers/gpu/drm/renesas/rcar-du/Kconfig
> > +++ b/drivers/gpu/drm/renesas/rcar-du/Kconfig
> > @@ -76,7 +76,7 @@ config DRM_RCAR_MIPI_DSI
> >  	select RESET_CONTROLLER
> >  
> >  config DRM_RCAR_VSP
> > -	bool "R-Car DU VSP Compositor Support" if ARM
> > +	bool "R-Car DU VSP Compositor Support" if ARM || COMPILE_TEST
> 
> This will make it possible to disable the option when ARM64 &&
> COMPILE_TEST, which will break operation on Gen3 and up. It would be
> best to avoid that.

Break operation, as in when running on real hardware?
People actually do that with COMPILE_TEST=y kernels?

> 
> >  	default y if ARM64
> >  	depends on DRM_RCAR_DU
> >  	depends on VIDEO_RENESAS_VSP1=y || (VIDEO_RENESAS_VSP1 && DRM_RCAR_DU=m)
> > @@ -85,5 +85,5 @@ config DRM_RCAR_VSP
> >  
> >  config DRM_RCAR_WRITEBACK
> >  	bool
> > -	default y if ARM64
> > +	default y if ARM64 || COMPILE_TEST
> >  	depends on DRM_RCAR_DU
> 
> -- 
> Regards,
> 
> Laurent Pinchart

-- 
Ville Syrjälä
Intel

^ permalink raw reply	[flat|nested] 36+ messages in thread

* Re: [PATCH v4 20/20] drm/rcar-du: Allow VSP and writeback build with COMPILE_TEST=y
  2026-11-18 22:32     ` Ville Syrjälä
@ 2026-11-18 22:39       ` Laurent Pinchart
  2026-11-18 22:57         ` Ville Syrjälä
  0 siblings, 1 reply; 36+ messages in thread
From: Laurent Pinchart @ 2026-11-18 22:39 UTC (permalink / raw)
  To: Ville Syrjälä
  Cc: dri-devel, Tomi Valkeinen, Kieran Bingham, Geert Uytterhoeven,
	Magnus Damm, linux-renesas-soc

On Thu, Nov 19, 2026 at 12:32:15AM +0200, Ville Syrjälä wrote:
> On Thu, Nov 19, 2026 at 12:11:14AM +0200, Laurent Pinchart wrote:
> > On Wed, Nov 18, 2026 at 11:45:17PM +0200, Ville Syrjala wrote:
> > > From: Ville Syrjälä <ville.syrjala@linux.intel.com>
> > > 
> > > Make DRM_RCAR_VSP selectable with COMPILE_TEST=y, and enable the hidden
> > > DRM_RCAR_WRITEBACK option by default, so the VSP and writeback code gets
> > > built for greater coverage. Builds fine on x86, x86_64, arm and arm64 at
> > > least.
> > 
> > Let's see what build bots will report on more exotic architectures. I
> > may volunteer you to send a fix if I'm too busy when the builders start
> > yelling :-)
> > 
> > > Cc: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
> > > Cc: Tomi Valkeinen <tomi.valkeinen+renesas@ideasonboard.com>
> > > Cc: Kieran Bingham <kieran.bingham+renesas@ideasonboard.com>
> > > Cc: Geert Uytterhoeven <geert+renesas@glider.be>
> > > Cc: Magnus Damm <magnus.damm@gmail.com>
> > > Cc: linux-renesas-soc@vger.kernel.org
> > > Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
> > > ---
> > >  drivers/gpu/drm/renesas/rcar-du/Kconfig | 4 ++--
> > >  1 file changed, 2 insertions(+), 2 deletions(-)
> > > 
> > > diff --git a/drivers/gpu/drm/renesas/rcar-du/Kconfig b/drivers/gpu/drm/renesas/rcar-du/Kconfig
> > > index 80bd770ae0f6..5394f131c58f 100644
> > > --- a/drivers/gpu/drm/renesas/rcar-du/Kconfig
> > > +++ b/drivers/gpu/drm/renesas/rcar-du/Kconfig
> > > @@ -76,7 +76,7 @@ config DRM_RCAR_MIPI_DSI
> > >  	select RESET_CONTROLLER
> > >  
> > >  config DRM_RCAR_VSP
> > > -	bool "R-Car DU VSP Compositor Support" if ARM
> > > +	bool "R-Car DU VSP Compositor Support" if ARM || COMPILE_TEST
> > 
> > This will make it possible to disable the option when ARM64 &&
> > COMPILE_TEST, which will break operation on Gen3 and up. It would be
> > best to avoid that.
> 
> Break operation, as in when running on real hardware?
> People actually do that with COMPILE_TEST=y kernels?

Hopefully not in production, but that may be the case during development
(I don't though, there are enough kernel options to deal with without
COMPILE_TEST).

This should do:

	bool "R-Car DU VSP Compositor Support" if ARM || (COMPILE_TEST && !ARM64)

(untested though). Do you think it brings too much complexity ?

> > >  	default y if ARM64
> > >  	depends on DRM_RCAR_DU
> > >  	depends on VIDEO_RENESAS_VSP1=y || (VIDEO_RENESAS_VSP1 && DRM_RCAR_DU=m)
> > > @@ -85,5 +85,5 @@ config DRM_RCAR_VSP
> > >  
> > >  config DRM_RCAR_WRITEBACK
> > >  	bool
> > > -	default y if ARM64
> > > +	default y if ARM64 || COMPILE_TEST
> > >  	depends on DRM_RCAR_DU

-- 
Regards,

Laurent Pinchart

^ permalink raw reply	[flat|nested] 36+ messages in thread

* Re: [PATCH v4 20/20] drm/rcar-du: Allow VSP and writeback build with COMPILE_TEST=y
  2026-11-18 22:39       ` Laurent Pinchart
@ 2026-11-18 22:57         ` Ville Syrjälä
  2026-11-18 23:14           ` Laurent Pinchart
  2026-11-19  7:38           ` Geert Uytterhoeven
  0 siblings, 2 replies; 36+ messages in thread
From: Ville Syrjälä @ 2026-11-18 22:57 UTC (permalink / raw)
  To: Laurent Pinchart
  Cc: dri-devel, Tomi Valkeinen, Kieran Bingham, Geert Uytterhoeven,
	Magnus Damm, linux-renesas-soc

On Thu, Nov 19, 2026 at 12:39:11AM +0200, Laurent Pinchart wrote:
> On Thu, Nov 19, 2026 at 12:32:15AM +0200, Ville Syrjälä wrote:
> > On Thu, Nov 19, 2026 at 12:11:14AM +0200, Laurent Pinchart wrote:
> > > On Wed, Nov 18, 2026 at 11:45:17PM +0200, Ville Syrjala wrote:
> > > > From: Ville Syrjälä <ville.syrjala@linux.intel.com>
> > > > 
> > > > Make DRM_RCAR_VSP selectable with COMPILE_TEST=y, and enable the hidden
> > > > DRM_RCAR_WRITEBACK option by default, so the VSP and writeback code gets
> > > > built for greater coverage. Builds fine on x86, x86_64, arm and arm64 at
> > > > least.
> > > 
> > > Let's see what build bots will report on more exotic architectures. I
> > > may volunteer you to send a fix if I'm too busy when the builders start
> > > yelling :-)
> > > 
> > > > Cc: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
> > > > Cc: Tomi Valkeinen <tomi.valkeinen+renesas@ideasonboard.com>
> > > > Cc: Kieran Bingham <kieran.bingham+renesas@ideasonboard.com>
> > > > Cc: Geert Uytterhoeven <geert+renesas@glider.be>
> > > > Cc: Magnus Damm <magnus.damm@gmail.com>
> > > > Cc: linux-renesas-soc@vger.kernel.org
> > > > Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
> > > > ---
> > > >  drivers/gpu/drm/renesas/rcar-du/Kconfig | 4 ++--
> > > >  1 file changed, 2 insertions(+), 2 deletions(-)
> > > > 
> > > > diff --git a/drivers/gpu/drm/renesas/rcar-du/Kconfig b/drivers/gpu/drm/renesas/rcar-du/Kconfig
> > > > index 80bd770ae0f6..5394f131c58f 100644
> > > > --- a/drivers/gpu/drm/renesas/rcar-du/Kconfig
> > > > +++ b/drivers/gpu/drm/renesas/rcar-du/Kconfig
> > > > @@ -76,7 +76,7 @@ config DRM_RCAR_MIPI_DSI
> > > >  	select RESET_CONTROLLER
> > > >  
> > > >  config DRM_RCAR_VSP
> > > > -	bool "R-Car DU VSP Compositor Support" if ARM
> > > > +	bool "R-Car DU VSP Compositor Support" if ARM || COMPILE_TEST
> > > 
> > > This will make it possible to disable the option when ARM64 &&
> > > COMPILE_TEST, which will break operation on Gen3 and up. It would be
> > > best to avoid that.
> > 
> > Break operation, as in when running on real hardware?
> > People actually do that with COMPILE_TEST=y kernels?
> 
> Hopefully not in production, but that may be the case during development
> (I don't though, there are enough kernel options to deal with without
> COMPILE_TEST).
> 
> This should do:
> 
> 	bool "R-Car DU VSP Compositor Support" if ARM || (COMPILE_TEST && !ARM64)
> 
> (untested though). Do you think it brings too much complexity ?

Doesn't bother me.

All I'm really after is being able to have COMPILE_TEST=y
enabled in my x86-32 and x86-64 drm-misc test build configs
so that I don't have to build on all the other architectures,
while still having reasonable confidence that I've not
completely broken some non-x86 drm driver.

-- 
Ville Syrjälä
Intel

^ permalink raw reply	[flat|nested] 36+ messages in thread

* Re: [PATCH v4 20/20] drm/rcar-du: Allow VSP and writeback build with COMPILE_TEST=y
  2026-11-18 22:57         ` Ville Syrjälä
@ 2026-11-18 23:14           ` Laurent Pinchart
  2026-11-19  7:38           ` Geert Uytterhoeven
  1 sibling, 0 replies; 36+ messages in thread
From: Laurent Pinchart @ 2026-11-18 23:14 UTC (permalink / raw)
  To: Ville Syrjälä
  Cc: dri-devel, Tomi Valkeinen, Kieran Bingham, Geert Uytterhoeven,
	Magnus Damm, linux-renesas-soc

On Thu, Nov 19, 2026 at 12:57:05AM +0200, Ville Syrjälä wrote:
> On Thu, Nov 19, 2026 at 12:39:11AM +0200, Laurent Pinchart wrote:
> > On Thu, Nov 19, 2026 at 12:32:15AM +0200, Ville Syrjälä wrote:
> > > On Thu, Nov 19, 2026 at 12:11:14AM +0200, Laurent Pinchart wrote:
> > > > On Wed, Nov 18, 2026 at 11:45:17PM +0200, Ville Syrjala wrote:
> > > > > From: Ville Syrjälä <ville.syrjala@linux.intel.com>
> > > > > 
> > > > > Make DRM_RCAR_VSP selectable with COMPILE_TEST=y, and enable the hidden
> > > > > DRM_RCAR_WRITEBACK option by default, so the VSP and writeback code gets
> > > > > built for greater coverage. Builds fine on x86, x86_64, arm and arm64 at
> > > > > least.
> > > > 
> > > > Let's see what build bots will report on more exotic architectures. I
> > > > may volunteer you to send a fix if I'm too busy when the builders start
> > > > yelling :-)
> > > > 
> > > > > Cc: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
> > > > > Cc: Tomi Valkeinen <tomi.valkeinen+renesas@ideasonboard.com>
> > > > > Cc: Kieran Bingham <kieran.bingham+renesas@ideasonboard.com>
> > > > > Cc: Geert Uytterhoeven <geert+renesas@glider.be>
> > > > > Cc: Magnus Damm <magnus.damm@gmail.com>
> > > > > Cc: linux-renesas-soc@vger.kernel.org
> > > > > Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
> > > > > ---
> > > > >  drivers/gpu/drm/renesas/rcar-du/Kconfig | 4 ++--
> > > > >  1 file changed, 2 insertions(+), 2 deletions(-)
> > > > > 
> > > > > diff --git a/drivers/gpu/drm/renesas/rcar-du/Kconfig b/drivers/gpu/drm/renesas/rcar-du/Kconfig
> > > > > index 80bd770ae0f6..5394f131c58f 100644
> > > > > --- a/drivers/gpu/drm/renesas/rcar-du/Kconfig
> > > > > +++ b/drivers/gpu/drm/renesas/rcar-du/Kconfig
> > > > > @@ -76,7 +76,7 @@ config DRM_RCAR_MIPI_DSI
> > > > >  	select RESET_CONTROLLER
> > > > >  
> > > > >  config DRM_RCAR_VSP
> > > > > -	bool "R-Car DU VSP Compositor Support" if ARM
> > > > > +	bool "R-Car DU VSP Compositor Support" if ARM || COMPILE_TEST
> > > > 
> > > > This will make it possible to disable the option when ARM64 &&
> > > > COMPILE_TEST, which will break operation on Gen3 and up. It would be
> > > > best to avoid that.
> > > 
> > > Break operation, as in when running on real hardware?
> > > People actually do that with COMPILE_TEST=y kernels?
> > 
> > Hopefully not in production, but that may be the case during development
> > (I don't though, there are enough kernel options to deal with without
> > COMPILE_TEST).
> > 
> > This should do:
> > 
> > 	bool "R-Car DU VSP Compositor Support" if ARM || (COMPILE_TEST && !ARM64)
> > 
> > (untested though). Do you think it brings too much complexity ?
> 
> Doesn't bother me.
> 
> All I'm really after is being able to have COMPILE_TEST=y
> enabled in my x86-32 and x86-64 drm-misc test build configs
> so that I don't have to build on all the other architectures,
> while still having reasonable confidence that I've not
> completely broken some non-x86 drm driver.

Sure. Let's go for the above change then. The commit message could also
be updated, maybe to

----
Make DRM_RCAR_VSP selectable with COMPILE_TEST=y (but keep it
unconditionally enabled on ARM64 to avoid breaking operation), and
enable the hidden DRM_RCAR_WRITEBACK option by default, so the VSP and
writeback code gets built for greater coverage. Builds fine on x86,
x86_64, arm and arm64 at least.
----

With that,

Reviewed-by: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>

I assume you'll merge this with the rest of the series.

-- 
Regards,

Laurent Pinchart

^ permalink raw reply	[flat|nested] 36+ messages in thread

* Re: [PATCH v4 20/20] drm/rcar-du: Allow VSP and writeback build with COMPILE_TEST=y
  2026-11-18 21:45 ` [PATCH v4 20/20] drm/rcar-du: Allow VSP and writeback " Ville Syrjala
  2026-11-18 22:11   ` Laurent Pinchart
@ 2026-11-19  7:33   ` Geert Uytterhoeven
  1 sibling, 0 replies; 36+ messages in thread
From: Geert Uytterhoeven @ 2026-11-19  7:33 UTC (permalink / raw)
  To: Ville Syrjala
  Cc: dri-devel, Laurent Pinchart, Tomi Valkeinen, Kieran Bingham,
	Geert Uytterhoeven, Magnus Damm, linux-renesas-soc

Hi Ville,

On Wed, 18 Nov 2026 at 22:47, Ville Syrjala
<ville.syrjala@linux.intel.com> wrote:
> From: Ville Syrjälä <ville.syrjala@linux.intel.com>
>
> Make DRM_RCAR_VSP selectable with COMPILE_TEST=y, and enable the hidden
> DRM_RCAR_WRITEBACK option by default, so the VSP and writeback code gets
> built for greater coverage. Builds fine on x86, x86_64, arm and arm64 at
> least.

> Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>

Thanks for your patch!

> --- a/drivers/gpu/drm/renesas/rcar-du/Kconfig
> +++ b/drivers/gpu/drm/renesas/rcar-du/Kconfig

> @@ -85,5 +85,5 @@ config DRM_RCAR_VSP
>
>  config DRM_RCAR_WRITEBACK
>         bool
> -       default y if ARM64
> +       default y if ARM64 || COMPILE_TEST
>         depends on DRM_RCAR_DU

Merely enabling COMPILE_TEST should not enable extra functionality.

Gr{oetje,eeting}s,

                        Geert

-- 
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds

^ permalink raw reply	[flat|nested] 36+ messages in thread

* Re: [PATCH v4 20/20] drm/rcar-du: Allow VSP and writeback build with COMPILE_TEST=y
  2026-11-18 22:57         ` Ville Syrjälä
  2026-11-18 23:14           ` Laurent Pinchart
@ 2026-11-19  7:38           ` Geert Uytterhoeven
  2026-11-19  8:16             ` Laurent Pinchart
  1 sibling, 1 reply; 36+ messages in thread
From: Geert Uytterhoeven @ 2026-11-19  7:38 UTC (permalink / raw)
  To: Ville Syrjälä
  Cc: Laurent Pinchart, dri-devel, Tomi Valkeinen, Kieran Bingham,
	Geert Uytterhoeven, Magnus Damm, linux-renesas-soc

On Wed, 18 Nov 2026 at 23:57, Ville Syrjälä
<ville.syrjala@linux.intel.com> wrote:
> On Thu, Nov 19, 2026 at 12:39:11AM +0200, Laurent Pinchart wrote:
> > On Thu, Nov 19, 2026 at 12:32:15AM +0200, Ville Syrjälä wrote:
> > > On Thu, Nov 19, 2026 at 12:11:14AM +0200, Laurent Pinchart wrote:
> > > > On Wed, Nov 18, 2026 at 11:45:17PM +0200, Ville Syrjala wrote:
> > > > > From: Ville Syrjälä <ville.syrjala@linux.intel.com>
> > > > >
> > > > > Make DRM_RCAR_VSP selectable with COMPILE_TEST=y, and enable the hidden
> > > > > DRM_RCAR_WRITEBACK option by default, so the VSP and writeback code gets
> > > > > built for greater coverage. Builds fine on x86, x86_64, arm and arm64 at
> > > > > least.
> > > >
> > > > Let's see what build bots will report on more exotic architectures. I
> > > > may volunteer you to send a fix if I'm too busy when the builders start
> > > > yelling :-)
> > > >
> > > > > Cc: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
> > > > > Cc: Tomi Valkeinen <tomi.valkeinen+renesas@ideasonboard.com>
> > > > > Cc: Kieran Bingham <kieran.bingham+renesas@ideasonboard.com>
> > > > > Cc: Geert Uytterhoeven <geert+renesas@glider.be>
> > > > > Cc: Magnus Damm <magnus.damm@gmail.com>
> > > > > Cc: linux-renesas-soc@vger.kernel.org
> > > > > Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
> > > > > ---
> > > > >  drivers/gpu/drm/renesas/rcar-du/Kconfig | 4 ++--
> > > > >  1 file changed, 2 insertions(+), 2 deletions(-)
> > > > >
> > > > > diff --git a/drivers/gpu/drm/renesas/rcar-du/Kconfig b/drivers/gpu/drm/renesas/rcar-du/Kconfig
> > > > > index 80bd770ae0f6..5394f131c58f 100644
> > > > > --- a/drivers/gpu/drm/renesas/rcar-du/Kconfig
> > > > > +++ b/drivers/gpu/drm/renesas/rcar-du/Kconfig
> > > > > @@ -76,7 +76,7 @@ config DRM_RCAR_MIPI_DSI
> > > > >         select RESET_CONTROLLER
> > > > >
> > > > >  config DRM_RCAR_VSP
> > > > > -       bool "R-Car DU VSP Compositor Support" if ARM
> > > > > +       bool "R-Car DU VSP Compositor Support" if ARM || COMPILE_TEST
> > > >
> > > > This will make it possible to disable the option when ARM64 &&
> > > > COMPILE_TEST, which will break operation on Gen3 and up. It would be
> > > > best to avoid that.
> > >
> > > Break operation, as in when running on real hardware?
> > > People actually do that with COMPILE_TEST=y kernels?

COMPILE_TEST=y kernels should still work.

> > Hopefully not in production, but that may be the case during development
> > (I don't though, there are enough kernel options to deal with without
> > COMPILE_TEST).
> >
> > This should do:
> >
> >       bool "R-Car DU VSP Compositor Support" if ARM || (COMPILE_TEST && !ARM64)
> >
> > (untested though). Do you think it brings too much complexity ?
>
> Doesn't bother me.

It does bother me, as it breaks the safeguards encoded in Kconfig rules.
I think it would be better if DRM_RCAR_DU would select DRM_RCAR_VSP
if ARM64, so DRM_RCAR_VSP cannot be disabled where it is needed.

Gr{oetje,eeting}s,

                        Geert

-- 
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds

^ permalink raw reply	[flat|nested] 36+ messages in thread

* Re: [PATCH v4 20/20] drm/rcar-du: Allow VSP and writeback build with COMPILE_TEST=y
  2026-11-19  7:38           ` Geert Uytterhoeven
@ 2026-11-19  8:16             ` Laurent Pinchart
  2026-11-19  8:23               ` Geert Uytterhoeven
  0 siblings, 1 reply; 36+ messages in thread
From: Laurent Pinchart @ 2026-11-19  8:16 UTC (permalink / raw)
  To: Geert Uytterhoeven
  Cc: Ville Syrjälä, dri-devel, Tomi Valkeinen,
	Kieran Bingham, Geert Uytterhoeven, Magnus Damm,
	linux-renesas-soc

Hi Geert,

On Thu, Nov 19, 2026 at 08:38:21AM +0100, Geert Uytterhoeven wrote:
> On Wed, 18 Nov 2026 at 23:57, Ville Syrjälä wrote:
> > On Thu, Nov 19, 2026 at 12:39:11AM +0200, Laurent Pinchart wrote:
> > > On Thu, Nov 19, 2026 at 12:32:15AM +0200, Ville Syrjälä wrote:
> > > > On Thu, Nov 19, 2026 at 12:11:14AM +0200, Laurent Pinchart wrote:
> > > > > On Wed, Nov 18, 2026 at 11:45:17PM +0200, Ville Syrjala wrote:
> > > > > > From: Ville Syrjälä <ville.syrjala@linux.intel.com>
> > > > > >
> > > > > > Make DRM_RCAR_VSP selectable with COMPILE_TEST=y, and enable the hidden
> > > > > > DRM_RCAR_WRITEBACK option by default, so the VSP and writeback code gets
> > > > > > built for greater coverage. Builds fine on x86, x86_64, arm and arm64 at
> > > > > > least.
> > > > >
> > > > > Let's see what build bots will report on more exotic architectures. I
> > > > > may volunteer you to send a fix if I'm too busy when the builders start
> > > > > yelling :-)
> > > > >
> > > > > > Cc: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
> > > > > > Cc: Tomi Valkeinen <tomi.valkeinen+renesas@ideasonboard.com>
> > > > > > Cc: Kieran Bingham <kieran.bingham+renesas@ideasonboard.com>
> > > > > > Cc: Geert Uytterhoeven <geert+renesas@glider.be>
> > > > > > Cc: Magnus Damm <magnus.damm@gmail.com>
> > > > > > Cc: linux-renesas-soc@vger.kernel.org
> > > > > > Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
> > > > > > ---
> > > > > >  drivers/gpu/drm/renesas/rcar-du/Kconfig | 4 ++--
> > > > > >  1 file changed, 2 insertions(+), 2 deletions(-)
> > > > > >
> > > > > > diff --git a/drivers/gpu/drm/renesas/rcar-du/Kconfig b/drivers/gpu/drm/renesas/rcar-du/Kconfig
> > > > > > index 80bd770ae0f6..5394f131c58f 100644
> > > > > > --- a/drivers/gpu/drm/renesas/rcar-du/Kconfig
> > > > > > +++ b/drivers/gpu/drm/renesas/rcar-du/Kconfig
> > > > > > @@ -76,7 +76,7 @@ config DRM_RCAR_MIPI_DSI
> > > > > >         select RESET_CONTROLLER
> > > > > >
> > > > > >  config DRM_RCAR_VSP
> > > > > > -       bool "R-Car DU VSP Compositor Support" if ARM
> > > > > > +       bool "R-Car DU VSP Compositor Support" if ARM || COMPILE_TEST
> > > > >
> > > > > This will make it possible to disable the option when ARM64 &&
> > > > > COMPILE_TEST, which will break operation on Gen3 and up. It would be
> > > > > best to avoid that.
> > > >
> > > > Break operation, as in when running on real hardware?
> > > > People actually do that with COMPILE_TEST=y kernels?
> 
> COMPILE_TEST=y kernels should still work.
> 
> > > Hopefully not in production, but that may be the case during development
> > > (I don't though, there are enough kernel options to deal with without
> > > COMPILE_TEST).
> > >
> > > This should do:
> > >
> > >       bool "R-Car DU VSP Compositor Support" if ARM || (COMPILE_TEST && !ARM64)
> > >
> > > (untested though). Do you think it brings too much complexity ?
> >
> > Doesn't bother me.
> 
> It does bother me, as it breaks the safeguards encoded in Kconfig rules.
> I think it would be better if DRM_RCAR_DU would select DRM_RCAR_VSP
> if ARM64, so DRM_RCAR_VSP cannot be disabled where it is needed.

That's possibly an option too. But does

	bool "R-Car DU VSP Compositor Support" if ARM || (COMPILE_TEST && !ARM64)

break the safeguard ?

-- 
Regards,

Laurent Pinchart

^ permalink raw reply	[flat|nested] 36+ messages in thread

* Re: [PATCH v4 20/20] drm/rcar-du: Allow VSP and writeback build with COMPILE_TEST=y
  2026-11-19  8:16             ` Laurent Pinchart
@ 2026-11-19  8:23               ` Geert Uytterhoeven
  0 siblings, 0 replies; 36+ messages in thread
From: Geert Uytterhoeven @ 2026-11-19  8:23 UTC (permalink / raw)
  To: Laurent Pinchart
  Cc: Ville Syrjälä, dri-devel, Tomi Valkeinen,
	Kieran Bingham, Geert Uytterhoeven, Magnus Damm,
	linux-renesas-soc

Hi Laurent,

On Thu, 19 Nov 2026 at 09:16, Laurent Pinchart
<laurent.pinchart@ideasonboard.com> wrote:
> On Thu, Nov 19, 2026 at 08:38:21AM +0100, Geert Uytterhoeven wrote:
> > On Wed, 18 Nov 2026 at 23:57, Ville Syrjälä wrote:
> > > On Thu, Nov 19, 2026 at 12:39:11AM +0200, Laurent Pinchart wrote:
> > > > On Thu, Nov 19, 2026 at 12:32:15AM +0200, Ville Syrjälä wrote:
> > > > > On Thu, Nov 19, 2026 at 12:11:14AM +0200, Laurent Pinchart wrote:
> > > > > > On Wed, Nov 18, 2026 at 11:45:17PM +0200, Ville Syrjala wrote:
> > > > > > > From: Ville Syrjälä <ville.syrjala@linux.intel.com>
> > > > > > >
> > > > > > > Make DRM_RCAR_VSP selectable with COMPILE_TEST=y, and enable the hidden
> > > > > > > DRM_RCAR_WRITEBACK option by default, so the VSP and writeback code gets
> > > > > > > built for greater coverage. Builds fine on x86, x86_64, arm and arm64 at
> > > > > > > least.
> > > > > >
> > > > > > Let's see what build bots will report on more exotic architectures. I
> > > > > > may volunteer you to send a fix if I'm too busy when the builders start
> > > > > > yelling :-)
> > > > > >
> > > > > > > Cc: Laurent Pinchart <laurent.pinchart+renesas@ideasonboard.com>
> > > > > > > Cc: Tomi Valkeinen <tomi.valkeinen+renesas@ideasonboard.com>
> > > > > > > Cc: Kieran Bingham <kieran.bingham+renesas@ideasonboard.com>
> > > > > > > Cc: Geert Uytterhoeven <geert+renesas@glider.be>
> > > > > > > Cc: Magnus Damm <magnus.damm@gmail.com>
> > > > > > > Cc: linux-renesas-soc@vger.kernel.org
> > > > > > > Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
> > > > > > > ---
> > > > > > >  drivers/gpu/drm/renesas/rcar-du/Kconfig | 4 ++--
> > > > > > >  1 file changed, 2 insertions(+), 2 deletions(-)
> > > > > > >
> > > > > > > diff --git a/drivers/gpu/drm/renesas/rcar-du/Kconfig b/drivers/gpu/drm/renesas/rcar-du/Kconfig
> > > > > > > index 80bd770ae0f6..5394f131c58f 100644
> > > > > > > --- a/drivers/gpu/drm/renesas/rcar-du/Kconfig
> > > > > > > +++ b/drivers/gpu/drm/renesas/rcar-du/Kconfig
> > > > > > > @@ -76,7 +76,7 @@ config DRM_RCAR_MIPI_DSI
> > > > > > >         select RESET_CONTROLLER
> > > > > > >
> > > > > > >  config DRM_RCAR_VSP
> > > > > > > -       bool "R-Car DU VSP Compositor Support" if ARM
> > > > > > > +       bool "R-Car DU VSP Compositor Support" if ARM || COMPILE_TEST
> > > > > >
> > > > > > This will make it possible to disable the option when ARM64 &&
> > > > > > COMPILE_TEST, which will break operation on Gen3 and up. It would be
> > > > > > best to avoid that.
> > > > >
> > > > > Break operation, as in when running on real hardware?
> > > > > People actually do that with COMPILE_TEST=y kernels?
> >
> > COMPILE_TEST=y kernels should still work.
> >
> > > > Hopefully not in production, but that may be the case during development
> > > > (I don't though, there are enough kernel options to deal with without
> > > > COMPILE_TEST).
> > > >
> > > > This should do:
> > > >
> > > >       bool "R-Car DU VSP Compositor Support" if ARM || (COMPILE_TEST && !ARM64)
> > > >
> > > > (untested though). Do you think it brings too much complexity ?
> > >
> > > Doesn't bother me.
> >
> > It does bother me, as it breaks the safeguards encoded in Kconfig rules.
> > I think it would be better if DRM_RCAR_DU would select DRM_RCAR_VSP
> > if ARM64, so DRM_RCAR_VSP cannot be disabled where it is needed.
>
> That's possibly an option too. But does
>
>         bool "R-Car DU VSP Compositor Support" if ARM || (COMPILE_TEST && !ARM64)
>
> break the safeguard ?

No it does not. But it is more complicated and less readable.

Gr{oetje,eeting}s,

                        Geert

-- 
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds

^ permalink raw reply	[flat|nested] 36+ messages in thread

* Re: [PATCH v4 18/20] drm/ingenic: Allow ingenic-dw-hdmi build with COMPILE_TEST=y
  2026-11-18 21:45 ` [PATCH v4 18/20] drm/ingenic: Allow ingenic-dw-hdmi build with COMPILE_TEST=y Ville Syrjala
@ 2026-11-19  9:31   ` Paul Cercueil
  0 siblings, 0 replies; 36+ messages in thread
From: Paul Cercueil @ 2026-11-19  9:31 UTC (permalink / raw)
  To: Ville Syrjala, dri-devel; +Cc: linux-mips

Hi Ville,

Le mercredi 18 novembre 2026 à 23:45 +0200, Ville Syrjala a écrit :
> From: Ville Syrjälä <ville.syrjala@linux.intel.com>
> 
> Allow DRM_INGENIC_DW_HDMI to be built with COMPILE_TEST=y for greater
> coverage. The rest of the ingenic driver already allows this. Builds
> fine on x86, x86_64, arm and arm64 at least.
> 
> Cc: Paul Cercueil <paul@crapouillou.net>
> Cc: linux-mips@vger.kernel.org
> Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>

Reviewed-by: Paul Cercueil <paul@crapouillou.net>

Cheers,
-Paul

> ---
>  drivers/gpu/drm/ingenic/Kconfig | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/drivers/gpu/drm/ingenic/Kconfig b/drivers/gpu/drm/ingenic/Kconfig
> index 987ed42f6f7a..60bff9ebee5d 100644
> --- a/drivers/gpu/drm/ingenic/Kconfig
> +++ b/drivers/gpu/drm/ingenic/Kconfig
> @@ -31,7 +31,7 @@ config DRM_INGENIC_IPU
>  
>  config DRM_INGENIC_DW_HDMI
>  	tristate "Ingenic specific support for Synopsys DW HDMI"
> -	depends on MACH_JZ4780
> +	depends on MACH_JZ4780 || COMPILE_TEST
>  	select DRM_DW_HDMI
>  	help
>  	  Choose this option to enable Synopsys DesignWare HDMI based driver.

^ permalink raw reply	[flat|nested] 36+ messages in thread

* Re: [PATCH v4 04/20] drm/tilcdc: Fix endpoint node names in the legacy panel overlay
  2026-11-18 21:45 ` [PATCH v4 04/20] drm/tilcdc: Fix endpoint node names in the legacy panel overlay Ville Syrjala
@ 2026-11-19  9:32   ` Kory Maincent
  0 siblings, 0 replies; 36+ messages in thread
From: Kory Maincent @ 2026-11-19  9:32 UTC (permalink / raw)
  To: Ville Syrjala, dri-devel; +Cc: Jyri Sarha, Tomi Valkeinen



On 11/18/26 22:45, Ville Syrjala wrote:
> From: Ville Syrjälä <ville.syrjala@linux.intel.com>
> 
> The overlay names its two graph endpoints "endpoint@0" but gives them
> no reg property, so dtc complains with W=1:
> 
>    Warning (unit_address_vs_reg): .../endpoint@0: node has a unit name,
>    but no reg or ranges property
> 
> There is only one endpoint per port, and the overlay refers to the
> nodes through the panel_in and lcd_0 labels, not by name. Use plain
> "endpoint" for the node names, which is the usual form for a single
> endpoint, and drop the bogus unit address. The remote-endpoint
> phandle links between the two nodes are unchanged.

Reviewed-by: Kory Maincent <kory.maincent@bootlin.com>

Thank you!
-- 
Köry Maincent, Bootlin
Embedded Linux and kernel engineering
https://bootlin.com


^ permalink raw reply	[flat|nested] 36+ messages in thread

end of thread, other threads:[~2026-11-19  9:32 UTC | newest]

Thread overview: 36+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2026-11-18 21:44 [PATCH v4 00/20] drm: Increase COMPILE_TEST=y coverage Ville Syrjala
2026-11-18 21:44 ` [PATCH v4 01/20] drm/tilcdc: Allow build without __iowmb() Ville Syrjala
2026-11-18 21:44 ` [PATCH v4 02/20] drm/tilcdc: Drop volatile from the mmio pointer in tilcdc_write64() Ville Syrjala
2026-11-18 21:45 ` [PATCH v4 03/20] drm/tilcdc: Use __le64 for the fallback 64bit write Ville Syrjala
2026-11-18 21:45 ` [PATCH v4 04/20] drm/tilcdc: Fix endpoint node names in the legacy panel overlay Ville Syrjala
2026-11-19  9:32   ` Kory Maincent
2026-11-18 21:45 ` [PATCH v4 05/20] drm/tilcdc: Allow build with COMPILE_TEST=y Ville Syrjala
2026-11-18 21:45 ` [PATCH v4 06/20] drm/armada: Fix printk arguments Ville Syrjala
2026-11-18 21:45 ` [PATCH v4 07/20] drm/armada: Fix armada_debugfs_crtc_reg_write() return type Ville Syrjala
2026-11-18 21:45 ` [PATCH v4 08/20] drm/armada: Fix sparse __iomem address space warnings Ville Syrjala
2026-11-18 21:45 ` [PATCH v4 09/20] drm/armada: Allow build with COMPILE_TEST=y Ville Syrjala
2026-11-18 21:45 ` [PATCH v4 10/20] drm/fsl-dcu: " Ville Syrjala
2026-11-18 21:45 ` [PATCH v4 11/20] drm/adp: Include linux/bitfield.h Ville Syrjala
2026-11-18 21:45 ` [PATCH v4 12/20] drm/adp: Allow build with COMPILE_TEST=y Ville Syrjala
2026-11-18 21:45 ` [PATCH v4 13/20] drm/imagination: Include io-64-nonatomic-lo-hi.h Ville Syrjala
2026-11-18 21:45 ` [PATCH v4 14/20] drm/imagination: Use rcu_access_pointer() for dma_fence::ops Ville Syrjala
2026-11-18 21:45 ` [PATCH v4 15/20] drm/imagination: Allow build with COMPILE_TEST=y Ville Syrjala
2026-11-18 21:58   ` sashiko-bot
2026-11-18 22:30     ` Ville Syrjälä
2026-11-18 21:45 ` [PATCH v4 16/20] drm/imagination: Drop stale kernel-doc for pvr_mmu_backing_page.pvr_dev Ville Syrjala
2026-11-18 21:45 ` [PATCH v4 17/20] drm/imagination: Fix kernel-doc for pvr_mmu_op_context anonymous structs Ville Syrjala
2026-11-18 21:45 ` [PATCH v4 18/20] drm/ingenic: Allow ingenic-dw-hdmi build with COMPILE_TEST=y Ville Syrjala
2026-11-19  9:31   ` Paul Cercueil
2026-11-18 21:45 ` [PATCH v4 19/20] drm/nouveau: Allow platform driver " Ville Syrjala
2026-11-18 21:57   ` sashiko-bot
2026-11-18 22:28     ` Ville Syrjälä
2026-11-18 21:45 ` [PATCH v4 20/20] drm/rcar-du: Allow VSP and writeback " Ville Syrjala
2026-11-18 22:11   ` Laurent Pinchart
2026-11-18 22:32     ` Ville Syrjälä
2026-11-18 22:39       ` Laurent Pinchart
2026-11-18 22:57         ` Ville Syrjälä
2026-11-18 23:14           ` Laurent Pinchart
2026-11-19  7:38           ` Geert Uytterhoeven
2026-11-19  8:16             ` Laurent Pinchart
2026-11-19  8:23               ` Geert Uytterhoeven
2026-11-19  7:33   ` Geert Uytterhoeven

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox