[PATCH v2 0/4] media: qcom: flip the switch between Venus and Iris drivers

[PATCH v2 0/4] media: qcom: flip the switch between Venus and Iris drivers

[Dmitry Baryshkov]

As the Venus and Iris drivers are close to the "feature parity" for the
common platforms (SC7280, SM8250), in order to get more attention to
squashing bugs from the Iris driver, flip the switch and default to the
Iris driver if both are enabled. The Iris driver has several
regressions, but hopefully they can be fixed through the development
cycle by the respective team. Also it is better to fail the test than
crash the device (which Venus driver does a lot).

Note: then intention is to land this in 6.21, which might let us to
drop those platforms from the Venus driver in 6.22+.

Testing methodology: fluster test-suite, single-threaded mode, SM8250
device (RB5).

$ ./fluster.py r -j 1 -d  GStreamer-H.264-V4L2-Gst1.0 -s -so out-h264.md -ts JVT-AVC_V1
$ ./fluster.py r -j 1 -d  GStreamer-H.265-V4L2-Gst1.0 -s -so out-h265.md -ts JCT-VC-HEVC_V1
$ ./fluster.py r -j 1 -d  GStreamer-VP9-V4L2-Gst1.0 -s -so out-vp9.md

For H.264:

Only the main set was used on both platforms

On Venus pass 126/135
On Iris  pass 78/135

- Fail on Iris, but pass on Venus:
  cabac_mot_fld0_full, cabac_mot_mbaff0_full, cabac_mot_picaff0_full,
  CAFI1_SVA_C, CAMA1_Sony_C, cama1_vtc_c, cama2_vtc_b, cama3_vtc_b,
  CAMACI3_Sony_C, CAMASL3_Sony_B, CAMP_MOT_MBAFF_L30,
  CAMP_MOT_MBAFF_L31, CANLMA2_Sony_C, CANLMA3_Sony_C,
  cavlc_mot_fld0_full_B, cavlc_mot_mbaff0_full_B,
  cavlc_mot_picaff0_full_B, CVCANLMA2_Sony_C, CVFI1_Sony_D, CVFI1_SVA_C,
  CVFI2_Sony_H, CVFI2_SVA_C, CVMA1_Sony_D, CVMAPAQP3_Sony_E,
  CVMAQP2_Sony_G, CVMAQP3_Sony_D, CVMP_MOT_FLD_L30_B,
  CVMP_MOT_FRM_L31_B, CVNLFI1_Sony_C, CVNLFI2_Sony_H, FI1_Sony_E,
  MR9_BT_B, Sharp_MP_Field_1_B, Sharp_MP_Field_2_B, Sharp_MP_Field_3_B,
  Sharp_MP_PAFF_1r2, Sharp_MP_PAFF_2r

- Fail on Venus, but pass on Iris:
  BA3_SVA_C

For Venus several tests crash the firmware or crash the device,
so test-suite was executed with extra:

-sv CABREF3_Sand_D CAMANL1_TOSHIBA_B CAMA1_TOSHIBA_B \
  CAMA3_Sand_E CAMANL2_TOSHIBA_B CAMANL3_Sand_E CAPA1_TOSHIBA_B \
  CAPAMA3_Sand_F CVMA1_TOSHIBA_B CVMANL1_TOSHIBA_B CVMANL2_TOSHIBA_B \
  CVPA1_TOSHIBA_B

For H.265:

Only the main set was used on both platforms

On Venus pass 135/147
On Iris  pass 133/147

- Fail on Iris, but pass on Venus:
  INITQP_B_Main10_Sony_1, WP_A_MAIN10_Toshiba_3,
  WPP_A_ericsson_MAIN10_2, WPP_B_ericsson_MAIN10_2,
  WPP_C_ericsson_MAIN10_2, WPP_E_ericsson_MAIN10_2,
  WPP_F_ericsson_MAIN10_2

- Fail on Venus, but pass on Iris:
  ipcm_A_NEC_3, ipcm_B_NEC_3, ipcm_C_NEC_3, ipcm_D_NEC_3, ipcm_E_NEC_2,
  IPRED_B_Nokia_3, VPSSPSPPS_A_MainConcept_1

For VP9 codec:

On Venus pass 174/311
On Iris  pass 232/311

- Fail on Venus, but pass on Iris:
  vp90-2-05-resize.ivf, vp90-2-14-resize-10frames-fp-tiles-1-2-4-8,
  vp90-2-14-resize-10frames-fp-tiles-1-2,
  vp90-2-14-resize-10frames-fp-tiles-1-4,
  vp90-2-14-resize-10frames-fp-tiles-1-8,
  vp90-2-14-resize-10frames-fp-tiles-2-1,
  vp90-2-14-resize-10frames-fp-tiles-2-4,
  vp90-2-14-resize-10frames-fp-tiles-2-8,
  vp90-2-14-resize-10frames-fp-tiles-4-1,
  vp90-2-14-resize-10frames-fp-tiles-4-2,
  vp90-2-14-resize-10frames-fp-tiles-4-8,
  vp90-2-14-resize-10frames-fp-tiles-8-1,
  vp90-2-14-resize-10frames-fp-tiles-8-2,
  vp90-2-14-resize-10frames-fp-tiles-8-4-2-1,
  vp90-2-14-resize-10frames-fp-tiles-8-4, vp90-2-14-resize-fp-tiles-1-16,
  vp90-2-14-resize-fp-tiles-1-2-4-8-16, vp90-2-14-resize-fp-tiles-1-2,
  vp90-2-14-resize-fp-tiles-1-4, vp90-2-14-resize-fp-tiles-16-1,
  vp90-2-14-resize-fp-tiles-16-2, vp90-2-14-resize-fp-tiles-16-4,
  vp90-2-14-resize-fp-tiles-16-8-4-2-1, vp90-2-14-resize-fp-tiles-16-8,
  vp90-2-14-resize-fp-tiles-1-8, vp90-2-14-resize-fp-tiles-2-16,
  vp90-2-14-resize-fp-tiles-2-1, vp90-2-14-resize-fp-tiles-2-4,
  vp90-2-14-resize-fp-tiles-2-8, vp90-2-14-resize-fp-tiles-4-16,
  vp90-2-14-resize-fp-tiles-4-1, vp90-2-14-resize-fp-tiles-4-2,
  vp90-2-14-resize-fp-tiles-4-8, vp90-2-14-resize-fp-tiles-8-16,
  vp90-2-14-resize-fp-tiles-8-1, vp90-2-14-resize-fp-tiles-8-2,
  vp90-2-14-resize-fp-tiles-8-4, vp90-2-15-segkey,
  vp90-2-21-resize_inter_1280x720_5_1-2, vp90-2-21-resize_inter_1280x720_5_3-4,
  vp90-2-21-resize_inter_1280x720_7_1-2, vp90-2-21-resize_inter_1280x720_7_3-4,
  vp90-2-21-resize_inter_1920x1080_5_1-2,
  vp90-2-21-resize_inter_1920x1080_5_3-4,
  vp90-2-21-resize_inter_1920x1080_7_1-2,
  vp90-2-21-resize_inter_1920x1080_7_3-4,
  vp90-2-21-resize_inter_320x180_5_3-4, vp90-2-21-resize_inter_320x180_7_3-4,
  vp90-2-21-resize_inter_320x240_5_3-4, vp90-2-21-resize_inter_320x240_7_3-4,
  vp90-2-21-resize_inter_640x360_5_1-2, vp90-2-21-resize_inter_640x360_5_3-4,
  vp90-2-21-resize_inter_640x360_7_1-2, vp90-2-21-resize_inter_640x360_7_3-4,
  vp90-2-21-resize_inter_640x480_5_1-2, vp90-2-21-resize_inter_640x480_5_3-4,
  vp90-2-21-resize_inter_640x480_7_1-2, vp90-2-21-resize_inter_640x480_7_3-4

Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com>
---
Changes in v2:
- Sorted out SC7280 Chromebook bindings (Stephan)
- Dropped most of ifdef from venus (Vikash, Konrad)
- Added testign results to the cover letter.
- Link to v1: https://lore.kernel.org/r/20251119-venus-iris-flip-switch-v1-1-852369f66e36@oss.qualcomm.com

---
Dmitry Baryshkov (4):
      arm64: dts: qcom: sc7280-chrome-common: disable Venus
      media: dt-bindings: qcom,sc7280-venus: drop non-PAS support
      media: dt-bindings: qcom-sc7180-venus: move video-firmware here
      media: qcom: venus: flip the venus/iris switch

 .../devicetree/bindings/media/qcom,sc7180-venus.yaml      | 15 +++++++++++++++
 .../devicetree/bindings/media/qcom,sc7280-venus.yaml      | 10 ++--------
 .../devicetree/bindings/media/qcom,venus-common.yaml      | 15 ---------------
 arch/arm64/boot/dts/qcom/sc7280-chrome-common.dtsi        | 11 -----------
 drivers/media/platform/qcom/iris/iris_probe.c             |  4 ----
 drivers/media/platform/qcom/venus/core.c                  |  6 ++++++
 drivers/media/platform/qcom/venus/core.h                  | 11 +++++++++++
 7 files changed, 34 insertions(+), 38 deletions(-)
---
base-commit: ca3a02fda4da8e2c1cb6baee5d72352e9e2cfaea
change-id: 20251119-venus-iris-flip-switch-d59a3fbc6a4b
prerequisite-change-id: 20260120-iris-sc8280xp-85d13bc60536:v3
prerequisite-patch-id: d34f95892aabfca4695d8b41c4001a23688e72ad
prerequisite-patch-id: 0dd97291d4aa94feb2e915a6a6324ac0031deca8
prerequisite-patch-id: 9de15d81588208683f296b6d48363226df41fbd6
prerequisite-patch-id: 6a89a4c1709b0b543a7d2a03954b7a628f549aa5
prerequisite-patch-id: 668a6b7fec7ed8302ab23c65ad76481be06847f6
prerequisite-patch-id: e423b8d12b47c7414e694625204977e37414ce4e
prerequisite-patch-id: e5be869bdd8a73b5c3d3407b1a47c66c944dc832

Best regards,
-- 
With best wishes
Dmitry

[PATCH v2 1/4] arm64: dts: qcom: sc7280-chrome-common: disable Venus

[Dmitry Baryshkov]

Iris driver doesn't support non-PAS/TZ setup for setting up the Iris
core. It's expected that the ABI that is going to be used in a long-term
would be different from the defined venus-firmware subnode.

Granted that the SC7280 Chromebooks were cancelled before reaching wide
audience and granted the feature ABI changes, drop venus configuration
for SC7280 ChromeOS devices.

Suggested-by: Stephan Gerhold <stephan.gerhold@linaro.org>
Cc: Douglas Anderson <dianders@chromium.org>
Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com>
---
 arch/arm64/boot/dts/qcom/sc7280-chrome-common.dtsi | 11 -----------
 1 file changed, 11 deletions(-)

diff --git a/arch/arm64/boot/dts/qcom/sc7280-chrome-common.dtsi b/arch/arm64/boot/dts/qcom/sc7280-chrome-common.dtsi
index 617a39d32488..debf62baec9b 100644
--- a/arch/arm64/boot/dts/qcom/sc7280-chrome-common.dtsi
+++ b/arch/arm64/boot/dts/qcom/sc7280-chrome-common.dtsi
@@ -140,17 +140,6 @@ &scm {
 	dma-coherent;
 };
 
-&venus {
-	iommus = <&apps_smmu 0x2180 0x20>,
-		 <&apps_smmu 0x2184 0x20>;
-
-	status = "okay";
-
-	video-firmware {
-		iommus = <&apps_smmu 0x21a2 0x0>;
-	};
-};
-
 &watchdog {
 	status = "okay";
 };

-- 
2.47.3

[PATCH v2 2/4] media: dt-bindings: qcom,sc7280-venus: drop non-PAS support

[Dmitry Baryshkov]

The only users of the non-PAS setup on SC7280 platform are the ChromeOS
devices, which were cancelled before reaching end users. Iris, the
alternative driver for the same hardware, does not support non-PAS
setup. It is expected that in future both Venus and Iris devices will
use different ABI for non-PAS (EL2) setup.

In order to declare only the future-proof hardware description drop
support for non-PAS setup from the SC7280 Venus schema (breaking almost
non-existing SC7280 ChromeOS devices).

Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com>
---
 Documentation/devicetree/bindings/media/qcom,sc7280-venus.yaml | 10 ++--------
 1 file changed, 2 insertions(+), 8 deletions(-)

diff --git a/Documentation/devicetree/bindings/media/qcom,sc7280-venus.yaml b/Documentation/devicetree/bindings/media/qcom,sc7280-venus.yaml
index 413c5b4ee650..9725fcb761dc 100644
--- a/Documentation/devicetree/bindings/media/qcom,sc7280-venus.yaml
+++ b/Documentation/devicetree/bindings/media/qcom,sc7280-venus.yaml
@@ -43,8 +43,7 @@ properties:
       - const: vcodec_bus
 
   iommus:
-    minItems: 1
-    maxItems: 2
+    maxItems: 1
 
   interconnects:
     maxItems: 2
@@ -120,12 +119,7 @@ examples:
                         <&mmss_noc MASTER_VIDEO_P0 0 &mc_virt SLAVE_EBI1 0>;
         interconnect-names = "cpu-cfg", "video-mem";
 
-        iommus = <&apps_smmu 0x2180 0x20>,
-                 <&apps_smmu 0x2184 0x20>;
+        iommus = <&apps_smmu 0x2180 0x20>;
 
         memory-region = <&video_mem>;
-
-        video-firmware {
-            iommus = <&apps_smmu 0x21a2 0x0>;
-        };
     };

-- 
2.47.3

[PATCH v2 3/4] media: dt-bindings: qcom-sc7180-venus: move video-firmware here

[Dmitry Baryshkov]

As SC7180 is the only remaining user of the non-TZ / non-PAS setup which
uses the video-firmware subnode, move its definition from the common
schema to the SC7180-specific one.

Future platforms that are going to support non-TZ setup will use
different semantics and different DT ABI.

Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com>
---
 .../devicetree/bindings/media/qcom,sc7180-venus.yaml      | 15 +++++++++++++++
 .../devicetree/bindings/media/qcom,venus-common.yaml      | 15 ---------------
 2 files changed, 15 insertions(+), 15 deletions(-)

diff --git a/Documentation/devicetree/bindings/media/qcom,sc7180-venus.yaml b/Documentation/devicetree/bindings/media/qcom,sc7180-venus.yaml
index bfd8b1ad4731..b21bed314848 100644
--- a/Documentation/devicetree/bindings/media/qcom,sc7180-venus.yaml
+++ b/Documentation/devicetree/bindings/media/qcom,sc7180-venus.yaml
@@ -91,6 +91,21 @@ properties:
     deprecated: true
     additionalProperties: false
 
+  video-firmware:
+    type: object
+    additionalProperties: false
+
+    description: |
+      Firmware subnode is needed when the platform does not
+      have TrustZone.
+
+    properties:
+      iommus:
+        maxItems: 1
+
+    required:
+      - iommus
+
 required:
   - compatible
   - power-domain-names
diff --git a/Documentation/devicetree/bindings/media/qcom,venus-common.yaml b/Documentation/devicetree/bindings/media/qcom,venus-common.yaml
index 3153d91f9d18..59a3fde846d2 100644
--- a/Documentation/devicetree/bindings/media/qcom,venus-common.yaml
+++ b/Documentation/devicetree/bindings/media/qcom,venus-common.yaml
@@ -47,21 +47,6 @@ properties:
     minItems: 1
     maxItems: 4
 
-  video-firmware:
-    type: object
-    additionalProperties: false
-
-    description: |
-      Firmware subnode is needed when the platform does not
-      have TrustZone.
-
-    properties:
-      iommus:
-        maxItems: 1
-
-    required:
-      - iommus
-
 required:
   - reg
   - clocks

-- 
2.47.3

[PATCH v2 4/4] media: qcom: venus: flip the venus/iris switch

[Dmitry Baryshkov]

With the Iris and Venus driver having more or less feature parity for
"HFI 6xx" platforms and with Iris gaining support for SC7280, flip the
switch. Use Iris by default for SM8250 and SC7280, the platforms which
are supported by both drivers, and use Venus only if Iris is not
compiled at all. Use IS_ENABLED to strip out the code and data
structures which are used by the disabled platforms.

Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com>
---
 drivers/media/platform/qcom/iris/iris_probe.c |  4 ----
 drivers/media/platform/qcom/venus/core.c      |  6 ++++++
 drivers/media/platform/qcom/venus/core.h      | 11 +++++++++++
 3 files changed, 17 insertions(+), 4 deletions(-)

diff --git a/drivers/media/platform/qcom/iris/iris_probe.c b/drivers/media/platform/qcom/iris/iris_probe.c
index 10b00d9808d2..3b662ba90259 100644
--- a/drivers/media/platform/qcom/iris/iris_probe.c
+++ b/drivers/media/platform/qcom/iris/iris_probe.c
@@ -352,22 +352,18 @@ static const struct of_device_id iris_dt_match[] = {
 		.compatible = "qcom,qcs8300-iris",
 		.data = &qcs8300_data,
 	},
-#if (!IS_ENABLED(CONFIG_VIDEO_QCOM_VENUS))
 	{
 		.compatible = "qcom,sc7280-venus",
 		.data = &sc7280_data,
 	},
-#endif
 	{
 		.compatible = "qcom,sc8280xp-venus",
 		.data = &sc8280xp_data,
 	},
-#if (!IS_ENABLED(CONFIG_VIDEO_QCOM_VENUS))
 	{
 		.compatible = "qcom,sm8250-venus",
 		.data = &sm8250_data,
 	},
-#endif
 	{
 		.compatible = "qcom,sm8350-venus",
 		.data = &sm8350_data,
diff --git a/drivers/media/platform/qcom/venus/core.c b/drivers/media/platform/qcom/venus/core.c
index 24d2b2fd0340..646dae3407b4 100644
--- a/drivers/media/platform/qcom/venus/core.c
+++ b/drivers/media/platform/qcom/venus/core.c
@@ -949,6 +949,7 @@ static const struct venus_resources sc7180_res = {
 	.enc_nodename = "video-encoder",
 };
 
+#if (!IS_ENABLED(CONFIG_VIDEO_QCOM_IRIS))
 static const struct freq_tbl sm8250_freq_table[] = {
 	{ 0, 444000000 },
 	{ 0, 366000000 },
@@ -1069,6 +1070,7 @@ static const struct venus_resources sc7280_res = {
 	.dec_nodename = "video-decoder",
 	.enc_nodename = "video-encoder",
 };
+#endif
 
 static const struct bw_tbl qcm2290_bw_table_dec[] = {
 	{ 352800, 597000, 0, 746000, 0 }, /* 1080p@30 + 720p@30 */
@@ -1125,11 +1127,15 @@ static const struct of_device_id venus_dt_match[] = {
 	{ .compatible = "qcom,msm8998-venus", .data = &msm8998_res, },
 	{ .compatible = "qcom,qcm2290-venus", .data = &qcm2290_res, },
 	{ .compatible = "qcom,sc7180-venus", .data = &sc7180_res, },
+#if (!IS_ENABLED(CONFIG_VIDEO_QCOM_IRIS))
 	{ .compatible = "qcom,sc7280-venus", .data = &sc7280_res, },
+#endif
 	{ .compatible = "qcom,sdm660-venus", .data = &sdm660_res, },
 	{ .compatible = "qcom,sdm845-venus", .data = &sdm845_res, },
 	{ .compatible = "qcom,sdm845-venus-v2", .data = &sdm845_res_v2, },
+#if (!IS_ENABLED(CONFIG_VIDEO_QCOM_IRIS))
 	{ .compatible = "qcom,sm8250-venus", .data = &sm8250_res, },
+#endif
 	{ }
 };
 MODULE_DEVICE_TABLE(of, venus_dt_match);
diff --git a/drivers/media/platform/qcom/venus/core.h b/drivers/media/platform/qcom/venus/core.h
index 7506f5d0f609..c7acacaa53b8 100644
--- a/drivers/media/platform/qcom/venus/core.h
+++ b/drivers/media/platform/qcom/venus/core.h
@@ -54,8 +54,10 @@ enum vpu_version {
 	VPU_VERSION_AR50,
 	VPU_VERSION_AR50_LITE,
 	VPU_VERSION_IRIS1,
+#if (!IS_ENABLED(CONFIG_VIDEO_QCOM_IRIS))
 	VPU_VERSION_IRIS2,
 	VPU_VERSION_IRIS2_1,
+#endif
 };
 
 struct firmware_version {
@@ -525,13 +527,22 @@ struct venus_inst {
 #define IS_V1(core)	((core)->res->hfi_version == HFI_VERSION_1XX)
 #define IS_V3(core)	((core)->res->hfi_version == HFI_VERSION_3XX)
 #define IS_V4(core)	((core)->res->hfi_version == HFI_VERSION_4XX)
+#if (!IS_ENABLED(CONFIG_VIDEO_QCOM_IRIS))
 #define IS_V6(core)	((core)->res->hfi_version == HFI_VERSION_6XX)
+#else
+#define IS_V6(core)	(0)
+#endif
 
 #define IS_AR50(core)		((core)->res->vpu_version == VPU_VERSION_AR50)
 #define IS_AR50_LITE(core)	((core)->res->vpu_version == VPU_VERSION_AR50_LITE)
 #define IS_IRIS1(core)		((core)->res->vpu_version == VPU_VERSION_IRIS1)
+#if (!IS_ENABLED(CONFIG_VIDEO_QCOM_IRIS))
 #define IS_IRIS2(core)		((core)->res->vpu_version == VPU_VERSION_IRIS2)
 #define IS_IRIS2_1(core)	((core)->res->vpu_version == VPU_VERSION_IRIS2_1)
+#else
+#define IS_IRIS2(core)		(0)
+#define IS_IRIS2_1(core)	(0)
+#endif
 
 static inline bool is_lite(struct venus_core *core)
 {

-- 
2.47.3

Re: [PATCH v2 1/4] arm64: dts: qcom: sc7280-chrome-common: disable Venus

[Doug Anderson]

Hi,

On Sun, Jan 25, 2026 at 6:09 PM Dmitry Baryshkov
<dmitry.baryshkov@oss.qualcomm.com> wrote:
>
> Iris driver doesn't support non-PAS/TZ setup for setting up the Iris
> core. It's expected that the ABI that is going to be used in a long-term
> would be different from the defined venus-firmware subnode.
>
> Granted that the SC7280 Chromebooks were cancelled before reaching wide
> audience and granted the feature ABI changes, drop venus configuration
> for SC7280 ChromeOS devices.
>
> Suggested-by: Stephan Gerhold <stephan.gerhold@linaro.org>
> Cc: Douglas Anderson <dianders@chromium.org>
> Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com>
> ---
>  arch/arm64/boot/dts/qcom/sc7280-chrome-common.dtsi | 11 -----------
>  1 file changed, 11 deletions(-)

As far as I'm aware, anyone still using a sc7280 Chromebook dev
platform doesn't care about the video encoder / decoder working.

Acked-by: Douglas Anderson <dianders@chromium.org>

Re: [PATCH v2 1/4] arm64: dts: qcom: sc7280-chrome-common: disable Venus

[Konrad Dybcio]

On 1/26/26 3:09 AM, Dmitry Baryshkov wrote:
> Iris driver doesn't support non-PAS/TZ setup for setting up the Iris
> core. It's expected that the ABI that is going to be used in a long-term
> would be different from the defined venus-firmware subnode.
> 
> Granted that the SC7280 Chromebooks were cancelled before reaching wide
> audience and granted the feature ABI changes, drop venus configuration
> for SC7280 ChromeOS devices.
> 
> Suggested-by: Stephan Gerhold <stephan.gerhold@linaro.org>
> Cc: Douglas Anderson <dianders@chromium.org>
> Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com>
> ---

Reviewed-by: Konrad Dybcio <konrad.dybcio@oss.qualcomm.com>

Konrad

Re: [PATCH v2 2/4] media: dt-bindings: qcom,sc7280-venus: drop non-PAS support

[Konrad Dybcio]

On 1/26/26 3:09 AM, Dmitry Baryshkov wrote:
> The only users of the non-PAS setup on SC7280 platform are the ChromeOS
> devices, which were cancelled before reaching end users. Iris, the
> alternative driver for the same hardware, does not support non-PAS
> setup. It is expected that in future both Venus and Iris devices will
> use different ABI for non-PAS (EL2) setup.
> 
> In order to declare only the future-proof hardware description drop
> support for non-PAS setup from the SC7280 Venus schema (breaking almost
> non-existing SC7280 ChromeOS devices).
> 
> Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com>
> ---

The commit message could highlight the *actual* problem ("iommus"
entries aren't treated like separate entities which is exactly what
we need to avoid for this complex hw)

Konrad

Re: [PATCH v2 3/4] media: dt-bindings: qcom-sc7180-venus: move video-firmware here

[Konrad Dybcio]

On 1/26/26 3:09 AM, Dmitry Baryshkov wrote:
> As SC7180 is the only remaining user of the non-TZ / non-PAS setup which
> uses the video-firmware subnode, move its definition from the common
> schema to the SC7180-specific one.
> 
> Future platforms that are going to support non-TZ setup will use
> different semantics and different DT ABI.
> 
> Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com>
> ---

It's worth highlighting that this set of properties is not enough to
accurately describe the full capabilities of the hardware and will
likely one day be replaced with iommu-map anyway

Konrad

Re: [PATCH v2 0/4] media: qcom: flip the switch between Venus and Iris drivers

[Konrad Dybcio]

On 1/26/26 3:09 AM, Dmitry Baryshkov wrote:
> As the Venus and Iris drivers are close to the "feature parity" for the
> common platforms (SC7280, SM8250), in order to get more attention to
> squashing bugs from the Iris driver, flip the switch and default to the
> Iris driver if both are enabled. The Iris driver has several
> regressions, but hopefully they can be fixed through the development
> cycle by the respective team. Also it is better to fail the test than
> crash the device (which Venus driver does a lot).
> 
> Note: then intention is to land this in 6.21, which might let us to
> drop those platforms from the Venus driver in 6.22+.
> 
> Testing methodology: fluster test-suite, single-threaded mode, SM8250
> device (RB5).
> 
> $ ./fluster.py r -j 1 -d  GStreamer-H.264-V4L2-Gst1.0 -s -so out-h264.md -ts JVT-AVC_V1
> $ ./fluster.py r -j 1 -d  GStreamer-H.265-V4L2-Gst1.0 -s -so out-h265.md -ts JCT-VC-HEVC_V1
> $ ./fluster.py r -j 1 -d  GStreamer-VP9-V4L2-Gst1.0 -s -so out-vp9.md
> 
> For H.264:
> 
> Only the main set was used on both platforms
> 
> On Venus pass 126/135
> On Iris  pass 78/135

This sounds a little bad..

[...]

> For Venus several tests crash the firmware or crash the device,
> so test-suite was executed with extra:

But this sounds way way worse

[...]

> For H.265:
> 
> Only the main set was used on both platforms
> 
> On Venus pass 135/147
> On Iris  pass 133/147

[...]

> For VP9 codec:
> 
> On Venus pass 174/311
> On Iris  pass 232/311

These two give me hope..

Ultimately, I think it's the right thing to do, if only to prevent the
crashes you've described..

Konrad

Re: [PATCH v2 4/4] media: qcom: venus: flip the venus/iris switch

[Konrad Dybcio]

On 1/26/26 3:09 AM, Dmitry Baryshkov wrote:
> With the Iris and Venus driver having more or less feature parity for
> "HFI 6xx" platforms and with Iris gaining support for SC7280, flip the
> switch. Use Iris by default for SM8250 and SC7280, the platforms which
> are supported by both drivers, and use Venus only if Iris is not
> compiled at all. Use IS_ENABLED to strip out the code and data
> structures which are used by the disabled platforms.
> 
> Signed-off-by: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com>
> ---

Reviewed-by: Konrad Dybcio <konrad.dybcio@oss.qualcomm.com>

Konrad

Re: [PATCH v2 0/4] media: qcom: flip the switch between Venus and Iris drivers

[Dmitry Baryshkov]

On Tue, Jan 27, 2026 at 12:43:36PM +0100, Konrad Dybcio wrote:
> On 1/26/26 3:09 AM, Dmitry Baryshkov wrote:
> > As the Venus and Iris drivers are close to the "feature parity" for the
> > common platforms (SC7280, SM8250), in order to get more attention to
> > squashing bugs from the Iris driver, flip the switch and default to the
> > Iris driver if both are enabled. The Iris driver has several
> > regressions, but hopefully they can be fixed through the development
> > cycle by the respective team. Also it is better to fail the test than
> > crash the device (which Venus driver does a lot).
> > 
> > Note: then intention is to land this in 6.21, which might let us to
> > drop those platforms from the Venus driver in 6.22+.
> > 
> > Testing methodology: fluster test-suite, single-threaded mode, SM8250
> > device (RB5).
> > 
> > $ ./fluster.py r -j 1 -d  GStreamer-H.264-V4L2-Gst1.0 -s -so out-h264.md -ts JVT-AVC_V1
> > $ ./fluster.py r -j 1 -d  GStreamer-H.265-V4L2-Gst1.0 -s -so out-h265.md -ts JCT-VC-HEVC_V1
> > $ ./fluster.py r -j 1 -d  GStreamer-VP9-V4L2-Gst1.0 -s -so out-vp9.md
> > 
> > For H.264:
> > 
> > Only the main set was used on both platforms
> > 
> > On Venus pass 126/135
> > On Iris  pass 78/135
> 
> This sounds a little bad..

I'd let Vikash comment here, we were promised feature parity for Venus /
Iris drivers before accepting the latter one.

The list of tests passing for Venus, but failing for Iris:

> |BA3_SVA_C|✔️|
< |cabac_mot_fld0_full|✔️|
< |cabac_mot_mbaff0_full|✔️|
< |cabac_mot_picaff0_full|✔️|
< |CAFI1_SVA_C|✔️|
< |CAMA1_Sony_C|✔️|
< |cama1_vtc_c|✔️|
< |cama2_vtc_b|✔️|
< |cama3_vtc_b|✔️|
< |CAMACI3_Sony_C|✔️|
< |CAMASL3_Sony_B|✔️|
< |CAMP_MOT_MBAFF_L30|✔️|
< |CAMP_MOT_MBAFF_L31|✔️|
< |CANLMA2_Sony_C|✔️|
< |CANLMA3_Sony_C|✔️|
< |cavlc_mot_fld0_full_B|✔️|
< |cavlc_mot_mbaff0_full_B|✔️|
< |cavlc_mot_picaff0_full_B|✔️|
< |CVCANLMA2_Sony_C|✔️|
< |CVFI1_Sony_D|✔️|
< |CVFI1_SVA_C|✔️|
< |CVFI2_Sony_H|✔️|
< |CVFI2_SVA_C|✔️|
< |CVMA1_Sony_D|✔️|
< |CVMAPAQP3_Sony_E|✔️|
< |CVMAQP2_Sony_G|✔️|
< |CVMAQP3_Sony_D|✔️|
< |CVMP_MOT_FLD_L30_B|✔️|
< |CVMP_MOT_FRM_L31_B|✔️|
< |CVNLFI1_Sony_C|✔️|
< |CVNLFI2_Sony_H|✔️|
< |FI1_Sony_E|✔️|
< |MR9_BT_B|✔️|
< |Sharp_MP_Field_1_B|✔️|
< |Sharp_MP_Field_2_B|✔️|
< |Sharp_MP_Field_3_B|✔️|
< |Sharp_MP_PAFF_1r2|✔️|
< |Sharp_MP_PAFF_2r|✔️|


> 
> [...]
> 
> > For Venus several tests crash the firmware or crash the device,
> > so test-suite was executed with extra:
> 
> But this sounds way way worse
> 
> [...]
> 
> > For H.265:
> > 
> > Only the main set was used on both platforms
> > 
> > On Venus pass 135/147
> > On Iris  pass 133/147
> 
> [...]
> 
> > For VP9 codec:
> > 
> > On Venus pass 174/311
> > On Iris  pass 232/311
> 
> These two give me hope..
> 
> Ultimately, I think it's the right thing to do, if only to prevent the
> crashes you've described..

Yep.

-- 
With best wishes
Dmitry

Re: [PATCH v2 0/4] media: qcom: flip the switch between Venus and Iris drivers

[Vikash Garodia]

On 1/27/2026 5:35 PM, Dmitry Baryshkov wrote:
> On Tue, Jan 27, 2026 at 12:43:36PM +0100, Konrad Dybcio wrote:
>> On 1/26/26 3:09 AM, Dmitry Baryshkov wrote:
>>> As the Venus and Iris drivers are close to the "feature parity" for the
>>> common platforms (SC7280, SM8250), in order to get more attention to
>>> squashing bugs from the Iris driver, flip the switch and default to the
>>> Iris driver if both are enabled. The Iris driver has several
>>> regressions, but hopefully they can be fixed through the development
>>> cycle by the respective team. Also it is better to fail the test than
>>> crash the device (which Venus driver does a lot).
>>>
>>> Note: then intention is to land this in 6.21, which might let us to
>>> drop those platforms from the Venus driver in 6.22+.
>>>
>>> Testing methodology: fluster test-suite, single-threaded mode, SM8250
>>> device (RB5).
>>>
>>> $ ./fluster.py r -j 1 -d  GStreamer-H.264-V4L2-Gst1.0 -s -so out-h264.md -ts JVT-AVC_V1
>>> $ ./fluster.py r -j 1 -d  GStreamer-H.265-V4L2-Gst1.0 -s -so out-h265.md -ts JCT-VC-HEVC_V1
>>> $ ./fluster.py r -j 1 -d  GStreamer-VP9-V4L2-Gst1.0 -s -so out-vp9.md
>>>
>>> For H.264:
>>>
>>> Only the main set was used on both platforms
>>>
>>> On Venus pass 126/135
>>> On Iris  pass 78/135
>>
>> This sounds a little bad..
> 
> I'd let Vikash comment here, we were promised feature parity for Venus /
> Iris drivers before accepting the latter one.
> 
> The list of tests passing for Venus, but failing for Iris:
> 
>> |BA3_SVA_C|✔️|

This need to be checked. Similar failure was seen in SM8550 due to 
unssuported profile V4L2_MPEG_VIDEO_H264_PROFILE_EXTENDED.

> < |cabac_mot_fld0_full|✔️|
> < |cabac_mot_mbaff0_full|✔️|
> < |cabac_mot_picaff0_full|✔️|
> < |CAFI1_SVA_C|✔️|
> < |CAMA1_Sony_C|✔️|
> < |cama1_vtc_c|✔️|
> < |cama2_vtc_b|✔️|
> < |cama3_vtc_b|✔️|
> < |CAMACI3_Sony_C|✔️|
> < |CAMASL3_Sony_B|✔️|
> < |CAMP_MOT_MBAFF_L30|✔️|
> < |CAMP_MOT_MBAFF_L31|✔️|
> < |CANLMA2_Sony_C|✔️|
> < |CANLMA3_Sony_C|✔️|
> < |cavlc_mot_fld0_full_B|✔️|
> < |cavlc_mot_mbaff0_full_B|✔️|
> < |cavlc_mot_picaff0_full_B|✔️|
> < |CVCANLMA2_Sony_C|✔️|
> < |CVFI1_Sony_D|✔️|
> < |CVFI1_SVA_C|✔️|
> < |CVFI2_Sony_H|✔️|
> < |CVFI2_SVA_C|✔️|
> < |CVMA1_Sony_D|✔️|
> < |CVMAPAQP3_Sony_E|✔️|
> < |CVMAQP2_Sony_G|✔️|
> < |CVMAQP3_Sony_D|✔️|
> < |CVMP_MOT_FLD_L30_B|✔️|
> < |CVMP_MOT_FRM_L31_B|✔️|
> < |CVNLFI1_Sony_C|✔️|
> < |CVNLFI2_Sony_H|✔️|
> < |FI1_Sony_E|✔️|
> < |MR9_BT_B|✔️|
> < |Sharp_MP_Field_1_B|✔️|
> < |Sharp_MP_Field_2_B|✔️|
> < |Sharp_MP_Field_3_B|✔️|
> < |Sharp_MP_PAFF_1r2|✔️|
> < |Sharp_MP_PAFF_2r|✔️|

Above tests are related to interlace support. As of now, interlace 
support is not there in iris driver, as it is gradually evolving w.r.t 
enabling available hardware codecs.

> 
> 
>>
>> [...]
>>
>>> For Venus several tests crash the firmware or crash the device,
>>> so test-suite was executed with extra:
>>
>> But this sounds way way worse
>>
>> [...]
>>
>>> For H.265:
>>>
>>> Only the main set was used on both platforms
>>>
>>> On Venus pass 135/147
>>> On Iris  pass 133/147
>>
>> [...]
>>
>>> For VP9 codec:
>>>
>>> On Venus pass 174/311
>>> On Iris  pass 232/311
>>
>> These two give me hope..
>>
>> Ultimately, I think it's the right thing to do, if only to prevent the
>> crashes you've described..
> 
> Yep.
> 

Regards,
Vikash