Skip to contentSkip to conversation

Use INTEGER64 type for MEI CSI LINK_FREQ control

PATCH In review Sakari Ailus 4 participants 9 messages Apr 26, 2024 – Apr 29, 2024

Subsystems (2)
lore
Share & tools
lore

Sakari Ailus

author maintainer
#0 lore raw

v2 cover Use INTEGER64 type for MEI CSI LINK_FREQ control

Hi folks,

An integer menu isn't a useful control type for conveying the frequency depending on an external device to the receiver. Instead, in the MEI CSI driver, just obtain the link frequency from the upsteam sub-device and pass it on to the receiver.

The v4l2_get_link_freq() is changed to add support for this and it's documented as well.

since v1:

- Add a new 64-bit integer control V4L2_CID_CUR_LINK_FREQ instead of re-using V4L2_CID_LINK_FREQ.

Sakari Ailus (3):
media: Documentation: v4l: Add CUR_LINK_FREQ control
media: v4l: Support obtaining link frequency from CUR_LINK_FREQ
control
media: ivsc: csi: Fix link frequency control behaviour

 .../media/v4l/ext-ctrls-image-process.rst     |  4 ++++
 drivers/media/pci/intel/ivsc/mei_csi.c        | 24 ++++++-------------
 drivers/media/v4l2-core/v4l2-common.c         |  6 ++++-
 drivers/media/v4l2-core/v4l2-ctrls-defs.c     |  5 ++++
 include/uapi/linux/v4l2-controls.h            |  1 +
 5 files changed, 22 insertions(+), 18 deletions(-)

base-commit: 6612d1ae37b52d37d1d515d7c5110b9f3afe94d2

Signature · 1 line
2.39.2
author maintainer
#1 lore raw

PATCH v2 1/3 media: Documentation: v4l: Add CUR_LINK_FREQ control

Add a 64-bit integer control for devices that are passing through CSI-2 (or other) data, to signal the link frequency used by also the upstream sub-device. The control is volatile and read-only.

---
 .../userspace-api/media/v4l/ext-ctrls-image-process.rst      | 4 ++++
 drivers/media/v4l2-core/v4l2-ctrls-defs.c                    | 5 +++++
 include/uapi/linux/v4l2-controls.h                           | 1 +
 3 files changed, 10 insertions(+)
Documentation/userspace-api/media/v4l/ext-ctrls-image-process.rst +4 -0
index b1c2ab2854af..31d5045f32e7 100644
@@ -55,3 +55,7 @@ Image Process Control IDs
     control value divided by e.g. 0x100, meaning that to get no
     digital gain the control value needs to be 0x100. The no-gain
     configuration is also typically the default.
+
+``V4L2_CID_CUR_LINK_FREQ (64-bit integer)``
+    The current link frequency on a data interface (e.g. parallel or
+    CSI-2). This control is read-only.
drivers/media/v4l2-core/v4l2-ctrls-defs.c +5 -0
index 8696eb1cdd61..1e8e02818f9f 100644
@@ -1163,6 +1163,7 @@ const char *v4l2_ctrl_get_name(u32 id)
 	case V4L2_CID_TEST_PATTERN:		return "Test Pattern";
 	case V4L2_CID_DEINTERLACING_MODE:	return "Deinterlacing Mode";
 	case V4L2_CID_DIGITAL_GAIN:		return "Digital Gain";
+	case V4L2_CID_CUR_LINK_FREQ:		return "Link Frequency";
 
 	/* DV controls */
 	/* Keep the order of the 'case's the same as in v4l2-controls.h! */
@@ -1425,6 +1426,10 @@ void v4l2_ctrl_fill(u32 id, const char **name, enum v4l2_ctrl_type *type,
 	case V4L2_CID_LINK_FREQ:
 		*type = V4L2_CTRL_TYPE_INTEGER_MENU;
 		break;
+	case V4L2_CID_CUR_LINK_FREQ:
+		*flags |= V4L2_CTRL_FLAG_VOLATILE | V4L2_CTRL_FLAG_READ_ONLY;
+		*type = V4L2_CTRL_TYPE_INTEGER64;
+		break;
 	case V4L2_CID_RDS_TX_PS_NAME:
 	case V4L2_CID_RDS_TX_RADIO_TEXT:
 	case V4L2_CID_RDS_RX_PS_NAME:
include/uapi/linux/v4l2-controls.h +1 -0
index 99c3f5e99da7..1c63d2e51e8d 100644
@@ -1223,6 +1223,7 @@ enum v4l2_jpeg_chroma_subsampling {
 #define V4L2_CID_TEST_PATTERN			(V4L2_CID_IMAGE_PROC_CLASS_BASE + 3)
 #define V4L2_CID_DEINTERLACING_MODE		(V4L2_CID_IMAGE_PROC_CLASS_BASE + 4)
 #define V4L2_CID_DIGITAL_GAIN			(V4L2_CID_IMAGE_PROC_CLASS_BASE + 5)
+#define V4L2_CID_CUR_LINK_FREQ			(V4L2_CID_IMAGE_PROC_CLASS_BASE + 6)
 
 /*  DV-class control IDs defined by V4L2 */
 #define V4L2_CID_DV_CLASS_BASE			(V4L2_CTRL_CLASS_DV | 0x900)
Signature · 1 line
2.39.2
author maintainer
#2 lore raw

PATCH v2 2/3 media: v4l: Support obtaining link frequency from CUR_LINK_FREQ control

In reply to Sakari Ailus

Support V4L2_CID_CUR_LINK_FREQ INTEGER64 control in v4l2_get_link_freq().

---
 drivers/media/v4l2-core/v4l2-common.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)
drivers/media/v4l2-core/v4l2-common.c +5 -1
index 4165c815faef..bf1923314d15 100644
@@ -483,6 +483,10 @@ s64 v4l2_get_link_freq(struct v4l2_ctrl_handler *handler, unsigned int mul,
 
 		freq = qm.value;
 	} else {
+		ctrl = v4l2_ctrl_find(handler, V4L2_CID_CUR_LINK_FREQ);
+		if (ctrl)
+			return v4l2_ctrl_g_ctrl_int64(ctrl);
+
 		if (!mul || !div)
 			return -ENOENT;
 
@@ -494,7 +498,7 @@ s64 v4l2_get_link_freq(struct v4l2_ctrl_handler *handler, unsigned int mul,
 
 		pr_warn("%s: Link frequency estimated using pixel rate: result might be inaccurate\n",
 			__func__);
-		pr_warn("%s: Consider implementing support for V4L2_CID_LINK_FREQ in the transmitter driver\n",
+		pr_warn("%s: Consider implementing support for V4L2_CID_LINK_FREQ or V4L2_CID_CUR_LINK_FREQ in the transmitter driver\n",
 			__func__);
 	}
 
Signature · 1 line
2.39.2
author maintainer
#3 lore raw

PATCH v2 3/3 media: ivsc: csi: Fix link frequency control behaviour

In reply to Sakari Ailus

The link frequency control was an integer menu control the value of which was hard-coded in the MEI CSI driver itself. Instead obtain the control value from the upstream sub-device and use the CUR_LINK_FREQ INTEGER64 control type for the purpose.

Fixes 29006e196a56 ("media: pci: intel: ivsc: Add CSI submodule")
---
 drivers/media/pci/intel/ivsc/mei_csi.c | 24 +++++++-----------------
 1 file changed, 7 insertions(+), 17 deletions(-)
drivers/media/pci/intel/ivsc/mei_csi.c +7 -17
index 55e0c60c420c..4a35932f1cd3 100644
@@ -15,6 +15,7 @@
 #include <linux/completion.h>
 #include <linux/delay.h>
 #include <linux/kernel.h>
+#include <linux/limits.h>
 #include <linux/math64.h>
 #include <linux/mei_cl_bus.h>
 #include <linux/module.h>
@@ -32,8 +33,6 @@
 
 #define MEI_CSI_ENTITY_NAME "Intel IVSC CSI"
 
-#define MEI_CSI_LINK_FREQ_400MHZ 400000000ULL
-
 /* the 5s used here is based on experiment */
 #define CSI_CMD_TIMEOUT (5 * HZ)
 /* to setup CSI-2 link an extra delay needed and determined experimentally */
@@ -145,10 +144,6 @@ static const struct v4l2_mbus_framefmt mei_csi_format_mbus_default = {
 	.field = V4L2_FIELD_NONE,
 };
 
-static s64 link_freq_menu_items[] = {
-	MEI_CSI_LINK_FREQ_400MHZ
-};
-
 static inline struct mei_csi *notifier_to_csi(struct v4l2_async_notifier *n)
 {
 	return container_of(n, struct mei_csi, notifier);
@@ -470,7 +465,7 @@ static int mei_csi_g_volatile_ctrl(struct v4l2_ctrl *ctrl)
 	struct mei_csi *csi = ctrl_to_csi(ctrl);
 	s64 freq;
 
-	if (ctrl->id == V4L2_CID_LINK_FREQ) {
+	if (ctrl->id == V4L2_CID_CUR_LINK_FREQ) {
 		if (!csi->remote)
 			return -EINVAL;
 
@@ -481,8 +476,7 @@ static int mei_csi_g_volatile_ctrl(struct v4l2_ctrl *ctrl)
 			return -EINVAL;
 		}
 
-		link_freq_menu_items[0] = freq;
-		ctrl->val = 0;
+		ctrl->p_new.p_s64[0] = csi->link_freq = freq;
 
 		return 0;
 	}
@@ -553,7 +547,6 @@ static const struct v4l2_async_notifier_operations mei_csi_notify_ops = {
 
 static int mei_csi_init_controls(struct mei_csi *csi)
 {
-	u32 max;
 	int ret;
 
 	ret = v4l2_ctrl_handler_init(&csi->ctrl_handler, 2);
@@ -562,13 +555,10 @@ static int mei_csi_init_controls(struct mei_csi *csi)
 
 	csi->ctrl_handler.lock = &csi->lock;
 
-	max = ARRAY_SIZE(link_freq_menu_items) - 1;
-	csi->freq_ctrl = v4l2_ctrl_new_int_menu(&csi->ctrl_handler,
-						&mei_csi_ctrl_ops,
-						V4L2_CID_LINK_FREQ,
-						max,
-						0,
-						link_freq_menu_items);
+	csi->freq_ctrl = v4l2_ctrl_new_std(&csi->ctrl_handler,
+					   &mei_csi_ctrl_ops,
+					   V4L2_CID_CUR_LINK_FREQ, 1, S64_MAX,
+					   1, 1);
 	if (csi->freq_ctrl)
 		csi->freq_ctrl->flags |= V4L2_CTRL_FLAG_READ_ONLY |
 					 V4L2_CTRL_FLAG_VOLATILE;
Signature · 1 line
2.39.2
Acked-by on v2 3/3
#4 lore raw

RE: [PATCH v2 3/3] media: ivsc: csi: Fix link frequency control behaviour

Quoted 9 lines

From: Sakari Ailus <[email protected]>

The link frequency control was an integer menu control the value of which was hard-coded in the MEI CSI driver itself. Instead obtain the control value from the upstream sub-device and use the CUR_LINK_FREQ INTEGER64 control type for the purpose.

Fixes 29006e196a56 ("media: pci: intel: ivsc: Add CSI submodule")
Quoted 85 lines
---
 drivers/media/pci/intel/ivsc/mei_csi.c | 24 +++++++-----------------
 1 file changed, 7 insertions(+), 17 deletions(-)
drivers/media/pci/intel/ivsc/mei_csi.c +1 -2
index 55e0c60c420c..4a35932f1cd3 100644
@@ -15,6 +15,7 @@
 #include <linux/completion.h>
 #include <linux/delay.h>
 #include <linux/kernel.h>
+#include <linux/limits.h>
 #include <linux/math64.h>
 #include <linux/mei_cl_bus.h>
 #include <linux/module.h>
@@ -32,8 +33,6 @@
 
 #define MEI_CSI_ENTITY_NAME "Intel IVSC CSI"
 
-#define MEI_CSI_LINK_FREQ_400MHZ 400000000ULL
-
 /* the 5s used here is based on experiment */  #define CSI_CMD_TIMEOUT
 (5 * HZ)
 /* to setup CSI-2 link an extra delay needed and determined experimentally

*/ @@ -145,10 +144,6 @@ static const struct v4l2_mbus_framefmt mei_csi_format_mbus_default = {

drivers/media/pci/intel/ivsc/mei_csi.c +0 -4
 	.field = V4L2_FIELD_NONE,
 };
 
-static s64 link_freq_menu_items[] = {
-	MEI_CSI_LINK_FREQ_400MHZ
-};
-
 static inline struct mei_csi *notifier_to_csi(struct v4l2_async_notifier *n)  {
 	return container_of(n, struct mei_csi, notifier); @@ -470,7 +465,7

@@ static int mei_csi_g_volatile_ctrl(struct v4l2_ctrl *ctrl)

drivers/media/pci/intel/ivsc/mei_csi.c +6 -11
 	struct mei_csi *csi = ctrl_to_csi(ctrl);
 	s64 freq;
 
-	if (ctrl->id == V4L2_CID_LINK_FREQ) {
+	if (ctrl->id == V4L2_CID_CUR_LINK_FREQ) {
 		if (!csi->remote)
 			return -EINVAL;
@@ -481,8 +476,7 @@ static int mei_csi_g_volatile_ctrl(struct v4l2_ctrl *ctrl)
 			return -EINVAL;
 		}
 
-		link_freq_menu_items[0] = freq;
-		ctrl->val = 0;
+		ctrl->p_new.p_s64[0] = csi->link_freq = freq;
 
 		return 0;
 	}
@@ -553,7 +547,6 @@ static const struct v4l2_async_notifier_operations
 mei_csi_notify_ops = {
 
 static int mei_csi_init_controls(struct mei_csi *csi)  {
-	u32 max;
 	int ret;
 
 	ret = v4l2_ctrl_handler_init(&csi->ctrl_handler, 2); @@ -562,13
+555,10 @@ static int mei_csi_init_controls(struct mei_csi *csi)
 
 	csi->ctrl_handler.lock = &csi->lock;
 
-	max = ARRAY_SIZE(link_freq_menu_items) - 1;
-	csi->freq_ctrl = v4l2_ctrl_new_int_menu(&csi->ctrl_handler,
-						&mei_csi_ctrl_ops,
-						V4L2_CID_LINK_FREQ,
-						max,
-						0,
-						link_freq_menu_items);
+	csi->freq_ctrl = v4l2_ctrl_new_std(&csi->ctrl_handler,
+					   &mei_csi_ctrl_ops,
+					   V4L2_CID_CUR_LINK_FREQ, 1,

S64_MAX,

drivers/media/pci/intel/ivsc/mei_csi.c +1 -1
+					   1, 1);
 	if (csi->freq_ctrl)
 		csi->freq_ctrl->flags |= V4L2_CTRL_FLAG_READ_ONLY |
 					 V4L2_CTRL_FLAG_VOLATILE;
--

2.39.2

#5 lore raw
In reply to Sakari Ailus

Hi Sakari,

Thank you for the patches.

Quoted 9 lines · Sakari Ailus

Hi folks,

An integer menu isn't a useful control type for conveying the frequency depending on an external device to the receiver. Instead, in the MEI CSI driver, just obtain the link frequency from the upsteam sub-device and pass it on to the receiver.

The v4l2_get_link_freq() is changed to add support for this and it's documented as well.

Using a control for this seems to be a bit of a hack :-S Wouldn't it fit more nicely in the .get_mbus_config() operation ? That would avoid exposing this to userspace, and would also have the nice advantage that the value could be queried per pad.

Quoted 20 lines

since v1:

- Add a new 64-bit integer control V4L2_CID_CUR_LINK_FREQ instead of re-using V4L2_CID_LINK_FREQ.

Sakari Ailus (3):
media: Documentation: v4l: Add CUR_LINK_FREQ control
media: v4l: Support obtaining link frequency from CUR_LINK_FREQ
control
media: ivsc: csi: Fix link frequency control behaviour

 .../media/v4l/ext-ctrls-image-process.rst     |  4 ++++
 drivers/media/pci/intel/ivsc/mei_csi.c        | 24 ++++++-------------
 drivers/media/v4l2-core/v4l2-common.c         |  6 ++++-
 drivers/media/v4l2-core/v4l2-ctrls-defs.c     |  5 ++++
 include/uapi/linux/v4l2-controls.h            |  1 +
 5 files changed, 22 insertions(+), 18 deletions(-)

base-commit: 6612d1ae37b52d37d1d515d7c5110b9f3afe94d2

Signature · 3 lines
Regards,

Laurent Pinchart
author maintainer
#6 lore raw

Hi Laurent,

Quoted 12 lines · Laurent Pinchart

Hi Sakari,

Thank you for the patches.

Quoted 7 lines · Sakari Ailus

Hi folks,

An integer menu isn't a useful control type for conveying the frequency depending on an external device to the receiver. Instead, in the MEI CSI driver, just obtain the link frequency from the upsteam sub-device and pass it on to the receiver.

Quoted 7 lines
Quoted 2 lines

The v4l2_get_link_freq() is changed to add support for this and it's documented as well.

Using a control for this seems to be a bit of a hack :-S Wouldn't it fit more nicely in the .get_mbus_config() operation ? That would avoid exposing this to userspace, and would also have the nice advantage that the value could be queried per pad.

Works for me. I also think it's better than using a new control for this actually.

Quoted 22 lines
Quoted 21 lines

since v1:

- Add a new 64-bit integer control V4L2_CID_CUR_LINK_FREQ instead of re-using V4L2_CID_LINK_FREQ.

Sakari Ailus (3):
media: Documentation: v4l: Add CUR_LINK_FREQ control
media: v4l: Support obtaining link frequency from CUR_LINK_FREQ
control
media: ivsc: csi: Fix link frequency control behaviour

 .../media/v4l/ext-ctrls-image-process.rst     |  4 ++++
 drivers/media/pci/intel/ivsc/mei_csi.c        | 24 ++++++-------------
 drivers/media/v4l2-core/v4l2-common.c         |  6 ++++-
 drivers/media/v4l2-core/v4l2-ctrls-defs.c     |  5 ++++
 include/uapi/linux/v4l2-controls.h            |  1 +
 5 files changed, 22 insertions(+), 18 deletions(-)

base-commit: 6612d1ae37b52d37d1d515d7c5110b9f3afe94d2

Signature · 3 lines
Regards,

Sakari Ailus
#7 lore raw
Quoted 18 lines · Sakari Ailus

Hi Laurent,

Quoted 15 lines · Laurent Pinchart

Hi Sakari,

Thank you for the patches.

Quoted 9 lines · Sakari Ailus

Hi folks,

An integer menu isn't a useful control type for conveying the frequency depending on an external device to the receiver. Instead, in the MEI CSI driver, just obtain the link frequency from the upsteam sub-device and pass it on to the receiver.

The v4l2_get_link_freq() is changed to add support for this and it's documented as well.

Quoted 7 lines
Quoted 4 lines

Using a control for this seems to be a bit of a hack :-S Wouldn't it fit more nicely in the .get_mbus_config() operation ? That would avoid exposing this to userspace, and would also have the nice advantage that the value could be queried per pad.

Works for me. I also think it's better than using a new control for this actually.

Same for me. As you gathered, I was not very enthusiastic about using a control for this either.

Regards,

Hans

Quoted 24 lines
Quoted 23 lines
Quoted 22 lines

since v1:

- Add a new 64-bit integer control V4L2_CID_CUR_LINK_FREQ instead of re-using V4L2_CID_LINK_FREQ.

Sakari Ailus (3):
media: Documentation: v4l: Add CUR_LINK_FREQ control
media: v4l: Support obtaining link frequency from CUR_LINK_FREQ
control
media: ivsc: csi: Fix link frequency control behaviour

 .../media/v4l/ext-ctrls-image-process.rst     |  4 ++++
 drivers/media/pci/intel/ivsc/mei_csi.c        | 24 ++++++-------------
 drivers/media/v4l2-core/v4l2-common.c         |  6 ++++-
 drivers/media/v4l2-core/v4l2-ctrls-defs.c     |  5 ++++
 include/uapi/linux/v4l2-controls.h            |  1 +
 5 files changed, 22 insertions(+), 18 deletions(-)

base-commit: 6612d1ae37b52d37d1d515d7c5110b9f3afe94d2

author maintainer
#8 lore raw

Hi Hans,

Quoted 24 lines · Hans Verkuil
Quoted 23 lines · Sakari Ailus

Hi Laurent,

Quoted 19 lines · Laurent Pinchart

Hi Sakari,

Thank you for the patches.

Quoted 9 lines · Sakari Ailus

Hi folks,

An integer menu isn't a useful control type for conveying the frequency depending on an external device to the receiver. Instead, in the MEI CSI driver, just obtain the link frequency from the upsteam sub-device and pass it on to the receiver.

The v4l2_get_link_freq() is changed to add support for this and it's documented as well.

Using a control for this seems to be a bit of a hack :-S Wouldn't it fit more nicely in the .get_mbus_config() operation ? That would avoid exposing this to userspace, and would also have the nice advantage that the value could be queried per pad.

Quoted 5 lines
Quoted 2 lines

Works for me. I also think it's better than using a new control for this actually.

Same for me. As you gathered, I was not very enthusiastic about using a control for this either.

Ack.

I guess these are fine to be merged for 6.11 but still backporting to 6.10 should be done. Nothing breaks now as there's a single frequency used right now and that's 400 MHz.

Signature · 3 lines
Regards,

Sakari Ailus