Skip to contentSkip to conversation

Ignore bad graph port nodes on Dell XPS 9315

PATCH Applied Sakari Ailus 3 participants 8 messages Feb 13, 2024 – Feb 22, 2024 ACPI

lore
Share & tools
lore
#0 lore raw

v2 cover Ignore bad graph port nodes on Dell XPS 9315

Hi Rafael,

These two patches skip instantiating bad port nodes on Dell XPS 9315.

The ACPI firmware contains Linux specific camera declarations but they are missing the IVSC which sits between the camera and the IPU. The IPU bridge can be instead used to obtain a configuration that is usable in the system.

I presume there are other models that have the same issue.

since v1:

- Use a better subject for the second patch.

- Don't explicitly initialise static variable (second patch).

- Improve the commit message of the second patch slightly.

Sakari Ailus (2):
ACPI: utils: Make acpi_handle_path() not static
ACPI: property: Ignore bad graph port nodes on Dell XPS 9315

 drivers/acpi/internal.h       |  1 +
 drivers/acpi/mipi-disco-img.c | 71 +++++++++++++++++++++++++++++++++++
 drivers/acpi/property.c       |  3 ++
 drivers/acpi/utils.c          |  2 +-
 include/linux/acpi.h          |  1 +
 5 files changed, 77 insertions(+), 1 deletion(-)
Signature · 1 line
2.39.2
#1 lore raw

PATCH v2 1/2 ACPI: utils: Make acpi_handle_path() not static

acpi_handle_path() will soon be required for node name comparison elsewhere in ACPI framework. Remove the static keyword and add the prototype to include/linux/acpi.h.

---
 drivers/acpi/utils.c | 2 +-
 include/linux/acpi.h | 1 +
 2 files changed, 2 insertions(+), 1 deletion(-)
drivers/acpi/utils.c +1 -1
index abac5cc25477..202234ba54bd 100644
@@ -559,7 +559,7 @@ EXPORT_SYMBOL(acpi_evaluate_ost);
  *
  * Caller must free the returned buffer
  */
-static char *acpi_handle_path(acpi_handle handle)
+char *acpi_handle_path(acpi_handle handle)
 {
 	struct acpi_buffer buffer = {
 		.length = ACPI_ALLOCATE_BUFFER,
include/linux/acpi.h +1 -0
index b7165e52b3c6..a170c389dd74 100644
@@ -1170,6 +1170,7 @@ static inline void acpi_ec_set_gpe_wake_mask(u8 action) {}
 #endif
 
 #ifdef CONFIG_ACPI
+char *acpi_handle_path(acpi_handle handle);
 __printf(3, 4)
 void acpi_handle_printk(const char *level, acpi_handle handle,
 			const char *fmt, ...);
Signature · 1 line
2.39.2
#2 lore raw

PATCH v2 2/2 ACPI: property: Ignore bad graph port nodes on Dell XPS 9315

In reply to Sakari Ailus

Some systems were shipped with both Windows and Linux camera descriptions. In general, if Linux description exist, they will be used and Windows description ignored.

In this case the Linux descriptions were buggy so use Windows definition instead. This patch ignores the bad graph port nodes on Dell XPS 9315 and there are likely other such systems, too. The corresponding information has been added to ipu-bridge to cover the missing bits.

---
 drivers/acpi/internal.h       |  1 +
 drivers/acpi/mipi-disco-img.c | 71 +++++++++++++++++++++++++++++++++++
 drivers/acpi/property.c       |  3 ++
 3 files changed, 75 insertions(+)
drivers/acpi/internal.h +1 -0
index 6588525c45ef..e0145df519bd 100644
@@ -301,5 +301,6 @@ void acpi_mipi_check_crs_csi2(acpi_handle handle);
 void acpi_mipi_scan_crs_csi2(void);
 void acpi_mipi_init_crs_csi2_swnodes(void);
 void acpi_mipi_crs_csi2_cleanup(void);
+bool acpi_graph_ignore_port(acpi_handle handle);
 
 #endif /* _ACPI_INTERNAL_H_ */
drivers/acpi/mipi-disco-img.c +71 -0

80 lines not on this page · raw message

drivers/acpi/property.c +3 -0
index a6ead5204046..2b73580c9f36 100644
@@ -80,6 +80,9 @@ static bool acpi_nondev_subnode_extract(union acpi_object *desc,
 	struct acpi_data_node *dn;
 	bool result;
 
+	if (acpi_graph_ignore_port(handle))
+		return false;
+
 	dn = kzalloc(sizeof(*dn), GFP_KERNEL);
 	if (!dn)
 		return false;
Signature · 1 line
2.39.2
maintainer Applied
#3 lore raw

Re: [PATCH v2 2/2] ACPI: property: Ignore bad graph port nodes on Dell XPS 9315

Quoted 119 lines · Sakari Ailus

Some systems were shipped with both Windows and Linux camera descriptions. In general, if Linux description exist, they will be used and Windows description ignored.

In this case the Linux descriptions were buggy so use Windows definition instead. This patch ignores the bad graph port nodes on Dell XPS 9315 and there are likely other such systems, too. The corresponding information has been added to ipu-bridge to cover the missing bits.

---
 drivers/acpi/internal.h       |  1 +
 drivers/acpi/mipi-disco-img.c | 71 +++++++++++++++++++++++++++++++++++
 drivers/acpi/property.c       |  3 ++
 3 files changed, 75 insertions(+)
drivers/acpi/internal.h +1 -0
index 6588525c45ef..e0145df519bd 100644
@@ -301,5 +301,6 @@ void acpi_mipi_check_crs_csi2(acpi_handle handle);
 void acpi_mipi_scan_crs_csi2(void);
 void acpi_mipi_init_crs_csi2_swnodes(void);
 void acpi_mipi_crs_csi2_cleanup(void);
+bool acpi_graph_ignore_port(acpi_handle handle);
 
 #endif /* _ACPI_INTERNAL_H_ */
drivers/acpi/mipi-disco-img.c +71 -0
index 7286cf4579bc..da71eb4bf6a6 100644
@@ -19,6 +19,7 @@
  */
 
 #include <linux/acpi.h>
+#include <linux/dmi.h>
 #include <linux/limits.h>
 #include <linux/list.h>
 #include <linux/module.h>
@@ -723,3 +724,73 @@ void acpi_mipi_crs_csi2_cleanup(void)
        list_for_each_entry_safe(csi2, csi2_tmp, &acpi_mipi_crs_csi2_list, entry)
                acpi_mipi_del_crs_csi2(csi2);
 }
+
+static const struct dmi_system_id dmi_ignore_port_nodes[] = {
+       {
+               .matches = {
+                       DMI_EXACT_MATCH(DMI_SYS_VENDOR, "Dell Inc."),
+                       DMI_EXACT_MATCH(DMI_PRODUCT_NAME, "XPS 9315"),
+               },
+       },
+       { 0 }
+};
+
+static const char *strnext(const char *s1, const char *s2)
+{
+       s1 = strstr(s1, s2);
+
+       if (!s1)
+               return NULL;
+
+       return s1 + strlen(s2);
+}
+
+/**
+ * acpi_graph_ignore_port - Tell whether a port node should be ignored
+ * @handle: The ACPI handle of the node (which may be a port node)
+ *
+ * Returns true if a port node should be ignored and the data to that should
+ * come from other sources instead (Windows ACPI definitions and
+ * ipu-bridge). This is currently used to ignore bad port nodes related to IPU6
+ * ("IPU?") and camera sensor devices ("LNK?") in certain Dell systems with
+ * Intel VSC.
+ */
+bool acpi_graph_ignore_port(acpi_handle handle)
+{
+       const char *path = NULL, *orig_path;
+       static bool dmi_tested, ignore_port;
+
+       if (!dmi_tested) {
+               ignore_port = dmi_first_match(dmi_ignore_port_nodes);
+               dmi_tested = true;
+       }
+
+       if (!ignore_port)
+               return false;
+
+       /* Check if the device is either "IPU" or "LNK" (sensor). */
+       orig_path = acpi_handle_path(handle);
+       if (!orig_path)
+               return false;
+       path = strnext(orig_path, "IPU");
+       if (!path)
+               path = strnext(orig_path, "LNK");
+       if (!path)
+               goto out_free;
+
+       if (!(path[0] >= '0' && path[0] <= '9' && path[1] == '.'))
+               goto out_free;
+
+       /* Check if the node has a "PRT" prefix. */
+       path = strnext(path, "PRT");
+       if (path && path[0] >= '0' && path[0] <= '9' && !path[1]) {
+               acpi_handle_debug(handle, "ignoring data node\n");
+
+               kfree(orig_path);
+               return true;
+       }
+
+out_free:
+       kfree(orig_path);
+       return false;
+}
drivers/acpi/property.c +0 -0
index a6ead5204046..2b73580c9f36 100644

drivers/acpi/property.c static bool acpi_nondev_subnode_extract(union acpi_object *desc,

Quoted 10 lines
drivers/acpi/property.c +3 -1
@@ -80,6 +80,9 @@ static bool acpi_nondev_subnode_extract(union acpi_object *desc,
        struct acpi_data_node *dn;
        bool result;
 
+       if (acpi_graph_ignore_port(handle))
+               return false;
+
        dn = kzalloc(sizeof(*dn), GFP_KERNEL);
        if (!dn)
                return false;
--

Applied as 6.9 material along with the [1/2], thanks!

#4 lore raw

Re: [PATCH v2 2/2] ACPI: property: Ignore bad graph port nodes on Dell XPS 9315

In reply to Sakari Ailus

Tue, Feb 13, 2024 at 03:46:06PM +0200, Sakari Ailus kirjoitti:

Quoted 8 lines

Some systems were shipped with both Windows and Linux camera descriptions. In general, if Linux description exist, they will be used and Windows description ignored.

In this case the Linux descriptions were buggy so use Windows definition instead. This patch ignores the bad graph port nodes on Dell XPS 9315 and there are likely other such systems, too. The corresponding information has been added to ipu-bridge to cover the missing bits.

...

Quoted 3 lines

void acpi_mipi_scan_crs_csi2(void);
void acpi_mipi_init_crs_csi2_swnodes(void);
void acpi_mipi_crs_csi2_cleanup(void);

+ blank line?

Quoted 3 lines

+bool acpi_graph_ignore_port(acpi_handle handle);

#endif /* _ACPI_INTERNAL_H_ */

...

drivers/acpi/mipi-disco-img.c void acpi_mipi_crs_csi2_cleanup(void) · line 735

Quoted 8 lines
drivers/acpi/mipi-disco-img.c +8 -0
+static const struct dmi_system_id dmi_ignore_port_nodes[] = {
+	{
+		.matches = {
+			DMI_EXACT_MATCH(DMI_SYS_VENDOR, "Dell Inc."),
+			DMI_EXACT_MATCH(DMI_PRODUCT_NAME, "XPS 9315"),
+		},
+	},
+	{ 0 }

0 is not needed. Moreover, it's a bit unusual.

Quoted 1 line

+};

...

drivers/acpi/mipi-disco-img.c void acpi_mipi_crs_csi2_cleanup(void) · line 746

Quoted 9 lines
drivers/acpi/mipi-disco-img.c +9 -0
+static const char *strnext(const char *s1, const char *s2)
+{
+	s1 = strstr(s1, s2);
+
+	if (!s1)
+		return NULL;
+
+	return s1 + strlen(s2);
+}

NIH str_has_prefix() ?

...

Quoted 5 lines

+/**
+ * acpi_graph_ignore_port - Tell whether a port node should be ignored
+ * @handle: The ACPI handle of the node (which may be a port node)
+ *
+ * Returns true if a port node should be ignored and the data to that should

Please, run kernel-doc validation and fix the warnings.

Quoted 13 lines

+ * come from other sources instead (Windows ACPI definitions and
+ * ipu-bridge). This is currently used to ignore bad port nodes related to IPU6
+ * ("IPU?") and camera sensor devices ("LNK?") in certain Dell systems with
+ * Intel VSC.
+ */

drivers/acpi/mipi-disco-img.c +8 -0
+bool acpi_graph_ignore_port(acpi_handle handle)
+{
+	const char *path = NULL, *orig_path;
+	static bool dmi_tested, ignore_port;
+
+	if (!dmi_tested) {
+		ignore_port = dmi_first_match(dmi_ignore_port_nodes);
+		dmi_tested = true;

drivers/acpi/mipi-disco-img.c void acpi_mipi_crs_csi2_cleanup(void) · line 775

Quoted 10 lines
drivers/acpi/mipi-disco-img.c +10 -0
+	}
+
+	if (!ignore_port)
+		return false;
+
+	/* Check if the device is either "IPU" or "LNK" (sensor). */
+	orig_path = acpi_handle_path(handle);
+	if (!orig_path)
+		return false;
+	path = strnext(orig_path, "IPU");

IIUC this can be rewritten as

	path += str_has_prefix();
	if (path == orig_path)
		...

drivers/acpi/mipi-disco-img.c void acpi_mipi_crs_csi2_cleanup(void) · line 781

Quoted 6 lines
drivers/acpi/mipi-disco-img.c +6 -0
+	if (!path)
+		path = strnext(orig_path, "LNK");
+	if (!path)
+		goto out_free;
+
+	if (!(path[0] >= '0' && path[0] <= '9' && path[1] == '.'))

isdigit() ?

drivers/acpi/mipi-disco-img.c void acpi_mipi_crs_csi2_cleanup(void) · line 786

Quoted 5 lines
drivers/acpi/mipi-disco-img.c +5 -0
+		goto out_free;
+
+	/* Check if the node has a "PRT" prefix. */
+	path = strnext(path, "PRT");
+	if (path && path[0] >= '0' && path[0] <= '9' && !path[1]) {

Ditto.

Quoted 10 lines
drivers/acpi/mipi-disco-img.c +10 -0
+		acpi_handle_debug(handle, "ignoring data node\n");
+
+		kfree(orig_path);
+		return true;
+	}
+
+out_free:
+	kfree(orig_path);
+	return false;
+}
Signature · 2 lines
With Best Regards,
Andy Shevchenko
#5 lore raw

Re: [PATCH v2 2/2] ACPI: property: Ignore bad graph port nodes on Dell XPS 9315

Hi Andy,

Thanks for the review.

I think Rafael has already applied this but I can send a new patch...

Quoted 10 lines · On Wed, Feb 21, 2024 at 12:21:13AM +0200, [email protected] wrote:

Tue, Feb 13, 2024 at 03:46:06PM +0200, Sakari Ailus kirjoitti:

Quoted 8 lines

Some systems were shipped with both Windows and Linux camera descriptions. In general, if Linux description exist, they will be used and Windows description ignored.

In this case the Linux descriptions were buggy so use Windows definition instead. This patch ignores the bad graph port nodes on Dell XPS 9315 and there are likely other such systems, too. The corresponding information has been added to ipu-bridge to cover the missing bits.

Quoted 7 lines

...

Quoted 3 lines

void acpi_mipi_scan_crs_csi2(void);
void acpi_mipi_init_crs_csi2_swnodes(void);
void acpi_mipi_crs_csi2_cleanup(void);

+ blank line?

The usa of blank lines elsewhere in the file is consistent with the above. These are all related.

Quoted 7 lines
Quoted 4 lines

+bool acpi_graph_ignore_port(acpi_handle handle);

#endif /* _ACPI_INTERNAL_H_ */

...

Quoted 10 lines

drivers/acpi/mipi-disco-img.c void acpi_mipi_crs_csi2_cleanup(void) · line 735

Quoted 8 lines
drivers/acpi/mipi-disco-img.c +8 -0
+static const struct dmi_system_id dmi_ignore_port_nodes[] = {
+	{
+		.matches = {
+			DMI_EXACT_MATCH(DMI_SYS_VENDOR, "Dell Inc."),
+			DMI_EXACT_MATCH(DMI_PRODUCT_NAME, "XPS 9315"),
+		},
+	},
+	{ 0 }

0 is not needed. Moreover, it's a bit unusual.

But also required by the C standard.

Quoted 4 lines
Quoted 1 line

+};

...

Quoted 11 lines

drivers/acpi/mipi-disco-img.c void acpi_mipi_crs_csi2_cleanup(void) · line 746

Quoted 9 lines
drivers/acpi/mipi-disco-img.c +9 -0
+static const char *strnext(const char *s1, const char *s2)
+{
+	s1 = strstr(s1, s2);
+
+	if (!s1)
+		return NULL;
+
+	return s1 + strlen(s2);
+}

NIH str_has_prefix() ?

It doesn't do the same thing. Yes, it could be used, but the code looks cleaner with this.

Quoted 10 lines

...

Quoted 5 lines

+/**
+ * acpi_graph_ignore_port - Tell whether a port node should be ignored
+ * @handle: The ACPI handle of the node (which may be a port node)
+ *
+ * Returns true if a port node should be ignored and the data to that should

Please, run kernel-doc validation and fix the warnings.

Running kernel-doc on the file doesn't produce any here.

Quoted 27 lines
Quoted 14 lines

+ * come from other sources instead (Windows ACPI definitions and
+ * ipu-bridge). This is currently used to ignore bad port nodes related to IPU6
+ * ("IPU?") and camera sensor devices ("LNK?") in certain Dell systems with
+ * Intel VSC.
+ */

drivers/acpi/mipi-disco-img.c +8 -0
+bool acpi_graph_ignore_port(acpi_handle handle)
+{
+	const char *path = NULL, *orig_path;
+	static bool dmi_tested, ignore_port;
+
+	if (!dmi_tested) {
+		ignore_port = dmi_first_match(dmi_ignore_port_nodes);
+		dmi_tested = true;

drivers/acpi/mipi-disco-img.c void acpi_mipi_crs_csi2_cleanup(void) · line 775

Quoted 10 lines
drivers/acpi/mipi-disco-img.c +10 -0
+	}
+
+	if (!ignore_port)
+		return false;
+
+	/* Check if the device is either "IPU" or "LNK" (sensor). */
+	orig_path = acpi_handle_path(handle);
+	if (!orig_path)
+		return false;
+	path = strnext(orig_path, "IPU");

IIUC this can be rewritten as

Quoted 2 lines
	path += str_has_prefix();
	if (path == orig_path)

You could do that here, but now below, without introducing a new temporary variable and shuffling that and "path". I prefer to keep it as-is.

Quoted 10 lines
		...

drivers/acpi/mipi-disco-img.c void acpi_mipi_crs_csi2_cleanup(void) · line 781

Quoted 6 lines
drivers/acpi/mipi-disco-img.c +6 -0
+	if (!path)
+		path = strnext(orig_path, "LNK");
+	if (!path)
+		goto out_free;
+
+	if (!(path[0] >= '0' && path[0] <= '9' && path[1] == '.'))

isdigit() ?

Makes sense.

Quoted 8 lines

drivers/acpi/mipi-disco-img.c void acpi_mipi_crs_csi2_cleanup(void) · line 786

Quoted 6 lines
drivers/acpi/mipi-disco-img.c +5 -0
+		goto out_free;
+
+	/* Check if the node has a "PRT" prefix. */
+	path = strnext(path, "PRT");
+	if (path && path[0] >= '0' && path[0] <= '9' && !path[1]) {

Ditto.

Yes.

Quoted 12 lines
Quoted 11 lines
drivers/acpi/mipi-disco-img.c +10 -0
+		acpi_handle_debug(handle, "ignoring data node\n");
+
+		kfree(orig_path);
+		return true;
+	}
+
+out_free:
+	kfree(orig_path);
+	return false;
+}
Signature · 3 lines
Kind regards,

Sakari Ailus
#6 lore raw

Re: [PATCH v2 2/2] ACPI: property: Ignore bad graph port nodes on Dell XPS 9315

Quoted 1 line · Sakari Ailus

I think Rafael has already applied this but I can send a new patch...

Depends if it's final or can be dropped.
If the former is the case, follow ups, please.

Quoted 2 lines · On Wed, Feb 21, 2024 at 12:21:13AM +0200, [email protected] wrote:
Quoted 1 line

Tue, Feb 13, 2024 at 03:46:06PM +0200, Sakari Ailus kirjoitti:

...

Quoted 8 lines
Quoted 5 lines
Quoted 3 lines

void acpi_mipi_scan_crs_csi2(void);
void acpi_mipi_init_crs_csi2_swnodes(void);
void acpi_mipi_crs_csi2_cleanup(void);

+ blank line?

The usa of blank lines elsewhere in the file is consistent with the above. These are all related.

Naming does not suggest that. I see either naming or semantic tights issues here. Hence the proposal to have a blank line.

Quoted 1 line
Quoted 1 line
Quoted 1 line

+bool acpi_graph_ignore_port(acpi_handle handle);

...

Quoted 9 lines
Quoted 9 lines
Quoted 8 lines
drivers/acpi/mipi-disco-img.c +8 -0
+static const struct dmi_system_id dmi_ignore_port_nodes[] = {
+   {
+           .matches = {
+                   DMI_EXACT_MATCH(DMI_SYS_VENDOR, "Dell Inc."),
+                   DMI_EXACT_MATCH(DMI_PRODUCT_NAME, "XPS 9315"),
+           },
+   },
+   { 0 }
Quoted 3 lines
Quoted 1 line

0 is not needed. Moreover, it's a bit unusual.

But also required by the C standard.

We have a lot of code that doesn't use that (yet this is valid C23, and yes I know that we rely on C11).

Quoted 1 line
Quoted 1 line
Quoted 1 line

+};

...

Quoted 12 lines
Quoted 11 lines

drivers/acpi/mipi-disco-img.c void acpi_mipi_crs_csi2_cleanup(void) · line 746

Quoted 9 lines
drivers/acpi/mipi-disco-img.c +9 -0
+static const char *strnext(const char *s1, const char *s2)
+{
+   s1 = strstr(s1, s2);
+
+   if (!s1)
+           return NULL;
+
+   return s1 + strlen(s2);
+}

NIH str_has_prefix() ?

Quoted 2 lines

It doesn't do the same thing. Yes, it could be used, but the code looks cleaner with this.

"Not Invented Here" was even at Nokia times, why do we repeat our mistakes?

...

Quoted 9 lines
Quoted 7 lines
Quoted 5 lines

+/**
+ * acpi_graph_ignore_port - Tell whether a port node should be ignored
+ * @handle: The ACPI handle of the node (which may be a port node)
+ *
+ * Returns true if a port node should be ignored and the data to that should

Please, run kernel-doc validation and fix the warnings.

Running kernel-doc on the file doesn't produce any here.

You haven't run it correctly?

$ scripts/kernel-doc -v -none -Wall drivers/acpi/mipi-disco-img.c
drivers/acpi/mipi-disco-img.c:163: info: Scanning doc for function
acpi_mipi_check_crs_csi2
drivers/acpi/mipi-disco-img.c:384: info: Scanning doc for function
acpi_mipi_scan_crs_csi2
drivers/acpi/mipi-disco-img.c:703: info: Scanning doc for function
acpi_mipi_init_crs_csi2_swnodes
drivers/acpi/mipi-disco-img.c:718: info: Scanning doc for function
acpi_mipi_crs_csi2_cleanup
drivers/acpi/mipi-disco-img.c:749: info: Scanning doc for function
acpi_graph_ignore_port
drivers/acpi/mipi-disco-img.c:759: warning: No description found for
return value of 'acpi_graph_ignore_port'
1 warnings
Quoted 5 lines
Quoted 5 lines
Quoted 5 lines

+ * come from other sources instead (Windows ACPI definitions and
+ * ipu-bridge). This is currently used to ignore bad port nodes related to IPU6
+ * ("IPU?") and camera sensor devices ("LNK?") in certain Dell systems with
+ * Intel VSC.
+ */

Signature · 2 lines
With Best Regards,
Andy Shevchenko
#7 lore raw

Re: [PATCH v2 2/2] ACPI: property: Ignore bad graph port nodes on Dell XPS 9315

Quoted 37 lines · Andy Shevchenko
Quoted 1 line · Sakari Ailus

I think Rafael has already applied this but I can send a new patch...

Depends if it's final or can be dropped.
If the former is the case, follow ups, please.

Quoted 2 lines · On Wed, Feb 21, 2024 at 12:21:13AM +0200, [email protected] wrote:
Quoted 1 line

Tue, Feb 13, 2024 at 03:46:06PM +0200, Sakari Ailus kirjoitti:

...

Quoted 8 lines
Quoted 5 lines
Quoted 3 lines

void acpi_mipi_scan_crs_csi2(void);
void acpi_mipi_init_crs_csi2_swnodes(void);
void acpi_mipi_crs_csi2_cleanup(void);

+ blank line?

The usa of blank lines elsewhere in the file is consistent with the above. These are all related.

Naming does not suggest that. I see either naming or semantic tights issues here. Hence the proposal to have a blank line.

Quoted 1 line
Quoted 1 line
Quoted 1 line

+bool acpi_graph_ignore_port(acpi_handle handle);

...

Quoted 9 lines
Quoted 9 lines
Quoted 8 lines
drivers/acpi/mipi-disco-img.c +8 -0
+static const struct dmi_system_id dmi_ignore_port_nodes[] = {
+   {
+           .matches = {
+                   DMI_EXACT_MATCH(DMI_SYS_VENDOR, "Dell Inc."),
+                   DMI_EXACT_MATCH(DMI_PRODUCT_NAME, "XPS 9315"),
+           },
+   },
+   { 0 }
Quoted 6 lines
Quoted 3 lines
Quoted 1 line

0 is not needed. Moreover, it's a bit unusual.

But also required by the C standard.

We have a lot of code that doesn't use that (yet this is valid C23, and yes I know that we rely on C11).

The current compilers support it and I guess we'll switch to C23 at some point. I can drop the 0.

Quoted 17 lines
Quoted 2 lines
Quoted 2 lines
Quoted 2 lines

+};

...

Quoted 12 lines
Quoted 11 lines

drivers/acpi/mipi-disco-img.c void acpi_mipi_crs_csi2_cleanup(void) · line 746

Quoted 9 lines
drivers/acpi/mipi-disco-img.c +9 -0
+static const char *strnext(const char *s1, const char *s2)
+{
+   s1 = strstr(s1, s2);
+
+   if (!s1)
+           return NULL;
+
+   return s1 + strlen(s2);
+}

NIH str_has_prefix() ?

Quoted 4 lines
Quoted 2 lines

It doesn't do the same thing. Yes, it could be used, but the code looks cleaner with this.

"Not Invented Here" was even at Nokia times, why do we repeat our mistakes?

As I noted previously, while str_has_prefix() could be used, it would make the code in the caller harder to read. Changes may be needed to that code later on as this isn't the only system with the issue so readability does count.

Quoted 15 lines

...

Quoted 9 lines
Quoted 7 lines
Quoted 5 lines

+/**
+ * acpi_graph_ignore_port - Tell whether a port node should be ignored
+ * @handle: The ACPI handle of the node (which may be a port node)
+ *
+ * Returns true if a port node should be ignored and the data to that should

Please, run kernel-doc validation and fix the warnings.

Running kernel-doc on the file doesn't produce any here.

You haven't run it correctly?

Quoted 14 lines
$ scripts/kernel-doc -v -none -Wall drivers/acpi/mipi-disco-img.c
drivers/acpi/mipi-disco-img.c:163: info: Scanning doc for function
acpi_mipi_check_crs_csi2
drivers/acpi/mipi-disco-img.c:384: info: Scanning doc for function
acpi_mipi_scan_crs_csi2
drivers/acpi/mipi-disco-img.c:703: info: Scanning doc for function
acpi_mipi_init_crs_csi2_swnodes
drivers/acpi/mipi-disco-img.c:718: info: Scanning doc for function
acpi_mipi_crs_csi2_cleanup
drivers/acpi/mipi-disco-img.c:749: info: Scanning doc for function
acpi_graph_ignore_port
drivers/acpi/mipi-disco-img.c:759: warning: No description found for
return value of 'acpi_graph_ignore_port'
1 warnings

Ah, I guess it was -Wall option that was required to produce the warning. I'll address it.

Quoted 7 lines
Quoted 6 lines
Quoted 6 lines
Quoted 6 lines

+ * come from other sources instead (Windows ACPI definitions and
+ * ipu-bridge). This is currently used to ignore bad port nodes related to IPU6
+ * ("IPU?") and camera sensor devices ("LNK?") in certain Dell systems with
+ * Intel VSC.
+ */

Signature · 3 lines
Regards,

Sakari Ailus