Skip to contentSkip to conversation

acpi,pci: handle duplicate IRQ routing entries returned from _PRT

PATCH In review Mateusz Jończyk 3 participants 9 messages Sep 17, 2022 – Nov 23, 2022

Subsystems (4)
lore
Share & tools
lore
#0 lore raw

PATCH v2 1/1 acpi,pci: handle duplicate IRQ routing entries returned from _PRT

On some platforms, the ACPI _PRT function returns duplicate interrupt routing entries. Linux uses the first matching entry, but sometimes the second matching entry contains the correct interrupt vector.

This happens on a Dell Latitude E6500 laptop with the i2c-i801 Intel SMBus controller. This controller was nonfunctional unless its interrupt usage was disabled (using the "disable_features=0x10" module parameter).

After investigation, it turned out that the driver was using an incorrect interrupt vector: in lspci output for this device there was:
Interrupt: pin B routed to IRQ 19
but after running i2cdetect (without using any i2c-i801 module parameters) the following was logged to dmesg:

        [...]
        [  132.248657] i801_smbus 0000:00:1f.3: Timeout waiting for interrupt!
        [  132.248669] i801_smbus 0000:00:1f.3: Transaction timeout
        [  132.452649] i801_smbus 0000:00:1f.3: Timeout waiting for interrupt!
        [  132.452662] i801_smbus 0000:00:1f.3: Transaction timeout
        [  132.467682] irq 17: nobody cared (try booting with the "irqpoll" option)

Existence of duplicate entries in a table returned by the _PRT method was confirmed by disassembling the ACPI DSTD table.

Linux used the first matching entry, which was incorrect. In order not to disrupt existing systems, use the first matching entry unless the pci=prtlast kernel parameter is used or a Dell Latitude E6500 laptop is detected.

Disclaimer: there is nothing really interesting connected to the SMBus controller on this laptop, but this change may help other systems.

Cc Jonathan Corbet <[email protected]>
Cc Bjorn Helgaas <[email protected]>
Cc "Rafael J. Wysocki" <[email protected]>
Cc Len Brown <[email protected]>
Cc Borislav Petkov <[email protected]>

---
v2: do not quote the disassembled ACPI DSDT table - for copyright reasons.

---
 .../admin-guide/kernel-parameters.txt         |  8 ++
 drivers/acpi/pci_irq.c                        | 89 ++++++++++++++++++-
 drivers/pci/pci.c                             |  9 ++
 3 files changed, 102 insertions(+), 4 deletions(-)
Documentation/admin-guide/kernel-parameters.txt +8 -0
index 426fa892d311..2ff351db10b8 100644
@@ -4190,6 +4190,14 @@
 				bridge windows. This is the default on modern
 				hardware. If you need to use this, please report
 				a bug to <[email protected]>.
+		prtlast		If the _PRT ACPI method returns duplicate
+				IRQ routing entries, use the last matching entry
+				for a given device. If the platform may be
+				affected by this problem, an error message is
+				printed to dmesg - this parameter is
+				ineffective otherwise. If you need to use this,
+				please report a bug to
+				<[email protected]>.
 		routeirq	Do IRQ routing for all PCI devices.
 				This is normally done in pci_enable_device(),
 				so this option is a temporary workaround
drivers/acpi/pci_irq.c +85 -4

110 lines not on this page · raw message

drivers/pci/pci.c +9 -0
index 95bc329e74c0..a14a2e4e4197 100644
@@ -155,6 +155,11 @@ static bool pci_bridge_d3_disable;
 /* Force bridge_d3 for all PCIe ports */
 static bool pci_bridge_d3_force;
 
+#ifdef CONFIG_ACPI
+/* Use the last matching entry from the table returned by the _PRT ACPI method. */
+bool pci_prtlast;
+#endif
+
 static int __init pcie_port_pm_setup(char *str)
 {
 	if (!strcmp(str, "off"))
@@ -6896,6 +6901,10 @@ static int __init pci_setup(char *str)
 				pci_add_flags(PCI_SCAN_ALL_PCIE_DEVS);
 			} else if (!strncmp(str, "disable_acs_redir=", 18)) {
 				disable_acs_redir_param = str + 18;
+#ifdef CONFIG_ACPI
+			} else if (!strncmp(str, "prtlast", 7)) {
+				pci_prtlast = true;
+#endif
 			} else {
 				pr_err("PCI: Unknown option `%s'\n", str);
 			}

base-commit: 7e18e42e4b280c85b76967a9106a13ca61c16179

Signature · 1 line
2.25.1
#1 lore raw
Quoted 14 lines · Mateusz Jończyk

On some platforms, the ACPI _PRT function returns duplicate interrupt routing entries. Linux uses the first matching entry, but sometimes the second matching entry contains the correct interrupt vector.

This happens on a Dell Latitude E6500 laptop with the i2c-i801 Intel SMBus controller. This controller was nonfunctional unless its interrupt usage was disabled (using the "disable_features=0x10" module parameter).

After investigation, it turned out that the driver was using an incorrect interrupt vector: in lspci output for this device there was:
Interrupt: pin B routed to IRQ 19
but after running i2cdetect (without using any i2c-i801 module parameters) the following was logged to dmesg:

Quoted 6 lines
        [...]
        [  132.248657] i801_smbus 0000:00:1f.3: Timeout waiting for interrupt!
        [  132.248669] i801_smbus 0000:00:1f.3: Transaction timeout
        [  132.452649] i801_smbus 0000:00:1f.3: Timeout waiting for interrupt!
        [  132.452662] i801_smbus 0000:00:1f.3: Transaction timeout
        [  132.467682] irq 17: nobody cared (try booting with the "irqpoll" option)

Drop the timestamps; they add clutter but not useful information.

Quoted 7 lines

Existence of duplicate entries in a table returned by the _PRT method was confirmed by disassembling the ACPI DSTD table.

Linux used the first matching entry, which was incorrect. In order not to disrupt existing systems, use the first matching entry unless the pci=prtlast kernel parameter is used or a Dell Latitude E6500 laptop is detected.

Do we have a reason to believe that in general, using the first matching entry is incorrect? I don't see anything in the ACPI spec (r6.5, sec 6.2.13) that sheds light on this.

Presumably this works on Windows, and I doubt Windows would have a platform quirk for this, so I hypothesize that Windows treats _PRT entries as assignments, and the last one rules. Maybe Linux should adopt that rule?

Bjorn

#2 lore raw

W dniu 27.09.2022 o 21:28, Bjorn Helgaas pisze:

Quoted 8 lines · Mateusz Jończyk
Quoted 7 lines

On some platforms, the ACPI _PRT function returns duplicate interrupt routing entries. Linux uses the first matching entry, but sometimes the second matching entry contains the correct interrupt vector.

This happens on a Dell Latitude E6500 laptop with the i2c-i801 Intel SMBus controller. This controller was nonfunctional unless its interrupt usage was disabled (using the "disable_features=0x10" module parameter).

[...]

Quoted 10 lines
Quoted 7 lines

Existence of duplicate entries in a table returned by the _PRT method was confirmed by disassembling the ACPI DSTD table.

Linux used the first matching entry, which was incorrect. In order not to disrupt existing systems, use the first matching entry unless the pci=prtlast kernel parameter is used or a Dell Latitude E6500 laptop is detected.

Do we have a reason to believe that in general, using the first matching entry is incorrect? I don't see anything in the ACPI spec (r6.5, sec 6.2.13) that sheds light on this.

I meant that the entry was incorrect, not that Linux behaviour was incorrect.

I have also searched and browsed the ACPI spec, but have found no general rule that the OS should use the first or the last matching element from a list (in a general case, not just _PRT).

Quoted 4 lines

Presumably this works on Windows, and I doubt Windows would have a platform quirk for this, so I hypothesize that Windows treats _PRT entries as assignments, and the last one rules. Maybe Linux should adopt that rule?

I don't know whether this works on Windows, or just the laptop OEM did not care about the i2c bus on this model.

As a start, we may just print a warning when the _PRT table contains multiple matching entries for a given device - to see if there are any other devices that are affected (and which of the interrupt vectors for them is the correct one). This would be simpler then my proposed patch.

Greetings,

Mateusz

#3 lore raw
In reply to Mateusz Jończyk

[+cc Jean, linux-i2c]

Quoted 3 lines · Mateusz Jończyk

On some platforms, the ACPI _PRT function returns duplicate interrupt routing entries. Linux uses the first matching entry, but sometimes the second matching entry contains the correct interrupt vector.

Rafael, Jean, what do you think about this? It seems like kind of a lot of infrastructure to deal with this oddness, but I'm not really opposed to it.

This is in i2c-i801.c, which seems to have some support for polling; maybe it could make smart enough to complain and automatically switch to polling if a timeout occurs.

Or maybe we scan the entire _PRT and let the match win (instead of the first as we do today).

Or ...?

Google finds a lot of hits for "i801_smbus" "timeout waiting for interrupt", but I can't tell whether they're a similar _PRT issue or something else.

Quoted 210 lines

This happens on a Dell Latitude E6500 laptop with the i2c-i801 Intel SMBus controller. This controller was nonfunctional unless its interrupt usage was disabled (using the "disable_features=0x10" module parameter).

After investigation, it turned out that the driver was using an incorrect interrupt vector: in lspci output for this device there was:
Interrupt: pin B routed to IRQ 19
but after running i2cdetect (without using any i2c-i801 module parameters) the following was logged to dmesg:

        [...]
        [  132.248657] i801_smbus 0000:00:1f.3: Timeout waiting for interrupt!
        [  132.248669] i801_smbus 0000:00:1f.3: Transaction timeout
        [  132.452649] i801_smbus 0000:00:1f.3: Timeout waiting for interrupt!
        [  132.452662] i801_smbus 0000:00:1f.3: Transaction timeout
        [  132.467682] irq 17: nobody cared (try booting with the "irqpoll" option)

Existence of duplicate entries in a table returned by the _PRT method was confirmed by disassembling the ACPI DSTD table.

Linux used the first matching entry, which was incorrect. In order not to disrupt existing systems, use the first matching entry unless the pci=prtlast kernel parameter is used or a Dell Latitude E6500 laptop is detected.

Disclaimer: there is nothing really interesting connected to the SMBus controller on this laptop, but this change may help other systems.

Cc Jonathan Corbet <[email protected]>
Cc Bjorn Helgaas <[email protected]>
Cc "Rafael J. Wysocki" <[email protected]>
Cc Len Brown <[email protected]>
Cc Borislav Petkov <[email protected]>

---
v2: do not quote the disassembled ACPI DSDT table - for copyright reasons.

---
 .../admin-guide/kernel-parameters.txt         |  8 ++
 drivers/acpi/pci_irq.c                        | 89 ++++++++++++++++++-
 drivers/pci/pci.c                             |  9 ++
 3 files changed, 102 insertions(+), 4 deletions(-)
Documentation/admin-guide/kernel-parameters.txt +8 -0
index 426fa892d311..2ff351db10b8 100644
@@ -4190,6 +4190,14 @@
 				bridge windows. This is the default on modern
 				hardware. If you need to use this, please report
 				a bug to <[email protected]>.
+		prtlast		If the _PRT ACPI method returns duplicate
+				IRQ routing entries, use the last matching entry
+				for a given device. If the platform may be
+				affected by this problem, an error message is
+				printed to dmesg - this parameter is
+				ineffective otherwise. If you need to use this,
+				please report a bug to
+				<[email protected]>.
 		routeirq	Do IRQ routing for all PCI devices.
 				This is normally done in pci_enable_device(),
 				so this option is a temporary workaround
drivers/acpi/pci_irq.c +85 -4
index 08e15774fb9f..5cead840de0b 100644
@@ -196,12 +196,73 @@ static int acpi_pci_irq_check_entry(acpi_handle handle, struct pci_dev *dev,
 	return 0;
 }
 
+extern bool pci_prtlast;
+
+static const struct dmi_system_id pci_prtlast_dmi[] = {
+	{
+		.ident = "Dell Latitude E6500",
+		.matches = {
+			DMI_MATCH(DMI_SYS_VENDOR, "Dell Inc."),
+			DMI_MATCH(DMI_PRODUCT_NAME, "Latitude E6500"),
+		},
+	},
+	{ }
+};
+
+static bool acpi_pci_prt_use_last(struct acpi_prt_entry *curr,
+				  const char *current_source,
+				  const char *previous_match_source,
+				  int previous_match_index)
+{
+	bool ret;
+	const struct dmi_system_id *id;
+	const int msg_bufsize = 512;
+	char *msg = kmalloc(msg_bufsize, GFP_KERNEL);
+
+	if (!msg)
+		return false;
+
+	snprintf(msg, msg_bufsize,
+		 FW_BUG
+		 "ACPI _PRT returned duplicate IRQ routing entries for PCI device "
+		 "%04x:%02x:%02x[INT%c]: %s[%d] and %s[%d]. ",
+		 curr->id.segment, curr->id.bus, curr->id.device,
+		 pin_name(curr->pin),
+		 previous_match_source, previous_match_index,
+		 current_source, curr->index);
+
+	id = dmi_first_match(pci_prtlast_dmi);
+
+	if (id) {
+		pr_warn("%s%s detected, using last entry.\n",
+			msg, id->ident);
+
+		ret = true;
+	} else if (pci_prtlast) {
+		pr_err(
+"%sUsing last entry, as directed on the command line. If this helps, report a bug.\n",
+		       msg);
+
+		ret = true;
+	} else {
+		pr_err("%sIf necessary, use \"pci=prtlast\" and report a bug.\n",
+		       msg);
+
+		ret = false;
+	}
+
+	kfree(msg);
+	return ret;
+}
+
 static int acpi_pci_irq_find_prt_entry(struct pci_dev *dev,
-			  int pin, struct acpi_prt_entry **entry_ptr)
+			  int pin, struct acpi_prt_entry **entry_ptr_out)
 {
 	acpi_status status;
 	struct acpi_buffer buffer = { ACPI_ALLOCATE_BUFFER, NULL };
 	struct acpi_pci_routing_table *entry;
+	struct acpi_prt_entry *match = NULL;
+	const char *match_source = NULL;
 	acpi_handle handle = NULL;
 
 	if (dev->bus->bridge)
@@ -219,13 +280,33 @@ static int acpi_pci_irq_find_prt_entry(struct pci_dev *dev,
 
 	entry = buffer.pointer;
 	while (entry && (entry->length > 0)) {
-		if (!acpi_pci_irq_check_entry(handle, dev, pin,
-						 entry, entry_ptr))
-			break;
+		struct acpi_prt_entry *curr;
+
+		if (!acpi_pci_irq_check_entry(handle, dev, pin, entry, &curr)) {
+			if (!match) {
+				// first match
+				match = curr;
+				match_source = entry->source;
+			} else if (!acpi_pci_prt_use_last(curr,
+							  entry->source,
+							  match_source,
+							  match->index)) {
+				// duplicates found, use first entry
+				kfree(curr);
+			} else {
+				// duplicates found, use last entry
+				kfree(match);
+				match = curr;
+				match_source = entry->source;
+			}
+		}
+
 		entry = (struct acpi_pci_routing_table *)
 		    ((unsigned long)entry + entry->length);
 	}
 
+	*entry_ptr_out = match;
+
 	kfree(buffer.pointer);
 	return 0;
 }
drivers/pci/pci.c +9 -0
index 95bc329e74c0..a14a2e4e4197 100644
@@ -155,6 +155,11 @@ static bool pci_bridge_d3_disable;
 /* Force bridge_d3 for all PCIe ports */
 static bool pci_bridge_d3_force;
 
+#ifdef CONFIG_ACPI
+/* Use the last matching entry from the table returned by the _PRT ACPI method. */
+bool pci_prtlast;
+#endif
+
 static int __init pcie_port_pm_setup(char *str)
 {
 	if (!strcmp(str, "off"))
@@ -6896,6 +6901,10 @@ static int __init pci_setup(char *str)
 				pci_add_flags(PCI_SCAN_ALL_PCIE_DEVS);
 			} else if (!strncmp(str, "disable_acs_redir=", 18)) {
 				disable_acs_redir_param = str + 18;
+#ifdef CONFIG_ACPI
+			} else if (!strncmp(str, "prtlast", 7)) {
+				pci_prtlast = true;
+#endif
 			} else {
 				pr_err("PCI: Unknown option `%s'\n", str);
 			}

base-commit: 7e18e42e4b280c85b76967a9106a13ca61c16179

Signature · 1 line
2.25.1
#4 lore raw

W dniu 12.11.2022 o 01:20, Bjorn Helgaas pisze:

Quoted 14 lines

[+cc Jean, linux-i2c]

Quoted 3 lines · Mateusz Jończyk

On some platforms, the ACPI _PRT function returns duplicate interrupt routing entries. Linux uses the first matching entry, but sometimes the second matching entry contains the correct interrupt vector.

Rafael, Jean, what do you think about this? It seems like kind of a lot of infrastructure to deal with this oddness, but I'm not really opposed to it.

This is in i2c-i801.c, which seems to have some support for polling; maybe it could make smart enough to complain and automatically switch to polling if a timeout occurs.

Quoted 12 lines

Or maybe we scan the entire _PRT and let the match win (instead of the first as we do today).

Or ...?

Google finds a lot of hits for "i801_smbus" "timeout waiting for interrupt", but I can't tell whether they're a similar _PRT issue or something else.

Quoted 3 lines

This happens on a Dell Latitude E6500 laptop with the i2c-i801 Intel SMBus controller. This controller was nonfunctional unless its interrupt usage was disabled (using the "disable_features=0x10" module parameter).

Hello,

I have prepared a lean patch that only prints a warning when there are two matching entries in the table returned from _PRT (I will send it in the next e-mail). Perhaps it could be merged and then after a release or two it will be known how widespread this problem is.

Greetings,

Mateusz

#5 lore raw

PATCH 1/1 acpi,pci: warn about duplicate IRQ routing entries returned from _PRT

On some platforms, the ACPI _PRT function returns duplicate interrupt routing entries. Linux uses the first matching entry, but sometimes the second matching entry contains the correct interrupt vector.

To check how widespread this problem is, print a warning to dmesg if this is the case.

This happens on a Dell Latitude E6500 laptop with the i2c-i801 Intel SMBus controller. This controller was nonfunctional unless its interrupt usage was disabled (using the "disable_features=0x10" module parameter).

After investigation, it turned out that the driver was using an incorrect interrupt vector: in lspci output for this device there was:
Interrupt: pin B routed to IRQ 19
but after running i2cdetect (without using any i2c-i801 module parameters) the following was logged to dmesg:

	[...]
        i801_smbus 0000:00:1f.3: Timeout waiting for interrupt!
        i801_smbus 0000:00:1f.3: Transaction timeout
        irq 17: nobody cared (try booting with the "irqpoll" option)

Existence of duplicate entries in the table returned by the _PRT method was confirmed by disassembling the ACPI DSDT table.

Cc Bjorn Helgaas <[email protected]>
Cc "Rafael J. Wysocki" <[email protected]>
Cc Len Brown <[email protected]>
Cc Borislav Petkov <[email protected]>
Cc Jean Delvare <[email protected]>
---
 drivers/acpi/pci_irq.c | 25 ++++++++++++++++++++++---
 1 file changed, 22 insertions(+), 3 deletions(-)
drivers/acpi/pci_irq.c +22 -3
index 08e15774fb9f..c3168927446c 100644
@@ -203,6 +203,8 @@ static int acpi_pci_irq_find_prt_entry(struct pci_dev *dev,
 	struct acpi_buffer buffer = { ACPI_ALLOCATE_BUFFER, NULL };
 	struct acpi_pci_routing_table *entry;
 	acpi_handle handle = NULL;
+	struct acpi_prt_entry *match = NULL;
+	const char *match_int_source = NULL;
 
 	if (dev->bus->bridge)
 		handle = ACPI_HANDLE(dev->bus->bridge);
@@ -219,13 +221,30 @@ static int acpi_pci_irq_find_prt_entry(struct pci_dev *dev,
 
 	entry = buffer.pointer;
 	while (entry && (entry->length > 0)) {
-		if (!acpi_pci_irq_check_entry(handle, dev, pin,
-						 entry, entry_ptr))
-			break;
+		struct acpi_prt_entry *curr;
+
+		if (!acpi_pci_irq_check_entry(handle, dev, pin, entry, &curr)) {
+			if (match == NULL) {
+				match = curr;
+				match_int_source = entry->source;
+			} else {
+				pr_warn(FW_BUG
+				"ACPI _PRT returned duplicate IRQ routing entries for device "
+					"%04x:%02x:%02x[INT%c]: %s[%d] and %s[%d]. ",
+					curr->id.segment, curr->id.bus, curr->id.device,
+					pin_name(curr->pin),
+					match_int_source, match->index,
+					entry->source, curr->index);
+				// we use the first matching entry nonetheless
+			}
+		}
+
 		entry = (struct acpi_pci_routing_table *)
 		    ((unsigned long)entry + entry->length);
 	}
 
+	*entry_ptr = match;
+
 	kfree(buffer.pointer);
 	return 0;
 }

base-commit: f0c4d9fc9cc9462659728d168387191387e903cc

Signature · 1 line
2.25.1
#6 lore raw

PATCH v2 1/1 acpi,pci: warn about duplicate IRQ routing entries returned from _PRT

On some platforms, the ACPI _PRT function returns duplicate interrupt routing entries. Linux uses the first matching entry, but sometimes the second matching entry contains the correct interrupt vector.

Print a warning to dmesg if duplicate interrupt routing entries are present, so that we could check how many models are affected.

This happens on a Dell Latitude E6500 laptop with the i2c-i801 Intel SMBus controller. This controller was nonfunctional unless its interrupt usage was disabled (using the "disable_features=0x10" module parameter).

After investigation, it turned out that the driver was using an incorrect interrupt vector: in lspci output for this device there was:
Interrupt: pin B routed to IRQ 19
but after running i2cdetect (without using any i2c-i801 module parameters) the following was logged to dmesg:

        [...]
        i801_smbus 0000:00:1f.3: Timeout waiting for interrupt!
        i801_smbus 0000:00:1f.3: Transaction timeout
        i801_smbus 0000:00:1f.3: Timeout waiting for interrupt!
        i801_smbus 0000:00:1f.3: Transaction timeout
        irq 17: nobody cared (try booting with the "irqpoll" option)

Existence of duplicate entries in a table returned by the _PRT method was confirmed by disassembling the ACPI DSDT table.

Cc Bjorn Helgaas <[email protected]>
Cc "Rafael J. Wysocki" <[email protected]>
Cc Len Brown <[email protected]>
Cc Borislav Petkov <[email protected]>
Cc Jean Delvare <[email protected]>

--
v2: - add a newline at the end of the kernel log message,
- replace: "if (match == NULL)" -> "if (!match)"
- patch description tweaks.

Tested on two computers, including the affected Dell Latitude E6500 laptop.

 drivers/acpi/pci_irq.c | 25 ++++++++++++++++++++++---
 1 file changed, 22 insertions(+), 3 deletions(-)
drivers/acpi/pci_irq.c +22 -3
index 08e15774fb9f..a4e41b7b71ed 100644
@@ -203,6 +203,8 @@ static int acpi_pci_irq_find_prt_entry(struct pci_dev *dev,
 	struct acpi_buffer buffer = { ACPI_ALLOCATE_BUFFER, NULL };
 	struct acpi_pci_routing_table *entry;
 	acpi_handle handle = NULL;
+	struct acpi_prt_entry *match = NULL;
+	const char *match_int_source = NULL;
 
 	if (dev->bus->bridge)
 		handle = ACPI_HANDLE(dev->bus->bridge);
@@ -219,13 +221,30 @@ static int acpi_pci_irq_find_prt_entry(struct pci_dev *dev,
 
 	entry = buffer.pointer;
 	while (entry && (entry->length > 0)) {
-		if (!acpi_pci_irq_check_entry(handle, dev, pin,
-						 entry, entry_ptr))
-			break;
+		struct acpi_prt_entry *curr;
+
+		if (!acpi_pci_irq_check_entry(handle, dev, pin, entry, &curr)) {
+			if (!match) {
+				match = curr;
+				match_int_source = entry->source;
+			} else {
+				pr_warn(FW_BUG
+				"ACPI _PRT returned duplicate IRQ routing entries for device "
+					"%04x:%02x:%02x[INT%c]: %s[%d] and %s[%d].\n",
+					curr->id.segment, curr->id.bus, curr->id.device,
+					pin_name(curr->pin),
+					match_int_source, match->index,
+					entry->source, curr->index);
+				// we use the first matching entry nonetheless
+			}
+		}
+
 		entry = (struct acpi_pci_routing_table *)
 		    ((unsigned long)entry + entry->length);
 	}
 
+	*entry_ptr = match;
+
 	kfree(buffer.pointer);
 	return 0;
 }

base-commit: f0c4d9fc9cc9462659728d168387191387e903cc

Signature · 1 line
2.25.1
Reviewed-by Tested-by on v2 1/1
#7 lore raw

Re: [PATCH v2] acpi,pci: warn about duplicate IRQ routing entries returned from _PRT

Hi Mateusz,

Quoted 6 lines · Mateusz Jończyk

On some platforms, the ACPI _PRT function returns duplicate interrupt routing entries. Linux uses the first matching entry, but sometimes the second matching entry contains the correct interrupt vector.

Print a warning to dmesg if duplicate interrupt routing entries are present, so that we could check how many models are affected.

Excellent idea. We want hardware manufacturers to fix such bugs in the firmware, and the best way for this to happen is to report them whenever they are encountered.

Quoted 17 lines

This happens on a Dell Latitude E6500 laptop with the i2c-i801 Intel SMBus controller. This controller was nonfunctional unless its interrupt usage was disabled (using the "disable_features=0x10" module parameter).

After investigation, it turned out that the driver was using an incorrect interrupt vector: in lspci output for this device there was:
Interrupt: pin B routed to IRQ 19
but after running i2cdetect (without using any i2c-i801 module parameters) the following was logged to dmesg:

        [...]
        i801_smbus 0000:00:1f.3: Timeout waiting for interrupt!
        i801_smbus 0000:00:1f.3: Transaction timeout
        i801_smbus 0000:00:1f.3: Timeout waiting for interrupt!
        i801_smbus 0000:00:1f.3: Transaction timeout
        irq 17: nobody cared (try booting with the "irqpoll" option)
Quoted 2 lines

Existence of duplicate entries in a table returned by the _PRT method was confirmed by disassembling the ACPI DSDT table.

Excuse a probably stupid question, but what would happen if we would plain ignore the IRQ routing information from ACPI in this case? Would we fallback to some pure-PCI routing logic which may have a chance to find the right IRQ routing (matching the second ACPI routing entry in this case)?

Quoted 37 lines
Cc Bjorn Helgaas <[email protected]>
Cc "Rafael J. Wysocki" <[email protected]>
Cc Len Brown <[email protected]>
Cc Borislav Petkov <[email protected]>
Cc Jean Delvare <[email protected]>

--
v2: - add a newline at the end of the kernel log message,
- replace: "if (match == NULL)" -> "if (!match)"
- patch description tweaks.

Tested on two computers, including the affected Dell Latitude E6500 laptop.

 drivers/acpi/pci_irq.c | 25 ++++++++++++++++++++++---
 1 file changed, 22 insertions(+), 3 deletions(-)
drivers/acpi/pci_irq.c +2 -3
index 08e15774fb9f..a4e41b7b71ed 100644
@@ -203,6 +203,8 @@ static int acpi_pci_irq_find_prt_entry(struct pci_dev *dev,
 	struct acpi_buffer buffer = { ACPI_ALLOCATE_BUFFER, NULL };
 	struct acpi_pci_routing_table *entry;
 	acpi_handle handle = NULL;
+	struct acpi_prt_entry *match = NULL;
+	const char *match_int_source = NULL;
 
 	if (dev->bus->bridge)
 		handle = ACPI_HANDLE(dev->bus->bridge);
@@ -219,13 +221,30 @@ static int acpi_pci_irq_find_prt_entry(struct pci_dev *dev,
 
 	entry = buffer.pointer;
 	while (entry && (entry->length > 0)) {
-		if (!acpi_pci_irq_check_entry(handle, dev, pin,
-						 entry, entry_ptr))
-			break;

drivers/acpi/pci_irq.c static int acpi_pci_irq_find_prt_entry(struct pci_dev *dev, · line 233

Quoted 10 lines
drivers/acpi/pci_irq.c +10 -0
+		struct acpi_prt_entry *curr;
+
+		if (!acpi_pci_irq_check_entry(handle, dev, pin, entry, &curr)) {
+			if (!match) {
+				match = curr;
+				match_int_source = entry->source;
+			} else {
+				pr_warn(FW_BUG
+				"ACPI _PRT returned duplicate IRQ routing entries for device "
+					"%04x:%02x:%02x[INT%c]: %s[%d] and %s[%d].\n",

The beginning of the string should be aligned with the opening parenthesis, and the string should be on a single line (this is a encouraged exception to the 80-column rule). I would also omit the tailing dot for consistency.

drivers/acpi/pci_irq.c static int acpi_pci_irq_find_prt_entry(struct pci_dev *dev, · line 234

Quoted 1 line
drivers/acpi/pci_irq.c +1 -0
+					curr->id.segment, curr->id.bus, curr->id.device,

Is the IRQ per PCI device, or per PCI function? If the latter, then you should print "%02x.%x" instead of just "%02x", with the extra element being curr->id.function.

drivers/acpi/pci_irq.c static int acpi_pci_irq_find_prt_entry(struct pci_dev *dev, · line 238

Quoted 4 lines
drivers/acpi/pci_irq.c +4 -0
+					pin_name(curr->pin),
+					match_int_source, match->index,
+					entry->source, curr->index);
+				// we use the first matching entry nonetheless

The rest of the file uses /* C89-style comments */ so I would stick to that for consistency.

drivers/acpi/pci_irq.c static int acpi_pci_irq_find_prt_entry(struct pci_dev *dev, · line 250

Quoted 12 lines
drivers/acpi/pci_irq.c +5 -0
+			}
+		}
+
 		entry = (struct acpi_pci_routing_table *)
 		    ((unsigned long)entry + entry->length);
 	}
 
+	*entry_ptr = match;
+
 	kfree(buffer.pointer);
 	return 0;
 }

(Tested on a Dell OptiPlex 9020 not affected by the problem.)

Signature · 2 lines
Jean Delvare
SUSE L3 Support
#8 lore raw

Re: [PATCH v2] acpi,pci: warn about duplicate IRQ routing entries returned from _PRT

Hello,

W dniu 15.11.2022 o 09:36, Jean Delvare pisze:

Quoted 30 lines

Hi Mateusz,

Quoted 6 lines · Mateusz Jończyk

On some platforms, the ACPI _PRT function returns duplicate interrupt routing entries. Linux uses the first matching entry, but sometimes the second matching entry contains the correct interrupt vector.

Print a warning to dmesg if duplicate interrupt routing entries are present, so that we could check how many models are affected.

Excellent idea. We want hardware manufacturers to fix such bugs in the firmware, and the best way for this to happen is to report them whenever they are encountered.

Quoted 17 lines

This happens on a Dell Latitude E6500 laptop with the i2c-i801 Intel SMBus controller. This controller was nonfunctional unless its interrupt usage was disabled (using the "disable_features=0x10" module parameter).

After investigation, it turned out that the driver was using an incorrect interrupt vector: in lspci output for this device there was:
Interrupt: pin B routed to IRQ 19
but after running i2cdetect (without using any i2c-i801 module parameters) the following was logged to dmesg:

        [...]
        i801_smbus 0000:00:1f.3: Timeout waiting for interrupt!
        i801_smbus 0000:00:1f.3: Transaction timeout
        i801_smbus 0000:00:1f.3: Timeout waiting for interrupt!
        i801_smbus 0000:00:1f.3: Transaction timeout
        irq 17: nobody cared (try booting with the "irqpoll" option)
Quoted 7 lines
Quoted 2 lines

Existence of duplicate entries in a table returned by the _PRT method was confirmed by disassembling the ACPI DSDT table.

Excuse a probably stupid question, but what would happen if we would plain ignore the IRQ routing information from ACPI in this case? Would we fallback to some pure-PCI routing logic which may have a chance to find the right IRQ routing (matching the second ACPI routing entry in this case)?

From what I understand, the PCI IRQ routing information is not discoverable by probing the hardware (in the general case), it has to be obtained from the ACPI tables (or perhaps from the obsolete MP tables, also provided by firmware). See https://docs.kernel.org/PCI/acpi-info.html :

Quoted 5 lines

For example, there’s no standard hardware mechanism for enumerating PCI host bridges, so the ACPI namespace must describe each host bridge, the method for accessing PCI config space below it, the address space windows the host bridge forwards to PCI (using _CRS), and the routing of legacy INTx interrupts (using _PRT).

(a PCI host bridge connects the CPU cores to the PCI bus, it is the root of the PCI device tree. This patch concerns the "legacy INTx interrupts" as above).

In the case of this particular laptop, however, it should be possible to obtain the information by reading chipset registers, which are documented at
https://www.intel.com/content/www/us/en/io/io-controller-hub-9-datasheet.html
But this is difficult to implement in every case.

Quoted 43 lines
Quoted 43 lines
Cc Bjorn Helgaas <[email protected]>
Cc "Rafael J. Wysocki" <[email protected]>
Cc Len Brown <[email protected]>
Cc Borislav Petkov <[email protected]>
Cc Jean Delvare <[email protected]>

--
v2: - add a newline at the end of the kernel log message,
- replace: "if (match == NULL)" -> "if (!match)"
- patch description tweaks.

Tested on two computers, including the affected Dell Latitude E6500 laptop.

 drivers/acpi/pci_irq.c | 25 ++++++++++++++++++++++---
 1 file changed, 22 insertions(+), 3 deletions(-)
drivers/acpi/pci_irq.c +8 -3
index 08e15774fb9f..a4e41b7b71ed 100644
@@ -203,6 +203,8 @@ static int acpi_pci_irq_find_prt_entry(struct pci_dev *dev,
 	struct acpi_buffer buffer = { ACPI_ALLOCATE_BUFFER, NULL };
 	struct acpi_pci_routing_table *entry;
 	acpi_handle handle = NULL;
+	struct acpi_prt_entry *match = NULL;
+	const char *match_int_source = NULL;
 
 	if (dev->bus->bridge)
 		handle = ACPI_HANDLE(dev->bus->bridge);
@@ -219,13 +221,30 @@ static int acpi_pci_irq_find_prt_entry(struct pci_dev *dev,
 
 	entry = buffer.pointer;
 	while (entry && (entry->length > 0)) {
-		if (!acpi_pci_irq_check_entry(handle, dev, pin,
-						 entry, entry_ptr))
-			break;
+		struct acpi_prt_entry *curr;
+
+		if (!acpi_pci_irq_check_entry(handle, dev, pin, entry, &curr)) {
+			if (!match) {
+				match = curr;
+				match_int_source = entry->source;
Quoted 8 lines

drivers/acpi/pci_irq.c static int acpi_pci_irq_find_prt_entry(struct pci_dev *dev, · line 233

Quoted 4 lines
drivers/acpi/pci_irq.c +4 -0
+			} else {
+				pr_warn(FW_BUG
+				"ACPI _PRT returned duplicate IRQ routing entries for device "
+					"%04x:%02x:%02x[INT%c]: %s[%d] and %s[%d].\n",

The beginning of the string should be aligned with the opening parenthesis, and the string should be on a single line (this is a encouraged exception to the 80-column rule). I would also omit the tailing dot for consistency.

OK

Quoted 4 lines

drivers/acpi/pci_irq.c static int acpi_pci_irq_find_prt_entry(struct pci_dev *dev, · line 234

Quoted 1 line
drivers/acpi/pci_irq.c +1 -0
+					curr->id.segment, curr->id.bus, curr->id.device,

Is the IRQ per PCI device, or per PCI function? If the latter, then you should print "%02x.%x" instead of just "%02x", with the extra element being curr->id.function.

This is per PCI device.

[snip]

Quoted 5 lines

(Tested on a Dell OptiPlex 9020 not affected by the problem.)

Thank you for reviewing.

Greetings,

Mateusz