Skip to contentSkip to conversation

PCI: Avoid putting some root ports into D3 on TUXEDO Sirius Gen1

PATCH Superseded Werner Sembach 2 participants 6 messages Dec 20, 2024 – Jan 14, 2025 PCI

v4 of 6 Superseded Latest is v6 →

lore
Share & tools
lore
#0 lore raw

PATCH v4 1/1 PCI: Avoid putting some root ports into D3 on TUXEDO Sirius Gen1

From: Mario Limonciello <[email protected]>

commit 9d26d3a8f1b0 ("PCI: Put PCIe ports into D3 during suspend") sets the policy that all PCIe ports are allowed to use D3. When the system is suspended if the port is not power manageable by the platform and won't be used for wakeup via a PME this sets up the policy for these ports to go into D3hot.

This policy generally makes sense from an OSPM perspective but it leads to problems with wakeup from suspend on the TUXEDO Sirius 16 Gen 1 with a specific old BIOS. This manifests as a system hang.

On the affected Device + BIOS combination, add a quirk for the root port of the problematic controller to ensure that these root ports are not put into D3hot at suspend.

This patch is based on
https://lore.kernel.org/linux-pci/[email protected]/
but with the added condition both in the documentation and in the code to apply only to the TUXEDO Sirius 16 Gen 1 with a specific old BIOS and only the affected root ports.

Fixes 9d26d3a8f1b0 ("PCI: Put PCIe ports into D3 during suspend")
---
 drivers/pci/quirks.c | 30 ++++++++++++++++++++++++++++++
 1 file changed, 30 insertions(+)
drivers/pci/quirks.c +30 -0
index 76f4df75b08a1..d2f45c3e24c0a 100644
@@ -3908,6 +3908,36 @@ static void quirk_apple_poweroff_thunderbolt(struct pci_dev *dev)
 DECLARE_PCI_FIXUP_SUSPEND_LATE(PCI_VENDOR_ID_INTEL,
 			       PCI_DEVICE_ID_INTEL_CACTUS_RIDGE_4C,
 			       quirk_apple_poweroff_thunderbolt);
+
+/*
+ * Putting PCIe root ports on Ryzen SoCs with USB4 controllers into D3hot
+ * may cause problems when the system attempts wake up from s2idle.
+ *
+ * On the TUXEDO Sirius 16 Gen 1 with a specific old BIOS this manifests as
+ * a system hang.
+ */
+static const struct dmi_system_id quirk_ryzen_rp_d3_dmi_table[] = {
+	{
+		.matches = {
+			DMI_EXACT_MATCH(DMI_SYS_VENDOR, "TUXEDO"),
+			DMI_EXACT_MATCH(DMI_BOARD_NAME, "APX958"),
+			DMI_EXACT_MATCH(DMI_BIOS_VERSION, "V1.00A00_20240108"),
+		},
+	},
+	{}
+};
+
+static void quirk_ryzen_rp_d3(struct pci_dev *pdev)
+{
+	struct pci_dev *root_pdev;
+
+	if (dmi_check_system(quirk_ryzen_rp_d3_dmi_table)) {
+		root_pdev = pcie_find_root_port(pdev);
+		if (root_pdev && !acpi_pci_power_manageable(root_pdev))
+			root_pdev->dev_flags |= PCI_DEV_FLAGS_NO_D3;
+	}
+}
+DECLARE_PCI_FIXUP_FINAL(PCI_VENDOR_ID_AMD, 0x1502, quirk_ryzen_rp_d3);
 #endif
 
 /*
Signature · 1 line
2.43.0
#1 lore raw
Quoted 22 lines · Werner Sembach

From: Mario Limonciello <[email protected]>

commit 9d26d3a8f1b0 ("PCI: Put PCIe ports into D3 during suspend") sets the policy that all PCIe ports are allowed to use D3. When the system is suspended if the port is not power manageable by the platform and won't be used for wakeup via a PME this sets up the policy for these ports to go into D3hot.

This policy generally makes sense from an OSPM perspective but it leads to problems with wakeup from suspend on the TUXEDO Sirius 16 Gen 1 with a specific old BIOS. This manifests as a system hang.

On the affected Device + BIOS combination, add a quirk for the root port of the problematic controller to ensure that these root ports are not put into D3hot at suspend.

This patch is based on
https://lore.kernel.org/linux-pci/[email protected]/
but with the added condition both in the documentation and in the code to apply only to the TUXEDO Sirius 16 Gen 1 with a specific old BIOS and only the affected root ports.

Quoted 7 lines
Fixes 9d26d3a8f1b0 ("PCI: Put PCIe ports into D3 during suspend")

So I don't think this should have my S-o-b. At most it should Suggested-by: or Co-developed-by: since it was based on my original patch.

Quoted 2 lines
---
  drivers/pci/quirks.c | 30 ++++++++++++++++++++++++++++++

I think a better location for this is arch/x86/pci/fixup.c, similar to how we have https://git.kernel.org/torvalds/c/7d08f21f8c630

thoughts?

Quoted 44 lines
  1 file changed, 30 insertions(+)
drivers/pci/quirks.c +30 -0
index 76f4df75b08a1..d2f45c3e24c0a 100644
@@ -3908,6 +3908,36 @@ static void quirk_apple_poweroff_thunderbolt(struct pci_dev *dev)
  DECLARE_PCI_FIXUP_SUSPEND_LATE(PCI_VENDOR_ID_INTEL,
  			       PCI_DEVICE_ID_INTEL_CACTUS_RIDGE_4C,
  			       quirk_apple_poweroff_thunderbolt);
+
+/*
+ * Putting PCIe root ports on Ryzen SoCs with USB4 controllers into D3hot
+ * may cause problems when the system attempts wake up from s2idle.
+ *
+ * On the TUXEDO Sirius 16 Gen 1 with a specific old BIOS this manifests as
+ * a system hang.
+ */
+static const struct dmi_system_id quirk_ryzen_rp_d3_dmi_table[] = {
+	{
+		.matches = {
+			DMI_EXACT_MATCH(DMI_SYS_VENDOR, "TUXEDO"),
+			DMI_EXACT_MATCH(DMI_BOARD_NAME, "APX958"),
+			DMI_EXACT_MATCH(DMI_BIOS_VERSION, "V1.00A00_20240108"),
+		},
+	},
+	{}
+};
+
+static void quirk_ryzen_rp_d3(struct pci_dev *pdev)
+{
+	struct pci_dev *root_pdev;
+
+	if (dmi_check_system(quirk_ryzen_rp_d3_dmi_table)) {
+		root_pdev = pcie_find_root_port(pdev);
+		if (root_pdev && !acpi_pci_power_manageable(root_pdev))
+			root_pdev->dev_flags |= PCI_DEV_FLAGS_NO_D3;
+	}
+}
+DECLARE_PCI_FIXUP_FINAL(PCI_VENDOR_ID_AMD, 0x1502, quirk_ryzen_rp_d3);
  #endif
 
  /*
#2 lore raw

Am 08.01.25 um 22:26 schrieb Mario Limonciello:

Quoted 32 lines · Werner Sembach
Quoted 23 lines

From: Mario Limonciello <[email protected]>

commit 9d26d3a8f1b0 ("PCI: Put PCIe ports into D3 during suspend") sets the policy that all PCIe ports are allowed to use D3.  When the system is suspended if the port is not power manageable by the platform and won't be used for wakeup via a PME this sets up the policy for these ports to go into D3hot.

This policy generally makes sense from an OSPM perspective but it leads to problems with wakeup from suspend on the TUXEDO Sirius 16 Gen 1 with a specific old BIOS. This manifests as a system hang.

On the affected Device + BIOS combination, add a quirk for the root port of the problematic controller to ensure that these root ports are not put into D3hot at suspend.

This patch is based on
https://lore.kernel.org/linux-pci/[email protected]/

but with the added condition both in the documentation and in the code to apply only to the TUXEDO Sirius 16 Gen 1 with a specific old BIOS and only the affected root ports.

Quoted 7 lines
Fixes 9d26d3a8f1b0 ("PCI: Put PCIe ports into D3 during suspend")
Quoted 1 line

So I don't think this should have my S-o-b.  At most it should Suggested-by: or Co-developed-by: since it was based on my original patch.

kk

Quoted 7 lines
Quoted 3 lines

---
  drivers/pci/quirks.c | 30 ++++++++++++++++++++++++++++++

I think a better location for this is arch/x86/pci/fixup.c, similar to how we have https://git.kernel.org/torvalds/c/7d08f21f8c630

thoughts?

Fine with me

I will make a v5

Quoted 46 lines
Quoted 45 lines
  1 file changed, 30 insertions(+)
drivers/pci/quirks.c +30 -0
index 76f4df75b08a1..d2f45c3e24c0a 100644
@@ -3908,6 +3908,36 @@ static void quirk_apple_poweroff_thunderbolt(struct 
 pci_dev *dev)
   DECLARE_PCI_FIXUP_SUSPEND_LATE(PCI_VENDOR_ID_INTEL,
                      PCI_DEVICE_ID_INTEL_CACTUS_RIDGE_4C,
                      quirk_apple_poweroff_thunderbolt);
+
+/*
+ * Putting PCIe root ports on Ryzen SoCs with USB4 controllers into D3hot
+ * may cause problems when the system attempts wake up from s2idle.
+ *
+ * On the TUXEDO Sirius 16 Gen 1 with a specific old BIOS this manifests as
+ * a system hang.
+ */
+static const struct dmi_system_id quirk_ryzen_rp_d3_dmi_table[] = {
+    {
+        .matches = {
+            DMI_EXACT_MATCH(DMI_SYS_VENDOR, "TUXEDO"),
+            DMI_EXACT_MATCH(DMI_BOARD_NAME, "APX958"),
+            DMI_EXACT_MATCH(DMI_BIOS_VERSION, "V1.00A00_20240108"),
+        },
+    },
+    {}
+};
+
+static void quirk_ryzen_rp_d3(struct pci_dev *pdev)
+{
+    struct pci_dev *root_pdev;
+
+    if (dmi_check_system(quirk_ryzen_rp_d3_dmi_table)) {
+        root_pdev = pcie_find_root_port(pdev);
+        if (root_pdev && !acpi_pci_power_manageable(root_pdev))
+            root_pdev->dev_flags |= PCI_DEV_FLAGS_NO_D3;
+    }
+}
+DECLARE_PCI_FIXUP_FINAL(PCI_VENDOR_ID_AMD, 0x1502, quirk_ryzen_rp_d3);
   #endif
     /*
#3 lore raw

Am 10.01.25 um 18:15 schrieb Werner Sembach:

Quoted 40 lines

Am 08.01.25 um 22:26 schrieb Mario Limonciello:

Quoted 32 lines · Werner Sembach
Quoted 23 lines

From: Mario Limonciello <[email protected]>

commit 9d26d3a8f1b0 ("PCI: Put PCIe ports into D3 during suspend") sets the policy that all PCIe ports are allowed to use D3.  When the system is suspended if the port is not power manageable by the platform and won't be used for wakeup via a PME this sets up the policy for these ports to go into D3hot.

This policy generally makes sense from an OSPM perspective but it leads to problems with wakeup from suspend on the TUXEDO Sirius 16 Gen 1 with a specific old BIOS. This manifests as a system hang.

On the affected Device + BIOS combination, add a quirk for the root port of the problematic controller to ensure that these root ports are not put into D3hot at suspend.

This patch is based on
https://lore.kernel.org/linux-pci/[email protected]/

but with the added condition both in the documentation and in the code to apply only to the TUXEDO Sirius 16 Gen 1 with a specific old BIOS and only the affected root ports.

Quoted 7 lines
Fixes 9d26d3a8f1b0 ("PCI: Put PCIe ports into D3 during suspend")
Quoted 1 line

So I don't think this should have my S-o-b.  At most it should Suggested-by: or Co-developed-by: since it was based on my original patch.

kk

Quoted 4 lines
Quoted 3 lines

---
  drivers/pci/quirks.c | 30 ++++++++++++++++++++++++++++++

Quoted 7 lines
Quoted 3 lines

I think a better location for this is arch/x86/pci/fixup.c, similar to how we have https://git.kernel.org/torvalds/c/7d08f21f8c630

thoughts?

Fine with me

I will make a v5

In fixup.c i don't have access to acpi_pci_power_manageable, but since DECLARE_PCI_FIXUP_FINAL(PCI_VENDOR_ID_AMD, 0x1502, quirk_ryzen_rp_d3); matches to only one device anyways can i just skip it?

Quoted 47 lines
Quoted 47 lines
Quoted 46 lines

  1 file changed, 30 insertions(+)
drivers/pci/quirks.c +30 -0
index 76f4df75b08a1..d2f45c3e24c0a 100644
@@ -3908,6 +3908,36 @@ static void quirk_apple_poweroff_thunderbolt(struct 
 pci_dev *dev)
   DECLARE_PCI_FIXUP_SUSPEND_LATE(PCI_VENDOR_ID_INTEL,
                      PCI_DEVICE_ID_INTEL_CACTUS_RIDGE_4C,
                      quirk_apple_poweroff_thunderbolt);
+
+/*
+ * Putting PCIe root ports on Ryzen SoCs with USB4 controllers into D3hot
+ * may cause problems when the system attempts wake up from s2idle.
+ *
+ * On the TUXEDO Sirius 16 Gen 1 with a specific old BIOS this manifests as
+ * a system hang.
+ */
+static const struct dmi_system_id quirk_ryzen_rp_d3_dmi_table[] = {
+    {
+        .matches = {
+            DMI_EXACT_MATCH(DMI_SYS_VENDOR, "TUXEDO"),
+            DMI_EXACT_MATCH(DMI_BOARD_NAME, "APX958"),
+            DMI_EXACT_MATCH(DMI_BIOS_VERSION, "V1.00A00_20240108"),
+        },
+    },
+    {}
+};
+
+static void quirk_ryzen_rp_d3(struct pci_dev *pdev)
+{
+    struct pci_dev *root_pdev;
+
+    if (dmi_check_system(quirk_ryzen_rp_d3_dmi_table)) {
+        root_pdev = pcie_find_root_port(pdev);
+        if (root_pdev && !acpi_pci_power_manageable(root_pdev))
+            root_pdev->dev_flags |= PCI_DEV_FLAGS_NO_D3;
+    }
+}
+DECLARE_PCI_FIXUP_FINAL(PCI_VENDOR_ID_AMD, 0x1502, quirk_ryzen_rp_d3);
   #endif
     /*
#4 lore raw
Quoted 41 lines · Werner Sembach

Am 10.01.25 um 18:15 schrieb Werner Sembach:

Quoted 39 lines

Am 08.01.25 um 22:26 schrieb Mario Limonciello:

Quoted 31 lines · Werner Sembach
Quoted 22 lines

From: Mario Limonciello <[email protected]>

commit 9d26d3a8f1b0 ("PCI: Put PCIe ports into D3 during suspend") sets the policy that all PCIe ports are allowed to use D3.  When the system is suspended if the port is not power manageable by the platform and won't be used for wakeup via a PME this sets up the policy for these ports to go into D3hot.

This policy generally makes sense from an OSPM perspective but it leads to problems with wakeup from suspend on the TUXEDO Sirius 16 Gen 1 with a specific old BIOS. This manifests as a system hang.

On the affected Device + BIOS combination, add a quirk for the root port of the problematic controller to ensure that these root ports are not put into D3hot at suspend.

This patch is based on
https://lore.kernel.org/linux-pci/20230708214457.1229-2- [email protected]/ but with the added condition both in the documentation and in the code to apply only to the TUXEDO Sirius 16 Gen 1 with a specific old BIOS and only the affected root ports.

Quoted 7 lines
Fixes 9d26d3a8f1b0 ("PCI: Put PCIe ports into D3 during suspend")
Quoted 1 line

So I don't think this should have my S-o-b.  At most it should Suggested-by: or Co-developed-by: since it was based on my original patch.

kk

Quoted 4 lines
Quoted 3 lines

---
  drivers/pci/quirks.c | 30 ++++++++++++++++++++++++++++++

Quoted 8 lines
Quoted 7 lines
Quoted 3 lines

I think a better location for this is arch/x86/pci/fixup.c, similar to how we have https://git.kernel.org/torvalds/c/7d08f21f8c630

thoughts?

Fine with me

I will make a v5

In fixup.c i don't have access to acpi_pci_power_manageable, but since DECLARE_PCI_FIXUP_FINAL(PCI_VENDOR_ID_AMD, 0x1502, quirk_ryzen_rp_d3); matches to only one device anyways can i just skip it?

Is it just a header problem? Maybe you can just add the header?

I think if you want to drop it that should be ok, but as it's a problem in your BIOS (specifically) and only matching your platform combo I would suggest renaming the function and struct to quirk_tuxeo_rp_d3 and quirk_tuxedo_rp_d3_dmi_table.

Quoted 49 lines
Quoted 49 lines
Quoted 49 lines
Quoted 48 lines

  1 file changed, 30 insertions(+)
drivers/pci/quirks.c +3 -0
index 76f4df75b08a1..d2f45c3e24c0a 100644
@@ -3908,6 +3908,36 @@ static void 
 quirk_apple_poweroff_thunderbolt(struct pci_dev *dev)
   DECLARE_PCI_FIXUP_SUSPEND_LATE(PCI_VENDOR_ID_INTEL,
                      PCI_DEVICE_ID_INTEL_CACTUS_RIDGE_4C,
                      quirk_apple_poweroff_thunderbolt);
+
+/*
+ * Putting PCIe root ports on Ryzen SoCs with USB4 controllers into 

D3hot
+ * may cause problems when the system attempts wake up from s2idle.
+ *
+ * On the TUXEDO Sirius 16 Gen 1 with a specific old BIOS this manifests as
+ * a system hang.
+ */

drivers/pci/quirks.c +22 -0
+static const struct dmi_system_id quirk_ryzen_rp_d3_dmi_table[] = {
+    {
+        .matches = {
+            DMI_EXACT_MATCH(DMI_SYS_VENDOR, "TUXEDO"),
+            DMI_EXACT_MATCH(DMI_BOARD_NAME, "APX958"),
+            DMI_EXACT_MATCH(DMI_BIOS_VERSION, "V1.00A00_20240108"),
+        },
+    },
+    {}
+};
+
+static void quirk_ryzen_rp_d3(struct pci_dev *pdev)
+{
+    struct pci_dev *root_pdev;
+
+    if (dmi_check_system(quirk_ryzen_rp_d3_dmi_table)) {
+        root_pdev = pcie_find_root_port(pdev);
+        if (root_pdev && !acpi_pci_power_manageable(root_pdev))
+            root_pdev->dev_flags |= PCI_DEV_FLAGS_NO_D3;
+    }
+}
+DECLARE_PCI_FIXUP_FINAL(PCI_VENDOR_ID_AMD, 0x1502, quirk_ryzen_rp_d3);

  #endif
    /*

#5 lore raw

Am 10.01.25 um 19:20 schrieb Mario Limonciello:

Quoted 44 lines · Werner Sembach
Quoted 43 lines

Am 10.01.25 um 18:15 schrieb Werner Sembach:

Quoted 41 lines

Am 08.01.25 um 22:26 schrieb Mario Limonciello:

Quoted 31 lines · Werner Sembach
Quoted 22 lines

From: Mario Limonciello <[email protected]>

commit 9d26d3a8f1b0 ("PCI: Put PCIe ports into D3 during suspend") sets the policy that all PCIe ports are allowed to use D3.  When the system is suspended if the port is not power manageable by the platform and won't be used for wakeup via a PME this sets up the policy for these ports to go into D3hot.

This policy generally makes sense from an OSPM perspective but it leads to problems with wakeup from suspend on the TUXEDO Sirius 16 Gen 1 with a specific old BIOS. This manifests as a system hang.

On the affected Device + BIOS combination, add a quirk for the root port of the problematic controller to ensure that these root ports are not put into D3hot at suspend.

This patch is based on
https://lore.kernel.org/linux-pci/20230708214457.1229-2- [email protected]/ but with the added condition both in the documentation and in the code to apply only to the TUXEDO Sirius 16 Gen 1 with a specific old BIOS and only the affected root ports.

Quoted 7 lines
Fixes 9d26d3a8f1b0 ("PCI: Put PCIe ports into D3 during suspend")
Quoted 1 line

So I don't think this should have my S-o-b.  At most it should Suggested-by: or Co-developed-by: since it was based on my original patch.

kk

Quoted 6 lines
Quoted 3 lines

---
  drivers/pci/quirks.c | 30 ++++++++++++++++++++++++++++++

I think a better location for this is arch/x86/pci/fixup.c, similar to how we have https://git.kernel.org/torvalds/c/7d08f21f8c630

Quoted 8 lines
Quoted 6 lines
Quoted 5 lines
Quoted 1 line

thoughts?

Fine with me

I will make a v5

In fixup.c i don't have access to acpi_pci_power_manageable, but since DECLARE_PCI_FIXUP_FINAL(PCI_VENDOR_ID_AMD, 0x1502, quirk_ryzen_rp_d3); matches to only one device anyways can i just skip it?

Is it just a header problem?  Maybe you can just add the header?

It's a local header to the other folder. so no.

Quoted 2 lines

I think if you want to drop it that should be ok, but as it's a problem in your BIOS (specifically) and only matching your platform combo I would suggest renaming the function and struct to quirk_tuxeo_rp_d3 and quirk_tuxedo_rp_d3_dmi_table.

Forgot this in v5. I will send a v6.

Quoted 49 lines
Quoted 48 lines
Quoted 48 lines
Quoted 48 lines
Quoted 47 lines


  1 file changed, 30 insertions(+)
drivers/pci/quirks.c +30 -0
index 76f4df75b08a1..d2f45c3e24c0a 100644
@@ -3908,6 +3908,36 @@ static void quirk_apple_poweroff_thunderbolt(struct 
 pci_dev *dev)
   DECLARE_PCI_FIXUP_SUSPEND_LATE(PCI_VENDOR_ID_INTEL,
                      PCI_DEVICE_ID_INTEL_CACTUS_RIDGE_4C,
                      quirk_apple_poweroff_thunderbolt);
+
+/*
+ * Putting PCIe root ports on Ryzen SoCs with USB4 controllers into D3hot
+ * may cause problems when the system attempts wake up from s2idle.
+ *
+ * On the TUXEDO Sirius 16 Gen 1 with a specific old BIOS this manifests as
+ * a system hang.
+ */
+static const struct dmi_system_id quirk_ryzen_rp_d3_dmi_table[] = {
+    {
+        .matches = {
+            DMI_EXACT_MATCH(DMI_SYS_VENDOR, "TUXEDO"),
+            DMI_EXACT_MATCH(DMI_BOARD_NAME, "APX958"),
+            DMI_EXACT_MATCH(DMI_BIOS_VERSION, "V1.00A00_20240108"),
+        },
+    },
+    {}
+};
+
+static void quirk_ryzen_rp_d3(struct pci_dev *pdev)
+{
+    struct pci_dev *root_pdev;
+
+    if (dmi_check_system(quirk_ryzen_rp_d3_dmi_table)) {
+        root_pdev = pcie_find_root_port(pdev);
+        if (root_pdev && !acpi_pci_power_manageable(root_pdev))
+            root_pdev->dev_flags |= PCI_DEV_FLAGS_NO_D3;
+    }
+}
+DECLARE_PCI_FIXUP_FINAL(PCI_VENDOR_ID_AMD, 0x1502, quirk_ryzen_rp_d3);
   #endif
     /*