[PATCH 4.12 060/196] PCI: Work around poweroff & suspend-to-RAM issue on Macbook Pro 11

[Greg Kroah-Hartman]

4.12-stable review patch.  If anyone has any objections, please let me know.

--

From: Bjorn Helgaas 

commit 13cfc732160f7bc7e596128ce34cda361c556966 upstream.

Neither soft poweroff (transition to ACPI power state S5) nor
suspend-to-RAM (transition to state S3) works on the Macbook Pro 11,4 and
11,5.

The problem is related to the [mem 0x7fa0-0x7fbf] space.  When we
use that space, e.g., by assigning it to the 00:1c.0 Root Port, the ACPI
Power Management 1 Control Register (PM1_CNT) at [io 0x1804] doesn't work
anymore.

Linux does a soft poweroff (transition to S5) by writing to PM1_CNT.  The
theory about why this doesn't work is:

  - The write to PM1_CNT causes an SMI
  - The BIOS SMI handler depends on something in
[mem 0x7fa0-0x7fbf]
  - When Linux assigns [mem 0x7fa0-0x7fbf] to the 00:1c.0 Port, it
covers up whatever the SMI handler uses, so the SMI handler no longer
works correctly

Reserve the [mem 0x7fa0-0x7fbf] space so we don't assign it to
anything.

This is voodoo programming, since we don't know what the real conflict is,
but we've failed to find the root cause.

Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=103211
Tested-by: the...@gmail.com
Signed-off-by: Bjorn Helgaas 
Cc: Rafael J. Wysocki 
Cc: Lukas Wunner 
Cc: Chen Yu 
Signed-off-by: Greg Kroah-Hartman 

---
 arch/x86/pci/fixup.c |   32 
 1 file changed, 32 insertions(+)

--- a/arch/x86/pci/fixup.c
+++ b/arch/x86/pci/fixup.c
@@ -571,3 +571,35 @@ DECLARE_PCI_FIXUP_EARLY(PCI_VENDOR_ID_IN
 DECLARE_PCI_FIXUP_EARLY(PCI_VENDOR_ID_INTEL, 0x6f60, pci_invalid_bar);
 DECLARE_PCI_FIXUP_EARLY(PCI_VENDOR_ID_INTEL, 0x6fa0, pci_invalid_bar);
 DECLARE_PCI_FIXUP_EARLY(PCI_VENDOR_ID_INTEL, 0x6fc0, pci_invalid_bar);
+
+/*
+ * Apple MacBook Pro: Avoid [mem 0x7fa0-0x7fbf]
+ *
+ * Using the [mem 0x7fa0-0x7fbf] region, e.g., by assigning it to
+ * the 00:1c.0 Root Port, causes a conflict with [io 0x1804], which is used
+ * for soft poweroff and suspend-to-RAM.
+ *
+ * As far as we know, this is related to the address space, not to the Root
+ * Port itself.  Attaching the quirk to the Root Port is a convenience, but
+ * it could probably also be a standalone DMI quirk.
+ *
+ * https://bugzilla.kernel.org/show_bug.cgi?id=103211
+ */
+static void quirk_apple_mbp_poweroff(struct pci_dev *pdev)
+{
+   struct device *dev = >dev;
+   struct resource *res;
+
+   if ((!dmi_match(DMI_PRODUCT_NAME, "MacBookPro11,4") &&
+!dmi_match(DMI_PRODUCT_NAME, "MacBookPro11,5")) ||
+   pdev->bus->number != 0 || pdev->devfn != PCI_DEVFN(0x1c, 0))
+   return;
+
+   res = request_mem_region(0x7fa0, 0x20,
+"MacBook Pro poweroff workaround");
+   if (res)
+   dev_info(dev, "claimed %s %pR\n", res->name, res);
+   else
+   dev_info(dev, "can't work around MacBook Pro poweroff issue\n");
+}
+DECLARE_PCI_FIXUP_HEADER(PCI_VENDOR_ID_INTEL, 0x8c10, 
quirk_apple_mbp_poweroff);

[PATCH 4.12 060/196] PCI: Work around poweroff & suspend-to-RAM issue on Macbook Pro 11

[Greg Kroah-Hartman]

4.12-stable review patch.  If anyone has any objections, please let me know.

--

From: Bjorn Helgaas 

commit 13cfc732160f7bc7e596128ce34cda361c556966 upstream.

Neither soft poweroff (transition to ACPI power state S5) nor
suspend-to-RAM (transition to state S3) works on the Macbook Pro 11,4 and
11,5.

The problem is related to the [mem 0x7fa0-0x7fbf] space.  When we
use that space, e.g., by assigning it to the 00:1c.0 Root Port, the ACPI
Power Management 1 Control Register (PM1_CNT) at [io 0x1804] doesn't work
anymore.

Linux does a soft poweroff (transition to S5) by writing to PM1_CNT.  The
theory about why this doesn't work is:

  - The write to PM1_CNT causes an SMI
  - The BIOS SMI handler depends on something in
[mem 0x7fa0-0x7fbf]
  - When Linux assigns [mem 0x7fa0-0x7fbf] to the 00:1c.0 Port, it
covers up whatever the SMI handler uses, so the SMI handler no longer
works correctly

Reserve the [mem 0x7fa0-0x7fbf] space so we don't assign it to
anything.

This is voodoo programming, since we don't know what the real conflict is,
but we've failed to find the root cause.

Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=103211
Tested-by: the...@gmail.com
Signed-off-by: Bjorn Helgaas 
Cc: Rafael J. Wysocki 
Cc: Lukas Wunner 
Cc: Chen Yu 
Signed-off-by: Greg Kroah-Hartman 

---
 arch/x86/pci/fixup.c |   32 
 1 file changed, 32 insertions(+)

--- a/arch/x86/pci/fixup.c
+++ b/arch/x86/pci/fixup.c
@@ -571,3 +571,35 @@ DECLARE_PCI_FIXUP_EARLY(PCI_VENDOR_ID_IN
 DECLARE_PCI_FIXUP_EARLY(PCI_VENDOR_ID_INTEL, 0x6f60, pci_invalid_bar);
 DECLARE_PCI_FIXUP_EARLY(PCI_VENDOR_ID_INTEL, 0x6fa0, pci_invalid_bar);
 DECLARE_PCI_FIXUP_EARLY(PCI_VENDOR_ID_INTEL, 0x6fc0, pci_invalid_bar);
+
+/*
+ * Apple MacBook Pro: Avoid [mem 0x7fa0-0x7fbf]
+ *
+ * Using the [mem 0x7fa0-0x7fbf] region, e.g., by assigning it to
+ * the 00:1c.0 Root Port, causes a conflict with [io 0x1804], which is used
+ * for soft poweroff and suspend-to-RAM.
+ *
+ * As far as we know, this is related to the address space, not to the Root
+ * Port itself.  Attaching the quirk to the Root Port is a convenience, but
+ * it could probably also be a standalone DMI quirk.
+ *
+ * https://bugzilla.kernel.org/show_bug.cgi?id=103211
+ */
+static void quirk_apple_mbp_poweroff(struct pci_dev *pdev)
+{
+   struct device *dev = >dev;
+   struct resource *res;
+
+   if ((!dmi_match(DMI_PRODUCT_NAME, "MacBookPro11,4") &&
+!dmi_match(DMI_PRODUCT_NAME, "MacBookPro11,5")) ||
+   pdev->bus->number != 0 || pdev->devfn != PCI_DEVFN(0x1c, 0))
+   return;
+
+   res = request_mem_region(0x7fa0, 0x20,
+"MacBook Pro poweroff workaround");
+   if (res)
+   dev_info(dev, "claimed %s %pR\n", res->name, res);
+   else
+   dev_info(dev, "can't work around MacBook Pro poweroff issue\n");
+}
+DECLARE_PCI_FIXUP_HEADER(PCI_VENDOR_ID_INTEL, 0x8c10, 
quirk_apple_mbp_poweroff);