When the power rail gets cut off, the hardware can create some electric noise on the link that triggers AER. If IRQ is shared between AER with PME, such AER noise will cause a spurious wakeup on system suspend. When the power rail gets back, the firmware of the device resets itself and can create unexpected behavior like sending PTM messages. If DPC is enabled, the DPC reset happens before driver's resume routine. The DPC reset usually fails because the device is not in the right state, and the resume also fails because the device is being reset by hardware. If the scenario happens to device like NVMe, it means the whole system resume fails. As Per PCIe Base Spec 5.0, section 5.2, titled "Link State Power Management", TLP and DLLP transmission are disabled for a Link in L2/L3 Ready (D3hot), L2 (D3cold with aux power) and L3 (D3cold) states. So if the power will be turned off during suspend process, disable DPC service and re-enable it during the resume process. This should not affect the basic functionality. Furthermore, since DPC depends on AER to function, and AER is disabled in previous patch, also disable DPC here. Link: https://bugzilla.kernel.org/show_bug.cgi?id=209149 Link: https://bugzilla.kernel.org/show_bug.cgi?id=216295 Link: https://bugzilla.kernel.org/show_bug.cgi?id=218090 Signed-off-by: Kai-Heng Feng <kai.heng.feng@xxxxxxxxxxxxx> --- v8: - Wording. - Add more bug reports. v7: - Wording. - Disable DPC completely (again) if power will be turned off v6: v5: - Wording. v4: v3: - No change. v2: - Only disable DPC IRQ. - No more check on PME IRQ#. drivers/pci/pcie/dpc.c | 57 ++++++++++++++++++++++++++++++++++-------- 1 file changed, 46 insertions(+), 11 deletions(-) diff --git a/drivers/pci/pcie/dpc.c b/drivers/pci/pcie/dpc.c index a668820696dc..7682ac4d6a89 100644 --- a/drivers/pci/pcie/dpc.c +++ b/drivers/pci/pcie/dpc.c @@ -14,6 +14,7 @@ #include <linux/interrupt.h> #include <linux/init.h> #include <linux/pci.h> +#include <linux/suspend.h> #include "portdrv.h" #include "../pci.h" @@ -412,13 +413,34 @@ void pci_dpc_init(struct pci_dev *pdev) } } +static void dpc_enable(struct pcie_device *dev) +{ + struct pci_dev *pdev = dev->port; + u16 ctl; + + pci_read_config_word(pdev, pdev->dpc_cap + PCI_EXP_DPC_CTL, &ctl); + ctl &= ~PCI_EXP_DPC_CTL_EN_MASK; + ctl |= PCI_EXP_DPC_CTL_EN_FATAL | PCI_EXP_DPC_CTL_INT_EN; + pci_write_config_word(pdev, pdev->dpc_cap + PCI_EXP_DPC_CTL, ctl); +} + +static void dpc_disable(struct pcie_device *dev) +{ + struct pci_dev *pdev = dev->port; + u16 ctl; + + pci_read_config_word(pdev, pdev->dpc_cap + PCI_EXP_DPC_CTL, &ctl); + ctl &= ~(PCI_EXP_DPC_CTL_EN_FATAL | PCI_EXP_DPC_CTL_INT_EN); + pci_write_config_word(pdev, pdev->dpc_cap + PCI_EXP_DPC_CTL, ctl); +} + #define FLAG(x, y) (((x) & (y)) ? '+' : '-') static int dpc_probe(struct pcie_device *dev) { struct pci_dev *pdev = dev->port; struct device *device = &dev->device; int status; - u16 ctl, cap; + u16 cap; if (!pcie_aer_is_native(pdev) && !pcie_ports_dpc_native) return -ENOTSUPP; @@ -433,11 +455,7 @@ static int dpc_probe(struct pcie_device *dev) } pci_read_config_word(pdev, pdev->dpc_cap + PCI_EXP_DPC_CAP, &cap); - - pci_read_config_word(pdev, pdev->dpc_cap + PCI_EXP_DPC_CTL, &ctl); - ctl &= ~PCI_EXP_DPC_CTL_EN_MASK; - ctl |= PCI_EXP_DPC_CTL_EN_FATAL | PCI_EXP_DPC_CTL_INT_EN; - pci_write_config_word(pdev, pdev->dpc_cap + PCI_EXP_DPC_CTL, ctl); + dpc_enable(dev); pci_info(pdev, "enabled with IRQ %d\n", dev->irq); pci_info(pdev, "error containment capabilities: Int Msg #%d, RPExt%c PoisonedTLP%c SwTrigger%c RP PIO Log %d, DL_ActiveErr%c\n", @@ -450,14 +468,29 @@ static int dpc_probe(struct pcie_device *dev) return status; } -static void dpc_remove(struct pcie_device *dev) +static int dpc_suspend(struct pcie_device *dev) { struct pci_dev *pdev = dev->port; - u16 ctl; - pci_read_config_word(pdev, pdev->dpc_cap + PCI_EXP_DPC_CTL, &ctl); - ctl &= ~(PCI_EXP_DPC_CTL_EN_FATAL | PCI_EXP_DPC_CTL_INT_EN); - pci_write_config_word(pdev, pdev->dpc_cap + PCI_EXP_DPC_CTL, ctl); + if (pci_ancestor_pr3_present(pdev) || pm_suspend_via_firmware()) + dpc_disable(dev); + + return 0; +} + +static int dpc_resume(struct pcie_device *dev) +{ + struct pci_dev *pdev = dev->port; + + if (pci_ancestor_pr3_present(pdev) || pm_resume_via_firmware()) + dpc_enable(dev); + + return 0; +} + +static void dpc_remove(struct pcie_device *dev) +{ + dpc_disable(dev); } static struct pcie_port_service_driver dpcdriver = { @@ -465,6 +498,8 @@ static struct pcie_port_service_driver dpcdriver = { .port_type = PCIE_ANY_PORT, .service = PCIE_PORT_SERVICE_DPC, .probe = dpc_probe, + .suspend = dpc_suspend, + .resume = dpc_resume, .remove = dpc_remove, }; -- 2.34.1
