On 10/28/19 4:27 PM, Bjorn Helgaas wrote:
On Thu, Oct 03, 2019 at 04:39:03PM -0700, sathyanarayanan.kuppuswamy@xxxxxxxxxxxxxxx wrote:
From: Kuppuswamy Sathyanarayanan <sathyanarayanan.kuppuswamy@xxxxxxxxxxxxxxx>
As per PCI firmware specification r3.2 Downstream Port Containment
Related Enhancements ECN,
Specific reference, please, e.g., the section/table/figure of the PCI
Firmware Spec being modified by the ECN.
Ok. I will include it.
OS is responsible for clearing the AER
registers in EDR mode. So clear AER registers in dpc_process_error()
function.
Signed-off-by: Kuppuswamy Sathyanarayanan <sathyanarayanan.kuppuswamy@xxxxxxxxxxxxxxx>
Acked-by: Keith Busch <keith.busch@xxxxxxxxx>
---
drivers/pci/pcie/dpc.c | 4 ++++
1 file changed, 4 insertions(+)
diff --git a/drivers/pci/pcie/dpc.c b/drivers/pci/pcie/dpc.c
index fafc55c00fe0..de2d892bc7c4 100644
--- a/drivers/pci/pcie/dpc.c
+++ b/drivers/pci/pcie/dpc.c
@@ -275,6 +275,10 @@ static void dpc_process_error(struct dpc_dev *dpc)
pci_aer_clear_fatal_status(pdev);
}
+ /* In EDR mode, OS is responsible for clearing AER registers */
+ if (dpc->firmware_dpc)
I guess "EDR mode" is effectively the same as "firmware-first mode"?
Yes, EDR mode is an upgrade to FF mode in which firmware allows
OS to share some of it job by sending ACPI notification. If you don't get
ACPI notification, EDR mode is effectively same as FF mode.
At least, the only place we set "firmware_dpc = 1" is:
+ if (pcie_aer_get_firmware_first(pdev))
+ dpc->firmware_dpc = 1;
If they're the same, why do we need two different names for it?
For better readability and performance, I tried to cache the value of
pcie_aer_get_firmware_first() result in DPC driver.
+ pci_cleanup_aer_error_status_regs(pdev);
+
/*
* Irrespective of whether the DPC event is triggered by
* ERR_FATAL or ERR_NONFATAL, since the link is already down,
--
2.21.0
--
Sathyanarayanan Kuppuswamy
Linux kernel developer
