On Wed, 2 Dec 2020 19:34:50 +0800
Chiqijun <chiqijun@xxxxxxxxxx> wrote:
When multiple VFs do FLR at the same time, the firmware is
processed serially, resulting in some VF FLRs being delayed more
than 100ms, when the virtual machine restarts and the device
driver is loaded, the firmware is doing the corresponding VF
FLR, causing the driver to fail to load.
To solve this problem, add host and firmware status synchronization
during FLR.
Signed-off-by: Chiqijun <chiqijun@xxxxxxxxxx>
---
v2:
- Update comments
- Use the HINIC_VF_FLR_CAP_BIT_SHIFT and HINIC_VF_FLR_PROC_BIT_SHIFT
macro instead of the magic number
---
drivers/pci/quirks.c | 75 ++++++++++++++++++++++++++++++++++++++++++++
1 file changed, 75 insertions(+)
diff --git a/drivers/pci/quirks.c b/drivers/pci/quirks.c
index f70692ac79c5..c9ad55709d03 100644
--- a/drivers/pci/quirks.c
+++ b/drivers/pci/quirks.c
@@ -3912,6 +3912,79 @@ static int delay_250ms_after_flr(struct pci_dev *dev, int probe)
return 0;
}
+#define PCI_DEVICE_ID_HINIC_VF 0x375E
+#define HINIC_VF_FLR_TYPE 0x1000
+#define HINIC_VF_FLR_CAP_BIT_SHIFT 6
+#define HINIC_VF_OP 0xE80
+#define HINIC_VF_FLR_PROC_BIT_SHIFT 10
+#define HINIC_OPERATION_TIMEOUT 15000
+
+/* Device-specific reset method for Huawei Intelligent NIC virtual functions */
+static int reset_hinic_vf_dev(struct pci_dev *pdev, int probe)
+{
+ unsigned long timeout;
+ void __iomem *bar;
+ u16 old_command;
+ u32 val;
+
+ if (probe)
+ return 0;
+
+ bar = pci_iomap(pdev, 0, 0);
+ if (!bar)
+ return -ENOTTY;
+
+ pci_read_config_word(pdev, PCI_COMMAND, &old_command);
+
+ /*
+ * FLR cap bit bit30, FLR processing bit: bit18, to avoid big-endian
+ * conversion the big-endian bit6, bit10 is directly operated here.
+ *
+ * Get and check firmware capabilities.
+ */
+ val = readl(bar + HINIC_VF_FLR_TYPE);
+ if (!(val & (1UL << HINIC_VF_FLR_CAP_BIT_SHIFT))) {
+ pci_iounmap(pdev, bar);
+ return -ENOTTY;
+ }
+
+ /*
+ * Set the processing bit for the start of FLR, which will be cleared
+ * by the firmware after FLR is completed.
+ */
+ val = readl(bar + HINIC_VF_OP);
+ val = val | (1UL << HINIC_VF_FLR_PROC_BIT_SHIFT);
+ writel(val, bar + HINIC_VF_OP);
+
+ /* Perform the actual device function reset */
+ pcie_flr(pdev);
+
+ pci_write_config_word(pdev, PCI_COMMAND,
+ old_command | PCI_COMMAND_MEMORY);
This quirk is specific to VFs, so isn't this command register handling
here, further above, and restore below unnecessary? VF memory space is
controlled by the VF MSE bit on the PF. On the VF this bit is
hardwired to zero. If that weren't the case then both the above
read/write to the BAR without testing the memory bit, and the below
polling of the BAR to watch for the reset to complete would be pretty
troubling. Thanks,
Alex
+
+ /* Waiting for device reset complete */
+ timeout = jiffies + msecs_to_jiffies(HINIC_OPERATION_TIMEOUT);
+ do {
+ val = readl(bar + HINIC_VF_OP);
+ if (!(val & (1UL << HINIC_VF_FLR_PROC_BIT_SHIFT)))
+ goto reset_complete;
+ msleep(20);
+ } while (time_before(jiffies, timeout));
+
+ val = readl(bar + HINIC_VF_OP);
+ if (!(val & (1UL << HINIC_VF_FLR_PROC_BIT_SHIFT)))
+ goto reset_complete;
+
+ pci_warn(pdev, "Reset dev timeout, flr ack reg: %x\n",
+ be32_to_cpu(val));
+
+reset_complete:
+ pci_write_config_word(pdev, PCI_COMMAND, old_command);
+ pci_iounmap(pdev, bar);
+
+ return 0;
+}
+
static const struct pci_dev_reset_methods pci_dev_reset_methods[] = {
{ PCI_VENDOR_ID_INTEL, PCI_DEVICE_ID_INTEL_82599_SFP_VF,
reset_intel_82599_sfp_virtfn },
@@ -3923,6 +3996,8 @@ static const struct pci_dev_reset_methods pci_dev_reset_methods[] = {
{ PCI_VENDOR_ID_INTEL, 0x0953, delay_250ms_after_flr },
{ PCI_VENDOR_ID_CHELSIO, PCI_ANY_ID,
reset_chelsio_generic_dev },
+ { PCI_VENDOR_ID_HUAWEI, PCI_DEVICE_ID_HINIC_VF,
+ reset_hinic_vf_dev },
{ 0 }
};
.