在 2024/12/4 22:03, Leon Romanovsky 写道:
On Tue, Nov 26, 2024 at 02:59:07PM +0800, Boshi Yu wrote:
Currently, the erdma driver supports both the iWARP and RoCEv2 protocols.
The erdma driver reads the ERDMA_REGS_DEV_PROTO_REG register to identify
the protocol used by the erdma device. Since each protocol requires
different ib_device_ops, we introduce the erdma_device_ops_iwarp and
erdma_device_ops_rocev2 for iWARP and RoCEv2 protocols, respectively.
Signed-off-by: Boshi Yu <boshiyu@xxxxxxxxxxxxxxxxx>
Reviewed-by: Cheng Xu <chengyou@xxxxxxxxxxxxxxxxx>
---
drivers/infiniband/hw/erdma/Kconfig | 2 +-
drivers/infiniband/hw/erdma/erdma.h | 3 +-
drivers/infiniband/hw/erdma/erdma_hw.h | 7 ++++
drivers/infiniband/hw/erdma/erdma_main.c | 47 ++++++++++++++++++-----
drivers/infiniband/hw/erdma/erdma_verbs.c | 16 +++++++-
drivers/infiniband/hw/erdma/erdma_verbs.h | 12 ++++++
6 files changed, 75 insertions(+), 12 deletions(-)
<...>
+++ b/drivers/infiniband/hw/erdma/erdma_main.c
@@ -172,6 +172,12 @@ static int erdma_device_init(struct erdma_dev *dev, struct pci_dev *pdev)
{
int ret;
+ dev->proto = erdma_reg_read32(dev, ERDMA_REGS_DEV_PROTO_REG);
+ if (!erdma_device_iwarp(dev) && !erdma_device_rocev2(dev)) {
Why do you need this check? Your old driver which supports only iwarp
doesn't have this check, so why did you suddenly need it for roce?
Hi, Leon,
We initially try to avoid reading an invalid value from the hardware.
Thank you for your question; I found that the check is unnecessary since
the value of ERDMA_REGS_DEV_PROTO_REG can only be 0 or 1. We will remove
this check in patch v2.
Thanks,
Boshi Yu
+ dev_err(&pdev->dev, "Unsupported protocol: %d\n", dev->proto);
+ return -ENODEV;
+ }
+
Thanks
