Commit 9b55bb03 authored by Yi Zou's avatar Yi Zou Committed by Jeff Kirsher

ixgbe: make sure FCoE DDP user buffers are really released by the HW

When invalidating the DDP context is invalidated, the HW may not be done
with the user buffer right away. In which case, we poll the FCBUFF register
to check if the buffer valid bit is cleared or not, if not, we wait for max
100us that is guaranteed by the HW.
Signed-off-by: default avatarYi Zou <yi.zou@intel.com>
Tested-by: default avatarRoss Brattain <ross.b.brattain@intel.com>
Signed-off-by: default avatarJeff Kirsher <jeffrey.t.kirsher@intel.com>
parent 8ca371e4
......@@ -92,6 +92,7 @@ int ixgbe_fcoe_ddp_put(struct net_device *netdev, u16 xid)
struct ixgbe_fcoe *fcoe;
struct ixgbe_adapter *adapter;
struct ixgbe_fcoe_ddp *ddp;
u32 fcbuff;
if (!netdev)
goto out_ddp_put;
......@@ -115,7 +116,14 @@ int ixgbe_fcoe_ddp_put(struct net_device *netdev, u16 xid)
IXGBE_WRITE_REG(&adapter->hw, IXGBE_FCBUFF, 0);
IXGBE_WRITE_REG(&adapter->hw, IXGBE_FCDMARW,
(xid | IXGBE_FCDMARW_WE));
/* guaranteed to be invalidated after 100us */
IXGBE_WRITE_REG(&adapter->hw, IXGBE_FCDMARW,
(xid | IXGBE_FCDMARW_RE));
fcbuff = IXGBE_READ_REG(&adapter->hw, IXGBE_FCBUFF);
spin_unlock_bh(&fcoe->lock);
if (fcbuff & IXGBE_FCBUFF_VALID)
udelay(100);
}
if (ddp->sgl)
pci_unmap_sg(adapter->pdev, ddp->sgl, ddp->sgc,
......
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment