cnic: Zero out status block and Event Queue indices.
To prevent stale indices from causing spurious events when restarting the bnx2x devices. Signed-off-by: Michael Chan <mchan@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
This commit is contained in:
parent
1bcdc32cf4
commit
4e9c4fd3e7
|
@ -1104,6 +1104,8 @@ static int cnic_alloc_bnx2x_resc(struct cnic_dev *dev)
|
|||
cp->bnx2x_status_blk = cp->status_blk;
|
||||
cp->bnx2x_def_status_blk = cp->ethdev->irq_arr[1].status_blk;
|
||||
|
||||
memset(cp->bnx2x_status_blk, 0, sizeof(struct host_status_block));
|
||||
|
||||
cp->l2_rx_ring_size = 15;
|
||||
|
||||
ret = cnic_alloc_l2_rings(dev, 4);
|
||||
|
@ -4295,6 +4297,9 @@ static void cnic_stop_bnx2x_hw(struct cnic_dev *dev)
|
|||
offsetof(struct cstorm_status_block_c,
|
||||
index_values[HC_INDEX_C_ISCSI_EQ_CONS]),
|
||||
0);
|
||||
CNIC_WR(dev, BAR_CSTRORM_INTMEM +
|
||||
CSTORM_ISCSI_EQ_CONS_OFFSET(cp->func, 0), 0);
|
||||
CNIC_WR16(dev, cp->kcq_io_addr, 0);
|
||||
cnic_free_resc(dev);
|
||||
}
|
||||
|
||||
|
|
Loading…
Reference in a new issue