kernel/linux.git/drivers/net/ethernet/intel/ice, branch linux-4.20.y

ice: Do not enable NAPI on q_vectors that have no rings

2019-02-12T19:02:26+00:00

[ Upstream commit eec903769b4ea476591ffff73bb7359f14f38c51 ] If ice driver has q_vectors w/ active NAPI that has no rings, then this will result in a divide by zero error. To correct it I am updating the driver code so that we only support NAPI on q_vectors that have 1 or more rings allocated to them. See commit 13a8cd191a2b ("i40e: Do not enable NAPI on q_vectors that have no rings") for detail. Signed-off-by: Young Xiao Acked-by: Anirudh Venkataramanan Tested-by: Andrew Bowers Signed-off-by: Jeff Kirsher Signed-off-by: Sasha Levin

ice: Change req_speeds to be u16

2018-11-07T17:37:28+00:00

Since the req_speeds field in struct ice_link_status is a u8, req_speeds & ICE_AQ_LINK_SPEED_40GB always returns 0. This was caught by a coverity scan. Fix this by changing req_speeds to be u16. Reported-by: Bruce Allan Signed-off-by: Chinh T Cao Signed-off-by: Anirudh Venkataramanan Tested-by: Andrew Bowers Signed-off-by: Jeff Kirsher

ice: Fix the bytecount sent to netdev_tx_sent_queue

2018-11-06T20:46:47+00:00

Currently if the driver does a TSO offload the bytecount sent to netdev_tx_sent_queue will be incorrect. This is because in ice_tso we overwrite the initial value that we set in ice_tx_map. This creates a mismatch between the Tx and Tx clean flow. In the Tx clean flow we calculate the bytecount (called total_bytes) as we clean the descriptors so the value used in the Tx clean path is correct. Fix this by using += in ice_tso instead of =. This fixes the mismatch in bytecount mentioned above. Signed-off-by: Brett Creeley Signed-off-by: Anirudh Venkataramanan Tested-by: Andrew Bowers Signed-off-by: Jeff Kirsher

ice: Fix tx_timeout in PF driver

2018-11-06T20:46:47+00:00

Prior to this commit the driver was running into tx_timeouts when a queue was stressed enough. This was happening because the HW tail and SW tail (NTU) were incorrectly out of sync. Consequently this was causing the HW head to collide with the HW tail, which to the hardware means that all descriptors posted for Tx have been processed. Due to the Tx logic used in the driver SW tail and HW tail are allowed to be out of sync. This is done as an optimization because it allows the driver to write HW tail as infrequently as possible, while still updating the SW tail index to keep track. However, there are situations where this results in the tail never getting updated, resulting in Tx timeouts. Tx HW tail write condition: if (netif_xmit_stopped(txring_txq(tx_ring) || !skb->xmit_more) writel(sw_tail, tx_ring->tail); An issue was found in the Tx logic that was causing the afore mentioned condition for updating HW tail to never happen, causing tx_timeouts. In ice_xmit_frame_ring we calculate how many descriptors we need for the Tx transaction based on the skb the kernel hands us. This is then passed into ice_maybe_stop_tx along with some extra padding to determine if we have enough descriptors available for this transaction. If we don't then we return -EBUSY to the stack, otherwise we move on and eventually prepare the Tx descriptors accordingly in ice_tx_map and set next_to_watch. In ice_tx_map we make another call to ice_maybe_stop_tx with a value of MAX_SKB_FRAGS + 4. The key here is that this value is possibly less than the value we sent in the first call to ice_maybe_stop_tx in ice_xmit_frame_ring. Now, if the number of unused descriptors is between MAX_SKB_FRAGS + 4 and the value used in the first call to ice_maybe_stop_tx in ice_xmit_frame_ring then we do not update the HW tail because of the "Tx HW tail write condition" above. This is because in ice_maybe_stop_tx we return success from ice_maybe_stop_tx instead of calling __ice_maybe_stop_tx and subsequently calling netif_stop_subqueue, which sets the __QUEUE_STATE_DEV_XOFF bit. This bit is then checked in the "Tx HW tail write condition" by calling netif_xmit_stopped and subsequently updating HW tail if the afore mentioned bit is set. In ice_clean_tx_irq, if next_to_watch is not NULL, we end up cleaning the descriptors that HW sets the DD bit on and we have the budget. The HW head will eventually run into the HW tail in response to the description in the paragraph above. The next time through ice_xmit_frame_ring we make the initial call to ice_maybe_stop_tx with another skb from the stack. This time we do not have enough descriptors available and we return NETDEV_TX_BUSY to the stack and end up setting next_to_watch to NULL. This is where we are stuck. In ice_clean_tx_irq we never clean anything because next_to_watch is always NULL and in ice_xmit_frame_ring we never update HW tail because we already return NETDEV_TX_BUSY to the stack and eventually we hit a tx_timeout. This issue was fixed by making sure that the second call to ice_maybe_stop_tx in ice_tx_map is passed a value that is >= the value that was used on the initial call to ice_maybe_stop_tx in ice_xmit_frame_ring. This was done by adding the following defines to make the logic more clear and to reduce the chance of mucking this up again: ICE_CACHE_LINE_BYTES 64 ICE_DESCS_PER_CACHE_LINE (ICE_CACHE_LINE_BYTES / \ sizeof(struct ice_tx_desc)) ICE_DESCS_FOR_CTX_DESC 1 ICE_DESCS_FOR_SKB_DATA_PTR 1 The ICE_CACHE_LINE_BYTES being 64 is an assumption being made so we don't have to figure this out on every pass through the Tx path. Instead I added a sanity check in ice_probe to verify cache line size and print a message if it's not 64 Bytes. This will make it easier to file issues if they are seen when the cache line size is not 64 Bytes when reading from the GLPCI_CNF2 register. Signed-off-by: Brett Creeley Signed-off-by: Anirudh Venkataramanan Tested-by: Andrew Bowers Signed-off-by: Jeff Kirsher

ice: Fix napi delete calls for remove

2018-11-06T20:46:47+00:00

In the remove path, the vsi->netdev is being set to NULL before the call to free vectors. This is causing the netif_napi_del call to never be made. Add a call to ice_napi_del to the same location as the calls to unregister_netdev and just prior to them. This will use the reverse flow as the register and netif_napi_add calls. Signed-off-by: Dave Ertman Signed-off-by: Anirudh Venkataramanan Tested-by: Andrew Bowers Signed-off-by: Jeff Kirsher

ice: Fix typo in error message

2018-11-06T20:46:47+00:00

Print should say "Enabling" instead of "Enaabling" Signed-off-by: Akeem G Abodunrin Signed-off-by: Anirudh Venkataramanan Tested-by: Andrew Bowers Signed-off-by: Jeff Kirsher

ice: Fix flags for port VLAN

2018-11-06T20:46:47+00:00

According to the spec, whenever insert PVID field is set, the VLAN driver insertion mode should be set to 01b which isn't done currently. Fix it. Signed-off-by: Md Fahad Iqbal Polash Signed-off-by: Anirudh Venkataramanan Tested-by: Andrew Bowers Signed-off-by: Jeff Kirsher

ice: Remove duplicate addition of VLANs in replay path

2018-11-06T20:46:47+00:00

ice_restore_vlan and active_vlans were originally put in place to reprogram VLAN filters in the replay path. This is now done as part of the much broader VSI rebuild/replay framework. So remove both ice_restore_vlan and active_vlans Signed-off-by: Anirudh Venkataramanan Tested-by: Andrew Bowers Signed-off-by: Jeff Kirsher

ice: Free VSI contexts during for unload

2018-11-06T20:46:47+00:00

In the unload path, all VSIs are freed. Also free the related VSI contexts to prevent memory leaks. Signed-off-by: Victor Raj Signed-off-by: Anirudh Venkataramanan Tested-by: Andrew Bowers Signed-off-by: Jeff Kirsher

ice: Fix dead device link issue with flow control

2018-11-06T20:46:47+00:00

Setting Rx or Tx pause parameter currently results in link loss on the interface, requiring the platform/host to be cold power cycled. Fix it. Signed-off-by: Akeem G Abodunrin Signed-off-by: Anirudh Venkataramanan Tested-by: Andrew Bowers Signed-off-by: Jeff Kirsher