- 12 Mar, 2018 31 commits
-
-
Gal Pressman authored
Same as LRO, hardware GRO cannot be enabled with RX-FCS. When both are requested, hardware GRO will be dropped. Suggested-by:
David Miller <davem@davemloft.net> Signed-off-by:
Gal Pressman <galp@mellanox.com> Signed-off-by:
-
Salvatore Mesoraca authored
Avoid a VLA[1] by using a real constant expression instead of a variable. The compiler should be able to optimize the original code and avoid using an actual VLA. Anyway this change is useful because it will avoid a false positive with -Wvla, it might also help the compiler generating better code. [1] https://lkml.org/lkml/2018/3/7/621Signed-off-by:
Salvatore Mesoraca <s.mesoraca16@gmail.com> Signed-off-by:
-
Colin Ian King authored
Functions lan743x_csr_read and lan743x_csr_read are local to the source and do not need to be in global scope, so make them static. Cleans up sparse warning: drivers/net/ethernet/microchip/lan743x_main.c:56:5: warning: symbol lan743x_csr_read' was not declared. Should it be static? drivers/net/ethernet/microchip/lan743x_main.c:61:6: warning: symbol 'lan743x_csr_write' was not declared. Should it be static? Signed-off-by:
Colin Ian King <colin.king@canonical.com> Signed-off-by:
-
Colin Ian King authored
Function lan743x_phy_init assigns pointer 'netdev' but this is never read and hence it can be removed. The return error code handling can also be cleaned up to remove the variable 'ret'. Function lan743x_phy_link_status_change assigns pointer 'phy' twice and this is never read, so it also can be removed. Finally, function lan743x_tx_napi_poll initializes pointer 'adapter' and then re-assigns the same value into this pointer a little later on so this second assignment is redundant and can be also removed. Cleans up clang warnings: drivers/net/ethernet/microchip/lan743x_main.c:951:2: warning: Value stored to 'netdev' is never read drivers/net/ethernet/microchip/lan743x_main.c:971:3: warning: Value stored to 'phy' is never read drivers/net/ethernet/microchip/lan743x_main.c:1583:26: warning: Value stored to 'adapter' during its initialization is never read Signed-off-by:
-
Colin Ian King authored
Variable sg_off is assigned a value but it is never read, hence it is redundant and can be removed. Cleans up clang warning: net/rds/message.c:373:2: warning: Value stored to 'sg_off' is never read Signed-off-by:
Sowmini Varadhan <sowmini.varadhan@oracle.com> Acked-by:
Santosh Shilimkar <santosh.shilimkar@oracle.com> Signed-off-by:
-
David S. Miller authored
Make use of the new helper. Suggested-by:
David Ahern <dsahern@gmail.com> Signed-off-by:
-
David S. Miller authored
Ido Schimmel says: ==================== mlxsw: Removing dependency of mlxsw on GRE Petr says: mlxsw_spectrum supports offloading of a tc action mirred egress mirror to a gretap or ip6gretap netdevice, which necessitates calls to functions defined in ip_gre, ip6_gre and ip6_tunnel modules. Previously this was enabled by introducing a hard dependency of MLXSW_SPECTRUM on NET_IPGRE and IPV6_GRE. However the rest of mlxsw is careful about picking which modules are absolutely required, and therefore the better approach is to make mlxsw_spectrum tolerant of absence of one or both of the GRE flavors. One way this might be resolved is by keeping the code in mlxsw_spectrum intact, and defining defaults for functions that mlxsw_spectrum depends on. The downsides are that other modules end up littered with these do-nothing defaults; that the driver ends up carrying quite a bit of dead code; and that the driver ends up having to explicitly depend on IPV6_TUNNEL to prevent configurations where mlxsw_spectrum is compiled in and and ip6_tunnel is a module, something that it currently can treat as an implementation detail of the IPV6_GRE dependency. Alternatively, the driver should just bite the bullet and ifdef-out the code that handles configurations that are not supported. Since that's what we are doing for IPv6 dependency, let's do the same for the GRE flavors. Patch #1 introduces a wrapper function for determining the value of ipv6.sysctl.multipath_hash_policy, which defaults to 0 on non-IPv6 builds. That function is then used from spectrum_router.c, instead of the direct variable reference that was introduced there during the short window when the Spectrum driver had a hard dependency on IPv6. Patch #2 moves one function to keep together in one block all the callbacks for handling (IPv4) gretap mirroring. Patch #3 then introduces the ifdefs to hide the irrelevant code. ==================== Signed-off-by: -
Petr Machata authored
mlxsw_spectrum supports offloading of a tc action mirred egress mirror to a gretap or an ip6gretap netdevice, which necessitates calls to functions defined in ip_gre, ip6_gre and ip6_tunnel modules. Previously this was enabled by introducing a hard dependency of MLXSW_SPECTRUM on NET_IPGRE and IPV6_GRE. However the rest of mlxsw is careful about picking which modules are absolutely required, and therefore the better approach is to make mlxsw_spectrum tolerant of absence of one or both of the GRE flavors. Hence rework the NET_IPGRE and IPV6_GRE dependencies to just guard matching modularity, and hide the corresponding code in spectrum_span.c in an #if IS_ENABLED. Mark mlxsw_sp_span_entry_tunnel_parms_common as maybe unused, to muffle warnings if neither GRE flavor is selected, which seems cleaner than introducing a composite #if. Signed-off-by:
Petr Machata <petrm@mellanox.com> Signed-off-by:
Ido Schimmel <idosch@mellanox.com> Signed-off-by:
-
Petr Machata authored
Move the function next to the rest of gretap4 functions. Thus the generic functions shared between gretap4 and gretap6 are in one block at the beginning, followed by a gretap4 block, followed by a gretap6 block. Signed-off-by:
-
Petr Machata authored
In order to abstract away access to the ipv6.sysctl.multipath_hash_policy variable, which is not available on systems compiled without IPv6 support, introduce a wrapper function ip6_multipath_hash_policy() that falls back to 0 on non-IPv6 systems. Use this wrapper from mlxsw/spectrum_router instead of a direct reference. Signed-off-by:
-
Ganesh Goudar authored
check firmware capabilities before setting ethtool link mode mask, also add few missing speeds. Signed-off-by:
Ganesh Goudar <ganeshgr@chelsio.com> Signed-off-by:
-
Ganesh Goudar authored
50Gbps is a supported speed, Stop reporting it as unsupported speed. Signed-off-by:
-
Veerasenareddy Burru authored
In a scenario where the command queued to firmware get dropped or times out, MTU change from host will not propagate to firmware. So, it is required for host driver to wait for response from firmware or timeout and then return correct status to caller of ndo_change_mtu. Also moved the common code for MTU change from PF and VF driver files to common file lio_core.c Signed-off-by:
Veerasenareddy Burru <veerasenareddy.burru@cavium.com> Signed-off-by:
Felix Manlunas <felix.manlunas@cavium.com> Signed-off-by:
-
David S. Miller authored
Peng Li says: ==================== fix some bugs for HNS3 driver This patchset fixes some bugs for HNS3 driver: [Patch 1/12 - Patch 8/12] fix various bugs for PF driver. [Patch 9/12 - Patch 12/12] fix issues when change the us mac address of PF/VF device to an existent one in the mac_vlan table. ==================== Signed-off-by: -
Jian Shen authored
VF changes unicast mac address by sending mailbox msg to PF, then PF completes the mac address modification. It may fail when the target uc mac address is already in the mac_vlan table. VF should be aware of it by reading the message result. Signed-off-by:
Jian Shen <shenjian15@huawei.com> Signed-off-by:
Peng Li <lipeng321@huawei.com> Signed-off-by:
-
Jian Shen authored
It's not allowed to add two same unicast mac address entries to the mac_vlan table. When modify the uc mac address of a VF device to the same value with the PF device's, the PF device will lose its entry of the mac_vlan table. Lookup the mac address in the mac_vlan table, and add it if the entry is inexistent. Fixes: 46a3df9f ("net: hns3: Add HNS3 Acceleration Engine & Compatibility Layer Support") Signed-off-by:
-
Jian Shen authored
Error code -EIO was used to indicate mutilple errors in function hclge_get_mac_vlan_cmd_status().This patch fixes it by using error code depending on the error type. For no space error, return -ENOSPC. For entry not found, return -ENOENT. For command send fail, return -EIO. For invalid op code, return -EINVAL. Fixes: 46a3df9f ("net: hns3: Add HNS3 Acceleration Engine & Compatibility Layer Support") Signed-off-by:
-
Jian Shen authored
An enum type variable was used to store an "int" type return value. This patch fixes it. Fixes: 46a3df9f ("net: hns3: Add HNS3 Acceleration Engine & Compatibility Layer Support") Signed-off-by:
-
Yunsheng Lin authored
This patch fixes the buffer overflow warning by refactoring hclgevf_bind_ring_to_vector and hclge_get_ring_chain_from_mbx. Fixes: e2cb1dec ("net: hns3: Add HNS3 VF HCL(Hardware Compatibility Layer) Support") Fixes: dde1a86e ("net: hns3: Add mailbox support to PF driver") Signed-off-by:
Yunsheng Lin <linyunsheng@huawei.com> Signed-off-by:
-
Yunsheng Lin authored
When vlan ctag filter is enabled, the loopback selftest fails because loopback selftest does not support vlan. This patch fixes it by disabling the vlan ctag filter when runnig loopback selftest. Signed-off-by:
-
Peng Li authored
This patch add support for querying pfc puase packets statistic in hclge_ieee_getpfc, which is used to tell user how many pfc puase packets have been sent and received by this mac port. Signed-off-by:
-
Peng Li authored
Original skb->truesize reports the received packet size, not the actual buffer size NIC driver allocated(1 Page). The linux net protocol will misjudge the true size of rx queue. Signed-off-by:
-
Fuyun Liang authored
Since the firmware cmd to setup mac pause params is the same as the firmware cmd to pfc pause params, this patch unifies the pause params setup function. Signed-off-by:
Fuyun Liang <liangfuyun1@huawei.com> Signed-off-by:
-
Fuyun Liang authored
The function of dev_close and dev_open is just likes ifconfig <netif> down and ifconfig <netif> up. The ipv6 address will be lost after dev_close and dev_open are called. This patch uses hns3_nic_net_stop to replace dev_close and uses hns3_nic_net_open to replace dev_open. Signed-off-by:
-
Fuyun Liang authored
The link status update function is called by timer every second. But net_stop and net_open may be called with very short intervals. The link status update function can not detect the link state has changed. It causes the netdev not running problem. This patch fixes it by updating the link state in ae_stop function. Signed-off-by:
-
Fuyun Liang authored
When driver is in initial state, the mac_vlan table table is empty. So the delete operation for mac address must fail. Existence check is needed here. Otherwise, the error message will make user confused. Fixes: 46a3df9f ("net: hns3: Add HNS3 Acceleration Engine & Compatibility Layer Support") Signed-off-by:
-
David S. Miller authored
Ido Schimmel says: ==================== selftests: forwarding: Tweaks and a new test First patch adds a new test for VLAN-unaware bridges. Next two patches make the tests fail in case they are missing interfaces or dependencies. Last patch allows one to create the veth interfaces even without the optional configuration file. ==================== Signed-off-by: -
Ido Schimmel authored
Some users want to be able to run the tests without a configuration file which is useful when one needs to test both virtual and physical interfaces on the same machine. Move the defines that set the type of interface to create and whether to create it away from the optional configuration file to the library like the rest of the defines. Signed-off-by:
Jiri Pirko <jiri@mellanox.com> Reviewed-by:
-
Ido Schimmel authored
Returning 0 gives a false sense of success when the required modules did not even manage to be initialized and register the required net devices. Signed-off-by:
-
Ido Schimmel authored
We already return an error when some dependencies (e.g., 'jq') are missing so lets be consistent and do that for all. Signed-off-by:
-
Ido Schimmel authored
Similar to the VLAN-aware bridge test, test the VLAN-unaware bridge and make sure that ping, FDB learning and flooding work as expected. Signed-off-by:
-
- 09 Mar, 2018 9 commits
-
-
David S. Miller authored
Julian Wiedmann says: ==================== s390/qeth: updates 2018-03-09 here is the current pile of qeth patches for net-next. Just the usual small updates and clean ups. Please apply. ==================== Signed-off-by: -
Julian Wiedmann authored
Using up 8 bytes in every ipaddr object to store SETIP/DELIP flags is rather wasteful. Except for takeover eligibility, the flag values all just depend on the address type, so determine them on demand. While at it reorder the struct to fill an alignment hole. Signed-off-by:
Julian Wiedmann <jwi@linux.vnet.ibm.com> Signed-off-by:
-
Julian Wiedmann authored
Reduce code duplication. Signed-off-by:
-
Julian Wiedmann authored
On an IP event, current code tries to determine if the netdev belongs to a L3 card by walking all qeth cards in the system, and then all of their VLAN devices too. Short-cut the whole thing by identifying a L3 device through its netdev_ops. Signed-off-by:
-
Julian Wiedmann authored
Extract a helper that does the actual work & returns the right NOTIFY_* responses, and start putting the temporary ipaddr container objects on the stack rather than kmalloc'ing them. They are small, and this reduces the confusion of which objects actually get added to qeth's IP tables. Signed-off-by:
-
Julian Wiedmann authored
init_qdio_queues() resets the Input Queue's overall QDIO state, and positions the buffer cursor back to 0. So this is the obvious place to also reset the queue's NAPI context (in contrast to doing it rather randomly in the middle of the big set_online() path). No functional change. Signed-off-by:
-
Julian Wiedmann authored
Newly-allocated skbs default to PACKET_HOST, and eth_type_trans() is smart enough to determine any other packet type from the frame's destination address. So except for the IQD sniffer case, there is no need to set up skb->pkt_type manually. Signed-off-by:
-
Julian Wiedmann authored
napi_alloc_skb() doesn't need to disable IRQs during the allocation, and thus may save us a few cycles. Doing so requires a small fix-up in the HiperTransport path, which currently assumes a fixed NET_SKB_PAD headroom padding. napi_alloc_skb() adds an additional NET_IP_ALIGN padding, so use the proper helper for setting up the mac_header offset. Use this opportunity to convert the non-NAPI path to netdev_alloc_skb(), which means that skb->dev is now always set-up during allocation and doesn't need to be assigned manually. Signed-off-by:
-
Julian Wiedmann authored
We need to pass the *payload* length, not the L2 address length. For qeth (using eth_header()) this is merely a cosmetic change: the parameter only matters when building headers for ETH_P_802_2 or ETH_P_802_3, whereas our fake headers are built with ETH_P_IP / ETH_P_IPV6 / 0. Signed-off-by:
-
