- 28 May, 2020 40 commits
-
-
Huazhong Tan authored
Remove HNAE3_RESTORE_CLIENT which is not needed now. Signed-off-by:
Huazhong Tan <tanhuazhong@huawei.com> Signed-off-by:
David S. Miller <davem@davemloft.net>
-
Huazhong Tan authored
Remove some fileds which defined in struct hns3_nic_priv, but not used, and remove the related definition of struct hns3_udp_tunnel and enum hns3_udp_tnl_type. Signed-off-by:
-
Huazhong Tan authored
Modify field .gro_en in struct hclgevf_cfg_gro_status_cmd to u8 according to the UM, otherwise, it will overwrite the reserved byte which may be used for other purpose. Signed-off-by:
-
Huazhong Tan authored
Modify field .gro_en in struct hclge_cfg_gro_status_cmd to u8 according to the UM, otherwise, it will overwrite the reserved byte which may be used for other purpose. Signed-off-by:
-
Huazhong Tan authored
In order to improve code maintainability and readability, rewrite the process of BDs' initialization in hclge_query_bd_num_cmd_send(). Signed-off-by:
-
Huazhong Tan authored
Since parameters 'tso_mss_min' and 'tso_mss_max' only indicate the minimum and maximum MSS, the hnae3_set_field() calls are meaningless, remove them and change the type of these two parameters to u16. Signed-off-by:
-
Huazhong Tan authored
Add a mutex destroy call in hclge_init_ae_dev() when fails. Signed-off-by:
-
Huazhong Tan authored
Remove the redundant 'goto' and return -ENOMEM directly, when allocating memory for 'hdev' fails in hclge_init_ae_dev(). Signed-off-by:
-
David S. Miller authored
Marek Vasut says: ==================== net: ks8851: Unify KS8851 SPI and MLL drivers The KS8851SNL/SNLI and KS8851-16MLL/MLLI/MLLU are very much the same pieces of silicon, except the former has an SPI interface, while the later has a parallel bus interface. Thus far, Linux has two separate drivers for each and they are diverging considerably. This series unifies them into a single driver with small SPI and parallel bus specific parts. The approach here is to first separate out the SPI specific parts into a separate file, then add parallel bus accessors in another separate file and then finally remove the old parallel bus driver. The reason for replacing the old parallel bus driver is because the SPI bus driver is much higher quality. Note that I dropped "net: ks8851: Drop define debug and pr_fmt()" for now, will send it separatelly later. ==================== Signed-off-by: -
Marek Vasut authored
The ks8851_mll.c is replaced by ks8851_par.c, which is using common code from ks8851.c, just like ks8851_spi.c . Remove this old ad-hoc driver. Signed-off-by:
Marek Vasut <marex@denx.de> Cc: David S. Miller <davem@davemloft.net> Cc: Lukas Wunner <lukas@wunner.de> Cc: Petr Stetiar <ynezz@true.cz> Cc: YueHaibing <yuehaibing@huawei.com> Reviewed-by:
Andrew Lunn <andrew@lunn.ch> Signed-off-by:
-
Marek Vasut authored
Implement accessors for KS8851-16MLL/MLLI/MLLU parallel bus variant of the KS8851. This is based off the ks8851_mll.c , which is a driver for exactly the same hardware, however the ks8851.c code is much higher quality. Hence, this patch pulls out the relevant information from the ks8851_mll.c on how to access the bus, but uses the common ks8851.c code. To make this patch reviewable, instead of rewriting ks8851_mll.c, ks8851_mll.c is removed in a separate subsequent patch. Signed-off-by:
-
Marek Vasut authored
Pull all the SPI bus specific code into a separate file, so that it is not mixed with the common code. Rename ks8851.c to ks8851_common.c. The ks8851_common.c is linked with ks8851_spi.c now, so it can call the accessors in the ks8851_spi.c without any pointer indirection. Signed-off-by:
-
Marek Vasut authored
The register and FIFO accessors are bus specific, so is locking. Implement callbacks so that each variant of the KS8851 can implement matching accessors and locking, and use the rest of the common code. Reviewed-by:
-
Marek Vasut authored
The parallel bus variant does not need to use the TX interrupt at all as it writes the TX FIFO directly with in .ndo_start_xmit, permit the drivers to configure the interrupt enable bits. Reviewed-by:
-
Marek Vasut authored
While the SPI version of the KS8851 requires a TX worker thread to pump data via SPI, the parallel bus version can write data into the TX FIFO directly in .ndo_start_xmit, as the parallel bus access is much faster and does not sleep. Factor out this TX work flush part, so it can be overridden by the parallel bus driver. Reviewed-by:
-
Marek Vasut authored
Factor out common code into ks8851_probe_common() and ks8851_remove_common() to permit both SPI and parallel bus driver variants to use the common code path for both probing and removal. There should be no functional change. Reviewed-by:
-
Marek Vasut authored
Add a new struct ks8851_net_spi, which embeds the original struct ks8851_net and contains the entries specific only to the SPI variant of KS8851. There should be no functional change. Reviewed-by:
-
Marek Vasut authored
Factor out this netif_rx_ni(), so it could be overridden by the parallel bus variant of the KS8851 driver. Signed-off-by:
-
Marek Vasut authored
Pull out bus access locking code into separate functions, this is done in preparation for unifying the driver with the parallel bus one. The parallel bus driver does not need heavy mutex locking of the bus and works better with spinlocks, hence prepare these locking functions to be overridden then. Reviewed-by:
-
Marek Vasut authored
The RXFC register is the only one being read using 8-bit accessors. To make it easier to support the 16-bit accesses used by the parallel bus variant of KS8851, use 16-bit accessor to read RXFC register as well as neighboring RXFCTR register. Remove ks8851_rdreg8() as it is not used anywhere anymore. There should be no functional change. Signed-off-by:
-
Marek Vasut authored
On the SPI variant of KS8851, the MAC address can be programmed with either 8/16/32-bit writes. To make it easier to support the 16-bit parallel option of KS8851 too, switch both the MAC address programming and readout to 16-bit operations. Remove ks8851_wrreg8() as it is not used anywhere anymore. There should be no functional change. Reviewed-by:
-
Marek Vasut authored
The ks8851_rdreg32() is used only in one place, to read two registers using a single read. To make it easier to support 16-bit accesses via parallel bus later on, replace this single read with two 16-bit reads from each of the registers and drop the ks8851_rdreg32() altogether. If this has noticeable performance impact on the SPI variant of KS8851, then we should consider using regmap to abstract the SPI and parallel bus options and in case of SPI, permit regmap to merge register reads of neighboring registers into single, longer, read. Signed-off-by:
-
Marek Vasut authored
Replace spi_{get,set}_drvdata() with dev_{get,set}_drvdata(), which works for both SPI and platform drivers. This is done in preparation for unifying the KS8851 SPI and parallel bus drivers. There should be no functional change. Reviewed-by: -
Marek Vasut authored
Use device managed version of alloc_etherdev() to simplify the code. No functional change intended. Reviewed-by:
-
Marek Vasut authored
Since the driver probe function already has a struct device *dev pointer and can easily derive of_node pointer from it, pass the of_node pointer as a parameter to ks8851_init_mac() to avoid fishing it out from ks->spidev. This is the only reference to spidev in the function, so get rid of it. This is done in preparation for unifying the KS8851 SPI and parallel bus drivers. No functional change. Signed-off-by:
-
Marek Vasut authored
Use netdev_err() instead of dev_err() to avoid accessing the spidev->dev in the interrupt handler. This is the only place which uses the spidev in this function, so replace it with netdev_err() to get rid of it. This is done in preparation for unifying the KS8851 SPI and parallel drivers. Reviewed-by:
-
Marek Vasut authored
Rename ndev variable to netdev for the sake of consistency. No functional change. Reviewed-by:
-
Marek Vasut authored
Pull out the spi->dev into one common place in the function instead of having it repeated over and over again. This is done in preparation for unifying ks8851 and ks8851-mll drivers. No functional change. Reviewed-by:
-
David S. Miller authored
Ronak Doshi says: ==================== vmxnet3: upgrade to version 4 vmxnet3 emulation has recently added several new features which includes offload support for tunnel packets, support for new commands the driver can issue to emulation, change in descriptor fields, etc. This patch series extends the vmxnet3 driver to leverage these new features. Compatibility is maintained using existing vmxnet3 versioning mechanism as follows: - new features added to vmxnet3 emulation are associated with new vmxnet3 version viz. vmxnet3 version 4. - emulation advertises all the versions it supports to the driver. - during initialization, vmxnet3 driver picks the highest version number supported by both the emulation and the driver and configures emulation to run at that version. In particular, following changes are introduced: Patch 1: This patch introduces utility macros for vmxnet3 version 4 comparison and updates Copyright information. Patch 2: This patch implements get_rss_hash_opts and set_rss_hash_opts methods to allow querying and configuring different Rx flow hash configurations which can be used to support UDP/ESP RSS. Patch 3: This patch introduces segmentation and checksum offload support for encapsulated packets. This avoids segmenting and calculating checksum for each segment and hence gives performance boost. Patch 4: With all vmxnet3 version 4 changes incorporated in the vmxnet3 driver, with this patch, the driver can configure emulation to run at vmxnet3 version 4. Changes in v3 -> v4: - Replaced BUG_ON() with WARN_ON_ONCE() Changes in v2 -> v3: - fixed get_rss_hash_opts to return correct values for udp rss Changes in v2: - Fixed compilation issue due to missing closed brace - added fallthrough comment ==================== Signed-off-by: -
Ronak Doshi authored
With all vmxnet3 version 4 changes incorporated in the vmxnet3 driver, the driver can configure emulation to run at vmxnet3 version 4, provided the emulation advertises support for version 4. Signed-off-by:
Ronak Doshi <doshir@vmware.com> Signed-off-by:
-
Ronak Doshi authored
Vmxnet3 version 3 device supports checksum/TSO offload. Thus, vNIC to pNIC traffic can leverage hardware checksum/TSO offloads. However, vmxnet3 does not support checksum/TSO offload for Geneve/VXLAN encapsulated packets. Thus, for a vNIC configured with an overlay, the guest stack must first segment the inner packet, compute the inner checksum for each segment and encapsulate each segment before transmitting the packet via the vNIC. This results in significant performance penalty. This patch will enhance vmxnet3 to support Geneve/VXLAN TSO as well as checksum offload. Signed-off-by:
-
Ronak Doshi authored
With vmxnet3 version 4, the emulation supports multiqueue(RSS) for UDP and ESP traffic. A guest can enable/disable RSS for UDP/ESP over IPv4/IPv6 by issuing commands introduced in this patch. ESP ipv6 is not yet supported in this patch. This patch implements get_rss_hash_opts and set_rss_hash_opts methods to allow querying and configuring different Rx flow hash configurations. Signed-off-by:
-
Ronak Doshi authored
vmxnet3 is currently at version 3 and this patch initiates the preparation to accommodate changes for version 4. Introduced utility macros for vmxnet3 version 4 comparison and update Copyright information. Signed-off-by:
-
Arnd Bergmann authored
'nic_data' is no longer used outside of the #ifdef block in efx_ef10_set_mac_address: drivers/net/ethernet/sfc/ef10.c:3231:28: error: unused variable 'nic_data' [-Werror,-Wunused-variable] struct efx_ef10_nic_data *nic_data = efx->nic_data; Move the variable into a local scope. Fixes: dfcabb07 ("sfc: move vport_id to struct efx_nic") Signed-off-by:Arnd Bergmann <arnd@arndb.de> Signed-off-by:
-
git://git.kernel.org/pub/scm/linux/kernel/git/jkirsher/next-queueDavid S. Miller authored
Jeff Kirsher says: ==================== 100GbE Intel Wired LAN Driver Updates 2020-05-27 This series contains updates to the ice driver only. Jesse fixes a number of issues, starting with fixing the remaining signed versus unsigned comparison issues. Cleaned up an unused code define. Fixed the implementation of the manage MAC write command, to simplify it by using a simple array to represent the MAC address when writing it. Paul fixes the setting of the VF default LAN address, by removing a check that assumed that the address had been deleted and zeroed. Surabhi prevents a memory leak on filter management initialization failures and during queue initialization and buffer allocation failures. Brett adds additional receive error counters that are reported by ethtool. Fixed the enabling and disabling of VLAN stripping when the PVID has been set. Evan fixes a race condition between the firmware and software, which can occur between the admin queue setup and the first command sent. Marta fixes the driver when XDP transmit rings are destroyed, also make sure the XDP transmit queues are also destroyed. Update the statistics when XDP transmit programs are loaded and packets are sent. Changed the number of XDP transmit queues to match the number of receive queues, instead of matching the number of transmit queues. Bruce avoids undefined behavior by not writing the 8-bit element init_q_state with the associated internal-to-hardware field which is 122-bits. Anirudh (Ani) refactors the receive checksum checks. Krzysztof notifies the user if the fill queue is not long enough to prepare all buffers before packet processing starts and allocates the buffers during the NAPI poll. ==================== Signed-off-by: -
David S. Miller authored
Christoph Hellwig says: ==================== remove most callers of kernel_setsockopt v3 this series removes most callers of the kernel_setsockopt functions, and instead switches their users to small functions that implement setting a sockopt directly using a normal kernel function call with type safety and all the other benefits of not having a function call. In some cases these functions seem pretty heavy handed as they do a lock_sock even for just setting a single variable, but this mirrors the real setsockopt implementation unlike a few drivers that just set set the fields directly. Changes since v2: - drop the separately merged kernel_getopt_removal - drop the sctp patches, as there is conflicting cleanup going on - add an additional ACK for the rxrpc changes Changes since v1: - use ->getname for sctp sockets in dlm - add a new ->bind_add struct proto method for dlm/sctp - switch the ipv6 and remaining sctp helpers to inline function so that the ipv6 and sctp modules are not pulled in by any module that could potentially use ipv6 or sctp connections - remove arguments to various sock_* helpers that are always used with the same constant arguments ==================== Signed-off-by: -
Christoph Hellwig authored
Avoid using kernel_setsockopt for the TIPC_IMPORTANCE option when we can just use the internal helper. The only change needed is to pass a struct sock instead of tipc_sock, which is private to socket.c Signed-off-by:
Christoph Hellwig <hch@lst.de> Signed-off-by:
-
Christoph Hellwig authored
Add a helper to directly set the RXRPC_MIN_SECURITY_LEVEL sockopt from kernel space without going through a fake uaccess. Thanks to David Howells for the documentation updates. Signed-off-by:
David Howells <dhowells@redhat.com> Signed-off-by:
-
Christoph Hellwig authored
Add a helper to directly set the IPV6_RECVPKTINFO sockopt from kernel space without going through a fake uaccess. Signed-off-by:
-
Christoph Hellwig authored
Add a helper to directly set the IPV6_ADD_PREFERENCES sockopt from kernel space without going through a fake uaccess. Signed-off-by:
-
