- 12 Oct, 2020 14 commits
-
-
Vasundhara Volam authored
This patch adds FW versions stored in the flash to devlink info_get callback. Return the correct fw.psid running version using the newly added bp->nvm_cfg_ver. v2: Ensure stored pkg_name string is NULL terminated when copied to devlink. Return directly from the last call to bnxt_dl_info_put(). If the FW call to get stored version fails for any reason, return success immediately to devlink without the stored versions. Reviewed-by:
Andy Gospodarek <gospo@broadcom.com> Signed-off-by:
Vasundhara Volam <vasundhara-v.volam@broadcom.com> Signed-off-by:
Michael Chan <michael.chan@broadcom.com> Link: https://lore.kernel.org/r/1602493854-29283-10-git-send-email-michael.chan@broadcom.comSigned-off-by:
Jakub Kicinski <kuba@kernel.org>
-
Vasundhara Volam authored
Add a new function bnxt_dl_info_put() to simplify the code, as there are more stored firmware version fields to be added in the next patch. Also, rename fw_ver variable name to ncsi_ver for better naming while copying to devlink info_get cb. v2: Ensure active_pkg_name string is NULL terminated when copied to devlink. Return directly from the last call to bnxt_dl_info_put(). Reviewed-by:
Pavan Chebbi <pavan.chebbi@broadcom.com> Reviewed-by:
-
Vasundhara Volam authored
Add a new bnxt_hwrm_nvm_get_dev_info() to query firmware version information via NVM_GET_DEV_INFO firmware command. Use it to get the running version of the NVM configuration information. This new function will also be used in subsequent patches to get the stored firmware versions. Reviewed-by:
-
Michael Chan authored
If the VF virtual link is set to always enabled, the speed may be unknown when the physical link is down. The driver currently logs the link speed as 4294967295 Mbps which is SPEED_UNKNOWN. Modify the link up log message as "speed unknown" which makes more sense. Reviewed-by:
Edwin Peer <edwin.peer@broadcom.com> Signed-off-by:
-
Michael Chan authored
Log these values that contain useful firmware state information. Reviewed-by:
-
Michael Chan authored
event_data1 and event_data2 are used when processing most events. Store these in local variables at the beginning of the function to simplify many of the case statements. Reviewed-by:
-
Michael Chan authored
Currently, bp->msg_enable has default value of 0. It is more useful to have the commonly used NETIF_MSG_DRV and NETIF_MSG_HW enabled by default. v2: Change the fall back bnxt_reset_task() inside bnxt_rx_ring_reset() to silent mode. With older fw, we would take the fall back path and it would be very noisy. Reviewed-by:
-
Vasundhara Volam authored
Online self tests are not disruptive and can be run in NPAR mode and in multi-host NIC as well. Reviewed-by:
-
Vasundhara Volam authored
If NVRAM resources are locked, NVM writes are not permitted. In such scenarios, firmware returns HWRM_ERR_CODE_RESOURCE_LOCKED error to firmware commands. Reviewed-by:
-
Jakub Kicinski authored
Merge tag 'linux-can-next-for-5.10-20201012' of git://git.kernel.org/pub/scm/linux/kernel/git/mkl/linux-can-next Marc Kleine-Budde says: ==================== linux-can-next-for-5.10-20201012 Both patches are by Oliver Hartkopp, the first one addresses Jakub's review comments of the ISOTP protocol, the other one removes version strings from various CAN protocols. ==================== Signed-off-by:
-
Ondrej Zary authored
Use netdev_err for better device identification in syslog. Signed-off-by:
Ondrej Zary <linux@zary.sk> Signed-off-by:
-
Ondrej Zary authored
When the router is rebooted without a power cycle, the USB device remains connected but its configuration is reset. This results in a non-working ethernet connection with messages like this in syslog: usb 2-2: RX packet too long: 65535 B Re-enable ethernet mode when receiving a packet with invalid size of 0xffff. Signed-off-by:
-
Oliver Hartkopp authored
As pointed out by Jakub Kicinski here: http://lore.kernel.org/r/20201009175751.5c54097f@kicinski-fedora-pc1c0hjn.dhcp.thefacebook.com this patch removes the obsolete version information of the different CAN protocols and the AF_CAN core module. Signed-off-by:
Oliver Hartkopp <socketcan@hartkopp.net> Link: https://lore.kernel.org/r/20201012074354.25839-2-socketcan@hartkopp.netSigned-off-by:
Marc Kleine-Budde <mkl@pengutronix.de>
-
Oliver Hartkopp authored
As pointed out by Jakub Kicinski here: http://lore.kernel.org/r/20201009175751.5c54097f@kicinski-fedora-pc1c0hjn.dhcp.thefacebook.com this patch addresses the remarked issues: - remove empty line in comment - remove default=y for CAN_ISOTP in Kconfig - make use of pr_notice_once() - use GFP_ATOMIC instead of gfp_any() in soft hrtimer context The version strings in the CAN subsystem are removed by a separate patch. Signed-off-by:
-
- 11 Oct, 2020 9 commits
-
-
Jakub Kicinski authored
Vladimir Oltean says: ==================== Offload tc-vlan mangle to mscc_ocelot switch This series offloads one more action to the VCAP IS1 ingress TCAM, which is to change the classified VLAN for packets, according to the VCAP IS1 keys (VLAN, source MAC, source IP, EtherType, etc). ==================== Signed-off-by: -
Vladimir Oltean authored
Create a test that changes a VLAN ID from 200 to 300. We also need to modify the preferences of the filters installed for the other rules so that they are unique, because we now install the "tc-vlan modify" filter in VCAP IS1 only temporarily, and we need to perform the deletion by filter preference number. Signed-off-by:
Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by:
-
Vladimir Oltean authored
When the Extraction Frame Header contains a valid classified VLAN, use that instead of the VLAN header present in the packet. Signed-off-by:
-
Vladimir Oltean authored
The VCAP_IS1_ACT_VID_REPLACE_ENA action, from the VCAP IS1 ingress TCAM, changes the classified VLAN. We are only exposing this ability for switch ports that are under VLAN aware bridges. This is because in standalone ports mode and under a bridge with vlan_filtering=0, the ocelot driver configures the switch to operate as VLAN-unaware, so the classified VLAN is not derived from the 802.1Q header from the packet, but instead is always equal to the port-based VLAN ID of the ingress port. We _can_ still change the classified VLAN for packets when operating in this mode, but the end result will most likely be a drop, since both the ingress and the egress port need to be members of the modified VLAN. And even if we install the new classified VLAN into the VLAN table of the switch, the result would still not be as expected: we wouldn't see, on the output port, the modified VLAN tag, but the original one, even though the classified VLAN was indeed modified. This is because of how the hardware works: on egress, what is pushed to the frame is a "port tag", which gives us the following options: - Tag all frames with port tag (derived from the classified VLAN) - Tag all frames with port tag, except if the classified VLAN is 0 or equal to the native VLAN of the egress port - No port tag Needless to say, in VLAN-unaware mode we are disabling the port tag. Otherwise, the existing VLAN tag would be ignored, and a second VLAN tag (the port tag), holding the classified VLAN, would be pushed (instead of replacing the existing 802.1Q tag). This is definitely not what the user wanted when installing a "vlan modify" action. So it is simply not worth bothering with VLAN modify rules under other configurations except when the ports are fully VLAN-aware. Signed-off-by:
-
Jakub Kicinski authored
Claudiu Manoil says: ==================== enetc: Migrate to PHYLINK and PCS_LYNX Transitioning the enetc driver from phylib to phylink. Offloading the serdes configuration to the PCS_LYNX module is a mandatory part of this transition. Aiming for a cleaner, more maintainable design, and better code reuse. The first 2 patches are clean up prerequisites. Tested on a p1028rdb board. v2: validate() explicitly rejects now all interface modes not supported by the driver instead of relying on the device tree to provide only supported interfaces, and dropped redundant activation of pcs_poll (addressing Ioana's findings) ==================== Signed-off-by: -
Claudiu Manoil authored
This is a methodical transition of the driver from phylib to phylink, following the guidelines from sfp-phylink.rst. The MAC register configurations based on interface mode were moved from the probing path to the mac_config() hook. MAC enable and disable commands (enabling Rx and Tx paths at MAC level) were also extracted and assigned to their corresponding phylink hooks. As part of the migration to phylink, the serdes configuration from the driver was offloaded to the PCS_LYNX module, introduced in commit 0da4c3d3 ("net: phy: add Lynx PCS module"), the PCS_LYNX module being a mandatory component required to make the enetc driver work with phylink. Signed-off-by:
Claudiu Manoil <claudiu.manoil@nxp.com> Reviewed-by:
Ioana Ciornei <ioana.cionei@nxp.com> Signed-off-by:
-
Claudiu Manoil authored
As part of the transition of the enetc ethernet driver from phylib to phylink, the in-band operation mode of the SGMII interface from enetc port 0 needs to be specified explicitly for phylink. Signed-off-by:
-
Claudiu Manoil authored
Decouple internal mdio bus creation from serdes configuration, as a prerequisite to offloading serdes configuration to a different module. Group together mdio bus creation routines, cleanup. Signed-off-by:
Ioana Ciornei <ioana.ciornei@nxp.com> Signed-off-by:
-
Claudiu Manoil authored
Decouple level MAC configuration based on phy interface type from general port configuration. Group together MAC and link configuration code. Decouple external mdio bus creation from interface type parsing. No longer return an (unhandled) error code when phy_node not found, use phy_node to indicate whether the port has a phy or not. No longer fall-through when serdes configuration fails for the link modes that require internal link configuration. Signed-off-by:
-
- 10 Oct, 2020 17 commits
-
-
Xie He authored
Move the skb_headroom check out of fr_hard_header and into pvc_xmit. This has two benefits: 1. Originally we only do this check for skbs sent by users on Ethernet- emulating PVC devices. After the change we do this check for skbs sent on normal PVC devices, too. (Also add a comment to make it clear that this is only a protection against upper layers that don't take dev->needed_headroom into account. Such upper layers should be rare and I believe they should be fixed.) 2. After the change we can simplify the parameter list of fr_hard_header. We no longer need to use a pointer to pointers (skb_p) because we no longer need to replace the skb inside fr_hard_header. Cc: Krzysztof Halasa <khc@pm.waw.pl> Signed-off-by:
Xie He <xie.he.0141@gmail.com> Signed-off-by:
-
Linus Walleij authored
The MTU setting for this DSA switch is global so we need to keep track of the MTU set for each port, then as soon as any MTU changes, roof the MTU to the biggest common denominator and poke that into the switch MTU setting. To achieve this we need a per-chip-variant state container for the RTL8366RB to use for the RTL8366RB-specific stuff. Other SMI switches does seem to have per-port MTU setting capabilities. Fixes: 5f4a8ef3 ("net: dsa: rtl8366rb: Support setting MTU") Signed-off-by:
Linus Walleij <linus.walleij@linaro.org> Reviewed-by:
Florian Fainelli <f.fainelli@gmail.com> Signed-off-by:
-
Calvin Johnson authored
Better place for of_mdio.c is drivers/net/mdio. Move of_mdio.c from drivers/of to drivers/net/mdio Signed-off-by:
Calvin Johnson <calvin.johnson@oss.nxp.com> Acked-by:
Rob Herring <robh@kernel.org> Signed-off-by:
-
Maxim Kochetkov authored
When packets are received on the error queue, this function under net_ratelimit(): netif_err(priv, hw, net_dev, "Err FD status = 0x%08x\n"); does not get printed. Instead we only see: [ 3658.845592] net_ratelimit: 244 callbacks suppressed [ 3663.969535] net_ratelimit: 230 callbacks suppressed [ 3669.085478] net_ratelimit: 228 callbacks suppressed Enabling NETIF_MSG_HW fixes this issue, and we can see some information about the frame descriptors of packets. Signed-off-by:
Maxim Kochetkov <fido_max@inbox.ru> Signed-off-by:
Madalin Bucur <madalin.bucur@oss.nxp.com> Signed-off-by:
-
Heiner Kallweit authored
Factor out handling the private packet/byte counters to new functions rtl_get_priv_stats() and rtl_inc_priv_stats(). Signed-off-by:
Heiner Kallweit <hkallweit1@gmail.com> Signed-off-by:
-
Heiner Kallweit authored
Obviously this driver version doesn't make sense. Go with the default and let ethtool display the kernel version. Signed-off-by:
Greg Kroah-Hartman <gregkh@linuxfoundation.org> Signed-off-by:
-
Gustavo A. R. Silva authored
Make use of the new struct_size() helper instead of the offsetof() idiom. Signed-off-by:
Gustavo A. R. Silva <gustavoars@kernel.org> Signed-off-by:
-
Jakub Kicinski authored
Merge tag 'wireless-drivers-next-2020-10-09' of git://git.kernel.org/pub/scm/linux/kernel/git/kvalo/wireless-drivers-next Kalle Valo says: ==================== wireless-drivers-next patches for v5.10 Fourth and last set of patches for v5.10. Most of these are iwlwifi patches, but few small fixes to other drivers as well. Major changes: iwlwifi * PNVM support (platform-specific phy config data) * bump the FW API support to 59 ==================== Signed-off-by:
-
Jakub Kicinski authored
Merge tag 'mac80211-next-for-net-next-2020-10-08' of git://git.kernel.org/pub/scm/linux/kernel/git/jberg/mac80211-next Johannes Berg says: ==================== A handful of changes: * fixes for the recent S1G work * a docbook build time improvement * API to pass beacon rate to lower-level driver ==================== Signed-off-by:
-
Jakub Kicinski authored
Johannes Berg says: ==================== netlink: export policy on validation failures Export the policy used for attribute validation when it fails, so e.g. for an out-of-range attribute userspace immediately gets the valid ranges back. v2 incorporates the suggestion from Jakub to have a function to estimate the size (netlink_policy_dump_attr_size_estimate()) and check that it does the right thing on the *normal* policy dumps, not (just) when calling it from the error scenario. v3 only addresses a few minor style issues. v4 fixes up a forgotten 'git add' ... sorry. v5 is a resend, I messed up v4's cover letter subject (saying v3) and apparently the second patch didn't go out at all. Tested using nl80211/iw in a few scenarios, seems to work fine and return the policy back, e.g. kernel reports: integer out of range policy: 04 00 0b 00 0c 00 04 00 01 00 00 00 00 00 00 00 ^ padding ^ minimum allowed value policy: 04 00 0b 00 0c 00 05 00 ff ff ff ff 00 00 00 00 ^ padding ^ maximum allowed value policy: 08 00 01 00 04 00 00 00 ^ type 4 == U32 for an out-of-range case. ==================== Signed-off-by: -
Johannes Berg authored
Add a new attribute NLMSGERR_ATTR_POLICY to the extended ACK to advertise the policy, e.g. if an attribute was out of range, you'll know the range that's permissible. Add new NL_SET_ERR_MSG_ATTR_POL() and NL_SET_ERR_MSG_ATTR_POL() macros to set this, since realistically it's only useful to do this when the bad attribute (offset) is also returned. Use it in lib/nlattr.c which practically does all the policy validation. v2: - add and use netlink_policy_dump_attr_size_estimate() v3: - remove redundant break v4: - really remove redundant break ... sorry Reviewed-by:
Johannes Berg <johannes.berg@intel.com> Signed-off-by:
-
Johannes Berg authored
Refactor the per-attribute policy writing into a new helper function, to be used later for dumping out the policy of a rejected attribute. v2: - fix some indentation v3: - change variable order in netlink_policy_dump_write() Reviewed-by:
-
Jakub Kicinski authored
Karsten Graul says: ==================== net/smc: updates 2020-10-07 Patch 1 and 2 address warnings from static code checkers, and patch 3 handles a case when all proposed ISM V2 devices fail to init and no V1 devices are tried afterwards. ==================== Signed-off-by: -
Karsten Graul authored
Field ini->smcd_version is set to SMC_V2 before calling smc_listen_ism_init(). This clears the V1 bit that may be set. When all matching ISM V2 devices fail to initialize then the smcd_version field needs to get restored to allow any possible V1 devices to initialize. And be consistent, always go to the not_found label when no device was found. Signed-off-by:
Karsten Graul <kgraul@linux.ibm.com> Signed-off-by:
-
Karsten Graul authored
coccinelle informs about net/smc/af_smc.c:1770:10-11: WARNING: opportunity for kzfree/kvfree_sensitive Its not that kzfree() would help here, the memset() is done to prepare the buffer for another socket receive. Fix that warning message by reordering the calls, while at it eliminate the unneeded variable cclc2 and use sizeof(*buf) as above in the same function. No functional changes. Signed-off-by:
-
Karsten Graul authored
Static code checkers warn of inconsistent returns because the lgr mutex is locked in one function and unlocked in a function called by the locking function: net/smc/af_smc.c:823 smc_connect_rdma() warn: inconsistent returns 'smc_client_lgr_pending'. net/smc/af_smc.c:897 smc_connect_ism() warn: inconsistent returns 'smc_server_lgr_pending'. Make the code consistent by doing the unlock in the same function that fetches the lock. No functional changes. Signed-off-by:
-
Jakub Kicinski authored
Merge tag 'linux-can-next-for-5.10-20201007' of git://git.kernel.org/pub/scm/linux/kernel/git/mkl/linux-can-next Marc Kleine-Budde says: ==================== linux-can-next-for-5.10-20201007 The first 3 patches are by me and fix several warnings found when compiling the kernel with W=1. Lukas Bulwahn's patch adjusts the MAINTAINERS file, to accommodate the renaming of the mcp251xfd driver. Vincent Mailhol contributes 3 patches for the CAN networking layer. First error queue support is added the the CAN RAW protocol. The second patch converts the get_can_dlc() and get_canfd_dlc() in-Kernel-only macros from using __u8 to u8. The third patch adds a helper function to calculate the length of one bit in in multiple of time quanta. Oliver Hartkopp's patch add support for the ISO 15765-2:2016 transport protocol to the CAN stack. Three patches by Lad Prabhakar add documentation for various new rcar controllers to the device tree bindings of the rcar_can and rcan_canfd driver. Michael Walle's patch adds various processors to the flexcan driver binding documentation. The next two patches are by me and target the flexcan driver aswell. The remove the ack_grp and ack_bit from the fsl,stop-mode DT property and the driver, as they are not used anymore. As these are the last two arguments this change will not break existing device trees. The last three patches are by Srinivas Neeli and target the xilinx_can driver. The first one increases the lower limit for the bit rate prescaler to 2, the other two fix sparse and coverity findings. ==================== Signed-off-by:
-
