- 09 Jun, 2017 20 commits
-
-
Jakub Kicinski authored
The fact that RTsym table is cached inside nfp_cpp handle is a relic of old times when nfpcore was a library module. All the nfp_cpp "caches" are awkward to deal with because of concurrency and prone to keeping stale information. Make the run time symbol table be an object read out from the device and managed by whoever requested it. Since the driver loads FW at ->probe() and never reloads, we can hold onto the table for ever. Signed-off-by:
Jakub Kicinski <jakub.kicinski@netronome.com> Signed-off-by:
David S. Miller <davem@davemloft.net>
-
Jakub Kicinski authored
If very last stages of netdev registering and init fail some other netdevs and devlink ports may have been visible to user space before we torn them back down. In this case there is a slight chance user may have triggered port refresh. We need to make sure the async work is cancelled. We have to cancel after releasing pf->lock, so we will always try to cancel, regardless of which part of probe has failed. Signed-off-by:
-
Derek Chickles authored
Disallow enabling firmware debug from a VF. Only PF is allowed to do that. Signed-off-by:
Derek Chickles <derek.chickles@cavium.com> Signed-off-by:
Felix Manlunas <felix.manlunas@cavium.com> Signed-off-by:
-
Girish Moodalbail authored
There are few places on the receive path where packet drops and packet errors were not accounted for. This patch fixes that issue. Signed-off-by:
Girish Moodalbail <girish.moodalbail@oracle.com> Signed-off-by:
-
Krister Johansen authored
The ipvlan code already knows how to detect when a duplicate address is about to be assigned to an ipvlan device. However, that failure is not propogated outward and leads to a silent failure. Introduce a validation step at ip address creation time and allow device drivers to register to validate the incoming ip addresses. The ipvlan code is the first consumer. If it detects an address in use, we can return an error to the user before beginning to commit the new ifa in the networking code. This can be especially useful if it is necessary to provision many ipvlans in containers. The provisioning software (or operator) can use this to detect situations where an ip address is unexpectedly in use. Signed-off-by:
Krister Johansen <kjlx@templeofstupid.com> Signed-off-by:
-
David S. Miller authored
Stephen Hemminger says: ==================== netvsc: small cleanups These are all small optimizations found during development of later features. ==================== Signed-off-by: -
stephen hemminger authored
No longer need common code to find get_outbound_net_device. Signed-off-by:
Stephen Hemminger <sthemmin@microsoft.com> Signed-off-by:
-
stephen hemminger authored
Don't need to find netvsc_device structure, caller already had it. Also rearrange declarations. Signed-off-by:
-
stephen hemminger authored
Mark if() statements used for error handling only as unlikely() Signed-off-by:
-
stephen hemminger authored
The element netvsc_device:extension is always a pointer to RNDIS information. Signed-off-by:
-
stephen hemminger authored
Don't need need to look at write space in netvsc_close. Signed-off-by:
-
stephen hemminger authored
Speed up transmit check for fragmented packets by using existing macros to compute number of pages, and eliminate loop since skb fragments each take a page. Number of slots is also unsigned. Signed-off-by:
-
David S. Miller authored
Vivien Didelot says: ==================== net: dsa: mv88e6xxx: ops cosmetics This patchset brings no functional changes. It is a first step in a bigger cosmetics change to the driver. It simplifies print messages and polishes data types and chip operations. The next patchs will only prefix and document the port registers macros. Changes in v2: - KISS and simply use dev_* since chip->ds may not be initialized - add reviewers tags ==================== Signed-off-by: -
Vivien Didelot authored
Prefix the PHY_* macros with a Marvell specific MV88E6XXX_ prefix. There is no functional changes. Reviewed-by:
Andrew Lunn <andrew@lunn.ch> Signed-off-by:
Vivien Didelot <vivien.didelot@savoirfairelinux.com> Signed-off-by:
-
Vivien Didelot authored
Marvell chips have a Jumbo Mode to set the maximum frame size (MTU). The mv88e6xxx_ops structure is meant to contain generic functionalities, no driver logic. Change port_jumbo_config to port_set_jumbo_size setting the mode from a given maximum size value. There is no functional changes since we still use 10240 bytes. At the same time, correctly clear all Jumbo Mode bits before writing. Reviewed-by:
-
Vivien Didelot authored
All Marvell chips supporting Pause frames limiting use 1-byte value for input and output. Old chips have both bytes adjacent in a 16-bit register. New ones have an indirect table using 8-bit data. The mv88e6xxx library functions (such as in port.c) must not contain driver logic, but only generic helpers. This patch changes the port_pause_config operation for port_pause_limit taking two u8 arguments for input and output limits. There is no functional changes. Reviewed-by:
-
Vivien Didelot authored
The mv88e6xxx_ops describe functionalities, regardless their locations (which can be Global1, Global2, or whatever register set.) Rename the g1_set_cpu_port and g1_set_egress_port ops to set_cpu_port and set_egress_port. No functional changes. Signed-off-by:
-
Vivien Didelot authored
Reuse the BR_STATE_* values to abstract a port STP state value. This provides shorter names and better control over the DSA switch operation call. Signed-off-by:
-
Vivien Didelot authored
As for the frame mode, add a mv88e6xxx_egress_mode enumeration instead of a 16-bit register mask. Reviewed-by:
-
Vivien Didelot authored
The mv888e6xxx driver accesses a port's netdev mostly for printing. This is bad for 2 reasons: DSA and CPU ports do not have a netdev pointer; it doesn't give us a correct picture of why a DSA driver might need to access a port's netdev. Instead simply use dev_* printing functions with chip->dev (or ds->dev depending on the scope, both guaranteed to exist), with a p%d prefix for the target port. Signed-off-by:
-
- 08 Jun, 2017 20 commits
-
-
Willem de Bruijn authored
When inheriting tx_flags from one skbuff to another, always apply a mask to avoid overwriting unrelated other bits in the field. The two SKBTX_SHARED_FRAG cases clears all other bits. In practice, tx_flags are zero at this point now. But this is fragile. Timestamp flags are set, for instance, if in tcp_gso_segment, after this clear in skb_segment. The SKBTX_ANY_TSTAMP mask in __skb_tstamp_tx ensures that new skbs do not accidentally inherit flags such as SKBTX_SHARED_FRAG. Signed-off-by:
Willem de Bruijn <willemb@google.com> Acked-by:
Soheil Hassas Yeganeh <soheil@google.com> Acked-by:
Eric Dumazet <edumazet@google.com> Signed-off-by:
-
David S. Miller authored
Bhumika Goyal says: ==================== drivers: net: add const to mii_phy_ops structures The object references of mii_phy_ops structures are only stored in the ops field of a mii_phy_def structure. This ops field is of type const. So, mii_phy_ops structures having similar properties can be declared as const. ==================== Signed-off-by: -
Bhumika Goyal authored
The object references of mii_phy_ops structures are only stored in the ops field of a mii_phy_def structure. This ops field is of type const. So, mii_phy_ops structures having similar properties can be declared as const. Signed-off-by:
Bhumika Goyal <bhumirks@gmail.com> Signed-off-by:
-
Bhumika Goyal authored
The object references of mii_phy_ops structures are only stored in the ops field of a mii_phy_def structure. This ops field is of type const. So, mii_phy_ops structures having similar properties can be declared as const. Signed-off-by:
-
git://git.kernel.org/pub/scm/linux/kernel/git/jkirsher/next-queueDavid S. Miller authored
Jeff Kirsher says: ==================== 1GbE Intel Wired LAN Driver Updates 2017-06-07 This series contains a fix for e1000e and igb. Colin Ian King fixes sparse warnings in igb by making functions static. Chris Wilson provides a fix for a previous commit which is causing an issue during suspend "e1000e_pm_suspend()", where we need to run e1000e_pm_thaw() if __e1000_shutdown() is unsuccessful. ==================== Signed-off-by: -
Rahul Lakkireddy authored
Use PORT_REG for T4 and T5_PORT_REG for > T4 to write to correct register to bring down link during shutdown after adapter crash. Signed-off-by:
Rahul Lakkireddy <rahul.lakkireddy@chelsio.com> Signed-off-by:
Ganesh Goudar <ganeshgr@chelsio.com> Signed-off-by:
-
Nikolay Aleksandrov authored
Currently there's no way to dump the VIF table for an ipmr table other than the default (via proc). This is a major issue when debugging ipmr issues and in general it is good to know which interfaces are configured. This patch adds support for RTM_GETLINK for the ipmr family so we can dump the VIF table and the ipmr table's current config for each table. We're protected by rtnl so no need to acquire RCU or mrt_lock. Signed-off-by:
Nikolay Aleksandrov <nikolay@cumulusnetworks.com> Signed-off-by:
-
David S. Miller authored
Jiri Pirko says: ==================== mlxsw: Remove compatibility with old firmware Up until recently we couldn't enforce a minimal firmware version, which forced us to be compatible with old firmware versions. This patchset removes this code and simplifies the driver. ==================== Signed-off-by: -
Ido Schimmel authored
Previous patch made it unnecessary to map ports to modules before we allocate their struct. We can now therefore pass the port struct to these functions, thereby making them consistent with other functions that operate on ports. Signed-off-by:
Ido Schimmel <idosch@mellanox.com> Signed-off-by:
Jiri Pirko <jiri@mellanox.com> Signed-off-by:
-
Ido Schimmel authored
In commit be94535f ("mlxsw: spectrum: Make split flow match firmware requirements") we had to modify the port split flow to overcome quirks in the device's firmware. This resulted in asymmetrical code with regards to port creation and removal. The problem in the firmware is long gone and since we can now enforce a minimal firmware version, we can simplify the code and make it symmetric again. Signed-off-by:
-
Ido Schimmel authored
In new firmware versions (that we can now enforce via request_firmware()), only the first LPM tree is reserved and not the first two as in older versions. Signed-off-by:
-
David S. Miller authored
Jiri Pirko says: =================== net: Remove support from bridge bypass for mlxsw/rocker drivers Currently setting bridge port attributes and adding FDBs are done via setting the SELF flag which implies unconsistent offloading model. This patch-set fixes this behavior by making the bridge and drivers which are using it to be totally in sync. This implies several changes: - Offloading bridge flags from the bridge code. - Sending notification about FDB add/del to the software bridge in a similiar way it is done for the hardware externally learned FDBs. By making the offloading model more consistent a cleanup is done in the drivers supporting it. This is done in order to remove un-needed logic related to dump operation which is redundant. First add missing functionality to bridge, then clean up the mlxsw/rocker drivers. v1->v2 - Move bridge-switchdev related stuff to br_switchdev.c as suggested by Nik =================== Signed-off-by: -
Arkadi Sharshevsky authored
The FDB add/delete are now done through the notification chain. The FDBs are synced with the bridge and there is no need for extra dumping. Signed-off-by:
Arkadi Sharshevsky <arkadis@mellanox.com> Signed-off-by:
-
Arkadi Sharshevsky authored
The bridge port attributes/vlan for mlxsw devices should be set only from bridge code. The vlans are synced totally with the bridge so there is no need to special dump support. Signed-off-by:
-
Arkadi Sharshevsky authored
Add support for learning FDB through notification. The driver defers the hardware update via ordered work queue. Signed-off-by:
-
Arkadi Sharshevsky authored
Currently the switchdev_trans struct is embedded in the world_ops API. In order to add support for adding FDB via a notfication chain the API should be switchdev independent. Signed-off-by:
-
Arkadi Sharshevsky authored
Add support for querying supported bridge flags. Signed-off-by:
-
Arkadi Sharshevsky authored
Currently the rocker driver supports an option for disabling syncing the hardware learned FDBs with the software bridge. This behavior breaks the bridge offload model and thus it is removed. Signed-off-by:
-
Arkadi Sharshevsky authored
Remove support for bridge bypass ndos from stacked devices. At this point no driver which supports stack device behavior offload supports operation with SELF flag. The case for upper device is already taken care of in both of the following cases: 1. FDB add/del - driver should check at the notification cb if the stacked device contains his ports. 2. Port attribute - calls switchdev code directly which checks for case of stack device. Signed-off-by: -
Arkadi Sharshevsky authored
The FDB add/del are now done through the notification chain. The FDBs are synced with the bridge and there is no need for extra dumping. Signed-off-by:
-
