- 31 10月, 2016 10 次提交
-
-
由 Jiri Pirko 提交于
Currently, devlink register/unregister is done directly from spectrum/switchx2 port create/remove functions. With a need to introduce a port type change, the devlink port instances have to be persistent across type changes, therefore across port create/remove function calls. So do a bit of reshuffling to achieve that. Signed-off-by: NJiri Pirko <jiri@mellanox.com> Signed-off-by: NElad Raz <eladr@mellanox.com> Reviewed-by: NIdo Schimmel <idosch@mellanox.com> Signed-off-by: NJiri Pirko <jiri@mellanox.com> Signed-off-by: NDavid S. Miller <davem@davemloft.net>
-
由 Elad Raz 提交于
In order to change a port type to Infiniband port we should change his mapping from local-port to Infiniband. Adding the PLIB (Port Local to InfiniBand) allows this mapping. Signed-off-by: NElad Raz <eladr@mellanox.com> Reviewed-by: NJiri Pirko <jiri@mellanox.com> Reviewed-by: NIdo Schimmel <idosch@mellanox.com> Signed-off-by: NJiri Pirko <jiri@mellanox.com> Signed-off-by: NDavid S. Miller <davem@davemloft.net>
-
由 Elad Raz 提交于
In order to support Infiniband fabric, we need to introduce IB speeds and capabilities to PTYS emads. Signed-off-by: NElad Raz <eladr@mellanox.com> Reviewed-by: NJiri Pirko <jiri@mellanox.com> Reviewed-by: NIdo Schimmel <idosch@mellanox.com> Signed-off-by: NJiri Pirko <jiri@mellanox.com> Signed-off-by: NDavid S. Miller <davem@davemloft.net>
-
由 Elad Raz 提交于
We want to add Infiniband support to PTYS. In order to maintain proper conventions, we will change pack and unpack prefix to eth. Signed-off-by: NElad Raz <eladr@mellanox.com> Reviewed-by: NJiri Pirko <jiri@mellanox.com> Reviewed-by: NIdo Schimmel <idosch@mellanox.com> Signed-off-by: NJiri Pirko <jiri@mellanox.com> Signed-off-by: NDavid S. Miller <davem@davemloft.net>
-
由 Elad Raz 提交于
In SwitchX-2 we configure the port speed to negotiate with 40G link only. Add support for all other supported speeds. Fixes: 31557f0f ("mlxsw: Introduce Mellanox SwitchX-2 ASIC support") Signed-off-by: NElad Raz <eladr@mellanox.com> Reviewed-by: NJiri Pirko <jiri@mellanox.com> Reviewed-by: NIdo Schimmel <idosch@mellanox.com> Signed-off-by: NJiri Pirko <jiri@mellanox.com> Signed-off-by: NDavid S. Miller <davem@davemloft.net>
-
由 Elad Raz 提交于
Export to userspace the front panel name of the port, so that udev can rename the ports accordingly. The convention suggested by switchdev documentation is used: pX Signed-off-by: NElad Raz <eladr@mellanox.com> Reviewed-by: NJiri Pirko <jiri@mellanox.com> Reviewed-by: NIdo Schimmel <idosch@mellanox.com> Signed-off-by: NJiri Pirko <jiri@mellanox.com> Signed-off-by: NDavid S. Miller <davem@davemloft.net>
-
由 Jiri Pirko 提交于
Be symmentrical with create and do the check outside the remove function. Signed-off-by: NJiri Pirko <jiri@mellanox.com> Signed-off-by: NElad Raz <eladr@mellanox.com> Signed-off-by: NDavid S. Miller <davem@davemloft.net>
-
由 Jiri Pirko 提交于
Be symmentrical with create and do the check outside the remove function. Signed-off-by: NJiri Pirko <jiri@mellanox.com> Signed-off-by: NElad Raz <eladr@mellanox.com> Signed-off-by: NDavid S. Miller <davem@davemloft.net>
-
由 Jiri Pirko 提交于
Do it in a same way we do it in spectrum. Check if port is usable first and only in that case create a port instance. Signed-off-by: NJiri Pirko <jiri@mellanox.com> Reviewed-by: NElad Raz <eladr@mellanox.com> Reviewed-by: NIdo Schimmel <idosch@mellanox.com> Signed-off-by: NDavid S. Miller <davem@davemloft.net>
-
由 Elad Raz 提交于
We recently discovered a bug in the firmware in which a field's length in one of the registers was incorrectly set. This caused the firmware to access garbage data that wasn't initialized by the driver and therefore emit error messages. While the bug is already fixed and the driver usually zeros the buffers passed to the firmware, there are a handful of cases where this isn't done. Zero the buffer in these cases and prevent similar bugs from recurring, as they tend to be hard to debug. Fixes: 52581961 ("mlxsw: core: Implement fan control using hwmon") Signed-off-by: NElad Raz <eladr@mellanox.com> Reviewed-by: NJiri Pirko <jiri@mellanox.com> Reviewed-by: NIdo Schimmel <idosch@mellanox.com> Signed-off-by: NJiri Pirko <jiri@mellanox.com> Signed-off-by: NDavid S. Miller <davem@davemloft.net>
-
- 30 10月, 2016 23 次提交
-
-
由 Tariq Toukan 提交于
Following the previous patch, as an optimization, the slave will not even bother sending the DUMP_ETH_STATS command over the comm channel. Signed-off-by: NTariq Toukan <tariqt@mellanox.com> Signed-off-by: NDavid S. Miller <davem@davemloft.net>
-
由 Jack Morgenstein 提交于
mlx4_en_DUMP_ETH_STATS took the *counter mutex* and then called the FW command, with WRAPPED attribute. As a result, the fw command is wrapped on the Hypervisor when it calls mlx4_en_DUMP_ETH_STATS. The FW command wrapper flow on the hypervisor takes the *slave_cmd_mutex* during processing. At the same time, a VF could be in the process of coming up, and could call mlx4_QUERY_FUNC_CAP. On the hypervisor, the command flow takes the *slave_cmd_mutex*, then executes mlx4_QUERY_FUNC_CAP_wrapper. mlx4_QUERY_FUNC_CAP wrapper calls mlx4_get_default_counter_index(), which takes the *counter mutex*. DEADLOCK. The fix is that the DUMP_ETH_STATS fw command should be called with the NATIVE attribute, so that on the hypervisor, this command does not enter the wrapper flow. Since the Hypervisor no longer goes through the wrapper code, we also simply return 0 in mlx4_DUMP_ETH_STATS_wrapper (i.e.the function succeeds, but the returned data will be all zeroes). No need to test if it is the Hypervisor going through the wrapper. Fixes: f9baff50 ("mlx4_core: Add "native" argument to mlx4_cmd ...") Signed-off-by: NJack Morgenstein <jackm@dev.mellanox.co.il> Signed-off-by: NTariq Toukan <tariqt@mellanox.com> Signed-off-by: NDavid S. Miller <davem@davemloft.net>
-
由 Eugenia Emantayev 提交于
Currently interrupt test that is part of ethtool selftest runs the check over all interrupt vectors of the device. In mlx4_en package part of interrupt vectors are uninitialized since mlx4_ib doesn't exist. This causes NOP FW command to time out. Change logic to test current port interrupt vectors only. Signed-off-by: NEugenia Emantayev <eugenia@mellanox.com> Signed-off-by: NTariq Toukan <tariqt@mellanox.com> Signed-off-by: NDavid S. Miller <davem@davemloft.net>
-
由 Jack Morgenstein 提交于
In the Hypervisor, there are several FW commands which are invoked before the comm channel is initialized (in mlx4_multi_func_init). These include MOD_STAT_CONFIG, QUERY_DEV_CAP, INIT_HCA, and others. If any of these commands fails, say with a timeout, the Hypervisor driver enters the internal error reset flow. In this flow, the driver attempts to notify all slaves via the comm channel that an internal error has occurred. Since the comm channel has not yet been initialized (i.e., mapped via ioremap), this will cause dereferencing a NULL pointer. To fix this, do not access the comm channel in the internal error flow if it has not yet been initialized. Fixes: 55ad3592 ("net/mlx4_core: Enable device recovery flow with SRIOV") Fixes: ab9c17a0 ("mlx4_core: Modify driver initialization flow to accommodate SRIOV for Ethernet") Signed-off-by: NJack Morgenstein <jackm@dev.mellanox.co.il> Signed-off-by: NTariq Toukan <tariqt@mellanox.com> Signed-off-by: NDavid S. Miller <davem@davemloft.net>
-
由 Eugenia Emantayev 提交于
Fix a kernel panic that occurs as a result of an asynchronous event handled in roce_gid_mgmt: mlx4_en_get_drvinfo is called and accesses freed resources. This happens in a shutdown flow only, since pci device is destroyed while netdevice is still alive. Fixes: c27a02cd ("mlx4_en: Add driver for Mellanox ConnectX 10GbE NIC") Signed-off-by: NEugenia Emantayev <eugenia@mellanox.com> Signed-off-by: NTariq Toukan <tariqt@mellanox.com> Signed-off-by: NDavid S. Miller <davem@davemloft.net>
-
由 Erez Shitrit 提交于
Currently there is a race between incoming traffic and initialization flow. HW is able to receive the packets after INIT_PORT is done and unicast steering is configured. Before we set priv->port_up NAPI is not scheduled and receive queues become full. Therefore we never get new interrupts about the completions. This issue could happen if running heavy traffic during bringing port up. The resolution is to schedule NAPI once port_up is set. If receive queues were full this will process all cqes and release them. Fixes: c27a02cd ("mlx4_en: Add driver for Mellanox ConnectX 10GbE NIC") Signed-off-by: NErez Shitrit <erezsh@mellanox.com> Signed-off-by: NEugenia Emantayev <eugenia@mellanox.com> Signed-off-by: NTariq Toukan <tariqt@mellanox.com> Signed-off-by: NDavid S. Miller <davem@davemloft.net>
-
由 Eugenia Emantayev 提交于
When doing roundup_pow_of_two for large enough number with bit 31, an overflow will occur and a value equal to 1 will be returned. In this case 1 will be subtracted from the return value and division by zero will be reached. Fixes: 31c128b6 ("net/mlx4_en: Choose time-stamping shift value according to HW frequency") Signed-off-by: NEugenia Emantayev <eugenia@mellanox.com> Signed-off-by: NTariq Toukan <tariqt@mellanox.com> Signed-off-by: NDavid S. Miller <davem@davemloft.net>
-
由 Moshe Lazer 提交于
Change the default status of quality of service back to disabled, as it hurts performance in some cases. Fixes: 38438f7c ("net/mlx4: Set enhanced QoS support by default when ...") Signed-off-by: NMoshe Lazer <moshel@mellanox.com> Signed-off-by: NTariq Toukan <tariqt@mellanox.com> Signed-off-by: NDavid S. Miller <davem@davemloft.net>
-
由 Maor Gottlieb 提交于
When only one port type is supported, it should be read only. We reject changing requests, even to the auto sense mode. Fixes: 27bf91d6 ("mlx4_core: Add link type autosensing") Signed-off-by: NMaor Gottlieb <maorg@mellanox.com> Signed-off-by: NMoshe Shemesh <moshe@mellanox.com> Signed-off-by: NTariq Toukan <tariqt@mellanox.com> Signed-off-by: NDavid S. Miller <davem@davemloft.net>
-
由 Jack Morgenstein 提交于
The resource type enum in the resource tracker was incorrect. RES_EQ was put in the position of RES_NPORT_ID (a FC resource). Since the remaining resources maintain their current values, and RES_EQ is not passed from slaves to the hypervisor in any FW command, this change affects only the hypervisor. Therefore, there is no backwards-compatibility issue. Fixes: 623ed84b ("mlx4_core: initial header-file changes for SRIOV support") Signed-off-by: NJack Morgenstein <jackm@dev.mellanox.co.il> Signed-off-by: NMoshe Shemesh <moshe@mellanox.com> Signed-off-by: NTariq Toukan <tariqt@mellanox.com> Signed-off-by: NDavid S. Miller <davem@davemloft.net>
-
由 Arnd Bergmann 提交于
There is an old warning about mlx4_SW2HW_EQ_wrapper on x86: ethernet/mellanox/mlx4/resource_tracker.c: In function ‘mlx4_SW2HW_EQ_wrapper’: ethernet/mellanox/mlx4/resource_tracker.c:3071:10: error: ‘eq’ may be used uninitialized in this function [-Werror=maybe-uninitialized] The problem here is that gcc won't track the state of the variable across a spin_unlock. Moving the assignment out of the lock is safe here and avoids the warning. Signed-off-by: NArnd Bergmann <arnd@arndb.de> Reviewed-by: NYishai Hadas <yishaih@mellanox.com> Signed-off-by: NDavid S. Miller <davem@davemloft.net>
-
由 Noa Osherovich 提交于
Currently the firmware can't work with a page with dma address 0. Passing such an address to the firmware will cause the give_pages command to fail. To avoid this, in case we get a 0 dma address of a page from the dma engine, we avoid passing it to FW by remapping to get an address other than 0. Fixes: bf0bf77f ('mlx5: Support communicating arbitrary host...') Signed-off-by: NNoa Osherovich <noaos@mellanox.com> Signed-off-by: NSaeed Mahameed <saeedm@mellanox.com> Signed-off-by: NDavid S. Miller <davem@davemloft.net>
-
由 Mohamad Haj Yahia 提交于
In case that the kernel PCI error handlers are not called, we will trigger our own recovery flow. The health work will give priority to the kernel pci error handlers to recover the PCI by waiting for a small period, if the pci error handlers are not triggered the manual recovery flow will be executed. We don't save pci state in case of manual recovery because it will ruin the pci configuration space and we will lose dma sync. Fixes: 89d44f0a ('net/mlx5_core: Add pci error handlers to mlx5_core driver') Signed-off-by: NMohamad Haj Yahia <mohamad@mellanox.com> Signed-off-by: NSaeed Mahameed <saeedm@mellanox.com> Signed-off-by: NDavid S. Miller <davem@davemloft.net>
-
由 Mohamad Haj Yahia 提交于
Currently there is a race between the health care work and the kernel pci error handlers because both of them detect the error, the first one to be called will do the error handling. There is a chance that health care will disable the pci after resuming pci slot. Also create a separate WQ because now we will have two types of health works, one for the error detection and one for the recovery. Fixes: 89d44f0a ('net/mlx5_core: Add pci error handlers to mlx5_core driver') Signed-off-by: NMohamad Haj Yahia <mohamad@mellanox.com> Signed-off-by: NSaeed Mahameed <saeedm@mellanox.com> Signed-off-by: NDavid S. Miller <davem@davemloft.net>
-
由 Mohamad Haj Yahia 提交于
The health sick status should be cleared when we start the health poll. This is crucial for driver reload (unload + load) in order to behave right in case of health issue. Fixes: fd76ee4d ('net/mlx5_core: Fix internal error detection conditions') Signed-off-by: NMohamad Haj Yahia <mohamad@mellanox.com> Signed-off-by: NSaeed Mahameed <saeedm@mellanox.com> Signed-off-by: NDavid S. Miller <davem@davemloft.net>
-
由 Mohamad Haj Yahia 提交于
The Ingress/Egress ACL enable function may fail and it should return status to its caller to avoid NULL pointer dereference. Fixes: f942380c ('net/mlx5: E-Switch, Vport ingress/egress ACLs rules for spoofchk') Signed-off-by: NMohamad Haj Yahia <mohamad@mellanox.com> Signed-off-by: NSaeed Mahameed <saeedm@mellanox.com> Signed-off-by: NDavid S. Miller <davem@davemloft.net>
-
由 Mohamad Haj Yahia 提交于
Detaching the netdev before unregistering it cause some netdev cleanup ndos to fail because they check presence of the netdev, so we need to unregister the netdev first. Fixes: 26e59d80 ('net/mlx5e: Implement mlx5e interface attach/detach callbacks') Signed-off-by: NMohamad Haj Yahia <mohamad@mellanox.com> Signed-off-by: NSaeed Mahameed <saeedm@mellanox.com> Signed-off-by: NDavid S. Miller <davem@davemloft.net>
-
由 Saeed Mahameed 提交于
Instead of predicting the index of the wanted LRO timeout value from hardware capabilities, look for the nearest LRO timeout value. Fixes: 5c50368f ('net/mlx5e: Light-weight netdev open/stop') Signed-off-by: NSaeed Mahameed <saeedm@mellanox.com> Signed-off-by: NMohamad Haj Yahia <mohamad@mellanox.com> Signed-off-by: NDavid S. Miller <davem@davemloft.net>
-
由 Paul Blakey 提交于
Currently, last use timestamp is initialized to zero. This is not the expected value by higher layers such as when we do TC action offloading. To fix that, set it to the current time, e.g when the counter/rule is offloaded. This is the same behaviour of non-offloaded TC actions. Fixes: 43a335e0 ('mlx5_core: Flow counters infrastructure') Signed-off-by: NPaul Blakey <paulb@mellanox.com> Signed-off-by: NSaeed Mahameed <saeedm@mellanox.com> Signed-off-by: NDavid S. Miller <davem@davemloft.net>
-
由 Paul Blakey 提交于
Autogroups groups num is increased when creating a new flow group, but is never decreased. Now decreasing it when deleting a flow group. Fixes: f0d22d18 ('net/mlx5_core: Introduce flow steering autogrouped flow table') Signed-off-by: NPaul Blakey <paulb@mellanox.com> Signed-off-by: NSaeed Mahameed <saeedm@mellanox.com> Signed-off-by: NDavid S. Miller <davem@davemloft.net>
-
由 Paul Blakey 提交于
Finding a new autogroup range is done by going over a group list sorted by each group start index. The search is stopped after finding the first free range. Adding the newly created group to the list is wrongly added to the end of the list regardless of its start index as the parameter of where to insert it is ignored. This commit makes sure to use that unused parameter to insert it where requested. Fixes: f0d22d18 ('net/mlx5_core: Introduce flow steering autogrouped flow table') Signed-off-by: NPaul Blakey <paulb@mellanox.com> Signed-off-by: NSaeed Mahameed <saeedm@mellanox.com> Signed-off-by: NDavid S. Miller <davem@davemloft.net>
-
由 Daniel Jurgens 提交于
Always query the HCA caps after setting them to update the capablities data structures. Not doing so results in incorrect capabilities being reported including max_dc, max_qp and several others. Fixes: 59211bd3 ("net/mlx5: Split the load/unload flow into hardware and software flows") Signed-off-by: NDaniel Jurgens <danielj@mellanox.com> Signed-off-by: NSaeed Mahameed <saeedm@mellanox.com> Signed-off-by: NDavid S. Miller <davem@davemloft.net>
-
由 Daniel Jurgens 提交于
ARM 64B cache line systems have L1_CACHE_BYTES set to 128. cache_line_size() will return the correct size. Fixes: cf50b5efa2fe('net/mlx5_core/ib: New device capabilities handling.') Signed-off-by: NDaniel Jurgens <danielj@mellanox.com> Signed-off-by: NSaeed Mahameed <saeedm@mellanox.com> Signed-off-by: NDavid S. Miller <davem@davemloft.net>
-
- 29 10月, 2016 7 次提交
-
-
由 Jiri Pirko 提交于
Do this so the sysfs has "device" link correctly set. Signed-off-by: NJiri Pirko <jiri@mellanox.com> Reviewed-by: NIdo Schimmel <idosch@mellanox.com> Signed-off-by: NDavid S. Miller <davem@davemloft.net>
-
由 Jiri Pirko 提交于
Do this so the sysfs has "device" link correctly set. Signed-off-by: NJiri Pirko <jiri@mellanox.com> Reviewed-by: NIdo Schimmel <idosch@mellanox.com> Signed-off-by: NDavid S. Miller <davem@davemloft.net>
-
由 Jiri Pirko 提交于
So far, mlxsw_pci.ko is the module that registers PCI table for all drivers (spectrum and switchx2). That is problematic for example with dracut. Since mlxsw_spectrum.ko and mlxsw_switchx2.ko are loaded dynamically from within mlxsw_core.ko, dracut does not have track of them and avoids them from being included in initramfs. So make this in an ordinary way and define the PCI tables in individual driver modules, so it can be properly loaded and included in dracut initramfs image. As a side effect, this patch could remove no longer necessary driver "kind" strings which were used to link PCI ids with individual mlxsw drivers. Suggested-by: NIvan Vecera <ivecera@redhat.com> Tested-by: NIvan Vecera <ivecera@redhat.com> Signed-off-by: NJiri Pirko <jiri@mellanox.com> Reviewed-by: NIdo Schimmel <idosch@mellanox.com> Acked-by: NIvan Vecera <ivecera@redhat.com> Signed-off-by: NDavid S. Miller <davem@davemloft.net>
-
由 Jiri Pirko 提交于
pci.h needs to be used for inner function declarations. So move the original one to more appropriate name, pci_hw.h. Signed-off-by: NJiri Pirko <jiri@mellanox.com> Reviewed-by: NIdo Schimmel <idosch@mellanox.com> Signed-off-by: NDavid S. Miller <davem@davemloft.net>
-
由 Ido Schimmel 提交于
Signed-off-by: NIdo Schimmel <idosch@mellanox.com> Reviewed-by: NJiri Pirko <jiri@mellanox.com> Signed-off-by: NJiri Pirko <jiri@mellanox.com> Signed-off-by: NDavid S. Miller <davem@davemloft.net>
-
由 Jiri Pirko 提交于
Only trees which are in use should be compared to requested prefix usage. Fixes: 53342023 ("mlxsw: spectrum_router: Implement LPM trees management") Signed-off-by: NJiri Pirko <jiri@mellanox.com> Reviewed-by: NIdo Schimmel <idosch@mellanox.com> Signed-off-by: NDavid S. Miller <davem@davemloft.net>
-
由 Jiri Pirko 提交于
Currently, the prefix bitlist is not saved for LPM trees, causing the compare to always fail which causes the tree to be destroyed and created for every inserted and removed FIB entry. So fix this by saving the bitlist as it should have been done from the very beginning. Fixes: 53342023 ("mlxsw: spectrum_router: Implement LPM trees management") Signed-off-by: NJiri Pirko <jiri@mellanox.com> Reviewed-by: NIdo Schimmel <idosch@mellanox.com> Signed-off-by: NDavid S. Miller <davem@davemloft.net>
-