From 958b3d396d7f80755e2c2e6a8f873a669f38de10 Mon Sep 17 00:00:00 2001 From: Brenden Blanco Date: Thu, 13 Oct 2016 13:13:11 -0700 Subject: net/mlx4_en: fixup xdp tx irq to match rx In cases where the number of tx rings is not a multiple of the number of rx rings, the tx completion event will be handled on a different core from the transmit and population of the ring. Races on the ring will lead to a double-free of the page, and possibly other corruption. The rings are initialized by default with a valid multiple of rings, based on the number of cpus, therefore an invalid configuration requires ethtool to change the ring layout. For instance 'ethtool -L eth0 rx 9 tx 8' will cause packets received on rx0, and XDP_TX'd to tx48, to be completed on cpu3 (48 % 9 == 3). Resolve this discrepancy by shifting the irq for the xdp tx queues to start again from 0, modulo rx_ring_num. Fixes: 9ecc2d86171a ("net/mlx4_en: add xdp forwarding and data write support") Reported-by: Jesper Dangaard Brouer Signed-off-by: Brenden Blanco Signed-off-by: David S. Miller --- drivers/net/ethernet/mellanox/mlx4/en_cq.c | 10 +++++++++- 1 file changed, 9 insertions(+), 1 deletion(-) (limited to 'drivers/net/ethernet/mellanox/mlx4') diff --git a/drivers/net/ethernet/mellanox/mlx4/en_cq.c b/drivers/net/ethernet/mellanox/mlx4/en_cq.c index 132cea655920..e3be7e44ff51 100644 --- a/drivers/net/ethernet/mellanox/mlx4/en_cq.c +++ b/drivers/net/ethernet/mellanox/mlx4/en_cq.c @@ -127,7 +127,15 @@ int mlx4_en_activate_cq(struct mlx4_en_priv *priv, struct mlx4_en_cq *cq, /* For TX we use the same irq per ring we assigned for the RX */ struct mlx4_en_cq *rx_cq; - + int xdp_index; + + /* The xdp tx irq must align with the rx ring that forwards to + * it, so reindex these from 0. This should only happen when + * tx_ring_num is not a multiple of rx_ring_num. + */ + xdp_index = (priv->xdp_ring_num - priv->tx_ring_num) + cq_idx; + if (xdp_index >= 0) + cq_idx = xdp_index; cq_idx = cq_idx % priv->rx_ring_num; rx_cq = priv->rx_cq[cq_idx]; cq->vector = rx_cq->vector; -- cgit v1.2.3 From a4256bc9ec36159e84d710c7c44aaa244a6380a8 Mon Sep 17 00:00:00 2001 From: Arnd Bergmann Date: Tue, 25 Oct 2016 18:16:20 +0200 Subject: IB/mlx4: avoid a -Wmaybe-uninitialize warning MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit 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: Arnd Bergmann Reviewed-by: Yishai Hadas Signed-off-by: David S. Miller --- drivers/net/ethernet/mellanox/mlx4/resource_tracker.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx4') diff --git a/drivers/net/ethernet/mellanox/mlx4/resource_tracker.c b/drivers/net/ethernet/mellanox/mlx4/resource_tracker.c index 84d7857ccc27..c548beaaf910 100644 --- a/drivers/net/ethernet/mellanox/mlx4/resource_tracker.c +++ b/drivers/net/ethernet/mellanox/mlx4/resource_tracker.c @@ -1605,13 +1605,14 @@ static int eq_res_start_move_to(struct mlx4_dev *dev, int slave, int index, r->com.from_state = r->com.state; r->com.to_state = state; r->com.state = RES_EQ_BUSY; - if (eq) - *eq = r; } } spin_unlock_irq(mlx4_tlock(dev)); + if (!err && eq) + *eq = r; + return err; } -- cgit v1.2.3 From aa0c08feae8161b945520ada753d0dfe62b14fe7 Mon Sep 17 00:00:00 2001 From: Jack Morgenstein Date: Thu, 27 Oct 2016 16:27:13 +0300 Subject: net/mlx4_core: Fix the resource-type enum in res tracker to conform to FW spec 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: 623ed84b1f95 ("mlx4_core: initial header-file changes for SRIOV support") Signed-off-by: Jack Morgenstein Signed-off-by: Moshe Shemesh Signed-off-by: Tariq Toukan Signed-off-by: David S. Miller --- drivers/net/ethernet/mellanox/mlx4/mlx4.h | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'drivers/net/ethernet/mellanox/mlx4') diff --git a/drivers/net/ethernet/mellanox/mlx4/mlx4.h b/drivers/net/ethernet/mellanox/mlx4/mlx4.h index e4878f31e45d..7aad07644289 100644 --- a/drivers/net/ethernet/mellanox/mlx4/mlx4.h +++ b/drivers/net/ethernet/mellanox/mlx4/mlx4.h @@ -145,9 +145,10 @@ enum mlx4_resource { RES_MTT, RES_MAC, RES_VLAN, - RES_EQ, + RES_NPORT_ID, RES_COUNTER, RES_FS_RULE, + RES_EQ, MLX4_NUM_OF_RESOURCE_TYPE }; -- cgit v1.2.3 From 33a1f8b196dca933313c001866c4df3f3ca11f78 Mon Sep 17 00:00:00 2001 From: Maor Gottlieb Date: Thu, 27 Oct 2016 16:27:14 +0300 Subject: net/mlx4_core: Avoid setting ports to auto when only one port type is supported When only one port type is supported, it should be read only. We reject changing requests, even to the auto sense mode. Fixes: 27bf91d6a0d5 ("mlx4_core: Add link type autosensing") Signed-off-by: Maor Gottlieb Signed-off-by: Moshe Shemesh Signed-off-by: Tariq Toukan Signed-off-by: David S. Miller --- drivers/net/ethernet/mellanox/mlx4/main.c | 10 +++++++++- 1 file changed, 9 insertions(+), 1 deletion(-) (limited to 'drivers/net/ethernet/mellanox/mlx4') diff --git a/drivers/net/ethernet/mellanox/mlx4/main.c b/drivers/net/ethernet/mellanox/mlx4/main.c index 7183ac4135d2..6f4e67bc3538 100644 --- a/drivers/net/ethernet/mellanox/mlx4/main.c +++ b/drivers/net/ethernet/mellanox/mlx4/main.c @@ -1102,6 +1102,14 @@ static int __set_port_type(struct mlx4_port_info *info, int i; int err = 0; + if ((port_type & mdev->caps.supported_type[info->port]) != port_type) { + mlx4_err(mdev, + "Requested port type for port %d is not supported on this HCA\n", + info->port); + err = -EINVAL; + goto err_sup; + } + mlx4_stop_sense(mdev); mutex_lock(&priv->port_mutex); info->tmp_type = port_type; @@ -1147,7 +1155,7 @@ static int __set_port_type(struct mlx4_port_info *info, out: mlx4_start_sense(mdev); mutex_unlock(&priv->port_mutex); - +err_sup: return err; } -- cgit v1.2.3 From 72da2e911f79d4c7132d7431a97d46659ee862be Mon Sep 17 00:00:00 2001 From: Moshe Lazer Date: Thu, 27 Oct 2016 16:27:15 +0300 Subject: net/mlx4_core: Change the default value of enable_qos Change the default status of quality of service back to disabled, as it hurts performance in some cases. Fixes: 38438f7c7e8c ("net/mlx4: Set enhanced QoS support by default when ...") Signed-off-by: Moshe Lazer Signed-off-by: Tariq Toukan Signed-off-by: David S. Miller --- drivers/net/ethernet/mellanox/mlx4/fw.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx4') diff --git a/drivers/net/ethernet/mellanox/mlx4/fw.c b/drivers/net/ethernet/mellanox/mlx4/fw.c index c41ab31a39f8..84bab9f0732e 100644 --- a/drivers/net/ethernet/mellanox/mlx4/fw.c +++ b/drivers/net/ethernet/mellanox/mlx4/fw.c @@ -49,9 +49,9 @@ enum { extern void __buggy_use_of_MLX4_GET(void); extern void __buggy_use_of_MLX4_PUT(void); -static bool enable_qos = true; +static bool enable_qos; module_param(enable_qos, bool, 0444); -MODULE_PARM_DESC(enable_qos, "Enable Enhanced QoS support (default: on)"); +MODULE_PARM_DESC(enable_qos, "Enable Enhanced QoS support (default: off)"); #define MLX4_GET(dest, source, offset) \ do { \ -- cgit v1.2.3 From 4850cf4581578216468b7b3c3d06cc5abb0a697d Mon Sep 17 00:00:00 2001 From: Eugenia Emantayev Date: Thu, 27 Oct 2016 16:27:16 +0300 Subject: net/mlx4_en: Resolve dividing by zero in 32-bit system 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: 31c128b66e5b ("net/mlx4_en: Choose time-stamping shift value according to HW frequency") Signed-off-by: Eugenia Emantayev Signed-off-by: Tariq Toukan Signed-off-by: David S. Miller --- drivers/net/ethernet/mellanox/mlx4/en_clock.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) (limited to 'drivers/net/ethernet/mellanox/mlx4') diff --git a/drivers/net/ethernet/mellanox/mlx4/en_clock.c b/drivers/net/ethernet/mellanox/mlx4/en_clock.c index 08fc5fc56d43..a5fc46bbcbe2 100644 --- a/drivers/net/ethernet/mellanox/mlx4/en_clock.c +++ b/drivers/net/ethernet/mellanox/mlx4/en_clock.c @@ -245,8 +245,11 @@ static u32 freq_to_shift(u16 freq) { u32 freq_khz = freq * 1000; u64 max_val_cycles = freq_khz * 1000 * MLX4_EN_WRAP_AROUND_SEC; + u64 tmp_rounded = + roundup_pow_of_two(max_val_cycles) > max_val_cycles ? + roundup_pow_of_two(max_val_cycles) - 1 : UINT_MAX; u64 max_val_cycles_rounded = is_power_of_2(max_val_cycles + 1) ? - max_val_cycles : roundup_pow_of_two(max_val_cycles) - 1; + max_val_cycles : tmp_rounded; /* calculate max possible multiplier in order to fit in 64bit */ u64 max_mul = div_u64(0xffffffffffffffffULL, max_val_cycles_rounded); -- cgit v1.2.3 From 8d59de8f7bb3db296331c665779c653b0c8d13ba Mon Sep 17 00:00:00 2001 From: Erez Shitrit Date: Thu, 27 Oct 2016 16:27:17 +0300 Subject: net/mlx4_en: Process all completions in RX rings after port goes up 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: c27a02cd94d6 ("mlx4_en: Add driver for Mellanox ConnectX 10GbE NIC") Signed-off-by: Erez Shitrit Signed-off-by: Eugenia Emantayev Signed-off-by: Tariq Toukan Signed-off-by: David S. Miller --- drivers/net/ethernet/mellanox/mlx4/en_netdev.c | 7 +++++++ 1 file changed, 7 insertions(+) (limited to 'drivers/net/ethernet/mellanox/mlx4') diff --git a/drivers/net/ethernet/mellanox/mlx4/en_netdev.c b/drivers/net/ethernet/mellanox/mlx4/en_netdev.c index 7e703bed7b82..e25c11dff525 100644 --- a/drivers/net/ethernet/mellanox/mlx4/en_netdev.c +++ b/drivers/net/ethernet/mellanox/mlx4/en_netdev.c @@ -1733,6 +1733,13 @@ int mlx4_en_start_port(struct net_device *dev) udp_tunnel_get_rx_info(dev); priv->port_up = true; + + /* Process all completions if exist to prevent + * the queues freezing if they are full + */ + for (i = 0; i < priv->rx_ring_num; i++) + napi_schedule(&priv->rx_cq[i]->napi); + netif_tx_start_all_queues(dev); netif_device_attach(dev); -- cgit v1.2.3 From 9d2afba058722d40cc02f430229c91611c0e8d16 Mon Sep 17 00:00:00 2001 From: Eugenia Emantayev Date: Thu, 27 Oct 2016 16:27:18 +0300 Subject: net/mlx4_en: Fix panic during reboot 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: c27a02cd94d6 ("mlx4_en: Add driver for Mellanox ConnectX 10GbE NIC") Signed-off-by: Eugenia Emantayev Signed-off-by: Tariq Toukan Signed-off-by: David S. Miller --- drivers/net/ethernet/mellanox/mlx4/en_netdev.c | 1 + 1 file changed, 1 insertion(+) (limited to 'drivers/net/ethernet/mellanox/mlx4') diff --git a/drivers/net/ethernet/mellanox/mlx4/en_netdev.c b/drivers/net/ethernet/mellanox/mlx4/en_netdev.c index e25c11dff525..314f54c8dbed 100644 --- a/drivers/net/ethernet/mellanox/mlx4/en_netdev.c +++ b/drivers/net/ethernet/mellanox/mlx4/en_netdev.c @@ -2201,6 +2201,7 @@ void mlx4_en_destroy_netdev(struct net_device *dev) if (!shutdown) free_netdev(dev); + dev->ethtool_ops = NULL; } static int mlx4_en_change_mtu(struct net_device *dev, int new_mtu) -- cgit v1.2.3 From 81d184199e328fdad5633da139a10337327154e0 Mon Sep 17 00:00:00 2001 From: Jack Morgenstein Date: Thu, 27 Oct 2016 16:27:19 +0300 Subject: net/mlx4_core: Do not access comm channel if it has not yet been initialized 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: 55ad359225b2 ("net/mlx4_core: Enable device recovery flow with SRIOV") Fixes: ab9c17a009ee ("mlx4_core: Modify driver initialization flow to accommodate SRIOV for Ethernet") Signed-off-by: Jack Morgenstein Signed-off-by: Tariq Toukan Signed-off-by: David S. Miller --- drivers/net/ethernet/mellanox/mlx4/cmd.c | 9 +++++++++ 1 file changed, 9 insertions(+) (limited to 'drivers/net/ethernet/mellanox/mlx4') diff --git a/drivers/net/ethernet/mellanox/mlx4/cmd.c b/drivers/net/ethernet/mellanox/mlx4/cmd.c index b1cef7a0f7ca..e36bebcab3f2 100644 --- a/drivers/net/ethernet/mellanox/mlx4/cmd.c +++ b/drivers/net/ethernet/mellanox/mlx4/cmd.c @@ -2469,6 +2469,7 @@ err_comm_admin: kfree(priv->mfunc.master.slave_state); err_comm: iounmap(priv->mfunc.comm); + priv->mfunc.comm = NULL; err_vhcr: dma_free_coherent(&dev->persist->pdev->dev, PAGE_SIZE, priv->mfunc.vhcr, @@ -2537,6 +2538,13 @@ void mlx4_report_internal_err_comm_event(struct mlx4_dev *dev) int slave; u32 slave_read; + /* If the comm channel has not yet been initialized, + * skip reporting the internal error event to all + * the communication channels. + */ + if (!priv->mfunc.comm) + return; + /* Report an internal error event to all * communication channels. */ @@ -2571,6 +2579,7 @@ void mlx4_multi_func_cleanup(struct mlx4_dev *dev) } iounmap(priv->mfunc.comm); + priv->mfunc.comm = NULL; } void mlx4_cmd_cleanup(struct mlx4_dev *dev, int cleanup_mask) -- cgit v1.2.3 From 6f2e0d2c3bf0f8d322ab7516c57340c7189cca02 Mon Sep 17 00:00:00 2001 From: Eugenia Emantayev Date: Thu, 27 Oct 2016 16:27:20 +0300 Subject: net/mlx4: Fix firmware command timeout during interrupt test 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: Eugenia Emantayev Signed-off-by: Tariq Toukan Signed-off-by: David S. Miller --- drivers/net/ethernet/mellanox/mlx4/en_selftest.c | 26 +++++++++- drivers/net/ethernet/mellanox/mlx4/eq.c | 62 +++++++++++------------- include/linux/mlx4/device.h | 3 +- 3 files changed, 55 insertions(+), 36 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx4') diff --git a/drivers/net/ethernet/mellanox/mlx4/en_selftest.c b/drivers/net/ethernet/mellanox/mlx4/en_selftest.c index b66e03d9711f..c06346a82496 100644 --- a/drivers/net/ethernet/mellanox/mlx4/en_selftest.c +++ b/drivers/net/ethernet/mellanox/mlx4/en_selftest.c @@ -118,6 +118,29 @@ mlx4_en_test_loopback_exit: return !loopback_ok; } +static int mlx4_en_test_interrupts(struct mlx4_en_priv *priv) +{ + struct mlx4_en_dev *mdev = priv->mdev; + int err = 0; + int i = 0; + + err = mlx4_test_async(mdev->dev); + /* When not in MSI_X or slave, test only async */ + if (!(mdev->dev->flags & MLX4_FLAG_MSI_X) || mlx4_is_slave(mdev->dev)) + return err; + + /* A loop over all completion vectors of current port, + * for each vector check whether it works by mapping command + * completions to that vector and performing a NOP command + */ + for (i = 0; i < priv->rx_ring_num; i++) { + err = mlx4_test_interrupt(mdev->dev, priv->rx_cq[i]->vector); + if (err) + break; + } + + return err; +} static int mlx4_en_test_link(struct mlx4_en_priv *priv) { @@ -151,7 +174,6 @@ static int mlx4_en_test_speed(struct mlx4_en_priv *priv) void mlx4_en_ex_selftest(struct net_device *dev, u32 *flags, u64 *buf) { struct mlx4_en_priv *priv = netdev_priv(dev); - struct mlx4_en_dev *mdev = priv->mdev; int i, carrier_ok; memset(buf, 0, sizeof(u64) * MLX4_EN_NUM_SELF_TEST); @@ -177,7 +199,7 @@ void mlx4_en_ex_selftest(struct net_device *dev, u32 *flags, u64 *buf) netif_carrier_on(dev); } - buf[0] = mlx4_test_interrupts(mdev->dev); + buf[0] = mlx4_en_test_interrupts(priv); buf[1] = mlx4_en_test_link(priv); buf[2] = mlx4_en_test_speed(priv); diff --git a/drivers/net/ethernet/mellanox/mlx4/eq.c b/drivers/net/ethernet/mellanox/mlx4/eq.c index cf8f8a72a801..cd3638e6fe25 100644 --- a/drivers/net/ethernet/mellanox/mlx4/eq.c +++ b/drivers/net/ethernet/mellanox/mlx4/eq.c @@ -1361,53 +1361,49 @@ void mlx4_cleanup_eq_table(struct mlx4_dev *dev) kfree(priv->eq_table.uar_map); } -/* A test that verifies that we can accept interrupts on all - * the irq vectors of the device. +/* A test that verifies that we can accept interrupts + * on the vector allocated for asynchronous events + */ +int mlx4_test_async(struct mlx4_dev *dev) +{ + return mlx4_NOP(dev); +} +EXPORT_SYMBOL(mlx4_test_async); + +/* A test that verifies that we can accept interrupts + * on the given irq vector of the tested port. * Interrupts are checked using the NOP command. */ -int mlx4_test_interrupts(struct mlx4_dev *dev) +int mlx4_test_interrupt(struct mlx4_dev *dev, int vector) { struct mlx4_priv *priv = mlx4_priv(dev); - int i; int err; - err = mlx4_NOP(dev); - /* When not in MSI_X, there is only one irq to check */ - if (!(dev->flags & MLX4_FLAG_MSI_X) || mlx4_is_slave(dev)) - return err; - - /* A loop over all completion vectors, for each vector we will check - * whether it works by mapping command completions to that vector - * and performing a NOP command - */ - for(i = 0; !err && (i < dev->caps.num_comp_vectors); ++i) { - /* Make sure request_irq was called */ - if (!priv->eq_table.eq[i].have_irq) - continue; - - /* Temporary use polling for command completions */ - mlx4_cmd_use_polling(dev); - - /* Map the new eq to handle all asynchronous events */ - err = mlx4_MAP_EQ(dev, get_async_ev_mask(dev), 0, - priv->eq_table.eq[i].eqn); - if (err) { - mlx4_warn(dev, "Failed mapping eq for interrupt test\n"); - mlx4_cmd_use_events(dev); - break; - } + /* Temporary use polling for command completions */ + mlx4_cmd_use_polling(dev); - /* Go back to using events */ - mlx4_cmd_use_events(dev); - err = mlx4_NOP(dev); + /* Map the new eq to handle all asynchronous events */ + err = mlx4_MAP_EQ(dev, get_async_ev_mask(dev), 0, + priv->eq_table.eq[MLX4_CQ_TO_EQ_VECTOR(vector)].eqn); + if (err) { + mlx4_warn(dev, "Failed mapping eq for interrupt test\n"); + goto out; } + /* Go back to using events */ + mlx4_cmd_use_events(dev); + err = mlx4_NOP(dev); + /* Return to default */ + mlx4_cmd_use_polling(dev); +out: mlx4_MAP_EQ(dev, get_async_ev_mask(dev), 0, priv->eq_table.eq[MLX4_EQ_ASYNC].eqn); + mlx4_cmd_use_events(dev); + return err; } -EXPORT_SYMBOL(mlx4_test_interrupts); +EXPORT_SYMBOL(mlx4_test_interrupt); bool mlx4_is_eq_vector_valid(struct mlx4_dev *dev, u8 port, int vector) { diff --git a/include/linux/mlx4/device.h b/include/linux/mlx4/device.h index f6a164297358..3be7abd6e722 100644 --- a/include/linux/mlx4/device.h +++ b/include/linux/mlx4/device.h @@ -1399,7 +1399,8 @@ void mlx4_fmr_unmap(struct mlx4_dev *dev, struct mlx4_fmr *fmr, u32 *lkey, u32 *rkey); int mlx4_fmr_free(struct mlx4_dev *dev, struct mlx4_fmr *fmr); int mlx4_SYNC_TPT(struct mlx4_dev *dev); -int mlx4_test_interrupts(struct mlx4_dev *dev); +int mlx4_test_interrupt(struct mlx4_dev *dev, int vector); +int mlx4_test_async(struct mlx4_dev *dev); int mlx4_query_diag_counters(struct mlx4_dev *dev, u8 op_modifier, const u32 offset[], u32 value[], size_t array_len, u8 port); -- cgit v1.2.3 From d2582a03939ed0a80ffcd3ea5345505bc8067c54 Mon Sep 17 00:00:00 2001 From: Jack Morgenstein Date: Thu, 27 Oct 2016 16:27:21 +0300 Subject: net/mlx4_en: Fix potential deadlock in port statistics flow 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: f9baff509f8a ("mlx4_core: Add "native" argument to mlx4_cmd ...") Signed-off-by: Jack Morgenstein Signed-off-by: Tariq Toukan Signed-off-by: David S. Miller --- drivers/net/ethernet/mellanox/mlx4/en_port.c | 4 ++-- drivers/net/ethernet/mellanox/mlx4/mlx4.h | 2 -- drivers/net/ethernet/mellanox/mlx4/port.c | 13 +------------ 3 files changed, 3 insertions(+), 16 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx4') diff --git a/drivers/net/ethernet/mellanox/mlx4/en_port.c b/drivers/net/ethernet/mellanox/mlx4/en_port.c index 5aa8b751f417..59473a0ebcdf 100644 --- a/drivers/net/ethernet/mellanox/mlx4/en_port.c +++ b/drivers/net/ethernet/mellanox/mlx4/en_port.c @@ -166,7 +166,7 @@ int mlx4_en_DUMP_ETH_STATS(struct mlx4_en_dev *mdev, u8 port, u8 reset) return PTR_ERR(mailbox); err = mlx4_cmd_box(mdev->dev, 0, mailbox->dma, in_mod, 0, MLX4_CMD_DUMP_ETH_STATS, MLX4_CMD_TIME_CLASS_B, - MLX4_CMD_WRAPPED); + MLX4_CMD_NATIVE); if (err) goto out; @@ -322,7 +322,7 @@ int mlx4_en_DUMP_ETH_STATS(struct mlx4_en_dev *mdev, u8 port, u8 reset) err = mlx4_cmd_box(mdev->dev, 0, mailbox->dma, in_mod | MLX4_DUMP_ETH_STATS_FLOW_CONTROL, 0, MLX4_CMD_DUMP_ETH_STATS, - MLX4_CMD_TIME_CLASS_B, MLX4_CMD_WRAPPED); + MLX4_CMD_TIME_CLASS_B, MLX4_CMD_NATIVE); if (err) goto out; } diff --git a/drivers/net/ethernet/mellanox/mlx4/mlx4.h b/drivers/net/ethernet/mellanox/mlx4/mlx4.h index 7aad07644289..88ee7d8a5923 100644 --- a/drivers/net/ethernet/mellanox/mlx4/mlx4.h +++ b/drivers/net/ethernet/mellanox/mlx4/mlx4.h @@ -1330,8 +1330,6 @@ int mlx4_SET_VLAN_FLTR_wrapper(struct mlx4_dev *dev, int slave, struct mlx4_cmd_info *cmd); int mlx4_common_set_vlan_fltr(struct mlx4_dev *dev, int function, int port, void *buf); -int mlx4_common_dump_eth_stats(struct mlx4_dev *dev, int slave, u32 in_mod, - struct mlx4_cmd_mailbox *outbox); int mlx4_DUMP_ETH_STATS_wrapper(struct mlx4_dev *dev, int slave, struct mlx4_vhcr *vhcr, struct mlx4_cmd_mailbox *inbox, diff --git a/drivers/net/ethernet/mellanox/mlx4/port.c b/drivers/net/ethernet/mellanox/mlx4/port.c index c5b2064297a1..b656dd5772e5 100644 --- a/drivers/net/ethernet/mellanox/mlx4/port.c +++ b/drivers/net/ethernet/mellanox/mlx4/port.c @@ -1728,24 +1728,13 @@ int mlx4_SET_VLAN_FLTR_wrapper(struct mlx4_dev *dev, int slave, return err; } -int mlx4_common_dump_eth_stats(struct mlx4_dev *dev, int slave, - u32 in_mod, struct mlx4_cmd_mailbox *outbox) -{ - return mlx4_cmd_box(dev, 0, outbox->dma, in_mod, 0, - MLX4_CMD_DUMP_ETH_STATS, MLX4_CMD_TIME_CLASS_B, - MLX4_CMD_NATIVE); -} - int mlx4_DUMP_ETH_STATS_wrapper(struct mlx4_dev *dev, int slave, struct mlx4_vhcr *vhcr, struct mlx4_cmd_mailbox *inbox, struct mlx4_cmd_mailbox *outbox, struct mlx4_cmd_info *cmd) { - if (slave != dev->caps.function) - return 0; - return mlx4_common_dump_eth_stats(dev, slave, - vhcr->in_modifier, outbox); + return 0; } int mlx4_get_slave_from_roce_gid(struct mlx4_dev *dev, int port, u8 *gid, -- cgit v1.2.3 From eb4b678825992aa434d32b2f615d2090281e0f88 Mon Sep 17 00:00:00 2001 From: Tariq Toukan Date: Thu, 27 Oct 2016 16:27:22 +0300 Subject: net/mlx4_en: Save slave ethtool stats command 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: Tariq Toukan Signed-off-by: David S. Miller --- drivers/net/ethernet/mellanox/mlx4/en_netdev.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) (limited to 'drivers/net/ethernet/mellanox/mlx4') diff --git a/drivers/net/ethernet/mellanox/mlx4/en_netdev.c b/drivers/net/ethernet/mellanox/mlx4/en_netdev.c index 314f54c8dbed..12c99a2655f2 100644 --- a/drivers/net/ethernet/mellanox/mlx4/en_netdev.c +++ b/drivers/net/ethernet/mellanox/mlx4/en_netdev.c @@ -1917,8 +1917,9 @@ static void mlx4_en_clear_stats(struct net_device *dev) struct mlx4_en_dev *mdev = priv->mdev; int i; - if (mlx4_en_DUMP_ETH_STATS(mdev, priv->port, 1)) - en_dbg(HW, priv, "Failed dumping statistics\n"); + if (!mlx4_is_slave(mdev->dev)) + if (mlx4_en_DUMP_ETH_STATS(mdev, priv->port, 1)) + en_dbg(HW, priv, "Failed dumping statistics\n"); memset(&priv->pstats, 0, sizeof(priv->pstats)); memset(&priv->pkstats, 0, sizeof(priv->pkstats)); -- cgit v1.2.3