From 5a5e78cd706ce0577ed6356634a34552c866cc10 Mon Sep 17 00:00:00 2001 From: Caesar Wang Date: Wed, 18 May 2016 23:01:39 +0800 Subject: thermal: add the note for set_trip_temp Fixes commit 60f9ce3ada53 ("thermal: of-thermal: allow setting trip_temp on hardware") Signed-off-by: Caesar Wang Cc: Zhang Rui Cc: Eduardo Valentin Cc: linux-pm@vger.kernel.org Signed-off-by: Zhang Rui --- include/linux/thermal.h | 2 ++ 1 file changed, 2 insertions(+) (limited to 'include/linux') diff --git a/include/linux/thermal.h b/include/linux/thermal.h index e45abe7db9a6..ee517bef0db0 100644 --- a/include/linux/thermal.h +++ b/include/linux/thermal.h @@ -335,6 +335,8 @@ struct thermal_genl_event { * @get_trend: a pointer to a function that reads the sensor temperature trend. * @set_emul_temp: a pointer to a function that sets sensor emulated * temperature. + * @set_trip_temp: a pointer to a function that sets the trip temperature on + * hardware. */ struct thermal_zone_of_device_ops { int (*get_temp)(void *, int *); -- cgit v1.2.3 From 9bd616e3dbedfc103f158197c8ad93678849b1ed Mon Sep 17 00:00:00 2001 From: Catalin Marinas Date: Wed, 1 Jun 2016 18:52:16 +0100 Subject: cpuidle: Do not access cpuidle_devices when !CONFIG_CPU_IDLE The cpuidle_devices per-CPU variable is only defined when CPU_IDLE is enabled. Commit c8cc7d4de7a4 ("sched/idle: Reorganize the idle loop") removed the #ifdef CONFIG_CPU_IDLE around cpuidle_idle_call() with the compiler optimising away __this_cpu_read(cpuidle_devices). However, with CONFIG_UBSAN && !CONFIG_CPU_IDLE, this optimisation no longer happens and the kernel fails to link since cpuidle_devices is not defined. This patch introduces an accessor function for the current CPU cpuidle device (returning NULL when !CONFIG_CPU_IDLE) and uses it in cpuidle_idle_call(). Signed-off-by: Catalin Marinas Cc: 4.5+ # 4.5+ Signed-off-by: Rafael J. Wysocki --- include/linux/cpuidle.h | 3 +++ kernel/sched/idle.c | 2 +- 2 files changed, 4 insertions(+), 1 deletion(-) (limited to 'include/linux') diff --git a/include/linux/cpuidle.h b/include/linux/cpuidle.h index 786ad32631a6..07b83d32f66c 100644 --- a/include/linux/cpuidle.h +++ b/include/linux/cpuidle.h @@ -152,6 +152,8 @@ extern void cpuidle_disable_device(struct cpuidle_device *dev); extern int cpuidle_play_dead(void); extern struct cpuidle_driver *cpuidle_get_cpu_driver(struct cpuidle_device *dev); +static inline struct cpuidle_device *cpuidle_get_device(void) +{return __this_cpu_read(cpuidle_devices); } #else static inline void disable_cpuidle(void) { } static inline bool cpuidle_not_available(struct cpuidle_driver *drv, @@ -187,6 +189,7 @@ static inline void cpuidle_disable_device(struct cpuidle_device *dev) { } static inline int cpuidle_play_dead(void) {return -ENODEV; } static inline struct cpuidle_driver *cpuidle_get_cpu_driver( struct cpuidle_device *dev) {return NULL; } +static inline struct cpuidle_device *cpuidle_get_device(void) {return NULL; } #endif #if defined(CONFIG_CPU_IDLE) && defined(CONFIG_SUSPEND) diff --git a/kernel/sched/idle.c b/kernel/sched/idle.c index bd12c6c714ec..c5aeedf4e93a 100644 --- a/kernel/sched/idle.c +++ b/kernel/sched/idle.c @@ -127,7 +127,7 @@ static int call_cpuidle(struct cpuidle_driver *drv, struct cpuidle_device *dev, */ static void cpuidle_idle_call(void) { - struct cpuidle_device *dev = __this_cpu_read(cpuidle_devices); + struct cpuidle_device *dev = cpuidle_get_device(); struct cpuidle_driver *drv = cpuidle_get_cpu_driver(dev); int next_state, entered_state; -- cgit v1.2.3 From 55eed755c6e30a89be3a791a6b0ad208aadd9bdc Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Fri, 27 May 2016 13:11:17 +0200 Subject: locking/seqcount: Re-fix raw_read_seqcount_latch() Commit 50755bc1c305 ("seqlock: fix raw_read_seqcount_latch()") broke raw_read_seqcount_latch(). If you look at the comment that was modified; the thing that changes is the seq count, not the latch pointer. * void latch_modify(struct latch_struct *latch, ...) * { * smp_wmb(); <- Ensure that the last data[1] update is visible * latch->seq++; * smp_wmb(); <- Ensure that the seqcount update is visible * * modify(latch->data[0], ...); * * smp_wmb(); <- Ensure that the data[0] update is visible * latch->seq++; * smp_wmb(); <- Ensure that the seqcount update is visible * * modify(latch->data[1], ...); * } * * The query will have a form like: * * struct entry *latch_query(struct latch_struct *latch, ...) * { * struct entry *entry; * unsigned seq, idx; * * do { * seq = lockless_dereference(latch->seq); So here we have: seq = READ_ONCE(latch->seq); smp_read_barrier_depends(); Which is exactly what we want; the new code: seq = ({ p = READ_ONCE(latch); smp_read_barrier_depends(); p })->seq; is just wrong; because it looses the volatile read on seq, which can now be torn or worse 'optimized'. And the read_depend barrier is also placed wrong, we want it after the load of seq, to match the above data[] up-to-date wmb()s. Such that when we dereference latch->data[] below, we're guaranteed to observe the right data. * * idx = seq & 0x01; * entry = data_query(latch->data[idx], ...); * * smp_rmb(); * } while (seq != latch->seq); * * return entry; * } So yes, not passing a pointer is not pretty, but the code was correct, and isn't anymore now. Change to explicit READ_ONCE()+smp_read_barrier_depends() to avoid confusion and allow strict lockless_dereference() checking. Signed-off-by: Peter Zijlstra (Intel) Cc: Alexey Dobriyan Cc: Andrew Morton Cc: Linus Torvalds Cc: Paul E. McKenney Cc: Paul McKenney Cc: Peter Zijlstra Cc: Thomas Gleixner Fixes: 50755bc1c305 ("seqlock: fix raw_read_seqcount_latch()") Link: http://lkml.kernel.org/r/20160527111117.GL3192@twins.programming.kicks-ass.net Signed-off-by: Ingo Molnar --- include/linux/seqlock.h | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-) (limited to 'include/linux') diff --git a/include/linux/seqlock.h b/include/linux/seqlock.h index 7973a821ac58..ead97654c4e9 100644 --- a/include/linux/seqlock.h +++ b/include/linux/seqlock.h @@ -277,7 +277,10 @@ static inline void raw_write_seqcount_barrier(seqcount_t *s) static inline int raw_read_seqcount_latch(seqcount_t *s) { - return lockless_dereference(s)->sequence; + int seq = READ_ONCE(s->sequence); + /* Pairs with the first smp_wmb() in raw_write_seqcount_latch() */ + smp_read_barrier_depends(); + return seq; } /** @@ -331,7 +334,7 @@ static inline int raw_read_seqcount_latch(seqcount_t *s) * unsigned seq, idx; * * do { - * seq = lockless_dereference(latch)->seq; + * seq = raw_read_seqcount_latch(&latch->seq); * * idx = seq & 0x01; * entry = data_query(latch->data[idx], ...); -- cgit v1.2.3 From 55f1ea15216a5a14c96738bd5284100a00ffa9dc Mon Sep 17 00:00:00 2001 From: Vitaly Kuznetsov Date: Tue, 31 May 2016 11:23:43 +0100 Subject: efi: Fix for_each_efi_memory_desc_in_map() for empty memmaps Commit: 78ce248faa3c ("efi: Iterate over efi.memmap in for_each_efi_memory_desc()") introduced a regression for systems booted with the 'noefi' kernel option. In particular, I observed an early kernel hang in efi_find_mirror()'s for_each_efi_memory_desc() call. As we don't have efi memmap on this system we enter this iterator with the following parameters: efi.memmap.map = 0, efi.memmap.map_end = 0, efi.memmap.desc_size = 28 ... then for_each_efi_memory_desc_in_map() does the following comparison: (md) <= (efi_memory_desc_t *)((m)->map_end - (m)->desc_size); ... where md = 0, (m)->map_end = 0 and (m)->desc_size = 28 but when we subtract something from a NULL pointer wrap around happens and we end up returning invalid pointer and crash. Fix it by using the correct pointer arithmetics. Signed-off-by: Vitaly Kuznetsov Signed-off-by: Matt Fleming Cc: Ard Biesheuvel Cc: K. Y. Srinivasan Cc: Linus Torvalds Cc: Mark Salter Cc: Peter Zijlstra Cc: Thomas Gleixner Cc: linux-efi@vger.kernel.org Fixes: 78ce248faa3c ("efi: Iterate over efi.memmap in for_each_efi_memory_desc()") Link: http://lkml.kernel.org/r/1464690224-4503-2-git-send-email-matt@codeblueprint.co.uk [ Made the changelog more readable. ] Signed-off-by: Ingo Molnar --- include/linux/efi.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'include/linux') diff --git a/include/linux/efi.h b/include/linux/efi.h index c2db3ca22217..f196dd0b0f2f 100644 --- a/include/linux/efi.h +++ b/include/linux/efi.h @@ -1005,7 +1005,7 @@ extern int efi_memattr_apply_permissions(struct mm_struct *mm, /* Iterate through an efi_memory_map */ #define for_each_efi_memory_desc_in_map(m, md) \ for ((md) = (m)->map; \ - (md) <= (efi_memory_desc_t *)((m)->map_end - (m)->desc_size); \ + ((void *)(md) + (m)->desc_size) <= (m)->map_end; \ (md) = (void *)(md) + (m)->desc_size) /** -- cgit v1.2.3 From e93aeeae0bf8cba43e05484ae4ad17213ac6c3a7 Mon Sep 17 00:00:00 2001 From: Kuninori Morimoto Date: Wed, 25 May 2016 01:15:04 +0000 Subject: of: add missing const for of_parse_phandle_with_args() in !CONFIG_OF commit 93c667ca2598bd84f1bd3f2fa176af69707699fe ("of: *node argument to of_parse_phandle_with_args should be const") changed to const for struct device node *np, but it cares CONFIG_OF case only, !CONFIG_OF case need it too. Signed-off-by: Kuninori Morimoto Signed-off-by: Rob Herring --- include/linux/of.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'include/linux') diff --git a/include/linux/of.h b/include/linux/of.h index c7292e8ea080..74eb28cadbef 100644 --- a/include/linux/of.h +++ b/include/linux/of.h @@ -614,7 +614,7 @@ static inline struct device_node *of_parse_phandle(const struct device_node *np, return NULL; } -static inline int of_parse_phandle_with_args(struct device_node *np, +static inline int of_parse_phandle_with_args(const struct device_node *np, const char *list_name, const char *cells_name, int index, -- cgit v1.2.3 From d3ae2bdeba9bad8cb95301451aeaf03ce31e82f0 Mon Sep 17 00:00:00 2001 From: Noa Osherovich Date: Sat, 4 Jun 2016 15:15:36 +0300 Subject: IB/mlx5: Fix pkey_index length in the QP path record Pkey index fields in the QP context path record are extended to 16 bits, as required by IB spec (version 1.3). This change affects all QP commands which include path records. To enable this change, moved the free adaptive routing flag bit (free_ar) to the most significant byte of the QP path record. Fixes: e126ba97dba9e ('mlx5: Add driver for Mellanox Connect-IB ...') Signed-off-by: Noa Osherovich Reviewed-by: Jack Morgenstein Signed-off-by: Leon Romanovsky Signed-off-by: Doug Ledford --- drivers/infiniband/hw/mlx5/qp.c | 15 ++++++++------- include/linux/mlx5/qp.h | 5 ++--- 2 files changed, 10 insertions(+), 10 deletions(-) (limited to 'include/linux') diff --git a/drivers/infiniband/hw/mlx5/qp.c b/drivers/infiniband/hw/mlx5/qp.c index 43c1441b6fb8..6b90bfdea830 100644 --- a/drivers/infiniband/hw/mlx5/qp.c +++ b/drivers/infiniband/hw/mlx5/qp.c @@ -1859,7 +1859,7 @@ static int mlx5_set_path(struct mlx5_ib_dev *dev, struct mlx5_ib_qp *qp, int err; if (attr_mask & IB_QP_PKEY_INDEX) - path->pkey_index = attr->pkey_index; + path->pkey_index = cpu_to_be16(attr->pkey_index); if (ah->ah_flags & IB_AH_GRH) { if (ah->grh.sgid_index >= @@ -1879,9 +1879,9 @@ static int mlx5_set_path(struct mlx5_ib_dev *dev, struct mlx5_ib_qp *qp, ah->grh.sgid_index); path->dci_cfi_prio_sl = (ah->sl & 0x7) << 4; } else { - path->fl = (path_flags & MLX5_PATH_FLAG_FL) ? 0x80 : 0; - path->free_ar = (path_flags & MLX5_PATH_FLAG_FREE_AR) ? 0x80 : - 0; + path->fl_free_ar = (path_flags & MLX5_PATH_FLAG_FL) ? 0x80 : 0; + path->fl_free_ar |= + (path_flags & MLX5_PATH_FLAG_FREE_AR) ? 0x40 : 0; path->rlid = cpu_to_be16(ah->dlid); path->grh_mlid = ah->src_path_bits & 0x7f; if (ah->ah_flags & IB_AH_GRH) @@ -2266,7 +2266,7 @@ static int __mlx5_ib_modify_qp(struct ib_qp *ibqp, context->log_pg_sz_remote_qpn = cpu_to_be32(attr->dest_qp_num); if (attr_mask & IB_QP_PKEY_INDEX) - context->pri_path.pkey_index = attr->pkey_index; + context->pri_path.pkey_index = cpu_to_be16(attr->pkey_index); /* todo implement counter_index functionality */ @@ -4015,11 +4015,12 @@ static int query_qp_attr(struct mlx5_ib_dev *dev, struct mlx5_ib_qp *qp, if (qp->ibqp.qp_type == IB_QPT_RC || qp->ibqp.qp_type == IB_QPT_UC) { to_ib_ah_attr(dev, &qp_attr->ah_attr, &context->pri_path); to_ib_ah_attr(dev, &qp_attr->alt_ah_attr, &context->alt_path); - qp_attr->alt_pkey_index = context->alt_path.pkey_index & 0x7f; + qp_attr->alt_pkey_index = + be16_to_cpu(context->alt_path.pkey_index); qp_attr->alt_port_num = qp_attr->alt_ah_attr.port_num; } - qp_attr->pkey_index = context->pri_path.pkey_index & 0x7f; + qp_attr->pkey_index = be16_to_cpu(context->pri_path.pkey_index); qp_attr->port_num = context->pri_path.port; /* qp_attr->en_sqd_async_notify is only applicable in modify qp */ diff --git a/include/linux/mlx5/qp.h b/include/linux/mlx5/qp.h index 64221027bf1f..e4e29882fdfd 100644 --- a/include/linux/mlx5/qp.h +++ b/include/linux/mlx5/qp.h @@ -460,10 +460,9 @@ struct mlx5_core_qp { }; struct mlx5_qp_path { - u8 fl; + u8 fl_free_ar; u8 rsvd3; - u8 free_ar; - u8 pkey_index; + __be16 pkey_index; u8 rsvd0; u8 grh_mlid; __be16 rlid; -- cgit v1.2.3 From 5c1d3310d84309330264e2a06e5000eb289a44ad Mon Sep 17 00:00:00 2001 From: Robin Murphy Date: Tue, 7 Jun 2016 18:44:48 +0100 Subject: drivers: of: Fix of_pci.h header guard The compilation of of_pci.c is governed by CONFIG_OF_PCI, but the corresponding declarations in of_pci.h are inconsistently guarded by CONFIG_OF, with the result that if CONFIG_PCI is disabled for an OF platform, the dangling external declarations are still active and the inline stub definitions not. So far this has managed to go unnoticed since it happens that the only references to these functions are from code which itself depends on CONFIG_PCI or CONFIG_OF_PCI. Fix this with the appropriate config guard so that any new callers outside PCI-specific code don't start unexpectedly breaking under certain configs. Signed-off-by: Robin Murphy Signed-off-by: Rob Herring --- include/linux/of_pci.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'include/linux') diff --git a/include/linux/of_pci.h b/include/linux/of_pci.h index f6e9e85164e8..b969e9443962 100644 --- a/include/linux/of_pci.h +++ b/include/linux/of_pci.h @@ -8,7 +8,7 @@ struct pci_dev; struct of_phandle_args; struct device_node; -#ifdef CONFIG_OF +#ifdef CONFIG_OF_PCI int of_irq_parse_pci(const struct pci_dev *pdev, struct of_phandle_args *out_irq); struct device_node *of_pci_find_child_device(struct device_node *parent, unsigned int devfn); -- cgit v1.2.3 From 06dfeef88573cf032e5c27e37f80ff5237b3318a Mon Sep 17 00:00:00 2001 From: Ben Dooks Date: Thu, 9 Jun 2016 11:38:34 +0100 Subject: drivers: of: add definition of early_init_dt_alloc_reserved_memory_arch The function early_init_dt_alloc_reserved_memory_arch is defined in drivers/of/of_reserved_mem.c but is not declared in any of the header files. Add the declaration of this to avoid the warning: drivers/of/of_reserved_mem.c:31:19: warning: symbol 'early_init_dt_alloc_reserved_memory_arch' was not declared. Should it be static? Signed-off-by: Ben Dooks [robh: drop extern from declaration] Signed-off-by: Rob Herring --- include/linux/of_reserved_mem.h | 7 +++++++ 1 file changed, 7 insertions(+) (limited to 'include/linux') diff --git a/include/linux/of_reserved_mem.h b/include/linux/of_reserved_mem.h index ad2f67054372..c201060e0c6d 100644 --- a/include/linux/of_reserved_mem.h +++ b/include/linux/of_reserved_mem.h @@ -31,6 +31,13 @@ typedef int (*reservedmem_of_init_fn)(struct reserved_mem *rmem); int of_reserved_mem_device_init(struct device *dev); void of_reserved_mem_device_release(struct device *dev); +int early_init_dt_alloc_reserved_memory_arch(phys_addr_t size, + phys_addr_t align, + phys_addr_t start, + phys_addr_t end, + bool nomap, + phys_addr_t *res_base); + void fdt_init_reserved_mem(void); void fdt_reserved_mem_save_node(unsigned long node, const char *uname, phys_addr_t base, phys_addr_t size); -- cgit v1.2.3 From 418f8399a8bedf376ec13eb01088f04a76ebdd6f Mon Sep 17 00:00:00 2001 From: Majd Dibbiny Date: Fri, 10 Jun 2016 00:07:28 +0300 Subject: net/mlx5: Fix the size of modify QP mailbox Add 16 reserved bytes at the end of mlx5_modify_qp_mbox_in to match the hardware spec definition. Fixes: e126ba97dba9 ('mlx5: Add driver for Mellanox Connect-IB adapters') Signed-off-by: Majd Dibbiny Signed-off-by: Saeed Mahameed Signed-off-by: David S. Miller --- include/linux/mlx5/qp.h | 1 + 1 file changed, 1 insertion(+) (limited to 'include/linux') diff --git a/include/linux/mlx5/qp.h b/include/linux/mlx5/qp.h index 64221027bf1f..1532dcf6fc5e 100644 --- a/include/linux/mlx5/qp.h +++ b/include/linux/mlx5/qp.h @@ -560,6 +560,7 @@ struct mlx5_modify_qp_mbox_in { __be32 optparam; u8 rsvd0[4]; struct mlx5_qp_context ctx; + u8 rsvd2[16]; }; struct mlx5_modify_qp_mbox_out { -- cgit v1.2.3 From 86d56a1a6b7352542661d8a9463758c7f285fce3 Mon Sep 17 00:00:00 2001 From: Shahar Klein Date: Fri, 10 Jun 2016 00:07:30 +0300 Subject: net/mlx5: Fix MLX5_CMD_OP_MAX to be defined correctly Having MLX5_CMD_OP_MAX on another file causes us to repeatedly miss accounting new commands added to the driver and hence there're no entries for them in debugfs. To solve that, we integrate it into the commands enum as the last entry. Fixes: 34a40e689393 ('net/mlx5_core: Introduce modify flow table command') Signed-off-by: Shahar Klein Signed-off-by: Or Gerlitz Signed-off-by: Saeed Mahameed Signed-off-by: David S. Miller --- include/linux/mlx5/device.h | 2 -- include/linux/mlx5/mlx5_ifc.h | 3 ++- 2 files changed, 2 insertions(+), 3 deletions(-) (limited to 'include/linux') diff --git a/include/linux/mlx5/device.h b/include/linux/mlx5/device.h index 035abdf62cfe..51f0caf299d8 100644 --- a/include/linux/mlx5/device.h +++ b/include/linux/mlx5/device.h @@ -1240,8 +1240,6 @@ struct mlx5_destroy_psv_out { u8 rsvd[8]; }; -#define MLX5_CMD_OP_MAX 0x920 - enum { VPORT_STATE_DOWN = 0x0, VPORT_STATE_UP = 0x1, diff --git a/include/linux/mlx5/mlx5_ifc.h b/include/linux/mlx5/mlx5_ifc.h index 9a05cd7e5890..986a615f623c 100644 --- a/include/linux/mlx5/mlx5_ifc.h +++ b/include/linux/mlx5/mlx5_ifc.h @@ -205,7 +205,8 @@ enum { MLX5_CMD_OP_ALLOC_FLOW_COUNTER = 0x939, MLX5_CMD_OP_DEALLOC_FLOW_COUNTER = 0x93a, MLX5_CMD_OP_QUERY_FLOW_COUNTER = 0x93b, - MLX5_CMD_OP_MODIFY_FLOW_TABLE = 0x93c + MLX5_CMD_OP_MODIFY_FLOW_TABLE = 0x93c, + MLX5_CMD_OP_MAX }; struct mlx5_ifc_flow_table_fields_supported_bits { -- cgit v1.2.3 From 876d634d19e41603aab91455f2c52a78a28372d5 Mon Sep 17 00:00:00 2001 From: Maor Gottlieb Date: Fri, 10 Jun 2016 00:07:32 +0300 Subject: net/mlx5: Fix flow steering NIC capabilities check Flow steering infrastructure is currently used only on link layer ethernet, therefore the driver should initialize the flow steering when the device link layer is ethernet. In addition, add missing capability check before initializing the namespace of NIC RX flow tables. Fixes: 2530236303d9 ('net/mlx5_core: Flow steering tree initialization') Signed-off-by: Maor Gottlieb Signed-off-by: Saeed Mahameed Signed-off-by: David S. Miller --- drivers/net/ethernet/mellanox/mlx5/core/fs_core.c | 10 +++++++++- include/linux/mlx5/device.h | 6 ++++++ 2 files changed, 15 insertions(+), 1 deletion(-) (limited to 'include/linux') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c b/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c index fa6fec1930f5..c1efa5517d17 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/fs_core.c @@ -1767,6 +1767,9 @@ static void cleanup_root_ns(struct mlx5_core_dev *dev) void mlx5_cleanup_fs(struct mlx5_core_dev *dev) { + if (MLX5_CAP_GEN(dev, port_type) != MLX5_CAP_PORT_TYPE_ETH) + return; + cleanup_root_ns(dev); cleanup_single_prio_root_ns(dev, dev->priv.fdb_root_ns); cleanup_single_prio_root_ns(dev, dev->priv.esw_egress_root_ns); @@ -1828,15 +1831,20 @@ int mlx5_init_fs(struct mlx5_core_dev *dev) { int err = 0; + if (MLX5_CAP_GEN(dev, port_type) != MLX5_CAP_PORT_TYPE_ETH) + return 0; + err = mlx5_init_fc_stats(dev); if (err) return err; - if (MLX5_CAP_GEN(dev, nic_flow_table)) { + if (MLX5_CAP_GEN(dev, nic_flow_table) && + MLX5_CAP_FLOWTABLE_NIC_RX(dev, ft_support)) { err = init_root_ns(dev); if (err) goto err; } + if (MLX5_CAP_GEN(dev, eswitch_flow_table)) { err = init_fdb_root_ns(dev); if (err) diff --git a/include/linux/mlx5/device.h b/include/linux/mlx5/device.h index 51f0caf299d8..73a48479892d 100644 --- a/include/linux/mlx5/device.h +++ b/include/linux/mlx5/device.h @@ -1367,6 +1367,12 @@ enum mlx5_cap_type { #define MLX5_CAP_FLOWTABLE_MAX(mdev, cap) \ MLX5_GET(flow_table_nic_cap, mdev->hca_caps_max[MLX5_CAP_FLOW_TABLE], cap) +#define MLX5_CAP_FLOWTABLE_NIC_RX(mdev, cap) \ + MLX5_CAP_FLOWTABLE(mdev, flow_table_properties_nic_receive.cap) + +#define MLX5_CAP_FLOWTABLE_NIC_RX_MAX(mdev, cap) \ + MLX5_CAP_FLOWTABLE_MAX(mdev, flow_table_properties_nic_receive.cap) + #define MLX5_CAP_ESW_FLOWTABLE(mdev, cap) \ MLX5_GET(flow_table_eswitch_cap, \ mdev->hca_caps_cur[MLX5_CAP_ESWITCH_FLOW_TABLE], cap) -- cgit v1.2.3 From 23898c763f4af6f5c80b0230b1ea788a0ce3cf73 Mon Sep 17 00:00:00 2001 From: Noa Osherovich Date: Fri, 10 Jun 2016 00:07:37 +0300 Subject: net/mlx5: E-Switch, Modify node guid on vf set MAC In RoCE, the RDMA-CM needs the node guid to establish connection between nodes. Today, the node guid exposed to mlx5 Ethernet VFs is zero, therefore RDMA-CM on the VF is broken. Whenever the administrator sets a MAC for a VF, derive the node guid from it and set it as well in the following way: MAC: e4:1d:2d:b3:f4:01 -> node_guid: e4:1d:2d:ff:fe:b3:f4:01 Fixes: 77256579c6b43 ('net/mlx5: E-Switch, Introduce Vport...') Signed-off-by: Noa Osherovich Signed-off-by: Majd Dibbiny Signed-off-by: Saeed Mahameed Signed-off-by: David S. Miller --- drivers/net/ethernet/mellanox/mlx5/core/eswitch.c | 23 ++++++++++++-- drivers/net/ethernet/mellanox/mlx5/core/vport.c | 38 +++++++++++++++++++++++ include/linux/mlx5/mlx5_ifc.h | 9 ++++-- include/linux/mlx5/vport.h | 2 ++ 4 files changed, 68 insertions(+), 4 deletions(-) (limited to 'include/linux') diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c index cfec20cffd26..9b1855b199a1 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/eswitch.c @@ -1725,11 +1725,24 @@ void mlx5_eswitch_vport_event(struct mlx5_eswitch *esw, struct mlx5_eqe *eqe) (esw && MLX5_CAP_GEN(esw->dev, vport_group_manager) && mlx5_core_is_pf(esw->dev)) #define LEGAL_VPORT(esw, vport) (vport >= 0 && vport < esw->total_vports) +static void node_guid_gen_from_mac(u64 *node_guid, u8 mac[ETH_ALEN]) +{ + ((u8 *)node_guid)[7] = mac[0]; + ((u8 *)node_guid)[6] = mac[1]; + ((u8 *)node_guid)[5] = mac[2]; + ((u8 *)node_guid)[4] = 0xff; + ((u8 *)node_guid)[3] = 0xfe; + ((u8 *)node_guid)[2] = mac[3]; + ((u8 *)node_guid)[1] = mac[4]; + ((u8 *)node_guid)[0] = mac[5]; +} + int mlx5_eswitch_set_vport_mac(struct mlx5_eswitch *esw, int vport, u8 mac[ETH_ALEN]) { - int err = 0; struct mlx5_vport *evport; + u64 node_guid; + int err = 0; if (!ESW_ALLOWED(esw)) return -EPERM; @@ -1753,11 +1766,17 @@ int mlx5_eswitch_set_vport_mac(struct mlx5_eswitch *esw, return err; } + node_guid_gen_from_mac(&node_guid, mac); + err = mlx5_modify_nic_vport_node_guid(esw->dev, vport, node_guid); + if (err) + mlx5_core_warn(esw->dev, + "Failed to set vport %d node guid, err = %d. RDMA_CM will not function properly for this VF.\n", + vport, err); + mutex_lock(&esw->state_lock); if (evport->enabled) err = esw_vport_ingress_config(esw, evport); mutex_unlock(&esw->state_lock); - return err; } diff --git a/drivers/net/ethernet/mellanox/mlx5/core/vport.c b/drivers/net/ethernet/mellanox/mlx5/core/vport.c index b69dadcfb897..daf44cd4c566 100644 --- a/drivers/net/ethernet/mellanox/mlx5/core/vport.c +++ b/drivers/net/ethernet/mellanox/mlx5/core/vport.c @@ -508,6 +508,44 @@ int mlx5_query_nic_vport_node_guid(struct mlx5_core_dev *mdev, u64 *node_guid) } EXPORT_SYMBOL_GPL(mlx5_query_nic_vport_node_guid); +int mlx5_modify_nic_vport_node_guid(struct mlx5_core_dev *mdev, + u32 vport, u64 node_guid) +{ + int inlen = MLX5_ST_SZ_BYTES(modify_nic_vport_context_in); + void *nic_vport_context; + u8 *guid; + void *in; + int err; + + if (!vport) + return -EINVAL; + if (!MLX5_CAP_GEN(mdev, vport_group_manager)) + return -EACCES; + if (!MLX5_CAP_ESW(mdev, nic_vport_node_guid_modify)) + return -ENOTSUPP; + + in = mlx5_vzalloc(inlen); + if (!in) + return -ENOMEM; + + MLX5_SET(modify_nic_vport_context_in, in, + field_select.node_guid, 1); + MLX5_SET(modify_nic_vport_context_in, in, vport_number, vport); + MLX5_SET(modify_nic_vport_context_in, in, other_vport, !!vport); + + nic_vport_context = MLX5_ADDR_OF(modify_nic_vport_context_in, + in, nic_vport_context); + guid = MLX5_ADDR_OF(nic_vport_context, nic_vport_context, + node_guid); + MLX5_SET64(nic_vport_context, nic_vport_context, node_guid, node_guid); + + err = mlx5_modify_nic_vport_context(mdev, in, inlen); + + kvfree(in); + + return err; +} + int mlx5_query_nic_vport_qkey_viol_cntr(struct mlx5_core_dev *mdev, u16 *qkey_viol_cntr) { diff --git a/include/linux/mlx5/mlx5_ifc.h b/include/linux/mlx5/mlx5_ifc.h index 986a615f623c..e955a2859009 100644 --- a/include/linux/mlx5/mlx5_ifc.h +++ b/include/linux/mlx5/mlx5_ifc.h @@ -501,7 +501,9 @@ struct mlx5_ifc_e_switch_cap_bits { u8 vport_svlan_insert[0x1]; u8 vport_cvlan_insert_if_not_exist[0x1]; u8 vport_cvlan_insert_overwrite[0x1]; - u8 reserved_at_5[0x1b]; + u8 reserved_at_5[0x19]; + u8 nic_vport_node_guid_modify[0x1]; + u8 nic_vport_port_guid_modify[0x1]; u8 reserved_at_20[0x7e0]; }; @@ -4584,7 +4586,10 @@ struct mlx5_ifc_modify_nic_vport_context_out_bits { }; struct mlx5_ifc_modify_nic_vport_field_select_bits { - u8 reserved_at_0[0x19]; + u8 reserved_at_0[0x16]; + u8 node_guid[0x1]; + u8 port_guid[0x1]; + u8 reserved_at_18[0x1]; u8 mtu[0x1]; u8 change_event[0x1]; u8 promisc[0x1]; diff --git a/include/linux/mlx5/vport.h b/include/linux/mlx5/vport.h index 301da4a5e6bf..6c16c198f680 100644 --- a/include/linux/mlx5/vport.h +++ b/include/linux/mlx5/vport.h @@ -50,6 +50,8 @@ int mlx5_modify_nic_vport_mtu(struct mlx5_core_dev *mdev, u16 mtu); int mlx5_query_nic_vport_system_image_guid(struct mlx5_core_dev *mdev, u64 *system_image_guid); int mlx5_query_nic_vport_node_guid(struct mlx5_core_dev *mdev, u64 *node_guid); +int mlx5_modify_nic_vport_node_guid(struct mlx5_core_dev *mdev, + u32 vport, u64 node_guid); int mlx5_query_nic_vport_qkey_viol_cntr(struct mlx5_core_dev *mdev, u16 *qkey_viol_cntr); int mlx5_query_hca_vport_gid(struct mlx5_core_dev *dev, u8 other_vport, -- cgit v1.2.3