Commit dc2316eb authored by Yishai Hadas's avatar Yishai Hadas Committed by Doug Ledford

IB/mlx5: Fix device memory flows

Fix device memory flows so that only once there will be no live mmaped
VA to a given allocation the matching object will be destroyed.

This prevents a potential scenario that existing VA that was mmaped by
one process might still be used post its deallocation despite that it's
owned now by other process.

The above is achieved by integrating with IB core APIs to manage
mmap/munmap. Only once the refcount will become 0 the DM object and its
underlay area will be freed.

Fixes: 3b113a1e ("IB/mlx5: Support device memory type attribute")
Signed-off-by: default avatarYishai Hadas <yishaih@mellanox.com>
Signed-off-by: default avatarLeon Romanovsky <leonro@mellanox.com>
Link: https://lore.kernel.org/r/20191212100237.330654-3-leon@kernel.orgSigned-off-by: default avatarDoug Ledford <dledford@redhat.com>
parent 7a763d18
...@@ -157,7 +157,7 @@ int mlx5_cmd_alloc_memic(struct mlx5_dm *dm, phys_addr_t *addr, ...@@ -157,7 +157,7 @@ int mlx5_cmd_alloc_memic(struct mlx5_dm *dm, phys_addr_t *addr,
return -ENOMEM; return -ENOMEM;
} }
int mlx5_cmd_dealloc_memic(struct mlx5_dm *dm, phys_addr_t addr, u64 length) void mlx5_cmd_dealloc_memic(struct mlx5_dm *dm, phys_addr_t addr, u64 length)
{ {
struct mlx5_core_dev *dev = dm->dev; struct mlx5_core_dev *dev = dm->dev;
u64 hw_start_addr = MLX5_CAP64_DEV_MEM(dev, memic_bar_start_addr); u64 hw_start_addr = MLX5_CAP64_DEV_MEM(dev, memic_bar_start_addr);
...@@ -175,15 +175,13 @@ int mlx5_cmd_dealloc_memic(struct mlx5_dm *dm, phys_addr_t addr, u64 length) ...@@ -175,15 +175,13 @@ int mlx5_cmd_dealloc_memic(struct mlx5_dm *dm, phys_addr_t addr, u64 length)
MLX5_SET(dealloc_memic_in, in, memic_size, length); MLX5_SET(dealloc_memic_in, in, memic_size, length);
err = mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out)); err = mlx5_cmd_exec(dev, in, sizeof(in), out, sizeof(out));
if (err)
return;
if (!err) {
spin_lock(&dm->lock); spin_lock(&dm->lock);
bitmap_clear(dm->memic_alloc_pages, bitmap_clear(dm->memic_alloc_pages,
start_page_idx, num_pages); start_page_idx, num_pages);
spin_unlock(&dm->lock); spin_unlock(&dm->lock);
}
return err;
} }
int mlx5_cmd_query_ext_ppcnt_counters(struct mlx5_core_dev *dev, void *out) int mlx5_cmd_query_ext_ppcnt_counters(struct mlx5_core_dev *dev, void *out)
......
...@@ -46,7 +46,7 @@ int mlx5_cmd_modify_cong_params(struct mlx5_core_dev *mdev, ...@@ -46,7 +46,7 @@ int mlx5_cmd_modify_cong_params(struct mlx5_core_dev *mdev,
void *in, int in_size); void *in, int in_size);
int mlx5_cmd_alloc_memic(struct mlx5_dm *dm, phys_addr_t *addr, int mlx5_cmd_alloc_memic(struct mlx5_dm *dm, phys_addr_t *addr,
u64 length, u32 alignment); u64 length, u32 alignment);
int mlx5_cmd_dealloc_memic(struct mlx5_dm *dm, phys_addr_t addr, u64 length); void mlx5_cmd_dealloc_memic(struct mlx5_dm *dm, phys_addr_t addr, u64 length);
void mlx5_cmd_dealloc_pd(struct mlx5_core_dev *dev, u32 pdn, u16 uid); void mlx5_cmd_dealloc_pd(struct mlx5_core_dev *dev, u32 pdn, u16 uid);
void mlx5_cmd_destroy_tir(struct mlx5_core_dev *dev, u32 tirn, u16 uid); void mlx5_cmd_destroy_tir(struct mlx5_core_dev *dev, u32 tirn, u16 uid);
void mlx5_cmd_destroy_tis(struct mlx5_core_dev *dev, u32 tisn, u16 uid); void mlx5_cmd_destroy_tis(struct mlx5_core_dev *dev, u32 tisn, u16 uid);
......
...@@ -2074,6 +2074,24 @@ static int mlx5_ib_mmap_clock_info_page(struct mlx5_ib_dev *dev, ...@@ -2074,6 +2074,24 @@ static int mlx5_ib_mmap_clock_info_page(struct mlx5_ib_dev *dev,
virt_to_page(dev->mdev->clock_info)); virt_to_page(dev->mdev->clock_info));
} }
static void mlx5_ib_mmap_free(struct rdma_user_mmap_entry *entry)
{
struct mlx5_user_mmap_entry *mentry = to_mmmap(entry);
struct mlx5_ib_dev *dev = to_mdev(entry->ucontext->device);
struct mlx5_ib_dm *mdm;
switch (mentry->mmap_flag) {
case MLX5_IB_MMAP_TYPE_MEMIC:
mdm = container_of(mentry, struct mlx5_ib_dm, mentry);
mlx5_cmd_dealloc_memic(&dev->dm, mdm->dev_addr,
mdm->size);
kfree(mdm);
break;
default:
WARN_ON(true);
}
}
static int uar_mmap(struct mlx5_ib_dev *dev, enum mlx5_ib_mmap_cmd cmd, static int uar_mmap(struct mlx5_ib_dev *dev, enum mlx5_ib_mmap_cmd cmd,
struct vm_area_struct *vma, struct vm_area_struct *vma,
struct mlx5_ib_ucontext *context) struct mlx5_ib_ucontext *context)
...@@ -2186,26 +2204,55 @@ static int uar_mmap(struct mlx5_ib_dev *dev, enum mlx5_ib_mmap_cmd cmd, ...@@ -2186,26 +2204,55 @@ static int uar_mmap(struct mlx5_ib_dev *dev, enum mlx5_ib_mmap_cmd cmd,
return err; return err;
} }
static int dm_mmap(struct ib_ucontext *context, struct vm_area_struct *vma) static int add_dm_mmap_entry(struct ib_ucontext *context,
struct mlx5_ib_dm *mdm,
u64 address)
{ {
struct mlx5_ib_ucontext *mctx = to_mucontext(context); mdm->mentry.mmap_flag = MLX5_IB_MMAP_TYPE_MEMIC;
struct mlx5_ib_dev *dev = to_mdev(context->device); mdm->mentry.address = address;
u16 page_idx = get_extended_index(vma->vm_pgoff); return rdma_user_mmap_entry_insert_range(
size_t map_size = vma->vm_end - vma->vm_start; context, &mdm->mentry.rdma_entry,
u32 npages = map_size >> PAGE_SHIFT; mdm->size,
MLX5_IB_MMAP_DEVICE_MEM << 16,
(MLX5_IB_MMAP_DEVICE_MEM << 16) + (1UL << 16) - 1);
}
static unsigned long mlx5_vma_to_pgoff(struct vm_area_struct *vma)
{
unsigned long idx;
u8 command;
command = get_command(vma->vm_pgoff);
idx = get_extended_index(vma->vm_pgoff);
return (command << 16 | idx);
}
static int mlx5_ib_mmap_offset(struct mlx5_ib_dev *dev,
struct vm_area_struct *vma,
struct ib_ucontext *ucontext)
{
struct mlx5_user_mmap_entry *mentry;
struct rdma_user_mmap_entry *entry;
unsigned long pgoff;
pgprot_t prot;
phys_addr_t pfn; phys_addr_t pfn;
int ret;
if (find_next_zero_bit(mctx->dm_pages, page_idx + npages, page_idx) != pgoff = mlx5_vma_to_pgoff(vma);
page_idx + npages) entry = rdma_user_mmap_entry_get_pgoff(ucontext, pgoff);
if (!entry)
return -EINVAL; return -EINVAL;
pfn = ((dev->mdev->bar_addr + mentry = to_mmmap(entry);
MLX5_CAP64_DEV_MEM(dev->mdev, memic_bar_start_addr)) >> pfn = (mentry->address >> PAGE_SHIFT);
PAGE_SHIFT) + prot = pgprot_writecombine(vma->vm_page_prot);
page_idx; ret = rdma_user_mmap_io(ucontext, vma, pfn,
return rdma_user_mmap_io(context, vma, pfn, map_size, entry->npages * PAGE_SIZE,
pgprot_writecombine(vma->vm_page_prot), prot,
NULL); entry);
rdma_user_mmap_entry_put(&mentry->rdma_entry);
return ret;
} }
static int mlx5_ib_mmap(struct ib_ucontext *ibcontext, struct vm_area_struct *vma) static int mlx5_ib_mmap(struct ib_ucontext *ibcontext, struct vm_area_struct *vma)
...@@ -2248,11 +2295,8 @@ static int mlx5_ib_mmap(struct ib_ucontext *ibcontext, struct vm_area_struct *vm ...@@ -2248,11 +2295,8 @@ static int mlx5_ib_mmap(struct ib_ucontext *ibcontext, struct vm_area_struct *vm
case MLX5_IB_MMAP_CLOCK_INFO: case MLX5_IB_MMAP_CLOCK_INFO:
return mlx5_ib_mmap_clock_info_page(dev, vma, context); return mlx5_ib_mmap_clock_info_page(dev, vma, context);
case MLX5_IB_MMAP_DEVICE_MEM:
return dm_mmap(ibcontext, vma);
default: default:
return -EINVAL; return mlx5_ib_mmap_offset(dev, vma, ibcontext);
} }
return 0; return 0;
...@@ -2288,8 +2332,9 @@ static int handle_alloc_dm_memic(struct ib_ucontext *ctx, ...@@ -2288,8 +2332,9 @@ static int handle_alloc_dm_memic(struct ib_ucontext *ctx,
{ {
struct mlx5_dm *dm_db = &to_mdev(ctx->device)->dm; struct mlx5_dm *dm_db = &to_mdev(ctx->device)->dm;
u64 start_offset; u64 start_offset;
u32 page_idx; u16 page_idx;
int err; int err;
u64 address;
dm->size = roundup(attr->length, MLX5_MEMIC_BASE_SIZE); dm->size = roundup(attr->length, MLX5_MEMIC_BASE_SIZE);
...@@ -2298,28 +2343,30 @@ static int handle_alloc_dm_memic(struct ib_ucontext *ctx, ...@@ -2298,28 +2343,30 @@ static int handle_alloc_dm_memic(struct ib_ucontext *ctx,
if (err) if (err)
return err; return err;
page_idx = (dm->dev_addr - pci_resource_start(dm_db->dev->pdev, 0) - address = dm->dev_addr & PAGE_MASK;
MLX5_CAP64_DEV_MEM(dm_db->dev, memic_bar_start_addr)) >> err = add_dm_mmap_entry(ctx, dm, address);
PAGE_SHIFT; if (err)
goto err_dealloc;
page_idx = dm->mentry.rdma_entry.start_pgoff & 0xFFFF;
err = uverbs_copy_to(attrs, err = uverbs_copy_to(attrs,
MLX5_IB_ATTR_ALLOC_DM_RESP_PAGE_INDEX, MLX5_IB_ATTR_ALLOC_DM_RESP_PAGE_INDEX,
&page_idx, sizeof(page_idx)); &page_idx,
sizeof(page_idx));
if (err) if (err)
goto err_dealloc; goto err_copy;
start_offset = dm->dev_addr & ~PAGE_MASK; start_offset = dm->dev_addr & ~PAGE_MASK;
err = uverbs_copy_to(attrs, err = uverbs_copy_to(attrs,
MLX5_IB_ATTR_ALLOC_DM_RESP_START_OFFSET, MLX5_IB_ATTR_ALLOC_DM_RESP_START_OFFSET,
&start_offset, sizeof(start_offset)); &start_offset, sizeof(start_offset));
if (err) if (err)
goto err_dealloc; goto err_copy;
bitmap_set(to_mucontext(ctx)->dm_pages, page_idx,
DIV_ROUND_UP(dm->size, PAGE_SIZE));
return 0; return 0;
err_copy:
rdma_user_mmap_entry_remove(&dm->mentry.rdma_entry);
err_dealloc: err_dealloc:
mlx5_cmd_dealloc_memic(dm_db, dm->dev_addr, dm->size); mlx5_cmd_dealloc_memic(dm_db, dm->dev_addr, dm->size);
...@@ -2423,23 +2470,13 @@ int mlx5_ib_dealloc_dm(struct ib_dm *ibdm, struct uverbs_attr_bundle *attrs) ...@@ -2423,23 +2470,13 @@ int mlx5_ib_dealloc_dm(struct ib_dm *ibdm, struct uverbs_attr_bundle *attrs)
struct mlx5_ib_ucontext *ctx = rdma_udata_to_drv_context( struct mlx5_ib_ucontext *ctx = rdma_udata_to_drv_context(
&attrs->driver_udata, struct mlx5_ib_ucontext, ibucontext); &attrs->driver_udata, struct mlx5_ib_ucontext, ibucontext);
struct mlx5_core_dev *dev = to_mdev(ibdm->device)->mdev; struct mlx5_core_dev *dev = to_mdev(ibdm->device)->mdev;
struct mlx5_dm *dm_db = &to_mdev(ibdm->device)->dm;
struct mlx5_ib_dm *dm = to_mdm(ibdm); struct mlx5_ib_dm *dm = to_mdm(ibdm);
u32 page_idx;
int ret; int ret;
switch (dm->type) { switch (dm->type) {
case MLX5_IB_UAPI_DM_TYPE_MEMIC: case MLX5_IB_UAPI_DM_TYPE_MEMIC:
ret = mlx5_cmd_dealloc_memic(dm_db, dm->dev_addr, dm->size); rdma_user_mmap_entry_remove(&dm->mentry.rdma_entry);
if (ret) return 0;
return ret;
page_idx = (dm->dev_addr - pci_resource_start(dev->pdev, 0) -
MLX5_CAP64_DEV_MEM(dev, memic_bar_start_addr)) >>
PAGE_SHIFT;
bitmap_clear(ctx->dm_pages, page_idx,
DIV_ROUND_UP(dm->size, PAGE_SIZE));
break;
case MLX5_IB_UAPI_DM_TYPE_STEERING_SW_ICM: case MLX5_IB_UAPI_DM_TYPE_STEERING_SW_ICM:
ret = mlx5_dm_sw_icm_dealloc(dev, MLX5_SW_ICM_TYPE_STEERING, ret = mlx5_dm_sw_icm_dealloc(dev, MLX5_SW_ICM_TYPE_STEERING,
dm->size, ctx->devx_uid, dm->dev_addr, dm->size, ctx->devx_uid, dm->dev_addr,
...@@ -6235,6 +6272,7 @@ static const struct ib_device_ops mlx5_ib_dev_ops = { ...@@ -6235,6 +6272,7 @@ static const struct ib_device_ops mlx5_ib_dev_ops = {
.map_mr_sg = mlx5_ib_map_mr_sg, .map_mr_sg = mlx5_ib_map_mr_sg,
.map_mr_sg_pi = mlx5_ib_map_mr_sg_pi, .map_mr_sg_pi = mlx5_ib_map_mr_sg_pi,
.mmap = mlx5_ib_mmap, .mmap = mlx5_ib_mmap,
.mmap_free = mlx5_ib_mmap_free,
.modify_cq = mlx5_ib_modify_cq, .modify_cq = mlx5_ib_modify_cq,
.modify_device = mlx5_ib_modify_device, .modify_device = mlx5_ib_modify_device,
.modify_port = mlx5_ib_modify_port, .modify_port = mlx5_ib_modify_port,
......
...@@ -118,6 +118,10 @@ enum { ...@@ -118,6 +118,10 @@ enum {
MLX5_MEMIC_BASE_SIZE = 1 << MLX5_MEMIC_BASE_ALIGN, MLX5_MEMIC_BASE_SIZE = 1 << MLX5_MEMIC_BASE_ALIGN,
}; };
enum mlx5_ib_mmap_type {
MLX5_IB_MMAP_TYPE_MEMIC = 1,
};
#define MLX5_LOG_SW_ICM_BLOCK_SIZE(dev) \ #define MLX5_LOG_SW_ICM_BLOCK_SIZE(dev) \
(MLX5_CAP_DEV_MEM(dev, log_sw_icm_alloc_granularity)) (MLX5_CAP_DEV_MEM(dev, log_sw_icm_alloc_granularity))
#define MLX5_SW_ICM_BLOCK_SIZE(dev) (1 << MLX5_LOG_SW_ICM_BLOCK_SIZE(dev)) #define MLX5_SW_ICM_BLOCK_SIZE(dev) (1 << MLX5_LOG_SW_ICM_BLOCK_SIZE(dev))
...@@ -135,7 +139,6 @@ struct mlx5_ib_ucontext { ...@@ -135,7 +139,6 @@ struct mlx5_ib_ucontext {
u32 tdn; u32 tdn;
u64 lib_caps; u64 lib_caps;
DECLARE_BITMAP(dm_pages, MLX5_MAX_MEMIC_PAGES);
u16 devx_uid; u16 devx_uid;
/* For RoCE LAG TX affinity */ /* For RoCE LAG TX affinity */
atomic_t tx_port_affinity; atomic_t tx_port_affinity;
...@@ -556,6 +559,12 @@ enum mlx5_ib_mtt_access_flags { ...@@ -556,6 +559,12 @@ enum mlx5_ib_mtt_access_flags {
MLX5_IB_MTT_WRITE = (1 << 1), MLX5_IB_MTT_WRITE = (1 << 1),
}; };
struct mlx5_user_mmap_entry {
struct rdma_user_mmap_entry rdma_entry;
u8 mmap_flag;
u64 address;
};
struct mlx5_ib_dm { struct mlx5_ib_dm {
struct ib_dm ibdm; struct ib_dm ibdm;
phys_addr_t dev_addr; phys_addr_t dev_addr;
...@@ -567,6 +576,7 @@ struct mlx5_ib_dm { ...@@ -567,6 +576,7 @@ struct mlx5_ib_dm {
} icm_dm; } icm_dm;
/* other dm types specific params should be added here */ /* other dm types specific params should be added here */
}; };
struct mlx5_user_mmap_entry mentry;
}; };
#define MLX5_IB_MTT_PRESENT (MLX5_IB_MTT_READ | MLX5_IB_MTT_WRITE) #define MLX5_IB_MTT_PRESENT (MLX5_IB_MTT_READ | MLX5_IB_MTT_WRITE)
...@@ -1101,6 +1111,13 @@ to_mflow_act(struct ib_flow_action *ibact) ...@@ -1101,6 +1111,13 @@ to_mflow_act(struct ib_flow_action *ibact)
return container_of(ibact, struct mlx5_ib_flow_action, ib_action); return container_of(ibact, struct mlx5_ib_flow_action, ib_action);
} }
static inline struct mlx5_user_mmap_entry *
to_mmmap(struct rdma_user_mmap_entry *rdma_entry)
{
return container_of(rdma_entry,
struct mlx5_user_mmap_entry, rdma_entry);
}
int mlx5_ib_db_map_user(struct mlx5_ib_ucontext *context, int mlx5_ib_db_map_user(struct mlx5_ib_ucontext *context,
struct ib_udata *udata, unsigned long virt, struct ib_udata *udata, unsigned long virt,
struct mlx5_db *db); struct mlx5_db *db);
......
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment