If CONFIG_ARCH_DMA_ADDR_T_64BIT enabled for x86 systems and physical
memory is more than 4GB, dma_map_page may return a valid memory
address which greater than 0xffffffff. As a result, the mlx5 device page
allocator RB tree will be initialized with valid addresses greater than
0xfffffff.
However, (addr & PAGE_MASK) set the high four bytes to zeros. So, it's
impossible for the function, free_4k, to release the pages whose
addresses greater than 4GB. Memory leaks. And mlx5_ib module can't
release the pages when user try to remove the module, as a result,
system hang.
[root@rdma05 root]# dmesg | grep addr | head
addr = 3fe384000
addr & PAGE_MASK = fe384000
[root@rdma05 root]# rmmod mlx5_ib <---- hang on
---------------------- cosnole log -----------------
mlx5_ib 0000:04:00.0: irq 138 for MSI/MSI-X
alloc irq_desc for 139 on node -1
alloc kstat_irqs on node -1
mlx5_ib 0000:04:00.0: irq 139 for MSI/MSI-X
0000:04:00.0:free_4k:221:(pid 1519): page not found
0000:04:00.0:free_4k:221:(pid 1519): page not found
0000:04:00.0:free_4k:221:(pid 1519): page not found
0000:04:00.0:free_4k:221:(pid 1519): page not found
---------------------- cosnole log -----------------
Fixes: bf0bf77f6519 ('mlx5: Support communicating arbitrary host page size to firmware')
Signed-off-by: Honggang Li <[email protected]>
---
.../net/ethernet/mellanox/mlx5/core/pagealloc.c | 10 ++++++----
1 files changed, 6 insertions(+), 4 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx5/core/pagealloc.c b/drivers/net/ethernet/mellanox/mlx5/core/pagealloc.c
index df22383..8a64542 100644
--- a/drivers/net/ethernet/mellanox/mlx5/core/pagealloc.c
+++ b/drivers/net/ethernet/mellanox/mlx5/core/pagealloc.c
@@ -211,26 +211,28 @@ static int alloc_4k(struct mlx5_core_dev *dev, u64 *addr)
return 0;
}
+#define MLX5_U64_4K_PAGE_MASK ((~(u64)0U) << PAGE_SHIFT)
+
static void free_4k(struct mlx5_core_dev *dev, u64 addr)
{
struct fw_page *fwp;
int n;
- fwp = find_fw_page(dev, addr & PAGE_MASK);
+ fwp = find_fw_page(dev, addr & MLX5_U64_4K_PAGE_MASK);
if (!fwp) {
mlx5_core_warn(dev, "page not found\n");
return;
}
- n = (addr & ~PAGE_MASK) >> MLX5_ADAPTER_PAGE_SHIFT;
+ n = (addr & ~MLX5_U64_4K_PAGE_MASK) >> MLX5_ADAPTER_PAGE_SHIFT;
fwp->free_count++;
set_bit(n, &fwp->bitmask);
if (fwp->free_count == MLX5_NUM_4K_IN_PAGE) {
rb_erase(&fwp->rb_node, &dev->priv.page_root);
if (fwp->free_count != 1)
list_del(&fwp->list);
- dma_unmap_page(&dev->pdev->dev, addr & PAGE_MASK, PAGE_SIZE,
- DMA_BIDIRECTIONAL);
+ dma_unmap_page(&dev->pdev->dev, addr & MLX5_U64_4K_PAGE_MASK,
+ PAGE_SIZE, DMA_BIDIRECTIONAL);
__free_page(fwp->page);
kfree(fwp);
} else if (fwp->free_count == 1) {
--
1.7.1
On Wed, Apr 15, 2015 at 04:36:15PM +0800, Honggang Li wrote:
> If CONFIG_ARCH_DMA_ADDR_T_64BIT enabled for x86 systems and physical
> memory is more than 4GB, dma_map_page may return a valid memory
> address which greater than 0xffffffff. As a result, the mlx5 device page
> allocator RB tree will be initialized with valid addresses greater than
> 0xfffffff.
>
> However, (addr & PAGE_MASK) set the high four bytes to zeros. So, it's
> impossible for the function, free_4k, to release the pages whose
> addresses greater than 4GB. Memory leaks. And mlx5_ib module can't
> release the pages when user try to remove the module, as a result,
> system hang.
>
> [root@rdma05 root]# dmesg | grep addr | head
> addr = 3fe384000
> addr & PAGE_MASK = fe384000
> [root@rdma05 root]# rmmod mlx5_ib <---- hang on
>
> ---------------------- cosnole log -----------------
> mlx5_ib 0000:04:00.0: irq 138 for MSI/MSI-X
> alloc irq_desc for 139 on node -1
> alloc kstat_irqs on node -1
> mlx5_ib 0000:04:00.0: irq 139 for MSI/MSI-X
> 0000:04:00.0:free_4k:221:(pid 1519): page not found
> 0000:04:00.0:free_4k:221:(pid 1519): page not found
> 0000:04:00.0:free_4k:221:(pid 1519): page not found
> 0000:04:00.0:free_4k:221:(pid 1519): page not found
> ---------------------- cosnole log -----------------
>
> Fixes: bf0bf77f6519 ('mlx5: Support communicating arbitrary host page size to firmware')
> Signed-off-by: Honggang Li <[email protected]>
> ---
Acked-by: Eli Cohen <[email protected]>