[PATCH rdma-rc] RDMA/mlx5: Fix DevX umem page size truncation for pages >= 4G

[PATCH rdma-rc] RDMA/mlx5: Fix DevX umem page size truncation for pages >= 4G

[0z5a]

From: Dezhen Lu <dezhen.lu@student.uni-tuebingen.de>

devx_umem_find_best_pgsize() computes the page size in an unsigned long
but returns it as unsigned int, and devx_umem_reg_cmd_alloc() stores it
in an unsigned int. When the DMA mapping of the umem is contiguous and
aligned to 4 GiB or more, and the umem length is a multiple of that
size, ib_umem_find_best_pgoff() selects a page size >= 2^32. Since the
value is a power of two, the truncation turns it into 0 and the
registration fails with -EINVAL.

The umem log_page_size field is 5 bits wide on top of
MLX5_ADAPTER_PAGE_SHIFT, so the command can encode these page sizes.

Seen on two GH200 nodes (ConnectX-7, firmware 32.38.3056 and
32.43.1014) with a dma-buf exported from cudaMalloc memory: registering
4 GiB or 8 GiB umems at 4 GiB-aligned offsets of the dma-buf fails with
EINVAL every time, while 2 GiB umems at the same offsets and
ibv_reg_dmabuf_mr() of the same ranges succeed. Restricting the user
pgsz_bitmap to <= 2 GiB makes all of them succeed.

Other large physically contiguous buffers, such as a 16 GiB hugetlb
page on an arm64 kernel with 64K base pages, should hit the same path.

Use unsigned long for the page size, as commit f0c2427412b4
("RDMA/mlx5: Fix page_size variable overflow") already did for the MR
paths.

The bug was found by an AI coding assistant (Claude Opus 5.5, Anthropic)
during an audit of 32-bit truncations of page sizes and offsets in
drivers/infiniband/core and drivers/infiniband/hw/mlx5. The assistant
also wrote the reproducer and this patch. The reproducer was run on the
GH200 nodes described above.

Fixes: 7610ab57de56 ("RDMA/mlx5: Allow larger pages in DevX umem")
Cc: stable@vger.kernel.org
Signed-off-by: Dezhen Lu <dezhen.lu@student.uni-tuebingen.de>
---

Notes:
    Testing notes:
    - The failures above were reproduced with the DOCA-OFED 26.01 mlx5_ib,
      whose devx_umem_find_best_pgsize() is identical to upstream; I do not
      have root on these hosts, so I could not load a patched module.
    - So I have not verified that CREATE_UMEM accepts log_page_size values
      for pages >= 4 GiB. If it does not, capping pgsz_bitmap at 2^31 in
      devx_umem_find_best_pgsize() is the alternative fix.
    - Compile-tested on v7.4-rc1 (arm64 defconfig + MLX5_INFINIBAND, W=1).
    - A standalone reproducer (dma-buf from cuMemGetHandleForAddressRange()
      plus mlx5dv_devx_umem_reg_ex(), sweeping offsets) is available on
      request.

 drivers/infiniband/hw/mlx5/devx.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/drivers/infiniband/hw/mlx5/devx.c b/drivers/infiniband/hw/mlx5/devx.c
index fc6f793dc..2277160d9 100644
--- a/drivers/infiniband/hw/mlx5/devx.c
+++ b/drivers/infiniband/hw/mlx5/devx.c
@@ -2294,8 +2294,8 @@ static int devx_umem_get(struct mlx5_ib_dev *dev, struct ib_ucontext *ucontext,
 	return 0;
 }
 
-static unsigned int devx_umem_find_best_pgsize(struct ib_umem *umem,
-					       unsigned long pgsz_bitmap)
+static unsigned long devx_umem_find_best_pgsize(struct ib_umem *umem,
+						unsigned long pgsz_bitmap)
 {
 	unsigned long page_size;
 
@@ -2333,7 +2333,7 @@ static int devx_umem_reg_cmd_alloc(struct mlx5_ib_dev *dev,
 				   int access)
 {
 	unsigned long pgsz_bitmap;
-	unsigned int page_size;
+	unsigned long page_size;
 	__be64 *mtt;
 	void *umem;
 	int ret;
-- 
2.43.0

Re: [PATCH rdma-rc] RDMA/mlx5: Fix DevX umem page size truncation for pages >= 4G

[Leon Romanovsky]

On Tue, Nov 03, 2026 at 03:28:04PM +0100, 0z5a wrote:
> From: Dezhen Lu <dezhen.lu@student.uni-tuebingen.de>
> 
> devx_umem_find_best_pgsize() computes the page size in an unsigned long
> but returns it as unsigned int, and devx_umem_reg_cmd_alloc() stores it
> in an unsigned int. When the DMA mapping of the umem is contiguous and
> aligned to 4 GiB or more, and the umem length is a multiple of that
> size, ib_umem_find_best_pgoff() selects a page size >= 2^32. Since the
> value is a power of two, the truncation turns it into 0 and the
> registration fails with -EINVAL.
> 
> The umem log_page_size field is 5 bits wide on top of
> MLX5_ADAPTER_PAGE_SHIFT, so the command can encode these page sizes.
> 
> Seen on two GH200 nodes (ConnectX-7, firmware 32.38.3056 and
> 32.43.1014) with a dma-buf exported from cudaMalloc memory: registering
> 4 GiB or 8 GiB umems at 4 GiB-aligned offsets of the dma-buf fails with
> EINVAL every time, while 2 GiB umems at the same offsets and
> ibv_reg_dmabuf_mr() of the same ranges succeed. Restricting the user
> pgsz_bitmap to <= 2 GiB makes all of them succeed.
> 
> Other large physically contiguous buffers, such as a 16 GiB hugetlb
> page on an arm64 kernel with 64K base pages, should hit the same path.
> 
> Use unsigned long for the page size, as commit f0c2427412b4
> ("RDMA/mlx5: Fix page_size variable overflow") already did for the MR
> paths.
> 
> The bug was found by an AI coding assistant (Claude Opus 5.5, Anthropic)
> during an audit of 32-bit truncations of page sizes and offsets in
> drivers/infiniband/core and drivers/infiniband/hw/mlx5. The assistant
> also wrote the reproducer and this patch. The reproducer was run on the
> GH200 nodes described above.
> 
> Fixes: 7610ab57de56 ("RDMA/mlx5: Allow larger pages in DevX umem")
> Cc: stable@vger.kernel.org
> Signed-off-by: Dezhen Lu <dezhen.lu@student.uni-tuebingen.de>
> ---
> 
> Notes:
>     Testing notes:
>     - The failures above were reproduced with the DOCA-OFED 26.01 mlx5_ib,
>       whose devx_umem_find_best_pgsize() is identical to upstream; I do not
>       have root on these hosts, so I could not load a patched module.
>     - So I have not verified that CREATE_UMEM accepts log_page_size values
>       for pages >= 4 GiB. If it does not, capping pgsz_bitmap at 2^31 in
>       devx_umem_find_best_pgsize() is the alternative fix.
>     - Compile-tested on v7.4-rc1 (arm64 defconfig + MLX5_INFINIBAND, W=1).
>     - A standalone reproducer (dma-buf from cuMemGetHandleForAddressRange()
>       plus mlx5dv_devx_umem_reg_ex(), sweeping offsets) is available on
>       request.


Isn't this fix enough?
https://patch.msgid.link/r/1-v1-88303e9e509f+f7-ib_umem_types_jgg@nvidia.com
15fe76e23615 ("RDMA/umem: Fix truncation for block sizes >= 4G")

Thanks

> 
>  drivers/infiniband/hw/mlx5/devx.c | 6 +++---
>  1 file changed, 3 insertions(+), 3 deletions(-)
> 
> diff --git a/drivers/infiniband/hw/mlx5/devx.c b/drivers/infiniband/hw/mlx5/devx.c
> index fc6f793dc..2277160d9 100644
> --- a/drivers/infiniband/hw/mlx5/devx.c
> +++ b/drivers/infiniband/hw/mlx5/devx.c
> @@ -2294,8 +2294,8 @@ static int devx_umem_get(struct mlx5_ib_dev *dev, struct ib_ucontext *ucontext,
>  	return 0;
>  }
>  
> -static unsigned int devx_umem_find_best_pgsize(struct ib_umem *umem,
> -					       unsigned long pgsz_bitmap)
> +static unsigned long devx_umem_find_best_pgsize(struct ib_umem *umem,
> +						unsigned long pgsz_bitmap)
>  {
>  	unsigned long page_size;
>  
> @@ -2333,7 +2333,7 @@ static int devx_umem_reg_cmd_alloc(struct mlx5_ib_dev *dev,
>  				   int access)
>  {
>  	unsigned long pgsz_bitmap;
> -	unsigned int page_size;
> +	unsigned long page_size;
>  	__be64 *mtt;
>  	void *umem;
>  	int ret;
> -- 
> 2.43.0
> 

Re: [PATCH rdma-rc] RDMA/mlx5: Fix DevX umem page size truncation for pages >= 4G

[Jason Gunthorpe]

On Wed, Nov 04, 2026 at 12:07:47PM +0200, Leon Romanovsky wrote:

> Isn't this fix enough?
> https://patch.msgid.link/r/1-v1-88303e9e509f+f7-ib_umem_types_jgg@nvidia.com
> 15fe76e23615 ("RDMA/umem: Fix truncation for block sizes >= 4G")

No, the long AI generated commit message buried what the bug here is:

> > -static unsigned int devx_umem_find_best_pgsize(struct ib_umem *umem,
> > -					       unsigned long pgsz_bitmap)
> > +static unsigned long devx_umem_find_best_pgsize(struct ib_umem *umem,
> > +						unsigned long pgsz_bitmap)
> >  {

That 'int' is an integer truncation if the page size is > 4G. The
entire page size path from the core code has to be unsigned long if
the pgsz_bitmap is unsigned long.

Similar root cause of truncate, but different code.

Jason

Re: [PATCH rdma-rc] RDMA/mlx5: Fix DevX umem page size truncation for pages >= 4G

[Jason Gunthorpe]

On Tue, Nov 03, 2026 at 03:28:04PM +0100, 0z5a wrote:
> From: Dezhen Lu <dezhen.lu@student.uni-tuebingen.de>
> 
> devx_umem_find_best_pgsize() computes the page size in an unsigned long
> but returns it as unsigned int, and devx_umem_reg_cmd_alloc() stores it
> in an unsigned int. When the DMA mapping of the umem is contiguous and
> aligned to 4 GiB or more, and the umem length is a multiple of that
> size, ib_umem_find_best_pgoff() selects a page size >= 2^32. Since the
> value is a power of two, the truncation turns it into 0 and the
> registration fails with -EINVAL.
> 
> The umem log_page_size field is 5 bits wide on top of
> MLX5_ADAPTER_PAGE_SHIFT, so the command can encode these page sizes.
> 
> Seen on two GH200 nodes (ConnectX-7, firmware 32.38.3056 and
> 32.43.1014) with a dma-buf exported from cudaMalloc memory: registering
> 4 GiB or 8 GiB umems at 4 GiB-aligned offsets of the dma-buf fails with
> EINVAL every time, while 2 GiB umems at the same offsets and
> ibv_reg_dmabuf_mr() of the same ranges succeed. Restricting the user
> pgsz_bitmap to <= 2 GiB makes all of them succeed.
> 
> Other large physically contiguous buffers, such as a 16 GiB hugetlb
> page on an arm64 kernel with 64K base pages, should hit the same path.
> 
> Use unsigned long for the page size, as commit f0c2427412b4
> ("RDMA/mlx5: Fix page_size variable overflow") already did for the MR
> paths.
> 
> The bug was found by an AI coding assistant (Claude Opus 5.5, Anthropic)
> during an audit of 32-bit truncations of page sizes and offsets in
> drivers/infiniband/core and drivers/infiniband/hw/mlx5. The assistant
> also wrote the reproducer and this patch. The reproducer was run on the
> GH200 nodes described above.
> 
> Fixes: 7610ab57de56 ("RDMA/mlx5: Allow larger pages in DevX umem")
> Cc: stable@vger.kernel.org
> Signed-off-by: Dezhen Lu <dezhen.lu@student.uni-tuebingen.de>
> ---

Reviewed-by: Jason Gunthorpe <jgg@nvidia.com>

>     - So I have not verified that CREATE_UMEM accepts log_page_size values
>       for pages >= 4 GiB. If it does not, capping pgsz_bitmap at 2^31 in
>       devx_umem_find_best_pgsize() is the alternative fix.

It is userspace's problem to provide a correct bitmap in
MLX5_IB_ATTR_DEVX_UMEM_REG_PGSZ_BITMAP
for whatever it is doing. FW can reject large sizes outside what it
supports if userspace sends the incorrect bitmap.

Jason

[PATCH rdma-rc] RDMA/mlx5: Fix DevX umem page size truncation for pages >= 4G

[0z5a]

From: Dezhen Lu <dezhen.lu@student.uni-tuebingen.de>

devx_umem_find_best_pgsize() computes the page size in an unsigned long
but returns it as unsigned int, and devx_umem_reg_cmd_alloc() stores it
in an unsigned int. When the DMA mapping of the umem is contiguous and
aligned to 4 GiB or more, and the umem length is a multiple of that
size, ib_umem_find_best_pgoff() selects a page size >= 2^32. Since the
value is a power of two, the truncation turns it into 0 and the
registration fails with -EINVAL.

The umem log_page_size field is 5 bits wide on top of
MLX5_ADAPTER_PAGE_SHIFT, so the command can encode these page sizes.

Seen on two GH200 nodes (ConnectX-7, firmware 32.38.3056 and
32.43.1014) with a dma-buf exported from cudaMalloc memory: registering
4 GiB or 8 GiB umems at 4 GiB-aligned offsets of the dma-buf fails with
EINVAL every time, while 2 GiB umems at the same offsets and
ibv_reg_dmabuf_mr() of the same ranges succeed. Restricting the user
pgsz_bitmap to <= 2 GiB makes all of them succeed.

Other large physically contiguous buffers, such as a 16 GiB hugetlb
page on an arm64 kernel with 64K base pages, should hit the same path.

Use unsigned long for the page size, as commit f0c2427412b4
("RDMA/mlx5: Fix page_size variable overflow") already did for the MR
paths.

The bug was found by an AI coding assistant (Claude Opus 5.5, Anthropic)
during an audit of 32-bit truncations of page sizes and offsets in
drivers/infiniband/core and drivers/infiniband/hw/mlx5. The assistant
also wrote the reproducer and this patch. The reproducer was run on the
GH200 nodes described above.

Fixes: 7610ab57de56 ("RDMA/mlx5: Allow larger pages in DevX umem")
Cc: stable@vger.kernel.org
Signed-off-by: Dezhen Lu <dezhen.lu@student.uni-tuebingen.de>
---

Notes:
    Testing notes:
    - The failures above were reproduced with the DOCA-OFED 26.01 mlx5_ib,
      whose devx_umem_find_best_pgsize() is identical to upstream; I do not
      have root on these hosts, so I could not load a patched module.
    - So I have not verified that CREATE_UMEM accepts log_page_size values
      for pages >= 4 GiB. If it does not, capping pgsz_bitmap at 2^31 in
      devx_umem_find_best_pgsize() is the alternative fix.
    - Compile-tested on v7.4-rc1 (arm64 defconfig + MLX5_INFINIBAND, W=1).
    - A standalone reproducer (dma-buf from cuMemGetHandleForAddressRange()
      plus mlx5dv_devx_umem_reg_ex(), sweeping offsets) is available on
      request.

 drivers/infiniband/hw/mlx5/devx.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/drivers/infiniband/hw/mlx5/devx.c b/drivers/infiniband/hw/mlx5/devx.c
index fc6f793dc..2277160d9 100644
--- a/drivers/infiniband/hw/mlx5/devx.c
+++ b/drivers/infiniband/hw/mlx5/devx.c
@@ -2294,8 +2294,8 @@ static int devx_umem_get(struct mlx5_ib_dev *dev, struct ib_ucontext *ucontext,
 	return 0;
 }
 
-static unsigned int devx_umem_find_best_pgsize(struct ib_umem *umem,
-					       unsigned long pgsz_bitmap)
+static unsigned long devx_umem_find_best_pgsize(struct ib_umem *umem,
+						unsigned long pgsz_bitmap)
 {
 	unsigned long page_size;
 
@@ -2333,7 +2333,7 @@ static int devx_umem_reg_cmd_alloc(struct mlx5_ib_dev *dev,
 				   int access)
 {
 	unsigned long pgsz_bitmap;
-	unsigned int page_size;
+	unsigned long page_size;
 	__be64 *mtt;
 	void *umem;
 	int ret;
-- 
2.43.0