2024-04-06 15:44:40

by Erick Archer

[permalink] [raw]
Subject: [PATCH v3 2/3] RDMA/mana_ib: Prefer struct_size over open coded arithmetic

This is an effort to get rid of all multiplications from allocation
functions in order to prevent integer overflows [1][2].

As the "req" variable is a pointer to "struct mana_cfg_rx_steer_req_v2"
and this structure ends in a flexible array:

struct mana_cfg_rx_steer_req_v2 {
[...]
mana_handle_t indir_tab[] __counted_by(num_indir_entries);
};

the preferred way in the kernel is to use the struct_size() helper to
do the arithmetic instead of the calculation "size + size * count" in
the kzalloc() function.

Moreover, use the "offsetof" helper to get the indirect table offset
instead of the "sizeof" operator and avoid the open-coded arithmetic in
pointers using the new flex member. This new structure member also allow
us to remove the "req_indir_tab" variable since it is no longer needed.

This way, the code is more readable and safer.

This code was detected with the help of Coccinelle, and audited and
modified manually.

Link: https://www.kernel.org/doc/html/latest/process/deprecated.html#open-coded-arithmetic-in-allocator-arguments [1]
Link: https://github.com/KSPP/linux/issues/160 [2]
Signed-off-by: Erick Archer <[email protected]>
---
drivers/infiniband/hw/mana/qp.c | 12 +++++-------
1 file changed, 5 insertions(+), 7 deletions(-)

diff --git a/drivers/infiniband/hw/mana/qp.c b/drivers/infiniband/hw/mana/qp.c
index ef0a6dc664d0..4cd8f8afe80d 100644
--- a/drivers/infiniband/hw/mana/qp.c
+++ b/drivers/infiniband/hw/mana/qp.c
@@ -15,15 +15,13 @@ static int mana_ib_cfg_vport_steering(struct mana_ib_dev *dev,
struct mana_port_context *mpc = netdev_priv(ndev);
struct mana_cfg_rx_steer_req_v2 *req;
struct mana_cfg_rx_steer_resp resp = {};
- mana_handle_t *req_indir_tab;
struct gdma_context *gc;
u32 req_buf_size;
int i, err;

gc = mdev_to_gc(dev);

- req_buf_size =
- sizeof(*req) + sizeof(mana_handle_t) * MANA_INDIRECT_TABLE_SIZE;
+ req_buf_size = struct_size(req, indir_tab, MANA_INDIRECT_TABLE_SIZE);
req = kzalloc(req_buf_size, GFP_KERNEL);
if (!req)
return -ENOMEM;
@@ -44,20 +42,20 @@ static int mana_ib_cfg_vport_steering(struct mana_ib_dev *dev,
req->rss_enable = true;

req->num_indir_entries = MANA_INDIRECT_TABLE_SIZE;
- req->indir_tab_offset = sizeof(*req);
+ req->indir_tab_offset = offsetof(struct mana_cfg_rx_steer_req_v2,
+ indir_tab);
req->update_indir_tab = true;
req->cqe_coalescing_enable = 1;

- req_indir_tab = (mana_handle_t *)(req + 1);
/* The ind table passed to the hardware must have
* MANA_INDIRECT_TABLE_SIZE entries. Adjust the verb
* ind_table to MANA_INDIRECT_TABLE_SIZE if required
*/
ibdev_dbg(&dev->ib_dev, "ind table size %u\n", 1 << log_ind_tbl_size);
for (i = 0; i < MANA_INDIRECT_TABLE_SIZE; i++) {
- req_indir_tab[i] = ind_table[i % (1 << log_ind_tbl_size)];
+ req->indir_tab[i] = ind_table[i % (1 << log_ind_tbl_size)];
ibdev_dbg(&dev->ib_dev, "index %u handle 0x%llx\n", i,
- req_indir_tab[i]);
+ req->indir_tab[i]);
}

req->update_hashkey = true;
--
2.25.1



2024-04-08 19:42:35

by Long Li

[permalink] [raw]
Subject: RE: [PATCH v3 2/3] RDMA/mana_ib: Prefer struct_size over open coded arithmetic

> Subject: [PATCH v3 2/3] RDMA/mana_ib: Prefer struct_size over open coded
> arithmetic
>
> This is an effort to get rid of all multiplications from allocation functions in order to
> prevent integer overflows [1][2].
>
> As the "req" variable is a pointer to "struct mana_cfg_rx_steer_req_v2"
> and this structure ends in a flexible array:
>
> struct mana_cfg_rx_steer_req_v2 {
> [...]
> mana_handle_t indir_tab[] __counted_by(num_indir_entries); };
>
> the preferred way in the kernel is to use the struct_size() helper to do the
> arithmetic instead of the calculation "size + size * count" in the kzalloc() function.
>
> Moreover, use the "offsetof" helper to get the indirect table offset instead of the
> "sizeof" operator and avoid the open-coded arithmetic in pointers using the new
> flex member. This new structure member also allow us to remove the
> "req_indir_tab" variable since it is no longer needed.
>
> This way, the code is more readable and safer.
>
> This code was detected with the help of Coccinelle, and audited and modified
> manually.
>
> Link:
> https://www.ker/
> nel.org%2Fdoc%2Fhtml%2Flatest%2Fprocess%2Fdeprecated.html%23open-
> coded-arithmetic-in-allocator-
> arguments&data=05%7C02%7Clongli%40microsoft.com%7Cfcf2a410393a429633
> ca08dc56506b01%7C72f988bf86f141af91ab2d7cd011db47%7C1%7C0%7C63848
> 0150654917952%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQI
> joiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C0%7C%7C%7C&sdata=2zSek
> zsyXsS1s9xw%2FwaSEl3h4s6AeiykFG4KiJLzXOc%3D&reserved=0 [1]
> Link:
> https://github.co/
> m%2FKSPP%2Flinux%2Fissues%2F160&data=05%7C02%7Clongli%40microsoft.co
> m%7Cfcf2a410393a429633ca08dc56506b01%7C72f988bf86f141af91ab2d7cd01
> 1db47%7C1%7C0%7C638480150654924997%7CUnknown%7CTWFpbGZsb3d8ey
> JWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C
> 0%7C%7C%7C&sdata=4pzQWVWVcIaeS07VgXY1I6%2FS%2FEFejUD4qv1D2Ouwf
> pA%3D&reserved=0 [2]
> Signed-off-by: Erick Archer <[email protected]>

Reviewed-by: Long Li <[email protected]>


2024-04-08 21:38:56

by Justin Stitt

[permalink] [raw]
Subject: Re: [PATCH v3 2/3] RDMA/mana_ib: Prefer struct_size over open coded arithmetic

Hi,

On Sat, Apr 06, 2024 at 04:23:36PM +0200, Erick Archer wrote:
> This is an effort to get rid of all multiplications from allocation
> functions in order to prevent integer overflows [1][2].
>
> As the "req" variable is a pointer to "struct mana_cfg_rx_steer_req_v2"
> and this structure ends in a flexible array:
>
> struct mana_cfg_rx_steer_req_v2 {
> [...]
> mana_handle_t indir_tab[] __counted_by(num_indir_entries);
> };
>
> the preferred way in the kernel is to use the struct_size() helper to
> do the arithmetic instead of the calculation "size + size * count" in
> the kzalloc() function.
>
> Moreover, use the "offsetof" helper to get the indirect table offset
> instead of the "sizeof" operator and avoid the open-coded arithmetic in
> pointers using the new flex member. This new structure member also allow
> us to remove the "req_indir_tab" variable since it is no longer needed.
>
> This way, the code is more readable and safer.
>
> This code was detected with the help of Coccinelle, and audited and
> modified manually.
>
> Link: https://www.kernel.org/doc/html/latest/process/deprecated.html#open-coded-arithmetic-in-allocator-arguments [1]
> Link: https://github.com/KSPP/linux/issues/160 [2]
> Signed-off-by: Erick Archer <[email protected]>

Reviewed-by: Justin Stitt <[email protected]>

> ---
> drivers/infiniband/hw/mana/qp.c | 12 +++++-------
> 1 file changed, 5 insertions(+), 7 deletions(-)
>
> diff --git a/drivers/infiniband/hw/mana/qp.c b/drivers/infiniband/hw/mana/qp.c
> index ef0a6dc664d0..4cd8f8afe80d 100644
> --- a/drivers/infiniband/hw/mana/qp.c
> +++ b/drivers/infiniband/hw/mana/qp.c
> @@ -15,15 +15,13 @@ static int mana_ib_cfg_vport_steering(struct mana_ib_dev *dev,
> struct mana_port_context *mpc = netdev_priv(ndev);
> struct mana_cfg_rx_steer_req_v2 *req;
> struct mana_cfg_rx_steer_resp resp = {};
> - mana_handle_t *req_indir_tab;
> struct gdma_context *gc;
> u32 req_buf_size;
> int i, err;
>
> gc = mdev_to_gc(dev);
>
> - req_buf_size =
> - sizeof(*req) + sizeof(mana_handle_t) * MANA_INDIRECT_TABLE_SIZE;
> + req_buf_size = struct_size(req, indir_tab, MANA_INDIRECT_TABLE_SIZE);
> req = kzalloc(req_buf_size, GFP_KERNEL);
> if (!req)
> return -ENOMEM;
> @@ -44,20 +42,20 @@ static int mana_ib_cfg_vport_steering(struct mana_ib_dev *dev,
> req->rss_enable = true;
>
> req->num_indir_entries = MANA_INDIRECT_TABLE_SIZE;
> - req->indir_tab_offset = sizeof(*req);
> + req->indir_tab_offset = offsetof(struct mana_cfg_rx_steer_req_v2,
> + indir_tab);
> req->update_indir_tab = true;
> req->cqe_coalescing_enable = 1;
>
> - req_indir_tab = (mana_handle_t *)(req + 1);
> /* The ind table passed to the hardware must have
> * MANA_INDIRECT_TABLE_SIZE entries. Adjust the verb
> * ind_table to MANA_INDIRECT_TABLE_SIZE if required
> */
> ibdev_dbg(&dev->ib_dev, "ind table size %u\n", 1 << log_ind_tbl_size);
> for (i = 0; i < MANA_INDIRECT_TABLE_SIZE; i++) {
> - req_indir_tab[i] = ind_table[i % (1 << log_ind_tbl_size)];
> + req->indir_tab[i] = ind_table[i % (1 << log_ind_tbl_size)];
> ibdev_dbg(&dev->ib_dev, "index %u handle 0x%llx\n", i,
> - req_indir_tab[i]);
> + req->indir_tab[i]);
> }
>
> req->update_hashkey = true;
> --
> 2.25.1
>

Thanks
Justin