net/mlx5: fix offset calculation for modify field action
PATCH 1/1 net/mlx5: fix offset calculation for modify field action
Offsets are not taken into account during MAC addresses manipulation for the MODIFY_FIELD action. That leads to a wrong split between 0-15 and 16-47 bits and corrupted data being copied to/from MAC addresses. Use both source and destination offsets to calcucate the proper modify header action specification.
--- drivers/net/mlx5/mlx5_flow_dv.c | 25 +++++++++++++++---------- 1 file changed, 15 insertions(+), 10 deletions(-)
drivers/net/mlx5/mlx5_flow_dv.c +15 -10
@@ -413,6 +413,7 @@ flow_dv_convert_modify_action(struct rte_flow_item *item, {uint32_t i = resource->actions_num;struct mlx5_modification_cmd *actions = resource->actions; +uint32_t carry_b =0 ;/* * The item and mask are provided in big-endian format.
@@ -422,8 +423,8 @@ flow_dv_convert_modify_action(struct rte_flow_item *item,MLX5_ASSERT (item->mask);MLX5_ASSERT (field->size); do { - unsigned int size_b; - unsigned int off_b; + uint32_t size_b; + uint32_t off_b; uint32_t mask; uint32_t data; bool next_field = true;
@@ -441,7 +442,7 @@ flow_dv_convert_modify_action(struct rte_flow_item *item,continue ; }/* Deduce actual data width in bits from mask value. */ - off_b =rte_bsf32 (mask); + off_b =rte_bsf32 (mask) + carry_b; size_b =sizeof (uint32_t ) *CHAR_BIT - off_b -__builtin_clz (mask);MLX5_ASSERT (size_b);
@@ -463,19 +464,23 @@ flow_dv_convert_modify_action(struct rte_flow_item *item,* Destination field overflow. Copy leftovers of * a source field to the next destination field. */ -if ((size_b > dcopy->size *CHAR_BIT ) && dcopy->size) { - actions[i].length = dcopy->size *CHAR_BIT ; - field->offset += dcopy->size; + carry_b =0 ; +if ((size_b > dcopy->size *CHAR_BIT - dcopy->offset) && + dcopy->size !=0 ) { + actions[i].length = + dcopy->size *CHAR_BIT - dcopy->offset; + carry_b = actions[i].length; next_field = false; }/* * Not enough bits in a source filed to fill a * destination field. Switch to the next source. */ -if (dcopy->size > field->size && - (size_b == field->size *CHAR_BIT )) { - actions[i].length = field->size *CHAR_BIT ; - dcopy->offset += field->size *CHAR_BIT ; +if ((size_b < dcopy->size *CHAR_BIT - dcopy->offset) && + (size_b == field->size *CHAR_BIT - off_b)) { + actions[i].length = + field->size *CHAR_BIT - off_b; + dcopy->offset += actions[i].length; next_dcopy = false; }if (next_dcopy)
Signature · 1 line
2.18.2
Hi,
Quoted 15 lines
-----Original Message-----
From: Alexander Kozyrev <[email protected]>
Sent: Monday, July 5, 2021 12:47 PM
To: [email protected]
Cc: [email protected]; Raslan Darawsheh <[email protected]>; Slava Ovsiienko <[email protected]>; Matan Azrad <[email protected]>
Subject: [PATCH] net/mlx5: fix offset calculation for modify field actionOffsets are not taken into account during MAC addresses manipulation for the
MODIFY_FIELDaction. That leads to a wrong split between 0-15 and 16-47 bits and corrupted data being copied to/from MAC addresses. Use both source and destination offsets to calcucate the proper modify header action specification.
Quoted 5 lines
Fixes fdd0c046f4 ("net/mlx5: fix modify field action order for MAC")Signed-off-by Alexander Kozyrev <[email protected]>
Patch applied to next-net-mlx,
Kindest regards,
Raslan Darawsheh