Skip to contentSkip to conversation

qed/qed_ll2: Replace one-element array with flexible-array member

PATCH Applied Gustavo A. R. Silva 2 participants 3 messages Sep 28, 2020 – Sep 29, 2020

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 qed/qed_ll2: Replace one-element array with flexible-array member

There is a regular need in the kernel to provide a way to declare having a dynamically sized set of trailing elements in a structure. Kernel code should always use “flexible array members”[1] for these cases. The older style of one-element or zero-length arrays should no longer be used[2].

Refactor the code according to the use of a flexible-array member in struct qed_ll2_tx_packet, instead of a one-element array and use the struct_size() helper to calculate the size for the allocations. Commit f5823fe6897c ("qed: Add ll2 option to limit the number of bds per packet") was used as a reference point for these changes.

Also, it's important to notice that flexible-array members should occur last in any structure, and structures containing such arrays and that are members of other structures, must also occur last in the containing structure. That's why _cur_completing_packet_ is now moved to the bottom in struct qed_ll2_tx_queue. _descq_mem_ and _cur_send_packet_ are also moved for unification.

[1] https://en.wikipedia.org/wiki/Flexible_array_member
[2] https://www.kernel.org/doc/html/v5.9-rc1/process/deprecated.html#zero-length-and-one-element-arrays

---
 drivers/net/ethernet/qlogic/qed/qed_ll2.c | 18 ++++++++----------
 drivers/net/ethernet/qlogic/qed/qed_ll2.h |  8 ++++----
 2 files changed, 12 insertions(+), 14 deletions(-)
drivers/net/ethernet/qlogic/qed/qed_ll2.c +8 -10
index 0452b728c527..49783f365079 100644
@@ -1185,7 +1185,7 @@ static int qed_ll2_acquire_connection_tx(struct qed_hwfn *p_hwfn,
 		.elem_size	= sizeof(struct core_tx_bd),
 	};
 	struct qed_ll2_tx_packet *p_descq;
-	u32 desc_size;
+	size_t desc_size;
 	u32 capacity;
 	int rc = 0;
 
@@ -1198,10 +1198,9 @@ static int qed_ll2_acquire_connection_tx(struct qed_hwfn *p_hwfn,
 		goto out;
 
 	capacity = qed_chain_get_capacity(&p_ll2_info->tx_queue.txq_chain);
-	/* First element is part of the packet, rest are flexibly added */
-	desc_size = (sizeof(*p_descq) +
-		     (p_ll2_info->input.tx_max_bds_per_packet - 1) *
-		     sizeof(p_descq->bds_set));
+	/* All bds_set elements are flexibily added. */
+	desc_size = struct_size(p_descq, bds_set,
+				p_ll2_info->input.tx_max_bds_per_packet);
 
 	p_descq = kcalloc(capacity, desc_size, GFP_KERNEL);
 	if (!p_descq) {
@@ -1524,7 +1523,7 @@ int qed_ll2_establish_connection(void *cxt, u8 connection_handle)
 	struct qed_ptt *p_ptt;
 	int rc = -EINVAL;
 	u32 i, capacity;
-	u32 desc_size;
+	size_t desc_size;
 	u8 qid;
 
 	p_ptt = qed_ptt_acquire(p_hwfn);
@@ -1558,10 +1557,9 @@ int qed_ll2_establish_connection(void *cxt, u8 connection_handle)
 	INIT_LIST_HEAD(&p_tx->sending_descq);
 	spin_lock_init(&p_tx->lock);
 	capacity = qed_chain_get_capacity(&p_tx->txq_chain);
-	/* First element is part of the packet, rest are flexibly added */
-	desc_size = (sizeof(*p_pkt) +
-		     (p_ll2_conn->input.tx_max_bds_per_packet - 1) *
-		     sizeof(p_pkt->bds_set));
+	/* All bds_set elements are flexibily added. */
+	desc_size = struct_size(p_pkt, bds_set,
+				p_ll2_conn->input.tx_max_bds_per_packet);
 
 	for (i = 0; i < capacity; i++) {
 		p_pkt = p_tx->descq_mem + desc_size * i;
drivers/net/ethernet/qlogic/qed/qed_ll2.h +4 -4
index 500d0c4f8077..df88d00053a2 100644
@@ -56,7 +56,7 @@ struct qed_ll2_tx_packet {
 		struct core_tx_bd *txq_bd;
 		dma_addr_t tx_frag;
 		u16 frag_len;
-	} bds_set[1];
+	} bds_set[];
 };
 
 struct qed_ll2_rx_queue {
@@ -86,9 +86,6 @@ struct qed_ll2_tx_queue {
 	struct list_head active_descq;
 	struct list_head free_descq;
 	struct list_head sending_descq;
-	void *descq_mem; /* memory for variable sized qed_ll2_tx_packet*/
-	struct qed_ll2_tx_packet *cur_send_packet;
-	struct qed_ll2_tx_packet cur_completing_packet;
 	u16 cur_completing_bd_idx;
 	void __iomem *doorbell_addr;
 	struct core_db_data db_msg;
@@ -96,6 +93,9 @@ struct qed_ll2_tx_queue {
 	u16 cur_send_frag_num;
 	u16 cur_completing_frag_num;
 	bool b_completing_packet;
+	void *descq_mem; /* memory for variable sized qed_ll2_tx_packet*/
+	struct qed_ll2_tx_packet *cur_send_packet;
+	struct qed_ll2_tx_packet cur_completing_packet;
 };
 
 struct qed_ll2_info {
Signature · 1 line
2.27.0
maintainer Applied
#1 lore raw

From: "Gustavo A. R. Silva" <[email protected]>
Date: Mon, 28 Sep 2020 10:16:17 -0500

Quoted 18 lines

There is a regular need in the kernel to provide a way to declare having a dynamically sized set of trailing elements in a structure. Kernel code should always use “flexible array members”[1] for these cases. The older style of one-element or zero-length arrays should no longer be used[2].

Refactor the code according to the use of a flexible-array member in struct qed_ll2_tx_packet, instead of a one-element array and use the struct_size() helper to calculate the size for the allocations. Commit f5823fe6897c ("qed: Add ll2 option to limit the number of bds per packet") was used as a reference point for these changes.

Also, it's important to notice that flexible-array members should occur last in any structure, and structures containing such arrays and that are members of other structures, must also occur last in the containing structure. That's why _cur_completing_packet_ is now moved to the bottom in struct qed_ll2_tx_queue. _descq_mem_ and _cur_send_packet_ are also moved for unification.

Quoted 4 lines

[1] https://en.wikipedia.org/wiki/Flexible_array_member
[2] https://www.kernel.org/doc/html/v5.9-rc1/process/deprecated.html#zero-length-and-one-element-arrays

I find such tags enormously misleading, because the kernel test robot didn't perform any functional testing of this change and honestly that's the part I'm more concerned about rather than "does it build".

Anyone can check test the build.

Quoted 2 lines

Applied.

#2 lore raw
Quoted 24 lines · David Miller

From: "Gustavo A. R. Silva" <[email protected]>
Date: Mon, 28 Sep 2020 10:16:17 -0500

Quoted 21 lines

There is a regular need in the kernel to provide a way to declare having a dynamically sized set of trailing elements in a structure. Kernel code should always use “flexible array members”[1] for these cases. The older style of one-element or zero-length arrays should no longer be used[2].

Refactor the code according to the use of a flexible-array member in struct qed_ll2_tx_packet, instead of a one-element array and use the struct_size() helper to calculate the size for the allocations. Commit f5823fe6897c ("qed: Add ll2 option to limit the number of bds per packet") was used as a reference point for these changes.

Also, it's important to notice that flexible-array members should occur last in any structure, and structures containing such arrays and that are members of other structures, must also occur last in the containing structure. That's why _cur_completing_packet_ is now moved to the bottom in struct qed_ll2_tx_queue. _descq_mem_ and _cur_send_packet_ are also moved for unification.

[1] https://en.wikipedia.org/wiki/Flexible_array_member
[2] https://www.kernel.org/doc/html/v5.9-rc1/process/deprecated.html#zero-length-and-one-element-arrays

Quoted 8 lines
Quoted 1 line

I find such tags enormously misleading, because the kernel test robot didn't perform any functional testing of this change and honestly that's the part I'm more concerned about rather than "does it build".

Anyone can check test the build.

I agree. I should add a Built-Tested-by tag instead next time.

Quoted 4 lines
Quoted 2 lines

Applied.

Thanks
--
Gustavo