Skip to contentSkip to conversation

fm10k: use snprintf() instead of sprintf() to avoid buffer overflow

Intel-wired-lan Jacob Keller 3 participants 8 messages Aug 25, 2015 – Sep 30, 2015

Subsystems (2)
lore
Share & tools
lore
#0 lore raw
---
 drivers/net/ethernet/intel/fm10k/fm10k_debugfs.c | 6 +++---
 drivers/net/ethernet/intel/fm10k/fm10k_ethtool.c | 8 ++++----
 2 files changed, 7 insertions(+), 7 deletions(-)
drivers/net/ethernet/intel/fm10k/fm10k_debugfs.c +3 -3
index 08ecf43dffc7..5304bc1fbecd 100644
@@ -176,7 +176,7 @@ void fm10k_dbg_q_vector_init(struct fm10k_q_vector *q_vector)
 		return;
 
 	/* Generate a folder for each q_vector */
-	sprintf(name, "q_vector.%03d", q_vector->v_idx);
+	snprintf(name, sizeof(name), "q_vector.%03d", q_vector->v_idx);
 
 	q_vector->dbg_q_vector = debugfs_create_dir(name, interface->dbg_intfc);
 	if (!q_vector->dbg_q_vector)
@@ -186,7 +186,7 @@ void fm10k_dbg_q_vector_init(struct fm10k_q_vector *q_vector)
 	for (i = 0; i < q_vector->tx.count; i++) {
 		struct fm10k_ring *ring = &q_vector->tx.ring[i];
 
-		sprintf(name, "tx_ring.%03d", ring->queue_index);
+		snprintf(name, sizeof(name), "tx_ring.%03d", ring->queue_index);
 
 		debugfs_create_file(name, 0600,
 				    q_vector->dbg_q_vector, ring,
@@ -197,7 +197,7 @@ void fm10k_dbg_q_vector_init(struct fm10k_q_vector *q_vector)
 	for (i = 0; i < q_vector->rx.count; i++) {
 		struct fm10k_ring *ring = &q_vector->rx.ring[i];
 
-		sprintf(name, "rx_ring.%03d", ring->queue_index);
+		snprintf(name, sizeof(name), "rx_ring.%03d", ring->queue_index);
 
 		debugfs_create_file(name, 0600,
 				    q_vector->dbg_q_vector, ring,
drivers/net/ethernet/intel/fm10k/fm10k_ethtool.c +4 -4
index 4ef2fbd22911..447a5f8da42f 100644
@@ -206,13 +206,13 @@ static void fm10k_get_stat_strings(struct net_device *dev, u8 *data)
 	}
 
 	for (i = 0; i < interface->hw.mac.max_queues; i++) {
-		sprintf(p, "tx_queue_%u_packets", i);
+		snprintf(p, ETH_GSTRING_LEN, "tx_queue_%u_packets", i);
 		p += ETH_GSTRING_LEN;
-		sprintf(p, "tx_queue_%u_bytes", i);
+		snprintf(p, ETH_GSTRING_LEN, "tx_queue_%u_bytes", i);
 		p += ETH_GSTRING_LEN;
-		sprintf(p, "rx_queue_%u_packets", i);
+		snprintf(p, ETH_GSTRING_LEN, "rx_queue_%u_packets", i);
 		p += ETH_GSTRING_LEN;
-		sprintf(p, "rx_queue_%u_bytes", i);
+		snprintf(p, ETH_GSTRING_LEN, "rx_queue_%u_bytes", i);
 		p += ETH_GSTRING_LEN;
 	}
 }
Signature · 1 line
2.5.0.280.g4aaba03
#1 lore raw

[Intel-wired-lan] [next-queue] fm10k: Correct MTU for jumbo frames

Based on hardware testing, the host interface can support up to 15368 bytes on the wire. Subtracting 8 for an internal switch tag and 14 for the L2 header shows that the MTU should not be allowed to be set above 15346 bytes.

---
 drivers/net/ethernet/intel/fm10k/fm10k.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/net/ethernet/intel/fm10k/fm10k.h +1 -1
index 14440200499b..9e05a36783e5 100644
@@ -33,7 +33,7 @@
 #include "fm10k_pf.h"
 #include "fm10k_vf.h"
 
-#define FM10K_MAX_JUMBO_FRAME_SIZE	15358	/* Maximum supported size 15K */
+#define FM10K_MAX_JUMBO_FRAME_SIZE	15346	/* Maximum supported size 15K */
 
 #define MAX_QUEUES	FM10K_MAX_QUEUES_PF
 
Signature · 1 line
2.5.0.280.g4aaba03
#2 lore raw

[Intel-wired-lan] [next-queue] fm10k: do not use enum as boolean

In reply to Jacob Keller

Check for actual value NETREG_UNINITIALIZED in case it ever changes from the current value of zero.

---
 drivers/net/ethernet/intel/fm10k/fm10k_main.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/net/ethernet/intel/fm10k/fm10k_main.c +1 -1
index cb5e3150048b..000f2b5ab3e5 100644
@@ -1905,7 +1905,7 @@ static void fm10k_init_reta(struct fm10k_intfc *interface)
 	u32 reta, base;
 
 	/* If the netdev is initialized we have to maintain table if possible */
-	if (interface->netdev->reg_state) {
+	if (interface->netdev->reg_state != NETREG_UNINITIALIZED) {
 		for (i = FM10K_RETA_SIZE; i--;) {
 			reta = interface->reta[i];
 			if ((((reta << 24) >> 24) < rss_i) &&
Signature · 1 line
2.5.0.280.g4aaba03
Tested-by Krishneil Singh
#3 lore raw

[Intel-wired-lan] [next-queue] fm10k: do not use enum as boolean

-----Original Message-----
From: Intel-wired-lan [mailto:intel-wired-lan-bounces at lists.osuosl.org] On Behalf Of Jacob Keller
Sent: Monday, August 24, 2015 5:02 PM
To: Intel Wired LAN <[email protected]>
Subject: [Intel-wired-lan] [next-queue] fm10k: do not use enum as boolean

Check for actual value NETREG_UNINITIALIZED in case it ever changes from the current value of zero.

---

#4 lore raw

[Intel-wired-lan] [next-queue] fm10k: do not assume VF always has 1 queue

In reply to Jacob Keller

It is possible that the PF has not yet assigned resources to the VF. Although rare, this could result in the VF attempting to read queues it does not own and result in FUM or THI faults in the PF. To prevent this, check queue 0 before we continue in init_hw_vf.

---
 drivers/net/ethernet/intel/fm10k/fm10k_vf.c | 7 ++++++-
 1 file changed, 6 insertions(+), 1 deletion(-)
drivers/net/ethernet/intel/fm10k/fm10k_vf.c +6 -1
index 36c8b0aa08fd..4d1da54a061b 100644
@@ -103,7 +103,12 @@ static s32 fm10k_init_hw_vf(struct fm10k_hw *hw)
 	s32 err;
 	u16 i;
 
-	/* assume we always have at least 1 queue */
+	/* verify we have at least 1 queue */
+	if (!~FM10K_READ_REG(hw, FM10K_TXQCTL(0)) ||
+	    !~FM10K_READ_REG(hw, FM10K_RXQCTL(0)))
+		return FM10K_ERR_NO_RESOURCES;
+
+	/* determine how many queues we have */
 	for (i = 1; tqdloc0 && (i < FM10K_MAX_QUEUES_POOL); i++) {
 		/* verify the Descriptor cache offsets are increasing */
 		tqdloc = ~fm10k_read_reg(hw, FM10K_TQDLOC(i));
Signature · 1 line
2.5.0.280.g4aaba03
#5 lore raw
In reply to Jacob Keller
Quoted 6 lines · On Mon, 24 Aug 2015 17:01:58 -0700 Jacob Keller <[email protected]>…
---
 drivers/net/ethernet/intel/fm10k/fm10k_debugfs.c | 6 +++---
 drivers/net/ethernet/intel/fm10k/fm10k_ethtool.c | 8 ++++----
 2 files changed, 7 insertions(+), 7 deletions(-)

The changes look fine, but I've seen multiple comments to the list that there needs to be a body in the commit message.

Suggest:

Subject:
fm10k: use snprintf()

Body:
Avoid buffer overflow and follow best practices by using snprintf() instead of sprintf().

#6 lore raw
Quoted 4 lines · Jesse Brandeburg

The changes look fine, but I've seen multiple comments to the list that
there needs to be a body in the commit message.

Really? Even for fairly trivial changes such as this?

I can re-work it.

Regards,
Jake

Tested-by Krishneil Singh
#7 lore raw
In reply to Jacob Keller

-----Original Message-----
From: Intel-wired-lan [mailto:intel-wired-lan-bounces at lists.osuosl.org] On Behalf Of Jacob Keller
Sent: Monday, August 24, 2015 5:02 PM
To: Intel Wired LAN <[email protected]>
Subject: [Intel-wired-lan] [next-queue] fm10k: use snprintf() instead of sprintf() to avoid buffer overflow

---