Skip to contentSkip to conversation

net: dcb: initialize PFC value before calling getpfccfg

PATCH Applied Joas Antonio dos Santos 3 participants 4 messages Oct 25 – Oct 28

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 net: dcb: initialize PFC value before calling getpfccfg

dcbnl_getpfccfg() and dcbnl_cee_fill() pass the address of an uninitialized stack u8 to the driver's getpfccfg() op and then put it into the netlink reply unconditionally. The op returns void, so the core cannot tell whether the driver wrote the value.

i40e_dcbnl_get_pfc_cfg() and ice_dcbnl_get_pfc_cfg() return without writing *setting unless the DCBX agent is host managed and in CEE mode. Both drivers normally run with firmware managed DCBX (DCB_CAP_DCBX_LLD_MANAGED), and i40e with FW LLDP disabled runs host managed in IEEE mode, so on a DCB capable i40e/ice port the PFC values in DCB_CMD_PFC_GCFG and DCB_CMD_CEE_GET replies, and in the CEE part of RTNLGRP_DCB notifications, come from uninitialized stack. The same u8 is reused for all eight priorities, so this is one byte per reply.

dcb_doit() only requires CAP_NET_ADMIN for RTM_SETDCB, so the GET commands are available to unprivileged users.

Set the value to 0 before each call, like the PG config path already presets its outputs to DCB_ATTR_VALUE_UNDEFINED. 0 is also what the reply contains today on kernels built with CONFIG_INIT_STACK_ALL_ZERO, so nothing changes for those.

Fixes b94b013eb626 ("ice: Implement DCBNL support")
Fixes 5effa78e7c94 ("i40e: Add netlink callbacks support for software based DCB")
Assisted-by Yaga Agent HackerSec Claude:claude-opus-5-5

---
Found by code inspection, compile-tested only (W=1). I don't have i40e/ice hardware to check the reply. On kernels built with CONFIG_INIT_STACK_ALL_ZERO the byte is already zero, so in practice it only leaks on builds without stack auto-initialization.

 net/dcb/dcbnl.c | 2 ++
 1 file changed, 2 insertions(+)
net/dcb/dcbnl.c +2 -0
index ab71f53128..ad0aebe0ad 100644
@@ -281,6 +281,7 @@ static int dcbnl_getpfccfg(struct net_device *netdev, struct nlmsghdr *nlh,
 		if (!getall && !data[i])
 			continue;
 
+		value = 0;
 		netdev->dcbnl_ops->getpfccfg(netdev, i - DCB_PFC_UP_ATTR_0,
 		                             &value);
 		ret = nla_put_u8(skb, i, value);
@@ -1395,6 +1396,7 @@ static int dcbnl_cee_fill(struct sk_buff *skb, struct net_device *netdev)
 			goto nla_put_failure;
 
 		for (i = DCB_PFC_UP_ATTR_0; i <= DCB_PFC_UP_ATTR_7; i++) {
+			value = 0;
 			ops->getpfccfg(netdev, i - DCB_PFC_UP_ATTR_0, &value);
 			if (nla_put_u8(skb, i, value))
 				goto nla_put_failure;
#1 lore raw
Quoted 17 lines

dcbnl_getpfccfg() and dcbnl_cee_fill() pass the address of an uninitialized stack u8 to the driver's getpfccfg() op and then put it into the netlink reply unconditionally. The op returns void, so the core cannot tell whether the driver wrote the value.

i40e_dcbnl_get_pfc_cfg() and ice_dcbnl_get_pfc_cfg() return without writing *setting unless the DCBX agent is host managed and in CEE mode. Both drivers normally run with firmware managed DCBX (DCB_CAP_DCBX_LLD_MANAGED), and i40e with FW LLDP disabled runs host managed in IEEE mode, so on a DCB capable i40e/ice port the PFC values in DCB_CMD_PFC_GCFG and DCB_CMD_CEE_GET replies, and in the CEE part of RTNLGRP_DCB notifications, come from uninitialized stack. The same u8 is reused for all eight priorities, so this is one byte per reply.

dcb_doit() only requires CAP_NET_ADMIN for RTM_SETDCB, so the GET commands are available to unprivileged users.

Quoted 7 lines

Set the value to 0 before each call, like the PG config path already presets its outputs to DCB_ATTR_VALUE_UNDEFINED. 0 is also what the reply contains today on kernels built with CONFIG_INIT_STACK_ALL_ZERO, so nothing changes for those.

Fixes b94b013eb626 ("ice: Implement DCBNL support")
Fixes 5effa78e7c94 ("i40e: Add netlink callbacks support for software based DCB")

I am not entirely sure, but should this patch add a fixes tag for a driver, when the change is only in the API?

Also, if this dates back to the introduction of DCBNL then I guess its not a regression and maybe a fix for net-next?

Quoted 16 lines
Assisted-by Yaga Agent HackerSec Claude:claude-opus-5-5

---
Found by code inspection, compile-tested only (W=1). I don't have i40e/ice hardware to check the reply. On kernels built with CONFIG_INIT_STACK_ALL_ZERO the byte is already zero, so in practice it only leaks on builds without stack auto-initialization.

 net/dcb/dcbnl.c | 2 ++
 1 file changed, 2 insertions(+)
net/dcb/dcbnl.c +0 -0
index ab71f53128..ad0aebe0ad 100644

net/dcb/dcbnl.c static int dcbnl_getpfccfg(struct net_device *netdev, struct nlmsghdr *nlh, · line 284

Quoted 4 lines
net/dcb/dcbnl.c +1 -0
@@ -281,6 +281,7 @@ static int dcbnl_getpfccfg(struct net_device *netdev, struct nlmsghdr *nlh,
 		if (!getall && !data[i])
 			continue;
 
+		value = 0;

Why noy initialize this at the declaration, like getall is?

net/dcb/dcbnl.c static int dcbnl_cee_fill(struct sk_buff *skb, struct net_device *netdev) · line 1399

Quoted 8 lines
net/dcb/dcbnl.c +1 -0
 		netdev->dcbnl_ops->getpfccfg(netdev, i - DCB_PFC_UP_ATTR_0,
 		                             &value);
 		ret = nla_put_u8(skb, i, value);
@@ -1395,6 +1396,7 @@ static int dcbnl_cee_fill(struct sk_buff *skb, struct net_device *netdev)
 			goto nla_put_failure;
 
 		for (i = DCB_PFC_UP_ATTR_0; i <= DCB_PFC_UP_ATTR_7; i++) {
+			value = 0;

Same.

net/dcb/dcbnl.c static int dcbnl_cee_fill(struct sk_buff *skb, struct net_device *netdev) · line 1402

Quoted 3 lines
net/dcb/dcbnl.c +0 -0
 			ops->getpfccfg(netdev, i - DCB_PFC_UP_ATTR_0, &value);
 			if (nla_put_u8(skb, i, value))
 				goto nla_put_failure;

/Daniel

#2 lore raw
Quoted 2 lines · Daniel Machon

I am not entirely sure, but should this patch add a fixes tag for a driver, when the change is only in the API?

Fair point. The uninitialized u8 is in dcbnl, the drivers only make it visible, so the driver commits are not the right Fixes. I'll drop them.

Quoted 2 lines

Also, if this dates back to the introduction of DCBNL then I guess its not a regression and maybe a fix for net-next?

Agreed. I'll send v2 for net-next once it reopens.

Quoted 1 line

Why noy initialize this at the declaration, like getall is?

That works. In both functions the PFC loop is the first user of value, so initializing it at the declaration is enough. I'll do that in v2.

Thanks for the review,
Joas

#3 lore raw
In reply to Joas Antonio dos Santos
This patch was applied to netdev/net.git (main) by Jakub Kicinski <[email protected]>:

Hello:

This patch was applied to netdev/net.git (main)
by Jakub Kicinski <[email protected]>:

Quoted 14 lines · On Sat, 24 Oct 2026 22:47:24 -0300 you wrote:

dcbnl_getpfccfg() and dcbnl_cee_fill() pass the address of an uninitialized stack u8 to the driver's getpfccfg() op and then put it into the netlink reply unconditionally. The op returns void, so the core cannot tell whether the driver wrote the value.

i40e_dcbnl_get_pfc_cfg() and ice_dcbnl_get_pfc_cfg() return without writing *setting unless the DCBX agent is host managed and in CEE mode. Both drivers normally run with firmware managed DCBX (DCB_CAP_DCBX_LLD_MANAGED), and i40e with FW LLDP disabled runs host managed in IEEE mode, so on a DCB capable i40e/ice port the PFC values in DCB_CMD_PFC_GCFG and DCB_CMD_CEE_GET replies, and in the CEE part of RTNLGRP_DCB notifications, come from uninitialized stack. The same u8 is reused for all eight priorities, so this is one byte per reply.

Quoted 1 line

[...]

Here is the summary with links:
- [net] net: dcb: initialize PFC value before calling getpfccfg

    https://git.kernel.org/netdev/net/c/ca87a2dffbb5

You are awesome, thank you!

Signature · 2 lines
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html