Skip to contentSkip to conversation

brcmfmac: NULL pointer dereference on tx statistic update

PATCH In review Marcel Hamer 3 participants 6 messages Jan 10, 2025 – Jan 15, 2025

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

PATCH v2 1/1 brcmfmac: NULL pointer dereference on tx statistic update

On removal of the device or unloading of the kernel module a potential NULL pointer dereference occurs.

The following sequence deletes the interface:

brcmf_detach()

    brcmf_remove_interface()
      brcmf_del_if()

Inside the brcmf_del_if() function the drvr->if2bss[ifidx] is updated to BRCMF_BSSIDX_INVALID (-1) if the bsscfgidx matches.

After brcmf_remove_interface() call the brcmf_proto_detach() function is called providing the following sequence:

brcmf_detach()

    brcmf_proto_detach()
      brcmf_proto_msgbuf_detach()
        brcmf_flowring_detach()
          brcmf_msgbuf_delete_flowring()
            brcmf_msgbuf_remove_flowring()
              brcmf_flowring_delete()
                brcmf_get_ifp()
                brcmf_txfinalize()

Since brcmf_get_ip() can and actually will return NULL in this case the call to brcmf_txfinalize() will result in a NULL pointer dereference inside brcmf_txfinalize() when trying to update ifp->ndev->stats.tx_errors.

This will only happen if a flowring still has an skb.

Although the NULL pointer dereference has only been seen when trying to update the tx statistic, all other uses of the ifp pointer have been guarded as well.

---
v1 -> v2: guard all uses of the ifp pointer inside brcmf_txfinalize()

---
 drivers/net/wireless/broadcom/brcm80211/brcmfmac/core.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/net/wireless/broadcom/brcm80211/brcmfmac/core.c +2 -2
index c3a57e30c855..791757a3ec13 100644
@@ -543,13 +543,13 @@ void brcmf_txfinalize(struct brcmf_if *ifp, struct sk_buff *txp, bool success)
 	eh = (struct ethhdr *)(txp->data);
 	type = ntohs(eh->h_proto);
 
-	if (type == ETH_P_PAE) {
+	if (type == ETH_P_PAE && ifp) {
 		atomic_dec(&ifp->pend_8021x_cnt);
 		if (waitqueue_active(&ifp->pend_8021x_wait))
 			wake_up(&ifp->pend_8021x_wait);
 	}
 
-	if (!success)
+	if (!success && ifp)
 		ifp->ndev->stats.tx_errors++;
 
 	brcmu_pkt_buf_free_skb(txp);
Signature · 1 line
2.34.1
maintainer
#1 lore raw
Quoted 30 lines · Marcel Hamer

On removal of the device or unloading of the kernel module a potential NULL pointer dereference occurs.

The following sequence deletes the interface:

   brcmf_detach()
     brcmf_remove_interface()
       brcmf_del_if()

Inside the brcmf_del_if() function the drvr->if2bss[ifidx] is updated to BRCMF_BSSIDX_INVALID (-1) if the bsscfgidx matches.

After brcmf_remove_interface() call the brcmf_proto_detach() function is called providing the following sequence:

   brcmf_detach()
     brcmf_proto_detach()
       brcmf_proto_msgbuf_detach()
         brcmf_flowring_detach()
           brcmf_msgbuf_delete_flowring()
             brcmf_msgbuf_remove_flowring()
               brcmf_flowring_delete()
                 brcmf_get_ifp()
                 brcmf_txfinalize()

Since brcmf_get_ip() can and actually will return NULL in this case the call to brcmf_txfinalize() will result in a NULL pointer dereference inside brcmf_txfinalize() when trying to update ifp->ndev->stats.tx_errors.

Quoted 4 lines

This will only happen if a flowring still has an skb.

Although the NULL pointer dereference has only been seen when trying to update the tx statistic, all other uses of the ifp pointer have been guarded as well.

Here my suggestion to make it a bit more simple...

drivers/net/wireless/broadcom/brcm80211/brcmfmac/core.c void brcmf_txfinalize(struct brcmf_if *ifp, struct sk_buff *txp, bool success)

Quoted 14 lines

---
v1 -> v2: guard all uses of the ifp pointer inside brcmf_txfinalize()

---
  drivers/net/wireless/broadcom/brcm80211/brcmfmac/core.c | 4 ++--
  1 file changed, 2 insertions(+), 2 deletions(-)
drivers/net/wireless/broadcom/brcm80211/brcmfmac/core.c +0 -0
index c3a57e30c855..791757a3ec13 100644
@@ -543,13 +543,13 @@ void brcmf_txfinalize(struct brcmf_if *ifp, struct sk_buff *txp, bool success)

Instead of checking ifp below you can simply do following here and be done with it:

	if (!ifp) {
		brcmu_pkt_buf_free_skb(txp);
		return;
	}
Quoted 15 lines
drivers/net/wireless/broadcom/brcm80211/brcmfmac/core.c +2 -2
  	eh = (struct ethhdr *)(txp->data);
  	type = ntohs(eh->h_proto);
 
-	if (type == ETH_P_PAE) {
+	if (type == ETH_P_PAE && ifp) {
  		atomic_dec(&ifp->pend_8021x_cnt);
  		if (waitqueue_active(&ifp->pend_8021x_wait))
  			wake_up(&ifp->pend_8021x_wait);
  	}
 
-	if (!success)
+	if (!success && ifp)
  		ifp->ndev->stats.tx_errors++;
 
  	brcmu_pkt_buf_free_skb(txp);
#2 lore raw

Hello,

Quoted 59 lines · Arend van Spriel

CAUTION: This email comes from a non Wind River email account!
Do not click links or open attachments unless you recognize the sender and know the content is safe.

Quoted 30 lines · Marcel Hamer

On removal of the device or unloading of the kernel module a potential NULL pointer dereference occurs.

The following sequence deletes the interface:

   brcmf_detach()
     brcmf_remove_interface()
       brcmf_del_if()

Inside the brcmf_del_if() function the drvr->if2bss[ifidx] is updated to BRCMF_BSSIDX_INVALID (-1) if the bsscfgidx matches.

After brcmf_remove_interface() call the brcmf_proto_detach() function is called providing the following sequence:

   brcmf_detach()
     brcmf_proto_detach()
       brcmf_proto_msgbuf_detach()
         brcmf_flowring_detach()
           brcmf_msgbuf_delete_flowring()
             brcmf_msgbuf_remove_flowring()
               brcmf_flowring_delete()
                 brcmf_get_ifp()
                 brcmf_txfinalize()

Since brcmf_get_ip() can and actually will return NULL in this case the call to brcmf_txfinalize() will result in a NULL pointer dereference inside brcmf_txfinalize() when trying to update ifp->ndev->stats.tx_errors.

Quoted 4 lines

This will only happen if a flowring still has an skb.

Although the NULL pointer dereference has only been seen when trying to update the tx statistic, all other uses of the ifp pointer have been guarded as well.

Here my suggestion to make it a bit more simple...

drivers/net/wireless/broadcom/brcm80211/brcmfmac/core.c void brcmf_txfinalize(struct brcmf_if *ifp, struct sk_buff *txp, bool success)

Quoted 14 lines

---
v1 -> v2: guard all uses of the ifp pointer inside brcmf_txfinalize()

---
  drivers/net/wireless/broadcom/brcm80211/brcmfmac/core.c | 4 ++--
  1 file changed, 2 insertions(+), 2 deletions(-)
drivers/net/wireless/broadcom/brcm80211/brcmfmac/core.c +0 -0
index c3a57e30c855..791757a3ec13 100644
@@ -543,13 +543,13 @@ void brcmf_txfinalize(struct brcmf_if *ifp, struct sk_buff *txp, bool success)

Instead of checking ifp below you can simply do following here and be done with it:

Quoted 5 lines
       if (!ifp) {
               brcmu_pkt_buf_free_skb(txp);
               return;
       }

Thank you for the suggestion and review. I considered that when doing my changes and then thought it would be more efficient to only check the validity of ifp when ifp is actually used. To skip checking it on each call to this function, since it seems unlikely ifp will be used on each and every call.

Before I create a patch with your suggested changes, I just wanted to check if you think it would be better to just wrap the two existing if statements in a new if statement in that case, something like this:

if (ifp) {
    ...
    if (type == ETH_P_PAE) {
    ...
    }

    if (!success)
        ifp->ndev->stats.tx_errors++;
}

And then keep a single call to brcmu_pkt_buf_free_skb(txp) at the end of the function.

Quoted 10 lines
Quoted 10 lines
drivers/net/wireless/broadcom/brcm80211/brcmfmac/core.c +1 -1
      eh = (struct ethhdr *)(txp->data);
      type = ntohs(eh->h_proto);
 
-     if (type == ETH_P_PAE) {
+     if (type == ETH_P_PAE && ifp) {
              atomic_dec(&ifp->pend_8021x_cnt);
              if (waitqueue_active(&ifp->pend_8021x_wait))
                      wake_up(&ifp->pend_8021x_wait);
      }
 
Quoted 6 lines

drivers/net/wireless/broadcom/brcm80211/brcmfmac/core.c void brcmf_txfinalize(struct brcmf_if *ifp, struct sk_buff *txp, bool success) · line 555

Quoted 5 lines
drivers/net/wireless/broadcom/brcm80211/brcmfmac/core.c +1 -1
-     if (!success)
+     if (!success && ifp)
              ifp->ndev->stats.tx_errors++;
 
      brcmu_pkt_buf_free_skb(txp);

Kind regards,

Marcel

maintainer
#3 lore raw
Quoted 68 lines · Marcel Hamer

Hello,

Quoted 60 lines · Arend van Spriel

CAUTION: This email comes from a non Wind River email account!
Do not click links or open attachments unless you recognize the sender and know the content is safe.

Quoted 30 lines · Marcel Hamer

On removal of the device or unloading of the kernel module a potential NULL pointer dereference occurs.

The following sequence deletes the interface:

brcmf_detach()
brcmf_remove_interface()
brcmf_del_if()

Inside the brcmf_del_if() function the drvr->if2bss[ifidx] is updated to BRCMF_BSSIDX_INVALID (-1) if the bsscfgidx matches.

After brcmf_remove_interface() call the brcmf_proto_detach() function is called providing the following sequence:

brcmf_detach()
brcmf_proto_detach()
brcmf_proto_msgbuf_detach()
brcmf_flowring_detach()

    brcmf_msgbuf_delete_flowring()
      brcmf_msgbuf_remove_flowring()
        brcmf_flowring_delete()
          brcmf_get_ifp()
          brcmf_txfinalize()

Since brcmf_get_ip() can and actually will return NULL in this case the call to brcmf_txfinalize() will result in a NULL pointer dereference inside brcmf_txfinalize() when trying to update ifp->ndev->stats.tx_errors.

Quoted 4 lines

This will only happen if a flowring still has an skb.

Although the NULL pointer dereference has only been seen when trying to update the tx statistic, all other uses of the ifp pointer have been guarded as well.

Here my suggestion to make it a bit more simple...

Quoted 14 lines

---
v1 -> v2: guard all uses of the ifp pointer inside brcmf_txfinalize()
---
drivers/net/wireless/broadcom/brcm80211/brcmfmac/core.c | 4 ++--

1 file changed, 2 insertions(+), 2 deletions(-)
drivers/net/wireless/broadcom/brcm80211/brcmfmac/core.c +0 -0
index c3a57e30c855..791757a3ec13 100644

drivers/net/wireless/broadcom/brcm80211/brcmfmac/core.c void brcmf_txfinalize(struct brcmf_if *ifp, struct · line 543

Quoted 1 line
drivers/net/wireless/broadcom/brcm80211/brcmfmac/core.c +0 -0
@@ -543,13 +543,13 @@ void brcmf_txfinalize(struct brcmf_if *ifp, struct 
 sk_buff *txp, bool success)

Instead of checking ifp below you can simply do following here and be done with it:

Quoted 4 lines
if (!ifp) {
      brcmu_pkt_buf_free_skb(txp);
      return;
}
Quoted 4 lines

Thank you for the suggestion and review. I considered that when doing my changes and then thought it would be more efficient to only check the validity of ifp when ifp is actually used. To skip checking it on each call to this function, since it seems unlikely ifp will be used on each and every call.

Actually, the function will mostly be called with a valid ifp instance.

Quoted 14 lines

Before I create a patch with your suggested changes, I just wanted to check if you think it would be better to just wrap the two existing if statements in a new if statement in that case, something like this:

if (ifp) {
   ...
   if (type == ETH_P_PAE) {
   ...
   }

   if (!success)
       ifp->ndev->stats.tx_errors++;
}
Quoted 2 lines

And then keep a single call to brcmu_pkt_buf_free_skb(txp) at the end of the function.

I understand your motivation. My generic approach is to bail out early and thereby avoid extra indentation levels. Although this function is in the data path I think the null-check for ifp is negligible. Could consider using unlikely(), but that requires profiling data to justify using it here.

Regards,
Arend

#4 lore raw
In reply to Marcel Hamer
Quoted 30 lines · Marcel Hamer

On removal of the device or unloading of the kernel module a potential NULL pointer dereference occurs.

The following sequence deletes the interface:

brcmf_detach()

    brcmf_remove_interface()
      brcmf_del_if()

Inside the brcmf_del_if() function the drvr->if2bss[ifidx] is updated to BRCMF_BSSIDX_INVALID (-1) if the bsscfgidx matches.

After brcmf_remove_interface() call the brcmf_proto_detach() function is called providing the following sequence:

brcmf_detach()

    brcmf_proto_detach()
      brcmf_proto_msgbuf_detach()
        brcmf_flowring_detach()
          brcmf_msgbuf_delete_flowring()
            brcmf_msgbuf_remove_flowring()
              brcmf_flowring_delete()
                brcmf_get_ifp()
                brcmf_txfinalize()

Since brcmf_get_ip() can and actually will return NULL in this case the call to brcmf_txfinalize() will result in a NULL pointer dereference inside brcmf_txfinalize() when trying to update ifp->ndev->stats.tx_errors.

Quoted 8 lines

This will only happen if a flowring still has an skb.

Although the NULL pointer dereference has only been seen when trying to update the tx statistic, all other uses of the ifp pointer have been guarded as well.

If you submit v3, please add 'wifi:'.

ERROR: 'wifi:' prefix missing: '[PATCH v2] brcmfmac: NULL pointer dereference on tx statistic update'

Patch set to Changes Requested.

Signature · 3 lines
https://patchwork.kernel.org/project/linux-wireless/patch/[email protected]/

https://wireless.wiki.kernel.org/en/developers/documentation/submittingpatches
maintainer
#5 lore raw
Quoted 37 lines · Kalle Valo
Quoted 35 lines · Marcel Hamer

On removal of the device or unloading of the kernel module a potential NULL pointer dereference occurs.

The following sequence deletes the interface:

brcmf_detach()
brcmf_remove_interface()
brcmf_del_if()

Inside the brcmf_del_if() function the drvr->if2bss[ifidx] is updated to BRCMF_BSSIDX_INVALID (-1) if the bsscfgidx matches.

After brcmf_remove_interface() call the brcmf_proto_detach() function is called providing the following sequence:

brcmf_detach()
brcmf_proto_detach()
brcmf_proto_msgbuf_detach()
brcmf_flowring_detach()
brcmf_msgbuf_delete_flowring()
brcmf_msgbuf_remove_flowring()

     brcmf_flowring_delete()
       brcmf_get_ifp()
       brcmf_txfinalize()

Since brcmf_get_ip() can and actually will return NULL in this case the call to brcmf_txfinalize() will result in a NULL pointer dereference inside brcmf_txfinalize() when trying to update ifp->ndev->stats.tx_errors.

This will only happen if a flowring still has an skb.

Although the NULL pointer dereference has only been seen when trying to update the tx statistic, all other uses of the ifp pointer have been guarded as well.

Quoted 7 lines
Quoted 3 lines

If you submit v3, please add 'wifi:'.

ERROR: 'wifi:' prefix missing: '[PATCH v2] brcmfmac: NULL pointer dereference on tx statistic update'

While at it maybe rephrase the subject to:

wifi: brcmfmac: fix NULL pointer dereference in brcmf_txfinalize()

Regards,
Arend