[PATCH net 0/2] airoha: fixes for sched HTB offload support

[PATCH net 0/2] airoha: fixes for sched HTB offload support

[Lorenzo Bianconi]

---
Lorenzo Bianconi (2):
      net: airoha: Fix off-by-one in airoha_tc_remove_htb_queue()
      net: airoha: fix netif_set_real_num_tx_queues for sparse QoS channels

 drivers/net/ethernet/airoha/airoha_eth.c | 17 +++++++++++++----
 1 file changed, 13 insertions(+), 4 deletions(-)
---
base-commit: 7d8297e26b4e20b5d1c3c3fe51fe81a1c7fbc823
change-id: 20260618-airoha-qos-fixes-b6460b085680

Best regards,
-- 
Lorenzo Bianconi <lorenzo@kernel.org>

[PATCH net 1/2] net: airoha: Fix off-by-one in airoha_tc_remove_htb_queue()

[Lorenzo Bianconi]

airoha_tc_htb_alloc_leaf_queue() computes the HTB QoS channel index
as opt->classid % AIROHA_NUM_QOS_CHANNELS and stores it in qos_sq_bmap.
However, airoha_tc_remove_htb_queue() clears the HTB configuration
using queue + 1 as the channel index, causing an off-by-one error.
Use queue directly as the QoS channel index to match the allocation
logic.

Fixes: ef1ca9271313b ("net: airoha: Add sched HTB offload support")
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
---
 drivers/net/ethernet/airoha/airoha_eth.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/net/ethernet/airoha/airoha_eth.c b/drivers/net/ethernet/airoha/airoha_eth.c
index 64dde6464f3f..aa98d1823ab6 100644
--- a/drivers/net/ethernet/airoha/airoha_eth.c
+++ b/drivers/net/ethernet/airoha/airoha_eth.c
@@ -3006,7 +3006,7 @@ static void airoha_tc_remove_htb_queue(struct net_device *netdev, int queue)
 	struct airoha_qdma *qdma = dev->qdma;
 
 	netif_set_real_num_tx_queues(netdev, netdev->real_num_tx_queues - 1);
-	airoha_qdma_set_tx_rate_limit(netdev, queue + 1, 0, 0);
+	airoha_qdma_set_tx_rate_limit(netdev, queue, 0, 0);
 
 	clear_bit(queue, qdma->qos_channel_map);
 	clear_bit(queue, dev->qos_sq_bmap);

-- 
2.54.0

[PATCH net 2/2] net: airoha: fix netif_set_real_num_tx_queues for sparse QoS channels

[Lorenzo Bianconi]

airoha_tc_htb_alloc_leaf_queue() assigns queue IDs based on the channel
index (opt->qid = AIROHA_NUM_TX_RING + channel), but updates
real_num_tx_queues with a simple increment (num_tx_queues + 1). When QoS
channels are allocated sparsely (e.g., channels 0 and 3 without 1 and
2), the returned qid can exceed real_num_tx_queues, causing out-of-bounds
accesses in the networking stack.
For example, allocating channel 0 then channel 3 results in
real_num_tx_queues = 34 but qid = 35, which is out of range [0, 34).
Fix this by computing real_num_tx_queues based on the highest active
channel index rather than using a simple counter, in both the allocation
and deletion paths.

Fixes: ef1ca9271313b ("net: airoha: Add sched HTB offload support")
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
---
 drivers/net/ethernet/airoha/airoha_eth.c | 15 ++++++++++++---
 1 file changed, 12 insertions(+), 3 deletions(-)

diff --git a/drivers/net/ethernet/airoha/airoha_eth.c b/drivers/net/ethernet/airoha/airoha_eth.c
index aa98d1823ab6..e2652cff67c0 100644
--- a/drivers/net/ethernet/airoha/airoha_eth.c
+++ b/drivers/net/ethernet/airoha/airoha_eth.c
@@ -2789,7 +2789,7 @@ static int airoha_tc_htb_alloc_leaf_queue(struct net_device *netdev,
 					  struct tc_htb_qopt_offload *opt)
 {
 	u32 channel = TC_H_MIN(opt->classid) % AIROHA_NUM_QOS_CHANNELS;
-	int err, num_tx_queues = netdev->real_num_tx_queues;
+	int err, num_tx_queues = AIROHA_NUM_TX_RING + channel + 1;
 	struct airoha_gdm_dev *dev = netdev_priv(netdev);
 	struct airoha_qdma *qdma = dev->qdma;
 
@@ -2806,7 +2806,10 @@ static int airoha_tc_htb_alloc_leaf_queue(struct net_device *netdev,
 	if (err)
 		goto error;
 
-	err = netif_set_real_num_tx_queues(netdev, num_tx_queues + 1);
+	if (num_tx_queues <= netdev->real_num_tx_queues)
+		goto set_qos_sq_bmap;
+
+	err = netif_set_real_num_tx_queues(netdev, num_tx_queues);
 	if (err) {
 		airoha_qdma_set_tx_rate_limit(netdev, channel, 0,
 					      opt->quantum);
@@ -2815,6 +2818,7 @@ static int airoha_tc_htb_alloc_leaf_queue(struct net_device *netdev,
 		goto error;
 	}
 
+set_qos_sq_bmap:
 	set_bit(channel, dev->qos_sq_bmap);
 	opt->qid = AIROHA_NUM_TX_RING + channel;
 
@@ -3003,13 +3007,18 @@ static int airoha_dev_setup_tc_block(struct net_device *dev,
 static void airoha_tc_remove_htb_queue(struct net_device *netdev, int queue)
 {
 	struct airoha_gdm_dev *dev = netdev_priv(netdev);
+	int num_tx_queues = AIROHA_NUM_TX_RING;
 	struct airoha_qdma *qdma = dev->qdma;
 
-	netif_set_real_num_tx_queues(netdev, netdev->real_num_tx_queues - 1);
 	airoha_qdma_set_tx_rate_limit(netdev, queue, 0, 0);
 
 	clear_bit(queue, qdma->qos_channel_map);
 	clear_bit(queue, dev->qos_sq_bmap);
+
+	if (!bitmap_empty(dev->qos_sq_bmap, AIROHA_NUM_QOS_CHANNELS))
+		num_tx_queues += find_last_bit(dev->qos_sq_bmap,
+					       AIROHA_NUM_QOS_CHANNELS) + 1;
+	netif_set_real_num_tx_queues(netdev, num_tx_queues);
 }
 
 static int airoha_tc_htb_delete_leaf_queue(struct net_device *netdev,

-- 
2.54.0

Re: [PATCH net 1/2] net: airoha: Fix off-by-one in airoha_tc_remove_htb_queue()

[Simon Horman]

On Thu, Jun 18, 2026 at 08:00:29AM +0200, Lorenzo Bianconi wrote:
> airoha_tc_htb_alloc_leaf_queue() computes the HTB QoS channel index
> as opt->classid % AIROHA_NUM_QOS_CHANNELS and stores it in qos_sq_bmap.
> However, airoha_tc_remove_htb_queue() clears the HTB configuration
> using queue + 1 as the channel index, causing an off-by-one error.
> Use queue directly as the QoS channel index to match the allocation
> logic.
> 
> Fixes: ef1ca9271313b ("net: airoha: Add sched HTB offload support")
> Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>

Reviewed-by: Simon Horman <horms@kernel.org>

Re: [PATCH net 2/2] net: airoha: fix netif_set_real_num_tx_queues for sparse QoS channels

[Simon Horman]

On Thu, Jun 18, 2026 at 08:00:30AM +0200, Lorenzo Bianconi wrote:
> airoha_tc_htb_alloc_leaf_queue() assigns queue IDs based on the channel
> index (opt->qid = AIROHA_NUM_TX_RING + channel), but updates
> real_num_tx_queues with a simple increment (num_tx_queues + 1). When QoS
> channels are allocated sparsely (e.g., channels 0 and 3 without 1 and
> 2), the returned qid can exceed real_num_tx_queues, causing out-of-bounds
> accesses in the networking stack.
> For example, allocating channel 0 then channel 3 results in
> real_num_tx_queues = 34 but qid = 35, which is out of range [0, 34).
> Fix this by computing real_num_tx_queues based on the highest active
> channel index rather than using a simple counter, in both the allocation
> and deletion paths.
> 
> Fixes: ef1ca9271313b ("net: airoha: Add sched HTB offload support")
> Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
> ---
>  drivers/net/ethernet/airoha/airoha_eth.c | 15 ++++++++++++---
>  1 file changed, 12 insertions(+), 3 deletions(-)
> 
> diff --git a/drivers/net/ethernet/airoha/airoha_eth.c b/drivers/net/ethernet/airoha/airoha_eth.c

...

> @@ -2806,7 +2806,10 @@ static int airoha_tc_htb_alloc_leaf_queue(struct net_device *netdev,
>  	if (err)
>  		goto error;
>  
> -	err = netif_set_real_num_tx_queues(netdev, num_tx_queues + 1);
> +	if (num_tx_queues <= netdev->real_num_tx_queues)
> +		goto set_qos_sq_bmap;
> +
> +	err = netif_set_real_num_tx_queues(netdev, num_tx_queues);
>  	if (err) {
>  		airoha_qdma_set_tx_rate_limit(netdev, channel, 0,
>  					      opt->quantum);
> @@ -2815,6 +2818,7 @@ static int airoha_tc_htb_alloc_leaf_queue(struct net_device *netdev,
>  		goto error;
>  	}
>  
> +set_qos_sq_bmap:

I would prefer if this could be achieved without a goto.

>  	set_bit(channel, dev->qos_sq_bmap);
>  	opt->qid = AIROHA_NUM_TX_RING + channel;
>  

...

Re: [PATCH net 2/2] net: airoha: fix netif_set_real_num_tx_queues for sparse QoS channels

[Lorenzo Bianconi]

[-- Attachment #1: Type: text/plain, Size: 2052 bytes --]

> On Thu, Jun 18, 2026 at 08:00:30AM +0200, Lorenzo Bianconi wrote:
> > airoha_tc_htb_alloc_leaf_queue() assigns queue IDs based on the channel
> > index (opt->qid = AIROHA_NUM_TX_RING + channel), but updates
> > real_num_tx_queues with a simple increment (num_tx_queues + 1). When QoS
> > channels are allocated sparsely (e.g., channels 0 and 3 without 1 and
> > 2), the returned qid can exceed real_num_tx_queues, causing out-of-bounds
> > accesses in the networking stack.
> > For example, allocating channel 0 then channel 3 results in
> > real_num_tx_queues = 34 but qid = 35, which is out of range [0, 34).
> > Fix this by computing real_num_tx_queues based on the highest active
> > channel index rather than using a simple counter, in both the allocation
> > and deletion paths.
> > 
> > Fixes: ef1ca9271313b ("net: airoha: Add sched HTB offload support")
> > Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
> > ---
> >  drivers/net/ethernet/airoha/airoha_eth.c | 15 ++++++++++++---
> >  1 file changed, 12 insertions(+), 3 deletions(-)
> > 
> > diff --git a/drivers/net/ethernet/airoha/airoha_eth.c b/drivers/net/ethernet/airoha/airoha_eth.c
> 
> ...
> 
> > @@ -2806,7 +2806,10 @@ static int airoha_tc_htb_alloc_leaf_queue(struct net_device *netdev,
> >  	if (err)
> >  		goto error;
> >  
> > -	err = netif_set_real_num_tx_queues(netdev, num_tx_queues + 1);
> > +	if (num_tx_queues <= netdev->real_num_tx_queues)
> > +		goto set_qos_sq_bmap;
> > +
> > +	err = netif_set_real_num_tx_queues(netdev, num_tx_queues);
> >  	if (err) {
> >  		airoha_qdma_set_tx_rate_limit(netdev, channel, 0,
> >  					      opt->quantum);
> > @@ -2815,6 +2818,7 @@ static int airoha_tc_htb_alloc_leaf_queue(struct net_device *netdev,
> >  		goto error;
> >  	}
> >  
> > +set_qos_sq_bmap:
> 
> I would prefer if this could be achieved without a goto.

ack, I will fix it in v2.

Regards,
Lorenzo

> 
> >  	set_bit(channel, dev->qos_sq_bmap);
> >  	opt->qid = AIROHA_NUM_TX_RING + channel;
> >  
> 
> ...

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 228 bytes --]