drivers/net/wan/x25_asy: Fix to make it work
Subsystems (3)
PATCH v2 1/1 drivers/net/wan/x25_asy: Fix to make it work
This driver is not working because of problems of its receiving code. This patch fixes it to make it work.
When the driver receives an LAPB frame, it should first pass the frame to the LAPB module to process. After processing, the LAPB module passes the data (the packet) back to the driver, the driver should then add a one-byte pseudo header and pass the data to upper layers.
The changes to the "x25_asy_bump" function and the "x25_asy_data_indication" function are to correctly implement this procedure.
Also, the "x25_asy_unesc" function ignores any frame that is shorter than 3 bytes. However the shortest frames are 2-byte long. So we need to change it to allow 2-byte frames to pass.
---
Change from v1:
Added skb_cow before skb_push to ensure skb_push will succeed according the suggestion of Eric Dumazet.
Hi Eric Dumazet and Martin Schiller,
Can you review this patch again and see if it is OK for me to include your names in a "Signed-off-by", "Reviewed-by" or "Acked-by" tag? Thank you!
Hi All,
I'm happy to answer any questions you might have and make improvements according to your suggestions. Thanks!
--- drivers/net/wan/x25_asy.c | 21 ++++++++++++++------- 1 file changed, 14 insertions(+), 7 deletions(-)
drivers/net/wan/x25_asy.c +14 -7
@@ -183,7 +183,7 @@ static inline void x25_asy_unlock(struct x25_asy *sl)netif_wake_queue (sl->dev); } -/* Send one completely decapsulated IP datagram to the IP layer. */ +/* Send an LAPB frame to the LAPB module to process. */ static void x25_asy_bump (struct x25_asy *sl) {
@@ -195,13 +195,12 @@ static void x25_asy_bump(struct x25_asy *sl) count = sl->rcount; dev->stats.rx_bytes += count; - skb =dev_alloc_skb (count+1 ); + skb =dev_alloc_skb (count);if (skb ==NULL ) {netdev_warn (sl->dev,"memory squeeze, dropping packet\n" ); dev->stats.rx_dropped++;return ; } -skb_push (skb,1 );/* LAPB internal control */ skb_put_data (skb, sl->rbuff, count); skb->protocol =x25_type_trans (skb, sl->dev); err =lapb_data_received (skb->dev, skb);
@@ -209,7 +208,6 @@ static void x25_asy_bump(struct x25_asy *sl)kfree_skb (skb);printk (KERN_DEBUG "x25_asy: data received err - %d\n" , err); }else { -netif_rx (skb); dev->stats.rx_packets++; } }
@@ -356,12 +354,21 @@ static netdev_tx_t x25_asy_xmit(struct sk_buff *skb,*/ /* -* Called when I frame data arrives. We did the work above - throw it -* at the net layer. +* Called when I frame data arrive. We add a pseudo header for upper +* layers and pass it to upper layers. */ static int x25_asy_data_indication (struct net_device *dev,struct sk_buff *skb) { +if (skb_cow (skb,1 )) { +kfree_skb (skb); +return NET_RX_DROP ; + } +skb_push (skb,1 ); + skb->data[0 ] =X25_IFACE_DATA ; + + skb->protocol =x25_type_trans (skb, dev); +return netif_rx (skb); }
@@ -657,7 +664,7 @@ static void x25_asy_unesc(struct x25_asy *sl, unsigned char s)switch (s) {case X25_END :if (!test_and_clear_bit (SLF_ERROR , &sl->flags) && - sl->rcount >2 ) + sl->rcount >=2 )x25_asy_bump (sl);clear_bit (SLF_ESCAPE , &sl->flags); sl->rcount =0 ;
Signature · 1 line
2.25.1
Quoted 110 lines · Xie He
This driver is not working because of problems of its receiving code. This patch fixes it to make it work.
When the driver receives an LAPB frame, it should first pass the frame to the LAPB module to process. After processing, the LAPB module passes the data (the packet) back to the driver, the driver should then add a one-byte pseudo header and pass the data to upper layers.
The changes to the "
x25_asy_bump" function and the "x25_asy_data_indication" function are to correctly implement this procedure.Also, the "
x25_asy_unesc" function ignores any frame that is shorter than 3 bytes. However the shortest frames are 2-byte long. So we need to change it to allow 2-byte frames to pass.Cc Eric Dumazet <[email protected]>Cc Martin Schiller <[email protected]>Signed-off-by Xie He <[email protected]>---
Change from v1:
Addedskb_cowbeforeskb_pushto ensureskb_pushwill succeed according the suggestion of Eric Dumazet.Hi Eric Dumazet and Martin Schiller,
Can you review this patch again and see if it is OK for me to include your names in a "Signed-off-by", "Reviewed-by" or "Acked-by" tag? Thank you!Hi All,
I'm happy to answer any questions you might have and make improvements according to your suggestions. Thanks!--- drivers/net/wan/x25_asy.c | 21 ++++++++++++++------- 1 file changed, 14 insertions(+), 7 deletions(-)drivers/net/wan/x25_asy.c +1
-1@@ -183,7 +183,7 @@ static inline void x25_asy_unlock(struct x25_asy *sl)netif_wake_queue (sl->dev); } -/* Send one completely decapsulated IP datagram to the IP layer. */ +/* Send an LAPB frame to the LAPB module to process. */ static void x25_asy_bump (struct x25_asy *sl){
drivers/net/wan/x25_asy.c +12
-5@@ -195,13 +195,12 @@ static void x25_asy_bump(struct x25_asy *sl) count = sl->rcount; dev->stats.rx_bytes += count; - skb =dev_alloc_skb (count+1 ); + skb =dev_alloc_skb (count);if (skb ==NULL ) {netdev_warn (sl->dev,"memory squeeze, dropping packet\n" ); dev->stats.rx_dropped++;return ; } -skb_push (skb,1 );/* LAPB internal control */ skb_put_data (skb, sl->rbuff, count); skb->protocol =x25_type_trans (skb, sl->dev); err =lapb_data_received (skb->dev, skb);@@ -209,7 +208,6 @@ static void x25_asy_bump(struct x25_asy *sl)kfree_skb (skb);printk (KERN_DEBUG "x25_asy: data received err - %d\n" , err); }else { -netif_rx (skb); dev->stats.rx_packets++; } }@@ -356,12 +354,21 @@ static netdev_tx_t x25_asy_xmit(struct sk_buff*skb, */ /* -* Called when I frame data arrives. We did the work above - throw it -* at the net layer. +* Called when I frame data arrive. We add a pseudo header for upper +* layers and pass it to upper layers. */ static int x25_asy_data_indication (struct net_device *dev,struct sk_buff *skb) { +if (skb_cow (skb,1 )) { +kfree_skb (skb); +return NET_RX_DROP ; + } +skb_push (skb,1 ); + skb->data[0 ] =X25_IFACE_DATA ; + + skb->protocol =x25_type_trans (skb, dev); +return netif_rx (skb);}
drivers/net/wan/x25_asy.c +1
-1@@ -657,7 +664,7 @@ static void x25_asy_unesc(struct x25_asy *sl, unsigned char s) switch (s) {case X25_END :if (!test_and_clear_bit (SLF_ERROR , &sl->flags) && - sl->rcount >2 ) + sl->rcount >=2 )x25_asy_bump (sl);clear_bit (SLF_ESCAPE , &sl->flags);
Quoted 1 line
sl->rcount= 0;
LGTM.
I have never used the driver, but the adjustments look conclusive. The functionality is now comparable to the one in the drivers lapbether or hdlc_x25.
Quoted 9 lines · On Mon, Jul 20, 2020 at 4:23 AM -0700 Martin Schiller <[email protected]> wrote:
LGTM.
I have never used the driver, but the adjustments look conclusive. The functionality is now comparable to the one in the drivers lapbether or
hdlc_x25.Reviewed-by Martin Schiller <[email protected]>
Thank you so much!
From: Xie He <[email protected]>
Date: Thu, 16 Jul 2020 16:44:33 -0700
Quoted 12 lines
This driver is not working because of problems of its receiving code. This patch fixes it to make it work.
When the driver receives an LAPB frame, it should first pass the frame to the LAPB module to process. After processing, the LAPB module passes the data (the packet) back to the driver, the driver should then add a one-byte pseudo header and pass the data to upper layers.
The changes to the "
x25_asy_bump" function and the "x25_asy_data_indication" function are to correctly implement this procedure.
Quoted 7 lines
Also, the "
x25_asy_unesc" function ignores any frame that is shorter than 3 bytes. However the shortest frames are 2-byte long. So we need to change it to allow 2-byte frames to pass.Cc Eric Dumazet <[email protected]>Cc Martin Schiller <[email protected]>Signed-off-by Xie He <[email protected]>
Applied, thank you.
Quoted 2 lines · On Tue, Jul 21, 2020 at 6:30 PM -0700 David Miller <[email protected]> wrote:
Applied, thank you.
Thank you so much, David!