* [PATCH net v2 0/2] net: dlink: handle copy_thresh allocation @ 2025-09-14 18:26 Yeounsu Moon 2025-09-14 18:26 ` [PATCH net v2 1/2] net: dlink: fix whitespace around function call Yeounsu Moon 2025-09-14 18:26 ` [PATCH net v2 2/2] net: dlink: handle copy_thresh allocation failure Yeounsu Moon 0 siblings, 2 replies; 5+ messages in thread From: Yeounsu Moon @ 2025-09-14 18:26 UTC (permalink / raw) To: Andrew Lunn, David S. Miller, Eric Dumazet, Jakub Kicinski, Paolo Abeni Cc: netdev, linux-kernel, Yeounsu Moon This series contains two patches: 1. clean up whitespace around function calls to follow coding style. (No functional change intended.) 2. Fix the memory handling issue with copybreak in the rx path. Both patches have been tested on hardware. Signed-off-by: Yeounsu Moon <yyyynoom@gmail.com> --- Changelog: v1: https://lore.kernel.org/netdev/20250912145339.67448-2-yyyynoom@gmail.com/ v2: - split into two patches: whitespace cleanup and functional fix --- Yeounsu Moon (2): net: dlink: fix whitespace around function call net: dlink: handle copy_thresh allocation failure drivers/net/ethernet/dlink/dl2k.c | 25 +++++++++++++++---------- 1 file changed, 15 insertions(+), 10 deletions(-) -- 2.51.0 ^ permalink raw reply [flat|nested] 5+ messages in thread
* [PATCH net v2 1/2] net: dlink: fix whitespace around function call 2025-09-14 18:26 [PATCH net v2 0/2] net: dlink: handle copy_thresh allocation Yeounsu Moon @ 2025-09-14 18:26 ` Yeounsu Moon 2025-09-14 18:55 ` Andrew Lunn 2025-09-14 18:26 ` [PATCH net v2 2/2] net: dlink: handle copy_thresh allocation failure Yeounsu Moon 1 sibling, 1 reply; 5+ messages in thread From: Yeounsu Moon @ 2025-09-14 18:26 UTC (permalink / raw) To: Andrew Lunn, David S. Miller, Eric Dumazet, Jakub Kicinski, Paolo Abeni Cc: netdev, linux-kernel, Yeounsu Moon Remove unnecessary whitespace between function names and the opening parenthesis to follow kernel coding style. No functional change intended. Tested-on: D-Link DGE-550T Rev-A3 Signed-off-by: Yeounsu Moon <yyyynoom@gmail.com> --- drivers/net/ethernet/dlink/dl2k.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/drivers/net/ethernet/dlink/dl2k.c b/drivers/net/ethernet/dlink/dl2k.c index 6bbf6e5584e5..faf8a9fc7ed1 100644 --- a/drivers/net/ethernet/dlink/dl2k.c +++ b/drivers/net/ethernet/dlink/dl2k.c @@ -970,17 +970,17 @@ receive_packet (struct net_device *dev) desc_to_dma(desc), np->rx_buf_sz, DMA_FROM_DEVICE); - skb_put (skb = np->rx_skbuff[entry], pkt_len); + skb_put(skb = np->rx_skbuff[entry], pkt_len); np->rx_skbuff[entry] = NULL; } else if ((skb = netdev_alloc_skb_ip_align(dev, pkt_len))) { dma_sync_single_for_cpu(&np->pdev->dev, desc_to_dma(desc), np->rx_buf_sz, DMA_FROM_DEVICE); - skb_copy_to_linear_data (skb, + skb_copy_to_linear_data(skb, np->rx_skbuff[entry]->data, pkt_len); - skb_put (skb, pkt_len); + skb_put(skb, pkt_len); dma_sync_single_for_device(&np->pdev->dev, desc_to_dma(desc), np->rx_buf_sz, -- 2.51.0 ^ permalink raw reply related [flat|nested] 5+ messages in thread
* Re: [PATCH net v2 1/2] net: dlink: fix whitespace around function call 2025-09-14 18:26 ` [PATCH net v2 1/2] net: dlink: fix whitespace around function call Yeounsu Moon @ 2025-09-14 18:55 ` Andrew Lunn 0 siblings, 0 replies; 5+ messages in thread From: Andrew Lunn @ 2025-09-14 18:55 UTC (permalink / raw) To: Yeounsu Moon Cc: Andrew Lunn, David S. Miller, Eric Dumazet, Jakub Kicinski, Paolo Abeni, netdev, linux-kernel On Mon, Sep 15, 2025 at 03:26:53AM +0900, Yeounsu Moon wrote: > Remove unnecessary whitespace between function names and the opening > parenthesis to follow kernel coding style. > > No functional change intended. > > Tested-on: D-Link DGE-550T Rev-A3 > Signed-off-by: Yeounsu Moon <yyyynoom@gmail.com> Reviewed-by: Andrew Lunn <andrew@lunn.ch> Andrew ^ permalink raw reply [flat|nested] 5+ messages in thread
* [PATCH net v2 2/2] net: dlink: handle copy_thresh allocation failure 2025-09-14 18:26 [PATCH net v2 0/2] net: dlink: handle copy_thresh allocation Yeounsu Moon 2025-09-14 18:26 ` [PATCH net v2 1/2] net: dlink: fix whitespace around function call Yeounsu Moon @ 2025-09-14 18:26 ` Yeounsu Moon 2025-09-14 18:53 ` Andrew Lunn 1 sibling, 1 reply; 5+ messages in thread From: Yeounsu Moon @ 2025-09-14 18:26 UTC (permalink / raw) To: Andrew Lunn, David S. Miller, Eric Dumazet, Jakub Kicinski, Paolo Abeni Cc: netdev, linux-kernel, Yeounsu Moon The driver did not handle failure of `netdev_alloc_skb_ip_align()`. If the allocation failed, dereferencing `skb->protocol` could lead to a NULL pointer dereference. This patch adds proper error handling by falling back to the `else` clause when the allocation fails. Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2") Tested-on: D-Link DGE-550T Rev-A3 Signed-off-by: Yeounsu Moon <yyyynoom@gmail.com> --- drivers/net/ethernet/dlink/dl2k.c | 21 +++++++++++++-------- 1 file changed, 13 insertions(+), 8 deletions(-) diff --git a/drivers/net/ethernet/dlink/dl2k.c b/drivers/net/ethernet/dlink/dl2k.c index faf8a9fc7ed1..a82e1fd01b92 100644 --- a/drivers/net/ethernet/dlink/dl2k.c +++ b/drivers/net/ethernet/dlink/dl2k.c @@ -965,14 +965,11 @@ receive_packet (struct net_device *dev) struct sk_buff *skb; /* Small skbuffs for short packets */ - if (pkt_len > copy_thresh) { - dma_unmap_single(&np->pdev->dev, - desc_to_dma(desc), - np->rx_buf_sz, - DMA_FROM_DEVICE); - skb_put(skb = np->rx_skbuff[entry], pkt_len); - np->rx_skbuff[entry] = NULL; - } else if ((skb = netdev_alloc_skb_ip_align(dev, pkt_len))) { + if (pkt_len <= copy_thresh) { + skb = netdev_alloc_skb_ip_align(dev, pkt_len); + if (!skb) + goto reuse_skbuff; + dma_sync_single_for_cpu(&np->pdev->dev, desc_to_dma(desc), np->rx_buf_sz, @@ -985,6 +982,14 @@ receive_packet (struct net_device *dev) desc_to_dma(desc), np->rx_buf_sz, DMA_FROM_DEVICE); + } else { +reuse_skbuff: + dma_unmap_single(&np->pdev->dev, + desc_to_dma(desc), + np->rx_buf_sz, + DMA_FROM_DEVICE); + skb_put(skb = np->rx_skbuff[entry], pkt_len); + np->rx_skbuff[entry] = NULL; } skb->protocol = eth_type_trans (skb, dev); #if 0 -- 2.51.0 ^ permalink raw reply related [flat|nested] 5+ messages in thread
* Re: [PATCH net v2 2/2] net: dlink: handle copy_thresh allocation failure 2025-09-14 18:26 ` [PATCH net v2 2/2] net: dlink: handle copy_thresh allocation failure Yeounsu Moon @ 2025-09-14 18:53 ` Andrew Lunn 0 siblings, 0 replies; 5+ messages in thread From: Andrew Lunn @ 2025-09-14 18:53 UTC (permalink / raw) To: Yeounsu Moon Cc: Andrew Lunn, David S. Miller, Eric Dumazet, Jakub Kicinski, Paolo Abeni, netdev, linux-kernel > This patch adds proper error handling by falling back to the `else` clause > when the allocation fails. > + if (pkt_len <= copy_thresh) { > + skb = netdev_alloc_skb_ip_align(dev, pkt_len); > + if (!skb) > + goto reuse_skbuff; > + > dma_sync_single_for_cpu(&np->pdev->dev, > desc_to_dma(desc), > np->rx_buf_sz, > @@ -985,6 +982,14 @@ receive_packet (struct net_device *dev) > desc_to_dma(desc), > np->rx_buf_sz, > DMA_FROM_DEVICE); > + } else { > +reuse_skbuff: To me, the name is confusing. What Ethernet drivers usually mean with reuse of an skbuf, is that they will give it straight back to the hardware for use. If you can successfully do copy break, this makes sense, the frame is no longer in the skbuf, it is in a new skbuf, so the old skbuf can be recycled. But that is not what is going on here. Copy break fails, and you fall back to the normal path. The data is still in the skbuf, so you cannot reuse it. Andrew --- pw-bot: cr ^ permalink raw reply [flat|nested] 5+ messages in thread
end of thread, other threads:[~2025-09-14 18:56 UTC | newest] Thread overview: 5+ messages (download: mbox.gz follow: Atom feed -- links below jump to the message on this page -- 2025-09-14 18:26 [PATCH net v2 0/2] net: dlink: handle copy_thresh allocation Yeounsu Moon 2025-09-14 18:26 ` [PATCH net v2 1/2] net: dlink: fix whitespace around function call Yeounsu Moon 2025-09-14 18:55 ` Andrew Lunn 2025-09-14 18:26 ` [PATCH net v2 2/2] net: dlink: handle copy_thresh allocation failure Yeounsu Moon 2025-09-14 18:53 ` Andrew Lunn
This is an external index of several public inboxes, see mirroring instructions on how to clone and mirror all data and code used by this external index.