Hello Marc, Thanks for your feedback. On Tue, 2024-04-30 at 11:10 +0200, Marc Kleine-Budde wrote: > On 11.03.2024 12:11:43, Vitor Soares wrote: > > From: Vitor Soares <vitor.soares@xxxxxxxxxxx> > > > > When the mcp251xfd_start_xmit() function fails, the driver stops > > processing messages, and the interrupt routine does not return, > > running indefinitely even after killing the running application. > > > > Error messages: > > [ 441.298819] mcp251xfd spi2.0 can0: ERROR in mcp251xfd_start_xmit: -16 > > [ 441.306498] mcp251xfd spi2.0 can0: Transmit Event FIFO buffer not empty. (seq=0x000017c7, > > tef_tail=0x000017cf, tef_head=0x000017d0, tx_head=0x000017d3). > > ... and repeat forever. > > > > The issue can be triggered when multiple devices share the same > > SPI interface. And there is concurrent access to the bus. > > > > The problem occurs because tx_ring->head increments even if > > mcp251xfd_start_xmit() fails. Consequently, the driver skips one > > TX package while still expecting a response in > > mcp251xfd_handle_tefif_one(). > > > > This patch resolves the issue by decreasing tx_ring->head if > > mcp251xfd_start_xmit() fails. With the fix, if we trigger the issue and > > the err = -EBUSY, the driver returns NETDEV_TX_BUSY. The network stack > > retries to transmit the message. > > Otherwise, it prints an error and discards the message. > > > > Fixes: 55e5b97f003e ("can: mcp25xxfd: add driver for Microchip MCP25xxFD SPI CAN") > > Cc: stable@xxxxxxxxxxxxxxx > > Signed-off-by: Vitor Soares <vitor.soares@xxxxxxxxxxx> > > --- > > > > V2->V3: > > - Add tx_dropped stats. > > - netdev_sent_queue() only if can_put_echo_skb() succeed. > > > > V1->V2: > > - Return NETDEV_TX_BUSY if mcp251xfd_tx_obj_write() == -EBUSY. > > - Rework the commit message to address the change above. > > - Change can_put_echo_skb() to be called after mcp251xfd_tx_obj_write() succeed. > > Otherwise, we get Kernel NULL pointer dereference error. > > > > drivers/net/can/spi/mcp251xfd/mcp251xfd-tx.c | 34 ++++++++++++-------- > > 1 file changed, 21 insertions(+), 13 deletions(-) > > > > diff --git a/drivers/net/can/spi/mcp251xfd/mcp251xfd-tx.c > > b/drivers/net/can/spi/mcp251xfd/mcp251xfd-tx.c > > index 160528d3cc26..146c44e47c60 100644 > > --- a/drivers/net/can/spi/mcp251xfd/mcp251xfd-tx.c > > +++ b/drivers/net/can/spi/mcp251xfd/mcp251xfd-tx.c > > @@ -166,6 +166,7 @@ netdev_tx_t mcp251xfd_start_xmit(struct sk_buff *skb, > > struct net_device *ndev) > > { > > struct mcp251xfd_priv *priv = netdev_priv(ndev); > > + struct net_device_stats *stats = &ndev->stats; > > struct mcp251xfd_tx_ring *tx_ring = priv->tx; > > struct mcp251xfd_tx_obj *tx_obj; > > unsigned int frame_len; > > @@ -181,25 +182,32 @@ netdev_tx_t mcp251xfd_start_xmit(struct sk_buff *skb, > > tx_obj = mcp251xfd_get_tx_obj_next(tx_ring); > > mcp251xfd_tx_obj_from_skb(priv, tx_obj, skb, tx_ring->head); > > > > - /* Stop queue if we occupy the complete TX FIFO */ > > tx_head = mcp251xfd_get_tx_head(tx_ring); > > - tx_ring->head++; > > - if (mcp251xfd_get_tx_free(tx_ring) == 0) > > - netif_stop_queue(ndev); > > - > > frame_len = can_skb_get_frame_len(skb); > > - err = can_put_echo_skb(skb, ndev, tx_head, frame_len); > > - if (!err) > > - netdev_sent_queue(priv->ndev, frame_len); > > + > > + tx_ring->head++; > > > > err = mcp251xfd_tx_obj_write(priv, tx_obj); > > - if (err) > > - goto out_err; > > + if (err) { > > + tx_ring->head--; > > > > - return NETDEV_TX_OK; > > + if (err == -EBUSY) > > + return NETDEV_TX_BUSY; > > > > - out_err: > > - netdev_err(priv->ndev, "ERROR in %s: %d\n", __func__, err); > > + stats->tx_dropped++; > > + > > + if (net_ratelimit()) > > + netdev_err(priv->ndev, > > + "ERROR in %s: %d\n", __func__, err); > > + } else { > > + err = can_put_echo_skb(skb, ndev, tx_head, frame_len); > > + if (!err) > > + netdev_sent_queue(priv->ndev, frame_len); > > + > > + /* Stop queue if we occupy the complete TX FIFO */ > > + if (mcp251xfd_get_tx_free(tx_ring) == 0) > > + netif_stop_queue(ndev); > > + } > > Once you have sent the TX objects to the chip, it can trigger a TX > complete IRQ. This is not very likely as the SPI sends asynchronously > and the xmit handler cannot sleep. > > This means you have to call can_put_echo_skb() and stop the queue if > needed, _before_ you call mcp251xfd_tx_obj_write(). In case of an error, > you have to roll-back these. Doing this way, I have to can_free_echo_skb() and it won't be possible to return NETDEV_TX_BUSY. Do you know how can we return NETDEV_TX_BUSY when spi_async() == EBUSY to allow the network stack to retry to transmit the package? Best regards, Vitor Soares > > regards, > Marc >