Hello Mikhail, Harald,
hoping there will be a v6 of this patch soon, a few comments:
On Monday, 8 January 2024 at 14:10:35 Mikhail Golubev-Ciuchea
<mikhail.golubev-ciuc...@opensynergy.com> wrote:
[...]
> +
> +/* virtio_can private data structure */
> +struct virtio_can_priv {
> + struct can_priv can; /* must be the first member */
> + /* NAPI for RX messages */
> + struct napi_struct napi;
> + /* NAPI for TX messages */
> + struct napi_struct napi_tx;
> + /* The network device we're associated with */
> + struct net_device *dev;
> + /* The virtio device we're associated with */
> + struct virtio_device *vdev;
> + /* The virtqueues */
> + struct virtqueue *vqs[VIRTIO_CAN_QUEUE_COUNT];
> + /* I/O callback function pointers for the virtqueues */
> + vq_callback_t *io_callbacks[VIRTIO_CAN_QUEUE_COUNT];
> + /* Lock for TX operations */
> + spinlock_t tx_lock;
> + /* Control queue lock. Defensive programming, may be not needed */
> + struct mutex ctrl_lock;
> + /* Wait for control queue processing without polling */
> + struct completion ctrl_done;
> + /* List of virtio CAN TX message */
> + struct list_head tx_list;
> + /* Array of receive queue messages */
> + struct virtio_can_rx rpkt[128];
This array should probably be allocated dynamically at probe - maybe
using a module parameter instead of a hardcoded value as length?
> + /* Those control queue messages cannot live on the stack! */
> + struct virtio_can_control_out cpkt_out;
> + struct virtio_can_control_in cpkt_in;
Consider using a container struct as you did for the tx message, e.g.:
struct virtio_can_control {
struct virtio_can_control_out ctrl_out;
struct virtio_can_control_in ctrl_in;
};
> + /* Data to get and maintain the putidx for local TX echo */
> + struct ida tx_putidx_ida;
> + /* In flight TX messages */
> + atomic_t tx_inflight;
> + /* BusOff pending. Reset after successful indication to upper layer */
> + bool busoff_pending;
> +};
> +
[...]
> +
> +/* Send a control message with message type either
> + *
> + * - VIRTIO_CAN_SET_CTRL_MODE_START or
> + * - VIRTIO_CAN_SET_CTRL_MODE_STOP.
> + *
> + * Unlike AUTOSAR CAN Driver Can_SetControllerMode() there is no requirement
> + * for this Linux driver to have an asynchronous implementation of the mode
> + * setting function so in order to keep things simple the function is
> + * implemented as synchronous function. Design pattern is
> + * virtio_console.c/__send_control_msg() &
> virtio_net.c/virtnet_send_command().
> + */
> +static u8 virtio_can_send_ctrl_msg(struct net_device *ndev, u16 msg_type)
> +{
> + struct scatterlist sg_out, sg_in, *sgs[2] = { &sg_out, &sg_in };
> + struct virtio_can_priv *priv = netdev_priv(ndev);
> + struct device *dev = &priv->vdev->dev;
> + struct virtqueue *vq;
> + unsigned int len;
> + int err;
> +
> + vq = priv->vqs[VIRTIO_CAN_QUEUE_CONTROL];
> +
> + /* The function may be serialized by rtnl lock. Not sure.
> + * Better safe than sorry.
> + */
> + mutex_lock(&priv->ctrl_lock);
> +
> + priv->cpkt_out.msg_type = cpu_to_le16(msg_type);
> + sg_init_one(&sg_out, &priv->cpkt_out, sizeof(priv->cpkt_out));
> + sg_init_one(&sg_in, &priv->cpkt_in, sizeof(priv->cpkt_in));
> +
> + err = virtqueue_add_sgs(vq, sgs, 1u, 1u, priv, GFP_ATOMIC);
> + if (err != 0) {
> + /* Not expected to happen */
> + dev_err(dev, "%s(): virtqueue_add_sgs() failed\n", __func__);
> + }
Here it should return VIRTIO_CAN_RESULT_NOT_OK after unlocking the
mutex, or it might wait for completion indefinitley below.
> +
> + if (!virtqueue_kick(vq)) {
> + /* Not expected to happen */
> + dev_err(dev, "%s(): Kick failed\n", __func__);
> + }
And here too.
> +
> + while (!virtqueue_get_buf(vq, &len) && !virtqueue_is_broken(vq))
> + wait_for_completion(&priv->ctrl_done);
> +
> + mutex_unlock(&priv->ctrl_lock);
> +
> + return priv->cpkt_in.result;
> +}
> +
[...]
> +static netdev_tx_t virtio_can_start_xmit(struct sk_buff *skb,
> + struct net_device *dev)
> +{
> + const unsigned int hdr_size = offsetof(struct virtio_can_tx_out, sdu);
> + struct scatterlist sg_out, sg_in, *sgs[2] = { &sg_out, &sg_in };
> + struct canfd_frame *cf = (struct canfd_frame *)skb->data;
> + struct virtio_can_priv *priv = netdev_priv(dev);
> + netdev_tx_t xmit_ret = NETDEV_TX_OK;
> + struct virtio_can_tx *can_tx_msg;
> + struct virtqueue *vq;
> + unsigned long flags;
> + u32 can_flags;
> + int putidx;
> + int err;
> +
> + vq = priv->vqs[VIRTIO_CAN_QUEUE_TX];
> +
> + if (can_dev_dropped_skb(dev, skb))
> + goto kick; /* No way to return NET_XMIT_DROP here */
> +
> + /* No local check for CAN_RTR_FLAG or FD frame against negotiated
> + * features. The device will reject those anyway if not supported.
> + */
> +
> + can_tx_msg = kzalloc(sizeof(*can_tx_msg), GFP_ATOMIC);
> + if (!can_tx_msg) {
> + dev->stats.tx_dropped++;
> + goto kick; /* No way to return NET_XMIT_DROP here */
> + }
> +
Since we are allocating tx messages dynamically, the sdu[64] array inside
struct virtio_can_tx_out can be converted to a flexible array and here
the allocation can become:
can_tx_msg = kzalloc(sizeof(*can_tx_msg) + cf->len, GFP_ATOMIC);
This would save memory in particular on CAN-CC interfaces, where 56 bytes
per message would otherwise be lost (not to mention the case if/when
CAN-XL will be supported).
> + can_tx_msg->tx_out.msg_type = cpu_to_le16(VIRTIO_CAN_TX);
> + can_flags = 0;
> +
> + if (cf->can_id & CAN_EFF_FLAG) {
> + can_flags |= VIRTIO_CAN_FLAGS_EXTENDED;
> + can_tx_msg->tx_out.can_id = cpu_to_le32(cf->can_id &
> CAN_EFF_MASK);
> + } else {
> + can_tx_msg->tx_out.can_id = cpu_to_le32(cf->can_id &
> CAN_SFF_MASK);
> + }
> + if (cf->can_id & CAN_RTR_FLAG)
> + can_flags |= VIRTIO_CAN_FLAGS_RTR;
> + else
> + memcpy(can_tx_msg->tx_out.sdu, cf->data, cf->len);
> + if (can_is_canfd_skb(skb))
> + can_flags |= VIRTIO_CAN_FLAGS_FD;
> +
> + can_tx_msg->tx_out.flags = cpu_to_le32(can_flags);
> + can_tx_msg->tx_out.length = cpu_to_le16(cf->len);
> +
> + /* Prepare sending of virtio message */
> + sg_init_one(&sg_out, &can_tx_msg->tx_out, hdr_size + cf->len);
> + sg_init_one(&sg_in, &can_tx_msg->tx_in, sizeof(can_tx_msg->tx_in));
> +
> + putidx = virtio_can_alloc_tx_idx(priv);
> +
> + if (unlikely(putidx < 0)) {
> + /* -ENOMEM or -ENOSPC here. -ENOSPC should not be possible as
> + * tx_inflight >= can.echo_skb_max is checked in flow control
> + */
> + WARN_ON_ONCE(putidx == -ENOSPC);
> + kfree(can_tx_msg);
> + dev->stats.tx_dropped++;
> + goto kick; /* No way to return NET_XMIT_DROP here */
> + }
> +
> + can_tx_msg->putidx = (unsigned int)putidx;
> +
> + /* Protect list operation */
> + spin_lock_irqsave(&priv->tx_lock, flags);
> + list_add_tail(&can_tx_msg->list, &priv->tx_list);
> + spin_unlock_irqrestore(&priv->tx_lock, flags);
> +
> + /* Push loopback echo. Will be looped back on TX interrupt/TX NAPI */
> + can_put_echo_skb(skb, dev, can_tx_msg->putidx, 0);
> +
> + /* Protect queue and list operations */
> + spin_lock_irqsave(&priv->tx_lock, flags);
> + err = virtqueue_add_sgs(vq, sgs, 1u, 1u, can_tx_msg, GFP_ATOMIC);
> + if (unlikely(err)) { /* checking vq->num_free in flow control */
> + list_del(&can_tx_msg->list);
> + can_free_echo_skb(dev, can_tx_msg->putidx, NULL);
> + virtio_can_free_tx_idx(priv, can_tx_msg->putidx);
> + spin_unlock_irqrestore(&priv->tx_lock, flags);
> + netif_stop_queue(dev);
> + kfree(can_tx_msg);
> + /* Expected never to be seen */
> + netdev_warn(dev, "TX: Stop queue, err = %d\n", err);
> + xmit_ret = NETDEV_TX_BUSY;
> + goto kick;
> + }
> +
> + /* Normal flow control: stop queue when no transmission slots left */
> + if (atomic_read(&priv->tx_inflight) >= priv->can.echo_skb_max ||
> + vq->num_free == 0 || (vq->num_free < ARRAY_SIZE(sgs) &&
> + !virtio_has_feature(vq->vdev, VIRTIO_RING_F_INDIRECT_DESC))) {
> + netif_stop_queue(dev);
> + netdev_dbg(dev, "TX: Normal stop queue\n");
> + }
> +
> + spin_unlock_irqrestore(&priv->tx_lock, flags);
> +
> +kick:
> + if (netif_queue_stopped(dev) || !netdev_xmit_more()) {
> + if (!virtqueue_kick(vq))
> + netdev_err(dev, "%s(): Kick failed\n", __func__);
> + }
> +
> + return xmit_ret;
> +}
> +
> +static const struct net_device_ops virtio_can_netdev_ops = {
> + .ndo_open = virtio_can_open,
> + .ndo_stop = virtio_can_close,
> + .ndo_start_xmit = virtio_can_start_xmit,
> + .ndo_change_mtu = can_change_mtu,
> +};
> +
> +static int register_virtio_can_dev(struct net_device *dev)
> +{
> + dev->flags |= IFF_ECHO; /* we support local echo */
> + dev->netdev_ops = &virtio_can_netdev_ops;
> +
> + return register_candev(dev);
> +}
> +
> +/* Compare with m_can.c/m_can_echo_tx_event() */
> +static int virtio_can_read_tx_queue(struct virtqueue *vq)
> +{
> + struct virtio_can_priv *can_priv = vq->vdev->priv;
> + struct net_device *dev = can_priv->dev;
> + struct virtio_can_tx *can_tx_msg;
> + struct net_device_stats *stats;
> + unsigned long flags;
> + unsigned int len;
> + u8 result;
> +
> + stats = &dev->stats;
> +
> + /* Protect list and virtio queue operations */
> + spin_lock_irqsave(&can_priv->tx_lock, flags);
> +
> + can_tx_msg = virtqueue_get_buf(vq, &len);
> + if (!can_tx_msg) {
> + spin_unlock_irqrestore(&can_priv->tx_lock, flags);
> + return 0; /* No more data */
> + }
> +
> + if (unlikely(len < sizeof(struct virtio_can_tx_in))) {
> + netdev_err(dev, "TX ACK: Device sent no result code\n");
> + result = VIRTIO_CAN_RESULT_NOT_OK; /* Keep things going */
> + } else {
> + result = can_tx_msg->tx_in.result;
> + }
> +
> + if (can_priv->can.state < CAN_STATE_BUS_OFF) {
> + /* Here also frames with result != VIRTIO_CAN_RESULT_OK are
> + * echoed. Intentional to bring a waiting process in an upper
> + * layer to an end.
> + * TODO: Any better means to indicate a problem here?
> + */
> + if (result != VIRTIO_CAN_RESULT_OK)
> + netdev_warn(dev, "TX ACK: Result = %u\n", result);
Maybe an error frame reporting CAN_ERR_CRTL_UNSPEC would be better?
For sure, counting the known errors as valid tx_packets and tx_bytes
is misleading.
> +
> + stats->tx_bytes += can_get_echo_skb(dev, can_tx_msg->putidx,
> + NULL);
> + stats->tx_packets++;
> + } else {
> + netdev_dbg(dev, "TX ACK: Controller inactive, drop echo\n");
> + can_free_echo_skb(dev, can_tx_msg->putidx, NULL);
> + }
> +
> + list_del(&can_tx_msg->list);
> + virtio_can_free_tx_idx(can_priv, can_tx_msg->putidx);
> +
> + /* Flow control */
> + if (netif_queue_stopped(dev)) {
> + netdev_dbg(dev, "TX ACK: Wake up stopped queue\n");
> + netif_wake_queue(dev);
> + }
> +
> + spin_unlock_irqrestore(&can_priv->tx_lock, flags);
> +
> + kfree(can_tx_msg);
> +
> + return 1; /* Queue was not empty so there may be more data */
> +}
> +
[...]
> +
> +static int virtio_can_find_vqs(struct virtio_can_priv *priv)
> +{
> + /* The order of RX and TX is exactly the opposite as in console and
> + * network. Does not play any role but is a bad trap.
> + */
> + static const char * const io_names[VIRTIO_CAN_QUEUE_COUNT] = {
> + "can-tx",
> + "can-rx",
> + "can-state-ctrl"
> + };
> +
> + priv->io_callbacks[VIRTIO_CAN_QUEUE_TX] = virtio_can_tx_intr;
> + priv->io_callbacks[VIRTIO_CAN_QUEUE_RX] = virtio_can_rx_intr;
> + priv->io_callbacks[VIRTIO_CAN_QUEUE_CONTROL] = virtio_can_control_intr;
> +
> + /* Find the queues. */
> + return virtio_find_vqs(priv->vdev, VIRTIO_CAN_QUEUE_COUNT, priv->vqs,
> + priv->io_callbacks, io_names, NULL);
> +}
Syntax of virtio_find_vqs changed a bit, here should now be:
struct virtqueue_info vqs_info[] = {
{ "can-tx", virtio_can_tx_intr },
{ "can-rx", virtio_can_rx_intr },
{ "can-state-ctrl", virtio_can_control_intr },
};
return virtio_find_vqs(priv->vdev, VIRTIO_CAN_QUEUE_COUNT, priv->vqs,
vqs_info, NULL);
> +
> +/* Function must not be called before virtio_can_find_vqs() has been run */
> +static void virtio_can_del_vq(struct virtio_device *vdev)
> +{
> + struct virtio_can_priv *priv = vdev->priv;
> + struct list_head *cursor, *next;
> + struct virtqueue *vq;
> +
> + /* Reset the device */
> + if (vdev->config->reset)
> + vdev->config->reset(vdev);
> +
> + /* From here we have dead silence from the device side so no locks
> + * are needed to protect against device side events.
> + */
> +
> + vq = priv->vqs[VIRTIO_CAN_QUEUE_CONTROL];
> + while (virtqueue_detach_unused_buf(vq))
> + ; /* Do nothing, content allocated statically */
> +
> + vq = priv->vqs[VIRTIO_CAN_QUEUE_RX];
> + while (virtqueue_detach_unused_buf(vq))
> + ; /* Do nothing, content allocated statically */
> +
> + vq = priv->vqs[VIRTIO_CAN_QUEUE_TX];
> + while (virtqueue_detach_unused_buf(vq))
> + ; /* Do nothing, content to be de-allocated separately */
> +
> + /* Is keeping track of allocated elements by an own linked list
> + * really necessary or may this be optimized using only
> + * virtqueue_detach_unused_buf()?
> + */
> + list_for_each_safe(cursor, next, &priv->tx_list) {
> + struct virtio_can_tx *can_tx;
> +
> + can_tx = list_entry(cursor, struct virtio_can_tx, list);
> + list_del(cursor);
> + kfree(can_tx);
> + }
I'd drop the tx_list entirely and rely on virtqueue_detach_unused_buf();
this would allow to remove at least one spinlock save/restore pair at
each transmission.
> +
> + if (vdev->config->del_vqs)
> + vdev->config->del_vqs(vdev);
> +}
> +
[...]
> diff --git a/include/uapi/linux/virtio_can.h b/include/uapi/linux/virtio_can.h
> new file mode 100644
> index 000000000000..7cf613bb3f1a
> --- /dev/null
> +++ b/include/uapi/linux/virtio_can.h
> @@ -0,0 +1,75 @@
> +/* SPDX-License-Identifier: BSD-3-Clause */
> +/*
> + * Copyright (C) 2021-2023 OpenSynergy GmbH
> + */
> +#ifndef _LINUX_VIRTIO_VIRTIO_CAN_H
> +#define _LINUX_VIRTIO_VIRTIO_CAN_H
> +
> +#include <linux/types.h>
> +#include <linux/virtio_types.h>
> +#include <linux/virtio_ids.h>
> +#include <linux/virtio_config.h>
> +
> +/* Feature bit numbers */
> +#define VIRTIO_CAN_F_CAN_CLASSIC 0
> +#define VIRTIO_CAN_F_CAN_FD 1
> +#define VIRTIO_CAN_F_LATE_TX_ACK 2
> +#define VIRTIO_CAN_F_RTR_FRAMES 3
> +
The values for VIRTIO_CAN_F_LATE_TX_ACK and VIRTIO_CAN_F_RTR_FRAMES are
inverted w.r.t. the merged virto-can spec [1].
Note that this is the only deviation from the spec I found.
> +/* CAN Result Types */
> +#define VIRTIO_CAN_RESULT_OK 0
> +#define VIRTIO_CAN_RESULT_NOT_OK 1
> +
> +/* CAN flags to determine type of CAN Id */
> +#define VIRTIO_CAN_FLAGS_EXTENDED 0x8000
> +#define VIRTIO_CAN_FLAGS_FD 0x4000
> +#define VIRTIO_CAN_FLAGS_RTR 0x2000
> +
> +struct virtio_can_config {
> +#define VIRTIO_CAN_S_CTRL_BUSOFF (1u << 0) /* Controller BusOff */
> + /* CAN controller status */
> + __le16 status;
> +};
> +
> +/* TX queue message types */
> +struct virtio_can_tx_out {
> +#define VIRTIO_CAN_TX 0x0001
> + __le16 msg_type;
> + __le16 length; /* 0..8 CC, 0..64 CAN-FD, 0..2048 CAN-XL, 12 bits */
> + __u8 reserved_classic_dlc; /* If CAN classic length = 8 then DLC can be
> 8..15 */
> + __u8 padding;
> + __le16 reserved_xl_priority; /* May be needed for CAN XL priority */
> + __le32 flags;
> + __le32 can_id;
> + __u8 sdu[64];
> +};
> +
sdu[] here might be a flexible array, if the driver allocates
virtio_can_tx_out structs dyncamically (see above). This would be
beneficial in case of CAN-XL frames (if/when they will be supported).
> +struct virtio_can_tx_in {
> + __u8 result;
> +};
> +
> +/* RX queue message types */
> +struct virtio_can_rx {
> +#define VIRTIO_CAN_RX 0x0101
> + __le16 msg_type;
> + __le16 length; /* 0..8 CC, 0..64 CAN-FD, 0..2048 CAN-XL, 12 bits */
> + __u8 reserved_classic_dlc; /* If CAN classic length = 8 then DLC can be
> 8..15 */
> + __u8 padding;
> + __le16 reserved_xl_priority; /* May be needed for CAN XL priority */
> + __le32 flags;
> + __le32 can_id;
> + __u8 sdu[64];
> +};
> +
Again, sdu[] might be a flexible array.
> +/* Control queue message types */
> +struct virtio_can_control_out {
> +#define VIRTIO_CAN_SET_CTRL_MODE_START 0x0201
> +#define VIRTIO_CAN_SET_CTRL_MODE_STOP 0x0202
> + __le16 msg_type;
> +};
> +
> +struct virtio_can_control_in {
> + __u8 result;
> +};
> +
> +#endif /* #ifndef _LINUX_VIRTIO_VIRTIO_CAN_H */
>
Thank you for your work!
Regards,
Francesco
[1]
https://github.com/oasis-tcs/virtio-spec/blob/virtio-1.4/device-types/can/description.tex#L45
