Re: [PATCH net-next v2] net: sysfs: Do not create sysfs for non BQL device
From: Jason Xing
Date: Sun Jun 09 2024 - 09:20:13 EST
On Sun, Jun 9, 2024 at 12:47 PM Jason Xing <kerneljasonxing@xxxxxxxxx> wrote:
>
> Hello Breno, Eric
>
> On Mon, Feb 19, 2024 at 5:47 PM Breno Leitao <leitao@xxxxxxxxxx> wrote:
> >
> > On Fri, Feb 16, 2024 at 07:45:37PM +0100, Eric Dumazet wrote:
> > > On Fri, Feb 16, 2024 at 7:41 PM Florian Fainelli <f.fainelli@xxxxxxxxx> wrote:
> > > >
> > > > On 2/16/24 09:29, Stephen Hemminger wrote:
> > > > > On Fri, 16 Feb 2024 01:41:52 -0800
> > > > > Breno Leitao <leitao@xxxxxxxxxx> wrote:
> > > > >
> > > > >> +static bool netdev_uses_bql(const struct net_device *dev)
> > > > >> +{
> > > > >> + if (dev->features & NETIF_F_LLTX ||
> > > > >> + dev->priv_flags & IFF_NO_QUEUE)
> > > > >> + return false;
> > > > >> +
> > > > >> + return IS_ENABLED(CONFIG_BQL);
> > > > >> +}
> > > > >
> > > > > Various compilers will warn about missing parens in that expression.
> > > > > It is valid but mixing & and || can be bug trap.
> > > > >
> > > > > if ((dev->features & NETIF_F_LLTX) || (dev->priv_flags & IFF_NO_QUEUE))
> > > > > return false;
> > > > >
> > > > > Not all drivers will be using bql, it requires driver to have that code.
> > > > > So really it means driver could be using BQL.
> > > > > Not sure if there is a way to find out if driver has the required BQL bits.
> > > >
> > > > There is not a feature flag to be keying off if that is what you are
> > > > after, you would need to audit the drivers and see whether they make
> > > > calls to netdev_tx_sent_queue(), netdev_tx_reset_queue(),
> > > > netdev_tx_completed_queue().
> > > >
> > > > I suppose you might be able to programmatically extract that information
> > > > by looking at whether a given driver object file has a reference to
> > > > dql_{reset,avail,completed} or do that at the source level, whichever is
> > > > easier.
> > >
> > > Note that the suggested patch does not change current functionality.
> > >
> > > Traditionally, we had sysfs entries fpr BQL for all netdev, regardless of them
> > > using BQL or not.
> > >
> > > The patch seems to be a good first step.
> >
> > Thanks Eric. I agree it solves the problem without creating a new
> > feature flag, that could also be done, but maybe less important than
> > this first step.
>
> When I'm reading and testing the dqs codes in my VM, I realize that
> the virtio_net driver should have been excluded from BQL drivers,
> which means VM using virtio_net driver should not have the
> /sys/class/net/eth1/queues/tx-1/byte_queue_limits directory because It
> has neither NETIF_F_LLTX nor IFF_NO_QUEUE.
>
> I'm trying to cook a patch to fix it without introducing a new feature
> like NETIF_F_LOOPBACK, but I failed to have a good patch.
>
> I have two options in my mind:
> 1) introduce a feature flag only for virtion_net [1]
> 2) introduce a BQL flag, and then apply it to all the drivers which
> has either NETIF_F_LLTX or IFF_NO_QUEUE bit, including virtio_net
> driver
I decided to use this method. Please review :
https://lore.kernel.org/all/20240609131732.73156-1-kerneljasonxing@xxxxxxxxx/
Thanks.
>
> Do you have any ideas or suggestions?
>
> Thanks in advance!
>
> [1]
> untested patch for now:
> diff --git a/drivers/net/virtio_net.c b/drivers/net/virtio_net.c
> index 61a57d134544..e39417d99ea8 100644
> --- a/drivers/net/virtio_net.c
> +++ b/drivers/net/virtio_net.c
> @@ -5634,7 +5634,7 @@ static int virtnet_probe(struct virtio_device *vdev)
> IFF_TX_SKB_NO_LINEAR;
> dev->netdev_ops = &virtnet_netdev;
> dev->stat_ops = &virtnet_stat_ops;
> - dev->features = NETIF_F_HIGHDMA;
> + dev->features = NETIF_F_HIGHDMA | NETIF_F_VIRTNET;
>
> dev->ethtool_ops = &virtnet_ethtool_ops;
> SET_NETDEV_DEV(dev, &vdev->dev);
> diff --git a/include/linux/netdev_features.h b/include/linux/netdev_features.h
> index 7c2d77d75a88..4ade9cdf079e 100644
> --- a/include/linux/netdev_features.h
> +++ b/include/linux/netdev_features.h
> @@ -14,7 +14,7 @@ typedef u64 netdev_features_t;
> enum {
> NETIF_F_SG_BIT, /* Scatter/gather IO. */
> NETIF_F_IP_CSUM_BIT, /* Can checksum TCP/UDP over IPv4. */
> - __UNUSED_NETIF_F_1,
> + //__UNUSED_NETIF_F_1,
> NETIF_F_HW_CSUM_BIT, /* Can checksum all the packets. */
> NETIF_F_IPV6_CSUM_BIT, /* Can checksum TCP/UDP over IPV6 */
> NETIF_F_HIGHDMA_BIT, /* Can DMA to high memory. */
> @@ -91,6 +91,7 @@ enum {
> NETIF_F_HW_HSR_FWD_BIT, /* Offload HSR forwarding */
> NETIF_F_HW_HSR_DUP_BIT, /* Offload HSR duplication */
>
> + NETIF_F_VIRTNET_BIT, /* Enable virtnet */
> /*
> * Add your fresh new feature above and remember to update
> * netdev_features_strings[] in net/ethtool/common.c and maybe
> @@ -122,6 +123,7 @@ enum {
> #define NETIF_F_IPV6_CSUM __NETIF_F(IPV6_CSUM)
> #define NETIF_F_LLTX __NETIF_F(LLTX)
> #define NETIF_F_LOOPBACK __NETIF_F(LOOPBACK)
> +#define NETIF_F_VIRTNET __NETIF_F(VIRTNET)
> #define NETIF_F_LRO __NETIF_F(LRO)
> #define NETIF_F_NETNS_LOCAL __NETIF_F(NETNS_LOCAL)
> #define NETIF_F_NOCACHE_COPY __NETIF_F(NOCACHE_COPY)
> diff --git a/net/core/net-sysfs.c b/net/core/net-sysfs.c
> index 4c27a360c294..d52d95ea6fb6 100644
> --- a/net/core/net-sysfs.c
> +++ b/net/core/net-sysfs.c
> @@ -1764,7 +1764,7 @@ static const struct kobj_type netdev_queue_ktype = {
>
> static bool netdev_uses_bql(const struct net_device *dev)
> {
> - if (dev->features & NETIF_F_LLTX ||
> + if (dev->features & (NETIF_F_LLTX | NETIF_F_VIRTNET) ||
> dev->priv_flags & IFF_NO_QUEUE)
> return false;
>
> diff --git a/net/ethtool/common.c b/net/ethtool/common.c
> index 6b2a360dcdf0..efb39a185e4b 100644
> --- a/net/ethtool/common.c
> +++ b/net/ethtool/common.c
> @@ -74,6 +74,7 @@ const char
> netdev_features_strings[NETDEV_FEATURE_COUNT][ETH_GSTRING_LEN] = {
> [NETIF_F_HW_HSR_TAG_RM_BIT] = "hsr-tag-rm-offload",
> [NETIF_F_HW_HSR_FWD_BIT] = "hsr-fwd-offload",
> [NETIF_F_HW_HSR_DUP_BIT] = "hsr-dup-offload",
> + [NETIF_F_VIRTNET_BIT] = "virtnet",
> };
>
> const char
>
> >
> > Hoping this is OK, I am planning to send a v2 adding the extra
> > parenthesis as reported above.
> >
> > Thanks
> >