net: dqs: introduce IFF_NO_BQL private flag for non-BQL drivers
Subsystems (3)
PATCH v2 1/1 net: dqs: introduce IFF_NO_BQL private flag for non-BQL drivers
From: Jason Xing <[email protected]>
Since commit 74293ea1c4db6 ("net: sysfs: Do not create sysfs for non BQL device") limits the non-BQL driver not creating byte_queue_limits directory, I found there is one exception, namely, virtio-net driver, which should also be limited in netdev_uses_bql(). Let me give it a try first.
I decided to introduce a NO_BQL bit because:
1) it can help us limit virtio-net driver for now.
2) if we found another non-BQL driver, we can take it into account.
3) we can replace all the driver meeting those two statements in netdev_uses_bql() in future.
For now, I would like to make the first step to use this new bit for dqs use instead of replacing/applying all the non-BQL drivers in one go.
As Jakub said, "netdev_uses_bql() is best effort", I think, we can add new non-BQL drivers as soon as we find one.
After this patch, there is no byte_queue_limits directory in virtio-net driver.
---
v2
1. chose to add the new bit into enum netdev_priv_flags() instead of breaking the room of device feature.
--- drivers/net/virtio_net.c | 2 +- include/linux/netdevice.h | 4 ++++ net/core/net-sysfs.c | 2 +- 3 files changed, 6 insertions(+), 2 deletions(-)
drivers/net/virtio_net.c +1 -1
@@ -5631,7 +5631,7 @@ static int virtnet_probe(struct virtio_device *vdev)/* Set up network device as normal. */ dev->priv_flags |=IFF_UNICAST_FLT |IFF_LIVE_ADDR_CHANGE | -IFF_TX_SKB_NO_LINEAR ; +IFF_TX_SKB_NO_LINEAR |IFF_NO_BQL ; dev->netdev_ops = &virtnet_netdev; dev->stat_ops = &virtnet_stat_ops; dev->features =NETIF_F_HIGHDMA ;
include/linux/netdevice.h +4 -0
@@ -1649,6 +1649,9 @@ struct net_device_ops {* @IFF_SEE_ALL_HWTSTAMP_REQUESTS: device wants to see calls to * ndo_hwtstamp_set() for all timestamp requests regardless of source, * even if those aren't HWTSTAMP_SOURCE_NETDEV. +* @IFF_NO_BQL: driver doesn't use BQL for flow control for now. It's used +* to check if we should create byte_queue_limits directory in dqs +* (see netdev_uses_bql()) */ enum netdev_priv_flags {IFF_802_1Q_VLAN =1 <<0 ,
@@ -1685,6 +1688,7 @@ enum netdev_priv_flags {IFF_TX_SKB_NO_LINEAR =BIT_ULL (31 ),IFF_CHANGE_PROTO_DOWN =BIT_ULL (32 ),IFF_SEE_ALL_HWTSTAMP_REQUESTS =BIT_ULL (33 ), + IFF_NO_BQL = BIT_ULL(34), };#define IFF_802_1Q_VLAN IFF_802_1Q_VLAN
net/core/net-sysfs.c +1 -1
@@ -1765,7 +1765,7 @@ static const struct kobj_type netdev_queue_ktype = {static bool netdev_uses_bql (const struct net_device *dev) {if (dev->features &NETIF_F_LLTX || - dev->priv_flags &IFF_NO_QUEUE ) + dev->priv_flags & (IFF_NO_QUEUE |IFF_NO_BQL ))return false;return IS_ENABLED (CONFIG_BQL );
Signature · 1 line
2.37.3
Jakub Kicinski
maintainerinclude/linux/netdevice.h struct net_device_ops { · line 1654
Quoted 8 lines · On Tue, 11 Jun 2024 11:32:03 +0800 Jason Xing wrote:
+++
b/include/linux/netdevice.hinclude/linux/netdevice.h +3
-0@@ -1649,6 +1649,9 @@ struct net_device_ops {* @IFF_SEE_ALL_HWTSTAMP_REQUESTS: device wants to see calls to * ndo_hwtstamp_set() for all timestamp requests regardless of source, * even if those aren't HWTSTAMP_SOURCE_NETDEV. +* @IFF_NO_BQL: driver doesn't use BQL for flow control for now. It's used +* to check if we should create byte_queue_limits directory in dqs +* (see netdev_uses_bql())
Sorry for nit but since it's netdevice.h.. can we rephrase the comment a bit? How about just:
+ * @IFF_NO_BQL: driver doesn't support BQL, don't create "byte_queue_limits"
+ * directories in sysfs.
Quoted 14 lines · Jakub Kicinski
include/linux/netdevice.h struct net_device_ops { · line 1654
Quoted 8 lines · On Tue, 11 Jun 2024 11:32:03 +0800 Jason Xing wrote:
+++
b/include/linux/netdevice.hinclude/linux/netdevice.h +3
-0@@ -1649,6 +1649,9 @@ struct net_device_ops {* @IFF_SEE_ALL_HWTSTAMP_REQUESTS: device wants to see calls to * ndo_hwtstamp_set() for all timestamp requests regardless of source, * even if those aren't HWTSTAMP_SOURCE_NETDEV. +* @IFF_NO_BQL: driver doesn't use BQL for flow control for now. It's used +* to check if we should create byte_queue_limits directory in dqs +* (see netdev_uses_bql()) Sorry for nit but since it's
netdevice.h.. can we rephrase the comment a bit? How about just:
Quoted 2 lines
+ * @IFF_NO_BQL: driver doesn't support BQL, don't create "
byte_queue_limits"
+ * directories in sysfs.
No problem. I will revise it as you said. Thanks.