With commit 34d21de99cea9 ("net: Move {l,t,d}stats allocation to core and convert veth & vrf"), stats allocation could be done on net core instead of in this driver.
With this new approach, the driver doesn't have to bother with error handling (allocation failure checking, making sure free happens in the right spot, etc). This is core responsibility now.
Remove the allocation in the gtp driver and leverage the network core allocation instead.
Signed-off-by: Breno Leitao leitao@debian.org --- drivers/net/gtp.c | 6 +----- 1 file changed, 1 insertion(+), 5 deletions(-)
diff --git a/drivers/net/gtp.c b/drivers/net/gtp.c index 2b46b7398ade..facbfba5d77f 100644 --- a/drivers/net/gtp.c +++ b/drivers/net/gtp.c @@ -717,10 +717,6 @@ static int gtp_dev_init(struct net_device *dev)
gtp->dev = dev;
- dev->tstats = netdev_alloc_pcpu_stats(struct pcpu_sw_netstats); - if (!dev->tstats) - return -ENOMEM; - return 0; }
@@ -729,7 +725,6 @@ static void gtp_dev_uninit(struct net_device *dev) struct gtp_dev *gtp = netdev_priv(dev);
gtp_encap_disable(gtp); - free_percpu(dev->tstats); }
static inline void gtp0_push_header(struct sk_buff *skb, struct pdp_ctx *pctx) @@ -970,6 +965,7 @@ static void gtp_link_setup(struct net_device *dev) dev->type = ARPHRD_NONE; dev->flags = IFF_POINTOPOINT | IFF_NOARP | IFF_MULTICAST;
+ dev->pcpu_stat_type = NETDEV_PCPU_STAT_TSTATS; dev->priv_flags |= IFF_NO_QUEUE; dev->features |= NETIF_F_LLTX; netif_keep_dst(dev);
Commit 3e2f544dd8a33 ("net: get stats64 if device if driver is configured") moved the callback to dev_get_tstats64() to net core, so, unless the driver is doing some custom stats collection, it does not need to set .ndo_get_stats64.
Since this driver is now relying in NETDEV_PCPU_STAT_TSTATS, then, it doesn't need to set the dev_get_tstats64() generic .ndo_get_stats64 function pointer.
Signed-off-by: Breno Leitao leitao@debian.org --- drivers/net/gtp.c | 1 - 1 file changed, 1 deletion(-)
diff --git a/drivers/net/gtp.c b/drivers/net/gtp.c index facbfba5d77f..a279f0cd1b4d 100644 --- a/drivers/net/gtp.c +++ b/drivers/net/gtp.c @@ -940,7 +940,6 @@ static const struct net_device_ops gtp_netdev_ops = { .ndo_init = gtp_dev_init, .ndo_uninit = gtp_dev_uninit, .ndo_start_xmit = gtp_dev_xmit, - .ndo_get_stats64 = dev_get_tstats64, };
static const struct device_type gtp_type = {
On Tue, Mar 05, 2024 at 04:15:22AM -0800, Breno Leitao wrote:
Commit 3e2f544dd8a33 ("net: get stats64 if device if driver is configured") moved the callback to dev_get_tstats64() to net core, so, unless the driver is doing some custom stats collection, it does not need to set .ndo_get_stats64.
Since this driver is now relying in NETDEV_PCPU_STAT_TSTATS, then, it doesn't need to set the dev_get_tstats64() generic .ndo_get_stats64 function pointer.
Signed-off-by: Breno Leitao leitao@debian.org
Acked-by: Pablo Neira Ayuso pablo@netfilter.org
Assign netdev to gtp->dev at setup time, so, we can get rid of gtp_dev_init() completely.
Signed-off-by: Breno Leitao leitao@debian.org --- drivers/net/gtp.c | 12 ++---------- 1 file changed, 2 insertions(+), 10 deletions(-)
diff --git a/drivers/net/gtp.c b/drivers/net/gtp.c index a279f0cd1b4d..ba4704c2c640 100644 --- a/drivers/net/gtp.c +++ b/drivers/net/gtp.c @@ -711,15 +711,6 @@ static int gtp_encap_recv(struct sock *sk, struct sk_buff *skb) return ret; }
-static int gtp_dev_init(struct net_device *dev) -{ - struct gtp_dev *gtp = netdev_priv(dev); - - gtp->dev = dev; - - return 0; -} - static void gtp_dev_uninit(struct net_device *dev) { struct gtp_dev *gtp = netdev_priv(dev); @@ -937,7 +928,6 @@ static netdev_tx_t gtp_dev_xmit(struct sk_buff *skb, struct net_device *dev) }
static const struct net_device_ops gtp_netdev_ops = { - .ndo_init = gtp_dev_init, .ndo_uninit = gtp_dev_uninit, .ndo_start_xmit = gtp_dev_xmit, }; @@ -951,6 +941,7 @@ static void gtp_link_setup(struct net_device *dev) unsigned int max_gtp_header_len = sizeof(struct iphdr) + sizeof(struct udphdr) + sizeof(struct gtp0_header); + struct gtp_dev *gtp = netdev_priv(dev);
dev->netdev_ops = >p_netdev_ops; dev->needs_free_netdev = true; @@ -970,6 +961,7 @@ static void gtp_link_setup(struct net_device *dev) netif_keep_dst(dev);
dev->needed_headroom = LL_MAX_HEADER + max_gtp_header_len; + gtp->dev = dev; }
static int gtp_hashtable_new(struct gtp_dev *gtp, int hsize);
On Tue, Mar 05, 2024 at 04:15:23AM -0800, Breno Leitao wrote:
Assign netdev to gtp->dev at setup time, so, we can get rid of gtp_dev_init() completely.
Signed-off-by: Breno Leitao leitao@debian.org
Acked-by: Pablo Neira Ayuso pablo@netfilter.org
On Tue, Mar 05, 2024 at 04:15:21AM -0800, Breno Leitao wrote:
With commit 34d21de99cea9 ("net: Move {l,t,d}stats allocation to core and convert veth & vrf"), stats allocation could be done on net core instead of in this driver.
With this new approach, the driver doesn't have to bother with error handling (allocation failure checking, making sure free happens in the right spot, etc). This is core responsibility now.
Remove the allocation in the gtp driver and leverage the network core allocation instead.
Signed-off-by: Breno Leitao leitao@debian.org
Acked-by: Pablo Neira Ayuso pablo@netfilter.org
Thanks, this was on my list.
Hello:
This series was applied to netdev/net-next.git (main) by Jakub Kicinski kuba@kernel.org:
On Tue, 5 Mar 2024 04:15:21 -0800 you wrote:
With commit 34d21de99cea9 ("net: Move {l,t,d}stats allocation to core and convert veth & vrf"), stats allocation could be done on net core instead of in this driver.
With this new approach, the driver doesn't have to bother with error handling (allocation failure checking, making sure free happens in the right spot, etc). This is core responsibility now.
[...]
Here is the summary with links: - [net-next,1/3] net: gtp: Leverage core stats allocator https://git.kernel.org/netdev/net-next/c/660e5aaea11d - [net-next,2/3] net: gtp: Remove generic .ndo_get_stats64 https://git.kernel.org/netdev/net-next/c/13957a0b0708 - [net-next,3/3] net: gtp: Move net_device assigned in setup https://git.kernel.org/netdev/net-next/c/81154bb83c44
You are awesome, thank you!
osmocom-net-gprs@lists.osmocom.org