| Message ID | 20241104092434.2677-1-dqfext@gmail.com (mailing list archive) |
|---|---|
| State | Rejected |
| Delegated to: | Netdev Maintainers |
| Headers | show |
| Series | [net-next] net: ppp: remove ppp->closing check | expand |
On 11/4/24 10:24, Qingfang Deng wrote: > ppp->closing was used to test if an interface is closing down. But upon > .ndo_uninit() where ppp->closing is set to 1, dev_close() is already > called to bring down an interface and a synchronize_net() guarantees > that no pending TX/RX can take place, so the check is unnecessary. > Remove the check. I'm unsure we can remote such check. The TX callback can be triggered even from a write on the controlling file, and it looks like such file will be untouched by uninit. Cheers, Paolo
Hi Paolo, On Thu, Nov 7, 2024 at 8:08 PM Paolo Abeni <pabeni@redhat.com> wrote: > > > > On 11/4/24 10:24, Qingfang Deng wrote: > > ppp->closing was used to test if an interface is closing down. But upon > > .ndo_uninit() where ppp->closing is set to 1, dev_close() is already > > called to bring down an interface and a synchronize_net() guarantees > > that no pending TX/RX can take place, so the check is unnecessary. > > Remove the check. > > I'm unsure we can remote such check. The TX callback can be triggered > even from a write on the controlling file, and it looks like such file > will be untouched by uninit. ppp_release (when the file is closed) calls unregister_netdevice, and no more writes can happen after that. > > Cheers, > > Paolo >
On 11/8/24 07:09, Qingfang Deng wrote: > On Thu, Nov 7, 2024 at 8:08 PM Paolo Abeni <pabeni@redhat.com> wrote: >> On 11/4/24 10:24, Qingfang Deng wrote: >>> ppp->closing was used to test if an interface is closing down. But upon >>> .ndo_uninit() where ppp->closing is set to 1, dev_close() is already >>> called to bring down an interface and a synchronize_net() guarantees >>> that no pending TX/RX can take place, so the check is unnecessary. >>> Remove the check. >> >> I'm unsure we can remote such check. The TX callback can be triggered >> even from a write on the controlling file, and it looks like such file >> will be untouched by uninit. > > ppp_release (when the file is closed) calls unregister_netdevice, and > no more writes can happen after that. AFAICS the device can be deleted even without closing the file, via netlink or deleting the namespace. In such cases, AFAICS, the file is still alive. In any case we need a more solid explanation describing why the change is safe (and possibly a test-case deleting the device in different ways). /P
diff --git a/drivers/net/ppp/ppp_generic.c b/drivers/net/ppp/ppp_generic.c index 4583e15ad03a..4ff94005f6c1 100644 --- a/drivers/net/ppp/ppp_generic.c +++ b/drivers/net/ppp/ppp_generic.c @@ -136,7 +136,6 @@ struct ppp { unsigned long last_xmit; /* jiffies when last pkt sent 9c */ unsigned long last_recv; /* jiffies when last pkt rcvd a0 */ struct net_device *dev; /* network interface device a4 */ - int closing; /* is device closing down? a8 */ #ifdef CONFIG_PPP_MULTILINK int nxchan; /* next channel to send something on */ u32 nxseq; /* next sequence number to send */ @@ -1569,10 +1568,6 @@ static void ppp_dev_uninit(struct net_device *dev) struct ppp *ppp = netdev_priv(dev); struct ppp_net *pn = ppp_pernet(ppp->ppp_net); - ppp_lock(ppp); - ppp->closing = 1; - ppp_unlock(ppp); - mutex_lock(&pn->all_ppp_mutex); unit_put(&pn->units_idr, ppp->file.index); mutex_unlock(&pn->all_ppp_mutex); @@ -1651,23 +1646,19 @@ static void ppp_setup(struct net_device *dev) static void __ppp_xmit_process(struct ppp *ppp, struct sk_buff *skb) { ppp_xmit_lock(ppp); - if (!ppp->closing) { - ppp_push(ppp); + ppp_push(ppp); - if (skb) - skb_queue_tail(&ppp->file.xq, skb); - while (!ppp->xmit_pending && - (skb = skb_dequeue(&ppp->file.xq))) - ppp_send_frame(ppp, skb); - /* If there's no work left to do, tell the core net - code that we can accept some more. */ - if (!ppp->xmit_pending && !skb_peek(&ppp->file.xq)) - netif_wake_queue(ppp->dev); - else - netif_stop_queue(ppp->dev); - } else { - kfree_skb(skb); - } + if (skb) + skb_queue_tail(&ppp->file.xq, skb); + while (!ppp->xmit_pending && + (skb = skb_dequeue(&ppp->file.xq))) + ppp_send_frame(ppp, skb); + /* If there's no work left to do, tell the core net + code that we can accept some more. */ + if (!ppp->xmit_pending && !skb_peek(&ppp->file.xq)) + netif_wake_queue(ppp->dev); + else + netif_stop_queue(ppp->dev); ppp_xmit_unlock(ppp); } @@ -2208,10 +2199,7 @@ static inline void ppp_do_recv(struct ppp *ppp, struct sk_buff *skb, struct channel *pch) { ppp_recv_lock(ppp); - if (!ppp->closing) - ppp_receive_frame(ppp, skb, pch); - else - kfree_skb(skb); + ppp_receive_frame(ppp, skb, pch); ppp_recv_unlock(ppp); }
ppp->closing was used to test if an interface is closing down. But upon .ndo_uninit() where ppp->closing is set to 1, dev_close() is already called to bring down an interface and a synchronize_net() guarantees that no pending TX/RX can take place, so the check is unnecessary. Remove the check. Signed-off-by: Qingfang Deng <dqfext@gmail.com> --- drivers/net/ppp/ppp_generic.c | 38 ++++++++++++----------------------- 1 file changed, 13 insertions(+), 25 deletions(-)