* ipv6 udp early demux breaks udp_l3mdev_accept=0
@ 2017-04-14 22:23 David Ahern
2017-04-15 0:38 ` Subash Abhinov Kasiviswanathan
0 siblings, 1 reply; 8+ messages in thread
From: David Ahern @ 2017-04-14 22:23 UTC (permalink / raw)
To: subashab; +Cc: netdev, Robert Shearman
Subash:
My understanding of early demux is that it should only match connected
sockets (src/sport + dst/dport are all set). The ipv6 udp early demux
code added in 5425077d73e0c is doing a generic socket lookup which does
not require an exact match and accordingly breaks udp_l3mdev_accept=0.
Please fix.
David
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: ipv6 udp early demux breaks udp_l3mdev_accept=0
2017-04-14 22:23 ipv6 udp early demux breaks udp_l3mdev_accept=0 David Ahern
@ 2017-04-15 0:38 ` Subash Abhinov Kasiviswanathan
2017-04-15 0:42 ` David Ahern
0 siblings, 1 reply; 8+ messages in thread
From: Subash Abhinov Kasiviswanathan @ 2017-04-15 0:38 UTC (permalink / raw)
To: David Ahern; +Cc: netdev, Robert Shearman
On 2017-04-14 16:23, David Ahern wrote:
> Subash:
>
> My understanding of early demux is that it should only match connected
> sockets (src/sport + dst/dport are all set). The ipv6 udp early demux
> code added in 5425077d73e0c is doing a generic socket lookup which does
> not require an exact match and accordingly breaks udp_l3mdev_accept=0.
> Please fix.
>
> David
Hi David
Would it be possible to disable udp_early_demux for your case
temporarily.
I'll look into this.
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: ipv6 udp early demux breaks udp_l3mdev_accept=0
2017-04-15 0:38 ` Subash Abhinov Kasiviswanathan
@ 2017-04-15 0:42 ` David Ahern
2017-04-15 13:48 ` David Miller
0 siblings, 1 reply; 8+ messages in thread
From: David Ahern @ 2017-04-15 0:42 UTC (permalink / raw)
To: Subash Abhinov Kasiviswanathan; +Cc: netdev, Robert Shearman
On 4/14/17 6:38 PM, Subash Abhinov Kasiviswanathan wrote:
> On 2017-04-14 16:23, David Ahern wrote:
>> Subash:
>>
>> My understanding of early demux is that it should only match connected
>> sockets (src/sport + dst/dport are all set). The ipv6 udp early demux
>> code added in 5425077d73e0c is doing a generic socket lookup which does
>> not require an exact match and accordingly breaks udp_l3mdev_accept=0.
>> Please fix.
>>
>> David
>
> Hi David
>
> Would it be possible to disable udp_early_demux for your case temporarily.
> I'll look into this.
>
It should be fixed for 4.12. Basically, __udp6_lib_demux_lookup needs to
be more like __udp4_lib_demux_lookup
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: ipv6 udp early demux breaks udp_l3mdev_accept=0
2017-04-15 0:42 ` David Ahern
@ 2017-04-15 13:48 ` David Miller
2017-04-16 2:23 ` Subash Abhinov Kasiviswanathan
0 siblings, 1 reply; 8+ messages in thread
From: David Miller @ 2017-04-15 13:48 UTC (permalink / raw)
To: dsa; +Cc: subashab, netdev, rshearma
From: David Ahern <dsa@cumulusnetworks.com>
Date: Fri, 14 Apr 2017 18:42:14 -0600
> On 4/14/17 6:38 PM, Subash Abhinov Kasiviswanathan wrote:
>> On 2017-04-14 16:23, David Ahern wrote:
>>> Subash:
>>>
>>> My understanding of early demux is that it should only match connected
>>> sockets (src/sport + dst/dport are all set). The ipv6 udp early demux
>>> code added in 5425077d73e0c is doing a generic socket lookup which does
>>> not require an exact match and accordingly breaks udp_l3mdev_accept=0.
>>> Please fix.
>>>
>>> David
>>
>> Hi David
>>
>> Would it be possible to disable udp_early_demux for your case temporarily.
>> I'll look into this.
>>
>
> It should be fixed for 4.12. Basically, __udp6_lib_demux_lookup needs to
> be more like __udp4_lib_demux_lookup
Agreed.
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: ipv6 udp early demux breaks udp_l3mdev_accept=0
2017-04-15 13:48 ` David Miller
@ 2017-04-16 2:23 ` Subash Abhinov Kasiviswanathan
2017-04-16 23:26 ` David Ahern
0 siblings, 1 reply; 8+ messages in thread
From: Subash Abhinov Kasiviswanathan @ 2017-04-16 2:23 UTC (permalink / raw)
To: dsa; +Cc: David Miller, netdev, rshearma
>> It should be fixed for 4.12. Basically, __udp6_lib_demux_lookup needs
>> to
>> be more like __udp4_lib_demux_lookup
Hi David
Would it be possible for you to test this with udp_l3mdev_accept=0
diff --git a/net/ipv6/udp.c b/net/ipv6/udp.c
index b793ed1..0e307e5 100644
--- a/net/ipv6/udp.c
+++ b/net/ipv6/udp.c
@@ -46,6 +46,7 @@
#include <net/tcp_states.h>
#include <net/ip6_checksum.h>
#include <net/xfrm.h>
+#include <net/inet_hashtables.h>
#include <net/inet6_hashtables.h>
#include <net/busy_poll.h>
#include <net/sock_reuseport.h>
@@ -864,21 +865,25 @@ int __udp6_lib_rcv(struct sk_buff *skb, struct
udp_table *udptable,
return 0;
}
+
static struct sock *__udp6_lib_demux_lookup(struct net *net,
__be16 loc_port, const struct in6_addr
*loc_addr,
__be16 rmt_port, const struct in6_addr
*rmt_addr,
int dif)
{
+ unsigned short hnum = ntohs(loc_port);
+ unsigned int hash2 = udp6_portaddr_hash(net, loc_addr, hnum);
+ unsigned int slot2 = hash2 & udp_table.mask;
+ struct udp_hslot *hslot2 = &udp_table.hash2[slot2];
+ const __portpair ports = INET_COMBINED_PORTS(rmt_port, hnum);
struct sock *sk;
- rcu_read_lock();
- sk = __udp6_lib_lookup(net, rmt_addr, rmt_port, loc_addr,
loc_port,
- dif, &udp_table, NULL);
- if (sk && !atomic_inc_not_zero(&sk->sk_refcnt))
- sk = NULL;
- rcu_read_unlock();
-
- return sk;
+ udp_portaddr_for_each_entry_rcu(sk, &hslot2->head) {
+ if (INET6_MATCH(sk, net, rmt_addr, loc_addr, ports,
dif))
+ return sk;
+ break;
+ }
+ return NULL;
}
static void udp_v6_early_demux(struct sk_buff *skb)
^ permalink raw reply related [flat|nested] 8+ messages in thread
* Re: ipv6 udp early demux breaks udp_l3mdev_accept=0
2017-04-16 2:23 ` Subash Abhinov Kasiviswanathan
@ 2017-04-16 23:26 ` David Ahern
2017-04-17 0:44 ` Subash Abhinov Kasiviswanathan
0 siblings, 1 reply; 8+ messages in thread
From: David Ahern @ 2017-04-16 23:26 UTC (permalink / raw)
To: Subash Abhinov Kasiviswanathan; +Cc: David Miller, netdev, rshearma
On 4/15/17 8:23 PM, Subash Abhinov Kasiviswanathan wrote:
>>> It should be fixed for 4.12. Basically, __udp6_lib_demux_lookup needs to
>>> be more like __udp4_lib_demux_lookup
>
> Hi David
>
> Would it be possible for you to test this with udp_l3mdev_accept=0
formatting was botched so it does not apply cleanly. please send a patch
file
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: ipv6 udp early demux breaks udp_l3mdev_accept=0
2017-04-16 23:26 ` David Ahern
@ 2017-04-17 0:44 ` Subash Abhinov Kasiviswanathan
2017-04-17 2:59 ` David Ahern
0 siblings, 1 reply; 8+ messages in thread
From: Subash Abhinov Kasiviswanathan @ 2017-04-17 0:44 UTC (permalink / raw)
To: dsa, davem, netdev, rshearma; +Cc: Subash Abhinov Kasiviswanathan
Can you try this
---
net/ipv6/udp.c | 21 +++++++++++++--------
1 file changed, 13 insertions(+), 8 deletions(-)
diff --git a/net/ipv6/udp.c b/net/ipv6/udp.c
index b793ed1..0e307e5 100644
--- a/net/ipv6/udp.c
+++ b/net/ipv6/udp.c
@@ -46,6 +46,7 @@
#include <net/tcp_states.h>
#include <net/ip6_checksum.h>
#include <net/xfrm.h>
+#include <net/inet_hashtables.h>
#include <net/inet6_hashtables.h>
#include <net/busy_poll.h>
#include <net/sock_reuseport.h>
@@ -864,21 +865,25 @@ int __udp6_lib_rcv(struct sk_buff *skb, struct udp_table *udptable,
return 0;
}
+
static struct sock *__udp6_lib_demux_lookup(struct net *net,
__be16 loc_port, const struct in6_addr *loc_addr,
__be16 rmt_port, const struct in6_addr *rmt_addr,
int dif)
{
+ unsigned short hnum = ntohs(loc_port);
+ unsigned int hash2 = udp6_portaddr_hash(net, loc_addr, hnum);
+ unsigned int slot2 = hash2 & udp_table.mask;
+ struct udp_hslot *hslot2 = &udp_table.hash2[slot2];
+ const __portpair ports = INET_COMBINED_PORTS(rmt_port, hnum);
struct sock *sk;
- rcu_read_lock();
- sk = __udp6_lib_lookup(net, rmt_addr, rmt_port, loc_addr, loc_port,
- dif, &udp_table, NULL);
- if (sk && !atomic_inc_not_zero(&sk->sk_refcnt))
- sk = NULL;
- rcu_read_unlock();
-
- return sk;
+ udp_portaddr_for_each_entry_rcu(sk, &hslot2->head) {
+ if (INET6_MATCH(sk, net, rmt_addr, loc_addr, ports, dif))
+ return sk;
+ break;
+ }
+ return NULL;
}
static void udp_v6_early_demux(struct sk_buff *skb)
--
1.9.1
^ permalink raw reply related [flat|nested] 8+ messages in thread
* Re: ipv6 udp early demux breaks udp_l3mdev_accept=0
2017-04-17 0:44 ` Subash Abhinov Kasiviswanathan
@ 2017-04-17 2:59 ` David Ahern
0 siblings, 0 replies; 8+ messages in thread
From: David Ahern @ 2017-04-17 2:59 UTC (permalink / raw)
To: Subash Abhinov Kasiviswanathan, davem, netdev, rshearma
On 4/16/17 6:44 PM, Subash Abhinov Kasiviswanathan wrote:
> Can you try this
>
> ---
> net/ipv6/udp.c | 21 +++++++++++++--------
> 1 file changed, 13 insertions(+), 8 deletions(-)
>
> diff --git a/net/ipv6/udp.c b/net/ipv6/udp.c
> index b793ed1..0e307e5 100644
> --- a/net/ipv6/udp.c
> +++ b/net/ipv6/udp.c
> @@ -46,6 +46,7 @@
> #include <net/tcp_states.h>
> #include <net/ip6_checksum.h>
> #include <net/xfrm.h>
> +#include <net/inet_hashtables.h>
> #include <net/inet6_hashtables.h>
> #include <net/busy_poll.h>
> #include <net/sock_reuseport.h>
> @@ -864,21 +865,25 @@ int __udp6_lib_rcv(struct sk_buff *skb, struct udp_table *udptable,
> return 0;
> }
>
> +
> static struct sock *__udp6_lib_demux_lookup(struct net *net,
> __be16 loc_port, const struct in6_addr *loc_addr,
> __be16 rmt_port, const struct in6_addr *rmt_addr,
> int dif)
> {
> + unsigned short hnum = ntohs(loc_port);
> + unsigned int hash2 = udp6_portaddr_hash(net, loc_addr, hnum);
> + unsigned int slot2 = hash2 & udp_table.mask;
> + struct udp_hslot *hslot2 = &udp_table.hash2[slot2];
> + const __portpair ports = INET_COMBINED_PORTS(rmt_port, hnum);
> struct sock *sk;
>
> - rcu_read_lock();
> - sk = __udp6_lib_lookup(net, rmt_addr, rmt_port, loc_addr, loc_port,
> - dif, &udp_table, NULL);
> - if (sk && !atomic_inc_not_zero(&sk->sk_refcnt))
> - sk = NULL;
> - rcu_read_unlock();
> -
> - return sk;
> + udp_portaddr_for_each_entry_rcu(sk, &hslot2->head) {
> + if (INET6_MATCH(sk, net, rmt_addr, loc_addr, ports, dif))
> + return sk;
> + break;
> + }
> + return NULL;
> }
>
> static void udp_v6_early_demux(struct sk_buff *skb)
>
That works for me. Thanks for the fast turn around.
^ permalink raw reply [flat|nested] 8+ messages in thread
end of thread, other threads:[~2017-04-17 2:59 UTC | newest]
Thread overview: 8+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2017-04-14 22:23 ipv6 udp early demux breaks udp_l3mdev_accept=0 David Ahern
2017-04-15 0:38 ` Subash Abhinov Kasiviswanathan
2017-04-15 0:42 ` David Ahern
2017-04-15 13:48 ` David Miller
2017-04-16 2:23 ` Subash Abhinov Kasiviswanathan
2017-04-16 23:26 ` David Ahern
2017-04-17 0:44 ` Subash Abhinov Kasiviswanathan
2017-04-17 2:59 ` David Ahern
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.