[net-next] net/smc: Introduce a hook to modify syn_smc at runtime

Commit Message

D. Wythe Oct. 10, 2024, 3:58 a.m. UTC

From: "D. Wythe" <[e-mail]>

The introduction of IPPROTO_SMC enables eBPF programs to determine
whether to use SMC based on the context of socket creation, such as
network namespaces, PID and comm name, etc.

As a subsequent enhancement, this patch introduces a new hook for eBPF
programs that allows decisions on whether to use SMC or not at runtime,
including but not limited to local/remote IP address or ports. In
simpler words, this feature allows modifications to syn_smc through eBPF
programs before the TCP three-way handshake got established.

Thanks to kfunc for making it easier for us to implement this feature in
SMC.

Signed-off-by: D. Wythe <[e-mail]>
---
 include/linux/tcp.h  |  4 ++-
 net/ipv4/tcp_input.c |  4 +--
 net/smc/af_smc.c     | 75 ++++++++++++++++++++++++++++++++++++++++++++++------
 3 files changed, 72 insertions(+), 11 deletions(-)

Comments

Alexei Starovoitov Oct. 10, 2024, 4:21 p.m. UTC | #1

On Wed, Oct 9, 2024 at 8:58 PM D. Wythe <[e-mail]> wrote:
>
>
> +__bpf_hook_start();
> +
> +__weak noinline int select_syn_smc(const struct sock *sk, struct sockaddr *peer)
> +{
> +       return 1;
> +}
> +
> +__bpf_hook_end();
> +
>  int smc_nl_dump_hs_limitation(struct sk_buff *skb, struct netlink_callback *cb)
>  {
>         struct smc_nl_dmp_ctx *cb_ctx = smc_nl_dmp_ctx(cb);
> @@ -156,19 +165,43 @@ static struct sock *smc_tcp_syn_recv_sock(const struct sock *sk,
>         return NULL;
>  }
>
> -static bool smc_hs_congested(const struct sock *sk)
> +static void smc_openreq_init(struct request_sock *req,
> +                            const struct tcp_options_received *rx_opt,
> +                            struct sk_buff *skb, const struct sock *sk)
>  {
> +       struct inet_request_sock *ireq = inet_rsk(req);
> +       struct sockaddr_storage rmt_sockaddr = {};
>         const struct smc_sock *smc;
>
>         smc = smc_clcsock_user_data(sk);
>
>         if (!smc)
> -               return true;
> +               return;
>
> -       if (workqueue_congested(WORK_CPU_UNBOUND, smc_hs_wq))
> -               return true;
> +       if (smc->limit_smc_hs && workqueue_congested(WORK_CPU_UNBOUND, smc_hs_wq))
> +               goto out_no_smc;
>
> -       return false;
> +       rmt_sockaddr.ss_family = sk->sk_family;
> +
> +       if (rmt_sockaddr.ss_family == AF_INET) {
> +               struct sockaddr_in *rmt4_sockaddr =  (struct sockaddr_in *)&rmt_sockaddr;
> +
> +               rmt4_sockaddr->sin_addr.s_addr = ireq->ir_rmt_addr;
> +               rmt4_sockaddr->sin_port = ireq->ir_rmt_port;
> +#if IS_ENABLED(CONFIG_IPV6)
> +       } else {
> +               struct sockaddr_in6 *rmt6_sockaddr =  (struct sockaddr_in6 *)&rmt_sockaddr;
> +
> +               rmt6_sockaddr->sin6_addr = ireq->ir_v6_rmt_addr;
> +               rmt6_sockaddr->sin6_port = ireq->ir_rmt_port;
> +#endif /* CONFIG_IPV6 */
> +       }
> +
> +       ireq->smc_ok = select_syn_smc(sk, (struct sockaddr *)&rmt_sockaddr);
> +       return;
> +out_no_smc:
> +       ireq->smc_ok = 0;
> +       return;
>  }
>
>  struct smc_hashinfo smc_v4_hashinfo = {
> @@ -1671,7 +1704,7 @@ int smc_connect(struct socket *sock, struct sockaddr *addr,
>         }
>
>         smc_copy_sock_settings_to_clc(smc);
> -       tcp_sk(smc->clcsock->sk)->syn_smc = 1;
> +       tcp_sk(smc->clcsock->sk)->syn_smc = select_syn_smc(sk, addr);
>         if (smc->connect_nonblock) {
>                 rc = -EALREADY;
>                 goto out;
> @@ -2650,8 +2683,7 @@ int smc_listen(struct socket *sock, int backlog)
>
>         inet_csk(smc->clcsock->sk)->icsk_af_ops = &smc->af_ops;
>
> -       if (smc->limit_smc_hs)
> -               tcp_sk(smc->clcsock->sk)->smc_hs_congested = smc_hs_congested;
> +       tcp_sk(smc->clcsock->sk)->smc_openreq_init = smc_openreq_init;
>
>         rc = kernel_listen(smc->clcsock, backlog);
>         if (rc) {
> @@ -3475,6 +3507,24 @@ static void __net_exit smc_net_stat_exit(struct net *net)
>         .exit = smc_net_stat_exit,
>  };
>
> +#if IS_ENABLED(CONFIG_BPF_SYSCALL)
> +BTF_SET8_START(bpf_smc_fmodret_ids)
> +BTF_ID_FLAGS(func, select_syn_smc)
> +BTF_SET8_END(bpf_smc_fmodret_ids)
> +
> +static const struct btf_kfunc_id_set bpf_smc_fmodret_set = {
> +       .owner = THIS_MODULE,
> +       .set   = &bpf_smc_fmodret_ids,
> +};
> +
> +static int bpf_smc_kfunc_init(void)
> +{
> +       return register_btf_fmodret_id_set(&bpf_smc_fmodret_set);
> +}

fmodret was an approach that hid-bpf took initially,
but eventually they removed it all and switched to struct-ops approach.
Please learn that lesson.
Use struct_ops from the beginning.

I did a presentation recently explaining the motivation behind
struct_ops and tips on how to extend the kernel.
TLDR: the step one is to design the extension _without_ bpf.
The interface should be usable for kernel modules.
And then when you have *_ops style api in place
the bpf progs will plug-in without extra work.

Slides:
https://github.com/4ast/docs/blob/main/BPF%20struct-ops.pdf

D. Wythe Oct. 11, 2024, 6:43 a.m. UTC | #2

On 10/11/24 12:21 AM, Alexei Starovoitov wrote:
> On Wed, Oct 9, 2024 at 8:58 PM D. Wythe <[e-mail]> wrote:
>>
>>
>> +__bpf_hook_start();
>> +
>> +__weak noinline int select_syn_smc(const struct sock *sk, struct sockaddr *peer)
>> +{
>> +       return 1;
>> +}
>> +
>> +__bpf_hook_end();
>> +
>>   int smc_nl_dump_hs_limitation(struct sk_buff *skb, struct netlink_callback *cb)
>>   {
>>          struct smc_nl_dmp_ctx *cb_ctx = smc_nl_dmp_ctx(cb);
>> @@ -156,19 +165,43 @@ static struct sock *smc_tcp_syn_recv_sock(const struct sock *sk,
>>          return NULL;
>>   }
>>
>> -static bool smc_hs_congested(const struct sock *sk)
>> +static void smc_openreq_init(struct request_sock *req,
>> +                            const struct tcp_options_received *rx_opt,
>> +                            struct sk_buff *skb, const struct sock *sk)
>>   {
>> +       struct inet_request_sock *ireq = inet_rsk(req);
>> +       struct sockaddr_storage rmt_sockaddr = {};
>>          const struct smc_sock *smc;
>>
>>          smc = smc_clcsock_user_data(sk);
>>
>>          if (!smc)
>> -               return true;
>> +               return;
>>
>> -       if (workqueue_congested(WORK_CPU_UNBOUND, smc_hs_wq))
>> -               return true;
>> +       if (smc->limit_smc_hs && workqueue_congested(WORK_CPU_UNBOUND, smc_hs_wq))
>> +               goto out_no_smc;
>>
>> -       return false;
>> +       rmt_sockaddr.ss_family = sk->sk_family;
>> +
>> +       if (rmt_sockaddr.ss_family == AF_INET) {
>> +               struct sockaddr_in *rmt4_sockaddr =  (struct sockaddr_in *)&rmt_sockaddr;
>> +
>> +               rmt4_sockaddr->sin_addr.s_addr = ireq->ir_rmt_addr;
>> +               rmt4_sockaddr->sin_port = ireq->ir_rmt_port;
>> +#if IS_ENABLED(CONFIG_IPV6)
>> +       } else {
>> +               struct sockaddr_in6 *rmt6_sockaddr =  (struct sockaddr_in6 *)&rmt_sockaddr;
>> +
>> +               rmt6_sockaddr->sin6_addr = ireq->ir_v6_rmt_addr;
>> +               rmt6_sockaddr->sin6_port = ireq->ir_rmt_port;
>> +#endif /* CONFIG_IPV6 */
>> +       }
>> +
>> +       ireq->smc_ok = select_syn_smc(sk, (struct sockaddr *)&rmt_sockaddr);
>> +       return;
>> +out_no_smc:
>> +       ireq->smc_ok = 0;
>> +       return;
>>   }
>>
>>   struct smc_hashinfo smc_v4_hashinfo = {
>> @@ -1671,7 +1704,7 @@ int smc_connect(struct socket *sock, struct sockaddr *addr,
>>          }
>>
>>          smc_copy_sock_settings_to_clc(smc);
>> -       tcp_sk(smc->clcsock->sk)->syn_smc = 1;
>> +       tcp_sk(smc->clcsock->sk)->syn_smc = select_syn_smc(sk, addr);
>>          if (smc->connect_nonblock) {
>>                  rc = -EALREADY;
>>                  goto out;
>> @@ -2650,8 +2683,7 @@ int smc_listen(struct socket *sock, int backlog)
>>
>>          inet_csk(smc->clcsock->sk)->icsk_af_ops = &smc->af_ops;
>>
>> -       if (smc->limit_smc_hs)
>> -               tcp_sk(smc->clcsock->sk)->smc_hs_congested = smc_hs_congested;
>> +       tcp_sk(smc->clcsock->sk)->smc_openreq_init = smc_openreq_init;
>>
>>          rc = kernel_listen(smc->clcsock, backlog);
>>          if (rc) {
>> @@ -3475,6 +3507,24 @@ static void __net_exit smc_net_stat_exit(struct net *net)
>>          .exit = smc_net_stat_exit,
>>   };
>>
>> +#if IS_ENABLED(CONFIG_BPF_SYSCALL)
>> +BTF_SET8_START(bpf_smc_fmodret_ids)
>> +BTF_ID_FLAGS(func, select_syn_smc)
>> +BTF_SET8_END(bpf_smc_fmodret_ids)
>> +
>> +static const struct btf_kfunc_id_set bpf_smc_fmodret_set = {
>> +       .owner = THIS_MODULE,
>> +       .set   = &bpf_smc_fmodret_ids,
>> +};
>> +
>> +static int bpf_smc_kfunc_init(void)
>> +{
>> +       return register_btf_fmodret_id_set(&bpf_smc_fmodret_set);
>> +}
> 
> fmodret was an approach that hid-bpf took initially,
> but eventually they removed it all and switched to struct-ops approach.
> Please learn that lesson.
> Use struct_ops from the beginning.
> 
> I did a presentation recently explaining the motivation behind
> struct_ops and tips on how to extend the kernel.
> TLDR: the step one is to design the extension _without_ bpf.
> The interface should be usable for kernel modules.
> And then when you have *_ops style api in place
> the bpf progs will plug-in without extra work.
> 
> Slides:
> https://github.com/4ast/docs/blob/main/BPF%20struct-ops.pdf


Hi Alexei,

Thanks very much for your suggestion.

In fact, I tried struct_ops in SMC about a year ago. Unfortunately, at that time struct_ops did not 
support registration from modules, and I had to move some smc dependencies into bpf, which met with 
community opposition. However, I noticed that this feature is now supported, so perhaps this is an 
opportunity.

But on the other hand, given the current functionality, I wonder if struct_ops might be an overkill. 
I haven't been able to come up with a suitable abstraction to define this ops, and in the future, 
this ops might only contain the very one callback (select_syn_smc).

Looking forward for your advises.

Thanks,
D. Wythe

Alexei Starovoitov Oct. 11, 2024, 3:37 p.m. UTC | #3

On Thu, Oct 10, 2024 at 11:44 PM D. Wythe <[e-mail]> wrote:
>
>
>
> On 10/11/24 12:21 AM, Alexei Starovoitov wrote:
> > On Wed, Oct 9, 2024 at 8:58 PM D. Wythe <[e-mail]> wrote:
> >>
> >>
> >> +__bpf_hook_start();
> >> +
> >> +__weak noinline int select_syn_smc(const struct sock *sk, struct sockaddr *peer)
> >> +{
> >> +       return 1;
> >> +}
> >> +
> >> +__bpf_hook_end();
> >> +
> >>   int smc_nl_dump_hs_limitation(struct sk_buff *skb, struct netlink_callback *cb)
> >>   {
> >>          struct smc_nl_dmp_ctx *cb_ctx = smc_nl_dmp_ctx(cb);
> >> @@ -156,19 +165,43 @@ static struct sock *smc_tcp_syn_recv_sock(const struct sock *sk,
> >>          return NULL;
> >>   }
> >>
> >> -static bool smc_hs_congested(const struct sock *sk)
> >> +static void smc_openreq_init(struct request_sock *req,
> >> +                            const struct tcp_options_received *rx_opt,
> >> +                            struct sk_buff *skb, const struct sock *sk)
> >>   {
> >> +       struct inet_request_sock *ireq = inet_rsk(req);
> >> +       struct sockaddr_storage rmt_sockaddr = {};
> >>          const struct smc_sock *smc;
> >>
> >>          smc = smc_clcsock_user_data(sk);
> >>
> >>          if (!smc)
> >> -               return true;
> >> +               return;
> >>
> >> -       if (workqueue_congested(WORK_CPU_UNBOUND, smc_hs_wq))
> >> -               return true;
> >> +       if (smc->limit_smc_hs && workqueue_congested(WORK_CPU_UNBOUND, smc_hs_wq))
> >> +               goto out_no_smc;
> >>
> >> -       return false;
> >> +       rmt_sockaddr.ss_family = sk->sk_family;
> >> +
> >> +       if (rmt_sockaddr.ss_family == AF_INET) {
> >> +               struct sockaddr_in *rmt4_sockaddr =  (struct sockaddr_in *)&rmt_sockaddr;
> >> +
> >> +               rmt4_sockaddr->sin_addr.s_addr = ireq->ir_rmt_addr;
> >> +               rmt4_sockaddr->sin_port = ireq->ir_rmt_port;
> >> +#if IS_ENABLED(CONFIG_IPV6)
> >> +       } else {
> >> +               struct sockaddr_in6 *rmt6_sockaddr =  (struct sockaddr_in6 *)&rmt_sockaddr;
> >> +
> >> +               rmt6_sockaddr->sin6_addr = ireq->ir_v6_rmt_addr;
> >> +               rmt6_sockaddr->sin6_port = ireq->ir_rmt_port;
> >> +#endif /* CONFIG_IPV6 */
> >> +       }
> >> +
> >> +       ireq->smc_ok = select_syn_smc(sk, (struct sockaddr *)&rmt_sockaddr);
> >> +       return;
> >> +out_no_smc:
> >> +       ireq->smc_ok = 0;
> >> +       return;
> >>   }
> >>
> >>   struct smc_hashinfo smc_v4_hashinfo = {
> >> @@ -1671,7 +1704,7 @@ int smc_connect(struct socket *sock, struct sockaddr *addr,
> >>          }
> >>
> >>          smc_copy_sock_settings_to_clc(smc);
> >> -       tcp_sk(smc->clcsock->sk)->syn_smc = 1;
> >> +       tcp_sk(smc->clcsock->sk)->syn_smc = select_syn_smc(sk, addr);
> >>          if (smc->connect_nonblock) {
> >>                  rc = -EALREADY;
> >>                  goto out;
> >> @@ -2650,8 +2683,7 @@ int smc_listen(struct socket *sock, int backlog)
> >>
> >>          inet_csk(smc->clcsock->sk)->icsk_af_ops = &smc->af_ops;
> >>
> >> -       if (smc->limit_smc_hs)
> >> -               tcp_sk(smc->clcsock->sk)->smc_hs_congested = smc_hs_congested;
> >> +       tcp_sk(smc->clcsock->sk)->smc_openreq_init = smc_openreq_init;
> >>
> >>          rc = kernel_listen(smc->clcsock, backlog);
> >>          if (rc) {
> >> @@ -3475,6 +3507,24 @@ static void __net_exit smc_net_stat_exit(struct net *net)
> >>          .exit = smc_net_stat_exit,
> >>   };
> >>
> >> +#if IS_ENABLED(CONFIG_BPF_SYSCALL)
> >> +BTF_SET8_START(bpf_smc_fmodret_ids)
> >> +BTF_ID_FLAGS(func, select_syn_smc)
> >> +BTF_SET8_END(bpf_smc_fmodret_ids)
> >> +
> >> +static const struct btf_kfunc_id_set bpf_smc_fmodret_set = {
> >> +       .owner = THIS_MODULE,
> >> +       .set   = &bpf_smc_fmodret_ids,
> >> +};
> >> +
> >> +static int bpf_smc_kfunc_init(void)
> >> +{
> >> +       return register_btf_fmodret_id_set(&bpf_smc_fmodret_set);
> >> +}
> >
> > fmodret was an approach that hid-bpf took initially,
> > but eventually they removed it all and switched to struct-ops approach.
> > Please learn that lesson.
> > Use struct_ops from the beginning.
> >
> > I did a presentation recently explaining the motivation behind
> > struct_ops and tips on how to extend the kernel.
> > TLDR: the step one is to design the extension _without_ bpf.
> > The interface should be usable for kernel modules.
> > And then when you have *_ops style api in place
> > the bpf progs will plug-in without extra work.
> >
> > Slides:
> > https://github.com/4ast/docs/blob/main/BPF%20struct-ops.pdf
>
>
> Hi Alexei,
>
> Thanks very much for your suggestion.
>
> In fact, I tried struct_ops in SMC about a year ago. Unfortunately, at that time struct_ops did not
> support registration from modules, and I had to move some smc dependencies into bpf, which met with
> community opposition. However, I noticed that this feature is now supported, so perhaps this is an
> opportunity.
>
> But on the other hand, given the current functionality, I wonder if struct_ops might be an overkill.
> I haven't been able to come up with a suitable abstraction to define this ops, and in the future,
> this ops might only contain the very one callback (select_syn_smc).
>
> Looking forward for your advises.

I guess I wasn't clear. It's a Nack to the current fmodret approach.

D. Wythe Oct. 14, 2024, 5:49 a.m. UTC | #4

On 10/11/24 11:37 PM, Alexei Starovoitov wrote:
> On Thu, Oct 10, 2024 at 11:44 PM D. Wythe <[e-mail]> wrote:
>>
>>
>>
>> On 10/11/24 12:21 AM, Alexei Starovoitov wrote:
>>> On Wed, Oct 9, 2024 at 8:58 PM D. Wythe <[e-mail]> wrote:
>>>>
>>>>
>>>> +__bpf_hook_start();
>>>> +
>>>> +__weak noinline int select_syn_smc(const struct sock *sk, struct sockaddr *peer)
>>>> +{
>>>> +       return 1;
>>>> +}
>>>> +
>>>> +__bpf_hook_end();
>>>> +
>>>>    int smc_nl_dump_hs_limitation(struct sk_buff *skb, struct netlink_callback *cb)
>>>>    {
>>>>           struct smc_nl_dmp_ctx *cb_ctx = smc_nl_dmp_ctx(cb);
>>>> @@ -156,19 +165,43 @@ static struct sock *smc_tcp_syn_recv_sock(const struct sock *sk,
>>>>           return NULL;
>>>>    }
>>>>
>>>> -static bool smc_hs_congested(const struct sock *sk)
>>>> +static void smc_openreq_init(struct request_sock *req,
>>>> +                            const struct tcp_options_received *rx_opt,
>>>> +                            struct sk_buff *skb, const struct sock *sk)
>>>>    {
>>>> +       struct inet_request_sock *ireq = inet_rsk(req);
>>>> +       struct sockaddr_storage rmt_sockaddr = {};
>>>>           const struct smc_sock *smc;
>>>>
>>>>           smc = smc_clcsock_user_data(sk);
>>>>
>>>>           if (!smc)
>>>> -               return true;
>>>> +               return;
>>>>
>>>> -       if (workqueue_congested(WORK_CPU_UNBOUND, smc_hs_wq))
>>>> -               return true;
>>>> +       if (smc->limit_smc_hs && workqueue_congested(WORK_CPU_UNBOUND, smc_hs_wq))
>>>> +               goto out_no_smc;
>>>>
>>>> -       return false;
>>>> +       rmt_sockaddr.ss_family = sk->sk_family;
>>>> +
>>>> +       if (rmt_sockaddr.ss_family == AF_INET) {
>>>> +               struct sockaddr_in *rmt4_sockaddr =  (struct sockaddr_in *)&rmt_sockaddr;
>>>> +
>>>> +               rmt4_sockaddr->sin_addr.s_addr = ireq->ir_rmt_addr;
>>>> +               rmt4_sockaddr->sin_port = ireq->ir_rmt_port;
>>>> +#if IS_ENABLED(CONFIG_IPV6)
>>>> +       } else {
>>>> +               struct sockaddr_in6 *rmt6_sockaddr =  (struct sockaddr_in6 *)&rmt_sockaddr;
>>>> +
>>>> +               rmt6_sockaddr->sin6_addr = ireq->ir_v6_rmt_addr;
>>>> +               rmt6_sockaddr->sin6_port = ireq->ir_rmt_port;
>>>> +#endif /* CONFIG_IPV6 */
>>>> +       }
>>>> +
>>>> +       ireq->smc_ok = select_syn_smc(sk, (struct sockaddr *)&rmt_sockaddr);
>>>> +       return;
>>>> +out_no_smc:
>>>> +       ireq->smc_ok = 0;
>>>> +       return;
>>>>    }
>>>>
>>>>    struct smc_hashinfo smc_v4_hashinfo = {
>>>> @@ -1671,7 +1704,7 @@ int smc_connect(struct socket *sock, struct sockaddr *addr,
>>>>           }
>>>>
>>>>           smc_copy_sock_settings_to_clc(smc);
>>>> -       tcp_sk(smc->clcsock->sk)->syn_smc = 1;
>>>> +       tcp_sk(smc->clcsock->sk)->syn_smc = select_syn_smc(sk, addr);
>>>>           if (smc->connect_nonblock) {
>>>>                   rc = -EALREADY;
>>>>                   goto out;
>>>> @@ -2650,8 +2683,7 @@ int smc_listen(struct socket *sock, int backlog)
>>>>
>>>>           inet_csk(smc->clcsock->sk)->icsk_af_ops = &smc->af_ops;
>>>>
>>>> -       if (smc->limit_smc_hs)
>>>> -               tcp_sk(smc->clcsock->sk)->smc_hs_congested = smc_hs_congested;
>>>> +       tcp_sk(smc->clcsock->sk)->smc_openreq_init = smc_openreq_init;
>>>>
>>>>           rc = kernel_listen(smc->clcsock, backlog);
>>>>           if (rc) {
>>>> @@ -3475,6 +3507,24 @@ static void __net_exit smc_net_stat_exit(struct net *net)
>>>>           .exit = smc_net_stat_exit,
>>>>    };
>>>>
>>>> +#if IS_ENABLED(CONFIG_BPF_SYSCALL)
>>>> +BTF_SET8_START(bpf_smc_fmodret_ids)
>>>> +BTF_ID_FLAGS(func, select_syn_smc)
>>>> +BTF_SET8_END(bpf_smc_fmodret_ids)
>>>> +
>>>> +static const struct btf_kfunc_id_set bpf_smc_fmodret_set = {
>>>> +       .owner = THIS_MODULE,
>>>> +       .set   = &bpf_smc_fmodret_ids,
>>>> +};
>>>> +
>>>> +static int bpf_smc_kfunc_init(void)
>>>> +{
>>>> +       return register_btf_fmodret_id_set(&bpf_smc_fmodret_set);
>>>> +}
>>>
>>> fmodret was an approach that hid-bpf took initially,
>>> but eventually they removed it all and switched to struct-ops approach.
>>> Please learn that lesson.
>>> Use struct_ops from the beginning.
>>>
>>> I did a presentation recently explaining the motivation behind
>>> struct_ops and tips on how to extend the kernel.
>>> TLDR: the step one is to design the extension _without_ bpf.
>>> The interface should be usable for kernel modules.
>>> And then when you have *_ops style api in place
>>> the bpf progs will plug-in without extra work.
>>>
>>> Slides:
>>> https://github.com/4ast/docs/blob/main/BPF%20struct-ops.pdf
>>
>>
>> Hi Alexei,
>>
>> Thanks very much for your suggestion.
>>
>> In fact, I tried struct_ops in SMC about a year ago. Unfortunately, at that time struct_ops did not
>> support registration from modules, and I had to move some smc dependencies into bpf, which met with
>> community opposition. However, I noticed that this feature is now supported, so perhaps this is an
>> opportunity.
>>
>> But on the other hand, given the current functionality, I wonder if struct_ops might be an overkill.
>> I haven't been able to come up with a suitable abstraction to define this ops, and in the future,
>> this ops might only contain the very one callback (select_syn_smc).
>>
>> Looking forward for your advises.
> 
> I guess I wasn't clear. It's a Nack to the current fmodret approach.


Understood, we do not oppose the use of struct_ops, especially when modules registration
was already supported.

Patch

diff --git a/include/linux/tcp.h b/include/linux/tcp.h
index 6a5e08b..d028d76 100644
--- a/include/linux/tcp.h
+++ b/include/linux/tcp.h
@@ -478,7 +478,9 @@  struct tcp_sock {
 #endif
 #if IS_ENABLED(CONFIG_SMC)
 	bool	syn_smc;	/* SYN includes SMC */
-	bool	(*smc_hs_congested)(const struct sock *sk);
+	void	(*smc_openreq_init)(struct request_sock *req,
+			     const struct tcp_options_received *rx_opt,
+			     struct sk_buff *skb, const struct sock *sk);
 #endif
 
 #if defined(CONFIG_TCP_MD5SIG) || defined(CONFIG_TCP_AO)
diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c
index cc05ec1..15fe8b9 100644
--- a/net/ipv4/tcp_input.c
+++ b/net/ipv4/tcp_input.c
@@ -7036,8 +7036,8 @@  static void tcp_openreq_init(struct request_sock *req,
 	ireq->ir_num = ntohs(tcp_hdr(skb)->dest);
 	ireq->ir_mark = inet_request_mark(sk, skb);
 #if IS_ENABLED(CONFIG_SMC)
-	ireq->smc_ok = rx_opt->smc_ok && !(tcp_sk(sk)->smc_hs_congested &&
-			tcp_sk(sk)->smc_hs_congested(sk));
+	if (rx_opt->smc_ok && tcp_sk(sk)->smc_openreq_init)
+		tcp_sk(sk)->smc_openreq_init(req, rx_opt, skb, sk);
 #endif
 }
 
diff --git a/net/smc/af_smc.c b/net/smc/af_smc.c
index 0316217..550799c 100644
--- a/net/smc/af_smc.c
+++ b/net/smc/af_smc.c
@@ -70,6 +70,15 @@ 
 static void smc_tcp_listen_work(struct work_struct *);
 static void smc_connect_work(struct work_struct *);
 
+__bpf_hook_start();
+
+__weak noinline int select_syn_smc(const struct sock *sk, struct sockaddr *peer)
+{
+	return 1;
+}
+
+__bpf_hook_end();
+
 int smc_nl_dump_hs_limitation(struct sk_buff *skb, struct netlink_callback *cb)
 {
 	struct smc_nl_dmp_ctx *cb_ctx = smc_nl_dmp_ctx(cb);
@@ -156,19 +165,43 @@  static struct sock *smc_tcp_syn_recv_sock(const struct sock *sk,
 	return NULL;
 }
 
-static bool smc_hs_congested(const struct sock *sk)
+static void smc_openreq_init(struct request_sock *req,
+			     const struct tcp_options_received *rx_opt,
+			     struct sk_buff *skb, const struct sock *sk)
 {
+	struct inet_request_sock *ireq = inet_rsk(req);
+	struct sockaddr_storage rmt_sockaddr = {};
 	const struct smc_sock *smc;
 
 	smc = smc_clcsock_user_data(sk);
 
 	if (!smc)
-		return true;
+		return;
 
-	if (workqueue_congested(WORK_CPU_UNBOUND, smc_hs_wq))
-		return true;
+	if (smc->limit_smc_hs && workqueue_congested(WORK_CPU_UNBOUND, smc_hs_wq))
+		goto out_no_smc;
 
-	return false;
+	rmt_sockaddr.ss_family = sk->sk_family;
+
+	if (rmt_sockaddr.ss_family == AF_INET) {
+		struct sockaddr_in *rmt4_sockaddr =  (struct sockaddr_in *)&rmt_sockaddr;
+
+		rmt4_sockaddr->sin_addr.s_addr = ireq->ir_rmt_addr;
+		rmt4_sockaddr->sin_port	= ireq->ir_rmt_port;
+#if IS_ENABLED(CONFIG_IPV6)
+	} else {
+		struct sockaddr_in6 *rmt6_sockaddr =  (struct sockaddr_in6 *)&rmt_sockaddr;
+
+		rmt6_sockaddr->sin6_addr = ireq->ir_v6_rmt_addr;
+		rmt6_sockaddr->sin6_port = ireq->ir_rmt_port;
+#endif /* CONFIG_IPV6 */
+	}
+
+	ireq->smc_ok = select_syn_smc(sk, (struct sockaddr *)&rmt_sockaddr);
+	return;
+out_no_smc:
+	ireq->smc_ok = 0;
+	return;
 }
 
 struct smc_hashinfo smc_v4_hashinfo = {
@@ -1671,7 +1704,7 @@  int smc_connect(struct socket *sock, struct sockaddr *addr,
 	}
 
 	smc_copy_sock_settings_to_clc(smc);
-	tcp_sk(smc->clcsock->sk)->syn_smc = 1;
+	tcp_sk(smc->clcsock->sk)->syn_smc = select_syn_smc(sk, addr);
 	if (smc->connect_nonblock) {
 		rc = -EALREADY;
 		goto out;
@@ -2650,8 +2683,7 @@  int smc_listen(struct socket *sock, int backlog)
 
 	inet_csk(smc->clcsock->sk)->icsk_af_ops = &smc->af_ops;
 
-	if (smc->limit_smc_hs)
-		tcp_sk(smc->clcsock->sk)->smc_hs_congested = smc_hs_congested;
+	tcp_sk(smc->clcsock->sk)->smc_openreq_init = smc_openreq_init;
 
 	rc = kernel_listen(smc->clcsock, backlog);
 	if (rc) {
@@ -3475,6 +3507,24 @@  static void __net_exit smc_net_stat_exit(struct net *net)
 	.exit = smc_net_stat_exit,
 };
 
+#if IS_ENABLED(CONFIG_BPF_SYSCALL)
+BTF_SET8_START(bpf_smc_fmodret_ids)
+BTF_ID_FLAGS(func, select_syn_smc)
+BTF_SET8_END(bpf_smc_fmodret_ids)
+
+static const struct btf_kfunc_id_set bpf_smc_fmodret_set = {
+	.owner = THIS_MODULE,
+	.set   = &bpf_smc_fmodret_ids,
+};
+
+static int bpf_smc_kfunc_init(void)
+{
+	return register_btf_fmodret_id_set(&bpf_smc_fmodret_set);
+}
+#else
+static inline int bpf_smc_kfunc_init(void) { return 0; }
+#endif /* CONFIG_BPF_SYSCALL */
+
 static int __init smc_init(void)
 {
 	int rc;
@@ -3574,8 +3624,17 @@  static int __init smc_init(void)
 		pr_err("%s: smc_inet_init fails with %d\n", __func__, rc);
 		goto out_ulp;
 	}
+
+	rc = bpf_smc_kfunc_init();
+	if (rc) {
+		pr_err("%s: bpf_smc_kfunc_init fails with %d\n", __func__, rc);
+		goto out_inet;
+	}
+
 	static_branch_enable(&tcp_have_smc);
 	return 0;
+out_inet:
+	smc_inet_exit();
 out_ulp:
 	tcp_unregister_ulp(&smc_ulp_ops);
 out_lo: