From: John Fastabend <john.fastabend@gmail.com>
To: Jakub Sitnicki <jakub@cloudflare.com>,
John Fastabend <john.fastabend@gmail.com>
Cc: bpf@vger.kernel.org, netdev@vger.kernel.org, ast@kernel.org,
daniel@iogearbox.net
Subject: Re: [bpf PATCH 3/9] bpf: sockmap/tls, push write_space updates through ulp updates
Date: Thu, 09 Jan 2020 13:22:49 -0800 [thread overview]
Message-ID: <5e1799a913185_35982ae92e9d45bcc9@john-XPS-13-9370.notmuch> (raw)
In-Reply-To: <87tv54syv8.fsf@cloudflare.com>
Jakub Sitnicki wrote:
> On Wed, Jan 08, 2020 at 10:14 PM CET, John Fastabend wrote:
> > When sockmap sock with TLS enabled is removed we cleanup bpf/psock state
> > and call tcp_update_ulp() to push updates to TLS ULP on top. However, we
> > don't push the write_space callback up and instead simply overwrite the
> > op with the psock stored previous op. This may or may not be correct so
> > to ensure we don't overwrite the TLS write space hook pass this field to
> > the ULP and have it fixup the ctx.
> >
> > This completes a previous fix that pushed the ops through to the ULP
> > but at the time missed doing this for write_space, presumably because
> > write_space TLS hook was added around the same time.
> >
> > Fixes: 95fa145479fbc ("bpf: sockmap/tls, close can race with map free")
> > Signed-off-by: John Fastabend <john.fastabend@gmail.com>
> > ---
> > include/linux/skmsg.h | 12 ++++++++----
> > include/net/tcp.h | 6 ++++--
> > net/ipv4/tcp_ulp.c | 6 ++++--
> > net/tls/tls_main.c | 10 +++++++---
> > 4 files changed, 23 insertions(+), 11 deletions(-)
> >
> > diff --git a/include/linux/skmsg.h b/include/linux/skmsg.h
> > index b6afe01f8592..14d61bba0b79 100644
> > --- a/include/linux/skmsg.h
> > +++ b/include/linux/skmsg.h
> > @@ -359,17 +359,21 @@ static inline void sk_psock_restore_proto(struct sock *sk,
> > struct sk_psock *psock)
> > {
> > sk->sk_prot->unhash = psock->saved_unhash;
> > - sk->sk_write_space = psock->saved_write_space;
> >
> > if (psock->sk_proto) {
> > struct inet_connection_sock *icsk = inet_csk(sk);
> > bool has_ulp = !!icsk->icsk_ulp_data;
> >
> > - if (has_ulp)
> > - tcp_update_ulp(sk, psock->sk_proto);
> > - else
> > + if (has_ulp) {
> > + tcp_update_ulp(sk, psock->sk_proto,
> > + psock->saved_write_space);
> > + } else {
> > sk->sk_prot = psock->sk_proto;
> > + sk->sk_write_space = psock->saved_write_space;
> > + }
>
> I'm wondering if we need the above fallback branch for no-ULP case?
> tcp_update_ulp repeats the ULP check and has the same fallback. Perhaps
> it can be reduced to:
>
> if (psock->sk_proto) {
> tcp_update_ulp(sk, psock->sk_proto, psock->saved_write_space);
> psock->sk_proto = NULL;
> } else {
> sk->sk_write_space = psock->saved_write_space;
> }
Yeah that is a bit nicer. How about pushing it for bpf-next? I'm not
sure its needed for bpf and the patch I pushed is the minimal change
needed for the fix and pushes the saved_write_space around.
>
> Then there's the question if it's okay to leave psock->sk_proto set and
> potentially restore it more than once? Reading tls_update, the only user
> ULP 'update' callback, it looks fine.
>
> Can sk_psock_restore_proto be as simple as:
>
> static inline void sk_psock_restore_proto(struct sock *sk,
> struct sk_psock *psock)
> {
> tcp_update_ulp(sk, psock->sk_proto, psock->saved_write_space);
> }
>
> ... or am I missing something?
I think that is good. bpf-next?
>
> Asking becuase I have a patch [0] like this in the queue and haven't
> seen issues with it during testing.
+1 Want to push it after we sort out this series?
>
> -jkbs
>
> [0] https://github.com/jsitnicki/linux/commit/2d2152593c8e6c5f38548796501a81a6ba20b6dc
>
> > psock->sk_proto = NULL;
> > + } else {
> > + sk->sk_write_space = psock->saved_write_space;
> > }
> > }
> >
> > diff --git a/include/net/tcp.h b/include/net/tcp.h
> > index e460ea7f767b..e6f48384dc71 100644
> > --- a/include/net/tcp.h
> > +++ b/include/net/tcp.h
> > @@ -2147,7 +2147,8 @@ struct tcp_ulp_ops {
> > /* initialize ulp */
> > int (*init)(struct sock *sk);
> > /* update ulp */
> > - void (*update)(struct sock *sk, struct proto *p);
> > + void (*update)(struct sock *sk, struct proto *p,
> > + void (*write_space)(struct sock *sk));
> > /* cleanup ulp */
> > void (*release)(struct sock *sk);
> > /* diagnostic */
> > @@ -2162,7 +2163,8 @@ void tcp_unregister_ulp(struct tcp_ulp_ops *type);
> > int tcp_set_ulp(struct sock *sk, const char *name);
> > void tcp_get_available_ulp(char *buf, size_t len);
> > void tcp_cleanup_ulp(struct sock *sk);
> > -void tcp_update_ulp(struct sock *sk, struct proto *p);
> > +void tcp_update_ulp(struct sock *sk, struct proto *p,
> > + void (*write_space)(struct sock *sk));
> >
> > #define MODULE_ALIAS_TCP_ULP(name) \
> > __MODULE_INFO(alias, alias_userspace, name); \
> > diff --git a/net/ipv4/tcp_ulp.c b/net/ipv4/tcp_ulp.c
> > index 12ab5db2b71c..38d3ad141161 100644
> > --- a/net/ipv4/tcp_ulp.c
> > +++ b/net/ipv4/tcp_ulp.c
> > @@ -99,17 +99,19 @@ void tcp_get_available_ulp(char *buf, size_t maxlen)
> > rcu_read_unlock();
> > }
> >
> > -void tcp_update_ulp(struct sock *sk, struct proto *proto)
> > +void tcp_update_ulp(struct sock *sk, struct proto *proto,
> > + void (*write_space)(struct sock *sk))
> > {
> > struct inet_connection_sock *icsk = inet_csk(sk);
> >
> > if (!icsk->icsk_ulp_ops) {
> > + sk->sk_write_space = write_space;
> > sk->sk_prot = proto;
> > return;
> > }
> >
> > if (icsk->icsk_ulp_ops->update)
> > - icsk->icsk_ulp_ops->update(sk, proto);
> > + icsk->icsk_ulp_ops->update(sk, proto, write_space);
> > }
> >
> > void tcp_cleanup_ulp(struct sock *sk)
> > diff --git a/net/tls/tls_main.c b/net/tls/tls_main.c
> > index dac24c7aa7d4..94774c0e5ff3 100644
> > --- a/net/tls/tls_main.c
> > +++ b/net/tls/tls_main.c
> > @@ -732,15 +732,19 @@ static int tls_init(struct sock *sk)
> > return rc;
> > }
> >
> > -static void tls_update(struct sock *sk, struct proto *p)
> > +static void tls_update(struct sock *sk, struct proto *p,
> > + void (*write_space)(struct sock *sk))
> > {
> > struct tls_context *ctx;
> >
> > ctx = tls_get_ctx(sk);
> > - if (likely(ctx))
> > + if (likely(ctx)) {
> > + ctx->sk_write_space = write_space;
> > ctx->sk_proto = p;
> > - else
> > + } else {
> > sk->sk_prot = p;
> > + sk->sk_write_space = write_space;
> > + }
> > }
> >
> > static int tls_get_info(const struct sock *sk, struct sk_buff *skb)
next prev parent reply other threads:[~2020-01-09 21:22 UTC|newest]
Thread overview: 23+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-01-08 21:13 [bpf PATCH 0/9] Fixes for sockmap/tls from more complex BPF progs John Fastabend
2020-01-08 21:14 ` [bpf PATCH 1/9] bpf: sockmap/tls, during free we may call tcp_bpf_unhash() in loop John Fastabend
2020-01-09 1:34 ` Song Liu
2020-01-08 21:14 ` [bpf PATCH 2/9] bpf: sockmap, ensure sock lock held during tear down John Fastabend
2020-01-09 17:10 ` Song Liu
2020-01-08 21:14 ` [bpf PATCH 3/9] bpf: sockmap/tls, push write_space updates through ulp updates John Fastabend
2020-01-09 10:33 ` Jakub Sitnicki
2020-01-09 21:22 ` John Fastabend [this message]
2020-01-10 13:40 ` Jakub Sitnicki
2020-01-08 21:14 ` [bpf PATCH 4/9] bpf: sockmap, skmsg helper overestimates push, pull, and pop bounds John Fastabend
2020-01-09 18:37 ` Song Liu
2020-01-08 21:15 ` [bpf PATCH 5/9] bpf: sockmap/tls, msg_push_data may leave end mark in place John Fastabend
2020-01-09 18:51 ` Song Liu
2020-01-08 21:15 ` [bpf PATCH 6/9] bpf: sockmap/tls, tls_sw can create a plaintext buf > encrypt buf John Fastabend
2020-01-09 23:04 ` Jonathan Lemon
2020-01-08 21:15 ` [bpf PATCH 7/9] bpf: sockmap/tls, skmsg can have wrapped skmsg that needs extra chaining John Fastabend
2020-01-09 23:13 ` Jonathan Lemon
2020-01-08 21:16 ` [bpf PATCH 8/9] bpf: sockmap/tls, tls_push_record can not handle zero length skmsg John Fastabend
2020-01-09 20:08 ` Song Liu
2020-01-09 21:25 ` John Fastabend
2020-01-10 23:20 ` John Fastabend
2020-01-08 21:16 ` [bpf PATCH 9/9] bpf: sockmap/tls, fix pop data with SK_DROP return code John Fastabend
2020-01-09 23:28 ` Jonathan Lemon
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=5e1799a913185_35982ae92e9d45bcc9@john-XPS-13-9370.notmuch \
--to=john.fastabend@gmail.com \
--cc=ast@kernel.org \
--cc=bpf@vger.kernel.org \
--cc=daniel@iogearbox.net \
--cc=jakub@cloudflare.com \
--cc=netdev@vger.kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox