netdev - Re: [PATCH 3/3] [RFC] Bluetooth: enable bpf TX timestamping

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <3546b79d-a09b-4971-abd7-ce18696a9536@linux.dev>
Date: Tue, 1 Apr 2025 18:34:00 -0700
From: Martin KaFai Lau <martin.lau@...ux.dev>
To: Pauli Virtanen <pav@....fi>
Cc: netdev@...r.kernel.org, bpf@...r.kernel.org,
 linux-bluetooth@...r.kernel.org, willemdebruijn.kernel@...il.com,
 kerneljasonxing@...il.com
Subject: Re: [PATCH 3/3] [RFC] Bluetooth: enable bpf TX timestamping

On 3/30/25 5:23 AM, Pauli Virtanen wrote:
> Emit timestamps also for BPF timestamping.
> 
> ***
> 
> The tskey management here is not quite right: see cover letter.
> ---
>   include/net/bluetooth/bluetooth.h |  1 +
>   net/bluetooth/hci_conn.c          | 21 +++++++++++++++++++--
>   2 files changed, 20 insertions(+), 2 deletions(-)
> 
> diff --git a/include/net/bluetooth/bluetooth.h b/include/net/bluetooth/bluetooth.h
> index bbefde319f95..3b2e59cedd2d 100644
> --- a/include/net/bluetooth/bluetooth.h
> +++ b/include/net/bluetooth/bluetooth.h
> @@ -383,6 +383,7 @@ struct bt_sock {
>   	struct list_head accept_q;
>   	struct sock *parent;
>   	unsigned long flags;
> +	atomic_t bpf_tskey;
>   	void (*skb_msg_name)(struct sk_buff *, void *, int *);
>   	void (*skb_put_cmsg)(struct sk_buff *, struct msghdr *, struct sock *);
>   };
> diff --git a/net/bluetooth/hci_conn.c b/net/bluetooth/hci_conn.c
> index 95972fd4c784..7430df1c5822 100644
> --- a/net/bluetooth/hci_conn.c
> +++ b/net/bluetooth/hci_conn.c
> @@ -28,6 +28,7 @@
>   #include <linux/export.h>
>   #include <linux/debugfs.h>
>   #include <linux/errqueue.h>
> +#include <linux/bpf-cgroup.h>
>   
>   #include <net/bluetooth/bluetooth.h>
>   #include <net/bluetooth/hci_core.h>
> @@ -3072,6 +3073,7 @@ void hci_setup_tx_timestamp(struct sk_buff *skb, size_t key_offset,
>   			    const struct sockcm_cookie *sockc)
>   {
>   	struct sock *sk = skb ? skb->sk : NULL;
> +	bool have_tskey = false;
>   
>   	/* This shall be called on a single skb of those generated by user
>   	 * sendmsg(), and only when the sendmsg() does not return error to
> @@ -3096,6 +3098,20 @@ void hci_setup_tx_timestamp(struct sk_buff *skb, size_t key_offset,
>   
>   			skb_shinfo(skb)->tskey = key - 1;
>   		}
> +		have_tskey = true;
> +	}
> +
> +	if (cgroup_bpf_enabled(CGROUP_SOCK_OPS) &&
> +	    SK_BPF_CB_FLAG_TEST(sk, SK_BPF_CB_TX_TIMESTAMPING)) {
> +		struct bt_sock *bt_sk = container_of(sk, struct bt_sock, sk);
> +		int key = atomic_inc_return(&bt_sk->bpf_tskey);

I don't think it needs to add "atomic_t bpf_tskey". Allow the bpf to decide what 
the skb_shinfo(skb)->tskey should be if it is not set by the userspace.

> +
> +		if (!have_tskey)
> +			skb_shinfo(skb)->tskey = key - 1;
> +
> +		bpf_skops_tx_timestamping(sk, skb,
> +					  BPF_SOCK_OPS_TSTAMP_SENDMSG_CB);
> +
>   	}
>   }
>