Re: [PATCH net-next v10 11/14] tcp: RX path for devmem TCP
From: Paolo Abeni <pabeni@redhat.com>
Date: 2024-06-04 10:53:42
Also in:
bpf, dri-devel, linux-alpha, linux-arch, linux-doc, linux-kselftest, linux-media, linux-mips, linux-trace-kernel, lkml, sparclinux
On Thu, 2024-05-30 at 20:16 +0000, Mina Almasry wrote:
quoted hunk ↗ jump to hunk
@@ -2317,6 +2318,213 @@ static int tcp_inq_hint(struct sock *sk) return inq; } +/* batch __xa_alloc() calls and reduce xa_lock()/xa_unlock() overhead. */ +struct tcp_xa_pool { + u8 max; /* max <= MAX_SKB_FRAGS */ + u8 idx; /* idx <= max */ + __u32 tokens[MAX_SKB_FRAGS]; + netmem_ref netmems[MAX_SKB_FRAGS]; +}; + +static void tcp_xa_pool_commit(struct sock *sk, struct tcp_xa_pool *p, + bool lock) +{ + int i; + + if (!p->max) + return; + if (lock) + xa_lock_bh(&sk->sk_user_frags);
The conditional lock here confuses sparse. I think you can avoid it providing a unlocked version (no need to check for '!p->max' the only caller wanting the unlocked version already performs such check) and a locked one, calling the other. Cheers, Paolo