[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <6880e365b84d2_334c6729452@willemb.c.googlers.com.notmuch>
Date: Wed, 23 Jul 2025 09:28:05 -0400
From: Willem de Bruijn <willemdebruijn.kernel@...il.com>
To: Pengtao He <hept.hept.hept@...il.com>,
"David S. Miller" <davem@...emloft.net>,
Eric Dumazet <edumazet@...gle.com>,
Jakub Kicinski <kuba@...nel.org>,
Paolo Abeni <pabeni@...hat.com>,
Simon Horman <horms@...nel.org>,
Willem de Bruijn <willemb@...gle.com>,
Mina Almasry <almasrymina@...gle.com>,
Jason Xing <kerneljasonxing@...il.com>,
Michal Luczaj <mhal@...x.co>,
Eric Biggers <ebiggers@...gle.com>,
Alexander Lobakin <aleksander.lobakin@...el.com>
Cc: netdev@...r.kernel.org,
linux-kernel@...r.kernel.org,
Pengtao He <hept.hept.hept@...il.com>
Subject: Re: [PATCH] net/core: fix wrong return value in __splice_segment
Pengtao He wrote:
> Return true immediately when the last segment is processed,
> without waiting for the next segment.
This can use a bit more explanation. Which unnecessary wait is
avoided.
The boolean return from __skb_splice_bits has a bit odd semantics. But
is ignored by its only caller anyway.
The relevant __splice_segment that can cause an early return is in the
frags loop. But I see no waiting operation in here.
Aside from that, the commit also should target [PATCH net-next],
assuuming that this is an optimization, not a fix.
>
> Signed-off-by: Pengtao He <hept.hept.hept@...il.com>
> ---
> net/core/skbuff.c | 3 +++
> 1 file changed, 3 insertions(+)
>
> diff --git a/net/core/skbuff.c b/net/core/skbuff.c
> index ee0274417948..cc3339ab829a 100644
> --- a/net/core/skbuff.c
> +++ b/net/core/skbuff.c
> @@ -3114,6 +3114,9 @@ static bool __splice_segment(struct page *page, unsigned int poff,
> *len -= flen;
> } while (*len && plen);
>
> + if (!*len)
> + return true;
> +
> return false;
> }
>
> --
> 2.49.0
>
Powered by blists - more mailing lists