From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from smtp.kernel.org (aws-us-west-2-korg-mail-1.web.codeaurora.org [10.30.226.201]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id CAC36134DD for ; Thu, 7 Sep 2023 18:10:51 +0000 (UTC) Received: by smtp.kernel.org (Postfix) with ESMTPSA id 60D7AC433C8; Thu, 7 Sep 2023 18:10:51 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=kernel.org; s=k20201202; t=1694110251; bh=+GlcYEFOlEcjEmHh6gw6oifVtm8vLXx4BRTWPKzRheI=; h=Date:From:To:Cc:Subject:In-Reply-To:References:From; b=Tsg13g2YBmSKQiDxkDghbB47nKlLxshGwM0q0J1l2eMOjaZ9zcPTrLleZEuOsPBt3 ds4IqKjh/bXvztf1HhPQ1cT0eZWAXtop13Vd2kedHLgBYidfAYuIQRw9R9sy4BBV2w J7maewB4byc+NK2BqlNhKsvOKK9oGPde22gq3GugdpkLESjMk+nwSKualdICdkqifS GzJ8EtBH7D/Z0ciZMBH82nZ+R2hW4gCRFcl2EwpHdxQryJsS1qIpEaSj0Uq0/XbuAa VbbjkUciMdWIbpAna2LsyDWfEsYtKJdXNeDRtIBRo6jx484ImOX11dxrv9bx2ANFBd Xvkn7/9Qu+bWw== Date: Thu, 7 Sep 2023 11:10:50 -0700 From: Jakub Kicinski To: Eric Dumazet Cc: "David S . Miller" , Paolo Abeni , netdev@vger.kernel.org, eric.dumazet@gmail.com, Soheil Hassas Yeganeh , Neal Cardwell , Yuchung Cheng Subject: Re: [RFC net-next 4/4] tcp: defer regular ACK while processing socket backlog Message-ID: <20230907111050.30425753@kernel.org> In-Reply-To: References: <20230906201046.463236-1-edumazet@google.com> <20230906201046.463236-5-edumazet@google.com> <20230907100932.58daf8e5@kernel.org> <20230907110015.75fdcc5c@kernel.org> Precedence: bulk X-Mailing-List: netdev@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit On Thu, 7 Sep 2023 20:05:29 +0200 Eric Dumazet wrote: > > Interesting. Some folks at Meta were recently looking into parsing RPCs > > in the kernel to avoid unnecessary wakeups. Poor man's KCM using BPF > > sockmaps. Passing message size hints from the sender would solve so > > many problems.. > > Yes, RPC headers make things easier for sure. > > (we internally have something similar named autolowat, where we parse > headers to set sk->sk_rcvlowat dynamically) Could this be turned into a BPF hook, possibly? That's basically what I suggested the BPF people did but they like to rewrite everything.