From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from rtits2.realtek.com.tw (rtits2.realtek.com [211.75.126.72]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id 537B03A2540; Fri, 29 May 2026 05:44:28 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=211.75.126.72 ARC-Seal:i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1780033472; cv=none; b=EET8We71Mz/K+KfqX3zZ+3G9ommtQdzj4xPCawcGrKc2CQpOiQROYZT4/JADZVB9iVeHcnlBM2vsPJevQI1Y1IEG3vQvT4z0ZTfGMsa1q/6lMETZ+/6nx+9caENDe4Zr95HVYBIKnsZabySzd15NdANt2YmgCakHVWaMATrUECM= ARC-Message-Signature:i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1780033472; c=relaxed/simple; bh=lTVats1/3JAffJhdGb5lrzqUkzkUDggDPMsXbwHLjcA=; h=From:To:CC:Subject:Date:Message-ID:References:In-Reply-To: Content-Type:MIME-Version; b=I/7drIqmszNF/MZAmeXlaGcfIEvbrDnGkfvW7fKo1NzEA+Ri14Y8sVF8QpLh51yX/iSBwdNWivq5LLXMQZCx6YkoYCHeYhjmB1df8cHSguiaqnaEyE8knkb4S0UB1vCYAXYW5naSHZMhaRVeXUcvYefQBPvF68JSYSWhPt2B5uI= ARC-Authentication-Results:i=1; smtp.subspace.kernel.org; dmarc=pass (p=none dis=none) header.from=realsil.com.cn; spf=pass smtp.mailfrom=realsil.com.cn; dkim=pass (2048-bit key) header.d=realsil.com.cn header.i=@realsil.com.cn header.b=h3M40yN2; arc=none smtp.client-ip=211.75.126.72 Authentication-Results: smtp.subspace.kernel.org; dmarc=pass (p=none dis=none) header.from=realsil.com.cn Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=realsil.com.cn Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=realsil.com.cn header.i=@realsil.com.cn header.b="h3M40yN2" X-SpamFilter-By: ArmorX SpamTrap 5.80 with qID 64T5hqRjB3908671, This message is accepted by code: ctloc85258 DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=realsil.com.cn; s=dkim; t=1780033433; bh=rtjWEsp/o05lMGpREuo0QXlKafK0Yf4N90ixQE5WZQM=; h=From:To:CC:Subject:Date:Message-ID:References:In-Reply-To: Content-Type:Content-Transfer-Encoding:MIME-Version; b=h3M40yN2GjfbrORmd320Er8RXa0VLUxRT/GBFh8pb871376vJdyXgvPfbwsp3/tEv VTjnpdtFyRWm/x8RMrVtQfCujeuTmhk3/Qf1is0hzEwYWH3Ze5USAVc2QwfaS1Olpn QsVdAZXYUoudsgAfwfpeQV7KWhDR/CGTKbtNLzr+ajfSzGx6nvkDaoeQG4SqYEFcJY N66bOq6R7w0NQBWgalHNQVbTkiMYw8P135mZhHG4gYNfL5siJk3bCYU8kJFpsVvhhn UvzrwQfFc5lP4XO2cc8B+NheQDPyg1QkzVmPEt/DwfeILwl8dQM9ejJhaAAG09aXV8 9gh00MpcHx7ng== Received: from RS-EX-MBS4.realsil.com.cn ([172.29.17.104]) by rtits2.realtek.com.tw (8.15.2/3.28/5.94) with ESMTPS id 64T5hqRjB3908671 (version=TLSv1.2 cipher=ECDHE-RSA-AES256-GCM-SHA384 bits=256 verify=FAIL); Fri, 29 May 2026 13:43:53 +0800 Received: from RS-EX-MBS3.realsil.com.cn (172.29.17.103) by RS-EX-MBS4.realsil.com.cn (172.29.17.104) with Microsoft SMTP Server (version=TLS1_2, cipher=TLS_ECDHE_RSA_WITH_AES_256_GCM_SHA384) id 15.2.2562.17; Fri, 29 May 2026 13:43:52 +0800 Received: from RS-EX-MBS3.realsil.com.cn ([172.29.17.103]) by RS-EX-MBS3.realsil.com.cn ([172.29.17.103]) with mapi id 15.02.2562.017; Fri, 29 May 2026 13:43:52 +0800 From: Javen To: Jakub Kicinski CC: "hkallweit1@gmail.com" , "nic_swsd@realtek.com" , "andrew+netdev@lunn.ch" , "davem@davemloft.net" , "edumazet@google.com" , "pabeni@redhat.com" , "horms@kernel.org" , "netdev@vger.kernel.org" , "linux-kernel@vger.kernel.org" Subject: RE: [Patch net-next v6 1/7] r8169: add support for multi irqs Thread-Topic: [Patch net-next v6 1/7] r8169: add support for multi irqs Thread-Index: AQHc7Oc8L+opUxqbfEm3mE61Tomyi7YjrQwAgACrvJA= Date: Fri, 29 May 2026 05:43:52 +0000 Message-ID: <6f55096511784f66ae46a217014def79@realsil.com.cn> References: <20260526081117.173-1-javen_xu@realsil.com.cn> <20260526081117.173-2-javen_xu@realsil.com.cn> <20260528180004.58991104@kernel.org> In-Reply-To: <20260528180004.58991104@kernel.org> Accept-Language: zh-CN, en-US Content-Language: en-US Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: quoted-printable Precedence: bulk X-Mailing-List: linux-kernel@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 >On Tue, 26 May 2026 16:11:11 +0800 javen wrote: >> @@ -4820,7 +4838,7 @@ static int rtl_rx(struct net_device *dev, struct >rtl8169_private *tp, int budget >> goto release_descriptor; >> } >> >> - skb =3D napi_alloc_skb(&tp->napi, pkt_size); >> + skb =3D napi_alloc_skb(&tp->rtl8169_napi[0], pkt_size); > >the caller is the NAPI poll function, you should pass that NAPI as arg to = rtl_rx() >already instead of hardcoding [0] in this patch. > >> if (unlikely(!skb)) { >> dev->stats.rx_dropped++; >> goto release_descriptor; @@ -4844,7 +4862,7 @@ >> static int rtl_rx(struct net_device *dev, struct rtl8169_private *tp, in= t budget >> if (skb->pkt_type =3D=3D PACKET_MULTICAST) >> dev->stats.multicast++; >> >> - napi_gro_receive(&tp->napi, skb); >> + napi_gro_receive(&tp->rtl8169_napi[0], skb); >> >> dev_sw_netstats_rx_add(dev, pkt_size); >> release_descriptor: > >> +static int rtl8169_set_real_num_queues(struct rtl8169_private *tp) { >> + int ret; >> + >> + ret =3D netif_set_real_num_tx_queues(tp->dev, 1); >> + if (ret < 0) >> + return ret; >> + >> + return netif_set_real_num_rx_queues(tp->dev, tp->num_rx_rings); > >netif_set_real_num_queues() exists, just call it directly instead of addin= g your >own helper. > >> +} >> + >> static int rtl_jumbo_max(struct rtl8169_private *tp) { >> /* Non-GBit versions don't support jumbo frames */ @@ -5599,6 >> +5669,22 @@ static bool rtl_aspm_is_safe(struct rtl8169_private *tp) >> return false; >> } >> >> +static void r8169_del_napi_action(void *data) { >> + struct rtl8169_private *tp =3D data; >> + int i; >> + >> + for (i =3D 0; i < tp->irq_nvecs; i++) >> + netif_napi_del(&tp->rtl8169_napi[i]); >> +} >> + >> +static void r8169_init_napi(struct rtl8169_private *tp) { >> + for (int i =3D 0; i < tp->irq_nvecs; i++) >> + netif_napi_add(tp->dev, &tp->rtl8169_napi[i], rtl8169_poll= ); >> + devm_add_action_or_reset(&tp->pci_dev->dev, >> +r8169_del_napi_action, tp); > >devm_add_action_or_reset() can fail (as the AI bots point out) but this wh= ole >devm_ dance is entirely unnecessary networking stack will automatically >delete NAPI instances when device is unregistered. Thanks for your review. In patch v3, link: https://lore.kernel.org/netdev/20260513115543.1730-2-jav= en_xu@realsil.com.cn/ I tried to alloc struct rtl8169_napi dynamically for saving memory accordin= g to Heiner's suggestion. I agree with his suggestion because only 8127 rss= are enabled. And in this ai review, link: https://netdev-ai.bots.linux.dev/sashiko/#/pat= chset/20260520031603.700-1-javen_xu%40realsil.com.cn AI suggested that the lifetime of this devm_kcalloc'd napi array may be com= patible with the netdev's napi list. So I add devm_add_action_or_reset in p= atch v6. I checked the code and agree that the stack auto-deletes NAPI instances in = free_netdev() -> netdev_napi_exit(). However, because devres releases resou= rces in LIFO order: 1. kfree for the NAPI array (allocated via devm_kcalloc) will be called fir= st. 2. free_netdev() (registered via devm_alloc_etherdev) will be called second When free_netdev() calls netdev_napi_exit() to iterate over dev->napi_list,= the NAPI memory has already been freed by devm, which will cause a Use-Aft= er-Free. That's why I added the devm action to explicitly remove it before = the memory is freed. So I wanna know what should I do? Whether keep the action in this patch(dyn= amically allocate napi array) or patch v2(fix the array size), or any other= suggestion will be apperaciated. BRs, Javen