From patchwork Mon Apr 8 09:38:02 2024 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Julien Panis X-Patchwork-Id: 13620882 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org Received: from gabe.freedesktop.org (gabe.freedesktop.org [131.252.210.177]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.lore.kernel.org (Postfix) with ESMTPS id 5BC25C67861 for ; Mon, 8 Apr 2024 09:38:21 +0000 (UTC) Received: from gabe.freedesktop.org (localhost [127.0.0.1]) by gabe.freedesktop.org (Postfix) with ESMTP id 94AB610F17B; Mon, 8 Apr 2024 09:38:18 +0000 (UTC) Authentication-Results: gabe.freedesktop.org; dkim=pass (2048-bit key; unprotected) header.d=baylibre-com.20230601.gappssmtp.com header.i=@baylibre-com.20230601.gappssmtp.com header.b="PiNFgSlT"; dkim-atps=neutral Received: from mail-ed1-f51.google.com (mail-ed1-f51.google.com [209.85.208.51]) by gabe.freedesktop.org (Postfix) with ESMTPS id 702E610ED94 for ; Mon, 8 Apr 2024 09:38:14 +0000 (UTC) Received: by mail-ed1-f51.google.com with SMTP id 4fb4d7f45d1cf-56e47843cc7so1221205a12.0 for ; Mon, 08 Apr 2024 02:38:14 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=baylibre-com.20230601.gappssmtp.com; s=20230601; t=1712569093; x=1713173893; darn=lists.freedesktop.org; h=cc:to:in-reply-to:references:message-id:content-transfer-encoding :mime-version:subject:date:from:from:to:cc:subject:date:message-id :reply-to; bh=4AGag0SLGP6ZTz9ElordkjjCIrAOBl+sGfiLsyWtvf4=; b=PiNFgSlTBezk7MetximFmeqDg6+wMDUpM0bgv2x4ty0qi7bAOPlcNC1uAqI/zcR+J1 i155tDTp5kL4W/+HJfoPto8sF1g9dQj68HytzAbG5OU0BBsuVHaZdsFN3bx2bP9plip6 byCWOhcrTKRsAKiU4PugNWLuh8qnG7fQzQasvgt7svpgFYA+y3NRqMgvngPeBUJ/tmn0 Yt4NC3nKZwPPNKVb43ADPUlobcwPmQ+fQTfNP7K0HKtygq9RNJYFKrw2YfrZcbocRay/ yIAyaPKtkxx8BbEfdm/Uu9Z5wajqxsuoOdWDcm7Q08mixDxjWEZ2tKG4lqTq84f3mwP0 KKKg== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20230601; t=1712569093; x=1713173893; h=cc:to:in-reply-to:references:message-id:content-transfer-encoding :mime-version:subject:date:from:x-gm-message-state:from:to:cc :subject:date:message-id:reply-to; bh=4AGag0SLGP6ZTz9ElordkjjCIrAOBl+sGfiLsyWtvf4=; b=UGhlJ5079id9HKMEAZtjgYYsOHmhCR67sp45AXeXsssPSu6ZPdgiNy9zhSbhcaZH/Y 43tQoCHrOHkYhdCy4di6ykIXzElYJFmzw6hBjDDL6Eg+Pxjd4OoiUZFCaP3IrE4Ui4sW 5l9XPUYoX2eWzFT+gF3WBaIgpNk/DGrt3ZRV+F1puhebPN7wTDV2F0mDS8Q58w66i9hG puHFKlIWMIFD5JF5EMd62+fudHWonxvLYmbzzp85LEx16c1D+0J90B5rtxbVZg5XYFM6 DtyAUdqEV8iAxar0/o69croBpqQGDi7Z6OE2qg7NrS7uf/5T08zw9SwUSDwjQGFK2mZ/ wpVw== X-Forwarded-Encrypted: i=1; AJvYcCUDawQkXMELCdPsuFSqH3pk3Zx4ouWf2uZzaJ8Udl2j6v6ITl5KHOjEBEuboPC4gwbEpogliPxsmLyLJnbdwQycawAz5fg0CnrinaHdyn4n X-Gm-Message-State: AOJu0YzUvqNrNR5KuevqattUXa043cvdRAa3o+09aUFxzwc9U3YzZ/mR lkupwQP7mAorUsL2tZIohtTqhe0ps1hXyof2RI9aBdk8Mu8gAkBqRQxyaRPzL2E= X-Google-Smtp-Source: AGHT+IEYUxmFx883FdaPSd6GReaIHMgqz54jpNmxo/069Pc1WV+2FfwQmDbrYRFO0OGtj7S/xf4JHw== X-Received: by 2002:a17:906:53d6:b0:a4e:8f73:6d5f with SMTP id p22-20020a17090653d600b00a4e8f736d5fmr4562108ejo.66.1712569092719; Mon, 08 Apr 2024 02:38:12 -0700 (PDT) Received: from [127.0.1.1] ([84.102.31.74]) by smtp.gmail.com with ESMTPSA id ne6-20020a1709077b8600b00a4e5a6b57a2sm4175803ejc.163.2024.04.08.02.38.10 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Mon, 08 Apr 2024 02:38:12 -0700 (PDT) From: Julien Panis Date: Mon, 08 Apr 2024 11:38:02 +0200 Subject: [PATCH net-next v8 1/3] net: ethernet: ti: Add accessors for struct k3_cppi_desc_pool members MIME-Version: 1.0 Message-Id: <20240223-am65-cpsw-xdp-basic-v8-1-f3421b58da09@baylibre.com> References: <20240223-am65-cpsw-xdp-basic-v8-0-f3421b58da09@baylibre.com> In-Reply-To: <20240223-am65-cpsw-xdp-basic-v8-0-f3421b58da09@baylibre.com> To: "David S. Miller" , Eric Dumazet , Jakub Kicinski , Paolo Abeni , Russell King , Alexei Starovoitov , Daniel Borkmann , Jesper Dangaard Brouer , John Fastabend , Sumit Semwal , =?utf-8?q?Christian_K=C3=B6nig?= , Simon Horman , Andrew Lunn , Ratheesh Kannoth , Naveen Mamindlapalli Cc: danishanwar@ti.com, yuehaibing@huawei.com, rogerq@kernel.org, netdev@vger.kernel.org, linux-kernel@vger.kernel.org, bpf@vger.kernel.org, linux-media@vger.kernel.org, dri-devel@lists.freedesktop.org, linaro-mm-sig@lists.linaro.org, Julien Panis X-Mailer: b4 0.12.2 X-Developer-Signature: v=1; a=ed25519-sha256; t=1712569087; l=1876; i=jpanis@baylibre.com; s=20230526; h=from:subject:message-id; bh=fA9wRh9BhzewgFnx03QdynOb7Uu7To/l82seA5HFPKw=; b=4WAESX8DEdgb0LHBGuXuYqy64h3h9/4ecp+e+9P4Pv/WoeQSh43Dcrr0OIqVYtl+8ywfstiNK 3HM2D23qidDBgm9m1M1r9ppXZpD07onXSmxFAwvZvl8YqJHO9eSzosu X-Developer-Key: i=jpanis@baylibre.com; a=ed25519; pk=8eSM4/xkiHWz2M1Cw1U3m2/YfPbsUdEJPCWY3Mh9ekQ= X-BeenThere: dri-devel@lists.freedesktop.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Direct Rendering Infrastructure - Development List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: dri-devel-bounces@lists.freedesktop.org Sender: "dri-devel" This patch adds accessors for desc_size and cpumem members. They may be used, for instance, to compute a descriptor index. Signed-off-by: Julien Panis --- drivers/net/ethernet/ti/k3-cppi-desc-pool.c | 12 ++++++++++++ drivers/net/ethernet/ti/k3-cppi-desc-pool.h | 2 ++ 2 files changed, 14 insertions(+) diff --git a/drivers/net/ethernet/ti/k3-cppi-desc-pool.c b/drivers/net/ethernet/ti/k3-cppi-desc-pool.c index 05cc7aab1ec8..414bcac9dcc6 100644 --- a/drivers/net/ethernet/ti/k3-cppi-desc-pool.c +++ b/drivers/net/ethernet/ti/k3-cppi-desc-pool.c @@ -132,5 +132,17 @@ size_t k3_cppi_desc_pool_avail(struct k3_cppi_desc_pool *pool) } EXPORT_SYMBOL_GPL(k3_cppi_desc_pool_avail); +size_t k3_cppi_desc_pool_desc_size(const struct k3_cppi_desc_pool *pool) +{ + return pool->desc_size; +} +EXPORT_SYMBOL_GPL(k3_cppi_desc_pool_desc_size); + +void *k3_cppi_desc_pool_cpuaddr(const struct k3_cppi_desc_pool *pool) +{ + return pool->cpumem; +} +EXPORT_SYMBOL_GPL(k3_cppi_desc_pool_cpuaddr); + MODULE_LICENSE("GPL"); MODULE_DESCRIPTION("TI K3 CPPI5 descriptors pool API"); diff --git a/drivers/net/ethernet/ti/k3-cppi-desc-pool.h b/drivers/net/ethernet/ti/k3-cppi-desc-pool.h index a7e3fa5e7b62..3c6aed0bed71 100644 --- a/drivers/net/ethernet/ti/k3-cppi-desc-pool.h +++ b/drivers/net/ethernet/ti/k3-cppi-desc-pool.h @@ -26,5 +26,7 @@ k3_cppi_desc_pool_dma2virt(struct k3_cppi_desc_pool *pool, dma_addr_t dma); void *k3_cppi_desc_pool_alloc(struct k3_cppi_desc_pool *pool); void k3_cppi_desc_pool_free(struct k3_cppi_desc_pool *pool, void *addr); size_t k3_cppi_desc_pool_avail(struct k3_cppi_desc_pool *pool); +size_t k3_cppi_desc_pool_desc_size(const struct k3_cppi_desc_pool *pool); +void *k3_cppi_desc_pool_cpuaddr(const struct k3_cppi_desc_pool *pool); #endif /* K3_CPPI_DESC_POOL_H_ */ From patchwork Mon Apr 8 09:38:03 2024 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Julien Panis X-Patchwork-Id: 13620881 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org Received: from gabe.freedesktop.org (gabe.freedesktop.org [131.252.210.177]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.lore.kernel.org (Postfix) with ESMTPS id 55017CD128A for ; Mon, 8 Apr 2024 09:38:19 +0000 (UTC) Received: from gabe.freedesktop.org (localhost [127.0.0.1]) by gabe.freedesktop.org (Postfix) with ESMTP id 1F97D10ED94; Mon, 8 Apr 2024 09:38:18 +0000 (UTC) Authentication-Results: gabe.freedesktop.org; dkim=pass (2048-bit key; unprotected) header.d=baylibre-com.20230601.gappssmtp.com header.i=@baylibre-com.20230601.gappssmtp.com header.b="tSyN0zY7"; dkim-atps=neutral Received: from mail-ej1-f51.google.com (mail-ej1-f51.google.com [209.85.218.51]) by gabe.freedesktop.org (Postfix) with ESMTPS id EBA2610ED94 for ; Mon, 8 Apr 2024 09:38:16 +0000 (UTC) Received: by mail-ej1-f51.google.com with SMTP id a640c23a62f3a-a51d90def9eso70418666b.0 for ; Mon, 08 Apr 2024 02:38:16 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=baylibre-com.20230601.gappssmtp.com; s=20230601; t=1712569095; x=1713173895; darn=lists.freedesktop.org; h=cc:to:in-reply-to:references:message-id:content-transfer-encoding :mime-version:subject:date:from:from:to:cc:subject:date:message-id :reply-to; bh=wKnYVxHG6Ypqd9VKY8Y3OU7R3wT68BUjHb822YWnywA=; b=tSyN0zY7J7ircyyG0ijWmv9wdmWGw1kf1iwpU+KITMFp27UtjFQTyLuKAGyyZdVraP H57CnrF66W1d2xhArk1uS+LCNQmmWtGnq5/uqC857O8wK9/L5QodceHJQEhsO1PcCbqc 3p+XA/ArWS5+mt9rOyaDiT343C1ziNvf/+rsNQBUau+hE27w3zCrbACvreLmu4eJkkiH KluSH5kXdbrlYxMYvl4y85bLm+ZmBzyjoOQB33Zk4Qdbz1CQIee3srxIwf1oKiO77X8m ezlDSDBZgjWHaN063NwDRMDwwlV/hUrHRP1M1LBF8eDFl3HNgqrTyQ1LnexLe4bBiR2H SZCA== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20230601; t=1712569095; x=1713173895; h=cc:to:in-reply-to:references:message-id:content-transfer-encoding :mime-version:subject:date:from:x-gm-message-state:from:to:cc :subject:date:message-id:reply-to; bh=wKnYVxHG6Ypqd9VKY8Y3OU7R3wT68BUjHb822YWnywA=; b=Hu5Uhe6LJcmaufys8R3q7zA6H+qVxvlS3Lbj6BoApXn4EYZ8IfIHTxXch2vTDnqu1l 7lRDzR9uDAGjQugbEK3E49OAO80RexkrcfthQlgq/s51ErxlCHk2yxHOtWA0Z/1yYYVc MX9s4m/feYcYO8fv19N3tKH0qmomrpEH8kO8pjrsDZDNKKSowNp2E3DexHaWYDWJ59q9 0wQADZ7dCS7pSpnjfEkS/kRtZqvdOXfCnq7Yn+imLYox98rjyobLn9OjnOhHuB9P06Oi xc4rGxd36WDOMCiBc3bEBuis4zhBZ6TRjAlPrmgJHQ58CNYKbA4wSXtFpUPjHJ1vFfzW hx8g== X-Forwarded-Encrypted: i=1; AJvYcCVCto6FO/nB11OcGWztDapO6f3lVDSRTbp7xYtl2cxLSZ7vmiQ9JOi5R42AJDGIGSYf+dRZUD7oqd71ZNTKLj+2zT6/9ZSIPYk61sQfEY/o X-Gm-Message-State: AOJu0YwXHeL9sQOcWXqYO6it8+TehjHy7yrwPqNIyoaLT3wQjdCXrOcV NZHLN2FHA3Bgy9IrJE0ZbqTvgHtO/bJjm4AXB+elMBG9J3elT7Op2t8qjE2lQoc= X-Google-Smtp-Source: AGHT+IHgl+gOBCC18ARiTO3fEThbY4OkUyNz4i7KX80xwmn8UN/J5ysnLiL+7vakmrwgfOE5pdl2/Q== X-Received: by 2002:a17:906:c29a:b0:a51:b463:6b62 with SMTP id r26-20020a170906c29a00b00a51b4636b62mr4084293ejz.55.1712569095151; Mon, 08 Apr 2024 02:38:15 -0700 (PDT) Received: from [127.0.1.1] ([84.102.31.74]) by smtp.gmail.com with ESMTPSA id ne6-20020a1709077b8600b00a4e5a6b57a2sm4175803ejc.163.2024.04.08.02.38.12 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Mon, 08 Apr 2024 02:38:14 -0700 (PDT) From: Julien Panis Date: Mon, 08 Apr 2024 11:38:03 +0200 Subject: [PATCH net-next v8 2/3] net: ethernet: ti: Add desc_infos member to struct k3_cppi_desc_pool MIME-Version: 1.0 Message-Id: <20240223-am65-cpsw-xdp-basic-v8-2-f3421b58da09@baylibre.com> References: <20240223-am65-cpsw-xdp-basic-v8-0-f3421b58da09@baylibre.com> In-Reply-To: <20240223-am65-cpsw-xdp-basic-v8-0-f3421b58da09@baylibre.com> To: "David S. Miller" , Eric Dumazet , Jakub Kicinski , Paolo Abeni , Russell King , Alexei Starovoitov , Daniel Borkmann , Jesper Dangaard Brouer , John Fastabend , Sumit Semwal , =?utf-8?q?Christian_K=C3=B6nig?= , Simon Horman , Andrew Lunn , Ratheesh Kannoth , Naveen Mamindlapalli Cc: danishanwar@ti.com, yuehaibing@huawei.com, rogerq@kernel.org, netdev@vger.kernel.org, linux-kernel@vger.kernel.org, bpf@vger.kernel.org, linux-media@vger.kernel.org, dri-devel@lists.freedesktop.org, linaro-mm-sig@lists.linaro.org, Julien Panis X-Mailer: b4 0.12.2 X-Developer-Signature: v=1; a=ed25519-sha256; t=1712569087; l=3592; i=jpanis@baylibre.com; s=20230526; h=from:subject:message-id; bh=egsyeRYCkGmNZpNd4TFDJy0XrmDqk+g4czEF0/FsO7g=; b=EZWf2Dw1rLqJZCiv70aLLxj8z1Gtfoeh3NwBjUbr82+Ic+2uVOIbqWQyYlOfAKdNRJ20yticv 2WUxdg4OoUOC53dXGo6QsqD4OWoGu7P4b1nBE7uz2zXxD2EYb7qfdcg X-Developer-Key: i=jpanis@baylibre.com; a=ed25519; pk=8eSM4/xkiHWz2M1Cw1U3m2/YfPbsUdEJPCWY3Mh9ekQ= X-BeenThere: dri-devel@lists.freedesktop.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Direct Rendering Infrastructure - Development List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: dri-devel-bounces@lists.freedesktop.org Sender: "dri-devel" This patch introduces a member and the related accessors which can be used to store descriptor specific additional information. This member can store, for instance, an ID to differentiate a skb TX buffer type from a xdpf TX buffer type. Signed-off-by: Julien Panis --- drivers/net/ethernet/ti/k3-cppi-desc-pool.c | 29 +++++++++++++++++++++++++++++ drivers/net/ethernet/ti/k3-cppi-desc-pool.h | 4 ++++ 2 files changed, 33 insertions(+) diff --git a/drivers/net/ethernet/ti/k3-cppi-desc-pool.c b/drivers/net/ethernet/ti/k3-cppi-desc-pool.c index 414bcac9dcc6..e79d039fdaa5 100644 --- a/drivers/net/ethernet/ti/k3-cppi-desc-pool.c +++ b/drivers/net/ethernet/ti/k3-cppi-desc-pool.c @@ -22,6 +22,7 @@ struct k3_cppi_desc_pool { size_t mem_size; size_t num_desc; struct gen_pool *gen_pool; + void **desc_infos; }; void k3_cppi_desc_pool_destroy(struct k3_cppi_desc_pool *pool) @@ -37,6 +38,8 @@ void k3_cppi_desc_pool_destroy(struct k3_cppi_desc_pool *pool) dma_free_coherent(pool->dev, pool->mem_size, pool->cpumem, pool->dma_addr); + kfree(pool->desc_infos); + gen_pool_destroy(pool->gen_pool); /* frees pool->name */ } EXPORT_SYMBOL_GPL(k3_cppi_desc_pool_destroy); @@ -72,6 +75,16 @@ k3_cppi_desc_pool_create_name(struct device *dev, size_t size, goto gen_pool_create_fail; } + pool->desc_infos = kcalloc(pool->num_desc, + sizeof(*pool->desc_infos), GFP_KERNEL); + if (!pool->desc_infos) { + ret = -ENOMEM; + dev_err(pool->dev, + "pool descriptor infos alloc failed %d\n", ret); + kfree_const(pool_name); + goto gen_pool_desc_infos_alloc_fail; + } + pool->gen_pool->name = pool_name; pool->cpumem = dma_alloc_coherent(pool->dev, pool->mem_size, @@ -94,6 +107,8 @@ k3_cppi_desc_pool_create_name(struct device *dev, size_t size, dma_free_coherent(pool->dev, pool->mem_size, pool->cpumem, pool->dma_addr); dma_alloc_fail: + kfree(pool->desc_infos); +gen_pool_desc_infos_alloc_fail: gen_pool_destroy(pool->gen_pool); /* frees pool->name */ gen_pool_create_fail: devm_kfree(pool->dev, pool); @@ -144,5 +159,19 @@ void *k3_cppi_desc_pool_cpuaddr(const struct k3_cppi_desc_pool *pool) } EXPORT_SYMBOL_GPL(k3_cppi_desc_pool_cpuaddr); +void k3_cppi_desc_pool_desc_info_set(struct k3_cppi_desc_pool *pool, + int desc_idx, void *info) +{ + pool->desc_infos[desc_idx] = info; +} +EXPORT_SYMBOL_GPL(k3_cppi_desc_pool_desc_info_set); + +void *k3_cppi_desc_pool_desc_info(const struct k3_cppi_desc_pool *pool, + int desc_idx) +{ + return pool->desc_infos[desc_idx]; +} +EXPORT_SYMBOL_GPL(k3_cppi_desc_pool_desc_info); + MODULE_LICENSE("GPL"); MODULE_DESCRIPTION("TI K3 CPPI5 descriptors pool API"); diff --git a/drivers/net/ethernet/ti/k3-cppi-desc-pool.h b/drivers/net/ethernet/ti/k3-cppi-desc-pool.h index 3c6aed0bed71..851d352b338b 100644 --- a/drivers/net/ethernet/ti/k3-cppi-desc-pool.h +++ b/drivers/net/ethernet/ti/k3-cppi-desc-pool.h @@ -28,5 +28,9 @@ void k3_cppi_desc_pool_free(struct k3_cppi_desc_pool *pool, void *addr); size_t k3_cppi_desc_pool_avail(struct k3_cppi_desc_pool *pool); size_t k3_cppi_desc_pool_desc_size(const struct k3_cppi_desc_pool *pool); void *k3_cppi_desc_pool_cpuaddr(const struct k3_cppi_desc_pool *pool); +void k3_cppi_desc_pool_desc_info_set(struct k3_cppi_desc_pool *pool, + int desc_idx, void *info); +void *k3_cppi_desc_pool_desc_info(const struct k3_cppi_desc_pool *pool, + int desc_idx); #endif /* K3_CPPI_DESC_POOL_H_ */ From patchwork Mon Apr 8 09:38:04 2024 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Julien Panis X-Patchwork-Id: 13620883 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org Received: from gabe.freedesktop.org (gabe.freedesktop.org [131.252.210.177]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.lore.kernel.org (Postfix) with ESMTPS id 12FEEC67861 for ; Mon, 8 Apr 2024 09:38:25 +0000 (UTC) Received: from gabe.freedesktop.org (localhost [127.0.0.1]) by gabe.freedesktop.org (Postfix) with ESMTP id 26A8710F53B; Mon, 8 Apr 2024 09:38:24 +0000 (UTC) Authentication-Results: gabe.freedesktop.org; dkim=pass (2048-bit key; unprotected) header.d=baylibre-com.20230601.gappssmtp.com header.i=@baylibre-com.20230601.gappssmtp.com header.b="hHh3q6hJ"; dkim-atps=neutral Received: from mail-ed1-f47.google.com (mail-ed1-f47.google.com [209.85.208.47]) by gabe.freedesktop.org (Postfix) with ESMTPS id 2651610FDDD for ; Mon, 8 Apr 2024 09:38:20 +0000 (UTC) Received: by mail-ed1-f47.google.com with SMTP id 4fb4d7f45d1cf-56e346224bdso2244425a12.1 for ; Mon, 08 Apr 2024 02:38:20 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=baylibre-com.20230601.gappssmtp.com; s=20230601; t=1712569098; x=1713173898; darn=lists.freedesktop.org; h=cc:to:in-reply-to:references:message-id:content-transfer-encoding :mime-version:subject:date:from:from:to:cc:subject:date:message-id :reply-to; bh=3EX13voEuFOCXB84/e+34NypJwHVVQ76fT/pzlEGXOQ=; b=hHh3q6hJAt/QpK+AHayrd2pzfXgaXAj69cgNL89B47fUHydLx1qAxXSciNSO/Q8Z8m wfERmOohlzrfw0KugLKpuFLqoputuwZ/3VH2tZE9K+cc6QQtQUmIUocpQkiaY/Aj3YvJ TAiM5UvUZvH86Ah2xibIxcXmM818yHj49TicW6Vluq+bnfjitlUhdE+6bZG6RcaqTlMH HTYBpn+Aiy8kdw+wdEOFKEQfQpcENXpRGm10IXflYlfDc6VpaFMgaU9Hrk+lzcb/D7zT d3zOTAej+J5Ut7QbG8jnDBzoTzuxInYvqz1qqbXn4tWVu3QLJw/xoNi2BDgt+ERSzy9O pdKw== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20230601; t=1712569098; x=1713173898; h=cc:to:in-reply-to:references:message-id:content-transfer-encoding :mime-version:subject:date:from:x-gm-message-state:from:to:cc :subject:date:message-id:reply-to; bh=3EX13voEuFOCXB84/e+34NypJwHVVQ76fT/pzlEGXOQ=; b=Kb1O8Q1P6tZNN62ADWO5B4XktNBuLXw6vjDVsQWoCRgbEgc/cupC/IZvPrn41xqAcs bmMiT1NBwrPqm9ARWvSDimfebIXS4NEujNvXmw3WuWA/Edx8ohI8ZrmFq+rY4QpZueqK 0nHOIUkVBDGaY+Kuxb/AFhcpC3OzWxEuXnBBuIzPNV1WdeAErsBZRBbSfT5B4KGpN3Is zr/bgUC2vOKN1Cm1HVBLzMAtHgES7eaZSX+/srjsxsGMIlmLnnkt1vlLWTZf8fwJv53V CPb+2h+2P3J/Qcc3wjBijt24C4tf4wolWYTeFHcIwQuMHXdvogNIPw5L8QHQqN1ZyBr7 /qiw== X-Forwarded-Encrypted: i=1; AJvYcCUiHwYbUXhRFUD1Mqn8K8u7Xrxson8K82awfohJlXPGsSn7TFpuBdLPJTNlxjh71PFgwOoVMjuQnaGVLz/ua5T2ZklJ1RlEdvJy3AINdO8d X-Gm-Message-State: AOJu0Yxr/NMDsbgB+MUAB7bM+ktkxJ6uFhmmOvRxedmLDGUyRXZOeYzF K+v9Jq/Q0uPGqwNmQ5aBjagndpFNyxZVtYeLByNPlxRj323jSRuB/u3TOdmwG+I= X-Google-Smtp-Source: AGHT+IGmLymbdCIbmaX+OXEDUuZxjfB7x2nmsA5hMvEgJ1EOTY8RMgwc1sLoZFCMRmkBoX7w/xyfYw== X-Received: by 2002:a17:907:31c9:b0:a51:dd50:6a with SMTP id xf9-20020a17090731c900b00a51dd50006amr1001312ejb.66.1712569097969; Mon, 08 Apr 2024 02:38:17 -0700 (PDT) Received: from [127.0.1.1] ([84.102.31.74]) by smtp.gmail.com with ESMTPSA id ne6-20020a1709077b8600b00a4e5a6b57a2sm4175803ejc.163.2024.04.08.02.38.15 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Mon, 08 Apr 2024 02:38:17 -0700 (PDT) From: Julien Panis Date: Mon, 08 Apr 2024 11:38:04 +0200 Subject: [PATCH net-next v8 3/3] net: ethernet: ti: am65-cpsw: Add minimal XDP support MIME-Version: 1.0 Message-Id: <20240223-am65-cpsw-xdp-basic-v8-3-f3421b58da09@baylibre.com> References: <20240223-am65-cpsw-xdp-basic-v8-0-f3421b58da09@baylibre.com> In-Reply-To: <20240223-am65-cpsw-xdp-basic-v8-0-f3421b58da09@baylibre.com> To: "David S. Miller" , Eric Dumazet , Jakub Kicinski , Paolo Abeni , Russell King , Alexei Starovoitov , Daniel Borkmann , Jesper Dangaard Brouer , John Fastabend , Sumit Semwal , =?utf-8?q?Christian_K=C3=B6nig?= , Simon Horman , Andrew Lunn , Ratheesh Kannoth , Naveen Mamindlapalli Cc: danishanwar@ti.com, yuehaibing@huawei.com, rogerq@kernel.org, netdev@vger.kernel.org, linux-kernel@vger.kernel.org, bpf@vger.kernel.org, linux-media@vger.kernel.org, dri-devel@lists.freedesktop.org, linaro-mm-sig@lists.linaro.org, Julien Panis X-Mailer: b4 0.12.2 X-Developer-Signature: v=1; a=ed25519-sha256; t=1712569087; l=31457; i=jpanis@baylibre.com; s=20230526; h=from:subject:message-id; bh=sSV1L/eXlfMgiCDi1QDWRTNxszRkvwa6W7kr3TC0pTI=; b=1xh2nvD00JLU6V/XcT2HpRQNh7rOpwZPUyeu6ATD2E7VrMFE5EUTn2JqOoB+Y1yhkuFo+KsBY 5pjAOOP11y4DjTs3qhGBrFIK1pv78Z1w6hDY/m/RmjXJiJScxZGazp3 X-Developer-Key: i=jpanis@baylibre.com; a=ed25519; pk=8eSM4/xkiHWz2M1Cw1U3m2/YfPbsUdEJPCWY3Mh9ekQ= X-BeenThere: dri-devel@lists.freedesktop.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Direct Rendering Infrastructure - Development List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: dri-devel-bounces@lists.freedesktop.org Sender: "dri-devel" This patch adds XDP (eXpress Data Path) support to TI AM65 CPSW Ethernet driver. The following features are implemented: - NETDEV_XDP_ACT_BASIC (XDP_PASS, XDP_TX, XDP_DROP, XDP_ABORTED) - NETDEV_XDP_ACT_REDIRECT (XDP_REDIRECT) - NETDEV_XDP_ACT_NDO_XMIT (ndo_xdp_xmit callback) The page pool memory model is used to get better performance. Below are benchmark results obtained for the receiver with iperf3 default parameters: - Without page pool: 495 Mbits/sec - With page pool: 505 Mbits/sec (actually 510 Mbits/sec, with a 5 Mbits/sec loss due to extra processing in the hot path to handle XDP). Signed-off-by: Julien Panis --- drivers/net/ethernet/ti/am65-cpsw-nuss.c | 569 ++++++++++++++++++++++++++++--- drivers/net/ethernet/ti/am65-cpsw-nuss.h | 13 + 2 files changed, 531 insertions(+), 51 deletions(-) diff --git a/drivers/net/ethernet/ti/am65-cpsw-nuss.c b/drivers/net/ethernet/ti/am65-cpsw-nuss.c index 9d2f4ac783e4..98c2e0cd5087 100644 --- a/drivers/net/ethernet/ti/am65-cpsw-nuss.c +++ b/drivers/net/ethernet/ti/am65-cpsw-nuss.c @@ -5,6 +5,7 @@ * */ +#include #include #include #include @@ -30,6 +31,7 @@ #include #include #include +#include #include #include "cpsw_ale.h" @@ -138,6 +140,18 @@ #define AM65_CPSW_DEFAULT_TX_CHNS 8 +/* CPPI streaming packet interface */ +#define AM65_CPSW_CPPI_TX_FLOW_ID 0x3FFF +#define AM65_CPSW_CPPI_TX_PKT_TYPE 0x7 + +/* XDP */ +#define AM65_CPSW_XDP_CONSUMED 2 +#define AM65_CPSW_XDP_REDIRECT 1 +#define AM65_CPSW_XDP_PASS 0 + +/* Include headroom compatible with both skb and xdpf */ +#define AM65_CPSW_HEADROOM max(NET_SKB_PAD, XDP_PACKET_HEADROOM) + static void am65_cpsw_port_set_sl_mac(struct am65_cpsw_port *slave, const u8 *dev_addr) { @@ -369,6 +383,66 @@ static void am65_cpsw_init_host_port_emac(struct am65_cpsw_common *common); static void am65_cpsw_init_port_switch_ale(struct am65_cpsw_port *port); static void am65_cpsw_init_port_emac_ale(struct am65_cpsw_port *port); +static void am65_cpsw_destroy_xdp_rxqs(struct am65_cpsw_common *common) +{ + struct am65_cpsw_rx_chn *rx_chn = &common->rx_chns; + struct xdp_rxq_info *rxq; + int i; + + for (i = 0; i < common->port_num; i++) { + rxq = &common->ports[i].xdp_rxq; + + if (xdp_rxq_info_is_reg(rxq)) + xdp_rxq_info_unreg(rxq); + } + + if (rx_chn->page_pool) { + page_pool_destroy(rx_chn->page_pool); + rx_chn->page_pool = NULL; + } +} + +static int am65_cpsw_create_xdp_rxqs(struct am65_cpsw_common *common) +{ + struct am65_cpsw_rx_chn *rx_chn = &common->rx_chns; + struct page_pool_params pp_params = { + .flags = PP_FLAG_DMA_MAP, + .order = 0, + .pool_size = AM65_CPSW_MAX_RX_DESC, + .nid = dev_to_node(common->dev), + .dev = common->dev, + .dma_dir = DMA_BIDIRECTIONAL, + .napi = &common->napi_rx, + }; + struct xdp_rxq_info *rxq; + struct page_pool *pool; + int i, ret; + + pool = page_pool_create(&pp_params); + if (IS_ERR(pool)) + return PTR_ERR(pool); + + rx_chn->page_pool = pool; + + for (i = 0; i < common->port_num; i++) { + rxq = &common->ports[i].xdp_rxq; + + ret = xdp_rxq_info_reg(rxq, common->ports[i].ndev, i, 0); + if (ret) + goto err; + + ret = xdp_rxq_info_reg_mem_model(rxq, MEM_TYPE_PAGE_POOL, pool); + if (ret) + goto err; + } + + return 0; + +err: + am65_cpsw_destroy_xdp_rxqs(common); + return ret; +} + static void am65_cpsw_nuss_rx_cleanup(void *data, dma_addr_t desc_dma) { struct am65_cpsw_rx_chn *rx_chn = data; @@ -440,9 +514,41 @@ static void am65_cpsw_nuss_tx_cleanup(void *data, dma_addr_t desc_dma) dev_kfree_skb_any(skb); } +static struct sk_buff *am65_cpsw_alloc_skb(struct am65_cpsw_rx_chn *rx_chn, + struct net_device *ndev, + unsigned int len, + int desc_idx, + bool allow_direct) +{ + struct sk_buff *skb; + struct page *page; + + page = page_pool_dev_alloc_pages(rx_chn->page_pool); + if (unlikely(!page)) + return NULL; + + len += AM65_CPSW_HEADROOM; + + skb = build_skb(page_address(page), len); + if (unlikely(!skb)) { + page_pool_put_full_page(rx_chn->page_pool, page, allow_direct); + rx_chn->pages[desc_idx] = NULL; + return NULL; + } + + skb_reserve(skb, AM65_CPSW_HEADROOM + NET_IP_ALIGN); + skb->dev = ndev; + + rx_chn->pages[desc_idx] = page; + + return skb; +} + static int am65_cpsw_nuss_common_open(struct am65_cpsw_common *common) { struct am65_cpsw_host *host_p = am65_common_get_host(common); + struct am65_cpsw_rx_chn *rx_chn = &common->rx_chns; + struct am65_cpsw_tx_chn *tx_chn = common->tx_chns; int port_idx, i, ret, tx; struct sk_buff *skb; u32 val, port_mask; @@ -505,10 +611,15 @@ static int am65_cpsw_nuss_common_open(struct am65_cpsw_common *common) am65_cpsw_qos_tx_p0_rate_init(common); - for (i = 0; i < common->rx_chns.descs_num; i++) { - skb = __netdev_alloc_skb_ip_align(NULL, - AM65_CPSW_MAX_PACKET_SIZE, - GFP_KERNEL); + ret = am65_cpsw_create_xdp_rxqs(common); + if (ret) { + dev_err(common->dev, "Failed to create XDP rx queues\n"); + return ret; + } + + for (i = 0; i < rx_chn->descs_num; i++) { + skb = am65_cpsw_alloc_skb(rx_chn, NULL, + AM65_CPSW_MAX_PACKET_SIZE, i, false); if (!skb) { ret = -ENOMEM; dev_err(common->dev, "cannot allocate skb\n"); @@ -531,27 +642,27 @@ static int am65_cpsw_nuss_common_open(struct am65_cpsw_common *common) } } - ret = k3_udma_glue_enable_rx_chn(common->rx_chns.rx_chn); + ret = k3_udma_glue_enable_rx_chn(rx_chn->rx_chn); if (ret) { dev_err(common->dev, "couldn't enable rx chn: %d\n", ret); goto fail_rx; } for (tx = 0; tx < common->tx_ch_num; tx++) { - ret = k3_udma_glue_enable_tx_chn(common->tx_chns[tx].tx_chn); + ret = k3_udma_glue_enable_tx_chn(tx_chn[tx].tx_chn); if (ret) { dev_err(common->dev, "couldn't enable tx chn %d: %d\n", tx, ret); tx--; goto fail_tx; } - napi_enable(&common->tx_chns[tx].napi_tx); + napi_enable(&tx_chn[tx].napi_tx); } napi_enable(&common->napi_rx); if (common->rx_irq_disabled) { common->rx_irq_disabled = false; - enable_irq(common->rx_chns.irq); + enable_irq(rx_chn->irq); } dev_dbg(common->dev, "cpsw_nuss started\n"); @@ -559,22 +670,23 @@ static int am65_cpsw_nuss_common_open(struct am65_cpsw_common *common) fail_tx: while (tx >= 0) { - napi_disable(&common->tx_chns[tx].napi_tx); - k3_udma_glue_disable_tx_chn(common->tx_chns[tx].tx_chn); + napi_disable(&tx_chn[tx].napi_tx); + k3_udma_glue_disable_tx_chn(tx_chn[tx].tx_chn); tx--; } - k3_udma_glue_disable_rx_chn(common->rx_chns.rx_chn); + k3_udma_glue_disable_rx_chn(rx_chn->rx_chn); fail_rx: - k3_udma_glue_reset_rx_chn(common->rx_chns.rx_chn, 0, - &common->rx_chns, + k3_udma_glue_reset_rx_chn(rx_chn->rx_chn, 0, rx_chn, am65_cpsw_nuss_rx_cleanup, 0); return ret; } static int am65_cpsw_nuss_common_stop(struct am65_cpsw_common *common) { + struct am65_cpsw_rx_chn *rx_chn = &common->rx_chns; + struct am65_cpsw_tx_chn *tx_chn = common->tx_chns; int i; if (common->usage_count != 1) @@ -590,26 +702,25 @@ static int am65_cpsw_nuss_common_stop(struct am65_cpsw_common *common) reinit_completion(&common->tdown_complete); for (i = 0; i < common->tx_ch_num; i++) - k3_udma_glue_tdown_tx_chn(common->tx_chns[i].tx_chn, false); + k3_udma_glue_tdown_tx_chn(tx_chn[i].tx_chn, false); i = wait_for_completion_timeout(&common->tdown_complete, msecs_to_jiffies(1000)); if (!i) dev_err(common->dev, "tx timeout\n"); for (i = 0; i < common->tx_ch_num; i++) { - napi_disable(&common->tx_chns[i].napi_tx); - hrtimer_cancel(&common->tx_chns[i].tx_hrtimer); + napi_disable(&tx_chn[i].napi_tx); + hrtimer_cancel(&tx_chn[i].tx_hrtimer); } for (i = 0; i < common->tx_ch_num; i++) { - k3_udma_glue_reset_tx_chn(common->tx_chns[i].tx_chn, - &common->tx_chns[i], + k3_udma_glue_reset_tx_chn(tx_chn[i].tx_chn, &tx_chn[i], am65_cpsw_nuss_tx_cleanup); - k3_udma_glue_disable_tx_chn(common->tx_chns[i].tx_chn); + k3_udma_glue_disable_tx_chn(tx_chn[i].tx_chn); } reinit_completion(&common->tdown_complete); - k3_udma_glue_tdown_rx_chn(common->rx_chns.rx_chn, true); + k3_udma_glue_tdown_rx_chn(rx_chn->rx_chn, true); if (common->pdata.quirks & AM64_CPSW_QUIRK_DMA_RX_TDOWN_IRQ) { i = wait_for_completion_timeout(&common->tdown_complete, msecs_to_jiffies(1000)); @@ -621,17 +732,25 @@ static int am65_cpsw_nuss_common_stop(struct am65_cpsw_common *common) hrtimer_cancel(&common->rx_hrtimer); for (i = 0; i < AM65_CPSW_MAX_RX_FLOWS; i++) - k3_udma_glue_reset_rx_chn(common->rx_chns.rx_chn, i, - &common->rx_chns, + k3_udma_glue_reset_rx_chn(rx_chn->rx_chn, i, rx_chn, am65_cpsw_nuss_rx_cleanup, !!i); - k3_udma_glue_disable_rx_chn(common->rx_chns.rx_chn); + k3_udma_glue_disable_rx_chn(rx_chn->rx_chn); cpsw_ale_stop(common->ale); writel(0, common->cpsw_base + AM65_CPSW_REG_CTL); writel(0, common->cpsw_base + AM65_CPSW_REG_STAT_PORT_EN); + for (i = 0; i < rx_chn->descs_num; i++) { + if (rx_chn->pages[i]) { + page_pool_put_full_page(rx_chn->page_pool, + rx_chn->pages[i], false); + rx_chn->pages[i] = NULL; + } + } + am65_cpsw_destroy_xdp_rxqs(common); + dev_dbg(common->dev, "cpsw_nuss stopped\n"); return 0; } @@ -749,6 +868,186 @@ static int am65_cpsw_nuss_ndo_slave_open(struct net_device *ndev) return ret; } +static int am65_cpsw_nuss_desc_idx(struct k3_cppi_desc_pool *desc_pool, + void *desc, + unsigned char dsize_log2) +{ + void *pool_addr = k3_cppi_desc_pool_cpuaddr(desc_pool); + + return (desc - pool_addr) >> dsize_log2; +} + +static void am65_cpsw_nuss_set_buf_type(struct am65_cpsw_tx_chn *tx_chn, + struct cppi5_host_desc_t *desc, + enum am65_cpsw_tx_buf_type buf_type) +{ + int desc_idx; + + desc_idx = am65_cpsw_nuss_desc_idx(tx_chn->desc_pool, desc, + tx_chn->dsize_log2); + k3_cppi_desc_pool_desc_info_set(tx_chn->desc_pool, desc_idx, + (void *)buf_type); +} + +static enum am65_cpsw_tx_buf_type am65_cpsw_nuss_buf_type(struct am65_cpsw_tx_chn *tx_chn, + dma_addr_t desc_dma) +{ + struct cppi5_host_desc_t *desc_tx; + int desc_idx; + + desc_tx = k3_cppi_desc_pool_dma2virt(tx_chn->desc_pool, desc_dma); + desc_idx = am65_cpsw_nuss_desc_idx(tx_chn->desc_pool, desc_tx, + tx_chn->dsize_log2); + + return (enum am65_cpsw_tx_buf_type)k3_cppi_desc_pool_desc_info(tx_chn->desc_pool, + desc_idx); +} + +static int am65_cpsw_xdp_tx_frame(struct net_device *ndev, + struct am65_cpsw_tx_chn *tx_chn, + struct xdp_frame *xdpf, + enum am65_cpsw_tx_buf_type buf_type) +{ + struct am65_cpsw_common *common = am65_ndev_to_common(ndev); + struct am65_cpsw_port *port = am65_ndev_to_port(ndev); + struct cppi5_host_desc_t *host_desc; + struct netdev_queue *netif_txq; + dma_addr_t dma_desc, dma_buf; + u32 pkt_len = xdpf->len; + void **swdata; + int ret; + + host_desc = k3_cppi_desc_pool_alloc(tx_chn->desc_pool); + if (unlikely(!host_desc)) { + ndev->stats.tx_dropped++; + return -ENOMEM; + } + + am65_cpsw_nuss_set_buf_type(tx_chn, host_desc, buf_type); + + dma_buf = dma_map_single(tx_chn->dma_dev, xdpf->data, + pkt_len, DMA_TO_DEVICE); + if (unlikely(dma_mapping_error(tx_chn->dma_dev, dma_buf))) { + ndev->stats.tx_dropped++; + ret = -ENOMEM; + goto pool_free; + } + + cppi5_hdesc_init(host_desc, CPPI5_INFO0_HDESC_EPIB_PRESENT, + AM65_CPSW_NAV_PS_DATA_SIZE); + cppi5_hdesc_set_pkttype(host_desc, AM65_CPSW_CPPI_TX_PKT_TYPE); + cppi5_hdesc_set_pktlen(host_desc, pkt_len); + cppi5_desc_set_pktids(&host_desc->hdr, 0, AM65_CPSW_CPPI_TX_FLOW_ID); + cppi5_desc_set_tags_ids(&host_desc->hdr, 0, port->port_id); + + k3_udma_glue_tx_dma_to_cppi5_addr(tx_chn->tx_chn, &dma_buf); + cppi5_hdesc_attach_buf(host_desc, dma_buf, pkt_len, dma_buf, pkt_len); + + swdata = cppi5_hdesc_get_swdata(host_desc); + *(swdata) = xdpf; + + /* Report BQL before sending the packet */ + netif_txq = netdev_get_tx_queue(ndev, tx_chn->id); + netdev_tx_sent_queue(netif_txq, pkt_len); + + dma_desc = k3_cppi_desc_pool_virt2dma(tx_chn->desc_pool, host_desc); + if (AM65_CPSW_IS_CPSW2G(common)) { + ret = k3_udma_glue_push_tx_chn(tx_chn->tx_chn, host_desc, + dma_desc); + } else { + spin_lock_bh(&tx_chn->lock); + ret = k3_udma_glue_push_tx_chn(tx_chn->tx_chn, host_desc, + dma_desc); + spin_unlock_bh(&tx_chn->lock); + } + if (ret) { + /* Inform BQL */ + netdev_tx_completed_queue(netif_txq, 1, pkt_len); + ndev->stats.tx_errors++; + goto dma_unmap; + } + + return 0; + +dma_unmap: + k3_udma_glue_tx_cppi5_to_dma_addr(tx_chn->tx_chn, &dma_buf); + dma_unmap_single(tx_chn->dma_dev, dma_buf, pkt_len, DMA_TO_DEVICE); +pool_free: + k3_cppi_desc_pool_free(tx_chn->desc_pool, host_desc); + return ret; +} + +static int am65_cpsw_run_xdp(struct am65_cpsw_common *common, + struct am65_cpsw_port *port, + struct xdp_buff *xdp, + int desc_idx, int cpu, int *len) +{ + struct am65_cpsw_rx_chn *rx_chn = &common->rx_chns; + struct net_device *ndev = port->ndev; + int ret = AM65_CPSW_XDP_CONSUMED; + struct am65_cpsw_tx_chn *tx_chn; + struct netdev_queue *netif_txq; + struct xdp_frame *xdpf; + struct bpf_prog *prog; + struct page *page; + u32 act; + + prog = READ_ONCE(port->xdp_prog); + if (!prog) + return AM65_CPSW_XDP_PASS; + + act = bpf_prog_run_xdp(prog, xdp); + /* XDP prog might have changed packet data and boundaries */ + *len = xdp->data_end - xdp->data; + + switch (act) { + case XDP_PASS: + ret = AM65_CPSW_XDP_PASS; + goto out; + case XDP_TX: + tx_chn = &common->tx_chns[cpu % AM65_CPSW_MAX_TX_QUEUES]; + netif_txq = netdev_get_tx_queue(ndev, tx_chn->id); + + xdpf = xdp_convert_buff_to_frame(xdp); + if (unlikely(!xdpf)) + break; + + __netif_tx_lock(netif_txq, cpu); + ret = am65_cpsw_xdp_tx_frame(ndev, tx_chn, xdpf, + AM65_CPSW_TX_BUF_TYPE_XDP_TX); + __netif_tx_unlock(netif_txq); + if (ret) + break; + + ndev->stats.rx_bytes += *len; + ndev->stats.rx_packets++; + ret = AM65_CPSW_XDP_CONSUMED; + goto out; + case XDP_REDIRECT: + if (unlikely(xdp_do_redirect(ndev, xdp, prog))) + break; + + ndev->stats.rx_bytes += *len; + ndev->stats.rx_packets++; + ret = AM65_CPSW_XDP_REDIRECT; + goto out; + default: + bpf_warn_invalid_xdp_action(ndev, prog, act); + fallthrough; + case XDP_ABORTED: + trace_xdp_exception(ndev, prog, act); + fallthrough; + case XDP_DROP: + ndev->stats.rx_dropped++; + } + + page = virt_to_head_page(xdp->data); + page_pool_recycle_direct(rx_chn->page_pool, page); + rx_chn->pages[desc_idx] = NULL; +out: + return ret; +} + static void am65_cpsw_nuss_rx_ts(struct sk_buff *skb, u32 *psdata) { struct skb_shared_hwtstamps *ssh; @@ -795,7 +1094,7 @@ static void am65_cpsw_nuss_rx_csum(struct sk_buff *skb, u32 csum_info) } static int am65_cpsw_nuss_rx_packets(struct am65_cpsw_common *common, - u32 flow_idx) + u32 flow_idx, int cpu) { struct am65_cpsw_rx_chn *rx_chn = &common->rx_chns; u32 buf_dma_len, pkt_len, port_id = 0, csum_info; @@ -806,10 +1105,12 @@ static int am65_cpsw_nuss_rx_packets(struct am65_cpsw_common *common, struct sk_buff *skb, *new_skb; dma_addr_t desc_dma, buf_dma; struct am65_cpsw_port *port; + int headroom, desc_idx, ret; struct net_device *ndev; + struct xdp_buff xdp; + struct page *page; void **swdata; u32 *psdata; - int ret = 0; ret = k3_udma_glue_pop_rx_chn(rx_chn->rx_chn, flow_idx, &desc_dma); if (ret) { @@ -851,11 +1152,34 @@ static int am65_cpsw_nuss_rx_packets(struct am65_cpsw_common *common, k3_cppi_desc_pool_free(rx_chn->desc_pool, desc_rx); - new_skb = netdev_alloc_skb_ip_align(ndev, AM65_CPSW_MAX_PACKET_SIZE); + desc_idx = am65_cpsw_nuss_desc_idx(rx_chn->desc_pool, desc_rx, + rx_chn->dsize_log2); + + if (port->xdp_prog) { + xdp_init_buff(&xdp, AM65_CPSW_MAX_PACKET_SIZE, &port->xdp_rxq); + + page = virt_to_page(skb->data); + xdp_prepare_buff(&xdp, page_address(page), skb_headroom(skb), + pkt_len, false); + + ret = am65_cpsw_run_xdp(common, port, &xdp, desc_idx, + cpu, &pkt_len); + if (ret != AM65_CPSW_XDP_PASS) + return ret; + + /* Compute additional headroom to be reserved */ + headroom = (xdp.data - xdp.data_hard_start) - skb_headroom(skb); + skb_reserve(skb, headroom); + } + + /* Pass skb to netstack if no XDP prog or returned XDP_PASS */ + new_skb = am65_cpsw_alloc_skb(rx_chn, ndev, AM65_CPSW_MAX_PACKET_SIZE, + desc_idx, true); if (new_skb) { ndev_priv = netdev_priv(ndev); am65_cpsw_nuss_set_offload_fwd_mark(skb, ndev_priv->offload_fwd_mark); skb_put(skb, pkt_len); + skb_mark_for_recycle(skb); skb->protocol = eth_type_trans(skb, ndev); am65_cpsw_nuss_rx_csum(skb, csum_info); napi_gro_receive(&common->napi_rx, skb); @@ -901,6 +1225,8 @@ static int am65_cpsw_nuss_rx_poll(struct napi_struct *napi_rx, int budget) { struct am65_cpsw_common *common = am65_cpsw_napi_to_common(napi_rx); int flow = AM65_CPSW_MAX_RX_FLOWS; + int cpu = smp_processor_id(); + bool xdp_redirect = false; int cur_budget, ret; int num_rx = 0; @@ -909,9 +1235,12 @@ static int am65_cpsw_nuss_rx_poll(struct napi_struct *napi_rx, int budget) cur_budget = budget - num_rx; while (cur_budget--) { - ret = am65_cpsw_nuss_rx_packets(common, flow); - if (ret) + ret = am65_cpsw_nuss_rx_packets(common, flow, cpu); + if (ret) { + if (ret == AM65_CPSW_XDP_REDIRECT) + xdp_redirect = true; break; + } num_rx++; } @@ -919,6 +1248,9 @@ static int am65_cpsw_nuss_rx_poll(struct napi_struct *napi_rx, int budget) break; } + if (xdp_redirect) + xdp_do_flush(); + dev_dbg(common->dev, "%s num_rx:%d %d\n", __func__, num_rx, budget); if (num_rx < budget && napi_complete_done(napi_rx, num_rx)) { @@ -938,8 +1270,8 @@ static int am65_cpsw_nuss_rx_poll(struct napi_struct *napi_rx, int budget) } static struct sk_buff * -am65_cpsw_nuss_tx_compl_packet(struct am65_cpsw_tx_chn *tx_chn, - dma_addr_t desc_dma) +am65_cpsw_nuss_tx_compl_packet_skb(struct am65_cpsw_tx_chn *tx_chn, + dma_addr_t desc_dma) { struct am65_cpsw_ndev_priv *ndev_priv; struct am65_cpsw_ndev_stats *stats; @@ -968,6 +1300,39 @@ am65_cpsw_nuss_tx_compl_packet(struct am65_cpsw_tx_chn *tx_chn, return skb; } +static struct xdp_frame * +am65_cpsw_nuss_tx_compl_packet_xdp(struct am65_cpsw_common *common, + struct am65_cpsw_tx_chn *tx_chn, + dma_addr_t desc_dma, + struct net_device **ndev) +{ + struct am65_cpsw_ndev_priv *ndev_priv; + struct am65_cpsw_ndev_stats *stats; + struct cppi5_host_desc_t *desc_tx; + struct am65_cpsw_port *port; + struct xdp_frame *xdpf; + u32 port_id = 0; + void **swdata; + + desc_tx = k3_cppi_desc_pool_dma2virt(tx_chn->desc_pool, desc_dma); + cppi5_desc_get_tags_ids(&desc_tx->hdr, NULL, &port_id); + swdata = cppi5_hdesc_get_swdata(desc_tx); + xdpf = *(swdata); + am65_cpsw_nuss_xmit_free(tx_chn, desc_tx); + + port = am65_common_get_port(common, port_id); + *ndev = port->ndev; + + ndev_priv = netdev_priv(*ndev); + stats = this_cpu_ptr(ndev_priv->stats); + u64_stats_update_begin(&stats->syncp); + stats->tx_packets++; + stats->tx_bytes += xdpf->len; + u64_stats_update_end(&stats->syncp); + + return xdpf; +} + static void am65_cpsw_nuss_tx_wake(struct am65_cpsw_tx_chn *tx_chn, struct net_device *ndev, struct netdev_queue *netif_txq) { @@ -988,11 +1353,13 @@ static void am65_cpsw_nuss_tx_wake(struct am65_cpsw_tx_chn *tx_chn, struct net_d static int am65_cpsw_nuss_tx_compl_packets(struct am65_cpsw_common *common, int chn, unsigned int budget, bool *tdown) { + enum am65_cpsw_tx_buf_type buf_type; struct device *dev = common->dev; struct am65_cpsw_tx_chn *tx_chn; struct netdev_queue *netif_txq; unsigned int total_bytes = 0; struct net_device *ndev; + struct xdp_frame *xdpf; struct sk_buff *skb; dma_addr_t desc_dma; int res, num_tx = 0; @@ -1013,10 +1380,21 @@ static int am65_cpsw_nuss_tx_compl_packets(struct am65_cpsw_common *common, break; } - skb = am65_cpsw_nuss_tx_compl_packet(tx_chn, desc_dma); - total_bytes = skb->len; - ndev = skb->dev; - napi_consume_skb(skb, budget); + buf_type = am65_cpsw_nuss_buf_type(tx_chn, desc_dma); + if (buf_type == AM65_CPSW_TX_BUF_TYPE_SKB) { + skb = am65_cpsw_nuss_tx_compl_packet_skb(tx_chn, desc_dma); + ndev = skb->dev; + total_bytes = skb->len; + napi_consume_skb(skb, budget); + } else { + xdpf = am65_cpsw_nuss_tx_compl_packet_xdp(common, tx_chn, + desc_dma, &ndev); + total_bytes = xdpf->len; + if (buf_type == AM65_CPSW_TX_BUF_TYPE_XDP_TX) + xdp_return_frame_rx_napi(xdpf); + else + xdp_return_frame(xdpf); + } num_tx++; netif_txq = netdev_get_tx_queue(ndev, chn); @@ -1034,11 +1412,13 @@ static int am65_cpsw_nuss_tx_compl_packets(struct am65_cpsw_common *common, static int am65_cpsw_nuss_tx_compl_packets_2g(struct am65_cpsw_common *common, int chn, unsigned int budget, bool *tdown) { + enum am65_cpsw_tx_buf_type buf_type; struct device *dev = common->dev; struct am65_cpsw_tx_chn *tx_chn; struct netdev_queue *netif_txq; unsigned int total_bytes = 0; struct net_device *ndev; + struct xdp_frame *xdpf; struct sk_buff *skb; dma_addr_t desc_dma; int res, num_tx = 0; @@ -1057,11 +1437,21 @@ static int am65_cpsw_nuss_tx_compl_packets_2g(struct am65_cpsw_common *common, break; } - skb = am65_cpsw_nuss_tx_compl_packet(tx_chn, desc_dma); - - ndev = skb->dev; - total_bytes += skb->len; - napi_consume_skb(skb, budget); + buf_type = am65_cpsw_nuss_buf_type(tx_chn, desc_dma); + if (buf_type == AM65_CPSW_TX_BUF_TYPE_SKB) { + skb = am65_cpsw_nuss_tx_compl_packet_skb(tx_chn, desc_dma); + ndev = skb->dev; + total_bytes += skb->len; + napi_consume_skb(skb, budget); + } else { + xdpf = am65_cpsw_nuss_tx_compl_packet_xdp(common, tx_chn, + desc_dma, &ndev); + total_bytes += xdpf->len; + if (buf_type == AM65_CPSW_TX_BUF_TYPE_XDP_TX) + xdp_return_frame_rx_napi(xdpf); + else + xdp_return_frame(xdpf); + } num_tx++; } @@ -1183,10 +1573,13 @@ static netdev_tx_t am65_cpsw_nuss_ndo_slave_xmit(struct sk_buff *skb, goto busy_stop_q; } + am65_cpsw_nuss_set_buf_type(tx_chn, first_desc, + AM65_CPSW_TX_BUF_TYPE_SKB); + cppi5_hdesc_init(first_desc, CPPI5_INFO0_HDESC_EPIB_PRESENT, AM65_CPSW_NAV_PS_DATA_SIZE); - cppi5_desc_set_pktids(&first_desc->hdr, 0, 0x3FFF); - cppi5_hdesc_set_pkttype(first_desc, 0x7); + cppi5_desc_set_pktids(&first_desc->hdr, 0, AM65_CPSW_CPPI_TX_FLOW_ID); + cppi5_hdesc_set_pkttype(first_desc, AM65_CPSW_CPPI_TX_PKT_TYPE); cppi5_desc_set_tags_ids(&first_desc->hdr, 0, port->port_id); k3_udma_glue_tx_dma_to_cppi5_addr(tx_chn->tx_chn, &buf_dma); @@ -1225,6 +1618,9 @@ static netdev_tx_t am65_cpsw_nuss_ndo_slave_xmit(struct sk_buff *skb, goto busy_free_descs; } + am65_cpsw_nuss_set_buf_type(tx_chn, next_desc, + AM65_CPSW_TX_BUF_TYPE_SKB); + buf_dma = skb_frag_dma_map(tx_chn->dma_dev, frag, 0, frag_size, DMA_TO_DEVICE); if (unlikely(dma_mapping_error(tx_chn->dma_dev, buf_dma))) { @@ -1488,6 +1884,59 @@ static void am65_cpsw_nuss_ndo_get_stats(struct net_device *dev, stats->tx_dropped = dev->stats.tx_dropped; } +static int am65_cpsw_xdp_prog_setup(struct net_device *ndev, + struct bpf_prog *prog) +{ + struct am65_cpsw_port *port = am65_ndev_to_port(ndev); + bool running = netif_running(ndev); + struct bpf_prog *old_prog; + + if (running) + am65_cpsw_nuss_ndo_slave_stop(ndev); + + old_prog = xchg(&port->xdp_prog, prog); + if (old_prog) + bpf_prog_put(old_prog); + + if (running) + return am65_cpsw_nuss_ndo_slave_open(ndev); + + return 0; +} + +static int am65_cpsw_ndo_bpf(struct net_device *ndev, struct netdev_bpf *bpf) +{ + switch (bpf->command) { + case XDP_SETUP_PROG: + return am65_cpsw_xdp_prog_setup(ndev, bpf->prog); + default: + return -EINVAL; + } +} + +static int am65_cpsw_ndo_xdp_xmit(struct net_device *ndev, int n, + struct xdp_frame **frames, u32 flags) +{ + struct am65_cpsw_tx_chn *tx_chn; + struct netdev_queue *netif_txq; + int cpu = smp_processor_id(); + int i, nxmit = 0; + + tx_chn = &am65_ndev_to_common(ndev)->tx_chns[cpu % AM65_CPSW_MAX_TX_QUEUES]; + netif_txq = netdev_get_tx_queue(ndev, tx_chn->id); + + __netif_tx_lock(netif_txq, cpu); + for (i = 0; i < n; i++) { + if (am65_cpsw_xdp_tx_frame(ndev, tx_chn, frames[i], + AM65_CPSW_TX_BUF_TYPE_XDP_NDO)) + break; + nxmit++; + } + __netif_tx_unlock(netif_txq); + + return nxmit; +} + static const struct net_device_ops am65_cpsw_nuss_netdev_ops = { .ndo_open = am65_cpsw_nuss_ndo_slave_open, .ndo_stop = am65_cpsw_nuss_ndo_slave_stop, @@ -1502,6 +1951,8 @@ static const struct net_device_ops am65_cpsw_nuss_netdev_ops = { .ndo_eth_ioctl = am65_cpsw_nuss_ndo_slave_ioctl, .ndo_setup_tc = am65_cpsw_qos_ndo_setup_tc, .ndo_set_tx_maxrate = am65_cpsw_qos_ndo_tx_p0_set_maxrate, + .ndo_bpf = am65_cpsw_ndo_bpf, + .ndo_xdp_xmit = am65_cpsw_ndo_xdp_xmit, }; static void am65_cpsw_disable_phy(struct phy *phy) @@ -1772,7 +2223,7 @@ static int am65_cpsw_nuss_init_tx_chns(struct am65_cpsw_common *common) .mode = K3_RINGACC_RING_MODE_RING, .flags = 0 }; - u32 hdesc_size; + u32 hdesc_size, hdesc_size_out; int i, ret = 0; hdesc_size = cppi5_hdesc_calc_size(true, AM65_CPSW_NAV_PS_DATA_SIZE, @@ -1816,6 +2267,10 @@ static int am65_cpsw_nuss_init_tx_chns(struct am65_cpsw_common *common) goto err; } + hdesc_size_out = k3_cppi_desc_pool_desc_size(tx_chn->desc_pool); + tx_chn->dsize_log2 = __fls(hdesc_size_out); + WARN_ON(hdesc_size_out != (1 << tx_chn->dsize_log2)); + tx_chn->irq = k3_udma_glue_tx_get_irq(tx_chn->tx_chn); if (tx_chn->irq < 0) { dev_err(dev, "Failed to get tx dma irq %d\n", @@ -1862,8 +2317,8 @@ static void am65_cpsw_nuss_free_rx_chns(void *data) static void am65_cpsw_nuss_remove_rx_chns(void *data) { struct am65_cpsw_common *common = data; - struct am65_cpsw_rx_chn *rx_chn; struct device *dev = common->dev; + struct am65_cpsw_rx_chn *rx_chn; rx_chn = &common->rx_chns; devm_remove_action(dev, am65_cpsw_nuss_free_rx_chns, common); @@ -1873,11 +2328,7 @@ static void am65_cpsw_nuss_remove_rx_chns(void *data) netif_napi_del(&common->napi_rx); - if (!IS_ERR_OR_NULL(rx_chn->desc_pool)) - k3_cppi_desc_pool_destroy(rx_chn->desc_pool); - - if (!IS_ERR_OR_NULL(rx_chn->rx_chn)) - k3_udma_glue_release_rx_chn(rx_chn->rx_chn); + am65_cpsw_nuss_free_rx_chns(common); common->rx_flow_id_base = -1; } @@ -1888,7 +2339,7 @@ static int am65_cpsw_nuss_init_rx_chns(struct am65_cpsw_common *common) struct k3_udma_glue_rx_channel_cfg rx_cfg = { 0 }; u32 max_desc_num = AM65_CPSW_MAX_RX_DESC; struct device *dev = common->dev; - u32 hdesc_size; + u32 hdesc_size, hdesc_size_out; u32 fdqring_id; int i, ret = 0; @@ -1920,6 +2371,17 @@ static int am65_cpsw_nuss_init_rx_chns(struct am65_cpsw_common *common) goto err; } + hdesc_size_out = k3_cppi_desc_pool_desc_size(rx_chn->desc_pool); + rx_chn->dsize_log2 = __fls(hdesc_size_out); + WARN_ON(hdesc_size_out != (1 << rx_chn->dsize_log2)); + + rx_chn->page_pool = NULL; + + rx_chn->pages = devm_kcalloc(dev, rx_chn->descs_num, + sizeof(*rx_chn->pages), GFP_KERNEL); + if (!rx_chn->pages) + return -ENOMEM; + common->rx_flow_id_base = k3_udma_glue_rx_get_flow_id_base(rx_chn->rx_chn); dev_info(dev, "set new flow-id-base %u\n", common->rx_flow_id_base); @@ -2252,6 +2714,9 @@ am65_cpsw_nuss_init_port_ndev(struct am65_cpsw_common *common, u32 port_idx) NETIF_F_HW_TC; port->ndev->features = port->ndev->hw_features | NETIF_F_HW_VLAN_CTAG_FILTER; + port->ndev->xdp_features = NETDEV_XDP_ACT_BASIC | + NETDEV_XDP_ACT_REDIRECT | + NETDEV_XDP_ACT_NDO_XMIT; port->ndev->vlan_features |= NETIF_F_SG; port->ndev->netdev_ops = &am65_cpsw_nuss_netdev_ops; port->ndev->ethtool_ops = &am65_cpsw_ethtool_ops_slave; @@ -2315,6 +2780,8 @@ am65_cpsw_nuss_init_port_ndev(struct am65_cpsw_common *common, u32 port_idx) if (ret) dev_err(dev, "failed to add percpu stat free action %d\n", ret); + port->xdp_prog = NULL; + if (!common->dma_ndev) common->dma_ndev = port->ndev; @@ -2940,9 +3407,9 @@ static int am65_cpsw_nuss_probe(struct platform_device *pdev) struct device_node *node; struct resource *res; struct clk *clk; + int ale_entries; u64 id_temp; int ret, i; - int ale_entries; common = devm_kzalloc(dev, sizeof(struct am65_cpsw_common), GFP_KERNEL); if (!common) @@ -3154,10 +3621,10 @@ static int am65_cpsw_nuss_suspend(struct device *dev) static int am65_cpsw_nuss_resume(struct device *dev) { struct am65_cpsw_common *common = dev_get_drvdata(dev); + struct am65_cpsw_host *host_p = am65_common_get_host(common); struct am65_cpsw_port *port; struct net_device *ndev; int i, ret; - struct am65_cpsw_host *host_p = am65_common_get_host(common); ret = am65_cpsw_nuss_init_tx_chns(common); if (ret) diff --git a/drivers/net/ethernet/ti/am65-cpsw-nuss.h b/drivers/net/ethernet/ti/am65-cpsw-nuss.h index 7da0492dc091..d8ce88dc9c89 100644 --- a/drivers/net/ethernet/ti/am65-cpsw-nuss.h +++ b/drivers/net/ethernet/ti/am65-cpsw-nuss.h @@ -14,6 +14,7 @@ #include #include #include +#include #include "am65-cpsw-qos.h" struct am65_cpts; @@ -56,10 +57,18 @@ struct am65_cpsw_port { bool rx_ts_enabled; struct am65_cpsw_qos qos; struct devlink_port devlink_port; + struct bpf_prog *xdp_prog; + struct xdp_rxq_info xdp_rxq; /* Only for suspend resume context */ u32 vid_context; }; +enum am65_cpsw_tx_buf_type { + AM65_CPSW_TX_BUF_TYPE_SKB, + AM65_CPSW_TX_BUF_TYPE_XDP_TX, + AM65_CPSW_TX_BUF_TYPE_XDP_NDO, +}; + struct am65_cpsw_host { struct am65_cpsw_common *common; void __iomem *port_base; @@ -80,6 +89,7 @@ struct am65_cpsw_tx_chn { int irq; u32 id; u32 descs_num; + unsigned char dsize_log2; char tx_chn_name[128]; u32 rate_mbps; }; @@ -89,7 +99,10 @@ struct am65_cpsw_rx_chn { struct device *dma_dev; struct k3_cppi_desc_pool *desc_pool; struct k3_udma_glue_rx_channel *rx_chn; + struct page_pool *page_pool; + struct page **pages; u32 descs_num; + unsigned char dsize_log2; int irq; };