From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from us-smtp-delivery-124.mimecast.com (us-smtp-delivery-124.mimecast.com [170.10.129.124]) by mail.toke.dk (Postfix) with ESMTPS id 647E5A1C6EE for ; Thu, 27 Jul 2023 15:50:30 +0200 (CEST) Authentication-Results: mail.toke.dk; dkim=pass (1024-bit key; unprotected) header.d=redhat.com header.i=@redhat.com header.a=rsa-sha256 header.s=mimecast20190719 header.b=P8czg3dp DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=redhat.com; s=mimecast20190719; t=1690465828; h=from:from:reply-to:subject:subject:date:date:message-id:message-id: to:to:cc:cc:mime-version:mime-version:content-type:content-type: content-transfer-encoding:content-transfer-encoding: in-reply-to:in-reply-to:references:references; bh=1XiOfTE6mIe+no+efCzQfb1o2160qXLz3EBiZhKoPf8=; b=P8czg3dp6yrl1W3zXcIUK6neY18R83yW7igZkJQjTwvHN9lp0J3hAhx9NuCirWbmM86iFh IUmfpqBCRG/wK8bYb0RsObIwASAf3fdehNiCOkI8vQEvc8UYgHur5MU9pMqWD0pJU07ZDV jb2/za8ca0EtYNESQgTOFYVdvYrWe0o= Received: from mail-ed1-f71.google.com (mail-ed1-f71.google.com [209.85.208.71]) by relay.mimecast.com with ESMTP with STARTTLS (version=TLSv1.3, cipher=TLS_AES_256_GCM_SHA384) id us-mta-683-gSXmU96_PRyZagjRcQoWhQ-1; Thu, 27 Jul 2023 09:50:27 -0400 X-MC-Unique: gSXmU96_PRyZagjRcQoWhQ-1 Received: by mail-ed1-f71.google.com with SMTP id 4fb4d7f45d1cf-51a5296eb8eso626233a12.2 for ; Thu, 27 Jul 2023 06:50:27 -0700 (PDT) X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20221208; t=1690465826; x=1691070626; h=content-transfer-encoding:in-reply-to:references:to :content-language:subject:cc:user-agent:mime-version:date:message-id :from:x-gm-message-state:from:to:cc:subject:date:message-id:reply-to; bh=1XiOfTE6mIe+no+efCzQfb1o2160qXLz3EBiZhKoPf8=; b=HhS24iafhE8HodMhQ4Sr4xbfH4/noCVz25MFFbg0zjIJ460bzFhuh05SDo72ddN+26 AWw/PKDr+ElY6izKvcu/hvcbb6ljBZ4MpchtMv7uWMJHET9J8Ct5W6yFZ57ok/9tCNXi 47/Z15cv9lntngG7qb8De5AAu43lpQNNtIeEVClVWEPS2mDtApXsSjV5CZnJ3S4MZtON cF1kikx81SQWhfCJ0ZOllJltoFRRXrPcQ5/id2yOpmzR6+daKVO3AeeU1828tG53rbF/ xaUjM6gcIz/LiqGVE1q9hFTfdkwv5Ha+dpw+muHRe7LFzAVCwWDJPmcDofcseBov97F0 cxUQ== X-Gm-Message-State: ABy/qLaGDOK9d6DbMlrzWdJBaf4xFGlfxgFcG2higjL3wj6Uk2NySuq+ +J64UESESsIipqvq6ZFEwgXlVLsv+pXaNYVtq56hscVHmMnTLcRgPhDximCXZqVyKTFEMZJW42i wISvqpNwZy/p/6EUfLsiB X-Received: by 2002:a17:906:845c:b0:994:4f10:fb39 with SMTP id e28-20020a170906845c00b009944f10fb39mr1964632ejy.16.1690465826176; Thu, 27 Jul 2023 06:50:26 -0700 (PDT) X-Google-Smtp-Source: APBJJlHZM8PwoZjcMnC01i1MxwpCysmdHzYnZsReHy2hiY8stkm7Z/mE2+w3vCR/tyszu1A41BPdQg== X-Received: by 2002:a17:906:845c:b0:994:4f10:fb39 with SMTP id e28-20020a170906845c00b009944f10fb39mr1964610ejy.16.1690465825835; Thu, 27 Jul 2023 06:50:25 -0700 (PDT) Received: from [192.168.42.222] (194-45-78-10.static.kviknet.net. [194.45.78.10]) by smtp.gmail.com with ESMTPSA id gq15-20020a170906e24f00b00992b510089asm800108ejb.84.2023.07.27.06.50.24 (version=TLS1_3 cipher=TLS_AES_128_GCM_SHA256 bits=128/128); Thu, 27 Jul 2023 06:50:25 -0700 (PDT) From: Jesper Dangaard Brouer X-Google-Original-From: Jesper Dangaard Brouer Message-ID: Date: Thu, 27 Jul 2023 15:50:23 +0200 MIME-Version: 1.0 User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:102.0) Gecko/20100101 Thunderbird/102.13.0 To: Stanislav Fomichev , Jesper Dangaard Brouer References: <20230724235957.1953861-1-sdf@google.com> <20230724235957.1953861-3-sdf@google.com> <383cc1ce-3c87-4b80-9e70-e0c10a7c1dcc@redhat.com> In-Reply-To: X-Mimecast-Spam-Score: 0 X-Mimecast-Originator: redhat.com Content-Language: en-US Content-Type: text/plain; charset=UTF-8; format=flowed Content-Transfer-Encoding: 7bit Message-ID-Hash: PXU3E2B53GEPRM7LYFZWQPGRIV4WVXWI X-Message-ID-Hash: PXU3E2B53GEPRM7LYFZWQPGRIV4WVXWI X-MailFrom: jbrouer@redhat.com X-Mailman-Rule-Misses: dmarc-mitigation; no-senders; approved; emergency; loop; banned-address; member-moderation; nonmember-moderation; administrivia; implicit-dest; max-recipients; max-size; news-moderation; no-subject; digests; suspicious-header CC: brouer@redhat.com, bpf@vger.kernel.org, ast@kernel.org, daniel@iogearbox.net, andrii@kernel.org, martin.lau@linux.dev, song@kernel.org, yhs@fb.com, john.fastabend@gmail.com, kpsingh@kernel.org, haoluo@google.com, jolsa@kernel.org, kuba@kernel.org, toke@kernel.org, willemb@google.com, dsahern@kernel.org, magnus.karlsson@intel.com, bjorn@kernel.org, maciej.fijalkowski@intel.com, hawk@kernel.org, netdev@vger.kernel.org, xdp-hints@xdp-project.net X-Mailman-Version: 3.3.8 Precedence: list Subject: [xdp-hints] Re: [RFC net-next v4 2/8] xsk: add TX timestamp and TX checksum offload support List-Id: XDP hardware hints design discussion Archived-At: List-Archive: List-Help: List-Owner: List-Post: List-Subscribe: List-Unsubscribe: On 26/07/2023 23.25, Stanislav Fomichev wrote: > On 07/26, Jesper Dangaard Brouer wrote: >> >> >> On 25/07/2023 01.59, Stanislav Fomichev wrote: >>> diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h >>> index 11652e464f5d..8b40c80557aa 100644 >>> --- a/include/linux/netdevice.h >>> +++ b/include/linux/netdevice.h >>> @@ -1660,6 +1660,31 @@ struct xdp_metadata_ops { >>> enum xdp_rss_hash_type *rss_type); >>> }; >>> +/* >>> + * This structure defines the AF_XDP TX metadata hooks for network devices. >>> + * The following hooks can be defined; unless noted otherwise, they are >>> + * optional and can be filled with a null pointer. >>> + * >>> + * int (*tmo_request_timestamp)(void *priv) >>> + * This function is called when AF_XDP frame requested egress timestamp. >>> + * >>> + * int (*tmo_fill_timestamp)(void *priv) >>> + * This function is called when AF_XDP frame, that had requested >>> + * egress timestamp, received a completion. The hook needs to return >>> + * the actual HW timestamp. >>> + * >>> + * int (*tmo_request_timestamp)(u16 csum_start, u16 csum_offset, void *priv) >>> + * This function is called when AF_XDP frame requested HW checksum >>> + * offload. csum_start indicates position where checksumming should start. >>> + * csum_offset indicates position where checksum should be stored. >>> + * >>> + */ >>> +struct xsk_tx_metadata_ops { >>> + void (*tmo_request_timestamp)(void *priv); >>> + u64 (*tmo_fill_timestamp)(void *priv); >>> + void (*tmo_request_checksum)(u16 csum_start, u16 csum_offset, void *priv); >>> +}; >>> + >>> /** >>> * enum netdev_priv_flags - &struct net_device priv_flags >>> * >>> @@ -1844,6 +1869,7 @@ enum netdev_ml_priv_type { >>> * @netdev_ops: Includes several pointers to callbacks, >>> * if one wants to override the ndo_*() functions >>> * @xdp_metadata_ops: Includes pointers to XDP metadata callbacks. >>> + * @xsk_tx_metadata_ops: Includes pointers to AF_XDP TX metadata callbacks. >>> * @ethtool_ops: Management operations >>> * @l3mdev_ops: Layer 3 master device operations >>> * @ndisc_ops: Includes callbacks for different IPv6 neighbour >>> @@ -2100,6 +2126,7 @@ struct net_device { >>> unsigned long long priv_flags; >>> const struct net_device_ops *netdev_ops; >>> const struct xdp_metadata_ops *xdp_metadata_ops; >>> + const struct xsk_tx_metadata_ops *xsk_tx_metadata_ops; >>> int ifindex; >>> unsigned short gflags; >>> unsigned short hard_header_len; >>> diff --git a/include/linux/skbuff.h b/include/linux/skbuff.h >>> index faaba050f843..5febc1a5131e 100644 >>> --- a/include/linux/skbuff.h >>> +++ b/include/linux/skbuff.h >>> @@ -581,7 +581,10 @@ struct skb_shared_info { >>> /* Warning: this field is not always filled in (UFO)! */ >>> unsigned short gso_segs; >>> struct sk_buff *frag_list; >>> - struct skb_shared_hwtstamps hwtstamps; >>> + union { >>> + struct skb_shared_hwtstamps hwtstamps; >>> + struct xsk_tx_metadata *xsk_meta; >>> + }; >>> unsigned int gso_type; >>> u32 tskey; >>> diff --git a/include/net/xdp_sock.h b/include/net/xdp_sock.h >>> index 467b9fb56827..288fa58c4665 100644 >>> --- a/include/net/xdp_sock.h >>> +++ b/include/net/xdp_sock.h >>> @@ -90,6 +90,54 @@ int xsk_generic_rcv(struct xdp_sock *xs, struct xdp_buff *xdp); >>> int __xsk_map_redirect(struct xdp_sock *xs, struct xdp_buff *xdp); >>> void __xsk_map_flush(void); >>> +/** >>> + * xsk_tx_metadata_request - Evaluate AF_XDP TX metadata at submission >>> + * and call appropriate xsk_tx_metadata_ops operation. >>> + * @meta: pointer to AF_XDP metadata area >>> + * @ops: pointer to struct xsk_tx_metadata_ops >>> + * @priv: pointer to driver-private aread >>> + * >>> + * This function should be called by the networking device when >>> + * it prepares AF_XDP egress packet. >>> + */ >>> +static inline void xsk_tx_metadata_request(const struct xsk_tx_metadata *meta, >>> + const struct xsk_tx_metadata_ops *ops, >>> + void *priv) >> >> (As you mentioned) this gets inlined in drivers for performance. >> >>> +{ >>> + if (!meta) >>> + return; >>> + >>> + if (ops->tmo_request_timestamp) >>> + if (meta->flags & XDP_TX_METADATA_TIMESTAMP) >>> + ops->tmo_request_timestamp(priv); >> >> We still have the overhead of function pointer call. >> With RETPOLINE this is costly. >> >> Measured on my testlab CPU E5-1650 v4 @ 3.60GHz >> Type:function_call_cost: 3 cycles(tsc) 1.010 ns >> Type:func_ptr_call_cost: 30 cycles(tsc) 8.341 ns >> >> Given this get inlined in drivers, perhaps we can add some >> INDIRECT_CALL_1 macros to avoid these indirect calls? > > I'm assuming that the compiler is smart enough to resolve these const > struct ops definitions as long as they are in the same file. > > At least here is what I see for mlx5e_xmit_xdp_frame_mpwqe: those > indirect jumps are resolved and the calls themselves are unrolled. > TBF, I don't have retpolines enabled in the config, but I don't think > it will bring indirect jumps back? Am I missing anything? > I tried this with CONFIG_RETPOLINE and same results. The compiler is smart and inlines the call to mlx5e_xsk_request_checksum(). This is great, no need for crazy INDIRECT_CALL_1 macros :-) > > 0000000000001930 : > ; mlx5e_xmit_xdp_frame_mpwqe(): > ; ./drivers/net/ethernet/mellanox/mlx5/core/en/xdp.c:436 [...] > ; ./drivers/net/ethernet/mellanox/mlx5/core/en/xdp.c:381 > ; stats->mpwqe++; > 1b4a: 49 ff 44 24 08 incq 0x8(%r12) > ; ././include/net/xdp_sock.h:107 How do you get objdump to add these file:line annotations? I use: objdump -rS --visualize-jumps=color -Mintel | less -R > ; if (!meta) > 1b4f: 4d 85 ff testq %r15, %r15 > 1b52: 74 0e je 0x1b62 > ; ././include/net/xdp_sock.h:115 > ; if (meta->flags & XDP_TX_METADATA_CHECKSUM) > 1b54: 41 f6 07 02 testb $0x2, (%r15) > 1b58: 74 08 je 0x1b62 > ; ./drivers/net/ethernet/mellanox/mlx5/core/en/xdp.c:483 > ; xsk_tx_metadata_request(meta, &mlx5e_xsk_tx_metadata_ops, &session->wqe->eth); > 1b5a: 48 8b 43 50 movq 0x50(%rbx), %rax > ; ./drivers/net/ethernet/mellanox/mlx5/core/en/xdp.c:286 > ; eseg->cs_flags |= MLX5_ETH_WQE_L3_CSUM | MLX5_ETH_WQE_L4_CSUM; > 1b5e: 80 48 14 c0 orb $-0x40, 0x14(%rax) Yes, here mlx5e_xsk_request_checksum() gets inlined, as MLX5_ETH_WQE_L3_CSUM | MLX5_ETH_WQE_L4_CSUM = 192 192 = 0xC0 and signed byte value $-0x40, which we see in inst 'orb'. I usually prefer Intel ASM code output via objdump -Mintel and it decode ASM with unsigned value 0xc0 / 192: or BYTE PTR ds:0x4,0xc0 > ; ./drivers/net/ethernet/mellanox/mlx5/core/en/xdp.h:203 > ; (struct mlx5_wqe_data_seg *)session->wqe + session->ds_count; > 1b62: 48 8b 43 50 movq 0x50(%rbx), %rax [...] Thank you for checking up on this! :-) --Jesper