From patchwork Mon Jan 16 22:52:24 2023 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Bob Pearson X-Patchwork-Id: 13103814 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 vger.kernel.org (vger.kernel.org [23.128.96.18]) by smtp.lore.kernel.org (Postfix) with ESMTP id 51EF2C46467 for ; Mon, 16 Jan 2023 22:53:50 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S234743AbjAPWxs (ORCPT ); Mon, 16 Jan 2023 17:53:48 -0500 Received: from lindbergh.monkeyblade.net ([23.128.96.19]:38540 "EHLO lindbergh.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S235046AbjAPWx0 (ORCPT ); Mon, 16 Jan 2023 17:53:26 -0500 Received: from mail-oi1-x230.google.com (mail-oi1-x230.google.com [IPv6:2607:f8b0:4864:20::230]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id 0D45328868 for ; Mon, 16 Jan 2023 14:53:09 -0800 (PST) Received: by mail-oi1-x230.google.com with SMTP id i5so6321249oih.11 for ; Mon, 16 Jan 2023 14:53:08 -0800 (PST) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20210112; h=content-transfer-encoding:mime-version:references:in-reply-to :message-id:date:subject:cc:to:from:from:to:cc:subject:date :message-id:reply-to; bh=sFVv+FTu2viKrddF8zqbqpZJ1cote5/OKzz+8VqeQLQ=; b=CQiAquaWD/qixzw7ZE/zGrIC/taAPph0Iudxa6A7B3wGYRIv2lQAckPvl9DFUPPO4T Dhjt+2QI045ee2HpcItH9kaw9gQeabMbbi7u2d2tBuY+3vab7eqLrdUtgCCEHC0wXcQS dsJP6DNMfS1coRmUJvMePk/S3vfRyYT9Qd/xYcC9KFrM9BX2lfz8IkfIzJDqQsWtBPyJ 1aqwYBwyNcGx8lMG5FlaBBdiWspvwtVjIXdyfS6b5Oe1sBFD222sn1gc9XVEwzZxKBQl /+q7iZD2LC0Y2XLzaIeqGqdBJIuuLOCc+cy2DWXp3djBULiYt6pf5vhNKnf09umRqQ13 SmjQ== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20210112; h=content-transfer-encoding:mime-version:references:in-reply-to :message-id:date:subject:cc:to:from:x-gm-message-state:from:to:cc :subject:date:message-id:reply-to; bh=sFVv+FTu2viKrddF8zqbqpZJ1cote5/OKzz+8VqeQLQ=; b=ry83jDiZ4oAJp5B5a9O7F+E/xd9uMviAjKp7iK70AN2B2JiCdH94drf8cnRxVi62Av WrgHb+hPiPxeZeTC+WK/fkvSQ/WNPjnzKEUi1B9WLArsKyLqiKZYgL19hj/wjw80x7lY oGLOtoYYrjmxcboM4efCfbQfzagHpjRepf0VhnsF6XQQp+fGlWYbZVP5T78ClPbi5ioM itCMlQyDSpby/8g9h5hkzStIn6zkW7GbW6+w45P5WMZpj4tkDQyZnSZN3ylioyYVG98+ XDTKdDmsmTSUzWTXYlxIAceRq04Vl7K40v1BkI+8KeeE1RwgLNyY8C8LWmA25aZv8DSy FQ+w== X-Gm-Message-State: AFqh2koVKrkoo615rKWaSVNs32R50isNjzmbputdpPtN377TLQIPmfFl f/QxrP1kOjZ7DJjoi9G6Q/s= X-Google-Smtp-Source: AMrXdXvH2IloXJ3HWrabd6e2loj5wB1wmkfZVO3sZ0JYvDf78OuIMSXABTZMMawfia5ot9ayX8oNvQ== X-Received: by 2002:a54:4816:0:b0:361:25e8:c6d8 with SMTP id j22-20020a544816000000b0036125e8c6d8mr482424oij.34.1673909586923; Mon, 16 Jan 2023 14:53:06 -0800 (PST) Received: from rpearson-X570-AORUS-PRO-WIFI.tx.rr.com (2603-8081-140c-1a00-ea18-3ee9-26d1-7526.res6.spectrum.com. [2603:8081:140c:1a00:ea18:3ee9:26d1:7526]) by smtp.gmail.com with ESMTPSA id s4-20020a056808008400b0035028730c90sm13651937oic.1.2023.01.16.14.53.06 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Mon, 16 Jan 2023 14:53:06 -0800 (PST) From: Bob Pearson To: jgg@nvidia.com, zyjzyj2000@gmail.com, leonro@nvidia.com, yangx.jy@fujitsu.com, linux-rdma@vger.kernel.org Cc: Bob Pearson Subject: [PATCH for-next v4 1/6] RDMA/rxe: Cleanup mr_check_range Date: Mon, 16 Jan 2023 16:52:24 -0600 Message-Id: <20230116225227.21163-2-rpearsonhpe@gmail.com> X-Mailer: git-send-email 2.37.2 In-Reply-To: <20230116225227.21163-1-rpearsonhpe@gmail.com> References: <20230116225227.21163-1-rpearsonhpe@gmail.com> MIME-Version: 1.0 Precedence: bulk List-ID: X-Mailing-List: linux-rdma@vger.kernel.org Remove blank lines and replace EFAULT by EINVAL when an invalid mr type is used. Signed-off-by: Bob Pearson --- drivers/infiniband/sw/rxe/rxe_mr.c | 4 +--- 1 file changed, 1 insertion(+), 3 deletions(-) diff --git a/drivers/infiniband/sw/rxe/rxe_mr.c b/drivers/infiniband/sw/rxe/rxe_mr.c index 072eac4b65d2..632ee1e516a1 100644 --- a/drivers/infiniband/sw/rxe/rxe_mr.c +++ b/drivers/infiniband/sw/rxe/rxe_mr.c @@ -26,8 +26,6 @@ u8 rxe_get_next_key(u32 last_key) int mr_check_range(struct rxe_mr *mr, u64 iova, size_t length) { - - switch (mr->ibmr.type) { case IB_MR_TYPE_DMA: return 0; @@ -41,7 +39,7 @@ int mr_check_range(struct rxe_mr *mr, u64 iova, size_t length) default: rxe_dbg_mr(mr, "type (%d) not supported\n", mr->ibmr.type); - return -EFAULT; + return -EINVAL; } } From patchwork Mon Jan 16 22:52:25 2023 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Bob Pearson X-Patchwork-Id: 13103815 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 vger.kernel.org (vger.kernel.org [23.128.96.18]) by smtp.lore.kernel.org (Postfix) with ESMTP id DCBABC54EBE for ; Mon, 16 Jan 2023 22:53:52 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S233504AbjAPWxu (ORCPT ); Mon, 16 Jan 2023 17:53:50 -0500 Received: from lindbergh.monkeyblade.net ([23.128.96.19]:38764 "EHLO lindbergh.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S235205AbjAPWx3 (ORCPT ); Mon, 16 Jan 2023 17:53:29 -0500 Received: from mail-oi1-x236.google.com (mail-oi1-x236.google.com [IPv6:2607:f8b0:4864:20::236]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id 185A927D57 for ; Mon, 16 Jan 2023 14:53:16 -0800 (PST) Received: by mail-oi1-x236.google.com with SMTP id v17so3235931oie.5 for ; Mon, 16 Jan 2023 14:53:16 -0800 (PST) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20210112; h=content-transfer-encoding:mime-version:references:in-reply-to :message-id:date:subject:cc:to:from:from:to:cc:subject:date :message-id:reply-to; bh=MSc+ZeVhmJVAyrOad5GEnPdKs11Mmr6EK22kGWHehxY=; b=Q4ZwSmDckDbKQ5RNQAamkNYNd9PFNmJECHhVMEtUkbXjd0qDU72+CFNW6Yc2RbGRBg vPaur+tRIG5bG8/oTYu2S2JnQ4PepTvJW6G2uhjkDpi1TvE5yi21P72kVaYzHEpUXfzc Wujx/rMWtukzh9qHnPgBS8T7HJbd0yZQCGTQ6GWttB3GQRmgewvXxLYIp+HTplfvkQ8u Eu+cbbI09bYWZ9tDjilAzJ4Old6pACftHHzNHZho9Bpdv/t5Ma2W9eadL9CSipKkVCci X+lah3C9p9lK38zgjWddZIo+Wb6XcIGZy/wJ3aqORoVL4u2J39n2mwfPdMuuhpqalLOU IU9Q== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20210112; h=content-transfer-encoding:mime-version:references:in-reply-to :message-id:date:subject:cc:to:from:x-gm-message-state:from:to:cc :subject:date:message-id:reply-to; bh=MSc+ZeVhmJVAyrOad5GEnPdKs11Mmr6EK22kGWHehxY=; b=tStLttW/sfHR2pRzLP41O2kUmpwqCwooC1I7p1o5Cn7oT6tQfX1mudk2H6RFn6RWW5 hrzrHQ7O2bv+vxfb3NWH6Pl7nmEoC3CwknmuloYCJlApP6nn9TJs6MFpuwNBXKU/ysxi bIvnB9fZ9NE1cSspJ+IM3MLDGy85FsYoNlxMqSqyC60pZvD2tT8/eywfQKZLSbPW+Npc CqmR5a3C7FxefZOZMmUKeHJpw1rLfOar7wQw3hXu9XnU9toFLy81nwDDvO3ZanFFt6Ag 88IEvZWqLhiqb7AhZ/elFS4ydiAz39Lri4oCobSEkTwAzcORryq7sx33pK+o271b40sS zrmg== X-Gm-Message-State: AFqh2kq5gkOI1TQFhWKht8rirlxiYLMlszsH6lP4d/RDCY03NviNgI/5 vx1KG37MPzOk5U7qEJh0S1lfjSqvIUJ0dg== X-Google-Smtp-Source: AMrXdXsea6zR/RQJWYfUeINxo7TpNO3KSsJQkQUWmLXjh97KJO6E9ga4Qy8A20FqqlJtIZ9agWUibg== X-Received: by 2002:a54:4816:0:b0:361:25e8:c6d8 with SMTP id j22-20020a544816000000b0036125e8c6d8mr482544oij.34.1673909595494; Mon, 16 Jan 2023 14:53:15 -0800 (PST) Received: from rpearson-X570-AORUS-PRO-WIFI.tx.rr.com (2603-8081-140c-1a00-ea18-3ee9-26d1-7526.res6.spectrum.com. [2603:8081:140c:1a00:ea18:3ee9:26d1:7526]) by smtp.gmail.com with ESMTPSA id s4-20020a056808008400b0035028730c90sm13651937oic.1.2023.01.16.14.53.14 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Mon, 16 Jan 2023 14:53:15 -0800 (PST) From: Bob Pearson To: jgg@nvidia.com, zyjzyj2000@gmail.com, leonro@nvidia.com, yangx.jy@fujitsu.com, linux-rdma@vger.kernel.org Cc: Bob Pearson Subject: [PATCH for-next v4 2/6] RDMA/rxe: Move rxe_map_mr_sg to rxe_mr.c Date: Mon, 16 Jan 2023 16:52:25 -0600 Message-Id: <20230116225227.21163-3-rpearsonhpe@gmail.com> X-Mailer: git-send-email 2.37.2 In-Reply-To: <20230116225227.21163-1-rpearsonhpe@gmail.com> References: <20230116225227.21163-1-rpearsonhpe@gmail.com> MIME-Version: 1.0 Precedence: bulk List-ID: X-Mailing-List: linux-rdma@vger.kernel.org Move rxe_map_mr_sg() to rxe_mr.c where it makes a little more sense. Signed-off-by: Bob Pearson --- drivers/infiniband/sw/rxe/rxe_loc.h | 2 ++ drivers/infiniband/sw/rxe/rxe_mr.c | 36 +++++++++++++++++++++++++++ drivers/infiniband/sw/rxe/rxe_verbs.c | 36 --------------------------- 3 files changed, 38 insertions(+), 36 deletions(-) diff --git a/drivers/infiniband/sw/rxe/rxe_loc.h b/drivers/infiniband/sw/rxe/rxe_loc.h index 948ce4902b10..29b6c2143045 100644 --- a/drivers/infiniband/sw/rxe/rxe_loc.h +++ b/drivers/infiniband/sw/rxe/rxe_loc.h @@ -69,6 +69,8 @@ int rxe_mr_copy(struct rxe_mr *mr, u64 iova, void *addr, int length, enum rxe_mr_copy_dir dir); int copy_data(struct rxe_pd *pd, int access, struct rxe_dma_info *dma, void *addr, int length, enum rxe_mr_copy_dir dir); +int rxe_map_mr_sg(struct ib_mr *ibmr, struct scatterlist *sg, + int sg_nents, unsigned int *sg_offset); void *iova_to_vaddr(struct rxe_mr *mr, u64 iova, int length); struct rxe_mr *lookup_mr(struct rxe_pd *pd, int access, u32 key, enum rxe_mr_lookup_type type); diff --git a/drivers/infiniband/sw/rxe/rxe_mr.c b/drivers/infiniband/sw/rxe/rxe_mr.c index 632ee1e516a1..229c7259644c 100644 --- a/drivers/infiniband/sw/rxe/rxe_mr.c +++ b/drivers/infiniband/sw/rxe/rxe_mr.c @@ -223,6 +223,42 @@ int rxe_mr_init_fast(int max_pages, struct rxe_mr *mr) return err; } +static int rxe_set_page(struct ib_mr *ibmr, u64 addr) +{ + struct rxe_mr *mr = to_rmr(ibmr); + struct rxe_map *map; + struct rxe_phys_buf *buf; + + if (unlikely(mr->nbuf == mr->num_buf)) + return -ENOMEM; + + map = mr->map[mr->nbuf / RXE_BUF_PER_MAP]; + buf = &map->buf[mr->nbuf % RXE_BUF_PER_MAP]; + + buf->addr = addr; + buf->size = ibmr->page_size; + mr->nbuf++; + + return 0; +} + +int rxe_map_mr_sg(struct ib_mr *ibmr, struct scatterlist *sg, + int sg_nents, unsigned int *sg_offset) +{ + struct rxe_mr *mr = to_rmr(ibmr); + int n; + + mr->nbuf = 0; + + n = ib_sg_to_pages(ibmr, sg, sg_nents, sg_offset, rxe_set_page); + + mr->page_shift = ilog2(ibmr->page_size); + mr->page_mask = ibmr->page_size - 1; + mr->offset = ibmr->iova & mr->page_mask; + + return n; +} + static void lookup_iova(struct rxe_mr *mr, u64 iova, int *m_out, int *n_out, size_t *offset_out) { diff --git a/drivers/infiniband/sw/rxe/rxe_verbs.c b/drivers/infiniband/sw/rxe/rxe_verbs.c index 025b35bf014e..7a902e0a0607 100644 --- a/drivers/infiniband/sw/rxe/rxe_verbs.c +++ b/drivers/infiniband/sw/rxe/rxe_verbs.c @@ -948,42 +948,6 @@ static struct ib_mr *rxe_alloc_mr(struct ib_pd *ibpd, enum ib_mr_type mr_type, return ERR_PTR(err); } -static int rxe_set_page(struct ib_mr *ibmr, u64 addr) -{ - struct rxe_mr *mr = to_rmr(ibmr); - struct rxe_map *map; - struct rxe_phys_buf *buf; - - if (unlikely(mr->nbuf == mr->num_buf)) - return -ENOMEM; - - map = mr->map[mr->nbuf / RXE_BUF_PER_MAP]; - buf = &map->buf[mr->nbuf % RXE_BUF_PER_MAP]; - - buf->addr = addr; - buf->size = ibmr->page_size; - mr->nbuf++; - - return 0; -} - -static int rxe_map_mr_sg(struct ib_mr *ibmr, struct scatterlist *sg, - int sg_nents, unsigned int *sg_offset) -{ - struct rxe_mr *mr = to_rmr(ibmr); - int n; - - mr->nbuf = 0; - - n = ib_sg_to_pages(ibmr, sg, sg_nents, sg_offset, rxe_set_page); - - mr->page_shift = ilog2(ibmr->page_size); - mr->page_mask = ibmr->page_size - 1; - mr->offset = ibmr->iova & mr->page_mask; - - return n; -} - static ssize_t parent_show(struct device *device, struct device_attribute *attr, char *buf) { From patchwork Mon Jan 16 22:52:26 2023 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Bob Pearson X-Patchwork-Id: 13103816 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 vger.kernel.org (vger.kernel.org [23.128.96.18]) by smtp.lore.kernel.org (Postfix) with ESMTP id EB650C46467 for ; Mon, 16 Jan 2023 22:53:54 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S232058AbjAPWxv (ORCPT ); Mon, 16 Jan 2023 17:53:51 -0500 Received: from lindbergh.monkeyblade.net ([23.128.96.19]:38568 "EHLO lindbergh.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S235208AbjAPWxa (ORCPT ); Mon, 16 Jan 2023 17:53:30 -0500 Received: from mail-oi1-x232.google.com (mail-oi1-x232.google.com [IPv6:2607:f8b0:4864:20::232]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id 0A1BF298C0 for ; Mon, 16 Jan 2023 14:53:17 -0800 (PST) Received: by mail-oi1-x232.google.com with SMTP id s66so23911034oib.7 for ; Mon, 16 Jan 2023 14:53:17 -0800 (PST) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20210112; h=content-transfer-encoding:mime-version:references:in-reply-to :message-id:date:subject:cc:to:from:from:to:cc:subject:date :message-id:reply-to; bh=6mvGPH/5e0Acrw0ibUA1HCYdUhROlXsosKvd0Ei2KI0=; b=L9FRW7aTw8w39+rpDxLRr8awWvuOl6U5edc5wclaPInYsIuXloVguVyWvIkDgyI7Dr fkR3gGdgN2N2kpaFAT+sjQE4ZkZ4PkvjfZN5MJ7BjjPzI++U2ovS2hGoPCRjofyD8XPO ZQBIEhevkOJ0fG5WOkvg8wT3JdyCTJ6dQAKEhZv0hrjMSH7+dcltY8mkBjNeZQRB1EGY +bqEq/NnKTZLNxkpitNkZGtuRA9Otp4ry1hvFkqP5UBGNoIrXxDc7XGGZXRzlDxHC7qc v6WvMnXP721OZkxJ1xtCyxroH6fmfWAbu3GMt/CNJklWQz+H5CezwwHtd3RWP3OMY6Fb yN9A== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20210112; h=content-transfer-encoding:mime-version:references:in-reply-to :message-id:date:subject:cc:to:from:x-gm-message-state:from:to:cc :subject:date:message-id:reply-to; bh=6mvGPH/5e0Acrw0ibUA1HCYdUhROlXsosKvd0Ei2KI0=; b=RgIjATql6hVbUZZunI+zviQpTB38Rh4cgWtWDKakc6h9Ku3mo0CO+hHCUoDlmGTr88 GykxU/Sl+HMGzeae7HN27MJnhrqIv7mj/sVrt8rQA/n5WpJYs6eytBVv9Q6YeWjNhBNV TU18eYC4olCuFTomg/RSm68sHQQxkSc/1iXO6MocyipRD4WvrcJRANAGfOBtfiZYYk76 b8b8T3kp1CYF3+ss1KK/WXjAa3ZS/LYS/rseL1/oDppRpjP4NEkfXelk2OOPDmKm1591 UQ7lfQlxTxMXfiEwTdkf6vznQcnA0E7OdQxx/GPhtdm1I/Ct+LM7UGy5K98UaZy9e93Q 9Q1A== X-Gm-Message-State: AFqh2kp5MtnA1Xh+mWBxBrMUhtGkl3SsLuabrrKq+5vnDLrZG/TW8Fj1 W4KbHnsQJNIApDJEqA+5Y+8= X-Google-Smtp-Source: AMrXdXsJ4Mh1W+eGwTZfTrzAeJwf1tHSFGyOnLrXCrcc1Phat9wZPgna7YML5r05GnX5UNNcOZjDVA== X-Received: by 2002:a05:6808:228a:b0:364:4eb7:e873 with SMTP id bo10-20020a056808228a00b003644eb7e873mr479485oib.7.1673909596356; Mon, 16 Jan 2023 14:53:16 -0800 (PST) Received: from rpearson-X570-AORUS-PRO-WIFI.tx.rr.com (2603-8081-140c-1a00-ea18-3ee9-26d1-7526.res6.spectrum.com. [2603:8081:140c:1a00:ea18:3ee9:26d1:7526]) by smtp.gmail.com with ESMTPSA id s4-20020a056808008400b0035028730c90sm13651937oic.1.2023.01.16.14.53.15 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Mon, 16 Jan 2023 14:53:16 -0800 (PST) From: Bob Pearson To: jgg@nvidia.com, zyjzyj2000@gmail.com, leonro@nvidia.com, yangx.jy@fujitsu.com, linux-rdma@vger.kernel.org Cc: Bob Pearson Subject: [PATCH for-next v4 3/6] RDMA-rxe: Isolate mr code from atomic_reply() Date: Mon, 16 Jan 2023 16:52:26 -0600 Message-Id: <20230116225227.21163-4-rpearsonhpe@gmail.com> X-Mailer: git-send-email 2.37.2 In-Reply-To: <20230116225227.21163-1-rpearsonhpe@gmail.com> References: <20230116225227.21163-1-rpearsonhpe@gmail.com> MIME-Version: 1.0 Precedence: bulk List-ID: X-Mailing-List: linux-rdma@vger.kernel.org Isolate mr specific code from atomic_reply() in rxe_resp.c into a subroutine rxe_mr_do_atomic_op() in rxe_mr.c. Signed-off-by: Bob Pearson --- drivers/infiniband/sw/rxe/rxe_loc.h | 2 + drivers/infiniband/sw/rxe/rxe_mr.c | 83 ++++++++++++++++++--------- drivers/infiniband/sw/rxe/rxe_resp.c | 49 +++++----------- drivers/infiniband/sw/rxe/rxe_verbs.h | 5 ++ 4 files changed, 76 insertions(+), 63 deletions(-) diff --git a/drivers/infiniband/sw/rxe/rxe_loc.h b/drivers/infiniband/sw/rxe/rxe_loc.h index 29b6c2143045..bcb1bbcf50df 100644 --- a/drivers/infiniband/sw/rxe/rxe_loc.h +++ b/drivers/infiniband/sw/rxe/rxe_loc.h @@ -72,6 +72,8 @@ int copy_data(struct rxe_pd *pd, int access, struct rxe_dma_info *dma, int rxe_map_mr_sg(struct ib_mr *ibmr, struct scatterlist *sg, int sg_nents, unsigned int *sg_offset); void *iova_to_vaddr(struct rxe_mr *mr, u64 iova, int length); +int rxe_mr_do_atomic_op(struct rxe_mr *mr, u64 iova, int opcode, + u64 compare, u64 swap_add, u64 *orig_val); struct rxe_mr *lookup_mr(struct rxe_pd *pd, int access, u32 key, enum rxe_mr_lookup_type type); int mr_check_range(struct rxe_mr *mr, u64 iova, size_t length); diff --git a/drivers/infiniband/sw/rxe/rxe_mr.c b/drivers/infiniband/sw/rxe/rxe_mr.c index 229c7259644c..15a8d44daa35 100644 --- a/drivers/infiniband/sw/rxe/rxe_mr.c +++ b/drivers/infiniband/sw/rxe/rxe_mr.c @@ -32,13 +32,15 @@ int mr_check_range(struct rxe_mr *mr, u64 iova, size_t length) case IB_MR_TYPE_USER: case IB_MR_TYPE_MEM_REG: - if (iova < mr->ibmr.iova || length > mr->ibmr.length || - iova > mr->ibmr.iova + mr->ibmr.length - length) - return -EFAULT; + if (iova < mr->ibmr.iova || + iova + length > mr->ibmr.iova + mr->ibmr.length) { + rxe_dbg_mr(mr, "iova/length out of range"); + return -ERANGE; + } return 0; default: - rxe_dbg_mr(mr, "type (%d) not supported\n", mr->ibmr.type); + rxe_dbg_mr(mr, "mr type not supported\n"); return -EINVAL; } } @@ -299,37 +301,22 @@ void *iova_to_vaddr(struct rxe_mr *mr, u64 iova, int length) { size_t offset; int m, n; - void *addr; - if (mr->state != RXE_MR_STATE_VALID) { - rxe_dbg_mr(mr, "Not in valid state\n"); - addr = NULL; - goto out; - } + if (mr->state != RXE_MR_STATE_VALID) + return NULL; - if (!mr->map) { - addr = (void *)(uintptr_t)iova; - goto out; - } + if (mr->ibmr.type == IB_MR_TYPE_DMA) + return (void *)(uintptr_t)iova; - if (mr_check_range(mr, iova, length)) { - rxe_dbg_mr(mr, "Range violation\n"); - addr = NULL; - goto out; - } + if (mr_check_range(mr, iova, length)) + return NULL; lookup_iova(mr, iova, &m, &n, &offset); - if (offset + length > mr->map[m]->buf[n].size) { - rxe_dbg_mr(mr, "Crosses page boundary\n"); - addr = NULL; - goto out; - } - - addr = (void *)(uintptr_t)mr->map[m]->buf[n].addr + offset; + if (offset + length > mr->map[m]->buf[n].size) + return NULL; -out: - return addr; + return (void *)(uintptr_t)mr->map[m]->buf[n].addr + offset; } int rxe_flush_pmem_iova(struct rxe_mr *mr, u64 iova, int length) @@ -538,6 +525,46 @@ int copy_data( return err; } +/* Guarantee atomicity of atomic operations at the machine level. */ +static DEFINE_SPINLOCK(atomic_ops_lock); + +int rxe_mr_do_atomic_op(struct rxe_mr *mr, u64 iova, int opcode, + u64 compare, u64 swap_add, u64 *orig_val) +{ + u64 *va; + u64 value; + + if (mr->state != RXE_MR_STATE_VALID) { + rxe_dbg_mr(mr, "mr not in valid state"); + return -EINVAL; + } + + va = iova_to_vaddr(mr, iova, sizeof(u64)); + if (!va) { + rxe_dbg_mr(mr, "iova out of range"); + return -ERANGE; + } + + if ((uintptr_t)va & 0x7) { + rxe_dbg_mr(mr, "iova not aligned"); + return -RXE_ERR_NOT_ALIGNED; + } + + spin_lock_bh(&atomic_ops_lock); + value = *orig_val = *va; + + if (opcode == IB_OPCODE_RC_COMPARE_SWAP) { + if (value == compare) + *va = swap_add; + } else { + value += swap_add; + *va = value; + } + spin_unlock_bh(&atomic_ops_lock); + + return 0; +} + int advance_dma_data(struct rxe_dma_info *dma, unsigned int length) { struct rxe_sge *sge = &dma->sge[dma->cur_sge]; diff --git a/drivers/infiniband/sw/rxe/rxe_resp.c b/drivers/infiniband/sw/rxe/rxe_resp.c index c74972244f08..1e38e5da1f4c 100644 --- a/drivers/infiniband/sw/rxe/rxe_resp.c +++ b/drivers/infiniband/sw/rxe/rxe_resp.c @@ -725,17 +725,12 @@ static enum resp_states process_flush(struct rxe_qp *qp, return RESPST_ACKNOWLEDGE; } -/* Guarantee atomicity of atomic operations at the machine level. */ -static DEFINE_SPINLOCK(atomic_ops_lock); - static enum resp_states atomic_reply(struct rxe_qp *qp, - struct rxe_pkt_info *pkt) + struct rxe_pkt_info *pkt) { - u64 *vaddr; - enum resp_states ret; struct rxe_mr *mr = qp->resp.mr; struct resp_res *res = qp->resp.res; - u64 value; + int err; if (!res) { res = rxe_prepare_res(qp, pkt, RXE_ATOMIC_MASK); @@ -743,33 +738,19 @@ static enum resp_states atomic_reply(struct rxe_qp *qp, } if (!res->replay) { - if (mr->state != RXE_MR_STATE_VALID) { - ret = RESPST_ERR_RKEY_VIOLATION; - goto out; - } - - vaddr = iova_to_vaddr(mr, qp->resp.va + qp->resp.offset, - sizeof(u64)); - - /* check vaddr is 8 bytes aligned. */ - if (!vaddr || (uintptr_t)vaddr & 7) { - ret = RESPST_ERR_MISALIGNED_ATOMIC; - goto out; - } - - spin_lock_bh(&atomic_ops_lock); - res->atomic.orig_val = value = *vaddr; - - if (pkt->opcode == IB_OPCODE_RC_COMPARE_SWAP) { - if (value == atmeth_comp(pkt)) - value = atmeth_swap_add(pkt); - } else { - value += atmeth_swap_add(pkt); + u64 iova = qp->resp.va + qp->resp.offset; + + err = rxe_mr_do_atomic_op(mr, iova, pkt->opcode, + atmeth_comp(pkt), + atmeth_swap_add(pkt), + &res->atomic.orig_val); + if (unlikely(err)) { + if (err == -RXE_ERR_NOT_ALIGNED) + return RESPST_ERR_MISALIGNED_ATOMIC; + else + return RESPST_ERR_RKEY_VIOLATION; } - *vaddr = value; - spin_unlock_bh(&atomic_ops_lock); - qp->resp.msn++; /* next expected psn, read handles this separately */ @@ -780,9 +761,7 @@ static enum resp_states atomic_reply(struct rxe_qp *qp, qp->resp.status = IB_WC_SUCCESS; } - ret = RESPST_ACKNOWLEDGE; -out: - return ret; + return RESPST_ACKNOWLEDGE; } #ifdef CONFIG_64BIT diff --git a/drivers/infiniband/sw/rxe/rxe_verbs.h b/drivers/infiniband/sw/rxe/rxe_verbs.h index 19ddfa890480..691200d99d6b 100644 --- a/drivers/infiniband/sw/rxe/rxe_verbs.h +++ b/drivers/infiniband/sw/rxe/rxe_verbs.h @@ -273,6 +273,11 @@ enum rxe_mr_state { RXE_MR_STATE_VALID, }; +/* some rxe specific error conditions */ +enum rxe_mr_error { + RXE_ERR_NOT_ALIGNED = 0x1001, /* misaligned address */ +}; + enum rxe_mr_copy_dir { RXE_TO_MR_OBJ, RXE_FROM_MR_OBJ, From patchwork Mon Jan 16 22:52:27 2023 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Bob Pearson X-Patchwork-Id: 13103817 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 vger.kernel.org (vger.kernel.org [23.128.96.18]) by smtp.lore.kernel.org (Postfix) with ESMTP id 4BECDC54EBE for ; Mon, 16 Jan 2023 22:53:56 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S235115AbjAPWxx (ORCPT ); Mon, 16 Jan 2023 17:53:53 -0500 Received: from lindbergh.monkeyblade.net ([23.128.96.19]:38564 "EHLO lindbergh.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S235212AbjAPWxb (ORCPT ); Mon, 16 Jan 2023 17:53:31 -0500 Received: from mail-oi1-x22b.google.com (mail-oi1-x22b.google.com [IPv6:2607:f8b0:4864:20::22b]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id ECD99298C1 for ; Mon, 16 Jan 2023 14:53:17 -0800 (PST) Received: by mail-oi1-x22b.google.com with SMTP id s124so7749407oif.1 for ; Mon, 16 Jan 2023 14:53:17 -0800 (PST) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20210112; h=content-transfer-encoding:mime-version:references:in-reply-to :message-id:date:subject:cc:to:from:from:to:cc:subject:date :message-id:reply-to; bh=9q6oRI/jJlGEjhfTY3sZkztExTr0+SXp9sacBMOPdxI=; b=gotUCJVmpjcLdJxNTqKpMD2Uf+1ZavKyqXidGM5n4QcDfvcO30sq8bPwGIF//oWoA4 DIJZC7hk7vDRwFu5dOgeHOvZgzM804ciSD2Bw6uy2Nwwhu8Lq4thVMYE5qvJZvzY5nWz HIUm6PgLfYdbAomeXa6iA2BkLMpzmuzuWDg/jEusa5H/yfEr7QVmPBHsDv+90t2E9oH8 4+c7xVVuqkHQxWEHL5yexIKpEa0RHo4NNmwrsqKRNcCXrSIgcx4h8HXipMux7dy5MJaN hPzMTukc1iBZJC2QES4jerQoCvUai3Iyk366JMtNiNvQjX5YzPWYYl1mwBNsvRaQmVDc TKQg== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20210112; h=content-transfer-encoding:mime-version:references:in-reply-to :message-id:date:subject:cc:to:from:x-gm-message-state:from:to:cc :subject:date:message-id:reply-to; bh=9q6oRI/jJlGEjhfTY3sZkztExTr0+SXp9sacBMOPdxI=; b=sRi7goXQ5aWoXnunmQLoirXHTfJvhv8mgTZvvri+8TzEb4DkVXeM2QiBeoQ4FkHEto G2JdxiMC3IdmYF44CXMtiXUWedTVTmF8o+qCG3l1HF5ImJ5Wln5rshgiZevkt6XKHZ+U 2kNsedqxrKyXWXpdZWDHLx/tXBGRf1sB7sPYEFfj+MbZ17+UVcXdpG3seNsgLgPqlThP 7wCL1Yv5FwdgGU2/6kX1YTQDjcOuEsDN0ukB1ZzQj54Rbeyog5vdNWqAN2SwAowoWkzi 1L1pujoXHd9TbKcZy2JtZEa4MRc6xhiL0qgAeXyLDaTMc1OcWaIm3nnx7j2NshmaZwWd gK3w== X-Gm-Message-State: AFqh2koA7DRXTclbY8oR5VFc7r8a22/R0MJciGmaQvrCRP3iBvBi9dpf +0H5lEdzvrnf9signQgDj+2i62KBi96zWQ== X-Google-Smtp-Source: AMrXdXvHCwspTqY1BeiC6bNZ2AAXT7m1Vs3585y9/ArxN7QhjW8fj8qYJ60l8ORwJtlFchmacRcpPw== X-Received: by 2002:aca:5c09:0:b0:363:a9b1:54a6 with SMTP id q9-20020aca5c09000000b00363a9b154a6mr317886oib.40.1673909597309; Mon, 16 Jan 2023 14:53:17 -0800 (PST) Received: from rpearson-X570-AORUS-PRO-WIFI.tx.rr.com (2603-8081-140c-1a00-ea18-3ee9-26d1-7526.res6.spectrum.com. [2603:8081:140c:1a00:ea18:3ee9:26d1:7526]) by smtp.gmail.com with ESMTPSA id s4-20020a056808008400b0035028730c90sm13651937oic.1.2023.01.16.14.53.16 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Mon, 16 Jan 2023 14:53:16 -0800 (PST) From: Bob Pearson To: jgg@nvidia.com, zyjzyj2000@gmail.com, leonro@nvidia.com, yangx.jy@fujitsu.com, linux-rdma@vger.kernel.org Cc: Bob Pearson Subject: [PATCH for-next v4 4/6] RDMA-rxe: Isolate mr code from atomic_write_reply() Date: Mon, 16 Jan 2023 16:52:27 -0600 Message-Id: <20230116225227.21163-5-rpearsonhpe@gmail.com> X-Mailer: git-send-email 2.37.2 In-Reply-To: <20230116225227.21163-1-rpearsonhpe@gmail.com> References: <20230116225227.21163-1-rpearsonhpe@gmail.com> MIME-Version: 1.0 Precedence: bulk List-ID: X-Mailing-List: linux-rdma@vger.kernel.org Isolate mr specific code from atomic_write_reply() in rxe_resp.c into a subroutine rxe_mr_do_atomic_write() in rxe_mr.c. Check length for atomic write operation. Make iova_to_vaddr() static. Signed-off-by: Bob Pearson --- drivers/infiniband/sw/rxe/rxe_loc.h | 1 + drivers/infiniband/sw/rxe/rxe_mr.c | 34 ++++++++++++++ drivers/infiniband/sw/rxe/rxe_resp.c | 69 ++++++++++++---------------- 3 files changed, 64 insertions(+), 40 deletions(-) diff --git a/drivers/infiniband/sw/rxe/rxe_loc.h b/drivers/infiniband/sw/rxe/rxe_loc.h index bcb1bbcf50df..b1dda0cf891b 100644 --- a/drivers/infiniband/sw/rxe/rxe_loc.h +++ b/drivers/infiniband/sw/rxe/rxe_loc.h @@ -74,6 +74,7 @@ int rxe_map_mr_sg(struct ib_mr *ibmr, struct scatterlist *sg, void *iova_to_vaddr(struct rxe_mr *mr, u64 iova, int length); int rxe_mr_do_atomic_op(struct rxe_mr *mr, u64 iova, int opcode, u64 compare, u64 swap_add, u64 *orig_val); +int rxe_mr_do_atomic_write(struct rxe_mr *mr, u64 iova, u64 value); struct rxe_mr *lookup_mr(struct rxe_pd *pd, int access, u32 key, enum rxe_mr_lookup_type type); int mr_check_range(struct rxe_mr *mr, u64 iova, size_t length); diff --git a/drivers/infiniband/sw/rxe/rxe_mr.c b/drivers/infiniband/sw/rxe/rxe_mr.c index 15a8d44daa35..10484f671977 100644 --- a/drivers/infiniband/sw/rxe/rxe_mr.c +++ b/drivers/infiniband/sw/rxe/rxe_mr.c @@ -565,6 +565,40 @@ int rxe_mr_do_atomic_op(struct rxe_mr *mr, u64 iova, int opcode, return 0; } +/* only implemented for 64 bit architectures */ +int rxe_mr_do_atomic_write(struct rxe_mr *mr, u64 iova, u64 value) +{ +#if defined CONFIG_64BIT + u64 *va; + + /* See IBA oA19-28 */ + if (unlikely(mr->state != RXE_MR_STATE_VALID)) { + rxe_dbg_mr(mr, "mr not in valid state"); + return -EINVAL; + } + + va = iova_to_vaddr(mr, iova, sizeof(value)); + if (unlikely(!va)) { + rxe_dbg_mr(mr, "iova out of range"); + return -ERANGE; + } + + /* See IBA A19.4.2 */ + if (unlikely((uintptr_t)va & 0x7 || iova & 0x7)) { + rxe_dbg_mr(mr, "misaligned address"); + return -RXE_ERR_NOT_ALIGNED; + } + + /* Do atomic write after all prior operations have completed */ + smp_store_release(va, value); + + return 0; +#else + WARN_ON(1); + return -EINVAL; +#endif +} + int advance_dma_data(struct rxe_dma_info *dma, unsigned int length) { struct rxe_sge *sge = &dma->sge[dma->cur_sge]; diff --git a/drivers/infiniband/sw/rxe/rxe_resp.c b/drivers/infiniband/sw/rxe/rxe_resp.c index 1e38e5da1f4c..49298ff88d25 100644 --- a/drivers/infiniband/sw/rxe/rxe_resp.c +++ b/drivers/infiniband/sw/rxe/rxe_resp.c @@ -764,30 +764,40 @@ static enum resp_states atomic_reply(struct rxe_qp *qp, return RESPST_ACKNOWLEDGE; } -#ifdef CONFIG_64BIT -static enum resp_states do_atomic_write(struct rxe_qp *qp, - struct rxe_pkt_info *pkt) +static enum resp_states atomic_write_reply(struct rxe_qp *qp, + struct rxe_pkt_info *pkt) { - struct rxe_mr *mr = qp->resp.mr; - int payload = payload_size(pkt); - u64 src, *dst; - - if (mr->state != RXE_MR_STATE_VALID) - return RESPST_ERR_RKEY_VIOLATION; + struct resp_res *res = qp->resp.res; + struct rxe_mr *mr; + u64 value; + u64 iova; + int err; - memcpy(&src, payload_addr(pkt), payload); + if (!res) { + res = rxe_prepare_res(qp, pkt, RXE_ATOMIC_WRITE_MASK); + qp->resp.res = res; + } - dst = iova_to_vaddr(mr, qp->resp.va + qp->resp.offset, payload); - /* check vaddr is 8 bytes aligned. */ - if (!dst || (uintptr_t)dst & 7) - return RESPST_ERR_MISALIGNED_ATOMIC; + if (res->replay) + return RESPST_ACKNOWLEDGE; - /* Do atomic write after all prior operations have completed */ - smp_store_release(dst, src); + mr = qp->resp.mr; + value = *(u64 *)payload_addr(pkt); + iova = qp->resp.va + qp->resp.offset; - /* decrease resp.resid to zero */ - qp->resp.resid -= sizeof(payload); +#if defined CONFIG_64BIT + err = rxe_mr_do_atomic_write(mr, iova, value); + if (unlikely(err)) { + if (err == -RXE_ERR_NOT_ALIGNED) + return RESPST_ERR_MISALIGNED_ATOMIC; + else + return RESPST_ERR_RKEY_VIOLATION; + } +#else + return RESPST_ERR_UNSUPPORTED_OPCODE; +#endif + qp->resp.resid = 0; qp->resp.msn++; /* next expected psn, read handles this separately */ @@ -796,29 +806,8 @@ static enum resp_states do_atomic_write(struct rxe_qp *qp, qp->resp.opcode = pkt->opcode; qp->resp.status = IB_WC_SUCCESS; - return RESPST_ACKNOWLEDGE; -} -#else -static enum resp_states do_atomic_write(struct rxe_qp *qp, - struct rxe_pkt_info *pkt) -{ - return RESPST_ERR_UNSUPPORTED_OPCODE; -} -#endif /* CONFIG_64BIT */ -static enum resp_states atomic_write_reply(struct rxe_qp *qp, - struct rxe_pkt_info *pkt) -{ - struct resp_res *res = qp->resp.res; - - if (!res) { - res = rxe_prepare_res(qp, pkt, RXE_ATOMIC_WRITE_MASK); - qp->resp.res = res; - } - - if (res->replay) - return RESPST_ACKNOWLEDGE; - return do_atomic_write(qp, pkt); + return RESPST_ACKNOWLEDGE; } static struct sk_buff *prepare_ack_packet(struct rxe_qp *qp, From patchwork Mon Jan 16 22:52:28 2023 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Bob Pearson X-Patchwork-Id: 13103818 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 vger.kernel.org (vger.kernel.org [23.128.96.18]) by smtp.lore.kernel.org (Postfix) with ESMTP id 1FA22C67871 for ; Mon, 16 Jan 2023 22:53:58 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S235046AbjAPWxz (ORCPT ); Mon, 16 Jan 2023 17:53:55 -0500 Received: from lindbergh.monkeyblade.net ([23.128.96.19]:37636 "EHLO lindbergh.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S235211AbjAPWxb (ORCPT ); Mon, 16 Jan 2023 17:53:31 -0500 Received: from mail-oi1-x22a.google.com (mail-oi1-x22a.google.com [IPv6:2607:f8b0:4864:20::22a]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id E83432687B for ; Mon, 16 Jan 2023 14:53:18 -0800 (PST) Received: by mail-oi1-x22a.google.com with SMTP id o66so24552234oia.6 for ; Mon, 16 Jan 2023 14:53:18 -0800 (PST) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20210112; h=content-transfer-encoding:mime-version:references:in-reply-to :message-id:date:subject:cc:to:from:from:to:cc:subject:date :message-id:reply-to; bh=+tMrbES7iAPFERVwmrpXMmq8ACmpuWPOnqo2LwvtS/M=; b=SBC5QRpltLHvE77of/qNokBeZ/fqInXFvxvf9WfDPZgMXDl1II+T60Dey3lj0GH9zK qk5P7PQP7q07H4rEZBkCaa06mT6pT74IO0+yq5z0WG54Cryy14ArazEFdbRu23O/uYrf NDSqxUWI5KyQkpNJu5XN5gEVSZIicvNtXvi/gYbsWzkjyAzw2k8gfJ8rTRZsA4gqESbE X+DbnMjr1p4Rq4lCecc+cKZSEGv/x2r1NfKRTTjT5scCD79MYTAWYsllJF7Sbu9bRhC2 LugCXRUtAP2gPJ8vQoSNX1MFWmhzji5c60XvUdwYD3yr6HecAc08fdf9ykg/6sgWa5MK V0Kg== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20210112; h=content-transfer-encoding:mime-version:references:in-reply-to :message-id:date:subject:cc:to:from:x-gm-message-state:from:to:cc :subject:date:message-id:reply-to; bh=+tMrbES7iAPFERVwmrpXMmq8ACmpuWPOnqo2LwvtS/M=; b=jfvRtDh40EemaiUp1L6dsKaykC5PtUcdw1nCix6mFouXUriyr+mJ/EB4zNZ9/A7AM0 2Rz6mSOu+Q43ybTL1evBcb1EqGHCO39+cx1Q7ijxRQ21JoLx3ytBPt3/oI6orvyvl9L8 xvtaugNQHqF4p5Mwyz1qoDlcew/ApkSvFrej1mvx9lFvskbvBHw+ad50yzMPgmW4kz4E 6oBvNb68IVDCcaumfPrB4yVdO3I+ZGbCJ4TSFNmywhyppcYrt2dEqhZOAXvcV6aB8Lbn JaJODrjExaykRcDrL77h+tNoDv9AtH5Uc5yedv/51crlz4Xzpljpz2EN4DthSHpfnBik bbNg== X-Gm-Message-State: AFqh2kr6SwEecCuyU3aNdEE3M9D/cE2fdp++lajL/wHU5x8n0tq0XGTW cqCSH/t4HUuTyeF+CETsfCw= X-Google-Smtp-Source: AMrXdXvRaBXYVnvD/HiWG8Z0S2bumH7i3jL2zKBKFLlzeablxZC+RSc5RHvHlGw5HvgWeL1shvasFg== X-Received: by 2002:a05:6808:1a1d:b0:35e:646b:2cd6 with SMTP id bk29-20020a0568081a1d00b0035e646b2cd6mr490169oib.59.1673909598213; Mon, 16 Jan 2023 14:53:18 -0800 (PST) Received: from rpearson-X570-AORUS-PRO-WIFI.tx.rr.com (2603-8081-140c-1a00-ea18-3ee9-26d1-7526.res6.spectrum.com. [2603:8081:140c:1a00:ea18:3ee9:26d1:7526]) by smtp.gmail.com with ESMTPSA id s4-20020a056808008400b0035028730c90sm13651937oic.1.2023.01.16.14.53.17 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Mon, 16 Jan 2023 14:53:17 -0800 (PST) From: Bob Pearson To: jgg@nvidia.com, zyjzyj2000@gmail.com, leonro@nvidia.com, yangx.jy@fujitsu.com, linux-rdma@vger.kernel.org Cc: Bob Pearson Subject: [PATCH for-next v4 5/6] RDMA/rxe: Cleanup page variables in rxe_mr.c Date: Mon, 16 Jan 2023 16:52:28 -0600 Message-Id: <20230116225227.21163-6-rpearsonhpe@gmail.com> X-Mailer: git-send-email 2.37.2 In-Reply-To: <20230116225227.21163-1-rpearsonhpe@gmail.com> References: <20230116225227.21163-1-rpearsonhpe@gmail.com> MIME-Version: 1.0 Precedence: bulk List-ID: X-Mailing-List: linux-rdma@vger.kernel.org Cleanup usage of mr->page_shift and mr->page_mask and introduce an extractor for mr->ibmr.page_size. Normal usage in the kernel has page_mask masking out offset in page rather than masking out the page number. The rxe driver had reversed that which was confusing. Implicitly there can be a per mr page_size which was not uniformly supported. Signed-off-by: Bob Pearson --- drivers/infiniband/sw/rxe/rxe_mr.c | 31 ++++++++++++--------------- drivers/infiniband/sw/rxe/rxe_verbs.h | 11 +++++++--- 2 files changed, 22 insertions(+), 20 deletions(-) diff --git a/drivers/infiniband/sw/rxe/rxe_mr.c b/drivers/infiniband/sw/rxe/rxe_mr.c index 10484f671977..fdf76df4cf3e 100644 --- a/drivers/infiniband/sw/rxe/rxe_mr.c +++ b/drivers/infiniband/sw/rxe/rxe_mr.c @@ -62,6 +62,9 @@ static void rxe_mr_init(int access, struct rxe_mr *mr) mr->lkey = mr->ibmr.lkey = lkey; mr->rkey = mr->ibmr.rkey = rkey; + mr->ibmr.page_size = PAGE_SIZE; + mr->page_mask = PAGE_MASK; + mr->page_shift = PAGE_SHIFT; mr->state = RXE_MR_STATE_INVALID; } @@ -151,9 +154,6 @@ int rxe_mr_init_user(struct rxe_dev *rxe, u64 start, u64 length, u64 iova, goto err_release_umem; } - mr->page_shift = PAGE_SHIFT; - mr->page_mask = PAGE_SIZE - 1; - num_buf = 0; map = mr->map; if (length > 0) { @@ -182,7 +182,7 @@ int rxe_mr_init_user(struct rxe_dev *rxe, u64 start, u64 length, u64 iova, goto err_release_umem; } buf->addr = (uintptr_t)vaddr; - buf->size = PAGE_SIZE; + buf->size = mr_page_size(mr); num_buf++; buf++; @@ -191,10 +191,9 @@ int rxe_mr_init_user(struct rxe_dev *rxe, u64 start, u64 length, u64 iova, mr->umem = umem; mr->access = access; - mr->offset = ib_umem_offset(umem); + mr->page_offset = ib_umem_offset(umem); mr->state = RXE_MR_STATE_VALID; mr->ibmr.type = IB_MR_TYPE_USER; - mr->ibmr.page_size = PAGE_SIZE; return 0; @@ -248,29 +247,27 @@ int rxe_map_mr_sg(struct ib_mr *ibmr, struct scatterlist *sg, int sg_nents, unsigned int *sg_offset) { struct rxe_mr *mr = to_rmr(ibmr); - int n; - - mr->nbuf = 0; + unsigned int page_size = mr_page_size(mr); - n = ib_sg_to_pages(ibmr, sg, sg_nents, sg_offset, rxe_set_page); + mr->page_shift = ilog2(page_size); + mr->page_mask = ~((u64)page_size - 1); + mr->page_offset = ibmr->iova & (page_size - 1); - mr->page_shift = ilog2(ibmr->page_size); - mr->page_mask = ibmr->page_size - 1; - mr->offset = ibmr->iova & mr->page_mask; + mr->nbuf = 0; - return n; + return ib_sg_to_pages(ibmr, sg, sg_nents, sg_offset, rxe_set_page); } static void lookup_iova(struct rxe_mr *mr, u64 iova, int *m_out, int *n_out, size_t *offset_out) { - size_t offset = iova - mr->ibmr.iova + mr->offset; + size_t offset = iova - mr->ibmr.iova + mr->page_offset; int map_index; int buf_index; u64 length; if (likely(mr->page_shift)) { - *offset_out = offset & mr->page_mask; + *offset_out = offset & (mr_page_size(mr) - 1); offset >>= mr->page_shift; *n_out = offset & mr->map_mask; *m_out = offset >> mr->map_shift; @@ -329,7 +326,7 @@ int rxe_flush_pmem_iova(struct rxe_mr *mr, u64 iova, int length) if (mr->ibmr.type == IB_MR_TYPE_DMA) return -EFAULT; - offset = (iova - mr->ibmr.iova + mr->offset) & mr->page_mask; + offset = (iova - mr->ibmr.iova + mr->page_offset) & mr->page_mask; while (length > 0) { u8 *va; int bytes; diff --git a/drivers/infiniband/sw/rxe/rxe_verbs.h b/drivers/infiniband/sw/rxe/rxe_verbs.h index 691200d99d6b..5c3d1500ca68 100644 --- a/drivers/infiniband/sw/rxe/rxe_verbs.h +++ b/drivers/infiniband/sw/rxe/rxe_verbs.h @@ -315,11 +315,11 @@ struct rxe_mr { u32 lkey; u32 rkey; enum rxe_mr_state state; - u32 offset; int access; - int page_shift; - int page_mask; + unsigned int page_offset; + unsigned int page_shift; + u64 page_mask; int map_shift; int map_mask; @@ -334,6 +334,11 @@ struct rxe_mr { struct rxe_map **map; }; +static inline unsigned int mr_page_size(struct rxe_mr *mr) +{ + return mr ? mr->ibmr.page_size : PAGE_SIZE; +} + enum rxe_mw_state { RXE_MW_STATE_INVALID = RXE_MR_STATE_INVALID, RXE_MW_STATE_FREE = RXE_MR_STATE_FREE, From patchwork Mon Jan 16 22:52:29 2023 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Bob Pearson X-Patchwork-Id: 13103819 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 vger.kernel.org (vger.kernel.org [23.128.96.18]) by smtp.lore.kernel.org (Postfix) with ESMTP id E60DBC46467 for ; Mon, 16 Jan 2023 22:53:58 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S235157AbjAPWx4 (ORCPT ); Mon, 16 Jan 2023 17:53:56 -0500 Received: from lindbergh.monkeyblade.net ([23.128.96.19]:38778 "EHLO lindbergh.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S235215AbjAPWxb (ORCPT ); Mon, 16 Jan 2023 17:53:31 -0500 Received: from mail-oi1-x229.google.com (mail-oi1-x229.google.com [IPv6:2607:f8b0:4864:20::229]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id F323627D71 for ; Mon, 16 Jan 2023 14:53:19 -0800 (PST) Received: by mail-oi1-x229.google.com with SMTP id s66so23911094oib.7 for ; Mon, 16 Jan 2023 14:53:19 -0800 (PST) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20210112; h=content-transfer-encoding:mime-version:references:in-reply-to :message-id:date:subject:cc:to:from:from:to:cc:subject:date :message-id:reply-to; bh=mw45/+onr1pgoxHNB/goyJ5gd+/ZzFqP4Lt0hHkxXsM=; b=BwlW8ry4I4+iLBan3prn3Ia9/T2nJWRtnvLrQWydRybkbO/U53PeOOYX5OGhu/ZWHg o9lJ4sGZyFKfXKGffTV/zssBuXSo672GCaeH+mLucjXKRIdb7plrLkgCLHYlATifvEHr zfU/hjKUdsIq+/RXmn0Onmzzwt8hRyMBzMSo2z/Cznf3/gnHcny4M6j/ezd9YHK42Y1c aW49N0U77JPoL+5GjKiccW8E0hN8PapYWxRP2DGQfgH1C2HGC7noQE/sAq1+4PcTw8fh h5hMSsQjv3HvbDPdxgJ5bdavOYHYXKO+FnAdSsk3nM3ooQR8QmbEPaYAllCsO3ef7JQE bBOA== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20210112; h=content-transfer-encoding:mime-version:references:in-reply-to :message-id:date:subject:cc:to:from:x-gm-message-state:from:to:cc :subject:date:message-id:reply-to; bh=mw45/+onr1pgoxHNB/goyJ5gd+/ZzFqP4Lt0hHkxXsM=; b=KtKIBn+alTc7mnPH/smWlVNHUqs2Y5B+v6FCvntS6lUxZpBZcqx5ysH8NFLcDSscOQ kvV/B1OGKbWQ3XeLdWlCgpmQjKidgTZKT/3mBrSFdCZfTgRzjvGBuoCX07yzm3GqSTpr 62IM827YUvoeGVYfpaqGgn9tvuHstuwDEwHzswki30LQv7ESxNOEvkjw7VjkAoQYIs9I QxKl7rD1Fh7ZEu6AoS9nkkNaLzrTAHdbTwW9rJYhXVnGhYwqp9zdG7hD7QEC6rrip3zs 1EeGbKRr6oDRJUTrrF9H9ZjMdDsSjo4uuV3t5fnnW2JTNNWMGNuwwbOaGEF6dKUsOoRu oRbg== X-Gm-Message-State: AFqh2krtlqr3a1r0cx33fDsfDF8V2fTSXl1DISRiMazEE1eKkFQIh6Lm Wp2Ew0zM8FTwCpIzP8RQwvg= X-Google-Smtp-Source: AMrXdXsLb54aXf/T2ZoScnrkXR7e4H4m/G8EY+SoIco83NmxlIGAdzJSRLoRvQH8FoCEM3i2UulWmA== X-Received: by 2002:aca:5a0b:0:b0:35e:d1f9:d6e8 with SMTP id o11-20020aca5a0b000000b0035ed1f9d6e8mr388790oib.13.1673909599135; Mon, 16 Jan 2023 14:53:19 -0800 (PST) Received: from rpearson-X570-AORUS-PRO-WIFI.tx.rr.com (2603-8081-140c-1a00-ea18-3ee9-26d1-7526.res6.spectrum.com. [2603:8081:140c:1a00:ea18:3ee9:26d1:7526]) by smtp.gmail.com with ESMTPSA id s4-20020a056808008400b0035028730c90sm13651937oic.1.2023.01.16.14.53.18 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Mon, 16 Jan 2023 14:53:18 -0800 (PST) From: Bob Pearson To: jgg@nvidia.com, zyjzyj2000@gmail.com, leonro@nvidia.com, yangx.jy@fujitsu.com, linux-rdma@vger.kernel.org Cc: Bob Pearson Subject: [PATCH for-next v4 6/6] RDMA/rxe: Replace rxe_map and rxe_phys_buf by xarray Date: Mon, 16 Jan 2023 16:52:29 -0600 Message-Id: <20230116225227.21163-7-rpearsonhpe@gmail.com> X-Mailer: git-send-email 2.37.2 In-Reply-To: <20230116225227.21163-1-rpearsonhpe@gmail.com> References: <20230116225227.21163-1-rpearsonhpe@gmail.com> MIME-Version: 1.0 Precedence: bulk List-ID: X-Mailing-List: linux-rdma@vger.kernel.org Replace struct rxe-phys_buf and struct rxe_map by struct xarray in rxe_verbs.h. This allows using rcu locking on reads for the memory maps stored in each mr. This is based off of a sketch of a patch from Jason Gunthorpe in the link below. Some changes were needed to make this work. It applies cleanly to the current for-next and passes the pyverbs, perftest and the same blktests test cases which run today. Link: https://lore.kernel.org/linux-rdma/Y3gvZr6%2FNCii9Avy@nvidia.com/ Co-developed-by: Jason Gunthorpe Signed-off-by: Bob Pearson --- drivers/infiniband/sw/rxe/rxe_loc.h | 8 +- drivers/infiniband/sw/rxe/rxe_mr.c | 544 ++++++++++++-------------- drivers/infiniband/sw/rxe/rxe_verbs.h | 21 +- 3 files changed, 265 insertions(+), 308 deletions(-) diff --git a/drivers/infiniband/sw/rxe/rxe_loc.h b/drivers/infiniband/sw/rxe/rxe_loc.h index b1dda0cf891b..c1f78ed0f991 100644 --- a/drivers/infiniband/sw/rxe/rxe_loc.h +++ b/drivers/infiniband/sw/rxe/rxe_loc.h @@ -64,14 +64,14 @@ void rxe_mr_init_dma(int access, struct rxe_mr *mr); int rxe_mr_init_user(struct rxe_dev *rxe, u64 start, u64 length, u64 iova, int access, struct rxe_mr *mr); int rxe_mr_init_fast(int max_pages, struct rxe_mr *mr); -int rxe_flush_pmem_iova(struct rxe_mr *mr, u64 iova, int length); -int rxe_mr_copy(struct rxe_mr *mr, u64 iova, void *addr, int length, - enum rxe_mr_copy_dir dir); +int rxe_flush_pmem_iova(struct rxe_mr *mr, u64 iova, + unsigned int length); +int rxe_mr_copy(struct rxe_mr *mr, u64 iova, void *addr, + unsigned int length, enum rxe_mr_copy_dir dir); int copy_data(struct rxe_pd *pd, int access, struct rxe_dma_info *dma, void *addr, int length, enum rxe_mr_copy_dir dir); int rxe_map_mr_sg(struct ib_mr *ibmr, struct scatterlist *sg, int sg_nents, unsigned int *sg_offset); -void *iova_to_vaddr(struct rxe_mr *mr, u64 iova, int length); int rxe_mr_do_atomic_op(struct rxe_mr *mr, u64 iova, int opcode, u64 compare, u64 swap_add, u64 *orig_val); int rxe_mr_do_atomic_write(struct rxe_mr *mr, u64 iova, u64 value); diff --git a/drivers/infiniband/sw/rxe/rxe_mr.c b/drivers/infiniband/sw/rxe/rxe_mr.c index fdf76df4cf3e..cadc3469b64c 100644 --- a/drivers/infiniband/sw/rxe/rxe_mr.c +++ b/drivers/infiniband/sw/rxe/rxe_mr.c @@ -62,145 +62,113 @@ static void rxe_mr_init(int access, struct rxe_mr *mr) mr->lkey = mr->ibmr.lkey = lkey; mr->rkey = mr->ibmr.rkey = rkey; + mr->access = access; mr->ibmr.page_size = PAGE_SIZE; mr->page_mask = PAGE_MASK; mr->page_shift = PAGE_SHIFT; mr->state = RXE_MR_STATE_INVALID; } -static int rxe_mr_alloc(struct rxe_mr *mr, int num_buf) -{ - int i; - int num_map; - struct rxe_map **map = mr->map; - - num_map = (num_buf + RXE_BUF_PER_MAP - 1) / RXE_BUF_PER_MAP; - - mr->map = kmalloc_array(num_map, sizeof(*map), GFP_KERNEL); - if (!mr->map) - goto err1; - - for (i = 0; i < num_map; i++) { - mr->map[i] = kmalloc(sizeof(**map), GFP_KERNEL); - if (!mr->map[i]) - goto err2; - } - - BUILD_BUG_ON(!is_power_of_2(RXE_BUF_PER_MAP)); - - mr->map_shift = ilog2(RXE_BUF_PER_MAP); - mr->map_mask = RXE_BUF_PER_MAP - 1; - - mr->num_buf = num_buf; - mr->num_map = num_map; - mr->max_buf = num_map * RXE_BUF_PER_MAP; - - return 0; - -err2: - for (i--; i >= 0; i--) - kfree(mr->map[i]); - - kfree(mr->map); - mr->map = NULL; -err1: - return -ENOMEM; -} - void rxe_mr_init_dma(int access, struct rxe_mr *mr) { rxe_mr_init(access, mr); - mr->access = access; mr->state = RXE_MR_STATE_VALID; mr->ibmr.type = IB_MR_TYPE_DMA; } -static bool is_pmem_page(struct page *pg) +static unsigned long rxe_mr_iova_to_index(struct rxe_mr *mr, u64 iova) { - unsigned long paddr = page_to_phys(pg); + return (iova >> mr->page_shift) - (mr->ibmr.iova >> mr->page_shift); +} - return REGION_INTERSECTS == - region_intersects(paddr, PAGE_SIZE, IORESOURCE_MEM, - IORES_DESC_PERSISTENT_MEMORY); +static unsigned long rxe_mr_iova_to_page_offset(struct rxe_mr *mr, u64 iova) +{ + return iova & (mr_page_size(mr) - 1); +} + +static int rxe_mr_fill_pages_from_sgt(struct rxe_mr *mr, struct sg_table *sgt) +{ + XA_STATE(xas, &mr->page_list, 0); + struct sg_page_iter sg_iter; + + xa_init(&mr->page_list); + + __sg_page_iter_start(&sg_iter, sgt->sgl, sgt->orig_nents, 0); + if (!__sg_page_iter_next(&sg_iter)) + return 0; + + do { + xas_lock(&xas); + while (true) { + xas_store(&xas, sg_page_iter_page(&sg_iter)); + if (xas_error(&xas)) + break; + xas_next(&xas); + if (!__sg_page_iter_next(&sg_iter)) + break; + } + xas_unlock(&xas); + } while (xas_nomem(&xas, GFP_KERNEL)); + + return xas_error(&xas); } int rxe_mr_init_user(struct rxe_dev *rxe, u64 start, u64 length, u64 iova, int access, struct rxe_mr *mr) { - struct rxe_map **map; - struct rxe_phys_buf *buf = NULL; - struct ib_umem *umem; - struct sg_page_iter sg_iter; - int num_buf; - void *vaddr; + struct ib_umem *umem; int err; + rxe_mr_init(access, mr); + umem = ib_umem_get(&rxe->ib_dev, start, length, access); if (IS_ERR(umem)) { rxe_dbg_mr(mr, "Unable to pin memory region err = %d\n", (int)PTR_ERR(umem)); - err = PTR_ERR(umem); - goto err_out; + return PTR_ERR(umem); } - num_buf = ib_umem_num_pages(umem); - - rxe_mr_init(access, mr); - - err = rxe_mr_alloc(mr, num_buf); + err = rxe_mr_fill_pages_from_sgt(mr, &umem->sgt_append.sgt); if (err) { - rxe_dbg_mr(mr, "Unable to allocate memory for map\n"); - goto err_release_umem; + ib_umem_release(umem); + return err; } - num_buf = 0; - map = mr->map; - if (length > 0) { - bool persistent_access = access & IB_ACCESS_FLUSH_PERSISTENT; + mr->umem = umem; + mr->ibmr.type = IB_MR_TYPE_USER; + mr->state = RXE_MR_STATE_VALID; - buf = map[0]->buf; - for_each_sgtable_page (&umem->sgt_append.sgt, &sg_iter, 0) { - struct page *pg = sg_page_iter_page(&sg_iter); + return 0; +} - if (persistent_access && !is_pmem_page(pg)) { - rxe_dbg_mr(mr, "Unable to register persistent access to non-pmem device\n"); - err = -EINVAL; - goto err_release_umem; - } +static int rxe_mr_alloc(struct rxe_mr *mr, int num_buf) +{ + XA_STATE(xas, &mr->page_list, 0); + int i = 0; + int err; - if (num_buf >= RXE_BUF_PER_MAP) { - map++; - buf = map[0]->buf; - num_buf = 0; - } - - vaddr = page_address(pg); - if (!vaddr) { - rxe_dbg_mr(mr, "Unable to get virtual address\n"); - err = -ENOMEM; - goto err_release_umem; - } - buf->addr = (uintptr_t)vaddr; - buf->size = mr_page_size(mr); - num_buf++; - buf++; + xa_init(&mr->page_list); + do { + xas_lock(&xas); + while (i != num_buf) { + xas_store(&xas, XA_ZERO_ENTRY); + if (xas_error(&xas)) + break; + xas_next(&xas); + i++; } - } + xas_unlock(&xas); + } while (xas_nomem(&xas, GFP_KERNEL)); - mr->umem = umem; - mr->access = access; - mr->page_offset = ib_umem_offset(umem); - mr->state = RXE_MR_STATE_VALID; - mr->ibmr.type = IB_MR_TYPE_USER; + err = xas_error(&xas); + if (err) + return err; - return 0; + mr->num_buf = num_buf; -err_release_umem: - ib_umem_release(umem); -err_out: - return err; + return 0; } int rxe_mr_init_fast(int max_pages, struct rxe_mr *mr) @@ -214,7 +182,6 @@ int rxe_mr_init_fast(int max_pages, struct rxe_mr *mr) if (err) goto err1; - mr->max_buf = max_pages; mr->state = RXE_MR_STATE_FREE; mr->ibmr.type = IB_MR_TYPE_MEM_REG; @@ -224,206 +191,138 @@ int rxe_mr_init_fast(int max_pages, struct rxe_mr *mr) return err; } -static int rxe_set_page(struct ib_mr *ibmr, u64 addr) +static int rxe_set_page(struct ib_mr *ibmr, u64 iova) { struct rxe_mr *mr = to_rmr(ibmr); - struct rxe_map *map; - struct rxe_phys_buf *buf; + struct page *page = virt_to_page(iova & mr->page_mask); + XA_STATE(xas, &mr->page_list, mr->nbuf); + int err; if (unlikely(mr->nbuf == mr->num_buf)) return -ENOMEM; - map = mr->map[mr->nbuf / RXE_BUF_PER_MAP]; - buf = &map->buf[mr->nbuf % RXE_BUF_PER_MAP]; + do { + xas_lock(&xas); + xas_store(&xas, page); + xas_unlock(&xas); + } while (xas_nomem(&xas, GFP_KERNEL)); - buf->addr = addr; - buf->size = ibmr->page_size; - mr->nbuf++; + err = xas_error(&xas); + if (err) + return err; + mr->nbuf++; return 0; } -int rxe_map_mr_sg(struct ib_mr *ibmr, struct scatterlist *sg, +int rxe_map_mr_sg(struct ib_mr *ibmr, struct scatterlist *sgl, int sg_nents, unsigned int *sg_offset) { struct rxe_mr *mr = to_rmr(ibmr); unsigned int page_size = mr_page_size(mr); + mr->nbuf = 0; mr->page_shift = ilog2(page_size); mr->page_mask = ~((u64)page_size - 1); - mr->page_offset = ibmr->iova & (page_size - 1); + mr->page_offset = mr->ibmr.iova & (page_size - 1); - mr->nbuf = 0; - - return ib_sg_to_pages(ibmr, sg, sg_nents, sg_offset, rxe_set_page); + return ib_sg_to_pages(ibmr, sgl, sg_nents, sg_offset, rxe_set_page); } -static void lookup_iova(struct rxe_mr *mr, u64 iova, int *m_out, int *n_out, - size_t *offset_out) +/* + * TODO: Attempting to modify the mr page map between the time + * a packet is received and the map is referenced as here + * in xa_load(&mr->page_list) will cause problems. It is OK to + * deregister the mr since the mr reference counts will preserve + * it until memory accesses are complete. Currently reregister mr + * operations are not supported by the rxe driver but could be + * in the future. Invalidate followed by fast_reg mr will change + * the map and then the rkey so delayed packets arriving in the + * middle could use the wrong map entries. This isn't new but was + * already the case in the earlier implementation. This won't be + * a problem for well behaved programs which wait until all the + * outstanding packets for the first FMR before remapping to the + * second. + */ +static int rxe_mr_copy_xarray(struct rxe_mr *mr, u64 iova, void *addr, + unsigned int length, enum rxe_mr_copy_dir dir) { - size_t offset = iova - mr->ibmr.iova + mr->page_offset; - int map_index; - int buf_index; - u64 length; - - if (likely(mr->page_shift)) { - *offset_out = offset & (mr_page_size(mr) - 1); - offset >>= mr->page_shift; - *n_out = offset & mr->map_mask; - *m_out = offset >> mr->map_shift; - } else { - map_index = 0; - buf_index = 0; - - length = mr->map[map_index]->buf[buf_index].size; + unsigned int page_offset = rxe_mr_iova_to_page_offset(mr, iova); + unsigned long index = rxe_mr_iova_to_index(mr, iova); + unsigned int bytes; + struct page *page; + void *va; - while (offset >= length) { - offset -= length; - buf_index++; - - if (buf_index == RXE_BUF_PER_MAP) { - map_index++; - buf_index = 0; - } - length = mr->map[map_index]->buf[buf_index].size; - } + while (length) { + page = xa_load(&mr->page_list, index); + if (!page) + return -EFAULT; - *m_out = map_index; - *n_out = buf_index; - *offset_out = offset; + bytes = min_t(unsigned int, length, + mr_page_size(mr) - page_offset); + va = kmap_local_page(page); + if (dir == RXE_FROM_MR_OBJ) + memcpy(addr, va + page_offset, bytes); + else + memcpy(va + page_offset, addr, bytes); + kunmap_local(va); + + page_offset = 0; + addr += bytes; + length -= bytes; + index++; } -} - -void *iova_to_vaddr(struct rxe_mr *mr, u64 iova, int length) -{ - size_t offset; - int m, n; - - if (mr->state != RXE_MR_STATE_VALID) - return NULL; - - if (mr->ibmr.type == IB_MR_TYPE_DMA) - return (void *)(uintptr_t)iova; - - if (mr_check_range(mr, iova, length)) - return NULL; - - lookup_iova(mr, iova, &m, &n, &offset); - - if (offset + length > mr->map[m]->buf[n].size) - return NULL; - return (void *)(uintptr_t)mr->map[m]->buf[n].addr + offset; + return 0; } -int rxe_flush_pmem_iova(struct rxe_mr *mr, u64 iova, int length) +static void rxe_mr_copy_dma(struct rxe_mr *mr, u64 iova, void *addr, + unsigned int length, enum rxe_mr_copy_dir dir) { - size_t offset; - - if (length == 0) - return 0; + unsigned int page_offset = iova & (PAGE_SIZE - 1); + unsigned int bytes; + struct page *page; + u8 *va; - if (mr->ibmr.type == IB_MR_TYPE_DMA) - return -EFAULT; - - offset = (iova - mr->ibmr.iova + mr->page_offset) & mr->page_mask; - while (length > 0) { - u8 *va; - int bytes; - - bytes = mr->ibmr.page_size - offset; - if (bytes > length) - bytes = length; - - va = iova_to_vaddr(mr, iova, length); - if (!va) - return -EFAULT; - - arch_wb_cache_pmem(va, bytes); - - length -= bytes; + while (length) { + page = virt_to_page(iova & mr->page_mask); + bytes = min_t(unsigned int, length, + PAGE_SIZE - page_offset); + va = kmap_local_page(page); + + if (dir == RXE_TO_MR_OBJ) + memcpy(va + page_offset, addr, bytes); + else + memcpy(addr, va + page_offset, bytes); + + kunmap_local(va); + page_offset = 0; iova += bytes; - offset = 0; + addr += bytes; + length -= bytes; } - - return 0; } -/* copy data from a range (vaddr, vaddr+length-1) to or from - * a mr object starting at iova. - */ -int rxe_mr_copy(struct rxe_mr *mr, u64 iova, void *addr, int length, - enum rxe_mr_copy_dir dir) +int rxe_mr_copy(struct rxe_mr *mr, u64 iova, void *addr, + unsigned int length, enum rxe_mr_copy_dir dir) { - int err; - int bytes; - u8 *va; - struct rxe_map **map; - struct rxe_phys_buf *buf; - int m; - int i; - size_t offset; + int err; if (length == 0) return 0; if (mr->ibmr.type == IB_MR_TYPE_DMA) { - u8 *src, *dest; - - src = (dir == RXE_TO_MR_OBJ) ? addr : ((void *)(uintptr_t)iova); - - dest = (dir == RXE_TO_MR_OBJ) ? ((void *)(uintptr_t)iova) : addr; - - memcpy(dest, src, length); - + rxe_mr_copy_dma(mr, iova, addr, length, dir); return 0; } - WARN_ON_ONCE(!mr->map); - err = mr_check_range(mr, iova, length); - if (err) { - err = -EFAULT; - goto err1; - } - - lookup_iova(mr, iova, &m, &i, &offset); - - map = mr->map + m; - buf = map[0]->buf + i; - - while (length > 0) { - u8 *src, *dest; - - va = (u8 *)(uintptr_t)buf->addr + offset; - src = (dir == RXE_TO_MR_OBJ) ? addr : va; - dest = (dir == RXE_TO_MR_OBJ) ? va : addr; - - bytes = buf->size - offset; - - if (bytes > length) - bytes = length; - - memcpy(dest, src, bytes); - - length -= bytes; - addr += bytes; - - offset = 0; - buf++; - i++; - - if (i == RXE_BUF_PER_MAP) { - i = 0; - map++; - buf = map[0]->buf; - } + if (unlikely(err)) { + rxe_dbg_mr(mr, "iova out of range"); + return err; } - return 0; - -err1: - return err; + return rxe_mr_copy_xarray(mr, iova, addr, length, dir); } /* copy data in or out of a wqe, i.e. sg list @@ -495,7 +394,6 @@ int copy_data( if (bytes > 0) { iova = sge->addr + offset; - err = rxe_mr_copy(mr, iova, addr, bytes, dir); if (err) goto err2; @@ -522,43 +420,104 @@ int copy_data( return err; } +int rxe_flush_pmem_iova(struct rxe_mr *mr, u64 iova, unsigned int length) +{ + unsigned int page_offset; + unsigned long index; + struct page *page; + unsigned int bytes; + int err; + u8 *va; + + if (length == 0) + return 0; + + if (mr->ibmr.type == IB_MR_TYPE_DMA) + return -EFAULT; + + err = mr_check_range(mr, iova, length); + if (err) + return err; + + while (length > 0) { + index = rxe_mr_iova_to_index(mr, iova); + page = xa_load(&mr->page_list, index); + page_offset = rxe_mr_iova_to_page_offset(mr, iova); + if (!page) + return -EFAULT; + bytes = min_t(unsigned int, length, + mr_page_size(mr) - page_offset); + + va = kmap_local_page(page); + if (!va) + return -EFAULT; + + arch_wb_cache_pmem(va + page_offset, bytes); + + length -= bytes; + iova += bytes; + page_offset = 0; + } + + return 0; +} + /* Guarantee atomicity of atomic operations at the machine level. */ static DEFINE_SPINLOCK(atomic_ops_lock); int rxe_mr_do_atomic_op(struct rxe_mr *mr, u64 iova, int opcode, u64 compare, u64 swap_add, u64 *orig_val) { - u64 *va; + unsigned int page_offset; + struct page *page; u64 value; + u64 *va; - if (mr->state != RXE_MR_STATE_VALID) { + if (unlikely(mr->state != RXE_MR_STATE_VALID)) { rxe_dbg_mr(mr, "mr not in valid state"); return -EINVAL; } - va = iova_to_vaddr(mr, iova, sizeof(u64)); - if (!va) { - rxe_dbg_mr(mr, "iova out of range"); - return -ERANGE; + if (mr->ibmr.type == IB_MR_TYPE_DMA) { + page_offset = iova & (PAGE_SIZE - 1); + page = virt_to_page(iova & PAGE_MASK); + } else { + unsigned long index; + int err; + + err = mr_check_range(mr, iova, sizeof(value)); + if (err) { + rxe_dbg_mr(mr, "iova out of range"); + return -ERANGE; + } + page_offset = rxe_mr_iova_to_page_offset(mr, iova); + index = rxe_mr_iova_to_index(mr, iova); + page = xa_load(&mr->page_list, index); + if (!page) + return -EFAULT; } - if ((uintptr_t)va & 0x7) { + if (unlikely(page_offset & 0x7)) { rxe_dbg_mr(mr, "iova not aligned"); - return RXE_ERR_NOT_ALIGNED; + return -RXE_ERR_NOT_ALIGNED; } + va = kmap_local_page(page); + spin_lock_bh(&atomic_ops_lock); - value = *orig_val = *va; + value = *orig_val = va[page_offset >> 3]; if (opcode == IB_OPCODE_RC_COMPARE_SWAP) { if (value == compare) - *va = swap_add; + va[page_offset >> 3] = swap_add; } else { value += swap_add; - *va = value; + va[page_offset >> 3] = value; } spin_unlock_bh(&atomic_ops_lock); + kunmap_local(va); + return 0; } @@ -566,6 +525,8 @@ int rxe_mr_do_atomic_op(struct rxe_mr *mr, u64 iova, int opcode, int rxe_mr_do_atomic_write(struct rxe_mr *mr, u64 iova, u64 value) { #if defined CONFIG_64BIT + unsigned int page_offset; + struct page *page; u64 *va; /* See IBA oA19-28 */ @@ -574,20 +535,45 @@ int rxe_mr_do_atomic_write(struct rxe_mr *mr, u64 iova, u64 value) return -EINVAL; } - va = iova_to_vaddr(mr, iova, sizeof(value)); - if (unlikely(!va)) { - rxe_dbg_mr(mr, "iova out of range"); - return -ERANGE; + if (mr->ibmr.type == IB_MR_TYPE_DMA) { + page_offset = iova & (PAGE_SIZE - 1); + page = virt_to_page(iova & PAGE_MASK); + } else { + unsigned long index; + int err; + + /* See IBA oA19-28 */ + err = mr_check_range(mr, iova, sizeof(value)); + if (unlikely(err)) { + rxe_dbg_mr(mr, "iova out of range"); + return -ERANGE; + } + page_offset = rxe_mr_iova_to_page_offset(mr, iova); + index = rxe_mr_iova_to_index(mr, iova); + page = xa_load(&mr->page_list, index); + if (!page) + return -EFAULT; } /* See IBA A19.4.2 */ - if (unlikely((uintptr_t)va & 0x7 || iova & 0x7)) { + if (unlikely(page_offset & 0x7)) { rxe_dbg_mr(mr, "misaligned address"); return -RXE_ERR_NOT_ALIGNED; } + va = kmap_local_page(page); + /* Do atomic write after all prior operations have completed */ - smp_store_release(va, value); + /* TODO: This is what was chosen by the implementer but I am + * concerned it isn't what they want. This only guarantees that + * the write will complete before any subsequent reads but the + * comment says all prior operations have completed. That would + * require a full mb or matching acquire. + * Normal usage has a matching load_acquire and store release. + */ + smp_store_release(&va[page_offset >> 3], value); + + kunmap_local(va); return 0; #else @@ -629,12 +615,6 @@ int advance_dma_data(struct rxe_dma_info *dma, unsigned int length) return 0; } -/* (1) find the mr corresponding to lkey/rkey - * depending on lookup_type - * (2) verify that the (qp) pd matches the mr pd - * (3) verify that the mr can support the requested access - * (4) verify that mr state is valid - */ struct rxe_mr *lookup_mr(struct rxe_pd *pd, int access, u32 key, enum rxe_mr_lookup_type type) { @@ -755,15 +735,9 @@ int rxe_dereg_mr(struct ib_mr *ibmr, struct ib_udata *udata) void rxe_mr_cleanup(struct rxe_pool_elem *elem) { struct rxe_mr *mr = container_of(elem, typeof(*mr), elem); - int i; rxe_put(mr_pd(mr)); ib_umem_release(mr->umem); - if (mr->map) { - for (i = 0; i < mr->num_map; i++) - kfree(mr->map[i]); - - kfree(mr->map); - } + xa_destroy(&mr->page_list); } diff --git a/drivers/infiniband/sw/rxe/rxe_verbs.h b/drivers/infiniband/sw/rxe/rxe_verbs.h index 5c3d1500ca68..09437077ceee 100644 --- a/drivers/infiniband/sw/rxe/rxe_verbs.h +++ b/drivers/infiniband/sw/rxe/rxe_verbs.h @@ -288,17 +288,6 @@ enum rxe_mr_lookup_type { RXE_LOOKUP_REMOTE, }; -#define RXE_BUF_PER_MAP (PAGE_SIZE / sizeof(struct rxe_phys_buf)) - -struct rxe_phys_buf { - u64 addr; - u64 size; -}; - -struct rxe_map { - struct rxe_phys_buf buf[RXE_BUF_PER_MAP]; -}; - static inline int rkey_is_mw(u32 rkey) { u32 index = rkey >> 8; @@ -316,22 +305,16 @@ struct rxe_mr { u32 rkey; enum rxe_mr_state state; int access; + atomic_t num_mw; unsigned int page_offset; unsigned int page_shift; u64 page_mask; - int map_shift; - int map_mask; u32 num_buf; u32 nbuf; - u32 max_buf; - u32 num_map; - - atomic_t num_mw; - - struct rxe_map **map; + struct xarray page_list; }; static inline unsigned int mr_page_size(struct rxe_mr *mr)