From patchwork Mon Nov 18 01:08:02 2024 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Hou Tao X-Patchwork-Id: 13878001 X-Patchwork-Delegate: bpf@iogearbox.net Received: from dggsgout11.his.huawei.com (dggsgout11.his.huawei.com [45.249.212.51]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id 6A34E8BFF for ; Mon, 18 Nov 2024 01:13:55 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=45.249.212.51 ARC-Seal: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1731892437; cv=none; b=iAwmWv5rtdMh+m0vEDgQdJ/VZ2uxhQ0H5K2aH7KzCAdFQ4zqui5SD7dl+mQeQni07EmoClFeJRwF19Vdqv89I/ZfRRa0Poejan0UGkxbVbK0IXpdCa30mOCqv1LefK4RAni7jVFQzfMJ1wzIuErntSZ/buYiNFebCGSjBn+9KFI= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1731892437; c=relaxed/simple; bh=iT8ZUSEM7IlxGZyL8663ba3PW85q+uAow0c5s95HLEk=; h=From:To:Cc:Subject:Date:Message-Id:In-Reply-To:References: MIME-Version; b=twHxpkGUJUx8GqtVP6HqD0eTkt5AtcCqk2dsZO4qdQYF/1pm0beiB1EKSHqfVkq+J5hknQMjRp5nJOJ6oWkuCCD5DZzawdlaNzQG+NShmLp56e7WheSMgIn2IwMHS9I9VY/qzsB5/+NEXa5kmPCsx/nFpOF2u11JxeWhxVlJfCM= ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dmarc=none (p=none dis=none) header.from=huaweicloud.com; spf=pass smtp.mailfrom=huaweicloud.com; arc=none smtp.client-ip=45.249.212.51 Authentication-Results: smtp.subspace.kernel.org; dmarc=none (p=none dis=none) header.from=huaweicloud.com Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=huaweicloud.com Received: from mail.maildlp.com (unknown [172.19.163.235]) by dggsgout11.his.huawei.com (SkyGuard) with ESMTP id 4Xs8Lk1kt2z4f3lW5 for ; Mon, 18 Nov 2024 08:55:42 +0800 (CST) Received: from mail02.huawei.com (unknown [10.116.40.128]) by mail.maildlp.com (Postfix) with ESMTP id 8D1851A0568 for ; Mon, 18 Nov 2024 08:56:01 +0800 (CST) Received: from huaweicloud.com (unknown [10.175.124.27]) by APP4 (Coremail) with SMTP id gCh0CgCXc4eckDpn2G5fCA--.44635S8; Mon, 18 Nov 2024 08:56:01 +0800 (CST) From: Hou Tao To: bpf@vger.kernel.org Cc: Martin KaFai Lau , Alexei Starovoitov , Andrii Nakryiko , Eduard Zingerman , Song Liu , Hao Luo , Yonghong Song , Daniel Borkmann , KP Singh , Stanislav Fomichev , Jiri Olsa , John Fastabend , Sebastian Andrzej Siewior , Thomas Gleixner , =?utf-8?q?Thomas_Wei=C3=9Fschuh?= , houtao1@huawei.com, xukuohai@huawei.com Subject: [PATCH bpf-next 04/10] bpf: Handle in-place update for full LPM trie correctly Date: Mon, 18 Nov 2024 09:08:02 +0800 Message-Id: <20241118010808.2243555-5-houtao@huaweicloud.com> X-Mailer: git-send-email 2.29.2 In-Reply-To: <20241118010808.2243555-1-houtao@huaweicloud.com> References: <20241118010808.2243555-1-houtao@huaweicloud.com> Precedence: bulk X-Mailing-List: bpf@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 X-CM-TRANSID: gCh0CgCXc4eckDpn2G5fCA--.44635S8 X-Coremail-Antispam: 1UD129KBjvJXoWxCrWxKF1DCFWfXr48AF1DWrg_yoW5tryfpF WSkF90yr18Jr4UWr4Fva98JrZ8Zw4rG3yj9F93W347AFySkr93tF18uayFyrW5ArWUXrWY yF45tFyqgrZ5ZrJanT9S1TB71UUUUU7qnTZGkaVYY2UrUUUUjbIjqfuFe4nvWSU5nxnvy2 9KBjDU0xBIdaVrnRJUUUPSb4IE77IF4wAFF20E14v26rWj6s0DM7CY07I20VC2zVCF04k2 6cxKx2IYs7xG6rWj6s0DM7CIcVAFz4kK6r1j6r18M28IrcIa0xkI8VA2jI8067AKxVWUAV Cq3wA2048vs2IY020Ec7CjxVAFwI0_Xr0E3s1l8cAvFVAK0II2c7xJM28CjxkF64kEwVA0 rcxSw2x7M28EF7xvwVC0I7IYx2IY67AKxVW7JVWDJwA2z4x0Y4vE2Ix0cI8IcVCY1x0267 AKxVW8Jr0_Cr1UM28EF7xvwVC2z280aVAFwI0_GcCE3s1l84ACjcxK6I8E87Iv6xkF7I0E 14v26rxl6s0DM2AIxVAIcxkEcVAq07x20xvEncxIr21l5I8CrVACY4xI64kE6c02F40Ex7 xfMcIj6xIIjxv20xvE14v26r1j6r18McIj6I8E87Iv67AKxVWUJVW8JwAm72CE4IkC6x0Y z7v_Jr0_Gr1lF7xvr2IYc2Ij64vIr41lFIxGxcIEc7CjxVA2Y2ka0xkIwI1lc7CjxVAaw2 AFwI0_GFv_Wryl42xK82IYc2Ij64vIr41l4I8I3I0E4IkC6x0Yz7v_Jr0_Gr1lx2IqxVAq x4xG67AKxVWUJVWUGwC20s026x8GjcxK67AKxVWUGVWUWwC2zVAF1VAY17CE14v26r4a6r W5MIIYrxkI7VAKI48JMIIF0xvE2Ix0cI8IcVAFwI0_Jr0_JF4lIxAIcVC0I7IYx2IY6xkF 7I0E14v26F4j6r4UJwCI42IY6xAIw20EY4v20xvaj40_Jr0_JF4lIxAIcVC2z280aVAFwI 0_Jr0_Gr1lIxAIcVC2z280aVCY1x0267AKxVW8JVW8JrUvcSsGvfC2KfnxnUUI43ZEXa7I U0sqXPUUUUU== X-CM-SenderInfo: xkrx3t3r6k3tpzhluzxrxghudrp/ X-Patchwork-Delegate: bpf@iogearbox.net From: Hou Tao When a LPM trie is full, in-place updates of existing elements incorrectly return -ENOSPC. Fix this by deferring the check of trie->n_entries. For new insertions, n_entries must not exceed max_entries. However, in-place updates are allowed even when the trie is full. Signed-off-by: Hou Tao --- kernel/bpf/lpm_trie.c | 46 +++++++++++++++++++++---------------------- 1 file changed, 23 insertions(+), 23 deletions(-) diff --git a/kernel/bpf/lpm_trie.c b/kernel/bpf/lpm_trie.c index 4300bd51ec6e..ff57e35357ae 100644 --- a/kernel/bpf/lpm_trie.c +++ b/kernel/bpf/lpm_trie.c @@ -310,6 +310,15 @@ static struct lpm_trie_node *lpm_trie_node_alloc(const struct lpm_trie *trie, return node; } +static int trie_check_noreplace_update(const struct lpm_trie *trie, u64 flags) +{ + if (flags == BPF_EXIST) + return -ENOENT; + if (trie->n_entries == trie->map.max_entries) + return -ENOSPC; + return 0; +} + /* Called from syscall or from eBPF program */ static long trie_update_elem(struct bpf_map *map, void *_key, void *value, u64 flags) @@ -333,20 +342,12 @@ static long trie_update_elem(struct bpf_map *map, spin_lock_irqsave(&trie->lock, irq_flags); /* Allocate and fill a new node */ - - if (trie->n_entries == trie->map.max_entries) { - ret = -ENOSPC; - goto out; - } - new_node = lpm_trie_node_alloc(trie, value); if (!new_node) { ret = -ENOMEM; goto out; } - trie->n_entries++; - new_node->prefixlen = key->prefixlen; RCU_INIT_POINTER(new_node->child[0], NULL); RCU_INIT_POINTER(new_node->child[1], NULL); @@ -375,10 +376,11 @@ static long trie_update_elem(struct bpf_map *map, * simply assign the @new_node to that slot and be done. */ if (!node) { - if (flags == BPF_EXIST) { - ret = -ENOENT; + ret = trie_check_noreplace_update(trie, flags); + if (ret) goto out; - } + + trie->n_entries++; rcu_assign_pointer(*slot, new_node); goto out; } @@ -392,10 +394,11 @@ static long trie_update_elem(struct bpf_map *map, ret = -EEXIST; goto out; } - trie->n_entries--; - } else if (flags == BPF_EXIST) { - ret = -ENOENT; - goto out; + } else { + ret = trie_check_noreplace_update(trie, flags); + if (ret) + goto out; + trie->n_entries++; } new_node->child[0] = node->child[0]; @@ -407,15 +410,15 @@ static long trie_update_elem(struct bpf_map *map, goto out; } - if (flags == BPF_EXIST) { - ret = -ENOENT; + ret = trie_check_noreplace_update(trie, flags); + if (ret) goto out; - } /* If the new node matches the prefix completely, it must be inserted * as an ancestor. Simply insert it between @node and *@slot. */ if (matchlen == key->prefixlen) { + trie->n_entries++; next_bit = extract_bit(node->data, matchlen); rcu_assign_pointer(new_node->child[next_bit], node); rcu_assign_pointer(*slot, new_node); @@ -428,6 +431,7 @@ static long trie_update_elem(struct bpf_map *map, goto out; } + trie->n_entries++; im_node->prefixlen = matchlen; im_node->flags |= LPM_TREE_NODE_FLAG_IM; memcpy(im_node->data, node->data, trie->data_size); @@ -445,12 +449,8 @@ static long trie_update_elem(struct bpf_map *map, rcu_assign_pointer(*slot, im_node); out: - if (ret) { - if (new_node) - trie->n_entries--; + if (ret) kfree(new_node); - } - spin_unlock_irqrestore(&trie->lock, irq_flags); kfree_rcu(free_node, rcu);