From patchwork Mon Mar 11 21:00:48 2019 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Thomas Richard via openwrt-devel X-Patchwork-Id: 1054879 Return-Path: X-Original-To: incoming@patchwork.ozlabs.org Delivered-To: patchwork-incoming@bilbo.ozlabs.org Authentication-Results: ozlabs.org; spf=pass (mailfrom) smtp.mailfrom=lists.openwrt.org (client-ip=2607:7c80:54:e::133; helo=bombadil.infradead.org; envelope-from=openwrt-devel-bounces+incoming=patchwork.ozlabs.org@lists.openwrt.org; receiver=) Authentication-Results: ozlabs.org; dmarc=none (p=none dis=none) header.from=lists.openwrt.org Authentication-Results: ozlabs.org; dkim=pass (2048-bit key; unprotected) header.d=lists.infradead.org header.i=@lists.infradead.org header.b="Fwmtr/E8"; dkim-atps=neutral Received: from bombadil.infradead.org (bombadil.infradead.org [IPv6:2607:7c80:54:e::133]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by ozlabs.org (Postfix) with ESMTPS id 44J9W145wBz9s3l for ; Tue, 12 Mar 2019 08:01:01 +1100 (AEDT) DKIM-Signature: v=1; a=rsa-sha256; q=dns/txt; c=relaxed/relaxed; d=lists.infradead.org; s=bombadil.20170209; h=Date:Sender:Content-Type: Subject:List-Help:Reply-To:List-Archive:List-Unsubscribe:List-Subscribe:Cc: From:List-Post:List-Id:Message-ID:MIME-Version:References:In-Reply-To:To: Content-Transfer-Encoding:Content-ID:Content-Description:Resent-Date: Resent-From:Resent-Sender:Resent-To:Resent-Cc:Resent-Message-ID:List-Owner; bh=5c8QjhhxTgE1QOuovMlFUpxCMhnVh8dezs7tm9iK1SU=; b=Fwmtr/E83oeWOCQWP8MBW0PJZ +LXp8e1IUa2djM1aVtiS2MMu22vxyLrV0SwTKnFSXk2yiKRqFuhxcslVQ5ru4fho/i8zi0XrtghUt QHaczB76OufG/XAEzHxn/tw4pm/rsEpS2HtrJJPIXWp1J4luHJmeUnZ6NIQ5UODUbpC0D3Pp7QNi3 AFPZjpZeoKezM8wSlvPFtxJ0AwtWhVpWyIjRJXn7VOhSuxvkCbIzqTZu5nLiRkn2oV48yNNWMP5kx 0j2/pCRwC5B9SJ7YOUkmOU1d2Danff8SfHKpdGixBiSzqn6Sjk9tQfE519jYcUTeA8GPrnbfeeH8K S8lSkQvbA==; Received: from localhost ([127.0.0.1] helo=bombadil.infradead.org) by bombadil.infradead.org with esmtp (Exim 4.90_1 #2 (Red Hat Linux)) id 1h3S2S-0003sp-6a; Mon, 11 Mar 2019 21:00:48 +0000 To: openwrt-devel@lists.openwrt.org In-Reply-To: <20190311210017.15831-1-cote2004-github@yahoo.com> References: <20190311210017.15831-1-cote2004-github@yahoo.com> MIME-Version: 1.0 Message-ID: List-Id: List-Post: X-Patchwork-Original-From: Eneas U de Queiroz via openwrt-devel From: Thomas Richard via openwrt-devel Precedence: list Cc: Eneas U de Queiroz X-Mailman-Version: 2.1.21 X-BeenThere: openwrt-devel@lists.openwrt.org List-Subscribe: , List-Unsubscribe: , List-Archive: Reply-To: Eneas U de Queiroz List-Help: Subject: [OpenWrt-Devel] [PATCH] openssl: disable digests by default, misc fixes Sender: "openwrt-devel" Date: Mon, 11 Mar 2019 21:00:48 +0000 Errors-To: openwrt-devel-bounces+incoming=patchwork.ozlabs.org@lists.openwrt.org The sender domain has a DMARC Reject/Quarantine policy which disallows sending mailing list messages using the original "From" header. To mitigate this problem, the original message has been wrapped automatically by the mailing list software. Openssh uses digest contexts across forks, which is not supported by the /dev/crypto engine. The speed of digests is usually not worth enabling them anyway. This changes the default of the DIGESTS option to NONE, so the user still has the option to enable them. Added another patch related to the use of encryption contexts across forks, that ignores a failure to close a previous open session when reinitializing a context, instead of failing the reinitialization. Added a link to the Cryptographic Hardware Accelerators document to the engine pacakges description, to provide more detailed instructions to configure the engines. Signed-off-by: Eneas U de Queiroz diff --git a/package/libs/openssl/Config.in b/package/libs/openssl/Config.in index 235f38e787..72ff64634f 100644 --- a/package/libs/openssl/Config.in +++ b/package/libs/openssl/Config.in @@ -269,7 +269,7 @@ config OPENSSL_ENGINE_BUILTIN_AFALG select PACKAGE_libopenssl-conf help This enables use of hardware acceleration through the - AF_ALG kenrel interface. + AF_ALG kernel interface. config OPENSSL_ENGINE_BUILTIN_DEVCRYPTO bool @@ -279,6 +279,9 @@ config OPENSSL_ENGINE_BUILTIN_DEVCRYPTO help This enables use of hardware acceleration through OpenBSD Cryptodev API (/dev/crypto) interface. + Even though configuraion is not strictly needed, it is worth seeing + https://openwrt.org/docs/techref/hardware/cryptographic.hardware.accelerators + for information on how to configure the engine. config OPENSSL_ENGINE_BUILTIN_PADLOCK bool diff --git a/package/libs/openssl/Makefile b/package/libs/openssl/Makefile index ef840e28ad..cb25c5557c 100644 --- a/package/libs/openssl/Makefile +++ b/package/libs/openssl/Makefile @@ -11,7 +11,7 @@ PKG_NAME:=openssl PKG_BASE:=1.1.1 PKG_BUGFIX:=b PKG_VERSION:=$(PKG_BASE)$(PKG_BUGFIX) -PKG_RELEASE:=2 +PKG_RELEASE:=3 PKG_USE_MIPS16:=0 ENGINES_DIR=engines-1.1 @@ -147,6 +147,7 @@ This package adds an engine that enables hardware acceleration through the AF_ALG kernel interface. To use it, you need to configure the engine in /etc/ssl/openssl.cnf See https://www.openssl.org/docs/man1.1.1/man5/config.html#Engine-Configuration-Module +and https://openwrt.org/docs/techref/hardware/cryptographic.hardware.accelerators The engine_id is "afalg" endef @@ -163,6 +164,7 @@ This package adds an engine that enables hardware acceleration through the /dev/crypto kernel interface. To use it, you need to configure the engine in /etc/ssl/openssl.cnf See https://www.openssl.org/docs/man1.1.1/man5/config.html#Engine-Configuration-Module +and https://openwrt.org/docs/techref/hardware/cryptographic.hardware.accelerators The engine_id is "devcrypto" endef @@ -178,6 +180,7 @@ define Package/libopenssl-padlock/description This package adds an engine that enables VIA Padlock hardware acceleration. To use it, you need to configure it in /etc/ssl/openssl.cnf. See https://www.openssl.org/docs/man1.1.1/man5/config.html#Engine-Configuration-Module +and https://openwrt.org/docs/techref/hardware/cryptographic.hardware.accelerators The engine_id is "padlock" endef diff --git a/package/libs/openssl/patches/500-e_devcrypto-default-to-not-use-digests-in-engine.patch b/package/libs/openssl/patches/500-e_devcrypto-default-to-not-use-digests-in-engine.patch new file mode 100644 index 0000000000..52bc8d2118 --- /dev/null +++ b/package/libs/openssl/patches/500-e_devcrypto-default-to-not-use-digests-in-engine.patch @@ -0,0 +1,41 @@ +From 5d3be6bc8ed7d73ab2c4d389fb0f0a03dacd04b1 Mon Sep 17 00:00:00 2001 +From: Eneas U de Queiroz +Date: Mon, 11 Mar 2019 09:29:13 -0300 +Subject: [PATCH] e_devcrypto: default to not use digests in engine + +Digests are almost always slower when using /dev/crypto because of the +cost of the context switches. Only for large blocks it is worth it. + +Also, when forking, the open context structures are duplicated, but the +internal kernel sessions are still shared between forks, which means an +update/close operation in one fork affects all processes using that +session. + +This affects digests, especially for HMAC, where the session with the +key hash is used as a source for subsequent operations. At least one +popular application does this across a fork. Disabling digests by +default will mitigate the problem, while still allowing the user to +turn them on if it is safe and fast enough. + +Signed-off-by: Eneas U de Queiroz + +--- a/engines/e_devcrypto.c ++++ b/engines/e_devcrypto.c +@@ -854,7 +854,7 @@ static void prepare_digest_methods(void) + for (i = 0, known_digest_nids_amount = 0; i < OSSL_NELEM(digest_data); + i++) { + +- selected_digests[i] = 1; ++ selected_digests[i] = 0; + + /* + * Check that the digest is usable +@@ -1068,7 +1068,7 @@ static const ENGINE_CMD_DEFN devcrypto_cmds[] = { + #ifdef IMPLEMENT_DIGEST + {DEVCRYPTO_CMD_DIGESTS, + "DIGESTS", +- "either ALL, NONE, or a comma-separated list of digests to enable [default=ALL]", ++ "either ALL, NONE, or a comma-separated list of digests to enable [default=NONE]", + ENGINE_CMD_FLAG_STRING}, + #endif + diff --git a/package/libs/openssl/patches/510-e_devcrypto-ignore-error-when-closing-session.patch b/package/libs/openssl/patches/510-e_devcrypto-ignore-error-when-closing-session.patch new file mode 100644 index 0000000000..160c23df30 --- /dev/null +++ b/package/libs/openssl/patches/510-e_devcrypto-ignore-error-when-closing-session.patch @@ -0,0 +1,24 @@ +From b6e6d157367bae91a8015434769572e430257d40 Mon Sep 17 00:00:00 2001 +From: Eneas U de Queiroz +Date: Mon, 11 Mar 2019 10:15:14 -0300 +Subject: [PATCH] e_devcrypto: ignore error when closing session + +In cipher_init, ignore an eventual error when closing the previous +session. It may have been closed by another process after a fork. + +Signed-off-by: Eneas U de Queiroz + +--- a/engines/e_devcrypto.c ++++ b/engines/e_devcrypto.c +@@ -197,9 +197,8 @@ static int cipher_init(EVP_CIPHER_CTX *ctx, const unsigned char *key, + get_cipher_data(EVP_CIPHER_CTX_nid(ctx)); + + /* cleanup a previous session */ +- if (cipher_ctx->sess.ses != 0 && +- clean_devcrypto_session(&cipher_ctx->sess) == 0) +- return 0; ++ if (cipher_ctx->sess.ses != 0) ++ clean_devcrypto_session(&cipher_ctx->sess); + + cipher_ctx->sess.cipher = cipher_d->devcryptoid; + cipher_ctx->sess.keylen = cipher_d->keylen;