From patchwork Thu Jun 25 12:42:42 2020 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Ard Biesheuvel X-Patchwork-Id: 11625089 Return-Path: Received: from mail.kernel.org (pdx-korg-mail-1.web.codeaurora.org [172.30.200.123]) by pdx-korg-patchwork-2.web.codeaurora.org (Postfix) with ESMTP id EDA8514E3 for ; Thu, 25 Jun 2020 12:43:56 +0000 (UTC) Received: from merlin.infradead.org (merlin.infradead.org [205.233.59.134]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by mail.kernel.org (Postfix) with ESMTPS id C68552072E for ; Thu, 25 Jun 2020 12:43:56 +0000 (UTC) Authentication-Results: mail.kernel.org; dkim=pass (2048-bit key) header.d=lists.infradead.org header.i=@lists.infradead.org header.b="KvLaxWxw"; dkim=fail reason="signature verification failed" (1024-bit key) header.d=kernel.org header.i=@kernel.org header.b="XHnxyn+u" DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org C68552072E Authentication-Results: mail.kernel.org; dmarc=fail (p=none dis=none) header.from=kernel.org Authentication-Results: mail.kernel.org; spf=none smtp.mailfrom=linux-amlogic-bounces+patchwork-linux-amlogic=patchwork.kernel.org@lists.infradead.org DKIM-Signature: v=1; a=rsa-sha256; q=dns/txt; c=relaxed/relaxed; d=lists.infradead.org; s=merlin.20170209; h=Sender:Content-Transfer-Encoding: Content-Type:Cc:List-Subscribe:List-Help:List-Post:List-Archive: List-Unsubscribe:List-Id:MIME-Version:References:In-Reply-To:Message-Id:Date: Subject:To:From:Reply-To:Content-ID:Content-Description:Resent-Date: Resent-From:Resent-Sender:Resent-To:Resent-Cc:Resent-Message-ID:List-Owner; bh=E+ALKjxtOwHF4723khZRItSy+KQ06ePsAn1oH0AoKZ0=; b=KvLaxWxwZW+cz38/4iXKmHdVN OzIL7tiQ7Njabay7DFIBxrGt5O9nZrcKnpkavZM3l31bMCkIoOoIfws+HTuH61MLanLy0kq9TgVzi FGpbPbkqQsX301Ae9VtFdnSCZdpJ8CEwZqgROYyl+2gp1/gJPbztPXRQqLsQ+iL2FY3dpnv0151dj KQUqrPFe0paN51epozVyUnqjVwkWNEVtFxkHhccPG8tx2AkofAuZ/Ez6J+RkOISJcuWWs5buKRb1o 2xVlL4f6pxWWCM6Ac0Xg39DJ+LpkoTF8sKZ+Qwr2zOJxWCBRcWSqAvVltRv5HoDWX0nMFg9pFeRoY 3Ag3HN2/Q==; Received: from localhost ([::1] helo=merlin.infradead.org) by merlin.infradead.org with esmtp (Exim 4.92.3 #3 (Red Hat Linux)) id 1joRE0-0005LZ-1d; Thu, 25 Jun 2020 12:43:28 +0000 Received: from mail.kernel.org ([198.145.29.99]) by merlin.infradead.org with esmtps (Exim 4.92.3 #3 (Red Hat Linux)) id 1joRDv-0005Kk-P1; Thu, 25 Jun 2020 12:43:24 +0000 Received: from localhost.localdomain (lfbn-nic-1-188-42.w2-15.abo.wanadoo.fr [2.15.37.42]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by mail.kernel.org (Postfix) with ESMTPSA id 9368A207E8; Thu, 25 Jun 2020 12:43:19 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=kernel.org; s=default; t=1593089003; bh=c+inIYObGSZUVYp7oF5MvXqUUSrjOBVfIrZv9HY3hO8=; h=From:To:Cc:Subject:Date:In-Reply-To:References:From; b=XHnxyn+umL7xjsWbcPaEEpz9fmEaU36nRZlbaYhyeyHyJ7svn/xv3pTucjhOHQT/U W+0kAGtDwk48a+grcEYWi2/HaF3cgvOlAoPQtbXsUYqHoQzI1yQ8pRlCLjoSgWsjq+ SdGtX6T3Kq7hMNPi6uenO6mm5dmnvaEpgDvgDbdw= From: Ard Biesheuvel To: linux-crypto@vger.kernel.org Subject: [PATCH 01/12] crypto: amlogic-gxl - default to build as module Date: Thu, 25 Jun 2020 14:42:42 +0200 Message-Id: <20200625124253.1906557-2-ardb@kernel.org> X-Mailer: git-send-email 2.27.0 In-Reply-To: <20200625124253.1906557-1-ardb@kernel.org> References: <20200625124253.1906557-1-ardb@kernel.org> MIME-Version: 1.0 X-Spam-Note: CRM114 invocation failed X-Spam-Score: -5.2 (-----) X-Spam-Report: SpamAssassin version 3.4.4 on merlin.infradead.org summary: Content analysis details: (-5.2 points) pts rule name description ---- ---------------------- -------------------------------------------------- -5.0 RCVD_IN_DNSWL_HI RBL: Sender listed at https://www.dnswl.org/, high trust [198.145.29.99 listed in list.dnswl.org] 0.0 SPF_HELO_NONE SPF: HELO does not publish an SPF Record -0.0 SPF_PASS SPF: sender matches SPF record -0.1 DKIM_VALID Message has at least one valid DKIM or DK signature -0.1 DKIM_VALID_EF Message has a valid DKIM or DK signature from envelope-from domain -0.1 DKIM_VALID_AU Message has a valid DKIM or DK signature from author's domain 0.1 DKIM_SIGNED Message has a DKIM or DK signature, not necessarily valid -0.0 DKIMWL_WL_HIGH DKIMwl.org - Whitelisted High sender X-BeenThere: linux-amlogic@lists.infradead.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Cc: Eric Biggers , Tom Lendacky , Jamie Iles , Herbert Xu , Rohit Maheshwari , Fabio Estevam , Sascha Hauer , Vinay Kumar Yadav , Maxime Ripard , "David S. Miller" , Chen-Yu Tsai , Corentin Labbe , NXP Linux Team , Pengutronix Kernel Team , Ayush Sawal , linux-amlogic@lists.infradead.org, linux-omap@vger.kernel.org, Shawn Guo , Ard Biesheuvel , linux-arm-kernel@lists.infradead.org Sender: "linux-amlogic" Errors-To: linux-amlogic-bounces+patchwork-linux-amlogic=patchwork.kernel.org@lists.infradead.org The AmLogic GXL crypto accelerator driver is built into the kernel if ARCH_MESON is set. However, given the single image policy of arm64, its defconfig enables all platforms by default, and so ARCH_MESON is usually enabled. This means that the AmLogic driver causes the arm64 defconfig build to pull in a huge chunk of the crypto stack as a builtin as well, which is undesirable, so let's make the amlogic GXL driver default to 'm' instead. Signed-off-by: Ard Biesheuvel --- drivers/crypto/amlogic/Kconfig | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/crypto/amlogic/Kconfig b/drivers/crypto/amlogic/Kconfig index cf9547602670..cf2c676a7093 100644 --- a/drivers/crypto/amlogic/Kconfig +++ b/drivers/crypto/amlogic/Kconfig @@ -1,7 +1,7 @@ config CRYPTO_DEV_AMLOGIC_GXL tristate "Support for amlogic cryptographic offloader" depends on HAS_IOMEM - default y if ARCH_MESON + default m if ARCH_MESON select CRYPTO_SKCIPHER select CRYPTO_ENGINE select CRYPTO_ECB From patchwork Thu Jun 25 12:42:43 2020 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Ard Biesheuvel X-Patchwork-Id: 11625087 Return-Path: Received: from mail.kernel.org (pdx-korg-mail-1.web.codeaurora.org [172.30.200.123]) by pdx-korg-patchwork-2.web.codeaurora.org (Postfix) with ESMTP id EA6A16C1 for ; Thu, 25 Jun 2020 12:43:56 +0000 (UTC) Received: from merlin.infradead.org (merlin.infradead.org [205.233.59.134]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by mail.kernel.org (Postfix) with ESMTPS id C3CC220720 for ; Thu, 25 Jun 2020 12:43:56 +0000 (UTC) Authentication-Results: mail.kernel.org; dkim=pass (2048-bit key) header.d=lists.infradead.org header.i=@lists.infradead.org header.b="hUXYclXF"; dkim=fail reason="signature verification failed" (1024-bit key) header.d=kernel.org header.i=@kernel.org header.b="Mdi3f4DJ" DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org C3CC220720 Authentication-Results: mail.kernel.org; dmarc=fail (p=none dis=none) header.from=kernel.org Authentication-Results: mail.kernel.org; spf=none smtp.mailfrom=linux-amlogic-bounces+patchwork-linux-amlogic=patchwork.kernel.org@lists.infradead.org DKIM-Signature: v=1; a=rsa-sha256; q=dns/txt; c=relaxed/relaxed; d=lists.infradead.org; s=merlin.20170209; h=Sender:Content-Transfer-Encoding: Content-Type:Cc:List-Subscribe:List-Help:List-Post:List-Archive: List-Unsubscribe:List-Id:MIME-Version:References:In-Reply-To:Message-Id:Date: Subject:To:From:Reply-To:Content-ID:Content-Description:Resent-Date: Resent-From:Resent-Sender:Resent-To:Resent-Cc:Resent-Message-ID:List-Owner; bh=vi8ZUNsr0bMLFkgRRnmrhCShkArfJ1KVhQaGLTWYP64=; b=hUXYclXFyy3FHHiJ9Cu1NWzf+ VYmooz/UcjOCtQPqCIaqK7jhgSQiV3a53azuWuMooBSMdP1QK6bVdfc4jkR8Xky08oDlcL1ckitc0 ycI5V3WEhMcym+BFeLCbHYK+XKw9NWRScw5k9OqA8qhcgX3nYYg8iXpWxYo7X1fJAv/z+BAQ0EVmH dD1AxEqn9FZSlxXZJcL8R/p+nBuzf7+BKySbjJlVtEMZM1DTb1LNeIUrSdS2VTQvXdsB0v02i+LhD gXl/D98mqLbTj/QbCxs4a/g5zLN5Jt5WXJgqugB/8cZtEJ+I/WrvkkgT4YF+ZvvA1NYfnrsw8VTRl +8MVBHNFQ==; Received: from localhost ([::1] helo=merlin.infradead.org) by merlin.infradead.org with esmtp (Exim 4.92.3 #3 (Red Hat Linux)) id 1joRE2-0005MN-KU; Thu, 25 Jun 2020 12:43:30 +0000 Received: from mail.kernel.org ([198.145.29.99]) by merlin.infradead.org with esmtps (Exim 4.92.3 #3 (Red Hat Linux)) id 1joRDz-0005LU-RS; Thu, 25 Jun 2020 12:43:29 +0000 Received: from localhost.localdomain (lfbn-nic-1-188-42.w2-15.abo.wanadoo.fr [2.15.37.42]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by mail.kernel.org (Postfix) with ESMTPSA id 9B7BF20720; Thu, 25 Jun 2020 12:43:23 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=kernel.org; s=default; t=1593089007; bh=DOov/VzevH5sYOijwCFcdKlJxFMfsKQgVxpnIqtZ1KY=; h=From:To:Cc:Subject:Date:In-Reply-To:References:From; b=Mdi3f4DJwlCShMLNjVw/YISKiTF/bsVsClqi46r889TBYamhsHsUV5ZRijSdXc2KA iyLZyrlP3Z2H+NCyCL+XBbiQj1JZ+leALX5PLS8QNefnNwAh/v2kPpaB3dGSEfvEAC aHng0adVku9Rub+q43ZoNl6258gR5mLd7TtdPS+k= From: Ard Biesheuvel To: linux-crypto@vger.kernel.org Subject: [PATCH 02/12] crypto: amlogic-gxl - permit async skcipher as fallback Date: Thu, 25 Jun 2020 14:42:43 +0200 Message-Id: <20200625124253.1906557-3-ardb@kernel.org> X-Mailer: git-send-email 2.27.0 In-Reply-To: <20200625124253.1906557-1-ardb@kernel.org> References: <20200625124253.1906557-1-ardb@kernel.org> MIME-Version: 1.0 X-Spam-Note: CRM114 invocation failed X-Spam-Score: -5.2 (-----) X-Spam-Report: SpamAssassin version 3.4.4 on merlin.infradead.org summary: Content analysis details: (-5.2 points) pts rule name description ---- ---------------------- -------------------------------------------------- -5.0 RCVD_IN_DNSWL_HI RBL: Sender listed at https://www.dnswl.org/, high trust [198.145.29.99 listed in list.dnswl.org] 0.0 SPF_HELO_NONE SPF: HELO does not publish an SPF Record -0.0 SPF_PASS SPF: sender matches SPF record -0.1 DKIM_VALID Message has at least one valid DKIM or DK signature -0.1 DKIM_VALID_EF Message has a valid DKIM or DK signature from envelope-from domain -0.1 DKIM_VALID_AU Message has a valid DKIM or DK signature from author's domain 0.1 DKIM_SIGNED Message has a DKIM or DK signature, not necessarily valid -0.0 DKIMWL_WL_HIGH DKIMwl.org - Whitelisted High sender X-BeenThere: linux-amlogic@lists.infradead.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Cc: Eric Biggers , Tom Lendacky , Jamie Iles , Herbert Xu , Rohit Maheshwari , Fabio Estevam , Sascha Hauer , Vinay Kumar Yadav , Maxime Ripard , "David S. Miller" , Chen-Yu Tsai , Corentin Labbe , NXP Linux Team , Pengutronix Kernel Team , Ayush Sawal , linux-amlogic@lists.infradead.org, linux-omap@vger.kernel.org, Shawn Guo , Ard Biesheuvel , linux-arm-kernel@lists.infradead.org Sender: "linux-amlogic" Errors-To: linux-amlogic-bounces+patchwork-linux-amlogic=patchwork.kernel.org@lists.infradead.org Even though the amlogic-gxl driver implements asynchronous versions of ecb(aes) and cbc(aes), the fallbacks it allocates are required to be synchronous. Given that SIMD based software implementations are usually asynchronous as well, even though they rarely complete asynchronously (this typically only happens in cases where the request was made from softirq context, while SIMD was already in use in the task context that it interrupted), these implementations are disregarded, and either the generic C version or another table based version implemented in assembler is selected instead. Since falling back to synchronous AES is not only a performance issue, but potentially a security issue as well (due to the fact that table based AES is not time invariant), let's fix this, by allocating an ordinary skcipher as the fallback, and invoke it with the completion routine that was given to the outer request. Signed-off-by: Ard Biesheuvel --- drivers/crypto/amlogic/amlogic-gxl-cipher.c | 27 ++++++++++---------- drivers/crypto/amlogic/amlogic-gxl.h | 3 ++- 2 files changed, 15 insertions(+), 15 deletions(-) diff --git a/drivers/crypto/amlogic/amlogic-gxl-cipher.c b/drivers/crypto/amlogic/amlogic-gxl-cipher.c index 9819dd50fbad..5880b94dcb32 100644 --- a/drivers/crypto/amlogic/amlogic-gxl-cipher.c +++ b/drivers/crypto/amlogic/amlogic-gxl-cipher.c @@ -64,22 +64,20 @@ static int meson_cipher_do_fallback(struct skcipher_request *areq) #ifdef CONFIG_CRYPTO_DEV_AMLOGIC_GXL_DEBUG struct skcipher_alg *alg = crypto_skcipher_alg(tfm); struct meson_alg_template *algt; -#endif - SYNC_SKCIPHER_REQUEST_ON_STACK(req, op->fallback_tfm); -#ifdef CONFIG_CRYPTO_DEV_AMLOGIC_GXL_DEBUG algt = container_of(alg, struct meson_alg_template, alg.skcipher); algt->stat_fb++; #endif - skcipher_request_set_sync_tfm(req, op->fallback_tfm); - skcipher_request_set_callback(req, areq->base.flags, NULL, NULL); - skcipher_request_set_crypt(req, areq->src, areq->dst, + skcipher_request_set_tfm(&rctx->fallback_req, op->fallback_tfm); + skcipher_request_set_callback(&rctx->fallback_req, areq->base.flags, + areq->base.complete, areq->base.data); + skcipher_request_set_crypt(&rctx->fallback_req, areq->src, areq->dst, areq->cryptlen, areq->iv); + if (rctx->op_dir == MESON_DECRYPT) - err = crypto_skcipher_decrypt(req); + err = crypto_skcipher_decrypt(&rctx->fallback_req); else - err = crypto_skcipher_encrypt(req); - skcipher_request_zero(req); + err = crypto_skcipher_encrypt(&rctx->fallback_req); return err; } @@ -321,15 +319,16 @@ int meson_cipher_init(struct crypto_tfm *tfm) algt = container_of(alg, struct meson_alg_template, alg.skcipher); op->mc = algt->mc; - sktfm->reqsize = sizeof(struct meson_cipher_req_ctx); - - op->fallback_tfm = crypto_alloc_sync_skcipher(name, 0, CRYPTO_ALG_NEED_FALLBACK); + op->fallback_tfm = crypto_alloc_skcipher(name, 0, CRYPTO_ALG_NEED_FALLBACK); if (IS_ERR(op->fallback_tfm)) { dev_err(op->mc->dev, "ERROR: Cannot allocate fallback for %s %ld\n", name, PTR_ERR(op->fallback_tfm)); return PTR_ERR(op->fallback_tfm); } + sktfm->reqsize = sizeof(struct meson_cipher_req_ctx) + + crypto_skcipher_reqsize(op->fallback_tfm); + op->enginectx.op.do_one_request = meson_handle_cipher_request; op->enginectx.op.prepare_request = NULL; op->enginectx.op.unprepare_request = NULL; @@ -345,7 +344,7 @@ void meson_cipher_exit(struct crypto_tfm *tfm) memzero_explicit(op->key, op->keylen); kfree(op->key); } - crypto_free_sync_skcipher(op->fallback_tfm); + crypto_free_skcipher(op->fallback_tfm); } int meson_aes_setkey(struct crypto_skcipher *tfm, const u8 *key, @@ -377,5 +376,5 @@ int meson_aes_setkey(struct crypto_skcipher *tfm, const u8 *key, if (!op->key) return -ENOMEM; - return crypto_sync_skcipher_setkey(op->fallback_tfm, key, keylen); + return crypto_skcipher_setkey(op->fallback_tfm, key, keylen); } diff --git a/drivers/crypto/amlogic/amlogic-gxl.h b/drivers/crypto/amlogic/amlogic-gxl.h index b7f2de91ab76..dc0f142324a3 100644 --- a/drivers/crypto/amlogic/amlogic-gxl.h +++ b/drivers/crypto/amlogic/amlogic-gxl.h @@ -109,6 +109,7 @@ struct meson_dev { struct meson_cipher_req_ctx { u32 op_dir; int flow; + struct skcipher_request fallback_req; // keep at the end }; /* @@ -126,7 +127,7 @@ struct meson_cipher_tfm_ctx { u32 keylen; u32 keymode; struct meson_dev *mc; - struct crypto_sync_skcipher *fallback_tfm; + struct crypto_skcipher *fallback_tfm; }; /* From patchwork Thu Jun 25 12:42:44 2020 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Ard Biesheuvel X-Patchwork-Id: 11625081 Return-Path: Received: from mail.kernel.org (pdx-korg-mail-1.web.codeaurora.org [172.30.200.123]) by pdx-korg-patchwork-2.web.codeaurora.org (Postfix) with ESMTP id 07A306C1 for ; Thu, 25 Jun 2020 12:43:53 +0000 (UTC) Received: from merlin.infradead.org (merlin.infradead.org [205.233.59.134]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by mail.kernel.org (Postfix) with ESMTPS id D3375206BE for ; Thu, 25 Jun 2020 12:43:52 +0000 (UTC) Authentication-Results: mail.kernel.org; dkim=pass (2048-bit key) header.d=lists.infradead.org header.i=@lists.infradead.org header.b="oQhXqtDq"; dkim=fail reason="signature verification failed" (1024-bit key) header.d=kernel.org header.i=@kernel.org header.b="q3/XepTQ" DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org D3375206BE Authentication-Results: mail.kernel.org; dmarc=fail (p=none dis=none) header.from=kernel.org Authentication-Results: mail.kernel.org; spf=none smtp.mailfrom=linux-amlogic-bounces+patchwork-linux-amlogic=patchwork.kernel.org@lists.infradead.org DKIM-Signature: v=1; a=rsa-sha256; q=dns/txt; c=relaxed/relaxed; d=lists.infradead.org; s=merlin.20170209; h=Sender:Content-Transfer-Encoding: Content-Type:Cc:List-Subscribe:List-Help:List-Post:List-Archive: List-Unsubscribe:List-Id:MIME-Version:References:In-Reply-To:Message-Id:Date: Subject:To:From:Reply-To:Content-ID:Content-Description:Resent-Date: Resent-From:Resent-Sender:Resent-To:Resent-Cc:Resent-Message-ID:List-Owner; bh=iTy4n5EmW3t+DYyqvEd2TbnbiaVvI/65BKVJ6FAxKAQ=; b=oQhXqtDqFbrvkneWvbKPrHrqw h+12oxXuPmH1KMVKxj2Dhz/+9UJEufNRcM+P7B9qNCLkrkIQAjYFVsdWrl7LfBDavYl/p75oKrTQG A3KR2BpEdl4/k/iLicDmpsBkBOR16GxlNFc/d9RXxkORIhQFvv7KWGfYAyGrcP45I7yIg/8pg8jGU CV5YuyFDVC2Ah8Hm37WSbWxknbwF/ushhhZbVjS+wJWeeXQ+eMKzJh6I5wIWCRS+trQP66jvDNWEo 9ZpFc1c9aqYdFfA06+ALanIcy/QddfmuwHV63qL8OD933hqPD5KdYhvghuKGpoZnNOuu9ccJMIz1X 4Af43DAyA==; Received: from localhost ([::1] helo=merlin.infradead.org) by merlin.infradead.org with esmtp (Exim 4.92.3 #3 (Red Hat Linux)) id 1joRE8-0005OE-4U; Thu, 25 Jun 2020 12:43:36 +0000 Received: from mail.kernel.org ([198.145.29.99]) by merlin.infradead.org with esmtps (Exim 4.92.3 #3 (Red Hat Linux)) id 1joRE3-0005Mk-W9; Thu, 25 Jun 2020 12:43:33 +0000 Received: from localhost.localdomain (lfbn-nic-1-188-42.w2-15.abo.wanadoo.fr [2.15.37.42]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by mail.kernel.org (Postfix) with ESMTPSA id A40A92082F; Thu, 25 Jun 2020 12:43:27 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=kernel.org; s=default; t=1593089011; bh=7g45QWyrYUg0An4TDCghUfvpEHV1X2Se+dPPjGObVdQ=; h=From:To:Cc:Subject:Date:In-Reply-To:References:From; b=q3/XepTQLoJhJKhn8/LBvEE9s+FbxrQLKtjsvC6S+AamPxSAD+v3rq3k8ov1gZ8jm ix1UsEqvKt2dqP/shAQkAzhBZXgDTBzog0jga5Gp2bp6hCeUR85HLXrYJWdguao6tR Fnbw1pH9Vmu41hADjmc9EHfDvkHPgXFXiKH075Zs= From: Ard Biesheuvel To: linux-crypto@vger.kernel.org Subject: [PATCH 03/12] crypto: omap-aes - permit asynchronous skcipher as fallback Date: Thu, 25 Jun 2020 14:42:44 +0200 Message-Id: <20200625124253.1906557-4-ardb@kernel.org> X-Mailer: git-send-email 2.27.0 In-Reply-To: <20200625124253.1906557-1-ardb@kernel.org> References: <20200625124253.1906557-1-ardb@kernel.org> MIME-Version: 1.0 X-Spam-Note: CRM114 invocation failed X-Spam-Score: -5.2 (-----) X-Spam-Report: SpamAssassin version 3.4.4 on merlin.infradead.org summary: Content analysis details: (-5.2 points) pts rule name description ---- ---------------------- -------------------------------------------------- -5.0 RCVD_IN_DNSWL_HI RBL: Sender listed at https://www.dnswl.org/, high trust [198.145.29.99 listed in list.dnswl.org] 0.0 SPF_HELO_NONE SPF: HELO does not publish an SPF Record -0.0 SPF_PASS SPF: sender matches SPF record -0.1 DKIM_VALID Message has at least one valid DKIM or DK signature -0.1 DKIM_VALID_EF Message has a valid DKIM or DK signature from envelope-from domain -0.1 DKIM_VALID_AU Message has a valid DKIM or DK signature from author's domain 0.1 DKIM_SIGNED Message has a DKIM or DK signature, not necessarily valid -0.0 DKIMWL_WL_HIGH DKIMwl.org - Whitelisted High sender X-BeenThere: linux-amlogic@lists.infradead.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Cc: Eric Biggers , Tom Lendacky , Jamie Iles , Herbert Xu , Rohit Maheshwari , Fabio Estevam , Sascha Hauer , Vinay Kumar Yadav , Maxime Ripard , "David S. Miller" , Chen-Yu Tsai , Corentin Labbe , NXP Linux Team , Pengutronix Kernel Team , Ayush Sawal , linux-amlogic@lists.infradead.org, linux-omap@vger.kernel.org, Shawn Guo , Ard Biesheuvel , linux-arm-kernel@lists.infradead.org Sender: "linux-amlogic" Errors-To: linux-amlogic-bounces+patchwork-linux-amlogic=patchwork.kernel.org@lists.infradead.org Even though the omap-aes driver implements asynchronous versions of ecb(aes), cbc(aes) and ctr(aes), the fallbacks it allocates are required to be synchronous. Given that SIMD based software implementations are usually asynchronous as well, even though they rarely complete asynchronously (this typically only happens in cases where the request was made from softirq context, while SIMD was already in use in the task context that it interrupted), these implementations are disregarded, and either the generic C version or another table based version implemented in assembler is selected instead. Since falling back to synchronous AES is not only a performance issue, but potentially a security issue as well (due to the fact that table based AES is not time invariant), let's fix this, by allocating an ordinary skcipher as the fallback, and invoke it with the completion routine that was given to the outer request. Signed-off-by: Ard Biesheuvel --- drivers/crypto/omap-aes.c | 35 ++++++++++---------- drivers/crypto/omap-aes.h | 3 +- 2 files changed, 19 insertions(+), 19 deletions(-) diff --git a/drivers/crypto/omap-aes.c b/drivers/crypto/omap-aes.c index b5aff20c5900..25154b74dcc6 100644 --- a/drivers/crypto/omap-aes.c +++ b/drivers/crypto/omap-aes.c @@ -548,20 +548,18 @@ static int omap_aes_crypt(struct skcipher_request *req, unsigned long mode) !!(mode & FLAGS_CBC)); if (req->cryptlen < aes_fallback_sz) { - SYNC_SKCIPHER_REQUEST_ON_STACK(subreq, ctx->fallback); - - skcipher_request_set_sync_tfm(subreq, ctx->fallback); - skcipher_request_set_callback(subreq, req->base.flags, NULL, - NULL); - skcipher_request_set_crypt(subreq, req->src, req->dst, - req->cryptlen, req->iv); + skcipher_request_set_tfm(&rctx->fallback_req, ctx->fallback); + skcipher_request_set_callback(&rctx->fallback_req, + req->base.flags, + req->base.complete, + req->base.data); + skcipher_request_set_crypt(&rctx->fallback_req, req->src, + req->dst, req->cryptlen, req->iv); if (mode & FLAGS_ENCRYPT) - ret = crypto_skcipher_encrypt(subreq); + ret = crypto_skcipher_encrypt(&rctx->fallback_req); else - ret = crypto_skcipher_decrypt(subreq); - - skcipher_request_zero(subreq); + ret = crypto_skcipher_decrypt(&rctx->fallback_req); return ret; } dd = omap_aes_find_dev(rctx); @@ -590,11 +588,11 @@ static int omap_aes_setkey(struct crypto_skcipher *tfm, const u8 *key, memcpy(ctx->key, key, keylen); ctx->keylen = keylen; - crypto_sync_skcipher_clear_flags(ctx->fallback, CRYPTO_TFM_REQ_MASK); - crypto_sync_skcipher_set_flags(ctx->fallback, tfm->base.crt_flags & + crypto_skcipher_clear_flags(ctx->fallback, CRYPTO_TFM_REQ_MASK); + crypto_skcipher_set_flags(ctx->fallback, tfm->base.crt_flags & CRYPTO_TFM_REQ_MASK); - ret = crypto_sync_skcipher_setkey(ctx->fallback, key, keylen); + ret = crypto_skcipher_setkey(ctx->fallback, key, keylen); if (!ret) return 0; @@ -640,15 +638,16 @@ static int omap_aes_init_tfm(struct crypto_skcipher *tfm) { const char *name = crypto_tfm_alg_name(&tfm->base); struct omap_aes_ctx *ctx = crypto_skcipher_ctx(tfm); - struct crypto_sync_skcipher *blk; + struct crypto_skcipher *blk; - blk = crypto_alloc_sync_skcipher(name, 0, CRYPTO_ALG_NEED_FALLBACK); + blk = crypto_alloc_skcipher(name, 0, CRYPTO_ALG_NEED_FALLBACK); if (IS_ERR(blk)) return PTR_ERR(blk); ctx->fallback = blk; - crypto_skcipher_set_reqsize(tfm, sizeof(struct omap_aes_reqctx)); + crypto_skcipher_set_reqsize(tfm, sizeof(struct omap_aes_reqctx) + + crypto_skcipher_reqsize(blk)); ctx->enginectx.op.prepare_request = omap_aes_prepare_req; ctx->enginectx.op.unprepare_request = NULL; @@ -662,7 +661,7 @@ static void omap_aes_exit_tfm(struct crypto_skcipher *tfm) struct omap_aes_ctx *ctx = crypto_skcipher_ctx(tfm); if (ctx->fallback) - crypto_free_sync_skcipher(ctx->fallback); + crypto_free_skcipher(ctx->fallback); ctx->fallback = NULL; } diff --git a/drivers/crypto/omap-aes.h b/drivers/crypto/omap-aes.h index 2d111bf906e1..23d073e87bb8 100644 --- a/drivers/crypto/omap-aes.h +++ b/drivers/crypto/omap-aes.h @@ -97,7 +97,7 @@ struct omap_aes_ctx { int keylen; u32 key[AES_KEYSIZE_256 / sizeof(u32)]; u8 nonce[4]; - struct crypto_sync_skcipher *fallback; + struct crypto_skcipher *fallback; }; struct omap_aes_gcm_ctx { @@ -110,6 +110,7 @@ struct omap_aes_reqctx { unsigned long mode; u8 iv[AES_BLOCK_SIZE]; u32 auth_tag[AES_BLOCK_SIZE / sizeof(u32)]; + struct skcipher_request fallback_req; // keep at the end }; #define OMAP_AES_QUEUE_LENGTH 1 From patchwork Thu Jun 25 12:42:45 2020 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Ard Biesheuvel X-Patchwork-Id: 11625095 Return-Path: Received: from mail.kernel.org (pdx-korg-mail-1.web.codeaurora.org [172.30.200.123]) by pdx-korg-patchwork-2.web.codeaurora.org (Postfix) with ESMTP id B1CC014B7 for ; Thu, 25 Jun 2020 12:43:59 +0000 (UTC) Received: from merlin.infradead.org (merlin.infradead.org [205.233.59.134]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by mail.kernel.org (Postfix) with ESMTPS id 8689420720 for ; Thu, 25 Jun 2020 12:43:59 +0000 (UTC) Authentication-Results: mail.kernel.org; dkim=pass (2048-bit key) header.d=lists.infradead.org header.i=@lists.infradead.org header.b="z8XV7ayk"; dkim=fail reason="signature verification failed" (1024-bit key) header.d=kernel.org header.i=@kernel.org header.b="XQtmLNpj" DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org 8689420720 Authentication-Results: mail.kernel.org; dmarc=fail (p=none dis=none) header.from=kernel.org Authentication-Results: mail.kernel.org; spf=none smtp.mailfrom=linux-amlogic-bounces+patchwork-linux-amlogic=patchwork.kernel.org@lists.infradead.org DKIM-Signature: v=1; a=rsa-sha256; q=dns/txt; c=relaxed/relaxed; d=lists.infradead.org; s=merlin.20170209; h=Sender:Content-Transfer-Encoding: Content-Type:Cc:List-Subscribe:List-Help:List-Post:List-Archive: List-Unsubscribe:List-Id:MIME-Version:References:In-Reply-To:Message-Id:Date: Subject:To:From:Reply-To:Content-ID:Content-Description:Resent-Date: Resent-From:Resent-Sender:Resent-To:Resent-Cc:Resent-Message-ID:List-Owner; bh=pW3fPTfigRULS2DieSKb1ysoMTzL31QoAJ41Te7bd2M=; b=z8XV7ayk2Cmgvj613OMxBP325 O4xNG3jKC+K76m+Fpni6Pm95wSqeVhsI3WmuEj5O30/6yqDfiRBcwnFM+zepMJRHOxaV0nZxImuSc JeNT3zN5eeo9AP+Fg/09n9cCAAU9z4GrheEb82hEQiPc/3hc8vXZ2Hc7Sm46PBbsMESjo6GhVW5p1 wPuyimKmY7DZOBW7wcbNLibcetos6JYTSBad5FKwn+yiXb+zDbRW5Sx81eI1gZ9hsaSzfNAZNZGkh fyRGBaCXrcF06vYNxn5XoYTqMYft7Yrx8DVq4iNTDEwuXJO5bX/Qjuuf8iE11MjuFQw2tvduk1CVJ MGJsk5/kw==; Received: from localhost ([::1] helo=merlin.infradead.org) by merlin.infradead.org with esmtp (Exim 4.92.3 #3 (Red Hat Linux)) id 1joREE-0005Qv-OZ; Thu, 25 Jun 2020 12:43:42 +0000 Received: from mail.kernel.org ([198.145.29.99]) by merlin.infradead.org with esmtps (Exim 4.92.3 #3 (Red Hat Linux)) id 1joRE7-0005O8-Ul; Thu, 25 Jun 2020 12:43:39 +0000 Received: from localhost.localdomain (lfbn-nic-1-188-42.w2-15.abo.wanadoo.fr [2.15.37.42]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by mail.kernel.org (Postfix) with ESMTPSA id ACD8320857; Thu, 25 Jun 2020 12:43:31 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=kernel.org; s=default; t=1593089015; bh=hood+SBSrfD8rlAvUlszp6fbHnc3dUB49g4uRq5KZac=; h=From:To:Cc:Subject:Date:In-Reply-To:References:From; b=XQtmLNpj/XMLhmv27k6PpfLv9k5QCNBz3Op+LSMmPHZhuWFmbZv+z8CdBQBp9SURh 7OIIpUfsC8LIFh/0p4k+vFWwCzlWlmuEnHXC8aNhIj4MQgocQo/5bDL4imQYKwZyhS +NAgiY3UHxUm9isxqRhPWy6iKfEag1R5RVq5Pdmw= From: Ard Biesheuvel To: linux-crypto@vger.kernel.org Subject: [PATCH 04/12] crypto: sun4i - permit asynchronous skcipher as fallback Date: Thu, 25 Jun 2020 14:42:45 +0200 Message-Id: <20200625124253.1906557-5-ardb@kernel.org> X-Mailer: git-send-email 2.27.0 In-Reply-To: <20200625124253.1906557-1-ardb@kernel.org> References: <20200625124253.1906557-1-ardb@kernel.org> MIME-Version: 1.0 X-Spam-Note: CRM114 invocation failed X-Spam-Score: -5.2 (-----) X-Spam-Report: SpamAssassin version 3.4.4 on merlin.infradead.org summary: Content analysis details: (-5.2 points) pts rule name description ---- ---------------------- -------------------------------------------------- -5.0 RCVD_IN_DNSWL_HI RBL: Sender listed at https://www.dnswl.org/, high trust [198.145.29.99 listed in list.dnswl.org] 0.0 SPF_HELO_NONE SPF: HELO does not publish an SPF Record -0.0 SPF_PASS SPF: sender matches SPF record -0.1 DKIM_VALID Message has at least one valid DKIM or DK signature -0.1 DKIM_VALID_EF Message has a valid DKIM or DK signature from envelope-from domain -0.1 DKIM_VALID_AU Message has a valid DKIM or DK signature from author's domain 0.1 DKIM_SIGNED Message has a DKIM or DK signature, not necessarily valid -0.0 DKIMWL_WL_HIGH DKIMwl.org - Whitelisted High sender X-BeenThere: linux-amlogic@lists.infradead.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Cc: Eric Biggers , Tom Lendacky , Jamie Iles , Herbert Xu , Rohit Maheshwari , Fabio Estevam , Sascha Hauer , Vinay Kumar Yadav , Maxime Ripard , "David S. Miller" , Chen-Yu Tsai , Corentin Labbe , NXP Linux Team , Pengutronix Kernel Team , Ayush Sawal , linux-amlogic@lists.infradead.org, linux-omap@vger.kernel.org, Shawn Guo , Ard Biesheuvel , linux-arm-kernel@lists.infradead.org Sender: "linux-amlogic" Errors-To: linux-amlogic-bounces+patchwork-linux-amlogic=patchwork.kernel.org@lists.infradead.org Even though the sun4i driver implements asynchronous versions of ecb(aes) and cbc(aes), the fallbacks it allocates are required to be synchronous. Given that SIMD based software implementations are usually asynchronous as well, even though they rarely complete asynchronously (this typically only happens in cases where the request was made from softirq context, while SIMD was already in use in the task context that it interrupted), these implementations are disregarded, and either the generic C version or another table based version implemented in assembler is selected instead. Since falling back to synchronous AES is not only a performance issue, but potentially a security issue as well (due to the fact that table based AES is not time invariant), let's fix this, by allocating an ordinary skcipher as the fallback, and invoke it with the completion routine that was given to the outer request. Signed-off-by: Ard Biesheuvel --- drivers/crypto/allwinner/sun4i-ss/sun4i-ss-cipher.c | 46 ++++++++++---------- drivers/crypto/allwinner/sun4i-ss/sun4i-ss.h | 3 +- 2 files changed, 25 insertions(+), 24 deletions(-) diff --git a/drivers/crypto/allwinner/sun4i-ss/sun4i-ss-cipher.c b/drivers/crypto/allwinner/sun4i-ss/sun4i-ss-cipher.c index 7f22d305178e..b72de8939497 100644 --- a/drivers/crypto/allwinner/sun4i-ss/sun4i-ss-cipher.c +++ b/drivers/crypto/allwinner/sun4i-ss/sun4i-ss-cipher.c @@ -122,19 +122,17 @@ static int noinline_for_stack sun4i_ss_cipher_poll_fallback(struct skcipher_requ struct crypto_skcipher *tfm = crypto_skcipher_reqtfm(areq); struct sun4i_tfm_ctx *op = crypto_skcipher_ctx(tfm); struct sun4i_cipher_req_ctx *ctx = skcipher_request_ctx(areq); - SYNC_SKCIPHER_REQUEST_ON_STACK(subreq, op->fallback_tfm); int err; - skcipher_request_set_sync_tfm(subreq, op->fallback_tfm); - skcipher_request_set_callback(subreq, areq->base.flags, NULL, - NULL); - skcipher_request_set_crypt(subreq, areq->src, areq->dst, + skcipher_request_set_tfm(&ctx->fallback_req, op->fallback_tfm); + skcipher_request_set_callback(&ctx->fallback_req, areq->base.flags, + areq->base.complete, areq->base.data); + skcipher_request_set_crypt(&ctx->fallback_req, areq->src, areq->dst, areq->cryptlen, areq->iv); if (ctx->mode & SS_DECRYPTION) - err = crypto_skcipher_decrypt(subreq); + err = crypto_skcipher_decrypt(&ctx->fallback_req); else - err = crypto_skcipher_encrypt(subreq); - skcipher_request_zero(subreq); + err = crypto_skcipher_encrypt(&ctx->fallback_req); return err; } @@ -494,23 +492,25 @@ int sun4i_ss_cipher_init(struct crypto_tfm *tfm) alg.crypto.base); op->ss = algt->ss; - crypto_skcipher_set_reqsize(__crypto_skcipher_cast(tfm), - sizeof(struct sun4i_cipher_req_ctx)); - - op->fallback_tfm = crypto_alloc_sync_skcipher(name, 0, CRYPTO_ALG_NEED_FALLBACK); + op->fallback_tfm = crypto_alloc_skcipher(name, 0, CRYPTO_ALG_NEED_FALLBACK); if (IS_ERR(op->fallback_tfm)) { dev_err(op->ss->dev, "ERROR: Cannot allocate fallback for %s %ld\n", name, PTR_ERR(op->fallback_tfm)); return PTR_ERR(op->fallback_tfm); } + crypto_skcipher_set_reqsize(__crypto_skcipher_cast(tfm), + sizeof(struct sun4i_cipher_req_ctx) + + crypto_skcipher_reqsize(op->fallback_tfm)); + + err = pm_runtime_get_sync(op->ss->dev); if (err < 0) goto error_pm; return 0; error_pm: - crypto_free_sync_skcipher(op->fallback_tfm); + crypto_free_skcipher(op->fallback_tfm); return err; } @@ -518,7 +518,7 @@ void sun4i_ss_cipher_exit(struct crypto_tfm *tfm) { struct sun4i_tfm_ctx *op = crypto_tfm_ctx(tfm); - crypto_free_sync_skcipher(op->fallback_tfm); + crypto_free_skcipher(op->fallback_tfm); pm_runtime_put(op->ss->dev); } @@ -546,10 +546,10 @@ int sun4i_ss_aes_setkey(struct crypto_skcipher *tfm, const u8 *key, op->keylen = keylen; memcpy(op->key, key, keylen); - crypto_sync_skcipher_clear_flags(op->fallback_tfm, CRYPTO_TFM_REQ_MASK); - crypto_sync_skcipher_set_flags(op->fallback_tfm, tfm->base.crt_flags & CRYPTO_TFM_REQ_MASK); + crypto_skcipher_clear_flags(op->fallback_tfm, CRYPTO_TFM_REQ_MASK); + crypto_skcipher_set_flags(op->fallback_tfm, tfm->base.crt_flags & CRYPTO_TFM_REQ_MASK); - return crypto_sync_skcipher_setkey(op->fallback_tfm, key, keylen); + return crypto_skcipher_setkey(op->fallback_tfm, key, keylen); } /* check and set the DES key, prepare the mode to be used */ @@ -566,10 +566,10 @@ int sun4i_ss_des_setkey(struct crypto_skcipher *tfm, const u8 *key, op->keylen = keylen; memcpy(op->key, key, keylen); - crypto_sync_skcipher_clear_flags(op->fallback_tfm, CRYPTO_TFM_REQ_MASK); - crypto_sync_skcipher_set_flags(op->fallback_tfm, tfm->base.crt_flags & CRYPTO_TFM_REQ_MASK); + crypto_skcipher_clear_flags(op->fallback_tfm, CRYPTO_TFM_REQ_MASK); + crypto_skcipher_set_flags(op->fallback_tfm, tfm->base.crt_flags & CRYPTO_TFM_REQ_MASK); - return crypto_sync_skcipher_setkey(op->fallback_tfm, key, keylen); + return crypto_skcipher_setkey(op->fallback_tfm, key, keylen); } /* check and set the 3DES key, prepare the mode to be used */ @@ -586,9 +586,9 @@ int sun4i_ss_des3_setkey(struct crypto_skcipher *tfm, const u8 *key, op->keylen = keylen; memcpy(op->key, key, keylen); - crypto_sync_skcipher_clear_flags(op->fallback_tfm, CRYPTO_TFM_REQ_MASK); - crypto_sync_skcipher_set_flags(op->fallback_tfm, tfm->base.crt_flags & CRYPTO_TFM_REQ_MASK); + crypto_skcipher_clear_flags(op->fallback_tfm, CRYPTO_TFM_REQ_MASK); + crypto_skcipher_set_flags(op->fallback_tfm, tfm->base.crt_flags & CRYPTO_TFM_REQ_MASK); - return crypto_sync_skcipher_setkey(op->fallback_tfm, key, keylen); + return crypto_skcipher_setkey(op->fallback_tfm, key, keylen); } diff --git a/drivers/crypto/allwinner/sun4i-ss/sun4i-ss.h b/drivers/crypto/allwinner/sun4i-ss/sun4i-ss.h index 2b4c6333eb67..163962f9e284 100644 --- a/drivers/crypto/allwinner/sun4i-ss/sun4i-ss.h +++ b/drivers/crypto/allwinner/sun4i-ss/sun4i-ss.h @@ -170,11 +170,12 @@ struct sun4i_tfm_ctx { u32 keylen; u32 keymode; struct sun4i_ss_ctx *ss; - struct crypto_sync_skcipher *fallback_tfm; + struct crypto_skcipher *fallback_tfm; }; struct sun4i_cipher_req_ctx { u32 mode; + struct skcipher_request fallback_req; // keep at the end }; struct sun4i_req_ctx { From patchwork Thu Jun 25 12:42:46 2020 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Ard Biesheuvel X-Patchwork-Id: 11625101 Return-Path: Received: from mail.kernel.org (pdx-korg-mail-1.web.codeaurora.org [172.30.200.123]) by pdx-korg-patchwork-2.web.codeaurora.org (Postfix) with ESMTP id 7B2526C1 for ; Thu, 25 Jun 2020 12:44:05 +0000 (UTC) Received: from merlin.infradead.org (merlin.infradead.org [205.233.59.134]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by mail.kernel.org (Postfix) with ESMTPS id 52DA420724 for ; Thu, 25 Jun 2020 12:44:05 +0000 (UTC) Authentication-Results: mail.kernel.org; dkim=pass (2048-bit key) header.d=lists.infradead.org header.i=@lists.infradead.org header.b="AZf6MFTc"; dkim=fail reason="signature verification failed" (1024-bit key) header.d=kernel.org header.i=@kernel.org header.b="DV76en3u" DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org 52DA420724 Authentication-Results: mail.kernel.org; dmarc=fail (p=none dis=none) header.from=kernel.org Authentication-Results: mail.kernel.org; spf=none smtp.mailfrom=linux-amlogic-bounces+patchwork-linux-amlogic=patchwork.kernel.org@lists.infradead.org DKIM-Signature: v=1; a=rsa-sha256; q=dns/txt; c=relaxed/relaxed; d=lists.infradead.org; s=merlin.20170209; h=Sender:Content-Transfer-Encoding: Content-Type:Cc:List-Subscribe:List-Help:List-Post:List-Archive: List-Unsubscribe:List-Id:MIME-Version:References:In-Reply-To:Message-Id:Date: Subject:To:From:Reply-To:Content-ID:Content-Description:Resent-Date: Resent-From:Resent-Sender:Resent-To:Resent-Cc:Resent-Message-ID:List-Owner; bh=N/9kubKxjrry9zafhX+NlC7yX4Dv/vNLuSvryvAJxak=; b=AZf6MFTcXIoizroErq68tlV14 d+dmFWF3LVcLcRL6N6nejca/u2YCpE+0GRULb49qOkH5KQhi612Jff4DMpeclINHpA8Kd6ZjrTUjm E287OGzuV4I7Zf01vZupAOpCiP0knNF+yZpXBSP/qXSeAbatfYe0FJKaDiDd2KbmIXmVNhnIhuoax xriOIXs7odKf7UIpnyFoW2e2eSubhw0xGAI4L7RBO1pMlfrLc3fA24Yie+0a4GJzoaiq2FnKZKyng lEF2PWfwVWIhSNoHWFG1HBXXN4sBcln19SK/6xU82jt/d2d9AFgjZcwQKQMUd/TbEoPbRN/xj4ahe 94Gc6FyNA==; Received: from localhost ([::1] helo=merlin.infradead.org) by merlin.infradead.org with esmtp (Exim 4.92.3 #3 (Red Hat Linux)) id 1joREM-0005UU-UZ; Thu, 25 Jun 2020 12:43:50 +0000 Received: from mail.kernel.org ([198.145.29.99]) by merlin.infradead.org with esmtps (Exim 4.92.3 #3 (Red Hat Linux)) id 1joREB-0005Pc-VD; Thu, 25 Jun 2020 12:43:41 +0000 Received: from localhost.localdomain (lfbn-nic-1-188-42.w2-15.abo.wanadoo.fr [2.15.37.42]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by mail.kernel.org (Postfix) with ESMTPSA id B87AF207E8; Thu, 25 Jun 2020 12:43:35 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=kernel.org; s=default; t=1593089019; bh=P1hf02tXhdwPbGM+6yMTQXpLfUnkB8fw+ZcUm1FA84g=; h=From:To:Cc:Subject:Date:In-Reply-To:References:From; b=DV76en3uPtZZ7Xa/1aPEPLiw5Jp89vEwaOYr/X9IA80h8wvj0pm2M3EO2gSY9z3Q2 BcBy95c/maeddjPwdddMFrEJbgk5PhTEQ9nMMA/P/0vucdNQXlug4Ywlr024oOetV1 g58CgBtK8+KfobC/ZFKOBuH3LbGxsRXvxEDfcLrY= From: Ard Biesheuvel To: linux-crypto@vger.kernel.org Subject: [PATCH 05/12] crypto: sun8i-ce - permit asynchronous skcipher as fallback Date: Thu, 25 Jun 2020 14:42:46 +0200 Message-Id: <20200625124253.1906557-6-ardb@kernel.org> X-Mailer: git-send-email 2.27.0 In-Reply-To: <20200625124253.1906557-1-ardb@kernel.org> References: <20200625124253.1906557-1-ardb@kernel.org> MIME-Version: 1.0 X-Spam-Note: CRM114 invocation failed X-Spam-Score: -5.2 (-----) X-Spam-Report: SpamAssassin version 3.4.4 on merlin.infradead.org summary: Content analysis details: (-5.2 points) pts rule name description ---- ---------------------- -------------------------------------------------- -5.0 RCVD_IN_DNSWL_HI RBL: Sender listed at https://www.dnswl.org/, high trust [198.145.29.99 listed in list.dnswl.org] 0.0 SPF_HELO_NONE SPF: HELO does not publish an SPF Record -0.0 SPF_PASS SPF: sender matches SPF record -0.1 DKIM_VALID Message has at least one valid DKIM or DK signature -0.1 DKIM_VALID_EF Message has a valid DKIM or DK signature from envelope-from domain -0.1 DKIM_VALID_AU Message has a valid DKIM or DK signature from author's domain 0.1 DKIM_SIGNED Message has a DKIM or DK signature, not necessarily valid -0.0 DKIMWL_WL_HIGH DKIMwl.org - Whitelisted High sender X-BeenThere: linux-amlogic@lists.infradead.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Cc: Eric Biggers , Tom Lendacky , Jamie Iles , Herbert Xu , Rohit Maheshwari , Fabio Estevam , Sascha Hauer , Vinay Kumar Yadav , Maxime Ripard , "David S. Miller" , Chen-Yu Tsai , Corentin Labbe , NXP Linux Team , Pengutronix Kernel Team , Ayush Sawal , linux-amlogic@lists.infradead.org, linux-omap@vger.kernel.org, Shawn Guo , Ard Biesheuvel , linux-arm-kernel@lists.infradead.org Sender: "linux-amlogic" Errors-To: linux-amlogic-bounces+patchwork-linux-amlogic=patchwork.kernel.org@lists.infradead.org Even though the sun8i-ce driver implements asynchronous versions of ecb(aes) and cbc(aes), the fallbacks it allocates are required to be synchronous. Given that SIMD based software implementations are usually asynchronous as well, even though they rarely complete asynchronously (this typically only happens in cases where the request was made from softirq context, while SIMD was already in use in the task context that it interrupted), these implementations are disregarded, and either the generic C version or another table based version implemented in assembler is selected instead. Since falling back to synchronous AES is not only a performance issue, but potentially a security issue as well (due to the fact that table based AES is not time invariant), let's fix this, by allocating an ordinary skcipher as the fallback, and invoke it with the completion routine that was given to the outer request. Signed-off-by: Ard Biesheuvel --- drivers/crypto/allwinner/sun8i-ce/sun8i-ce-cipher.c | 41 ++++++++++---------- drivers/crypto/allwinner/sun8i-ce/sun8i-ce.h | 3 +- 2 files changed, 22 insertions(+), 22 deletions(-) diff --git a/drivers/crypto/allwinner/sun8i-ce/sun8i-ce-cipher.c b/drivers/crypto/allwinner/sun8i-ce/sun8i-ce-cipher.c index a6abb701bfc6..82c99da24dfd 100644 --- a/drivers/crypto/allwinner/sun8i-ce/sun8i-ce-cipher.c +++ b/drivers/crypto/allwinner/sun8i-ce/sun8i-ce-cipher.c @@ -58,23 +58,20 @@ static int sun8i_ce_cipher_fallback(struct skcipher_request *areq) #ifdef CONFIG_CRYPTO_DEV_SUN8I_CE_DEBUG struct skcipher_alg *alg = crypto_skcipher_alg(tfm); struct sun8i_ce_alg_template *algt; -#endif - SYNC_SKCIPHER_REQUEST_ON_STACK(subreq, op->fallback_tfm); -#ifdef CONFIG_CRYPTO_DEV_SUN8I_CE_DEBUG algt = container_of(alg, struct sun8i_ce_alg_template, alg.skcipher); algt->stat_fb++; #endif - skcipher_request_set_sync_tfm(subreq, op->fallback_tfm); - skcipher_request_set_callback(subreq, areq->base.flags, NULL, NULL); - skcipher_request_set_crypt(subreq, areq->src, areq->dst, + skcipher_request_set_tfm(&rctx->fallback_req, op->fallback_tfm); + skcipher_request_set_callback(&rctx->fallback_req, areq->base.flags, + areq->base.complete, areq->base.data); + skcipher_request_set_crypt(&rctx->fallback_req, areq->src, areq->dst, areq->cryptlen, areq->iv); if (rctx->op_dir & CE_DECRYPTION) - err = crypto_skcipher_decrypt(subreq); + err = crypto_skcipher_decrypt(&rctx->fallback_req); else - err = crypto_skcipher_encrypt(subreq); - skcipher_request_zero(subreq); + err = crypto_skcipher_encrypt(&rctx->fallback_req); return err; } @@ -335,18 +332,20 @@ int sun8i_ce_cipher_init(struct crypto_tfm *tfm) algt = container_of(alg, struct sun8i_ce_alg_template, alg.skcipher); op->ce = algt->ce; - sktfm->reqsize = sizeof(struct sun8i_cipher_req_ctx); - - op->fallback_tfm = crypto_alloc_sync_skcipher(name, 0, CRYPTO_ALG_NEED_FALLBACK); + op->fallback_tfm = crypto_alloc_skcipher(name, 0, CRYPTO_ALG_NEED_FALLBACK); if (IS_ERR(op->fallback_tfm)) { dev_err(op->ce->dev, "ERROR: Cannot allocate fallback for %s %ld\n", name, PTR_ERR(op->fallback_tfm)); return PTR_ERR(op->fallback_tfm); } + sktfm->reqsize = sizeof(struct sun8i_cipher_req_ctx) + + crypto_skcipher_reqsize(op->fallback_tfm); + + dev_info(op->ce->dev, "Fallback for %s is %s\n", crypto_tfm_alg_driver_name(&sktfm->base), - crypto_tfm_alg_driver_name(crypto_skcipher_tfm(&op->fallback_tfm->base))); + crypto_tfm_alg_driver_name(crypto_skcipher_tfm(op->fallback_tfm))); op->enginectx.op.do_one_request = sun8i_ce_handle_cipher_request; op->enginectx.op.prepare_request = NULL; @@ -358,7 +357,7 @@ int sun8i_ce_cipher_init(struct crypto_tfm *tfm) return 0; error_pm: - crypto_free_sync_skcipher(op->fallback_tfm); + crypto_free_skcipher(op->fallback_tfm); return err; } @@ -370,7 +369,7 @@ void sun8i_ce_cipher_exit(struct crypto_tfm *tfm) memzero_explicit(op->key, op->keylen); kfree(op->key); } - crypto_free_sync_skcipher(op->fallback_tfm); + crypto_free_skcipher(op->fallback_tfm); pm_runtime_put_sync_suspend(op->ce->dev); } @@ -400,10 +399,10 @@ int sun8i_ce_aes_setkey(struct crypto_skcipher *tfm, const u8 *key, if (!op->key) return -ENOMEM; - crypto_sync_skcipher_clear_flags(op->fallback_tfm, CRYPTO_TFM_REQ_MASK); - crypto_sync_skcipher_set_flags(op->fallback_tfm, tfm->base.crt_flags & CRYPTO_TFM_REQ_MASK); + crypto_skcipher_clear_flags(op->fallback_tfm, CRYPTO_TFM_REQ_MASK); + crypto_skcipher_set_flags(op->fallback_tfm, tfm->base.crt_flags & CRYPTO_TFM_REQ_MASK); - return crypto_sync_skcipher_setkey(op->fallback_tfm, key, keylen); + return crypto_skcipher_setkey(op->fallback_tfm, key, keylen); } int sun8i_ce_des3_setkey(struct crypto_skcipher *tfm, const u8 *key, @@ -425,8 +424,8 @@ int sun8i_ce_des3_setkey(struct crypto_skcipher *tfm, const u8 *key, if (!op->key) return -ENOMEM; - crypto_sync_skcipher_clear_flags(op->fallback_tfm, CRYPTO_TFM_REQ_MASK); - crypto_sync_skcipher_set_flags(op->fallback_tfm, tfm->base.crt_flags & CRYPTO_TFM_REQ_MASK); + crypto_skcipher_clear_flags(op->fallback_tfm, CRYPTO_TFM_REQ_MASK); + crypto_skcipher_set_flags(op->fallback_tfm, tfm->base.crt_flags & CRYPTO_TFM_REQ_MASK); - return crypto_sync_skcipher_setkey(op->fallback_tfm, key, keylen); + return crypto_skcipher_setkey(op->fallback_tfm, key, keylen); } diff --git a/drivers/crypto/allwinner/sun8i-ce/sun8i-ce.h b/drivers/crypto/allwinner/sun8i-ce/sun8i-ce.h index 0e9eac397e1b..4ac0f91e2800 100644 --- a/drivers/crypto/allwinner/sun8i-ce/sun8i-ce.h +++ b/drivers/crypto/allwinner/sun8i-ce/sun8i-ce.h @@ -187,6 +187,7 @@ struct sun8i_ce_dev { struct sun8i_cipher_req_ctx { u32 op_dir; int flow; + struct skcipher_request fallback_req; // keep at the end }; /* @@ -202,7 +203,7 @@ struct sun8i_cipher_tfm_ctx { u32 *key; u32 keylen; struct sun8i_ce_dev *ce; - struct crypto_sync_skcipher *fallback_tfm; + struct crypto_skcipher *fallback_tfm; }; /* From patchwork Thu Jun 25 12:42:47 2020 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Ard Biesheuvel X-Patchwork-Id: 11625107 Return-Path: Received: from mail.kernel.org (pdx-korg-mail-1.web.codeaurora.org [172.30.200.123]) by pdx-korg-patchwork-2.web.codeaurora.org (Postfix) with ESMTP id 3C6836C1 for ; Thu, 25 Jun 2020 12:44:17 +0000 (UTC) Received: from merlin.infradead.org (merlin.infradead.org [205.233.59.134]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by mail.kernel.org (Postfix) with ESMTPS id 1572D206BE for ; Thu, 25 Jun 2020 12:44:17 +0000 (UTC) Authentication-Results: mail.kernel.org; dkim=pass (2048-bit key) header.d=lists.infradead.org header.i=@lists.infradead.org header.b="jW2k8/f4"; dkim=fail reason="signature verification failed" (1024-bit key) header.d=kernel.org header.i=@kernel.org header.b="faSHLjmL" DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org 1572D206BE Authentication-Results: mail.kernel.org; dmarc=fail (p=none dis=none) header.from=kernel.org Authentication-Results: mail.kernel.org; spf=none smtp.mailfrom=linux-amlogic-bounces+patchwork-linux-amlogic=patchwork.kernel.org@lists.infradead.org DKIM-Signature: v=1; a=rsa-sha256; q=dns/txt; c=relaxed/relaxed; d=lists.infradead.org; s=merlin.20170209; h=Sender:Content-Transfer-Encoding: Content-Type:Cc:List-Subscribe:List-Help:List-Post:List-Archive: List-Unsubscribe:List-Id:MIME-Version:References:In-Reply-To:Message-Id:Date: Subject:To:From:Reply-To:Content-ID:Content-Description:Resent-Date: Resent-From:Resent-Sender:Resent-To:Resent-Cc:Resent-Message-ID:List-Owner; bh=uov9O5wq0emLvSRg02tPAIsLto0Pp4ECwxHKPboRmk0=; b=jW2k8/f4Ay+tNjcM5Mw4/9Kf8 NmixyvhGLCpOTdRN/yYEXzng+hjM1O0iBnK5K5tOk7XP7fThPBAEBoVnghEzZRGvn5Vx2x3FIxt9+ sqa6QGs5oxh6KCSu4GoaKqBDppDj5tQnKXIHWbUPQ5KVy6gwiw6nTUfNr383dUR4Q8I5A27v7lXV/ DIbGGQKFkTzLq6eFOQRbso36Nb1gnZIeshFRfm1XvUTPNsiLck+HXyeeVuNdihIqye2ylB0x8HJ++ eNJn8EyzJWXZPES67AOyBfAMB7nK5XUl+7gyyuD+adsmX0KqOVSFdJ4m3Dd5WrXgxDHnXilBh08DE 8sOufqUow==; Received: from localhost ([::1] helo=merlin.infradead.org) by merlin.infradead.org with esmtp (Exim 4.92.3 #3 (Red Hat Linux)) id 1joREZ-0005aB-OD; Thu, 25 Jun 2020 12:44:03 +0000 Received: from mail.kernel.org ([198.145.29.99]) by merlin.infradead.org with esmtps (Exim 4.92.3 #3 (Red Hat Linux)) id 1joREG-0005RY-6i; Thu, 25 Jun 2020 12:43:45 +0000 Received: from localhost.localdomain (lfbn-nic-1-188-42.w2-15.abo.wanadoo.fr [2.15.37.42]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by mail.kernel.org (Postfix) with ESMTPSA id C452C20724; Thu, 25 Jun 2020 12:43:39 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=kernel.org; s=default; t=1593089023; bh=tpcze3XSHyZPREkPUbsCkWhR8rfSX5Y6bvgG9Oth1A8=; h=From:To:Cc:Subject:Date:In-Reply-To:References:From; b=faSHLjmLd+sYk6+dbz3dDtMmv4gu21dpcl0Osqg+RPTeHNfKcw+NYwpetM1+GirGA qXdDkPinNhGH07GkQTx1BSt282xVq/fwVTk6JOOAkCe7aay+djvqIpwLOfnsb+zEOf MRe4obOkf0QbhU1zwaI/Kiw0J/F1fppxRNkfUtDE= From: Ard Biesheuvel To: linux-crypto@vger.kernel.org Subject: [PATCH 06/12] crypto: sun8i-ss - permit asynchronous skcipher as fallback Date: Thu, 25 Jun 2020 14:42:47 +0200 Message-Id: <20200625124253.1906557-7-ardb@kernel.org> X-Mailer: git-send-email 2.27.0 In-Reply-To: <20200625124253.1906557-1-ardb@kernel.org> References: <20200625124253.1906557-1-ardb@kernel.org> MIME-Version: 1.0 X-Spam-Note: CRM114 invocation failed X-Spam-Score: -5.2 (-----) X-Spam-Report: SpamAssassin version 3.4.4 on merlin.infradead.org summary: Content analysis details: (-5.2 points) pts rule name description ---- ---------------------- -------------------------------------------------- -5.0 RCVD_IN_DNSWL_HI RBL: Sender listed at https://www.dnswl.org/, high trust [198.145.29.99 listed in list.dnswl.org] 0.0 SPF_HELO_NONE SPF: HELO does not publish an SPF Record -0.0 SPF_PASS SPF: sender matches SPF record -0.1 DKIM_VALID Message has at least one valid DKIM or DK signature -0.1 DKIM_VALID_EF Message has a valid DKIM or DK signature from envelope-from domain -0.1 DKIM_VALID_AU Message has a valid DKIM or DK signature from author's domain 0.1 DKIM_SIGNED Message has a DKIM or DK signature, not necessarily valid -0.0 DKIMWL_WL_HIGH DKIMwl.org - Whitelisted High sender X-BeenThere: linux-amlogic@lists.infradead.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Cc: Eric Biggers , Tom Lendacky , Jamie Iles , Herbert Xu , Rohit Maheshwari , Fabio Estevam , Sascha Hauer , Vinay Kumar Yadav , Maxime Ripard , "David S. Miller" , Chen-Yu Tsai , Corentin Labbe , NXP Linux Team , Pengutronix Kernel Team , Ayush Sawal , linux-amlogic@lists.infradead.org, linux-omap@vger.kernel.org, Shawn Guo , Ard Biesheuvel , linux-arm-kernel@lists.infradead.org Sender: "linux-amlogic" Errors-To: linux-amlogic-bounces+patchwork-linux-amlogic=patchwork.kernel.org@lists.infradead.org Even though the sun8i-ss driver implements asynchronous versions of ecb(aes) and cbc(aes), the fallbacks it allocates are required to be synchronous. Given that SIMD based software implementations are usually asynchronous as well, even though they rarely complete asynchronously (this typically only happens in cases where the request was made from softirq context, while SIMD was already in use in the task context that it interrupted), these implementations are disregarded, and either the generic C version or another table based version implemented in assembler is selected instead. Since falling back to synchronous AES is not only a performance issue, but potentially a security issue as well (due to the fact that table based AES is not time invariant), let's fix this, by allocating an ordinary skcipher as the fallback, and invoke it with the completion routine that was given to the outer request. Signed-off-by: Ard Biesheuvel --- drivers/crypto/allwinner/sun8i-ss/sun8i-ss-cipher.c | 39 ++++++++++---------- drivers/crypto/allwinner/sun8i-ss/sun8i-ss.h | 3 +- 2 files changed, 22 insertions(+), 20 deletions(-) diff --git a/drivers/crypto/allwinner/sun8i-ss/sun8i-ss-cipher.c b/drivers/crypto/allwinner/sun8i-ss/sun8i-ss-cipher.c index c89cb2ee2496..7a131675a41c 100644 --- a/drivers/crypto/allwinner/sun8i-ss/sun8i-ss-cipher.c +++ b/drivers/crypto/allwinner/sun8i-ss/sun8i-ss-cipher.c @@ -73,7 +73,6 @@ static int sun8i_ss_cipher_fallback(struct skcipher_request *areq) struct sun8i_cipher_req_ctx *rctx = skcipher_request_ctx(areq); int err; - SYNC_SKCIPHER_REQUEST_ON_STACK(subreq, op->fallback_tfm); #ifdef CONFIG_CRYPTO_DEV_SUN8I_SS_DEBUG struct skcipher_alg *alg = crypto_skcipher_alg(tfm); struct sun8i_ss_alg_template *algt; @@ -81,15 +80,15 @@ static int sun8i_ss_cipher_fallback(struct skcipher_request *areq) algt = container_of(alg, struct sun8i_ss_alg_template, alg.skcipher); algt->stat_fb++; #endif - skcipher_request_set_sync_tfm(subreq, op->fallback_tfm); - skcipher_request_set_callback(subreq, areq->base.flags, NULL, NULL); - skcipher_request_set_crypt(subreq, areq->src, areq->dst, + skcipher_request_set_tfm(&rctx->fallback_req, op->fallback_tfm); + skcipher_request_set_callback(&rctx->fallback_req, areq->base.flags, + areq->base.complete, areq->base.data); + skcipher_request_set_crypt(&rctx->fallback_req, areq->src, areq->dst, areq->cryptlen, areq->iv); if (rctx->op_dir & SS_DECRYPTION) - err = crypto_skcipher_decrypt(subreq); + err = crypto_skcipher_decrypt(&rctx->fallback_req); else - err = crypto_skcipher_encrypt(subreq); - skcipher_request_zero(subreq); + err = crypto_skcipher_encrypt(&rctx->fallback_req); return err; } @@ -334,18 +333,20 @@ int sun8i_ss_cipher_init(struct crypto_tfm *tfm) algt = container_of(alg, struct sun8i_ss_alg_template, alg.skcipher); op->ss = algt->ss; - sktfm->reqsize = sizeof(struct sun8i_cipher_req_ctx); - - op->fallback_tfm = crypto_alloc_sync_skcipher(name, 0, CRYPTO_ALG_NEED_FALLBACK); + op->fallback_tfm = crypto_alloc_skcipher(name, 0, CRYPTO_ALG_NEED_FALLBACK); if (IS_ERR(op->fallback_tfm)) { dev_err(op->ss->dev, "ERROR: Cannot allocate fallback for %s %ld\n", name, PTR_ERR(op->fallback_tfm)); return PTR_ERR(op->fallback_tfm); } + sktfm->reqsize = sizeof(struct sun8i_cipher_req_ctx) + + crypto_skcipher_reqsize(op->fallback_tfm); + + dev_info(op->ss->dev, "Fallback for %s is %s\n", crypto_tfm_alg_driver_name(&sktfm->base), - crypto_tfm_alg_driver_name(crypto_skcipher_tfm(&op->fallback_tfm->base))); + crypto_tfm_alg_driver_name(crypto_skcipher_tfm(op->fallback_tfm))); op->enginectx.op.do_one_request = sun8i_ss_handle_cipher_request; op->enginectx.op.prepare_request = NULL; @@ -359,7 +360,7 @@ int sun8i_ss_cipher_init(struct crypto_tfm *tfm) return 0; error_pm: - crypto_free_sync_skcipher(op->fallback_tfm); + crypto_free_skcipher(op->fallback_tfm); return err; } @@ -371,7 +372,7 @@ void sun8i_ss_cipher_exit(struct crypto_tfm *tfm) memzero_explicit(op->key, op->keylen); kfree(op->key); } - crypto_free_sync_skcipher(op->fallback_tfm); + crypto_free_skcipher(op->fallback_tfm); pm_runtime_put_sync(op->ss->dev); } @@ -401,10 +402,10 @@ int sun8i_ss_aes_setkey(struct crypto_skcipher *tfm, const u8 *key, if (!op->key) return -ENOMEM; - crypto_sync_skcipher_clear_flags(op->fallback_tfm, CRYPTO_TFM_REQ_MASK); - crypto_sync_skcipher_set_flags(op->fallback_tfm, tfm->base.crt_flags & CRYPTO_TFM_REQ_MASK); + crypto_skcipher_clear_flags(op->fallback_tfm, CRYPTO_TFM_REQ_MASK); + crypto_skcipher_set_flags(op->fallback_tfm, tfm->base.crt_flags & CRYPTO_TFM_REQ_MASK); - return crypto_sync_skcipher_setkey(op->fallback_tfm, key, keylen); + return crypto_skcipher_setkey(op->fallback_tfm, key, keylen); } int sun8i_ss_des3_setkey(struct crypto_skcipher *tfm, const u8 *key, @@ -427,8 +428,8 @@ int sun8i_ss_des3_setkey(struct crypto_skcipher *tfm, const u8 *key, if (!op->key) return -ENOMEM; - crypto_sync_skcipher_clear_flags(op->fallback_tfm, CRYPTO_TFM_REQ_MASK); - crypto_sync_skcipher_set_flags(op->fallback_tfm, tfm->base.crt_flags & CRYPTO_TFM_REQ_MASK); + crypto_skcipher_clear_flags(op->fallback_tfm, CRYPTO_TFM_REQ_MASK); + crypto_skcipher_set_flags(op->fallback_tfm, tfm->base.crt_flags & CRYPTO_TFM_REQ_MASK); - return crypto_sync_skcipher_setkey(op->fallback_tfm, key, keylen); + return crypto_skcipher_setkey(op->fallback_tfm, key, keylen); } diff --git a/drivers/crypto/allwinner/sun8i-ss/sun8i-ss.h b/drivers/crypto/allwinner/sun8i-ss/sun8i-ss.h index 29c44f279112..42658b134228 100644 --- a/drivers/crypto/allwinner/sun8i-ss/sun8i-ss.h +++ b/drivers/crypto/allwinner/sun8i-ss/sun8i-ss.h @@ -159,6 +159,7 @@ struct sun8i_cipher_req_ctx { unsigned int ivlen; unsigned int keylen; void *biv; + struct skcipher_request fallback_req; // keep at the end }; /* @@ -174,7 +175,7 @@ struct sun8i_cipher_tfm_ctx { u32 *key; u32 keylen; struct sun8i_ss_dev *ss; - struct crypto_sync_skcipher *fallback_tfm; + struct crypto_skcipher *fallback_tfm; }; /* From patchwork Thu Jun 25 12:42:48 2020 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Ard Biesheuvel X-Patchwork-Id: 11625113 Return-Path: Received: from mail.kernel.org (pdx-korg-mail-1.web.codeaurora.org [172.30.200.123]) by pdx-korg-patchwork-2.web.codeaurora.org (Postfix) with ESMTP id 3811E14B7 for ; Thu, 25 Jun 2020 12:44:26 +0000 (UTC) Received: from merlin.infradead.org (merlin.infradead.org [205.233.59.134]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by mail.kernel.org (Postfix) with ESMTPS id 0FABC20720 for ; Thu, 25 Jun 2020 12:44:26 +0000 (UTC) Authentication-Results: mail.kernel.org; dkim=pass (2048-bit key) header.d=lists.infradead.org header.i=@lists.infradead.org header.b="kRMZaUE0"; dkim=fail reason="signature verification failed" (1024-bit key) header.d=kernel.org header.i=@kernel.org header.b="MmwesvC+" DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org 0FABC20720 Authentication-Results: mail.kernel.org; dmarc=fail (p=none dis=none) header.from=kernel.org Authentication-Results: mail.kernel.org; spf=none smtp.mailfrom=linux-amlogic-bounces+patchwork-linux-amlogic=patchwork.kernel.org@lists.infradead.org DKIM-Signature: v=1; a=rsa-sha256; q=dns/txt; c=relaxed/relaxed; d=lists.infradead.org; s=merlin.20170209; h=Sender:Content-Transfer-Encoding: Content-Type:Cc:List-Subscribe:List-Help:List-Post:List-Archive: List-Unsubscribe:List-Id:MIME-Version:References:In-Reply-To:Message-Id:Date: Subject:To:From:Reply-To:Content-ID:Content-Description:Resent-Date: Resent-From:Resent-Sender:Resent-To:Resent-Cc:Resent-Message-ID:List-Owner; bh=/q/rJTaqrRiEpNsDxralfZQln76f4wM4g0IKplguK38=; b=kRMZaUE0fWYU/PWKqoHlyPdFY 5l/rOxmpmB6ORFMdUfCqoL7pQNkJ8to/PZup6NZpGSxq3vvfaWE422ZjBS8ijkWtTauq5aoHnmLNo xLGnKdeLMmpaDeHx9yVAJiQx45hlwprROFbdLY6932LoMrdIF67utnqOzCviSxq77lWZtUVLueDp6 qw06YDFHiMoJYBTBaeBPixSqlCxv2vseKoQqYDhiikGPZLuyeG5WrLCxeNczeshE0+zfBqAYWVu3B VSkumw3Fq1V7IACT4rLpmnGP+YNd8uPRYJ4ctlh2Dk7zguPemi1MBbJChyFCVz5EAEf4ceiYVU33K QVbXjt5+g==; Received: from localhost ([::1] helo=merlin.infradead.org) by merlin.infradead.org with esmtp (Exim 4.92.3 #3 (Red Hat Linux)) id 1joREi-0005eh-LB; Thu, 25 Jun 2020 12:44:12 +0000 Received: from mail.kernel.org ([198.145.29.99]) by merlin.infradead.org with esmtps (Exim 4.92.3 #3 (Red Hat Linux)) id 1joREJ-0005T8-W7; Thu, 25 Jun 2020 12:43:49 +0000 Received: from localhost.localdomain (lfbn-nic-1-188-42.w2-15.abo.wanadoo.fr [2.15.37.42]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by mail.kernel.org (Postfix) with ESMTPSA id CD956206BE; Thu, 25 Jun 2020 12:43:43 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=kernel.org; s=default; t=1593089027; bh=wGffz0ZawEaorm/Ljcec1lJrVUrfVxGLZbekPHv2K80=; h=From:To:Cc:Subject:Date:In-Reply-To:References:From; b=MmwesvC+7kD10YqfQO/nPFbYUFpC7XeU11gAUcw3j/N/E6Ffqnx/faL0B/oJyiqoq UzW7VOQjL9KDjDU8aCn/pE2m18ZVENw0zuOGBIJ+2sVy4wdjXUdEeX6EQPphXhnaiK YjosleV9OLAFL1GVLhL77uJ8LvDbm00pVMqh6YRk= From: Ard Biesheuvel To: linux-crypto@vger.kernel.org Subject: [PATCH 07/12] crypto: ccp - permit asynchronous skcipher as fallback Date: Thu, 25 Jun 2020 14:42:48 +0200 Message-Id: <20200625124253.1906557-8-ardb@kernel.org> X-Mailer: git-send-email 2.27.0 In-Reply-To: <20200625124253.1906557-1-ardb@kernel.org> References: <20200625124253.1906557-1-ardb@kernel.org> MIME-Version: 1.0 X-Spam-Note: CRM114 invocation failed X-Spam-Score: -5.2 (-----) X-Spam-Report: SpamAssassin version 3.4.4 on merlin.infradead.org summary: Content analysis details: (-5.2 points) pts rule name description ---- ---------------------- -------------------------------------------------- -5.0 RCVD_IN_DNSWL_HI RBL: Sender listed at https://www.dnswl.org/, high trust [198.145.29.99 listed in list.dnswl.org] 0.0 SPF_HELO_NONE SPF: HELO does not publish an SPF Record -0.0 SPF_PASS SPF: sender matches SPF record -0.1 DKIM_VALID Message has at least one valid DKIM or DK signature -0.1 DKIM_VALID_EF Message has a valid DKIM or DK signature from envelope-from domain -0.1 DKIM_VALID_AU Message has a valid DKIM or DK signature from author's domain 0.1 DKIM_SIGNED Message has a DKIM or DK signature, not necessarily valid -0.0 DKIMWL_WL_HIGH DKIMwl.org - Whitelisted High sender X-BeenThere: linux-amlogic@lists.infradead.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Cc: Eric Biggers , Tom Lendacky , Jamie Iles , Herbert Xu , Rohit Maheshwari , Fabio Estevam , Sascha Hauer , Vinay Kumar Yadav , Maxime Ripard , "David S. Miller" , Chen-Yu Tsai , Corentin Labbe , NXP Linux Team , Pengutronix Kernel Team , Ayush Sawal , linux-amlogic@lists.infradead.org, linux-omap@vger.kernel.org, Shawn Guo , Ard Biesheuvel , linux-arm-kernel@lists.infradead.org Sender: "linux-amlogic" Errors-To: linux-amlogic-bounces+patchwork-linux-amlogic=patchwork.kernel.org@lists.infradead.org Even though the ccp driver implements an asynchronous version of xts(aes), the fallback it allocates is required to be synchronous. Given that SIMD based software implementations are usually asynchronous as well, even though they rarely complete asynchronously (this typically only happens in cases where the request was made from softirq context, while SIMD was already in use in the task context that it interrupted), these implementations are disregarded, and either the generic C version or another table based version implemented in assembler is selected instead. Since falling back to synchronous AES is not only a performance issue, but potentially a security issue as well (due to the fact that table based AES is not time invariant), let's fix this, by allocating an ordinary skcipher as the fallback, and invoke it with the completion routine that was given to the outer request. Signed-off-by: Ard Biesheuvel Reported-by: kernel test robot --- drivers/crypto/ccp/ccp-crypto-aes-xts.c | 31 ++++++++++---------- drivers/crypto/ccp/ccp-crypto.h | 4 ++- 2 files changed, 18 insertions(+), 17 deletions(-) diff --git a/drivers/crypto/ccp/ccp-crypto-aes-xts.c b/drivers/crypto/ccp/ccp-crypto-aes-xts.c index 04b2517df955..e0fb4e8f22fb 100644 --- a/drivers/crypto/ccp/ccp-crypto-aes-xts.c +++ b/drivers/crypto/ccp/ccp-crypto-aes-xts.c @@ -98,7 +98,7 @@ static int ccp_aes_xts_setkey(struct crypto_skcipher *tfm, const u8 *key, ctx->u.aes.key_len = key_len / 2; sg_init_one(&ctx->u.aes.key_sg, ctx->u.aes.key, key_len); - return crypto_sync_skcipher_setkey(ctx->u.aes.tfm_skcipher, key, key_len); + return crypto_skcipher_setkey(ctx->u.aes.tfm_skcipher, key, key_len); } static int ccp_aes_xts_crypt(struct skcipher_request *req, @@ -145,20 +145,19 @@ static int ccp_aes_xts_crypt(struct skcipher_request *req, (ctx->u.aes.key_len != AES_KEYSIZE_256)) fallback = 1; if (fallback) { - SYNC_SKCIPHER_REQUEST_ON_STACK(subreq, - ctx->u.aes.tfm_skcipher); - /* Use the fallback to process the request for any * unsupported unit sizes or key sizes */ - skcipher_request_set_sync_tfm(subreq, ctx->u.aes.tfm_skcipher); - skcipher_request_set_callback(subreq, req->base.flags, - NULL, NULL); - skcipher_request_set_crypt(subreq, req->src, req->dst, - req->cryptlen, req->iv); - ret = encrypt ? crypto_skcipher_encrypt(subreq) : - crypto_skcipher_decrypt(subreq); - skcipher_request_zero(subreq); + skcipher_request_set_tfm(&rctx->fallback_req, + ctx->u.aes.tfm_skcipher); + skcipher_request_set_callback(&rctx->fallback_req, + req->base.flags, + req->base.complete, + req->base.data); + skcipher_request_set_crypt(&rctx->fallback_req, req->src, + req->dst, req->cryptlen, req->iv); + ret = encrypt ? crypto_skcipher_encrypt(&rctx->fallback_req) : + crypto_skcipher_decrypt(&rctx->fallback_req); return ret; } @@ -198,13 +197,12 @@ static int ccp_aes_xts_decrypt(struct skcipher_request *req) static int ccp_aes_xts_init_tfm(struct crypto_skcipher *tfm) { struct ccp_ctx *ctx = crypto_skcipher_ctx(tfm); - struct crypto_sync_skcipher *fallback_tfm; + struct crypto_skcipher *fallback_tfm; ctx->complete = ccp_aes_xts_complete; ctx->u.aes.key_len = 0; - fallback_tfm = crypto_alloc_sync_skcipher("xts(aes)", 0, - CRYPTO_ALG_ASYNC | + fallback_tfm = crypto_alloc_skcipher("xts(aes)", 0, CRYPTO_ALG_NEED_FALLBACK); if (IS_ERR(fallback_tfm)) { pr_warn("could not load fallback driver xts(aes)\n"); @@ -212,7 +210,8 @@ static int ccp_aes_xts_init_tfm(struct crypto_skcipher *tfm) } ctx->u.aes.tfm_skcipher = fallback_tfm; - crypto_skcipher_set_reqsize(tfm, sizeof(struct ccp_aes_req_ctx)); + crypto_skcipher_set_reqsize(tfm, sizeof(struct ccp_aes_req_ctx) + + crypto_skcipher_reqsize(fallback_tfm)); return 0; } diff --git a/drivers/crypto/ccp/ccp-crypto.h b/drivers/crypto/ccp/ccp-crypto.h index 90a009e6b5c1..aed3d2192d01 100644 --- a/drivers/crypto/ccp/ccp-crypto.h +++ b/drivers/crypto/ccp/ccp-crypto.h @@ -89,7 +89,7 @@ static inline struct ccp_crypto_ahash_alg * /***** AES related defines *****/ struct ccp_aes_ctx { /* Fallback cipher for XTS with unsupported unit sizes */ - struct crypto_sync_skcipher *tfm_skcipher; + struct crypto_skcipher *tfm_skcipher; enum ccp_engine engine; enum ccp_aes_type type; @@ -121,6 +121,8 @@ struct ccp_aes_req_ctx { u8 rfc3686_iv[AES_BLOCK_SIZE]; struct ccp_cmd cmd; + + struct skcipher_request fallback_req; // keep at the end }; struct ccp_aes_cmac_req_ctx { From patchwork Thu Jun 25 12:42:49 2020 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Ard Biesheuvel X-Patchwork-Id: 11625115 Return-Path: Received: from mail.kernel.org (pdx-korg-mail-1.web.codeaurora.org [172.30.200.123]) by pdx-korg-patchwork-2.web.codeaurora.org (Postfix) with ESMTP id 6B1DC6C1 for ; Thu, 25 Jun 2020 12:44:32 +0000 (UTC) Received: from merlin.infradead.org (merlin.infradead.org [205.233.59.134]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by mail.kernel.org (Postfix) with ESMTPS id 42ECA206BE for ; Thu, 25 Jun 2020 12:44:32 +0000 (UTC) Authentication-Results: mail.kernel.org; dkim=pass (2048-bit key) header.d=lists.infradead.org header.i=@lists.infradead.org header.b="vMU2VeKd"; dkim=fail reason="signature verification failed" (1024-bit key) header.d=kernel.org header.i=@kernel.org header.b="XLJUCThr" DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org 42ECA206BE Authentication-Results: mail.kernel.org; dmarc=fail (p=none dis=none) header.from=kernel.org Authentication-Results: mail.kernel.org; spf=none smtp.mailfrom=linux-amlogic-bounces+patchwork-linux-amlogic=patchwork.kernel.org@lists.infradead.org DKIM-Signature: v=1; a=rsa-sha256; q=dns/txt; c=relaxed/relaxed; d=lists.infradead.org; s=merlin.20170209; h=Sender:Content-Transfer-Encoding: Content-Type:Cc:List-Subscribe:List-Help:List-Post:List-Archive: List-Unsubscribe:List-Id:MIME-Version:References:In-Reply-To:Message-Id:Date: Subject:To:From:Reply-To:Content-ID:Content-Description:Resent-Date: Resent-From:Resent-Sender:Resent-To:Resent-Cc:Resent-Message-ID:List-Owner; bh=/hRmBHFgDkCQJM+xwxxjbQGqRRTfgks+7u8VYYCKrno=; b=vMU2VeKd4nOIGDM5+uzbtnQFw Sc04/M06Bd/+IlK2joXhbfsehZlXUUI9VQY7EHvDu995GxGMUBJDE2wr0hErU7fhPSbNllurVzArv 0KcT08SyWikXA92b1H+c40KLcFB8JUKKLNmEAjEgzZlAdK43wtlxXEPU9AGSl28NHq/tF3u3dwRFh VC6obJJfPqSIcx9TTR7e+uE8dNEfZp4qwHaVwYiNvdeJA5U4ug4j/2GT71V3cpBsyfczeN6EHDZdA nOAow/izB62R23YmZV5mAZ78zHGK3IXtWPG3x9uOwAhPU5/g1RxHZfwUaluX/CsO0bW/pJ1E60Xun +qabhzyYA==; Received: from localhost ([::1] helo=merlin.infradead.org) by merlin.infradead.org with esmtp (Exim 4.92.3 #3 (Red Hat Linux)) id 1joREo-0005hf-DD; Thu, 25 Jun 2020 12:44:18 +0000 Received: from mail.kernel.org ([198.145.29.99]) by merlin.infradead.org with esmtps (Exim 4.92.3 #3 (Red Hat Linux)) id 1joREO-0005V4-63; Thu, 25 Jun 2020 12:43:53 +0000 Received: from localhost.localdomain (lfbn-nic-1-188-42.w2-15.abo.wanadoo.fr [2.15.37.42]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by mail.kernel.org (Postfix) with ESMTPSA id D4FEB207E8; Thu, 25 Jun 2020 12:43:47 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=kernel.org; s=default; t=1593089031; bh=u3/i0Ai3PMvuNylvohLoEd68csg/Is/hfnIaNyDzZf8=; h=From:To:Cc:Subject:Date:In-Reply-To:References:From; b=XLJUCThru6iwLOTic9LzGUq7TrHZm9SwEWCblY4Vq8peu8rT19wXpZmxZITPtlFOm z/ZQySk9C6qkuemwmHUz4/JBD2x+oxVL5o5lTkqOdRhuBjzGaKzZO3WtxD5PzlHLUX f4JizI95+ZbBsTvPz2H6GfYAGTrEueaOa0Znwzvk= From: Ard Biesheuvel To: linux-crypto@vger.kernel.org Subject: [PATCH 08/12] crypto: chelsio - permit asynchronous skcipher as fallback Date: Thu, 25 Jun 2020 14:42:49 +0200 Message-Id: <20200625124253.1906557-9-ardb@kernel.org> X-Mailer: git-send-email 2.27.0 In-Reply-To: <20200625124253.1906557-1-ardb@kernel.org> References: <20200625124253.1906557-1-ardb@kernel.org> MIME-Version: 1.0 X-Spam-Note: CRM114 invocation failed X-Spam-Score: -5.2 (-----) X-Spam-Report: SpamAssassin version 3.4.4 on merlin.infradead.org summary: Content analysis details: (-5.2 points) pts rule name description ---- ---------------------- -------------------------------------------------- -5.0 RCVD_IN_DNSWL_HI RBL: Sender listed at https://www.dnswl.org/, high trust [198.145.29.99 listed in list.dnswl.org] 0.0 SPF_HELO_NONE SPF: HELO does not publish an SPF Record -0.0 SPF_PASS SPF: sender matches SPF record -0.1 DKIM_VALID Message has at least one valid DKIM or DK signature -0.1 DKIM_VALID_EF Message has a valid DKIM or DK signature from envelope-from domain -0.1 DKIM_VALID_AU Message has a valid DKIM or DK signature from author's domain 0.1 DKIM_SIGNED Message has a DKIM or DK signature, not necessarily valid -0.0 DKIMWL_WL_HIGH DKIMwl.org - Whitelisted High sender X-BeenThere: linux-amlogic@lists.infradead.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Cc: Eric Biggers , Tom Lendacky , Jamie Iles , Herbert Xu , Rohit Maheshwari , Fabio Estevam , Sascha Hauer , Vinay Kumar Yadav , Maxime Ripard , "David S. Miller" , Chen-Yu Tsai , Corentin Labbe , NXP Linux Team , Pengutronix Kernel Team , Ayush Sawal , linux-amlogic@lists.infradead.org, linux-omap@vger.kernel.org, Shawn Guo , Ard Biesheuvel , linux-arm-kernel@lists.infradead.org Sender: "linux-amlogic" Errors-To: linux-amlogic-bounces+patchwork-linux-amlogic=patchwork.kernel.org@lists.infradead.org Even though the chelsio driver implements asynchronous versions of cbc(aes) and xts(aes), the fallbacks it allocates are required to be synchronous. Given that SIMD based software implementations are usually asynchronous as well, even though they rarely complete asynchronously (this typically only happens in cases where the request was made from softirq context, while SIMD was already in use in the task context that it interrupted), these implementations are disregarded, and either the generic C version or another table based version implemented in assembler is selected instead. Since falling back to synchronous AES is not only a performance issue, but potentially a security issue as well (due to the fact that table based AES is not time invariant), let's fix this, by allocating an ordinary skcipher as the fallback, and invoke it with the completion routine that was given to the outer request. Signed-off-by: Ard Biesheuvel --- drivers/crypto/chelsio/chcr_algo.c | 57 ++++++++------------ drivers/crypto/chelsio/chcr_crypto.h | 3 +- 2 files changed, 25 insertions(+), 35 deletions(-) diff --git a/drivers/crypto/chelsio/chcr_algo.c b/drivers/crypto/chelsio/chcr_algo.c index 4c2553672b6f..a6625b90fb1a 100644 --- a/drivers/crypto/chelsio/chcr_algo.c +++ b/drivers/crypto/chelsio/chcr_algo.c @@ -690,26 +690,22 @@ static int chcr_sg_ent_in_wr(struct scatterlist *src, return min(srclen, dstlen); } -static int chcr_cipher_fallback(struct crypto_sync_skcipher *cipher, - u32 flags, - struct scatterlist *src, - struct scatterlist *dst, - unsigned int nbytes, +static int chcr_cipher_fallback(struct crypto_skcipher *cipher, + struct skcipher_request *req, u8 *iv, unsigned short op_type) { + struct chcr_skcipher_req_ctx *reqctx = skcipher_request_ctx(req); int err; - SYNC_SKCIPHER_REQUEST_ON_STACK(subreq, cipher); - - skcipher_request_set_sync_tfm(subreq, cipher); - skcipher_request_set_callback(subreq, flags, NULL, NULL); - skcipher_request_set_crypt(subreq, src, dst, - nbytes, iv); + skcipher_request_set_tfm(&reqctx->fallback_req, cipher); + skcipher_request_set_callback(&reqctx->fallback_req, req->base.flags, + req->base.complete, req->base.data); + skcipher_request_set_crypt(&reqctx->fallback_req, req->src, req->dst, + req->cryptlen, iv); - err = op_type ? crypto_skcipher_decrypt(subreq) : - crypto_skcipher_encrypt(subreq); - skcipher_request_zero(subreq); + err = op_type ? crypto_skcipher_decrypt(&reqctx->fallback_req) : + crypto_skcipher_encrypt(&reqctx->fallback_req); return err; @@ -924,11 +920,11 @@ static int chcr_cipher_fallback_setkey(struct crypto_skcipher *cipher, { struct ablk_ctx *ablkctx = ABLK_CTX(c_ctx(cipher)); - crypto_sync_skcipher_clear_flags(ablkctx->sw_cipher, + crypto_skcipher_clear_flags(ablkctx->sw_cipher, CRYPTO_TFM_REQ_MASK); - crypto_sync_skcipher_set_flags(ablkctx->sw_cipher, + crypto_skcipher_set_flags(ablkctx->sw_cipher, cipher->base.crt_flags & CRYPTO_TFM_REQ_MASK); - return crypto_sync_skcipher_setkey(ablkctx->sw_cipher, key, keylen); + return crypto_skcipher_setkey(ablkctx->sw_cipher, key, keylen); } static int chcr_aes_cbc_setkey(struct crypto_skcipher *cipher, @@ -1206,13 +1202,8 @@ static int chcr_handle_cipher_resp(struct skcipher_request *req, req); memcpy(req->iv, reqctx->init_iv, IV); atomic_inc(&adap->chcr_stats.fallback); - err = chcr_cipher_fallback(ablkctx->sw_cipher, - req->base.flags, - req->src, - req->dst, - req->cryptlen, - req->iv, - reqctx->op); + err = chcr_cipher_fallback(ablkctx->sw_cipher, req, req->iv, + reqctx->op); goto complete; } @@ -1341,11 +1332,7 @@ static int process_cipher(struct skcipher_request *req, chcr_cipher_dma_unmap(&ULD_CTX(c_ctx(tfm))->lldi.pdev->dev, req); fallback: atomic_inc(&adap->chcr_stats.fallback); - err = chcr_cipher_fallback(ablkctx->sw_cipher, - req->base.flags, - req->src, - req->dst, - req->cryptlen, + err = chcr_cipher_fallback(ablkctx->sw_cipher, req, subtype == CRYPTO_ALG_SUB_TYPE_CTR_RFC3686 ? reqctx->iv : req->iv, @@ -1486,14 +1473,15 @@ static int chcr_init_tfm(struct crypto_skcipher *tfm) struct chcr_context *ctx = crypto_skcipher_ctx(tfm); struct ablk_ctx *ablkctx = ABLK_CTX(ctx); - ablkctx->sw_cipher = crypto_alloc_sync_skcipher(alg->base.cra_name, 0, + ablkctx->sw_cipher = crypto_alloc_skcipher(alg->base.cra_name, 0, CRYPTO_ALG_NEED_FALLBACK); if (IS_ERR(ablkctx->sw_cipher)) { pr_err("failed to allocate fallback for %s\n", alg->base.cra_name); return PTR_ERR(ablkctx->sw_cipher); } init_completion(&ctx->cbc_aes_aio_done); - crypto_skcipher_set_reqsize(tfm, sizeof(struct chcr_skcipher_req_ctx)); + crypto_skcipher_set_reqsize(tfm, sizeof(struct chcr_skcipher_req_ctx) + + crypto_skcipher_reqsize(ablkctx->sw_cipher)); return chcr_device_init(ctx); } @@ -1507,13 +1495,14 @@ static int chcr_rfc3686_init(struct crypto_skcipher *tfm) /*RFC3686 initialises IV counter value to 1, rfc3686(ctr(aes)) * cannot be used as fallback in chcr_handle_cipher_response */ - ablkctx->sw_cipher = crypto_alloc_sync_skcipher("ctr(aes)", 0, + ablkctx->sw_cipher = crypto_alloc_skcipher("ctr(aes)", 0, CRYPTO_ALG_NEED_FALLBACK); if (IS_ERR(ablkctx->sw_cipher)) { pr_err("failed to allocate fallback for %s\n", alg->base.cra_name); return PTR_ERR(ablkctx->sw_cipher); } - crypto_skcipher_set_reqsize(tfm, sizeof(struct chcr_skcipher_req_ctx)); + crypto_skcipher_set_reqsize(tfm, sizeof(struct chcr_skcipher_req_ctx) + + crypto_skcipher_reqsize(ablkctx->sw_cipher)); return chcr_device_init(ctx); } @@ -1523,7 +1512,7 @@ static void chcr_exit_tfm(struct crypto_skcipher *tfm) struct chcr_context *ctx = crypto_skcipher_ctx(tfm); struct ablk_ctx *ablkctx = ABLK_CTX(ctx); - crypto_free_sync_skcipher(ablkctx->sw_cipher); + crypto_free_skcipher(ablkctx->sw_cipher); } static int get_alg_config(struct algo_param *params, diff --git a/drivers/crypto/chelsio/chcr_crypto.h b/drivers/crypto/chelsio/chcr_crypto.h index 31e427e273f8..e89f9e0094b4 100644 --- a/drivers/crypto/chelsio/chcr_crypto.h +++ b/drivers/crypto/chelsio/chcr_crypto.h @@ -171,7 +171,7 @@ static inline struct chcr_context *h_ctx(struct crypto_ahash *tfm) } struct ablk_ctx { - struct crypto_sync_skcipher *sw_cipher; + struct crypto_skcipher *sw_cipher; __be32 key_ctx_hdr; unsigned int enckey_len; unsigned char ciph_mode; @@ -305,6 +305,7 @@ struct chcr_skcipher_req_ctx { u8 init_iv[CHCR_MAX_CRYPTO_IV_LEN]; u16 txqidx; u16 rxqidx; + struct skcipher_request fallback_req; // keep at the end }; struct chcr_alg_template { From patchwork Thu Jun 25 12:42:50 2020 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Ard Biesheuvel X-Patchwork-Id: 11625117 Return-Path: Received: from mail.kernel.org (pdx-korg-mail-1.web.codeaurora.org [172.30.200.123]) by pdx-korg-patchwork-2.web.codeaurora.org (Postfix) with ESMTP id 6664514B7 for ; Thu, 25 Jun 2020 12:44:46 +0000 (UTC) Received: from merlin.infradead.org (merlin.infradead.org [205.233.59.134]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by mail.kernel.org (Postfix) with ESMTPS id 36EA2206BE for ; Thu, 25 Jun 2020 12:44:46 +0000 (UTC) Authentication-Results: mail.kernel.org; dkim=pass (2048-bit key) header.d=lists.infradead.org header.i=@lists.infradead.org header.b="fpxgTAej"; dkim=fail reason="signature verification failed" (1024-bit key) header.d=kernel.org header.i=@kernel.org header.b="fSa0m+ge" DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org 36EA2206BE Authentication-Results: mail.kernel.org; dmarc=fail (p=none dis=none) header.from=kernel.org Authentication-Results: mail.kernel.org; spf=none smtp.mailfrom=linux-amlogic-bounces+patchwork-linux-amlogic=patchwork.kernel.org@lists.infradead.org DKIM-Signature: v=1; a=rsa-sha256; q=dns/txt; c=relaxed/relaxed; d=lists.infradead.org; s=merlin.20170209; h=Sender:Content-Transfer-Encoding: Content-Type:Cc:List-Subscribe:List-Help:List-Post:List-Archive: List-Unsubscribe:List-Id:MIME-Version:References:In-Reply-To:Message-Id:Date: Subject:To:From:Reply-To:Content-ID:Content-Description:Resent-Date: Resent-From:Resent-Sender:Resent-To:Resent-Cc:Resent-Message-ID:List-Owner; bh=tDacVEroQNuSEc3Kqy0/nQeG2dAuHuhhY76+q7eSj5Q=; b=fpxgTAej+1s5xfantU3p6JQFW aJClbc6Sll47pnXq6kaPXaL+BtYeogAO/6F11syw+ue74bvOCAR47h4JNb3YpoWcau4WmRuIrtYJD px0dsGZlNFCqtmu0JfyTMpUB2f3Ptt7KaUIi1cd/zwvnzJgdgHcrDowN9RCr8o8f/CWZ2ECk/J5uD ADT2xVlXNd3LwHG09TdMqYZNpm+uVjBQ/4C2KXdlE+scDvi+OSGdidMU5RwwqOk7j4aytaMg97USt 49Z12Vc7LoJ3InuPZCOuC16RszuvGqMpmHvBc9Bg2ZwEDs+tTHZUk65mVhH9W2d3Ls1+aNFORSO+d w9mS8l2Fw==; Received: from localhost ([::1] helo=merlin.infradead.org) by merlin.infradead.org with esmtp (Exim 4.92.3 #3 (Red Hat Linux)) id 1joREw-0005ma-10; Thu, 25 Jun 2020 12:44:26 +0000 Received: from mail.kernel.org ([198.145.29.99]) by merlin.infradead.org with esmtps (Exim 4.92.3 #3 (Red Hat Linux)) id 1joRES-0005Wk-6o; Thu, 25 Jun 2020 12:43:57 +0000 Received: from localhost.localdomain (lfbn-nic-1-188-42.w2-15.abo.wanadoo.fr [2.15.37.42]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by mail.kernel.org (Postfix) with ESMTPSA id E006C20724; Thu, 25 Jun 2020 12:43:51 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=kernel.org; s=default; t=1593089035; bh=mJ44RN1Ax7UO5kUEhNQNsOxr9UMAVZrtpGUzPdPjV5w=; h=From:To:Cc:Subject:Date:In-Reply-To:References:From; b=fSa0m+genEQoHZRSnzLuXGzBji6ZPg9bb3YkrnePsj99vW84ULC56q+FgTopfKnI8 vQAU4erv7QryNTjNwiCdGr6HsqA4hb7br2jW+tamvgaWVDhQeQDZ/UHGDBZ4QBy9jq vK+kAK/fLUw4f7Bwx6LRgAXL8G9mU5wNMktVPh3s= From: Ard Biesheuvel To: linux-crypto@vger.kernel.org Subject: [PATCH 09/12] crypto: mxs-dcp - permit asynchronous skcipher as fallback Date: Thu, 25 Jun 2020 14:42:50 +0200 Message-Id: <20200625124253.1906557-10-ardb@kernel.org> X-Mailer: git-send-email 2.27.0 In-Reply-To: <20200625124253.1906557-1-ardb@kernel.org> References: <20200625124253.1906557-1-ardb@kernel.org> MIME-Version: 1.0 X-Spam-Note: CRM114 invocation failed X-Spam-Score: -5.2 (-----) X-Spam-Report: SpamAssassin version 3.4.4 on merlin.infradead.org summary: Content analysis details: (-5.2 points) pts rule name description ---- ---------------------- -------------------------------------------------- -5.0 RCVD_IN_DNSWL_HI RBL: Sender listed at https://www.dnswl.org/, high trust [198.145.29.99 listed in list.dnswl.org] 0.0 SPF_HELO_NONE SPF: HELO does not publish an SPF Record -0.0 SPF_PASS SPF: sender matches SPF record -0.1 DKIM_VALID Message has at least one valid DKIM or DK signature -0.1 DKIM_VALID_EF Message has a valid DKIM or DK signature from envelope-from domain -0.1 DKIM_VALID_AU Message has a valid DKIM or DK signature from author's domain 0.1 DKIM_SIGNED Message has a DKIM or DK signature, not necessarily valid -0.0 DKIMWL_WL_HIGH DKIMwl.org - Whitelisted High sender X-BeenThere: linux-amlogic@lists.infradead.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Cc: Eric Biggers , Tom Lendacky , Jamie Iles , Herbert Xu , Rohit Maheshwari , Fabio Estevam , Sascha Hauer , Vinay Kumar Yadav , Maxime Ripard , "David S. Miller" , Chen-Yu Tsai , Corentin Labbe , NXP Linux Team , Pengutronix Kernel Team , Ayush Sawal , linux-amlogic@lists.infradead.org, linux-omap@vger.kernel.org, Shawn Guo , Ard Biesheuvel , linux-arm-kernel@lists.infradead.org Sender: "linux-amlogic" Errors-To: linux-amlogic-bounces+patchwork-linux-amlogic=patchwork.kernel.org@lists.infradead.org Even though the mxs-dcp driver implements asynchronous versions of ecb(aes) and cbc(aes), the fallbacks it allocates are required to be synchronous. Given that SIMD based software implementations are usually asynchronous as well, even though they rarely complete asynchronously (this typically only happens in cases where the request was made from softirq context, while SIMD was already in use in the task context that it interrupted), these implementations are disregarded, and either the generic C version or another table based version implemented in assembler is selected instead. Since falling back to synchronous AES is not only a performance issue, but potentially a security issue as well (due to the fact that table based AES is not time invariant), let's fix this, by allocating an ordinary skcipher as the fallback, and invoke it with the completion routine that was given to the outer request. Signed-off-by: Ard Biesheuvel --- drivers/crypto/mxs-dcp.c | 33 ++++++++++---------- 1 file changed, 17 insertions(+), 16 deletions(-) diff --git a/drivers/crypto/mxs-dcp.c b/drivers/crypto/mxs-dcp.c index d84530293036..909a7eb748e3 100644 --- a/drivers/crypto/mxs-dcp.c +++ b/drivers/crypto/mxs-dcp.c @@ -97,7 +97,7 @@ struct dcp_async_ctx { unsigned int hot:1; /* Crypto-specific context */ - struct crypto_sync_skcipher *fallback; + struct crypto_skcipher *fallback; unsigned int key_len; uint8_t key[AES_KEYSIZE_128]; }; @@ -105,6 +105,7 @@ struct dcp_async_ctx { struct dcp_aes_req_ctx { unsigned int enc:1; unsigned int ecb:1; + struct skcipher_request fallback_req; // keep at the end }; struct dcp_sha_req_ctx { @@ -426,21 +427,20 @@ static int dcp_chan_thread_aes(void *data) static int mxs_dcp_block_fallback(struct skcipher_request *req, int enc) { struct crypto_skcipher *tfm = crypto_skcipher_reqtfm(req); + struct dcp_aes_req_ctx *rctx = skcipher_request_ctx(req); struct dcp_async_ctx *ctx = crypto_skcipher_ctx(tfm); - SYNC_SKCIPHER_REQUEST_ON_STACK(subreq, ctx->fallback); int ret; - skcipher_request_set_sync_tfm(subreq, ctx->fallback); - skcipher_request_set_callback(subreq, req->base.flags, NULL, NULL); - skcipher_request_set_crypt(subreq, req->src, req->dst, + skcipher_request_set_tfm(&rctx->fallback_req, ctx->fallback); + skcipher_request_set_callback(&rctx->fallback_req, req->base.flags, + req->base.complete, req->base.data); + skcipher_request_set_crypt(&rctx->fallback_req, req->src, req->dst, req->cryptlen, req->iv); if (enc) - ret = crypto_skcipher_encrypt(subreq); + ret = crypto_skcipher_encrypt(&rctx->fallback_req); else - ret = crypto_skcipher_decrypt(subreq); - - skcipher_request_zero(subreq); + ret = crypto_skcipher_decrypt(&rctx->fallback_req); return ret; } @@ -510,24 +510,25 @@ static int mxs_dcp_aes_setkey(struct crypto_skcipher *tfm, const u8 *key, * but is supported by in-kernel software implementation, we use * software fallback. */ - crypto_sync_skcipher_clear_flags(actx->fallback, CRYPTO_TFM_REQ_MASK); - crypto_sync_skcipher_set_flags(actx->fallback, + crypto_skcipher_clear_flags(actx->fallback, CRYPTO_TFM_REQ_MASK); + crypto_skcipher_set_flags(actx->fallback, tfm->base.crt_flags & CRYPTO_TFM_REQ_MASK); - return crypto_sync_skcipher_setkey(actx->fallback, key, len); + return crypto_skcipher_setkey(actx->fallback, key, len); } static int mxs_dcp_aes_fallback_init_tfm(struct crypto_skcipher *tfm) { const char *name = crypto_tfm_alg_name(crypto_skcipher_tfm(tfm)); struct dcp_async_ctx *actx = crypto_skcipher_ctx(tfm); - struct crypto_sync_skcipher *blk; + struct crypto_skcipher *blk; - blk = crypto_alloc_sync_skcipher(name, 0, CRYPTO_ALG_NEED_FALLBACK); + blk = crypto_alloc_skcipher(name, 0, CRYPTO_ALG_NEED_FALLBACK); if (IS_ERR(blk)) return PTR_ERR(blk); actx->fallback = blk; - crypto_skcipher_set_reqsize(tfm, sizeof(struct dcp_aes_req_ctx)); + crypto_skcipher_set_reqsize(tfm, sizeof(struct dcp_aes_req_ctx) + + crypto_skcipher_reqsize(blk)); return 0; } @@ -535,7 +536,7 @@ static void mxs_dcp_aes_fallback_exit_tfm(struct crypto_skcipher *tfm) { struct dcp_async_ctx *actx = crypto_skcipher_ctx(tfm); - crypto_free_sync_skcipher(actx->fallback); + crypto_free_skcipher(actx->fallback); } /* From patchwork Thu Jun 25 12:42:51 2020 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Ard Biesheuvel X-Patchwork-Id: 11625119 Return-Path: Received: from mail.kernel.org (pdx-korg-mail-1.web.codeaurora.org [172.30.200.123]) by pdx-korg-patchwork-2.web.codeaurora.org (Postfix) with ESMTP id CD9D56C1 for ; Thu, 25 Jun 2020 12:45:13 +0000 (UTC) Received: from merlin.infradead.org (merlin.infradead.org [205.233.59.134]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by mail.kernel.org (Postfix) with ESMTPS id A4FF0206BE for ; Thu, 25 Jun 2020 12:45:13 +0000 (UTC) Authentication-Results: mail.kernel.org; dkim=pass (2048-bit key) header.d=lists.infradead.org header.i=@lists.infradead.org header.b="jSGnMBlq"; dkim=fail reason="signature verification failed" (1024-bit key) header.d=kernel.org header.i=@kernel.org header.b="K5OJzBeO" DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org A4FF0206BE Authentication-Results: mail.kernel.org; dmarc=fail (p=none dis=none) header.from=kernel.org Authentication-Results: mail.kernel.org; spf=none smtp.mailfrom=linux-amlogic-bounces+patchwork-linux-amlogic=patchwork.kernel.org@lists.infradead.org DKIM-Signature: v=1; a=rsa-sha256; q=dns/txt; c=relaxed/relaxed; d=lists.infradead.org; s=merlin.20170209; h=Sender:Content-Transfer-Encoding: Content-Type:Cc:List-Subscribe:List-Help:List-Post:List-Archive: List-Unsubscribe:List-Id:MIME-Version:References:In-Reply-To:Message-Id:Date: Subject:To:From:Reply-To:Content-ID:Content-Description:Resent-Date: Resent-From:Resent-Sender:Resent-To:Resent-Cc:Resent-Message-ID:List-Owner; bh=BQCnR2PBnBz7C3qHjugMR0qdyc0fb3ZqoUAEPxd+VbA=; b=jSGnMBlqxKBU7F/OkKyV++mRt m4ThT8WmJf0OvLYo99GLElTiOI23ks+YDbgNH/9e5/7MoyXqwNHesNaJHcfLb8Coz70iyTtYw2cg5 W2YHlqa3ojkEo+ozxVbGwgGIrVuOCYqJ11ahrYM3+tlpC5jrTbIjqwTfAzEvgFWpfX8KlyunVAlrU M+VGclYO0lkRlvYv+VmpsZoG6fAY/l4QOhx7Nkj4S32g184FlNsXxo6EavuB+lw+kMdA5aUdVFs3k dCccKAX9Gw9zdT8+NHslex9uucXeqOlGtv0ZAOP/LLlDlI7L9muQirLER0u5zyEoL9hygG4KZm2KZ d3cYaEJEw==; Received: from localhost ([::1] helo=merlin.infradead.org) by merlin.infradead.org with esmtp (Exim 4.92.3 #3 (Red Hat Linux)) id 1joRFJ-0005yu-IJ; Thu, 25 Jun 2020 12:44:49 +0000 Received: from mail.kernel.org ([198.145.29.99]) by merlin.infradead.org with esmtps (Exim 4.92.3 #3 (Red Hat Linux)) id 1joREW-0005Yi-Pz; Thu, 25 Jun 2020 12:44:01 +0000 Received: from localhost.localdomain (lfbn-nic-1-188-42.w2-15.abo.wanadoo.fr [2.15.37.42]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by mail.kernel.org (Postfix) with ESMTPSA id E74AF206BE; Thu, 25 Jun 2020 12:43:55 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=kernel.org; s=default; t=1593089039; bh=EGar2rBAa9p0EmyWbgXPPJUEijUxQLN5bxRPFqvMCTY=; h=From:To:Cc:Subject:Date:In-Reply-To:References:From; b=K5OJzBeOcSUeDEjh1okj8bcq6cYB3aK/P15BeoB0ExprP8ihrPd4/QiHu2av34haO LGwjzQyxMbSVi+0oONIr1Y9J8CCRC+VnonTr2ZpX+Jx7RCbpw966/wF85BNNI45cba 0AhJHLMGYhWrnW1LyTDga9/iRKcylC4lxBmDwPw4= From: Ard Biesheuvel To: linux-crypto@vger.kernel.org Subject: [PATCH 10/12] crypto: picoxcell - permit asynchronous skcipher as fallback Date: Thu, 25 Jun 2020 14:42:51 +0200 Message-Id: <20200625124253.1906557-11-ardb@kernel.org> X-Mailer: git-send-email 2.27.0 In-Reply-To: <20200625124253.1906557-1-ardb@kernel.org> References: <20200625124253.1906557-1-ardb@kernel.org> MIME-Version: 1.0 X-Spam-Note: CRM114 invocation failed X-Spam-Score: -5.2 (-----) X-Spam-Report: SpamAssassin version 3.4.4 on merlin.infradead.org summary: Content analysis details: (-5.2 points) pts rule name description ---- ---------------------- -------------------------------------------------- -5.0 RCVD_IN_DNSWL_HI RBL: Sender listed at https://www.dnswl.org/, high trust [198.145.29.99 listed in list.dnswl.org] 0.0 SPF_HELO_NONE SPF: HELO does not publish an SPF Record -0.0 SPF_PASS SPF: sender matches SPF record -0.1 DKIM_VALID Message has at least one valid DKIM or DK signature -0.1 DKIM_VALID_EF Message has a valid DKIM or DK signature from envelope-from domain -0.1 DKIM_VALID_AU Message has a valid DKIM or DK signature from author's domain 0.1 DKIM_SIGNED Message has a DKIM or DK signature, not necessarily valid -0.0 DKIMWL_WL_HIGH DKIMwl.org - Whitelisted High sender X-BeenThere: linux-amlogic@lists.infradead.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Cc: Eric Biggers , Tom Lendacky , Jamie Iles , Herbert Xu , Rohit Maheshwari , Fabio Estevam , Sascha Hauer , Vinay Kumar Yadav , Maxime Ripard , "David S. Miller" , Chen-Yu Tsai , Corentin Labbe , NXP Linux Team , Pengutronix Kernel Team , Ayush Sawal , linux-amlogic@lists.infradead.org, linux-omap@vger.kernel.org, Shawn Guo , Ard Biesheuvel , linux-arm-kernel@lists.infradead.org Sender: "linux-amlogic" Errors-To: linux-amlogic-bounces+patchwork-linux-amlogic=patchwork.kernel.org@lists.infradead.org Even though the picoxcell driver implements asynchronous versions of ecb(aes) and cbc(aes), the fallbacks it allocates are required to be synchronous. Given that SIMD based software implementations are usually asynchronous as well, even though they rarely complete asynchronously (this typically only happens in cases where the request was made from softirq context, while SIMD was already in use in the task context that it interrupted), these implementations are disregarded, and either the generic C version or another table based version implemented in assembler is selected instead. Since falling back to synchronous AES is not only a performance issue, but potentially a security issue as well (due to the fact that table based AES is not time invariant), let's fix this, by allocating an ordinary skcipher as the fallback, and invoke it with the completion routine that was given to the outer request. Signed-off-by: Ard Biesheuvel --- drivers/crypto/picoxcell_crypto.c | 34 +++++++++++--------- 1 file changed, 19 insertions(+), 15 deletions(-) diff --git a/drivers/crypto/picoxcell_crypto.c b/drivers/crypto/picoxcell_crypto.c index 7384e91c8b32..eea75c7cbdf2 100644 --- a/drivers/crypto/picoxcell_crypto.c +++ b/drivers/crypto/picoxcell_crypto.c @@ -86,6 +86,7 @@ struct spacc_req { dma_addr_t src_addr, dst_addr; struct spacc_ddt *src_ddt, *dst_ddt; void (*complete)(struct spacc_req *req); + struct skcipher_request fallback_req; // keep at the end }; struct spacc_aead { @@ -158,7 +159,7 @@ struct spacc_ablk_ctx { * The fallback cipher. If the operation can't be done in hardware, * fallback to a software version. */ - struct crypto_sync_skcipher *sw_cipher; + struct crypto_skcipher *sw_cipher; }; /* AEAD cipher context. */ @@ -792,13 +793,13 @@ static int spacc_aes_setkey(struct crypto_skcipher *cipher, const u8 *key, * Set the fallback transform to use the same request flags as * the hardware transform. */ - crypto_sync_skcipher_clear_flags(ctx->sw_cipher, + crypto_skcipher_clear_flags(ctx->sw_cipher, CRYPTO_TFM_REQ_MASK); - crypto_sync_skcipher_set_flags(ctx->sw_cipher, + crypto_skcipher_set_flags(ctx->sw_cipher, cipher->base.crt_flags & CRYPTO_TFM_REQ_MASK); - err = crypto_sync_skcipher_setkey(ctx->sw_cipher, key, len); + err = crypto_skcipher_setkey(ctx->sw_cipher, key, len); if (err) goto sw_setkey_failed; } @@ -900,7 +901,7 @@ static int spacc_ablk_do_fallback(struct skcipher_request *req, struct crypto_tfm *old_tfm = crypto_skcipher_tfm(crypto_skcipher_reqtfm(req)); struct spacc_ablk_ctx *ctx = crypto_tfm_ctx(old_tfm); - SYNC_SKCIPHER_REQUEST_ON_STACK(subreq, ctx->sw_cipher); + struct spacc_req *dev_req = skcipher_request_ctx(req); int err; /* @@ -908,13 +909,13 @@ static int spacc_ablk_do_fallback(struct skcipher_request *req, * the ciphering has completed, put the old transform back into the * request. */ - skcipher_request_set_sync_tfm(subreq, ctx->sw_cipher); - skcipher_request_set_callback(subreq, req->base.flags, NULL, NULL); - skcipher_request_set_crypt(subreq, req->src, req->dst, + skcipher_request_set_tfm(&dev_req->fallback_req, ctx->sw_cipher); + skcipher_request_set_callback(&dev_req->fallback_req, req->base.flags, + req->base.complete, req->base.data); + skcipher_request_set_crypt(&dev_req->fallback_req, req->src, req->dst, req->cryptlen, req->iv); - err = is_encrypt ? crypto_skcipher_encrypt(subreq) : - crypto_skcipher_decrypt(subreq); - skcipher_request_zero(subreq); + err = is_encrypt ? crypto_skcipher_encrypt(&dev_req->fallback_req) : + crypto_skcipher_decrypt(&dev_req->fallback_req); return err; } @@ -1007,19 +1008,22 @@ static int spacc_ablk_init_tfm(struct crypto_skcipher *tfm) ctx->generic.flags = spacc_alg->type; ctx->generic.engine = engine; if (alg->base.cra_flags & CRYPTO_ALG_NEED_FALLBACK) { - ctx->sw_cipher = crypto_alloc_sync_skcipher( + ctx->sw_cipher = crypto_alloc_skcipher( alg->base.cra_name, 0, CRYPTO_ALG_NEED_FALLBACK); if (IS_ERR(ctx->sw_cipher)) { dev_warn(engine->dev, "failed to allocate fallback for %s\n", alg->base.cra_name); return PTR_ERR(ctx->sw_cipher); } + crypto_skcipher_set_reqsize(tfm, sizeof(struct spacc_req) + + crypto_skcipher_reqsize(ctx->sw_cipher)); + } else { + crypto_skcipher_set_reqsize(tfm, sizeof(struct spacc_req)); } + ctx->generic.key_offs = spacc_alg->key_offs; ctx->generic.iv_offs = spacc_alg->iv_offs; - crypto_skcipher_set_reqsize(tfm, sizeof(struct spacc_req)); - return 0; } @@ -1027,7 +1031,7 @@ static void spacc_ablk_exit_tfm(struct crypto_skcipher *tfm) { struct spacc_ablk_ctx *ctx = crypto_skcipher_ctx(tfm); - crypto_free_sync_skcipher(ctx->sw_cipher); + crypto_free_skcipher(ctx->sw_cipher); } static int spacc_ablk_encrypt(struct skcipher_request *req) From patchwork Thu Jun 25 12:42:52 2020 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Ard Biesheuvel X-Patchwork-Id: 11625121 Return-Path: Received: from mail.kernel.org (pdx-korg-mail-1.web.codeaurora.org [172.30.200.123]) by pdx-korg-patchwork-2.web.codeaurora.org (Postfix) with ESMTP id 516F9138C for ; Thu, 25 Jun 2020 12:45:31 +0000 (UTC) Received: from merlin.infradead.org (merlin.infradead.org [205.233.59.134]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by mail.kernel.org (Postfix) with ESMTPS id 2919F206BE for ; Thu, 25 Jun 2020 12:45:31 +0000 (UTC) Authentication-Results: mail.kernel.org; dkim=pass (2048-bit key) header.d=lists.infradead.org header.i=@lists.infradead.org header.b="bJ8P0rK7"; dkim=fail reason="signature verification failed" (1024-bit key) header.d=kernel.org header.i=@kernel.org header.b="HcJIPIhJ" DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org 2919F206BE Authentication-Results: mail.kernel.org; dmarc=fail (p=none dis=none) header.from=kernel.org Authentication-Results: mail.kernel.org; spf=none smtp.mailfrom=linux-amlogic-bounces+patchwork-linux-amlogic=patchwork.kernel.org@lists.infradead.org DKIM-Signature: v=1; a=rsa-sha256; q=dns/txt; c=relaxed/relaxed; d=lists.infradead.org; s=merlin.20170209; h=Sender:Content-Transfer-Encoding: Content-Type:Cc:List-Subscribe:List-Help:List-Post:List-Archive: List-Unsubscribe:List-Id:MIME-Version:References:In-Reply-To:Message-Id:Date: Subject:To:From:Reply-To:Content-ID:Content-Description:Resent-Date: Resent-From:Resent-Sender:Resent-To:Resent-Cc:Resent-Message-ID:List-Owner; bh=oXshhcDDipPVdKGC48gpxaKbCby41P/r5z9MRaRtstg=; b=bJ8P0rK7crECCnRFwhGHCKuL7 qcFHE4fwSv9NA0DCP9oN9TedE5fDDbYA/omtS8Fagj4dVbn+i9KbF5ptfaS4hSd8+aAB0GRk151AM gSIDO9QKM09vhQLCsXzd0yHaFJTy3w6k2sMQU5mrC0lGT8N3REg7aXeRyH8timh9Ys9iNDxyKgT++ EGtUzifmeAh8h8mRH8yg+2aIOrpdNBnPdTQOIj5N87aM4g4PEcXeYs526GkZP6gIxW5696DjuTrbi tDLmRMyQPifUYs6UDXs9dY4054M8c4QRL7vkdTc6EnyOVCe1uepp1Nv9c1Rkw7XKspK8oc1TPrXc+ ENqPoiRHQ==; Received: from localhost ([::1] helo=merlin.infradead.org) by merlin.infradead.org with esmtp (Exim 4.92.3 #3 (Red Hat Linux)) id 1joRFi-0006Bw-Bw; Thu, 25 Jun 2020 12:45:14 +0000 Received: from mail.kernel.org ([198.145.29.99]) by merlin.infradead.org with esmtps (Exim 4.92.3 #3 (Red Hat Linux)) id 1joREa-0005ad-5t; Thu, 25 Jun 2020 12:44:05 +0000 Received: from localhost.localdomain (lfbn-nic-1-188-42.w2-15.abo.wanadoo.fr [2.15.37.42]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by mail.kernel.org (Postfix) with ESMTPSA id EF39B20720; Thu, 25 Jun 2020 12:43:59 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=kernel.org; s=default; t=1593089043; bh=SZYJbEe8mCowaOS/Sy7cfSLq4/DGTZWmD8C72wZYJ5s=; h=From:To:Cc:Subject:Date:In-Reply-To:References:From; b=HcJIPIhJS07N6knm3q/5jLrLdS+15P61eJyqq1jEhx3E47sdvDf7gPQJufxaGoW5A 2UYcdrHwnhK7jwXrITIPjUTaDx815+QV4vd2gemXmMDF/lzKwitNEwMv8SRIKyyZlC Pt+3hk096uXRabjFRZ0jk0Rp3BV5VqCsB+qjtDg0= From: Ard Biesheuvel To: linux-crypto@vger.kernel.org Subject: [PATCH 11/12] crypto: qce - permit asynchronous skcipher as fallback Date: Thu, 25 Jun 2020 14:42:52 +0200 Message-Id: <20200625124253.1906557-12-ardb@kernel.org> X-Mailer: git-send-email 2.27.0 In-Reply-To: <20200625124253.1906557-1-ardb@kernel.org> References: <20200625124253.1906557-1-ardb@kernel.org> MIME-Version: 1.0 X-Spam-Note: CRM114 invocation failed X-Spam-Score: -5.2 (-----) X-Spam-Report: SpamAssassin version 3.4.4 on merlin.infradead.org summary: Content analysis details: (-5.2 points) pts rule name description ---- ---------------------- -------------------------------------------------- -5.0 RCVD_IN_DNSWL_HI RBL: Sender listed at https://www.dnswl.org/, high trust [198.145.29.99 listed in list.dnswl.org] 0.0 SPF_HELO_NONE SPF: HELO does not publish an SPF Record -0.0 SPF_PASS SPF: sender matches SPF record -0.1 DKIM_VALID Message has at least one valid DKIM or DK signature -0.1 DKIM_VALID_EF Message has a valid DKIM or DK signature from envelope-from domain -0.1 DKIM_VALID_AU Message has a valid DKIM or DK signature from author's domain 0.1 DKIM_SIGNED Message has a DKIM or DK signature, not necessarily valid -0.0 DKIMWL_WL_HIGH DKIMwl.org - Whitelisted High sender X-BeenThere: linux-amlogic@lists.infradead.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Cc: Eric Biggers , Tom Lendacky , Jamie Iles , Herbert Xu , Rohit Maheshwari , Fabio Estevam , Sascha Hauer , Vinay Kumar Yadav , Maxime Ripard , "David S. Miller" , Chen-Yu Tsai , Corentin Labbe , NXP Linux Team , Pengutronix Kernel Team , Ayush Sawal , linux-amlogic@lists.infradead.org, linux-omap@vger.kernel.org, Shawn Guo , Ard Biesheuvel , linux-arm-kernel@lists.infradead.org Sender: "linux-amlogic" Errors-To: linux-amlogic-bounces+patchwork-linux-amlogic=patchwork.kernel.org@lists.infradead.org Even though the qce driver implements asynchronous versions of ecb(aes), cbc(aes)and xts(aes), the fallbacks it allocates are required to be synchronous. Given that SIMD based software implementations are usually asynchronous as well, even though they rarely complete asynchronously (this typically only happens in cases where the request was made from softirq context, while SIMD was already in use in the task context that it interrupted), these implementations are disregarded, and either the generic C version or another table based version implemented in assembler is selected instead. Since falling back to synchronous AES is not only a performance issue, but potentially a security issue as well (due to the fact that table based AES is not time invariant), let's fix this, by allocating an ordinary skcipher as the fallback, and invoke it with the completion routine that was given to the outer request. Signed-off-by: Ard Biesheuvel --- drivers/crypto/qce/cipher.h | 3 ++- drivers/crypto/qce/skcipher.c | 27 ++++++++++---------- 2 files changed, 16 insertions(+), 14 deletions(-) diff --git a/drivers/crypto/qce/cipher.h b/drivers/crypto/qce/cipher.h index 7770660bc853..cffa9fc628ff 100644 --- a/drivers/crypto/qce/cipher.h +++ b/drivers/crypto/qce/cipher.h @@ -14,7 +14,7 @@ struct qce_cipher_ctx { u8 enc_key[QCE_MAX_KEY_SIZE]; unsigned int enc_keylen; - struct crypto_sync_skcipher *fallback; + struct crypto_skcipher *fallback; }; /** @@ -43,6 +43,7 @@ struct qce_cipher_reqctx { struct sg_table src_tbl; struct scatterlist *src_sg; unsigned int cryptlen; + struct skcipher_request fallback_req; // keep at the end }; static inline struct qce_alg_template *to_cipher_tmpl(struct crypto_skcipher *tfm) diff --git a/drivers/crypto/qce/skcipher.c b/drivers/crypto/qce/skcipher.c index 9412433f3b21..265afae29901 100644 --- a/drivers/crypto/qce/skcipher.c +++ b/drivers/crypto/qce/skcipher.c @@ -178,7 +178,7 @@ static int qce_skcipher_setkey(struct crypto_skcipher *ablk, const u8 *key, break; } - ret = crypto_sync_skcipher_setkey(ctx->fallback, key, keylen); + ret = crypto_skcipher_setkey(ctx->fallback, key, keylen); if (!ret) ctx->enc_keylen = keylen; return ret; @@ -235,16 +235,15 @@ static int qce_skcipher_crypt(struct skcipher_request *req, int encrypt) req->cryptlen <= aes_sw_max_len) || (IS_XTS(rctx->flags) && req->cryptlen > QCE_SECTOR_SIZE && req->cryptlen % QCE_SECTOR_SIZE))) { - SYNC_SKCIPHER_REQUEST_ON_STACK(subreq, ctx->fallback); - - skcipher_request_set_sync_tfm(subreq, ctx->fallback); - skcipher_request_set_callback(subreq, req->base.flags, - NULL, NULL); - skcipher_request_set_crypt(subreq, req->src, req->dst, - req->cryptlen, req->iv); - ret = encrypt ? crypto_skcipher_encrypt(subreq) : - crypto_skcipher_decrypt(subreq); - skcipher_request_zero(subreq); + skcipher_request_set_tfm(&rctx->fallback_req, ctx->fallback); + skcipher_request_set_callback(&rctx->fallback_req, + req->base.flags, + req->base.complete, + req->base.data); + skcipher_request_set_crypt(&rctx->fallback_req, req->src, + req->dst, req->cryptlen, req->iv); + ret = encrypt ? crypto_skcipher_encrypt(&rctx->fallback_req) : + crypto_skcipher_decrypt(&rctx->fallback_req); return ret; } @@ -275,8 +274,10 @@ static int qce_skcipher_init_fallback(struct crypto_skcipher *tfm) struct qce_cipher_ctx *ctx = crypto_skcipher_ctx(tfm); qce_skcipher_init(tfm); - ctx->fallback = crypto_alloc_sync_skcipher(crypto_tfm_alg_name(&tfm->base), + ctx->fallback = crypto_alloc_skcipher(crypto_tfm_alg_name(&tfm->base), 0, CRYPTO_ALG_NEED_FALLBACK); + crypto_skcipher_set_reqsize(tfm, sizeof(struct qce_cipher_reqctx) + + crypto_skcipher_reqsize(ctx->fallback)); return PTR_ERR_OR_ZERO(ctx->fallback); } @@ -284,7 +285,7 @@ static void qce_skcipher_exit(struct crypto_skcipher *tfm) { struct qce_cipher_ctx *ctx = crypto_skcipher_ctx(tfm); - crypto_free_sync_skcipher(ctx->fallback); + crypto_free_skcipher(ctx->fallback); } struct qce_skcipher_def { From patchwork Thu Jun 25 12:42:53 2020 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Ard Biesheuvel X-Patchwork-Id: 11625125 Return-Path: Received: from mail.kernel.org (pdx-korg-mail-1.web.codeaurora.org [172.30.200.123]) by pdx-korg-patchwork-2.web.codeaurora.org (Postfix) with ESMTP id 85B4B14E3 for ; Thu, 25 Jun 2020 12:45:41 +0000 (UTC) Received: from merlin.infradead.org (merlin.infradead.org [205.233.59.134]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by mail.kernel.org (Postfix) with ESMTPS id 5D51A206BE for ; Thu, 25 Jun 2020 12:45:41 +0000 (UTC) Authentication-Results: mail.kernel.org; dkim=pass (2048-bit key) header.d=lists.infradead.org header.i=@lists.infradead.org header.b="WQGyWwMR"; dkim=fail reason="signature verification failed" (1024-bit key) header.d=kernel.org header.i=@kernel.org header.b="EtSqpxr5" DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org 5D51A206BE Authentication-Results: mail.kernel.org; dmarc=fail (p=none dis=none) header.from=kernel.org Authentication-Results: mail.kernel.org; spf=none smtp.mailfrom=linux-amlogic-bounces+patchwork-linux-amlogic=patchwork.kernel.org@lists.infradead.org DKIM-Signature: v=1; a=rsa-sha256; q=dns/txt; c=relaxed/relaxed; d=lists.infradead.org; s=merlin.20170209; h=Sender:Content-Transfer-Encoding: Content-Type:Cc:List-Subscribe:List-Help:List-Post:List-Archive: List-Unsubscribe:List-Id:MIME-Version:References:In-Reply-To:Message-Id:Date: Subject:To:From:Reply-To:Content-ID:Content-Description:Resent-Date: Resent-From:Resent-Sender:Resent-To:Resent-Cc:Resent-Message-ID:List-Owner; bh=UB+73g+JsXOvERiNz7gKqfrKEPiNbmRPqQT55pLpCtE=; b=WQGyWwMRH9sBtQj+aIBz7f+yH vSK1xM4CquDNQrv+U17E42vUxscv9ylJcO8nM9Zv4DtQGks4VZZQfpld44LOgK3uZI0Pa/8ujnpUl 8A8hyrwo9PZ19ifkfvdKAhlwXuPCJGGFiJA2n4HUBKyrSTj3938VX+U6ayrcI5Nro9ZKM7BX4vCmu UJkT+6+j6heeY5amKimlND6Lcfk3rRlJVyIaDHPz9frYkcVZv5pkUPCGUBPcRVxN5lSjlmHG0/gGy FfE2J1TkwV0AB5MNUVqyzGaYqlnp54Jyrm8lQJaUYcvlhahl4GW0RD75eI1CkaCqpMDKeUFsaekLj DXYOVJAyA==; Received: from localhost ([::1] helo=merlin.infradead.org) by merlin.infradead.org with esmtp (Exim 4.92.3 #3 (Red Hat Linux)) id 1joRFr-0006IZ-I0; Thu, 25 Jun 2020 12:45:23 +0000 Received: from mail.kernel.org ([198.145.29.99]) by merlin.infradead.org with esmtps (Exim 4.92.3 #3 (Red Hat Linux)) id 1joREe-0005cO-C7; Thu, 25 Jun 2020 12:44:09 +0000 Received: from localhost.localdomain (lfbn-nic-1-188-42.w2-15.abo.wanadoo.fr [2.15.37.42]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by mail.kernel.org (Postfix) with ESMTPSA id 03531206BE; Thu, 25 Jun 2020 12:44:03 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=kernel.org; s=default; t=1593089047; bh=1mKbpa4I4Sh07AGQPoHxLj1RNd0nCGZ8qWt4DEk3uDw=; h=From:To:Cc:Subject:Date:In-Reply-To:References:From; b=EtSqpxr5/UQ9Ntco1uUPIcviLv6ya5Mu71Nnggl7PMTeIeA0sjbtLYU14mpmVEN2o EbvVdfOb4JRPemi3FcNe2EsCV3/iZIaUYEiLW3bo1H/2x3PuvrMrV5O80eNtBL46o7 zWfXGJzR0Od3k+p6GdnIBmGt1X8YSZHg5+uk1HDo= From: Ard Biesheuvel To: linux-crypto@vger.kernel.org Subject: [PATCH 12/12] crypto: sahara - permit asynchronous skcipher as fallback Date: Thu, 25 Jun 2020 14:42:53 +0200 Message-Id: <20200625124253.1906557-13-ardb@kernel.org> X-Mailer: git-send-email 2.27.0 In-Reply-To: <20200625124253.1906557-1-ardb@kernel.org> References: <20200625124253.1906557-1-ardb@kernel.org> MIME-Version: 1.0 X-Spam-Note: CRM114 invocation failed X-Spam-Score: -5.2 (-----) X-Spam-Report: SpamAssassin version 3.4.4 on merlin.infradead.org summary: Content analysis details: (-5.2 points) pts rule name description ---- ---------------------- -------------------------------------------------- -5.0 RCVD_IN_DNSWL_HI RBL: Sender listed at https://www.dnswl.org/, high trust [198.145.29.99 listed in list.dnswl.org] 0.0 SPF_HELO_NONE SPF: HELO does not publish an SPF Record -0.0 SPF_PASS SPF: sender matches SPF record -0.1 DKIM_VALID Message has at least one valid DKIM or DK signature -0.1 DKIM_VALID_EF Message has a valid DKIM or DK signature from envelope-from domain -0.1 DKIM_VALID_AU Message has a valid DKIM or DK signature from author's domain 0.1 DKIM_SIGNED Message has a DKIM or DK signature, not necessarily valid -0.0 DKIMWL_WL_HIGH DKIMwl.org - Whitelisted High sender X-BeenThere: linux-amlogic@lists.infradead.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Cc: Eric Biggers , Tom Lendacky , Jamie Iles , Herbert Xu , Rohit Maheshwari , Fabio Estevam , Sascha Hauer , Vinay Kumar Yadav , Maxime Ripard , "David S. Miller" , Chen-Yu Tsai , Corentin Labbe , NXP Linux Team , Pengutronix Kernel Team , Ayush Sawal , linux-amlogic@lists.infradead.org, linux-omap@vger.kernel.org, Shawn Guo , Ard Biesheuvel , linux-arm-kernel@lists.infradead.org Sender: "linux-amlogic" Errors-To: linux-amlogic-bounces+patchwork-linux-amlogic=patchwork.kernel.org@lists.infradead.org Even though the sahara driver implements asynchronous versions of ecb(aes) and cbc(aes), the fallbacks it allocates are required to be synchronous. Given that SIMD based software implementations are usually asynchronous as well, even though they rarely complete asynchronously (this typically only happens in cases where the request was made from softirq context, while SIMD was already in use in the task context that it interrupted), these implementations are disregarded, and either the generic C version or another table based version implemented in assembler is selected instead. Since falling back to synchronous AES is not only a performance issue, but potentially a security issue as well (due to the fact that table based AES is not time invariant), let's fix this, by allocating an ordinary skcipher as the fallback, and invoke it with the completion routine that was given to the outer request. Signed-off-by: Ard Biesheuvel --- drivers/crypto/sahara.c | 96 +++++++++----------- 1 file changed, 45 insertions(+), 51 deletions(-) diff --git a/drivers/crypto/sahara.c b/drivers/crypto/sahara.c index 466e30bd529c..0c8cb23ae708 100644 --- a/drivers/crypto/sahara.c +++ b/drivers/crypto/sahara.c @@ -146,11 +146,12 @@ struct sahara_ctx { /* AES-specific context */ int keylen; u8 key[AES_KEYSIZE_128]; - struct crypto_sync_skcipher *fallback; + struct crypto_skcipher *fallback; }; struct sahara_aes_reqctx { unsigned long mode; + struct skcipher_request fallback_req; // keep at the end }; /* @@ -617,10 +618,10 @@ static int sahara_aes_setkey(struct crypto_skcipher *tfm, const u8 *key, /* * The requested key size is not supported by HW, do a fallback. */ - crypto_sync_skcipher_clear_flags(ctx->fallback, CRYPTO_TFM_REQ_MASK); - crypto_sync_skcipher_set_flags(ctx->fallback, tfm->base.crt_flags & + crypto_skcipher_clear_flags(ctx->fallback, CRYPTO_TFM_REQ_MASK); + crypto_skcipher_set_flags(ctx->fallback, tfm->base.crt_flags & CRYPTO_TFM_REQ_MASK); - return crypto_sync_skcipher_setkey(ctx->fallback, key, keylen); + return crypto_skcipher_setkey(ctx->fallback, key, keylen); } static int sahara_aes_crypt(struct skcipher_request *req, unsigned long mode) @@ -651,21 +652,19 @@ static int sahara_aes_crypt(struct skcipher_request *req, unsigned long mode) static int sahara_aes_ecb_encrypt(struct skcipher_request *req) { + struct sahara_aes_reqctx *rctx = skcipher_request_ctx(req); struct sahara_ctx *ctx = crypto_skcipher_ctx( crypto_skcipher_reqtfm(req)); - int err; if (unlikely(ctx->keylen != AES_KEYSIZE_128)) { - SYNC_SKCIPHER_REQUEST_ON_STACK(subreq, ctx->fallback); - - skcipher_request_set_sync_tfm(subreq, ctx->fallback); - skcipher_request_set_callback(subreq, req->base.flags, - NULL, NULL); - skcipher_request_set_crypt(subreq, req->src, req->dst, - req->cryptlen, req->iv); - err = crypto_skcipher_encrypt(subreq); - skcipher_request_zero(subreq); - return err; + skcipher_request_set_tfm(&rctx->fallback_req, ctx->fallback); + skcipher_request_set_callback(&rctx->fallback_req, + req->base.flags, + req->base.complete, + req->base.data); + skcipher_request_set_crypt(&rctx->fallback_req, req->src, + req->dst, req->cryptlen, req->iv); + return crypto_skcipher_encrypt(&rctx->fallback_req); } return sahara_aes_crypt(req, FLAGS_ENCRYPT); @@ -673,21 +672,19 @@ static int sahara_aes_ecb_encrypt(struct skcipher_request *req) static int sahara_aes_ecb_decrypt(struct skcipher_request *req) { + struct sahara_aes_reqctx *rctx = skcipher_request_ctx(req); struct sahara_ctx *ctx = crypto_skcipher_ctx( crypto_skcipher_reqtfm(req)); - int err; if (unlikely(ctx->keylen != AES_KEYSIZE_128)) { - SYNC_SKCIPHER_REQUEST_ON_STACK(subreq, ctx->fallback); - - skcipher_request_set_sync_tfm(subreq, ctx->fallback); - skcipher_request_set_callback(subreq, req->base.flags, - NULL, NULL); - skcipher_request_set_crypt(subreq, req->src, req->dst, - req->cryptlen, req->iv); - err = crypto_skcipher_decrypt(subreq); - skcipher_request_zero(subreq); - return err; + skcipher_request_set_tfm(&rctx->fallback_req, ctx->fallback); + skcipher_request_set_callback(&rctx->fallback_req, + req->base.flags, + req->base.complete, + req->base.data); + skcipher_request_set_crypt(&rctx->fallback_req, req->src, + req->dst, req->cryptlen, req->iv); + return crypto_skcipher_decrypt(&rctx->fallback_req); } return sahara_aes_crypt(req, 0); @@ -695,21 +692,19 @@ static int sahara_aes_ecb_decrypt(struct skcipher_request *req) static int sahara_aes_cbc_encrypt(struct skcipher_request *req) { + struct sahara_aes_reqctx *rctx = skcipher_request_ctx(req); struct sahara_ctx *ctx = crypto_skcipher_ctx( crypto_skcipher_reqtfm(req)); - int err; if (unlikely(ctx->keylen != AES_KEYSIZE_128)) { - SYNC_SKCIPHER_REQUEST_ON_STACK(subreq, ctx->fallback); - - skcipher_request_set_sync_tfm(subreq, ctx->fallback); - skcipher_request_set_callback(subreq, req->base.flags, - NULL, NULL); - skcipher_request_set_crypt(subreq, req->src, req->dst, - req->cryptlen, req->iv); - err = crypto_skcipher_encrypt(subreq); - skcipher_request_zero(subreq); - return err; + skcipher_request_set_tfm(&rctx->fallback_req, ctx->fallback); + skcipher_request_set_callback(&rctx->fallback_req, + req->base.flags, + req->base.complete, + req->base.data); + skcipher_request_set_crypt(&rctx->fallback_req, req->src, + req->dst, req->cryptlen, req->iv); + return crypto_skcipher_encrypt(&rctx->fallback_req); } return sahara_aes_crypt(req, FLAGS_ENCRYPT | FLAGS_CBC); @@ -717,21 +712,19 @@ static int sahara_aes_cbc_encrypt(struct skcipher_request *req) static int sahara_aes_cbc_decrypt(struct skcipher_request *req) { + struct sahara_aes_reqctx *rctx = skcipher_request_ctx(req); struct sahara_ctx *ctx = crypto_skcipher_ctx( crypto_skcipher_reqtfm(req)); - int err; if (unlikely(ctx->keylen != AES_KEYSIZE_128)) { - SYNC_SKCIPHER_REQUEST_ON_STACK(subreq, ctx->fallback); - - skcipher_request_set_sync_tfm(subreq, ctx->fallback); - skcipher_request_set_callback(subreq, req->base.flags, - NULL, NULL); - skcipher_request_set_crypt(subreq, req->src, req->dst, - req->cryptlen, req->iv); - err = crypto_skcipher_decrypt(subreq); - skcipher_request_zero(subreq); - return err; + skcipher_request_set_tfm(&rctx->fallback_req, ctx->fallback); + skcipher_request_set_callback(&rctx->fallback_req, + req->base.flags, + req->base.complete, + req->base.data); + skcipher_request_set_crypt(&rctx->fallback_req, req->src, + req->dst, req->cryptlen, req->iv); + return crypto_skcipher_decrypt(&rctx->fallback_req); } return sahara_aes_crypt(req, FLAGS_CBC); @@ -742,14 +735,15 @@ static int sahara_aes_init_tfm(struct crypto_skcipher *tfm) const char *name = crypto_tfm_alg_name(&tfm->base); struct sahara_ctx *ctx = crypto_skcipher_ctx(tfm); - ctx->fallback = crypto_alloc_sync_skcipher(name, 0, + ctx->fallback = crypto_alloc_skcipher(name, 0, CRYPTO_ALG_NEED_FALLBACK); if (IS_ERR(ctx->fallback)) { pr_err("Error allocating fallback algo %s\n", name); return PTR_ERR(ctx->fallback); } - crypto_skcipher_set_reqsize(tfm, sizeof(struct sahara_aes_reqctx)); + crypto_skcipher_set_reqsize(tfm, sizeof(struct sahara_aes_reqctx) + + crypto_skcipher_reqsize(ctx->fallback)); return 0; } @@ -758,7 +752,7 @@ static void sahara_aes_exit_tfm(struct crypto_skcipher *tfm) { struct sahara_ctx *ctx = crypto_skcipher_ctx(tfm); - crypto_free_sync_skcipher(ctx->fallback); + crypto_free_skcipher(ctx->fallback); } static u32 sahara_sha_init_hdr(struct sahara_dev *dev,