From patchwork Tue Jan 29 08:00:29 2019 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Eric Biggers X-Patchwork-Id: 10785643 Return-Path: Received: from mail.wl.linuxfoundation.org (pdx-wl-mail.web.codeaurora.org [172.30.200.125]) by pdx-korg-patchwork-2.web.codeaurora.org (Postfix) with ESMTP id 750AD1390 for ; Tue, 29 Jan 2019 08:02:26 +0000 (UTC) Received: from mail.wl.linuxfoundation.org (localhost [127.0.0.1]) by mail.wl.linuxfoundation.org (Postfix) with ESMTP id 622A82B7A2 for ; Tue, 29 Jan 2019 08:02:26 +0000 (UTC) Received: by mail.wl.linuxfoundation.org (Postfix, from userid 486) id 558CA2B92E; Tue, 29 Jan 2019 08:02:26 +0000 (UTC) X-Spam-Checker-Version: SpamAssassin 3.3.1 (2010-03-16) on pdx-wl-mail.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-5.2 required=2.0 tests=BAYES_00,DKIM_SIGNED, DKIM_VALID,MAILING_LIST_MULTI,RCVD_IN_DNSWL_MED autolearn=unavailable version=3.3.1 Received: from bombadil.infradead.org (bombadil.infradead.org [198.137.202.133]) (using TLSv1.2 with cipher AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by mail.wl.linuxfoundation.org (Postfix) with ESMTPS id F26CF2B7A2 for ; Tue, 29 Jan 2019 08:02:24 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; q=dns/txt; c=relaxed/relaxed; d=lists.infradead.org; s=bombadil.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=TGvDkSQOzlxS3O+g+3NpVBBt5g2BjQVrfoD0BN2NOlg=; b=a1mrorBdwQkQ+Y yM4dYSv+SlOlb0Jy90rHyx1zwuWihcoemjoxxajeUHanbWiziVjXtZHtsQphppqeJwJ8xQG+ciaIh 10/IeP+PCEC1T1eJbwwfdxTq3q7NMwxrfl4Mka3sGX7clqTl5AwvCY05Bj+3GAarC6PYRpHirKmrs v2s0hY9g2/CFRnOQ0RHrPxJRK6gvZRfHZHXchcvaUNPP28Zu/Tod3gsFeXBMaIEd8uRfDsjQsEKAO GZqmcNyDBm1rAalSnk9M7ApeITlVnqChLxsJ8siVSn89WegSxlkOpuhG9AiA+YYecAsnNtqPeMZfu YhjS8Xw3by8BPoRfIyyQ==; Received: from localhost ([127.0.0.1] helo=bombadil.infradead.org) by bombadil.infradead.org with esmtp (Exim 4.90_1 #2 (Red Hat Linux)) id 1goOLe-0000q9-68; Tue, 29 Jan 2019 08:02:22 +0000 Received: from mail.kernel.org ([198.145.29.99]) by bombadil.infradead.org with esmtps (Exim 4.90_1 #2 (Red Hat Linux)) id 1goOL7-0000Re-Am for linux-arm-kernel@lists.infradead.org; Tue, 29 Jan 2019 08:01:56 +0000 Received: from sol.localdomain (c-107-3-167-184.hsd1.ca.comcast.net [107.3.167.184]) (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 BEC082183F; Tue, 29 Jan 2019 08:01:48 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=kernel.org; s=default; t=1548748909; bh=Z2tQVCABypWRDGS9T35xbEdpUE9vTFCYfonuHgZ7zj4=; h=From:To:Cc:Subject:Date:In-Reply-To:References:From; b=PCZQRRD6tEiO6ikMD6EZYLcCtceK+DNOTXECF609EbXP0c3Cn0+GtdXgRfB1CYWkC 7SNAbtbGS/rT/tmIEFwbt+QyVbb0TQVHhC3cWKqbv8+tpGKl/JQvnGDVN22UlnOWTr mABni1koX4HXMC1lmGrCvvNPaQ/Y8SRMnQ+0BHVg= From: Eric Biggers To: linux-crypto@vger.kernel.org, Herbert Xu Subject: [PATCH v2 3/3] crypto: arm64/crct10dif-ce - cleanup and optimizations Date: Tue, 29 Jan 2019 00:00:29 -0800 Message-Id: <20190129080029.22261-4-ebiggers@kernel.org> X-Mailer: git-send-email 2.20.1 In-Reply-To: <20190129080029.22261-1-ebiggers@kernel.org> References: <20190129080029.22261-1-ebiggers@kernel.org> MIME-Version: 1.0 X-CRM114-Version: 20100106-BlameMichelson ( TRE 0.8.0 (BSD) ) MR-646709E3 X-CRM114-CacheID: sfid-20190129_000149_707708_9FD09957 X-CRM114-Status: GOOD ( 20.59 ) X-BeenThere: linux-arm-kernel@lists.infradead.org X-Mailman-Version: 2.1.21 Precedence: list List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Cc: Tim Chen , linux-arm-kernel@lists.infradead.org, Ard Biesheuvel Sender: "linux-arm-kernel" Errors-To: linux-arm-kernel-bounces+patchwork-linux-arm=patchwork.kernel.org@lists.infradead.org X-Virus-Scanned: ClamAV using ClamSMTP From: Eric Biggers The x86, arm, and arm64 asm implementations of crct10dif are very difficult to understand partly because many of the comments, labels, and macros are named incorrectly: the lengths mentioned are usually off by a factor of two from the actual code. Many other things are unnecessarily convoluted as well, e.g. there are many more fold constants than actually needed and some aren't fully reduced. This series therefore cleans up all these implementations to be much more maintainable. I also made some small optimizations where I saw opportunities, resulting in slightly better performance. This patch cleans up the arm64 version. Signed-off-by: Eric Biggers --- arch/arm64/crypto/crct10dif-ce-core.S | 496 ++++++++++++-------------- arch/arm64/crypto/crct10dif-ce-glue.c | 4 +- 2 files changed, 233 insertions(+), 267 deletions(-) diff --git a/arch/arm64/crypto/crct10dif-ce-core.S b/arch/arm64/crypto/crct10dif-ce-core.S index f7326259c40de..e545b42e6a468 100644 --- a/arch/arm64/crypto/crct10dif-ce-core.S +++ b/arch/arm64/crypto/crct10dif-ce-core.S @@ -2,12 +2,14 @@ // Accelerated CRC-T10DIF using arm64 NEON and Crypto Extensions instructions // // Copyright (C) 2016 Linaro Ltd +// Copyright (C) 2019 Google LLC // // This program is free software; you can redistribute it and/or modify // it under the terms of the GNU General Public License version 2 as // published by the Free Software Foundation. // +// Derived from the x86 version: // // Implement fast CRC-T10DIF computation with SSE and PCLMULQDQ instructions // @@ -54,19 +56,11 @@ // NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS // SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. // -// Function API: -// UINT16 crc_t10dif_pcl( -// UINT16 init_crc, //initial CRC value, 16 bits -// const unsigned char *buf, //buffer pointer to calculate CRC on -// UINT64 len //buffer length in bytes (64-bit data) -// ); -// // Reference paper titled "Fast CRC Computation for Generic // Polynomials Using PCLMULQDQ Instruction" // URL: http://www.intel.com/content/dam/www/public/us/en/documents // /white-papers/fast-crc-computation-generic-polynomials-pclmulqdq-paper.pdf // -// #include #include @@ -74,14 +68,14 @@ .text .cpu generic+crypto - arg1_low32 .req w19 - arg2 .req x20 - arg3 .req x21 + init_crc .req w19 + buf .req x20 + len .req x21 + fold_consts_ptr .req x22 - vzr .req v13 + fold_consts .req v10 ad .req v14 - bd .req v10 k00_16 .req v15 k32_48 .req v16 @@ -143,11 +137,11 @@ __pmull_p8_core: ext t5.8b, ad.8b, ad.8b, #2 // A2 ext t6.8b, ad.8b, ad.8b, #3 // A3 - pmull t4.8h, t4.8b, bd.8b // F = A1*B + pmull t4.8h, t4.8b, fold_consts.8b // F = A1*B pmull t8.8h, ad.8b, bd1.8b // E = A*B1 - pmull t5.8h, t5.8b, bd.8b // H = A2*B + pmull t5.8h, t5.8b, fold_consts.8b // H = A2*B pmull t7.8h, ad.8b, bd2.8b // G = A*B2 - pmull t6.8h, t6.8b, bd.8b // J = A3*B + pmull t6.8h, t6.8b, fold_consts.8b // J = A3*B pmull t9.8h, ad.8b, bd3.8b // I = A*B3 pmull t3.8h, ad.8b, bd4.8b // K = A*B4 b 0f @@ -157,11 +151,11 @@ __pmull_p8_core: tbl t5.16b, {ad.16b}, perm2.16b // A2 tbl t6.16b, {ad.16b}, perm3.16b // A3 - pmull2 t4.8h, t4.16b, bd.16b // F = A1*B + pmull2 t4.8h, t4.16b, fold_consts.16b // F = A1*B pmull2 t8.8h, ad.16b, bd1.16b // E = A*B1 - pmull2 t5.8h, t5.16b, bd.16b // H = A2*B + pmull2 t5.8h, t5.16b, fold_consts.16b // H = A2*B pmull2 t7.8h, ad.16b, bd2.16b // G = A*B2 - pmull2 t6.8h, t6.16b, bd.16b // J = A3*B + pmull2 t6.8h, t6.16b, fold_consts.16b // J = A3*B pmull2 t9.8h, ad.16b, bd3.16b // I = A*B3 pmull2 t3.8h, ad.16b, bd4.16b // K = A*B4 @@ -203,14 +197,14 @@ __pmull_p8_core: ENDPROC(__pmull_p8_core) .macro __pmull_p8, rq, ad, bd, i - .ifnc \bd, v10 + .ifnc \bd, fold_consts .err .endif mov ad.16b, \ad\().16b .ifb \i - pmull \rq\().8h, \ad\().8b, bd.8b // D = A*B + pmull \rq\().8h, \ad\().8b, \bd\().8b // D = A*B .else - pmull2 \rq\().8h, \ad\().16b, bd.16b // D = A*B + pmull2 \rq\().8h, \ad\().16b, \bd\().16b // D = A*B .endif bl .L__pmull_p8_core\i @@ -219,17 +213,19 @@ ENDPROC(__pmull_p8_core) eor \rq\().16b, \rq\().16b, t6.16b .endm - .macro fold64, p, reg1, reg2 - ldp q11, q12, [arg2], #0x20 + // Fold reg1, reg2 into the next 32 data bytes, storing the result back + // into reg1, reg2. + .macro fold_32_bytes, p, reg1, reg2 + ldp q11, q12, [buf], #0x20 - __pmull_\p v8, \reg1, v10, 2 - __pmull_\p \reg1, \reg1, v10 + __pmull_\p v8, \reg1, fold_consts, 2 + __pmull_\p \reg1, \reg1, fold_consts CPU_LE( rev64 v11.16b, v11.16b ) CPU_LE( rev64 v12.16b, v12.16b ) - __pmull_\p v9, \reg2, v10, 2 - __pmull_\p \reg2, \reg2, v10 + __pmull_\p v9, \reg2, fold_consts, 2 + __pmull_\p \reg2, \reg2, fold_consts CPU_LE( ext v11.16b, v11.16b, v11.16b, #8 ) CPU_LE( ext v12.16b, v12.16b, v12.16b, #8 ) @@ -240,15 +236,16 @@ CPU_LE( ext v12.16b, v12.16b, v12.16b, #8 ) eor \reg2\().16b, \reg2\().16b, v12.16b .endm - .macro fold16, p, reg, rk - __pmull_\p v8, \reg, v10 - __pmull_\p \reg, \reg, v10, 2 - .ifnb \rk - ldr_l q10, \rk, x8 - __pmull_pre_\p v10 + // Fold src_reg into dst_reg, optionally loading the next fold constants + .macro fold_16_bytes, p, src_reg, dst_reg, load_next_consts + __pmull_\p v8, \src_reg, fold_consts + __pmull_\p \src_reg, \src_reg, fold_consts, 2 + .ifnb \load_next_consts + ld1 {fold_consts.2d}, [fold_consts_ptr], #16 + __pmull_pre_\p fold_consts .endif - eor v7.16b, v7.16b, v8.16b - eor v7.16b, v7.16b, \reg\().16b + eor \dst_reg\().16b, \dst_reg\().16b, v8.16b + eor \dst_reg\().16b, \dst_reg\().16b, \src_reg\().16b .endm .macro __pmull_p64, rd, rn, rm, n @@ -260,40 +257,27 @@ CPU_LE( ext v12.16b, v12.16b, v12.16b, #8 ) .endm .macro crc_t10dif_pmull, p - frame_push 3, 128 + frame_push 4, 128 - mov arg1_low32, w0 - mov arg2, x1 - mov arg3, x2 - - movi vzr.16b, #0 // init zero register + mov init_crc, w0 + mov buf, x1 + mov len, x2 __pmull_init_\p - // adjust the 16-bit initial_crc value, scale it to 32 bits - lsl arg1_low32, arg1_low32, #16 - - // check if smaller than 256 - cmp arg3, #256 - - // for sizes less than 128, we can't fold 64B at a time... - b.lt .L_less_than_128_\@ + // For sizes less than 256 bytes, we can't fold 128 bytes at a time. + cmp len, #256 + b.lt .Lless_than_256_bytes_\@ - // load the initial crc value - // crc value does not need to be byte-reflected, but it needs - // to be moved to the high part of the register. - // because data will be byte-reflected and will align with - // initial crc at correct place. - movi v10.16b, #0 - mov v10.s[3], arg1_low32 // initial crc - - // receive the initial 64B data, xor the initial crc value - ldp q0, q1, [arg2] - ldp q2, q3, [arg2, #0x20] - ldp q4, q5, [arg2, #0x40] - ldp q6, q7, [arg2, #0x60] - add arg2, arg2, #0x80 + adr_l fold_consts_ptr, .Lfold_across_128_bytes_consts + // Load the first 128 data bytes. Byte swapping is necessary to make + // the bit order match the polynomial coefficient order. + ldp q0, q1, [buf] + ldp q2, q3, [buf, #0x20] + ldp q4, q5, [buf, #0x40] + ldp q6, q7, [buf, #0x60] + add buf, buf, #0x80 CPU_LE( rev64 v0.16b, v0.16b ) CPU_LE( rev64 v1.16b, v1.16b ) CPU_LE( rev64 v2.16b, v2.16b ) @@ -302,7 +286,6 @@ CPU_LE( rev64 v4.16b, v4.16b ) CPU_LE( rev64 v5.16b, v5.16b ) CPU_LE( rev64 v6.16b, v6.16b ) CPU_LE( rev64 v7.16b, v7.16b ) - CPU_LE( ext v0.16b, v0.16b, v0.16b, #8 ) CPU_LE( ext v1.16b, v1.16b, v1.16b, #8 ) CPU_LE( ext v2.16b, v2.16b, v2.16b, #8 ) @@ -312,36 +295,29 @@ CPU_LE( ext v5.16b, v5.16b, v5.16b, #8 ) CPU_LE( ext v6.16b, v6.16b, v6.16b, #8 ) CPU_LE( ext v7.16b, v7.16b, v7.16b, #8 ) - // XOR the initial_crc value - eor v0.16b, v0.16b, v10.16b - - ldr_l q10, rk3, x8 // xmm10 has rk3 and rk4 - // type of pmull instruction - // will determine which constant to use - __pmull_pre_\p v10 - - // - // we subtract 256 instead of 128 to save one instruction from the loop - // - sub arg3, arg3, #256 - - // at this section of the code, there is 64*x+y (0<=y<64) bytes of - // buffer. The _fold_64_B_loop will fold 64B at a time - // until we have 64+y Bytes of buffer + // XOR the first 16 data *bits* with the initial CRC value. + movi v8.16b, #0 + mov v8.h[7], init_crc + eor v0.16b, v0.16b, v8.16b - // fold 64B at a time. This section of the code folds 4 vector - // registers in parallel -.L_fold_64_B_loop_\@: + // Load the constants for folding across 128 bytes. + ld1 {fold_consts.2d}, [fold_consts_ptr] + __pmull_pre_\p fold_consts - fold64 \p, v0, v1 - fold64 \p, v2, v3 - fold64 \p, v4, v5 - fold64 \p, v6, v7 + // Subtract 128 for the 128 data bytes just consumed. Subtract another + // 128 to simplify the termination condition of the following loop. + sub len, len, #256 - subs arg3, arg3, #128 + // While >= 128 data bytes remain (not counting v0-v7), fold the 128 + // bytes v0-v7 into them, storing the result back into v0-v7. +.Lfold_128_bytes_loop_\@: + fold_32_bytes \p, v0, v1 + fold_32_bytes \p, v2, v3 + fold_32_bytes \p, v4, v5 + fold_32_bytes \p, v6, v7 - // check if there is another 64B in the buffer to be able to fold - b.lt .L_fold_64_B_end_\@ + subs len, len, #128 + b.lt .Lfold_128_bytes_loop_done_\@ if_will_cond_yield_neon stp q0, q1, [sp, #.Lframe_local_offset] @@ -353,217 +329,207 @@ CPU_LE( ext v7.16b, v7.16b, v7.16b, #8 ) ldp q2, q3, [sp, #.Lframe_local_offset + 32] ldp q4, q5, [sp, #.Lframe_local_offset + 64] ldp q6, q7, [sp, #.Lframe_local_offset + 96] - ldr_l q10, rk3, x8 - movi vzr.16b, #0 // init zero register + ld1 {fold_consts.2d}, [fold_consts_ptr] __pmull_init_\p - __pmull_pre_\p v10 + __pmull_pre_\p fold_consts endif_yield_neon - b .L_fold_64_B_loop_\@ - -.L_fold_64_B_end_\@: - // at this point, the buffer pointer is pointing at the last y Bytes - // of the buffer the 64B of folded data is in 4 of the vector - // registers: v0, v1, v2, v3 - - // fold the 8 vector registers to 1 vector register with different - // constants - - ldr_l q10, rk9, x8 - __pmull_pre_\p v10 - - fold16 \p, v0, rk11 - fold16 \p, v1, rk13 - fold16 \p, v2, rk15 - fold16 \p, v3, rk17 - fold16 \p, v4, rk19 - fold16 \p, v5, rk1 - fold16 \p, v6 - - // instead of 64, we add 48 to the loop counter to save 1 instruction - // from the loop instead of a cmp instruction, we use the negative - // flag with the jl instruction - adds arg3, arg3, #(128-16) - b.lt .L_final_reduction_for_128_\@ - - // now we have 16+y bytes left to reduce. 16 Bytes is in register v7 - // and the rest is in memory. We can fold 16 bytes at a time if y>=16 - // continue folding 16B at a time - -.L_16B_reduction_loop_\@: - __pmull_\p v8, v7, v10 - __pmull_\p v7, v7, v10, 2 + b .Lfold_128_bytes_loop_\@ + +.Lfold_128_bytes_loop_done_\@: + + // Now fold the 112 bytes in v0-v6 into the 16 bytes in v7. + + // Fold across 64 bytes. + add fold_consts_ptr, fold_consts_ptr, #16 + ld1 {fold_consts.2d}, [fold_consts_ptr], #16 + __pmull_pre_\p fold_consts + fold_16_bytes \p, v0, v4 + fold_16_bytes \p, v1, v5 + fold_16_bytes \p, v2, v6 + fold_16_bytes \p, v3, v7, 1 + // Fold across 32 bytes. + fold_16_bytes \p, v4, v6 + fold_16_bytes \p, v5, v7, 1 + // Fold across 16 bytes. + fold_16_bytes \p, v6, v7 + + // Add 128 to get the correct number of data bytes remaining in 0...127 + // (not counting v7), following the previous extra subtraction by 128. + // Then subtract 16 to simplify the termination condition of the + // following loop. + adds len, len, #(128-16) + + // While >= 16 data bytes remain (not counting v7), fold the 16 bytes v7 + // into them, storing the result back into v7. + b.lt .Lfold_16_bytes_loop_done_\@ +.Lfold_16_bytes_loop_\@: + __pmull_\p v8, v7, fold_consts + __pmull_\p v7, v7, fold_consts, 2 eor v7.16b, v7.16b, v8.16b - - ldr q0, [arg2], #16 + ldr q0, [buf], #16 CPU_LE( rev64 v0.16b, v0.16b ) CPU_LE( ext v0.16b, v0.16b, v0.16b, #8 ) eor v7.16b, v7.16b, v0.16b - subs arg3, arg3, #16 - - // instead of a cmp instruction, we utilize the flags with the - // jge instruction equivalent of: cmp arg3, 16-16 - // check if there is any more 16B in the buffer to be able to fold - b.ge .L_16B_reduction_loop_\@ - - // now we have 16+z bytes left to reduce, where 0<= z < 16. - // first, we reduce the data in the xmm7 register - -.L_final_reduction_for_128_\@: - // check if any more data to fold. If not, compute the CRC of - // the final 128 bits - adds arg3, arg3, #16 - b.eq .L_128_done_\@ - - // here we are getting data that is less than 16 bytes. - // since we know that there was data before the pointer, we can - // offset the input pointer before the actual point, to receive - // exactly 16 bytes. after that the registers need to be adjusted. -.L_get_last_two_regs_\@: - add arg2, arg2, arg3 - ldr q1, [arg2, #-16] -CPU_LE( rev64 v1.16b, v1.16b ) -CPU_LE( ext v1.16b, v1.16b, v1.16b, #8 ) - - // get rid of the extra data that was loaded before - // load the shift constant - adr_l x4, tbl_shf_table + 16 - sub x4, x4, arg3 - ld1 {v0.16b}, [x4] - - // shift v2 to the left by arg3 bytes - tbl v2.16b, {v7.16b}, v0.16b - - // shift v7 to the right by 16-arg3 bytes - movi v9.16b, #0x80 - eor v0.16b, v0.16b, v9.16b - tbl v7.16b, {v7.16b}, v0.16b - - // blend - sshr v0.16b, v0.16b, #7 // convert to 8-bit mask - bsl v0.16b, v2.16b, v1.16b - - // fold 16 Bytes - __pmull_\p v8, v7, v10 - __pmull_\p v7, v7, v10, 2 - eor v7.16b, v7.16b, v8.16b - eor v7.16b, v7.16b, v0.16b + subs len, len, #16 + b.ge .Lfold_16_bytes_loop_\@ + +.Lfold_16_bytes_loop_done_\@: + // Add 16 to get the correct number of data bytes remaining in 0...15 + // (not counting v7), following the previous extra subtraction by 16. + adds len, len, #16 + b.eq .Lreduce_final_16_bytes_\@ + +.Lhandle_partial_segment_\@: + // Reduce the last '16 + len' bytes where 1 <= len <= 15 and the first + // 16 bytes are in v7 and the rest are the remaining data in 'buf'. To + // do this without needing a fold constant for each possible 'len', + // redivide the bytes into a first chunk of 'len' bytes and a second + // chunk of 16 bytes, then fold the first chunk into the second. + + // v0 = last 16 original data bytes + add buf, buf, len + ldr q0, [buf, #-16] +CPU_LE( rev64 v0.16b, v0.16b ) +CPU_LE( ext v0.16b, v0.16b, v0.16b, #8 ) -.L_128_done_\@: - // compute crc of a 128-bit value - ldr_l q10, rk5, x8 // rk5 and rk6 in xmm10 - __pmull_pre_\p v10 + // v1 = high order part of second chunk: v7 left-shifted by 'len' bytes. + adr_l x4, .Lbyteshift_table + 16 + sub x4, x4, len + ld1 {v2.16b}, [x4] + tbl v1.16b, {v7.16b}, v2.16b - // 64b fold - ext v0.16b, vzr.16b, v7.16b, #8 - mov v7.d[0], v7.d[1] - __pmull_\p v7, v7, v10 - eor v7.16b, v7.16b, v0.16b + // v3 = first chunk: v7 right-shifted by '16-len' bytes. + movi v3.16b, #0x80 + eor v2.16b, v2.16b, v3.16b + tbl v3.16b, {v7.16b}, v2.16b - // 32b fold - ext v0.16b, v7.16b, vzr.16b, #4 - mov v7.s[3], vzr.s[0] - __pmull_\p v0, v0, v10, 2 - eor v7.16b, v7.16b, v0.16b + // Convert to 8-bit masks: 'len' 0x00 bytes, then '16-len' 0xff bytes. + sshr v2.16b, v2.16b, #7 - // barrett reduction - ldr_l q10, rk7, x8 - __pmull_pre_\p v10 - mov v0.d[0], v7.d[1] + // v2 = second chunk: 'len' bytes from v0 (low-order bytes), + // then '16-len' bytes from v1 (high-order bytes). + bsl v2.16b, v1.16b, v0.16b - __pmull_\p v0, v0, v10 - ext v0.16b, vzr.16b, v0.16b, #12 - __pmull_\p v0, v0, v10, 2 - ext v0.16b, vzr.16b, v0.16b, #12 + // Fold the first chunk into the second chunk, storing the result in v7. + __pmull_\p v0, v3, fold_consts + __pmull_\p v7, v3, fold_consts, 2 eor v7.16b, v7.16b, v0.16b - mov w0, v7.s[1] - -.L_cleanup_\@: - // scale the result back to 16 bits - lsr x0, x0, #16 + eor v7.16b, v7.16b, v2.16b + +.Lreduce_final_16_bytes_\@: + // Reduce the 128-bit value M(x), stored in v7, to the final 16-bit CRC. + + movi v2.16b, #0 // init zero register + + // Load 'x^48 * (x^48 mod G(x))' and 'x^48 * (x^80 mod G(x))'. + ld1 {fold_consts.2d}, [fold_consts_ptr], #16 + __pmull_pre_\p fold_consts + + // Fold the high 64 bits into the low 64 bits, while also multiplying by + // x^64. This produces a 128-bit value congruent to x^64 * M(x) and + // whose low 48 bits are 0. + ext v0.16b, v2.16b, v7.16b, #8 + __pmull_\p v7, v7, fold_consts, 2 // high bits * x^48 * (x^80 mod G(x)) + eor v0.16b, v0.16b, v7.16b // + low bits * x^64 + + // Fold the high 32 bits into the low 96 bits. This produces a 96-bit + // value congruent to x^64 * M(x) and whose low 48 bits are 0. + ext v1.16b, v0.16b, v2.16b, #12 // extract high 32 bits + mov v0.s[3], v2.s[0] // zero high 32 bits + __pmull_\p v1, v1, fold_consts // high 32 bits * x^48 * (x^48 mod G(x)) + eor v0.16b, v0.16b, v1.16b // + low bits + + // Load G(x) and floor(x^48 / G(x)). + ld1 {fold_consts.2d}, [fold_consts_ptr] + __pmull_pre_\p fold_consts + + // Use Barrett reduction to compute the final CRC value. + __pmull_\p v1, v0, fold_consts, 2 // high 32 bits * floor(x^48 / G(x)) + ushr v1.2d, v1.2d, #32 // /= x^32 + __pmull_\p v1, v1, fold_consts // *= G(x) + ushr v0.2d, v0.2d, #48 + eor v0.16b, v0.16b, v1.16b // + low 16 nonzero bits + // Final CRC value (x^16 * M(x)) mod G(x) is in low 16 bits of v0. + + umov w0, v0.h[0] frame_pop ret -.L_less_than_128_\@: - cbz arg3, .L_cleanup_\@ +.Lless_than_256_bytes_\@: + // Checksumming a buffer of length 16...255 bytes - movi v0.16b, #0 - mov v0.s[3], arg1_low32 // get the initial crc value + adr_l fold_consts_ptr, .Lfold_across_16_bytes_consts - ldr q7, [arg2], #0x10 + // Load the first 16 data bytes. + ldr q7, [buf], #0x10 CPU_LE( rev64 v7.16b, v7.16b ) CPU_LE( ext v7.16b, v7.16b, v7.16b, #8 ) - eor v7.16b, v7.16b, v0.16b // xor the initial crc value - - cmp arg3, #16 - b.eq .L_128_done_\@ // exactly 16 left - ldr_l q10, rk1, x8 // rk1 and rk2 in xmm10 - __pmull_pre_\p v10 + // XOR the first 16 data *bits* with the initial CRC value. + movi v0.16b, #0 + mov v0.h[7], init_crc + eor v7.16b, v7.16b, v0.16b - // update the counter. subtract 32 instead of 16 to save one - // instruction from the loop - subs arg3, arg3, #32 - b.ge .L_16B_reduction_loop_\@ + // Load the fold-across-16-bytes constants. + ld1 {fold_consts.2d}, [fold_consts_ptr], #16 + __pmull_pre_\p fold_consts - add arg3, arg3, #16 - b .L_get_last_two_regs_\@ + cmp len, #16 + b.eq .Lreduce_final_16_bytes_\@ // len == 16 + subs len, len, #32 + b.ge .Lfold_16_bytes_loop_\@ // 32 <= len <= 255 + add len, len, #16 + b .Lhandle_partial_segment_\@ // 17 <= len <= 31 .endm +// +// u16 crc_t10dif_pmull_p8(u16 init_crc, const u8 *buf, size_t len); +// +// Assumes len >= 16. +// ENTRY(crc_t10dif_pmull_p8) crc_t10dif_pmull p8 ENDPROC(crc_t10dif_pmull_p8) .align 5 +// +// u16 crc_t10dif_pmull_p64(u16 init_crc, const u8 *buf, size_t len); +// +// Assumes len >= 16. +// ENTRY(crc_t10dif_pmull_p64) crc_t10dif_pmull p64 ENDPROC(crc_t10dif_pmull_p64) -// precomputed constants -// these constants are precomputed from the poly: -// 0x8bb70000 (0x8bb7 scaled to 32 bits) .section ".rodata", "a" .align 4 -// Q = 0x18BB70000 -// rk1 = 2^(32*3) mod Q << 32 -// rk2 = 2^(32*5) mod Q << 32 -// rk3 = 2^(32*15) mod Q << 32 -// rk4 = 2^(32*17) mod Q << 32 -// rk5 = 2^(32*3) mod Q << 32 -// rk6 = 2^(32*2) mod Q << 32 -// rk7 = floor(2^64/Q) -// rk8 = Q - -rk1: .octa 0x06df0000000000002d56000000000000 -rk3: .octa 0x7cf50000000000009d9d000000000000 -rk5: .octa 0x13680000000000002d56000000000000 -rk7: .octa 0x000000018bb7000000000001f65a57f8 -rk9: .octa 0xbfd6000000000000ceae000000000000 -rk11: .octa 0x713c0000000000001e16000000000000 -rk13: .octa 0x80a6000000000000f7f9000000000000 -rk15: .octa 0xe658000000000000044c000000000000 -rk17: .octa 0xa497000000000000ad18000000000000 -rk19: .octa 0xe7b50000000000006ee3000000000000 - -tbl_shf_table: -// use these values for shift constants for the tbl/tbx instruction -// different alignments result in values as shown: -// DDQ 0x008f8e8d8c8b8a898887868584838281 # shl 15 (16-1) / shr1 -// DDQ 0x01008f8e8d8c8b8a8988878685848382 # shl 14 (16-3) / shr2 -// DDQ 0x0201008f8e8d8c8b8a89888786858483 # shl 13 (16-4) / shr3 -// DDQ 0x030201008f8e8d8c8b8a898887868584 # shl 12 (16-4) / shr4 -// DDQ 0x04030201008f8e8d8c8b8a8988878685 # shl 11 (16-5) / shr5 -// DDQ 0x0504030201008f8e8d8c8b8a89888786 # shl 10 (16-6) / shr6 -// DDQ 0x060504030201008f8e8d8c8b8a898887 # shl 9 (16-7) / shr7 -// DDQ 0x07060504030201008f8e8d8c8b8a8988 # shl 8 (16-8) / shr8 -// DDQ 0x0807060504030201008f8e8d8c8b8a89 # shl 7 (16-9) / shr9 -// DDQ 0x090807060504030201008f8e8d8c8b8a # shl 6 (16-10) / shr10 -// DDQ 0x0a090807060504030201008f8e8d8c8b # shl 5 (16-11) / shr11 -// DDQ 0x0b0a090807060504030201008f8e8d8c # shl 4 (16-12) / shr12 -// DDQ 0x0c0b0a090807060504030201008f8e8d # shl 3 (16-13) / shr13 -// DDQ 0x0d0c0b0a090807060504030201008f8e # shl 2 (16-14) / shr14 -// DDQ 0x0e0d0c0b0a090807060504030201008f # shl 1 (16-15) / shr15 +// Fold constants precomputed from the polynomial 0x18bb7 +// G(x) = x^16 + x^15 + x^11 + x^9 + x^8 + x^7 + x^5 + x^4 + x^2 + x^1 + x^0 +.Lfold_across_128_bytes_consts: + .quad 0x0000000000006123 // x^(8*128) mod G(x) + .quad 0x0000000000002295 // x^(8*128+64) mod G(x) +// .Lfold_across_64_bytes_consts: + .quad 0x0000000000001069 // x^(4*128) mod G(x) + .quad 0x000000000000dd31 // x^(4*128+64) mod G(x) +// .Lfold_across_32_bytes_consts: + .quad 0x000000000000857d // x^(2*128) mod G(x) + .quad 0x0000000000007acc // x^(2*128+64) mod G(x) +.Lfold_across_16_bytes_consts: + .quad 0x000000000000a010 // x^(1*128) mod G(x) + .quad 0x0000000000001faa // x^(1*128+64) mod G(x) +// .Lfinal_fold_consts: + .quad 0x1368000000000000 // x^48 * (x^48 mod G(x)) + .quad 0x2d56000000000000 // x^48 * (x^80 mod G(x)) +// .Lbarrett_reduction_consts: + .quad 0x0000000000018bb7 // G(x) + .quad 0x00000001f65a57f8 // floor(x^48 / G(x)) + +// For 1 <= len <= 15, the 16-byte vector beginning at &byteshift_table[16 - +// len] is the index vector to shift left by 'len' bytes, and is also {0x80, +// ..., 0x80} XOR the index vector to shift right by '16 - len' bytes. +.Lbyteshift_table: .byte 0x0, 0x81, 0x82, 0x83, 0x84, 0x85, 0x86, 0x87 .byte 0x88, 0x89, 0x8a, 0x8b, 0x8c, 0x8d, 0x8e, 0x8f .byte 0x0, 0x1, 0x2, 0x3, 0x4, 0x5, 0x6, 0x7 diff --git a/arch/arm64/crypto/crct10dif-ce-glue.c b/arch/arm64/crypto/crct10dif-ce-glue.c index 242757cc6da91..dd325829ee44f 100644 --- a/arch/arm64/crypto/crct10dif-ce-glue.c +++ b/arch/arm64/crypto/crct10dif-ce-glue.c @@ -22,8 +22,8 @@ #define CRC_T10DIF_PMULL_CHUNK_SIZE 16U -asmlinkage u16 crc_t10dif_pmull_p64(u16 init_crc, const u8 buf[], u64 len); -asmlinkage u16 crc_t10dif_pmull_p8(u16 init_crc, const u8 buf[], u64 len); +asmlinkage u16 crc_t10dif_pmull_p8(u16 init_crc, const u8 *buf, size_t len); +asmlinkage u16 crc_t10dif_pmull_p64(u16 init_crc, const u8 *buf, size_t len); static int crct10dif_init(struct shash_desc *desc) {