From patchwork Tue Jan 29 08:00:27 2019 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Eric Biggers X-Patchwork-Id: 10785647 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 54488922 for ; Tue, 29 Jan 2019 08:02:58 +0000 (UTC) Received: from mail.wl.linuxfoundation.org (localhost [127.0.0.1]) by mail.wl.linuxfoundation.org (Postfix) with ESMTP id 4166E2B7A2 for ; Tue, 29 Jan 2019 08:02:58 +0000 (UTC) Received: by mail.wl.linuxfoundation.org (Postfix, from userid 486) id 34F132B92E; Tue, 29 Jan 2019 08:02:58 +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 B47642B7A2 for ; Tue, 29 Jan 2019 08:02:56 +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=1gFS16DU7ngyRaKsxopat2poK73Pbikg5RunJfGfOac=; b=X2301xuzQL+7Kg K4K4mPaWN8STcdou40a9lZgOg120K4K9AC0MahK3bgNFc9QWpRVwiGdl9kAmxB0m7m95gqDTRV0Hf /l5oUEhmsCnzrszdjxOUcgUBiWZNLjGcPAxmxhwCvDYXG0vYerJ+fu17Fr3Vnoh2xC7dpFfM50Gqd T3opFkNtHrkIlbnUXeBFpwkH7To/fOcuDaqX2uJ5WOqaRxttgb+1Hy4v+NEQ3KS8ni7R+PG94LHNQ riELoT4JDiS7iev0VhCYynckENKHR42na16aSQzaCjtepkPM02zDxiOvdBC+3zUGGhHQZSfOCYVk5 SbCD0sTpkd9dLV4N68Cw==; 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 1goOMA-0001I7-4V; Tue, 29 Jan 2019 08:02:54 +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 1goOL6-0000QX-Qy for linux-arm-kernel@lists.infradead.org; Tue, 29 Jan 2019 08:01:58 +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 1010A2177E; 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=1548748908; bh=YGgas+iq/hrR0v+9N0H8jt9SmcrXzGDRqj94y5SExmQ=; h=From:To:Cc:Subject:Date:In-Reply-To:References:From; b=AlKe7V05q6FatS5wifgXqGvWw8dIO+ybSxSV7e/r9RCGQ8FT/NFCF4dYMJJx+5wqq 96gFj4kk76xmES40NkTz6xGzcNQ9LeItAdubNC7dJE5NFNZENrhdoKieghlujcPdEV aI1seMRE0/cTT5uEH1uDBYUvT130Y7x35BDYLGDw= From: Eric Biggers To: linux-crypto@vger.kernel.org, Herbert Xu Subject: [PATCH v2 1/3] crypto: x86/crct10dif-pcl - cleanup and optimizations Date: Tue, 29 Jan 2019 00:00:27 -0800 Message-Id: <20190129080029.22261-2-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_158971_D0A754AA X-CRM114-Status: GOOD ( 21.22 ) 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 x86 version. Note that I opted to retain the support for buffers < 16 bytes, but I simplified it to be much shorter and have fewer different cases. Performance on 2 and 3-byte inputs dropped slightly but the rest are actually faster now. Signed-off-by: Eric Biggers --- arch/x86/crypto/crct10dif-pcl-asm_64.S | 844 +++++++++--------------- arch/x86/crypto/crct10dif-pclmul_glue.c | 3 +- 2 files changed, 301 insertions(+), 546 deletions(-) diff --git a/arch/x86/crypto/crct10dif-pcl-asm_64.S b/arch/x86/crypto/crct10dif-pcl-asm_64.S index de04d3e98d8d3..e57afa35ac0f9 100644 --- a/arch/x86/crypto/crct10dif-pcl-asm_64.S +++ b/arch/x86/crypto/crct10dif-pcl-asm_64.S @@ -43,609 +43,365 @@ # LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING # 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 .text -#define arg1 %rdi -#define arg2 %rsi -#define arg3 %rdx - -#define arg1_low32 %edi +#define init_crc %di +#define init_crcl %edi +#define buf %rsi +#define len %rdx + +#define FOLD_CONSTS %xmm10 +#define BSWAP_MASK %xmm11 + +# Fold reg1, reg2 into the next 32 data bytes, storing the result back into +# reg1, reg2. +.macro fold_32_bytes offset, reg1, reg2 + movdqu \offset(buf), %xmm9 + movdqu \offset+16(buf), %xmm12 + pshufb BSWAP_MASK, %xmm9 + pshufb BSWAP_MASK, %xmm12 + movdqa \reg1, %xmm8 + movdqa \reg2, %xmm13 + pclmulqdq $0x00, FOLD_CONSTS, \reg1 + pclmulqdq $0x11, FOLD_CONSTS, %xmm8 + pclmulqdq $0x00, FOLD_CONSTS, \reg2 + pclmulqdq $0x11, FOLD_CONSTS, %xmm13 + pxor %xmm9 , \reg1 + xorps %xmm8 , \reg1 + pxor %xmm12, \reg2 + xorps %xmm13, \reg2 +.endm + +# Fold src_reg into dst_reg. +.macro fold_16_bytes src_reg, dst_reg + movdqa \src_reg, %xmm8 + pclmulqdq $0x11, FOLD_CONSTS, \src_reg + pclmulqdq $0x00, FOLD_CONSTS, %xmm8 + pxor %xmm8, \dst_reg + xorps \src_reg, \dst_reg +.endm -ENTRY(crc_t10dif_pcl) +# +# u16 crc_t10dif_pcl(u16 init_crc, const *u8 buf, size_t len); +# .align 16 - - # adjust the 16-bit initial_crc value, scale it to 32 bits - shl $16, arg1_low32 - - # Allocate Stack Space - mov %rsp, %rcx - sub $16*2, %rsp - # align stack to 16 byte boundary - and $~(0x10 - 1), %rsp - - # check if smaller than 256 - cmp $256, arg3 - - # for sizes less than 128, we can't fold 64B at a time... - jl _less_than_128 - - - # load the initial crc value - movd arg1_low32, %xmm10 # initial crc - - # 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. - pslldq $12, %xmm10 - - movdqa SHUF_MASK(%rip), %xmm11 - # receive the initial 64B data, xor the initial crc value - movdqu 16*0(arg2), %xmm0 - movdqu 16*1(arg2), %xmm1 - movdqu 16*2(arg2), %xmm2 - movdqu 16*3(arg2), %xmm3 - movdqu 16*4(arg2), %xmm4 - movdqu 16*5(arg2), %xmm5 - movdqu 16*6(arg2), %xmm6 - movdqu 16*7(arg2), %xmm7 - - pshufb %xmm11, %xmm0 - # XOR the initial_crc value - pxor %xmm10, %xmm0 - pshufb %xmm11, %xmm1 - pshufb %xmm11, %xmm2 - pshufb %xmm11, %xmm3 - pshufb %xmm11, %xmm4 - pshufb %xmm11, %xmm5 - pshufb %xmm11, %xmm6 - pshufb %xmm11, %xmm7 - - movdqa rk3(%rip), %xmm10 #xmm10 has rk3 and rk4 - #imm value of pclmulqdq instruction - #will determine which constant to use - - ################################################################# - # we subtract 256 instead of 128 to save one instruction from the loop - sub $256, arg3 - - # 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 - - - # fold 64B at a time. This section of the code folds 4 xmm - # registers in parallel -_fold_64_B_loop: - - # update the buffer pointer - add $128, arg2 # buf += 64# - - movdqu 16*0(arg2), %xmm9 - movdqu 16*1(arg2), %xmm12 - pshufb %xmm11, %xmm9 - pshufb %xmm11, %xmm12 - movdqa %xmm0, %xmm8 - movdqa %xmm1, %xmm13 - pclmulqdq $0x0 , %xmm10, %xmm0 - pclmulqdq $0x11, %xmm10, %xmm8 - pclmulqdq $0x0 , %xmm10, %xmm1 - pclmulqdq $0x11, %xmm10, %xmm13 - pxor %xmm9 , %xmm0 - xorps %xmm8 , %xmm0 - pxor %xmm12, %xmm1 - xorps %xmm13, %xmm1 - - movdqu 16*2(arg2), %xmm9 - movdqu 16*3(arg2), %xmm12 - pshufb %xmm11, %xmm9 - pshufb %xmm11, %xmm12 - movdqa %xmm2, %xmm8 - movdqa %xmm3, %xmm13 - pclmulqdq $0x0, %xmm10, %xmm2 - pclmulqdq $0x11, %xmm10, %xmm8 - pclmulqdq $0x0, %xmm10, %xmm3 - pclmulqdq $0x11, %xmm10, %xmm13 - pxor %xmm9 , %xmm2 - xorps %xmm8 , %xmm2 - pxor %xmm12, %xmm3 - xorps %xmm13, %xmm3 - - movdqu 16*4(arg2), %xmm9 - movdqu 16*5(arg2), %xmm12 - pshufb %xmm11, %xmm9 - pshufb %xmm11, %xmm12 - movdqa %xmm4, %xmm8 - movdqa %xmm5, %xmm13 - pclmulqdq $0x0, %xmm10, %xmm4 - pclmulqdq $0x11, %xmm10, %xmm8 - pclmulqdq $0x0, %xmm10, %xmm5 - pclmulqdq $0x11, %xmm10, %xmm13 - pxor %xmm9 , %xmm4 - xorps %xmm8 , %xmm4 - pxor %xmm12, %xmm5 - xorps %xmm13, %xmm5 - - movdqu 16*6(arg2), %xmm9 - movdqu 16*7(arg2), %xmm12 - pshufb %xmm11, %xmm9 - pshufb %xmm11, %xmm12 - movdqa %xmm6 , %xmm8 - movdqa %xmm7 , %xmm13 - pclmulqdq $0x0 , %xmm10, %xmm6 - pclmulqdq $0x11, %xmm10, %xmm8 - pclmulqdq $0x0 , %xmm10, %xmm7 - pclmulqdq $0x11, %xmm10, %xmm13 - pxor %xmm9 , %xmm6 - xorps %xmm8 , %xmm6 - pxor %xmm12, %xmm7 - xorps %xmm13, %xmm7 - - sub $128, arg3 - - # check if there is another 64B in the buffer to be able to fold - jge _fold_64_B_loop - ################################################################## - - - add $128, arg2 - # 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 xmm - # registers: xmm0, xmm1, xmm2, xmm3 - - - # fold the 8 xmm registers to 1 xmm register with different constants - - movdqa rk9(%rip), %xmm10 - movdqa %xmm0, %xmm8 - pclmulqdq $0x11, %xmm10, %xmm0 - pclmulqdq $0x0 , %xmm10, %xmm8 - pxor %xmm8, %xmm7 - xorps %xmm0, %xmm7 - - movdqa rk11(%rip), %xmm10 - movdqa %xmm1, %xmm8 - pclmulqdq $0x11, %xmm10, %xmm1 - pclmulqdq $0x0 , %xmm10, %xmm8 - pxor %xmm8, %xmm7 - xorps %xmm1, %xmm7 - - movdqa rk13(%rip), %xmm10 - movdqa %xmm2, %xmm8 - pclmulqdq $0x11, %xmm10, %xmm2 - pclmulqdq $0x0 , %xmm10, %xmm8 - pxor %xmm8, %xmm7 - pxor %xmm2, %xmm7 - - movdqa rk15(%rip), %xmm10 - movdqa %xmm3, %xmm8 - pclmulqdq $0x11, %xmm10, %xmm3 - pclmulqdq $0x0 , %xmm10, %xmm8 - pxor %xmm8, %xmm7 - xorps %xmm3, %xmm7 - - movdqa rk17(%rip), %xmm10 - movdqa %xmm4, %xmm8 - pclmulqdq $0x11, %xmm10, %xmm4 - pclmulqdq $0x0 , %xmm10, %xmm8 - pxor %xmm8, %xmm7 - pxor %xmm4, %xmm7 - - movdqa rk19(%rip), %xmm10 - movdqa %xmm5, %xmm8 - pclmulqdq $0x11, %xmm10, %xmm5 - pclmulqdq $0x0 , %xmm10, %xmm8 - pxor %xmm8, %xmm7 - xorps %xmm5, %xmm7 - - movdqa rk1(%rip), %xmm10 #xmm10 has rk1 and rk2 - #imm value of pclmulqdq instruction - #will determine which constant to use - movdqa %xmm6, %xmm8 - pclmulqdq $0x11, %xmm10, %xmm6 - pclmulqdq $0x0 , %xmm10, %xmm8 - pxor %xmm8, %xmm7 - pxor %xmm6, %xmm7 - - - # 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 - add $128-16, arg3 - jl _final_reduction_for_128 - - # now we have 16+y bytes left to reduce. 16 Bytes is in register xmm7 - # and the rest is in memory. We can fold 16 bytes at a time if y>=16 - # continue folding 16B at a time - -_16B_reduction_loop: +ENTRY(crc_t10dif_pcl) + # Allocate a 16-byte aligned 16-byte stack buffer. + mov %rsp, %rcx + sub $16, %rsp + and $~15, %rsp + + movdqa .Lbswap_mask(%rip), BSWAP_MASK + + # For sizes less than 256 bytes, we can't fold 128 bytes at a time. + cmp $256, len + jl .Lless_than_256_bytes + + # Load the first 128 data bytes. Byte swapping is necessary to make the + # bit order match the polynomial coefficient order. + movdqu 16*0(buf), %xmm0 + movdqu 16*1(buf), %xmm1 + movdqu 16*2(buf), %xmm2 + movdqu 16*3(buf), %xmm3 + movdqu 16*4(buf), %xmm4 + movdqu 16*5(buf), %xmm5 + movdqu 16*6(buf), %xmm6 + movdqu 16*7(buf), %xmm7 + add $128, buf + pshufb BSWAP_MASK, %xmm0 + pshufb BSWAP_MASK, %xmm1 + pshufb BSWAP_MASK, %xmm2 + pshufb BSWAP_MASK, %xmm3 + pshufb BSWAP_MASK, %xmm4 + pshufb BSWAP_MASK, %xmm5 + pshufb BSWAP_MASK, %xmm6 + pshufb BSWAP_MASK, %xmm7 + + # XOR the first 16 data *bits* with the initial CRC value. + pxor %xmm8, %xmm8 + pinsrw $7, init_crcl, %xmm8 + pxor %xmm8, %xmm0 + + movdqa .Lfold_across_128_bytes_consts(%rip), FOLD_CONSTS + + # Subtract 128 for the 128 data bytes just consumed. Subtract another + # 128 to simplify the termination condition of the following loop. + sub $256, len + + # While >= 128 data bytes remain (not counting xmm0-7), fold the 128 + # bytes xmm0-7 into them, storing the result back into xmm0-7. +.Lfold_128_bytes_loop: + fold_32_bytes 0, %xmm0, %xmm1 + fold_32_bytes 32, %xmm2, %xmm3 + fold_32_bytes 64, %xmm4, %xmm5 + fold_32_bytes 96, %xmm6, %xmm7 + add $128, buf + sub $128, len + jge .Lfold_128_bytes_loop + + # Now fold the 112 bytes in xmm0-xmm6 into the 16 bytes in xmm7. + + # Fold across 64 bytes. + movdqa .Lfold_across_64_bytes_consts(%rip), FOLD_CONSTS + fold_16_bytes %xmm0, %xmm4 + fold_16_bytes %xmm1, %xmm5 + fold_16_bytes %xmm2, %xmm6 + fold_16_bytes %xmm3, %xmm7 + # Fold across 32 bytes. + movdqa .Lfold_across_32_bytes_consts(%rip), FOLD_CONSTS + fold_16_bytes %xmm4, %xmm6 + fold_16_bytes %xmm5, %xmm7 + # Fold across 16 bytes. + movdqa .Lfold_across_16_bytes_consts(%rip), FOLD_CONSTS + fold_16_bytes %xmm6, %xmm7 + + # Add 128 to get the correct number of data bytes remaining in 0...127 + # (not counting xmm7), following the previous extra subtraction by 128. + # Then subtract 16 to simplify the termination condition of the + # following loop. + add $128-16, len + + # While >= 16 data bytes remain (not counting xmm7), fold the 16 bytes + # xmm7 into them, storing the result back into xmm7. + jl .Lfold_16_bytes_loop_done +.Lfold_16_bytes_loop: movdqa %xmm7, %xmm8 - pclmulqdq $0x11, %xmm10, %xmm7 - pclmulqdq $0x0 , %xmm10, %xmm8 + pclmulqdq $0x11, FOLD_CONSTS, %xmm7 + pclmulqdq $0x00, FOLD_CONSTS, %xmm8 pxor %xmm8, %xmm7 - movdqu (arg2), %xmm0 - pshufb %xmm11, %xmm0 + movdqu (buf), %xmm0 + pshufb BSWAP_MASK, %xmm0 pxor %xmm0 , %xmm7 - add $16, arg2 - sub $16, arg3 - # 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 - jge _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 - - -_final_reduction_for_128: - # check if any more data to fold. If not, compute the CRC of - # the final 128 bits - add $16, arg3 - je _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. -_get_last_two_xmms: + add $16, buf + sub $16, len + jge .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 xmm7), following the previous extra subtraction by 16. + add $16, len + je .Lreduce_final_16_bytes + +.Lhandle_partial_segment: + # Reduce the last '16 + len' bytes where 1 <= len <= 15 and the first 16 + # bytes are in xmm7 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. + movdqa %xmm7, %xmm2 - movdqu -16(arg2, arg3), %xmm1 - pshufb %xmm11, %xmm1 + # xmm1 = last 16 original data bytes + movdqu -16(buf, len), %xmm1 + pshufb BSWAP_MASK, %xmm1 - # get rid of the extra data that was loaded before - # load the shift constant - lea pshufb_shf_table+16(%rip), %rax - sub arg3, %rax + # xmm2 = high order part of second chunk: xmm7 left-shifted by 'len' bytes. + lea .Lbyteshift_table+16(%rip), %rax + sub len, %rax movdqu (%rax), %xmm0 - - # shift xmm2 to the left by arg3 bytes pshufb %xmm0, %xmm2 - # shift xmm7 to the right by 16-arg3 bytes - pxor mask1(%rip), %xmm0 + # xmm7 = first chunk: xmm7 right-shifted by '16-len' bytes. + pxor .Lmask1(%rip), %xmm0 pshufb %xmm0, %xmm7 + + # xmm1 = second chunk: 'len' bytes from xmm1 (low-order bytes), + # then '16-len' bytes from xmm2 (high-order bytes). pblendvb %xmm2, %xmm1 #xmm0 is implicit - # fold 16 Bytes - movdqa %xmm1, %xmm2 + # Fold the first chunk into the second chunk, storing the result in xmm7. movdqa %xmm7, %xmm8 - pclmulqdq $0x11, %xmm10, %xmm7 - pclmulqdq $0x0 , %xmm10, %xmm8 + pclmulqdq $0x11, FOLD_CONSTS, %xmm7 + pclmulqdq $0x00, FOLD_CONSTS, %xmm8 pxor %xmm8, %xmm7 - pxor %xmm2, %xmm7 + pxor %xmm1, %xmm7 -_128_done: - # compute crc of a 128-bit value - movdqa rk5(%rip), %xmm10 # rk5 and rk6 in xmm10 - movdqa %xmm7, %xmm0 +.Lreduce_final_16_bytes: + # Reduce the 128-bit value M(x), stored in xmm7, to the final 16-bit CRC - #64b fold - pclmulqdq $0x1, %xmm10, %xmm7 - pslldq $8 , %xmm0 - pxor %xmm0, %xmm7 + # Load 'x^48 * (x^48 mod G(x))' and 'x^48 * (x^80 mod G(x))'. + movdqa .Lfinal_fold_consts(%rip), FOLD_CONSTS - #32b fold + # 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. movdqa %xmm7, %xmm0 + pclmulqdq $0x11, FOLD_CONSTS, %xmm7 # high bits * x^48 * (x^80 mod G(x)) + pslldq $8, %xmm0 + pxor %xmm0, %xmm7 # + low bits * x^64 - pand mask2(%rip), %xmm0 - - psrldq $12, %xmm7 - pclmulqdq $0x10, %xmm10, %xmm7 - pxor %xmm0, %xmm7 - - #barrett reduction -_barrett: - movdqa rk7(%rip), %xmm10 # rk7 and rk8 in xmm10 + # 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. movdqa %xmm7, %xmm0 - pclmulqdq $0x01, %xmm10, %xmm7 - pslldq $4, %xmm7 - pclmulqdq $0x11, %xmm10, %xmm7 + pand .Lmask2(%rip), %xmm0 # zero high 32 bits + psrldq $12, %xmm7 # extract high 32 bits + pclmulqdq $0x00, FOLD_CONSTS, %xmm7 # high 32 bits * x^48 * (x^48 mod G(x)) + pxor %xmm0, %xmm7 # + low bits - pslldq $4, %xmm7 - pxor %xmm0, %xmm7 - pextrd $1, %xmm7, %eax +.Lbarrett_reduction: + # Load G(x) and floor(x^48 / G(x)). + movdqa .Lbarrett_reduction_consts(%rip), FOLD_CONSTS -_cleanup: - # scale the result back to 16 bits - shr $16, %eax - mov %rcx, %rsp + # Use Barrett reduction to compute the final CRC value. + movdqa %xmm7, %xmm0 + pclmulqdq $0x11, FOLD_CONSTS, %xmm7 # high 32 bits * floor(x^48 / G(x)) + psrlq $32, %xmm7 # /= x^32 + pclmulqdq $0x00, FOLD_CONSTS, %xmm7 # *= G(x) + psrlq $48, %xmm0 + pxor %xmm7, %xmm0 # + low 16 nonzero bits + # Final CRC value (x^16 * M(x)) mod G(x) is in low 16 bits of xmm0. + + pextrw $0, %xmm0, %eax +.Ldone: + mov %rcx, %rsp ret -######################################################################## - .align 16 -_less_than_128: +.Lless_than_256_bytes: + pxor %xmm0, %xmm0 + cmp $16, len + jl .Lless_than_16_bytes - # check if there is enough buffer to be able to fold 16B at a time - cmp $32, arg3 - jl _less_than_32 - movdqa SHUF_MASK(%rip), %xmm11 + # Checksumming a buffer of length 16...255 bytes - # now if there is, load the constants - movdqa rk1(%rip), %xmm10 # rk1 and rk2 in xmm10 + # Load the first 16 data bytes. + movdqu (buf), %xmm7 + pshufb BSWAP_MASK, %xmm7 + add $16, buf - movd arg1_low32, %xmm0 # get the initial crc value - pslldq $12, %xmm0 # align it to its correct place - movdqu (arg2), %xmm7 # load the plaintext - pshufb %xmm11, %xmm7 # byte-reflect the plaintext + # XOR the first 16 data *bits* with the initial CRC value. + pinsrw $7, init_crcl, %xmm0 pxor %xmm0, %xmm7 - - # update the buffer pointer - add $16, arg2 - - # update the counter. subtract 32 instead of 16 to save one - # instruction from the loop - sub $32, arg3 - - jmp _16B_reduction_loop - - -.align 16 -_less_than_32: - # mov initial crc to the return value. this is necessary for - # zero-length buffers. - mov arg1_low32, %eax - test arg3, arg3 - je _cleanup - - movdqa SHUF_MASK(%rip), %xmm11 - - movd arg1_low32, %xmm0 # get the initial crc value - pslldq $12, %xmm0 # align it to its correct place - - cmp $16, arg3 - je _exact_16_left - jl _less_than_16_left - - movdqu (arg2), %xmm7 # load the plaintext - pshufb %xmm11, %xmm7 # byte-reflect the plaintext - pxor %xmm0 , %xmm7 # xor the initial crc value - add $16, arg2 - sub $16, arg3 - movdqa rk1(%rip), %xmm10 # rk1 and rk2 in xmm10 - jmp _get_last_two_xmms - + movdqa .Lfold_across_16_bytes_consts(%rip), FOLD_CONSTS + cmp $16, len + je .Lreduce_final_16_bytes # len == 16 + sub $32, len + jge .Lfold_16_bytes_loop # 32 <= len <= 255 + add $16, len + jmp .Lhandle_partial_segment # 17 <= len <= 31 .align 16 -_less_than_16_left: - # use stack space to load data less than 16 bytes, zero-out - # the 16B in memory first. - - pxor %xmm1, %xmm1 - mov %rsp, %r11 - movdqa %xmm1, (%r11) - - cmp $4, arg3 - jl _only_less_than_4 - - # backup the counter value - mov arg3, %r9 - cmp $8, arg3 - jl _less_than_8_left - - # load 8 Bytes - mov (arg2), %rax - mov %rax, (%r11) - add $8, %r11 - sub $8, arg3 - add $8, arg2 -_less_than_8_left: - - cmp $4, arg3 - jl _less_than_4_left - - # load 4 Bytes - mov (arg2), %eax - mov %eax, (%r11) - add $4, %r11 - sub $4, arg3 - add $4, arg2 -_less_than_4_left: - - cmp $2, arg3 - jl _less_than_2_left - - # load 2 Bytes - mov (arg2), %ax - mov %ax, (%r11) - add $2, %r11 - sub $2, arg3 - add $2, arg2 -_less_than_2_left: - cmp $1, arg3 - jl _zero_left - - # load 1 Byte - mov (arg2), %al - mov %al, (%r11) -_zero_left: +.Lless_than_16_bytes: + cmp $2, len + jl .Lless_than_2_bytes + + # Checksumming a buffer of length 2...15 bytes. Copy the data into a + # 16-byte stack buffer, left-padding with zeroes. Then reduce the + # resulting 16-byte value. + + movdqa %xmm0, (%rsp) + lea 16(%rsp), %r8 + sub len, %r8 + mov %r8, %r10 + + test $8, len + jz 1f + movq (buf), %r9 + movq %r9, (%r8) + add $8, buf + add $8, %r8 +1: + test $4, len + jz 1f + movl (buf), %r9d + movl %r9d, (%r8) + add $4, buf + add $4, %r8 +1: + test $2, len + jz 1f + movw (buf), %r9w + movw %r9w, (%r8) + add $2, buf + add $2, %r8 +1: + test $1, len + jz 1f + movb (buf), %r9b + movb %r9b, (%r8) +1: + # XOR the first 16 data *bits* with the initial CRC value. + rol $0x8, init_crc + xor init_crc, (%r10) + + # Load the data and reduce it. movdqa (%rsp), %xmm7 - pshufb %xmm11, %xmm7 - pxor %xmm0 , %xmm7 # xor the initial crc value - - # shl r9, 4 - lea pshufb_shf_table+16(%rip), %rax - sub %r9, %rax - movdqu (%rax), %xmm0 - pxor mask1(%rip), %xmm0 - - pshufb %xmm0, %xmm7 - jmp _128_done + pshufb BSWAP_MASK, %xmm7 + jmp .Lreduce_final_16_bytes .align 16 -_exact_16_left: - movdqu (arg2), %xmm7 - pshufb %xmm11, %xmm7 - pxor %xmm0 , %xmm7 # xor the initial crc value - - jmp _128_done - -_only_less_than_4: - cmp $3, arg3 - jl _only_less_than_3 - - # load 3 Bytes - mov (arg2), %al - mov %al, (%r11) - - mov 1(arg2), %al - mov %al, 1(%r11) - - mov 2(arg2), %al - mov %al, 2(%r11) - - movdqa (%rsp), %xmm7 - pshufb %xmm11, %xmm7 - pxor %xmm0 , %xmm7 # xor the initial crc value - - psrldq $5, %xmm7 - - jmp _barrett -_only_less_than_3: - cmp $2, arg3 - jl _only_less_than_2 - - # load 2 Bytes - mov (arg2), %al - mov %al, (%r11) - - mov 1(arg2), %al - mov %al, 1(%r11) - - movdqa (%rsp), %xmm7 - pshufb %xmm11, %xmm7 - pxor %xmm0 , %xmm7 # xor the initial crc value - - psrldq $6, %xmm7 - - jmp _barrett -_only_less_than_2: - - # load 1 Byte - mov (arg2), %al - mov %al, (%r11) - - movdqa (%rsp), %xmm7 - pshufb %xmm11, %xmm7 - pxor %xmm0 , %xmm7 # xor the initial crc value - - psrldq $7, %xmm7 - - jmp _barrett +.Lless_than_2_bytes: + movzwl init_crc, %eax + test len, len + je .Ldone # len = 0. Nothing to do, just return init_crc. + + # len = 1. This is a special case because with only 1 data byte the + # 16-bit initial CRC value can't just be XOR'ed with the first 16 data + # bits as usual. Instead, directly build the scaled value 'x^64 * M(x)' + # and skip to the Barrett reduction step. + xor (buf), %ah + movd %eax, %xmm7 + pslldq $7, %xmm7 + jmp .Lbarrett_reduction ENDPROC(crc_t10dif_pcl) .section .rodata, "a", @progbits .align 16 -# precomputed constants -# these constants are precomputed from the poly: -# 0x8bb70000 (0x8bb7 scaled to 32 bits) -# 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: -.quad 0x2d56000000000000 -rk2: -.quad 0x06df000000000000 -rk3: -.quad 0x9d9d000000000000 -rk4: -.quad 0x7cf5000000000000 -rk5: -.quad 0x2d56000000000000 -rk6: -.quad 0x1368000000000000 -rk7: -.quad 0x00000001f65a57f8 -rk8: -.quad 0x000000018bb70000 - -rk9: -.quad 0xceae000000000000 -rk10: -.quad 0xbfd6000000000000 -rk11: -.quad 0x1e16000000000000 -rk12: -.quad 0x713c000000000000 -rk13: -.quad 0xf7f9000000000000 -rk14: -.quad 0x80a6000000000000 -rk15: -.quad 0x044c000000000000 -rk16: -.quad 0xe658000000000000 -rk17: -.quad 0xad18000000000000 -rk18: -.quad 0xa497000000000000 -rk19: -.quad 0x6ee3000000000000 -rk20: -.quad 0xe7b5000000000000 - +# 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)) .section .rodata.cst16.mask1, "aM", @progbits, 16 .align 16 -mask1: -.octa 0x80808080808080808080808080808080 +.Lmask1: + .octa 0x80808080808080808080808080808080 .section .rodata.cst16.mask2, "aM", @progbits, 16 .align 16 -mask2: -.octa 0x00000000FFFFFFFFFFFFFFFFFFFFFFFF +.Lmask2: + .octa 0x00000000FFFFFFFFFFFFFFFFFFFFFFFF + +.section .rodata.cst16.bswap_mask, "aM", @progbits, 16 +.align 16 +.Lbswap_mask: + .octa 0x000102030405060708090A0B0C0D0E0F -.section .rodata.cst16.SHUF_MASK, "aM", @progbits, 16 +.section .rodata.cst32.byteshift_table, "aM", @progbits, 32 .align 16 -SHUF_MASK: -.octa 0x000102030405060708090A0B0C0D0E0F - -.section .rodata.cst32.pshufb_shf_table, "aM", @progbits, 32 -.align 32 -pshufb_shf_table: -# use these values for shift constants for the pshufb 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 -.octa 0x8f8e8d8c8b8a89888786858483828100 -.octa 0x000e0d0c0b0a09080706050403020100 +# 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 + .byte 0x8, 0x9, 0xa, 0xb, 0xc, 0xd, 0xe , 0x0 diff --git a/arch/x86/crypto/crct10dif-pclmul_glue.c b/arch/x86/crypto/crct10dif-pclmul_glue.c index cd4df93225014..80cea60e9c6d6 100644 --- a/arch/x86/crypto/crct10dif-pclmul_glue.c +++ b/arch/x86/crypto/crct10dif-pclmul_glue.c @@ -33,8 +33,7 @@ #include #include -asmlinkage __u16 crc_t10dif_pcl(__u16 crc, const unsigned char *buf, - size_t len); +asmlinkage u16 crc_t10dif_pcl(u16 init_crc, const u8 *buf, size_t len); struct chksum_desc_ctx { __u16 crc;