diff options
Diffstat (limited to 'arch/x86')
62 files changed, 9260 insertions, 2292 deletions
diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index c320ce005b02..297789aef9fa 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -1,8 +1,8 @@ # SPDX-License-Identifier: GPL-2.0 # Select 32 or 64 bit config 64BIT - bool "64-bit kernel" if ARCH = "x86" - default ARCH != "i386" + bool "64-bit kernel" if "$(ARCH)" = "x86" + default "$(ARCH)" != "i386" ---help--- Say yes to build a 64-bit kernel - formerly known as x86_64 Say no to build a 32-bit kernel - formerly known as i386 @@ -60,6 +60,7 @@ config X86 select ARCH_HAS_KCOV if X86_64 select ARCH_HAS_MEMBARRIER_SYNC_CORE select ARCH_HAS_PMEM_API if X86_64 + select ARCH_HAS_PTE_SPECIAL select ARCH_HAS_REFCOUNT select ARCH_HAS_UACCESS_FLUSHCACHE if X86_64 select ARCH_HAS_UACCESS_MCSAFE if X86_64 @@ -140,7 +141,7 @@ config X86 select HAVE_DMA_CONTIGUOUS select HAVE_DYNAMIC_FTRACE select HAVE_DYNAMIC_FTRACE_WITH_REGS - select HAVE_EBPF_JIT if X86_64 + select HAVE_EBPF_JIT select HAVE_EFFICIENT_UNALIGNED_ACCESS select HAVE_EXIT_THREAD select HAVE_FENTRY if X86_64 || DYNAMIC_FTRACE @@ -334,6 +335,9 @@ config ARCH_SUPPORTS_UPROBES config FIX_EARLYCON_MEM def_bool y +config DYNAMIC_PHYSICAL_MASK + bool + config PGTABLE_LEVELS int default 5 if X86_5LEVEL @@ -1486,6 +1490,7 @@ config ARCH_HAS_MEM_ENCRYPT config AMD_MEM_ENCRYPT bool "AMD Secure Memory Encryption (SME) support" depends on X86_64 && CPU_SUP_AMD + select DYNAMIC_PHYSICAL_MASK ---help--- Say yes to enable support for the encryption of system memory. This requires an AMD processor that supports Secure Memory diff --git a/arch/x86/Kconfig.debug b/arch/x86/Kconfig.debug index 192e4d2f9efc..c6dd1d980081 100644 --- a/arch/x86/Kconfig.debug +++ b/arch/x86/Kconfig.debug @@ -72,6 +72,9 @@ config EARLY_PRINTK_USB_XDBC You should normally say N here, unless you want to debug early crashes or need a very simple printk logging facility. +config MCSAFE_TEST + def_bool n + config X86_PTDUMP_CORE def_bool n diff --git a/arch/x86/Makefile b/arch/x86/Makefile index 60135cbd905c..f0a6ea22429d 100644 --- a/arch/x86/Makefile +++ b/arch/x86/Makefile @@ -94,7 +94,7 @@ ifeq ($(CONFIG_X86_32),y) else BITS := 64 UTS_MACHINE := x86_64 - CHECKFLAGS += -D__x86_64__ -m64 + CHECKFLAGS += -D__x86_64__ biarch := -m64 KBUILD_AFLAGS += -m64 diff --git a/arch/x86/boot/compressed/kaslr_64.c b/arch/x86/boot/compressed/kaslr_64.c index 522d11431433..748456c365f4 100644 --- a/arch/x86/boot/compressed/kaslr_64.c +++ b/arch/x86/boot/compressed/kaslr_64.c @@ -69,6 +69,8 @@ static struct alloc_pgt_data pgt_data; /* The top level page table entry pointer. */ static unsigned long top_level_pgt; +phys_addr_t physical_mask = (1ULL << __PHYSICAL_MASK_SHIFT) - 1; + /* * Mapping information structure passed to kernel_ident_mapping_init(). * Due to relocation, pointers must be assigned at run time not build time. @@ -81,6 +83,9 @@ void initialize_identity_maps(void) /* If running as an SEV guest, the encryption mask is required. */ set_sev_encryption_mask(); + /* Exclude the encryption mask from __PHYSICAL_MASK */ + physical_mask &= ~sme_me_mask; + /* Init mapping_info with run-time function/buffer pointers. */ mapping_info.alloc_pgt_page = alloc_pgt_page; mapping_info.context = &pgt_data; diff --git a/arch/x86/crypto/Makefile b/arch/x86/crypto/Makefile index 5f07333bb224..a450ad573dcb 100644 --- a/arch/x86/crypto/Makefile +++ b/arch/x86/crypto/Makefile @@ -15,7 +15,6 @@ obj-$(CONFIG_CRYPTO_GLUE_HELPER_X86) += glue_helper.o obj-$(CONFIG_CRYPTO_AES_586) += aes-i586.o obj-$(CONFIG_CRYPTO_TWOFISH_586) += twofish-i586.o -obj-$(CONFIG_CRYPTO_SALSA20_586) += salsa20-i586.o obj-$(CONFIG_CRYPTO_SERPENT_SSE2_586) += serpent-sse2-i586.o obj-$(CONFIG_CRYPTO_AES_X86_64) += aes-x86_64.o @@ -24,7 +23,6 @@ obj-$(CONFIG_CRYPTO_CAMELLIA_X86_64) += camellia-x86_64.o obj-$(CONFIG_CRYPTO_BLOWFISH_X86_64) += blowfish-x86_64.o obj-$(CONFIG_CRYPTO_TWOFISH_X86_64) += twofish-x86_64.o obj-$(CONFIG_CRYPTO_TWOFISH_X86_64_3WAY) += twofish-x86_64-3way.o -obj-$(CONFIG_CRYPTO_SALSA20_X86_64) += salsa20-x86_64.o obj-$(CONFIG_CRYPTO_CHACHA20_X86_64) += chacha20-x86_64.o obj-$(CONFIG_CRYPTO_SERPENT_SSE2_X86_64) += serpent-sse2-x86_64.o obj-$(CONFIG_CRYPTO_AES_NI_INTEL) += aesni-intel.o @@ -38,6 +36,16 @@ obj-$(CONFIG_CRYPTO_SHA512_SSSE3) += sha512-ssse3.o obj-$(CONFIG_CRYPTO_CRCT10DIF_PCLMUL) += crct10dif-pclmul.o obj-$(CONFIG_CRYPTO_POLY1305_X86_64) += poly1305-x86_64.o +obj-$(CONFIG_CRYPTO_AEGIS128_AESNI_SSE2) += aegis128-aesni.o +obj-$(CONFIG_CRYPTO_AEGIS128L_AESNI_SSE2) += aegis128l-aesni.o +obj-$(CONFIG_CRYPTO_AEGIS256_AESNI_SSE2) += aegis256-aesni.o + +obj-$(CONFIG_CRYPTO_MORUS640_GLUE) += morus640_glue.o +obj-$(CONFIG_CRYPTO_MORUS1280_GLUE) += morus1280_glue.o + +obj-$(CONFIG_CRYPTO_MORUS640_SSE2) += morus640-sse2.o +obj-$(CONFIG_CRYPTO_MORUS1280_SSE2) += morus1280-sse2.o + # These modules require assembler to support AVX. ifeq ($(avx_supported),yes) obj-$(CONFIG_CRYPTO_CAMELLIA_AESNI_AVX_X86_64) += \ @@ -55,11 +63,12 @@ ifeq ($(avx2_supported),yes) obj-$(CONFIG_CRYPTO_SHA1_MB) += sha1-mb/ obj-$(CONFIG_CRYPTO_SHA256_MB) += sha256-mb/ obj-$(CONFIG_CRYPTO_SHA512_MB) += sha512-mb/ + + obj-$(CONFIG_CRYPTO_MORUS1280_AVX2) += morus1280-avx2.o endif aes-i586-y := aes-i586-asm_32.o aes_glue.o twofish-i586-y := twofish-i586-asm_32.o twofish_glue.o -salsa20-i586-y := salsa20-i586-asm_32.o salsa20_glue.o serpent-sse2-i586-y := serpent-sse2-i586-asm_32.o serpent_sse2_glue.o aes-x86_64-y := aes-x86_64-asm_64.o aes_glue.o @@ -68,10 +77,16 @@ camellia-x86_64-y := camellia-x86_64-asm_64.o camellia_glue.o blowfish-x86_64-y := blowfish-x86_64-asm_64.o blowfish_glue.o twofish-x86_64-y := twofish-x86_64-asm_64.o twofish_glue.o twofish-x86_64-3way-y := twofish-x86_64-asm_64-3way.o twofish_glue_3way.o -salsa20-x86_64-y := salsa20-x86_64-asm_64.o salsa20_glue.o chacha20-x86_64-y := chacha20-ssse3-x86_64.o chacha20_glue.o serpent-sse2-x86_64-y := serpent-sse2-x86_64-asm_64.o serpent_sse2_glue.o +aegis128-aesni-y := aegis128-aesni-asm.o aegis128-aesni-glue.o +aegis128l-aesni-y := aegis128l-aesni-asm.o aegis128l-aesni-glue.o +aegis256-aesni-y := aegis256-aesni-asm.o aegis256-aesni-glue.o + +morus640-sse2-y := morus640-sse2-asm.o morus640-sse2-glue.o +morus1280-sse2-y := morus1280-sse2-asm.o morus1280-sse2-glue.o + ifeq ($(avx_supported),yes) camellia-aesni-avx-x86_64-y := camellia-aesni-avx-asm_64.o \ camellia_aesni_avx_glue.o @@ -87,6 +102,8 @@ ifeq ($(avx2_supported),yes) camellia-aesni-avx2-y := camellia-aesni-avx2-asm_64.o camellia_aesni_avx2_glue.o chacha20-x86_64-y += chacha20-avx2-x86_64.o serpent-avx2-y := serpent-avx2-asm_64.o serpent_avx2_glue.o + + morus1280-avx2-y := morus1280-avx2-asm.o morus1280-avx2-glue.o endif aesni-intel-y := aesni-intel_asm.o aesni-intel_glue.o fpu.o diff --git a/arch/x86/crypto/aegis128-aesni-asm.S b/arch/x86/crypto/aegis128-aesni-asm.S new file mode 100644 index 000000000000..9254e0b6cc06 --- /dev/null +++ b/arch/x86/crypto/aegis128-aesni-asm.S @@ -0,0 +1,749 @@ +/* + * AES-NI + SSE2 implementation of AEGIS-128 + * + * Copyright (c) 2017-2018 Ondrej Mosnacek <omosnacek@gmail.com> + * Copyright (C) 2017-2018 Red Hat, Inc. All rights reserved. + * + * 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. + */ + +#include <linux/linkage.h> +#include <asm/frame.h> + +#define STATE0 %xmm0 +#define STATE1 %xmm1 +#define STATE2 %xmm2 +#define STATE3 %xmm3 +#define STATE4 %xmm4 +#define KEY %xmm5 +#define MSG %xmm5 +#define T0 %xmm6 +#define T1 %xmm7 + +#define STATEP %rdi +#define LEN %rsi +#define SRC %rdx +#define DST %rcx + +.section .rodata.cst16.aegis128_const, "aM", @progbits, 32 +.align 16 +.Laegis128_const_0: + .byte 0x00, 0x01, 0x01, 0x02, 0x03, 0x05, 0x08, 0x0d + .byte 0x15, 0x22, 0x37, 0x59, 0x90, 0xe9, 0x79, 0x62 +.Laegis128_const_1: + .byte 0xdb, 0x3d, 0x18, 0x55, 0x6d, 0xc2, 0x2f, 0xf1 + .byte 0x20, 0x11, 0x31, 0x42, 0x73, 0xb5, 0x28, 0xdd + +.section .rodata.cst16.aegis128_counter, "aM", @progbits, 16 +.align 16 +.Laegis128_counter: + .byte 0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07 + .byte 0x08, 0x09, 0x0a, 0x0b, 0x0c, 0x0d, 0x0e, 0x0f + +.text + +/* + * aegis128_update + * input: + * STATE[0-4] - input state + * output: + * STATE[0-4] - output state (shifted positions) + * changed: + * T0 + */ +.macro aegis128_update + movdqa STATE4, T0 + aesenc STATE0, STATE4 + aesenc STATE1, STATE0 + aesenc STATE2, STATE1 + aesenc STATE3, STATE2 + aesenc T0, STATE3 +.endm + +/* + * __load_partial: internal ABI + * input: + * LEN - bytes + * SRC - src + * output: + * MSG - message block + * changed: + * T0 + * %r8 + * %r9 + */ +__load_partial: + xor %r9, %r9 + pxor MSG, MSG + + mov LEN, %r8 + and $0x1, %r8 + jz .Lld_partial_1 + + mov LEN, %r8 + and $0x1E, %r8 + add SRC, %r8 + mov (%r8), %r9b + +.Lld_partial_1: + mov LEN, %r8 + and $0x2, %r8 + jz .Lld_partial_2 + + mov LEN, %r8 + and $0x1C, %r8 + add SRC, %r8 + shl $0x10, %r9 + mov (%r8), %r9w + +.Lld_partial_2: + mov LEN, %r8 + and $0x4, %r8 + jz .Lld_partial_4 + + mov LEN, %r8 + and $0x18, %r8 + add SRC, %r8 + shl $32, %r9 + mov (%r8), %r8d + xor %r8, %r9 + +.Lld_partial_4: + movq %r9, MSG + + mov LEN, %r8 + and $0x8, %r8 + jz .Lld_partial_8 + + mov LEN, %r8 + and $0x10, %r8 + add SRC, %r8 + pslldq $8, MSG + movq (%r8), T0 + pxor T0, MSG + +.Lld_partial_8: + ret +ENDPROC(__load_partial) + +/* + * __store_partial: internal ABI + * input: + * LEN - bytes + * DST - dst + * output: + * T0 - message block + * changed: + * %r8 + * %r9 + * %r10 + */ +__store_partial: + mov LEN, %r8 + mov DST, %r9 + + movq T0, %r10 + + cmp $8, %r8 + jl .Lst_partial_8 + + mov %r10, (%r9) + psrldq $8, T0 + movq T0, %r10 + + sub $8, %r8 + add $8, %r9 + +.Lst_partial_8: + cmp $4, %r8 + jl .Lst_partial_4 + + mov %r10d, (%r9) + shr $32, %r10 + + sub $4, %r8 + add $4, %r9 + +.Lst_partial_4: + cmp $2, %r8 + jl .Lst_partial_2 + + mov %r10w, (%r9) + shr $0x10, %r10 + + sub $2, %r8 + add $2, %r9 + +.Lst_partial_2: + cmp $1, %r8 + jl .Lst_partial_1 + + mov %r10b, (%r9) + +.Lst_partial_1: + ret +ENDPROC(__store_partial) + +/* + * void crypto_aegis128_aesni_init(void *state, const void *key, const void *iv); + */ +ENTRY(crypto_aegis128_aesni_init) + FRAME_BEGIN + + /* load IV: */ + movdqu (%rdx), T1 + + /* load key: */ + movdqa (%rsi), KEY + pxor KEY, T1 + movdqa T1, STATE0 + movdqa KEY, STATE3 + movdqa KEY, STATE4 + + /* load the constants: */ + movdqa .Laegis128_const_0, STATE2 + movdqa .Laegis128_const_1, STATE1 + pxor STATE2, STATE3 + pxor STATE1, STATE4 + + /* update 10 times with KEY / KEY xor IV: */ + aegis128_update; pxor KEY, STATE4 + aegis128_update; pxor T1, STATE3 + aegis128_update; pxor KEY, STATE2 + aegis128_update; pxor T1, STATE1 + aegis128_update; pxor KEY, STATE0 + aegis128_update; pxor T1, STATE4 + aegis128_update; pxor KEY, STATE3 + aegis128_update; pxor T1, STATE2 + aegis128_update; pxor KEY, STATE1 + aegis128_update; pxor T1, STATE0 + + /* store the state: */ + movdqu STATE0, 0x00(STATEP) + movdqu STATE1, 0x10(STATEP) + movdqu STATE2, 0x20(STATEP) + movdqu STATE3, 0x30(STATEP) + movdqu STATE4, 0x40(STATEP) + + FRAME_END + ret +ENDPROC(crypto_aegis128_aesni_init) + +/* + * void crypto_aegis128_aesni_ad(void *state, unsigned int length, + * const void *data); + */ +ENTRY(crypto_aegis128_aesni_ad) + FRAME_BEGIN + + cmp $0x10, LEN + jb .Lad_out + + /* load the state: */ + movdqu 0x00(STATEP), STATE0 + movdqu 0x10(STATEP), STATE1 + movdqu 0x20(STATEP), STATE2 + movdqu 0x30(STATEP), STATE3 + movdqu 0x40(STATEP), STATE4 + + mov SRC, %r8 + and $0xF, %r8 + jnz .Lad_u_loop + +.align 8 +.Lad_a_loop: + movdqa 0x00(SRC), MSG + aegis128_update + pxor MSG, STATE4 + sub $0x10, LEN + cmp $0x10, LEN + jl .Lad_out_1 + + movdqa 0x10(SRC), MSG + aegis128_update + pxor MSG, STATE3 + sub $0x10, LEN + cmp $0x10, LEN + jl .Lad_out_2 + + movdqa 0x20(SRC), MSG + aegis128_update + pxor MSG, STATE2 + sub $0x10, LEN + cmp $0x10, LEN + jl .Lad_out_3 + + movdqa 0x30(SRC), MSG + aegis128_update + pxor MSG, STATE1 + sub $0x10, LEN + cmp $0x10, LEN + jl .Lad_out_4 + + movdqa 0x40(SRC), MSG + aegis128_update + pxor MSG, STATE0 + sub $0x10, LEN + cmp $0x10, LEN + jl .Lad_out_0 + + add $0x50, SRC + jmp .Lad_a_loop + +.align 8 +.Lad_u_loop: + movdqu 0x00(SRC), MSG + aegis128_update + pxor MSG, STATE4 + sub $0x10, LEN + cmp $0x10, LEN + jl .Lad_out_1 + + movdqu 0x10(SRC), MSG + aegis128_update + pxor MSG, STATE3 + sub $0x10, LEN + cmp $0x10, LEN + jl .Lad_out_2 + + movdqu 0x20(SRC), MSG + aegis128_update + pxor MSG, STATE2 + sub $0x10, LEN + cmp $0x10, LEN + jl .Lad_out_3 + + movdqu 0x30(SRC), MSG + aegis128_update + pxor MSG, STATE1 + sub $0x10, LEN + cmp $0x10, LEN + jl .Lad_out_4 + + movdqu 0x40(SRC), MSG + aegis128_update + pxor MSG, STATE0 + sub $0x10, LEN + cmp $0x10, LEN + jl .Lad_out_0 + + add $0x50, SRC + jmp .Lad_u_loop + + /* store the state: */ +.Lad_out_0: + movdqu STATE0, 0x00(STATEP) + movdqu STATE1, 0x10(STATEP) + movdqu STATE2, 0x20(STATEP) + movdqu STATE3, 0x30(STATEP) + movdqu STATE4, 0x40(STATEP) + FRAME_END + ret + +.Lad_out_1: + movdqu STATE4, 0x00(STATEP) + movdqu STATE0, 0x10(STATEP) + movdqu STATE1, 0x20(STATEP) + movdqu STATE2, 0x30(STATEP) + movdqu STATE3, 0x40(STATEP) + FRAME_END + ret + +.Lad_out_2: + movdqu STATE3, 0x00(STATEP) + movdqu STATE4, 0x10(STATEP) + movdqu STATE0, 0x20(STATEP) + movdqu STATE1, 0x30(STATEP) + movdqu STATE2, 0x40(STATEP) + FRAME_END + ret + +.Lad_out_3: + movdqu STATE2, 0x00(STATEP) + movdqu STATE3, 0x10(STATEP) + movdqu STATE4, 0x20(STATEP) + movdqu STATE0, 0x30(STATEP) + movdqu STATE1, 0x40(STATEP) + FRAME_END + ret + +.Lad_out_4: + movdqu STATE1, 0x00(STATEP) + movdqu STATE2, 0x10(STATEP) + movdqu STATE3, 0x20(STATEP) + movdqu STATE4, 0x30(STATEP) + movdqu STATE0, 0x40(STATEP) + FRAME_END + ret + +.Lad_out: + FRAME_END + ret +ENDPROC(crypto_aegis128_aesni_ad) + +.macro encrypt_block a s0 s1 s2 s3 s4 i + movdq\a (\i * 0x10)(SRC), MSG + movdqa MSG, T0 + pxor \s1, T0 + pxor \s4, T0 + movdqa \s2, T1 + pand \s3, T1 + pxor T1, T0 + movdq\a T0, (\i * 0x10)(DST) + + aegis128_update + pxor MSG, \s4 + + sub $0x10, LEN + cmp $0x10, LEN + jl .Lenc_out_\i +.endm + +/* + * void crypto_aegis128_aesni_enc(void *state, unsigned int length, + * const void *src, void *dst); + */ +ENTRY(crypto_aegis128_aesni_enc) + FRAME_BEGIN + + cmp $0x10, LEN + jb .Lenc_out + + /* load the state: */ + movdqu 0x00(STATEP), STATE0 + movdqu 0x10(STATEP), STATE1 + movdqu 0x20(STATEP), STATE2 + movdqu 0x30(STATEP), STATE3 + movdqu 0x40(STATEP), STATE4 + + mov SRC, %r8 + or DST, %r8 + and $0xF, %r8 + jnz .Lenc_u_loop + +.align 8 +.Lenc_a_loop: + encrypt_block a STATE0 STATE1 STATE2 STATE3 STATE4 0 + encrypt_block a STATE4 STATE0 STATE1 STATE2 STATE3 1 + encrypt_block a STATE3 STATE4 STATE0 STATE1 STATE2 2 + encrypt_block a STATE2 STATE3 STATE4 STATE0 STATE1 3 + encrypt_block a STATE1 STATE2 STATE3 STATE4 STATE0 4 + + add $0x50, SRC + add $0x50, DST + jmp .Lenc_a_loop + +.align 8 +.Lenc_u_loop: + encrypt_block u STATE0 STATE1 STATE2 STATE3 STATE4 0 + encrypt_block u STATE4 STATE0 STATE1 STATE2 STATE3 1 + encrypt_block u STATE3 STATE4 STATE0 STATE1 STATE2 2 + encrypt_block u STATE2 STATE3 STATE4 STATE0 STATE1 3 + encrypt_block u STATE1 STATE2 STATE3 STATE4 STATE0 4 + + add $0x50, SRC + add $0x50, DST + jmp .Lenc_u_loop + + /* store the state: */ +.Lenc_out_0: + movdqu STATE4, 0x00(STATEP) + movdqu STATE0, 0x10(STATEP) + movdqu STATE1, 0x20(STATEP) + movdqu STATE2, 0x30(STATEP) + movdqu STATE3, 0x40(STATEP) + FRAME_END + ret + +.Lenc_out_1: + movdqu STATE3, 0x00(STATEP) + movdqu STATE4, 0x10(STATEP) + movdqu STATE0, 0x20(STATEP) + movdqu STATE1, 0x30(STATEP) + movdqu STATE2, 0x40(STATEP) + FRAME_END + ret + +.Lenc_out_2: + movdqu STATE2, 0x00(STATEP) + movdqu STATE3, 0x10(STATEP) + movdqu STATE4, 0x20(STATEP) + movdqu STATE0, 0x30(STATEP) + movdqu STATE1, 0x40(STATEP) + FRAME_END + ret + +.Lenc_out_3: + movdqu STATE1, 0x00(STATEP) + movdqu STATE2, 0x10(STATEP) + movdqu STATE3, 0x20(STATEP) + movdqu STATE4, 0x30(STATEP) + movdqu STATE0, 0x40(STATEP) + FRAME_END + ret + +.Lenc_out_4: + movdqu STATE0, 0x00(STATEP) + movdqu STATE1, 0x10(STATEP) + movdqu STATE2, 0x20(STATEP) + movdqu STATE3, 0x30(STATEP) + movdqu STATE4, 0x40(STATEP) + FRAME_END + ret + +.Lenc_out: + FRAME_END + ret +ENDPROC(crypto_aegis128_aesni_enc) + +/* + * void crypto_aegis128_aesni_enc_tail(void *state, unsigned int length, + * const void *src, void *dst); + */ +ENTRY(crypto_aegis128_aesni_enc_tail) + FRAME_BEGIN + + /* load the state: */ + movdqu 0x00(STATEP), STATE0 + movdqu 0x10(STATEP), STATE1 + movdqu 0x20(STATEP), STATE2 + movdqu 0x30(STATEP), STATE3 + movdqu 0x40(STATEP), STATE4 + + /* encrypt message: */ + call __load_partial + + movdqa MSG, T0 + pxor STATE1, T0 + pxor STATE4, T0 + movdqa STATE2, T1 + pand STATE3, T1 + pxor T1, T0 + + call __store_partial + + aegis128_update + pxor MSG, STATE4 + + /* store the state: */ + movdqu STATE4, 0x00(STATEP) + movdqu STATE0, 0x10(STATEP) + movdqu STATE1, 0x20(STATEP) + movdqu STATE2, 0x30(STATEP) + movdqu STATE3, 0x40(STATEP) + + FRAME_END +ENDPROC(crypto_aegis128_aesni_enc_tail) + +.macro decrypt_block a s0 s1 s2 s3 s4 i + movdq\a (\i * 0x10)(SRC), MSG + pxor \s1, MSG + pxor \s4, MSG + movdqa \s2, T1 + pand \s3, T1 + pxor T1, MSG + movdq\a MSG, (\i * 0x10)(DST) + + aegis128_update + pxor MSG, \s4 + + sub $0x10, LEN + cmp $0x10, LEN + jl .Ldec_out_\i +.endm + +/* + * void crypto_aegis128_aesni_dec(void *state, unsigned int length, + * const void *src, void *dst); + */ +ENTRY(crypto_aegis128_aesni_dec) + FRAME_BEGIN + + cmp $0x10, LEN + jb .Ldec_out + + /* load the state: */ + movdqu 0x00(STATEP), STATE0 + movdqu 0x10(STATEP), STATE1 + movdqu 0x20(STATEP), STATE2 + movdqu 0x30(STATEP), STATE3 + movdqu 0x40(STATEP), STATE4 + + mov SRC, %r8 + or DST, %r8 + and $0xF, %r8 + jnz .Ldec_u_loop + +.align 8 +.Ldec_a_loop: + decrypt_block a STATE0 STATE1 STATE2 STATE3 STATE4 0 + decrypt_block a STATE4 STATE0 STATE1 STATE2 STATE3 1 + decrypt_block a STATE3 STATE4 STATE0 STATE1 STATE2 2 + decrypt_block a STATE2 STATE3 STATE4 STATE0 STATE1 3 + decrypt_block a STATE1 STATE2 STATE3 STATE4 STATE0 4 + + add $0x50, SRC + add $0x50, DST + jmp .Ldec_a_loop + +.align 8 +.Ldec_u_loop: + decrypt_block u STATE0 STATE1 STATE2 STATE3 STATE4 0 + decrypt_block u STATE4 STATE0 STATE1 STATE2 STATE3 1 + decrypt_block u STATE3 STATE4 STATE0 STATE1 STATE2 2 + decrypt_block u STATE2 STATE3 STATE4 STATE0 STATE1 3 + decrypt_block u STATE1 STATE2 STATE3 STATE4 STATE0 4 + + add $0x50, SRC + add $0x50, DST + jmp .Ldec_u_loop + + /* store the state: */ +.Ldec_out_0: + movdqu STATE4, 0x00(STATEP) + movdqu STATE0, 0x10(STATEP) + movdqu STATE1, 0x20(STATEP) + movdqu STATE2, 0x30(STATEP) + movdqu STATE3, 0x40(STATEP) + FRAME_END + ret + +.Ldec_out_1: + movdqu STATE3, 0x00(STATEP) + movdqu STATE4, 0x10(STATEP) + movdqu STATE0, 0x20(STATEP) + movdqu STATE1, 0x30(STATEP) + movdqu STATE2, 0x40(STATEP) + FRAME_END + ret + +.Ldec_out_2: + movdqu STATE2, 0x00(STATEP) + movdqu STATE3, 0x10(STATEP) + movdqu STATE4, 0x20(STATEP) + movdqu STATE0, 0x30(STATEP) + movdqu STATE1, 0x40(STATEP) + FRAME_END + ret + +.Ldec_out_3: + movdqu STATE1, 0x00(STATEP) + movdqu STATE2, 0x10(STATEP) + movdqu STATE3, 0x20(STATEP) + movdqu STATE4, 0x30(STATEP) + movdqu STATE0, 0x40(STATEP) + FRAME_END + ret + +.Ldec_out_4: + movdqu STATE0, 0x00(STATEP) + movdqu STATE1, 0x10(STATEP) + movdqu STATE2, 0x20(STATEP) + movdqu STATE3, 0x30(STATEP) + movdqu STATE4, 0x40(STATEP) + FRAME_END + ret + +.Ldec_out: + FRAME_END + ret +ENDPROC(crypto_aegis128_aesni_dec) + +/* + * void crypto_aegis128_aesni_dec_tail(void *state, unsigned int length, + * const void *src, void *dst); + */ +ENTRY(crypto_aegis128_aesni_dec_tail) + FRAME_BEGIN + + /* load the state: */ + movdqu 0x00(STATEP), STATE0 + movdqu 0x10(STATEP), STATE1 + movdqu 0x20(STATEP), STATE2 + movdqu 0x30(STATEP), STATE3 + movdqu 0x40(STATEP), STATE4 + + /* decrypt message: */ + call __load_partial + + pxor STATE1, MSG + pxor STATE4, MSG + movdqa STATE2, T1 + pand STATE3, T1 + pxor T1, MSG + + movdqa MSG, T0 + call __store_partial + + /* mask with byte count: */ + movq LEN, T0 + punpcklbw T0, T0 + punpcklbw T0, T0 + punpcklbw T0, T0 + punpcklbw T0, T0 + movdqa .Laegis128_counter, T1 + pcmpgtb T1, T0 + pand T0, MSG + + aegis128_update + pxor MSG, STATE4 + + /* store the state: */ + movdqu STATE4, 0x00(STATEP) + movdqu STATE0, 0x10(STATEP) + movdqu STATE1, 0x20(STATEP) + movdqu STATE2, 0x30(STATEP) + movdqu STATE3, 0x40(STATEP) + + FRAME_END + ret +ENDPROC(crypto_aegis128_aesni_dec_tail) + +/* + * void crypto_aegis128_aesni_final(void *state, void *tag_xor, + * u64 assoclen, u64 cryptlen); + */ +ENTRY(crypto_aegis128_aesni_final) + FRAME_BEGIN + + /* load the state: */ + movdqu 0x00(STATEP), STATE0 + movdqu 0x10(STATEP), STATE1 + movdqu 0x20(STATEP), STATE2 + movdqu 0x30(STATEP), STATE3 + movdqu 0x40(STATEP), STATE4 + + /* prepare length block: */ + movq %rdx, MSG + movq %rcx, T0 + pslldq $8, T0 + pxor T0, MSG + psllq $3, MSG /* multiply by 8 (to get bit count) */ + + pxor STATE3, MSG + + /* update state: */ + aegis128_update; pxor MSG, STATE4 + aegis128_update; pxor MSG, STATE3 + aegis128_update; pxor MSG, STATE2 + aegis128_update; pxor MSG, STATE1 + aegis128_update; pxor MSG, STATE0 + aegis128_update; pxor MSG, STATE4 + aegis128_update; pxor MSG, STATE3 + + /* xor tag: */ + movdqu (%rsi), MSG + + pxor STATE0, MSG + pxor STATE1, MSG + pxor STATE2, MSG + pxor STATE3, MSG + pxor STATE4, MSG + + movdqu MSG, (%rsi) + + FRAME_END + ret +ENDPROC(crypto_aegis128_aesni_final) diff --git a/arch/x86/crypto/aegis128-aesni-glue.c b/arch/x86/crypto/aegis128-aesni-glue.c new file mode 100644 index 000000000000..5de7c0d46edf --- /dev/null +++ b/arch/x86/crypto/aegis128-aesni-glue.c @@ -0,0 +1,407 @@ +/* + * The AEGIS-128 Authenticated-Encryption Algorithm + * Glue for AES-NI + SSE2 implementation + * + * Copyright (c) 2017-2018 Ondrej Mosnacek <omosnacek@gmail.com> + * Copyright (C) 2017-2018 Red Hat, Inc. All rights reserved. + * + * This program is free software; you can redistribute it and/or modify it + * under the terms of the GNU General Public License as published by the Free + * Software Foundation; either version 2 of the License, or (at your option) + * any later version. + */ + +#include <crypto/cryptd.h> +#include <crypto/internal/aead.h> +#include <crypto/internal/skcipher.h> +#include <crypto/scatterwalk.h> +#include <linux/module.h> +#include <asm/fpu/api.h> +#include <asm/cpu_device_id.h> + +#define AEGIS128_BLOCK_ALIGN 16 +#define AEGIS128_BLOCK_SIZE 16 +#define AEGIS128_NONCE_SIZE 16 +#define AEGIS128_STATE_BLOCKS 5 +#define AEGIS128_KEY_SIZE 16 +#define AEGIS128_MIN_AUTH_SIZE 8 +#define AEGIS128_MAX_AUTH_SIZE 16 + +asmlinkage void crypto_aegis128_aesni_init(void *state, void *key, void *iv); + +asmlinkage void crypto_aegis128_aesni_ad( + void *state, unsigned int length, const void *data); + +asmlinkage void crypto_aegis128_aesni_enc( + void *state, unsigned int length, const void *src, void *dst); + +asmlinkage void crypto_aegis128_aesni_dec( + void *state, unsigned int length, const void *src, void *dst); + +asmlinkage void crypto_aegis128_aesni_enc_tail( + void *state, unsigned int length, const void *src, void *dst); + +asmlinkage void crypto_aegis128_aesni_dec_tail( + void *state, unsigned int length, const void *src, void *dst); + +asmlinkage void crypto_aegis128_aesni_final( + void *state, void *tag_xor, unsigned int cryptlen, + unsigned int assoclen); + +struct aegis_block { + u8 bytes[AEGIS128_BLOCK_SIZE] __aligned(AEGIS128_BLOCK_ALIGN); +}; + +struct aegis_state { + struct aegis_block blocks[AEGIS128_STATE_BLOCKS]; +}; + +struct aegis_ctx { + struct aegis_block key; +}; + +struct aegis_crypt_ops { + int (*skcipher_walk_init)(struct skcipher_walk *walk, + struct aead_request *req, bool atomic); + + void (*crypt_blocks)(void *state, unsigned int length, const void *src, + void *dst); + void (*crypt_tail)(void *state, unsigned int length, const void *src, + void *dst); +}; + +static void crypto_aegis128_aesni_process_ad( + struct aegis_state *state, struct scatterlist *sg_src, + unsigned int assoclen) +{ + struct scatter_walk walk; + struct aegis_block buf; + unsigned int pos = 0; + + scatterwalk_start(&walk, sg_src); + while (assoclen != 0) { + unsigned int size = scatterwalk_clamp(&walk, assoclen); + unsigned int left = size; + void *mapped = scatterwalk_map(&walk); + const u8 *src = (const u8 *)mapped; + + if (pos + size >= AEGIS128_BLOCK_SIZE) { + if (pos > 0) { + unsigned int fill = AEGIS128_BLOCK_SIZE - pos; + memcpy(buf.bytes + pos, src, fill); + crypto_aegis128_aesni_ad(state, + AEGIS128_BLOCK_SIZE, + buf.bytes); + pos = 0; + left -= fill; + src += fill; + } + + crypto_aegis128_aesni_ad(state, left, src); + + src += left & ~(AEGIS128_BLOCK_SIZE - 1); + left &= AEGIS128_BLOCK_SIZE - 1; + } + + memcpy(buf.bytes + pos, src, left); + pos += left; + assoclen -= size; + + scatterwalk_unmap(mapped); + scatterwalk_advance(&walk, size); + scatterwalk_done(&walk, 0, assoclen); + } + + if (pos > 0) { + memset(buf.bytes + pos, 0, AEGIS128_BLOCK_SIZE - pos); + crypto_aegis128_aesni_ad(state, AEGIS128_BLOCK_SIZE, buf.bytes); + } +} + +static void crypto_aegis128_aesni_process_crypt( + struct aegis_state *state, struct aead_request *req, + const struct aegis_crypt_ops *ops) +{ + struct skcipher_walk walk; + u8 *src, *dst; + unsigned int chunksize, base; + + ops->skcipher_walk_init(&walk, req, false); + + while (walk.nbytes) { + src = walk.src.virt.addr; + dst = walk.dst.virt.addr; + chunksize = walk.nbytes; + + ops->crypt_blocks(state, chunksize, src, dst); + + base = chunksize & ~(AEGIS128_BLOCK_SIZE - 1); + src += base; + dst += base; + chunksize &= AEGIS128_BLOCK_SIZE - 1; + + if (chunksize > 0) + ops->crypt_tail(state, chunksize, src, dst); + + skcipher_walk_done(&walk, 0); + } +} + +static struct aegis_ctx *crypto_aegis128_aesni_ctx(struct crypto_aead *aead) +{ + u8 *ctx = crypto_aead_ctx(aead); + ctx = PTR_ALIGN(ctx, __alignof__(struct aegis_ctx)); + return (void *)ctx; +} + +static int crypto_aegis128_aesni_setkey(struct crypto_aead *aead, const u8 *key, + unsigned int keylen) +{ + struct aegis_ctx *ctx = crypto_aegis128_aesni_ctx(aead); + + if (keylen != AEGIS128_KEY_SIZE) { + crypto_aead_set_flags(aead, CRYPTO_TFM_RES_BAD_KEY_LEN); + return -EINVAL; + } + + memcpy(ctx->key.bytes, key, AEGIS128_KEY_SIZE); + + return 0; +} + +static int crypto_aegis128_aesni_setauthsize(struct crypto_aead *tfm, + unsigned int authsize) +{ + if (authsize > AEGIS128_MAX_AUTH_SIZE) + return -EINVAL; + if (authsize < AEGIS128_MIN_AUTH_SIZE) + return -EINVAL; + return 0; +} + +static void crypto_aegis128_aesni_crypt(struct aead_request *req, + struct aegis_block *tag_xor, + unsigned int cryptlen, + const struct aegis_crypt_ops *ops) +{ + struct crypto_aead *tfm = crypto_aead_reqtfm(req); + struct aegis_ctx *ctx = crypto_aegis128_aesni_ctx(tfm); + struct aegis_state state; + + kernel_fpu_begin(); + + crypto_aegis128_aesni_init(&state, ctx->key.bytes, req->iv); + crypto_aegis128_aesni_process_ad(&state, req->src, req->assoclen); + crypto_aegis128_aesni_process_crypt(&state, req, ops); + crypto_aegis128_aesni_final(&state, tag_xor, req->assoclen, cryptlen); + + kernel_fpu_end(); +} + +static int crypto_aegis128_aesni_encrypt(struct aead_request *req) +{ + static const struct aegis_crypt_ops OPS = { + .skcipher_walk_init = skcipher_walk_aead_encrypt, + .crypt_blocks = crypto_aegis128_aesni_enc, + .crypt_tail = crypto_aegis128_aesni_enc_tail, + }; + + struct crypto_aead *tfm = crypto_aead_reqtfm(req); + struct aegis_block tag = {}; + unsigned int authsize = crypto_aead_authsize(tfm); + unsigned int cryptlen = req->cryptlen; + + crypto_aegis128_aesni_crypt(req, &tag, cryptlen, &OPS); + + scatterwalk_map_and_copy(tag.bytes, req->dst, + req->assoclen + cryptlen, authsize, 1); + return 0; +} + +static int crypto_aegis128_aesni_decrypt(struct aead_request *req) +{ + static const struct aegis_block zeros = {}; + + static const struct aegis_crypt_ops OPS = { + .skcipher_walk_init = skcipher_walk_aead_decrypt, + .crypt_blocks = crypto_aegis128_aesni_dec, + .crypt_tail = crypto_aegis128_aesni_dec_tail, + }; + + struct crypto_aead *tfm = crypto_aead_reqtfm(req); + struct aegis_block tag; + unsigned int authsize = crypto_aead_authsize(tfm); + unsigned int cryptlen = req->cryptlen - authsize; + + scatterwalk_map_and_copy(tag.bytes, req->src, + req->assoclen + cryptlen, authsize, 0); + + crypto_aegis128_aesni_crypt(req, &tag, cryptlen, &OPS); + + return crypto_memneq(tag.bytes, zeros.bytes, authsize) ? -EBADMSG : 0; +} + +static int crypto_aegis128_aesni_init_tfm(struct crypto_aead *aead) +{ + return 0; +} + +static void crypto_aegis128_aesni_exit_tfm(struct crypto_aead *aead) +{ +} + +static int cryptd_aegis128_aesni_setkey(struct crypto_aead *aead, + const u8 *key, unsigned int keylen) +{ + struct cryptd_aead **ctx = crypto_aead_ctx(aead); + struct cryptd_aead *cryptd_tfm = *ctx; + + return crypto_aead_setkey(&cryptd_tfm->base, key, keylen); +} + +static int cryptd_aegis128_aesni_setauthsize(struct crypto_aead *aead, + unsigned int authsize) +{ + struct cryptd_aead **ctx = crypto_aead_ctx(aead); + struct cryptd_aead *cryptd_tfm = *ctx; + + return crypto_aead_setauthsize(&cryptd_tfm->base, authsize); +} + +static int cryptd_aegis128_aesni_encrypt(struct aead_request *req) +{ + struct crypto_aead *aead = crypto_aead_reqtfm(req); + struct cryptd_aead **ctx = crypto_aead_ctx(aead); + struct cryptd_aead *cryptd_tfm = *ctx; + + aead = &cryptd_tfm->base; + if (irq_fpu_usable() && (!in_atomic() || + !cryptd_aead_queued(cryptd_tfm))) + aead = cryptd_aead_child(cryptd_tfm); + + aead_request_set_tfm(req, aead); + + return crypto_aead_encrypt(req); +} + +static int cryptd_aegis128_aesni_decrypt(struct aead_request *req) +{ + struct crypto_aead *aead = crypto_aead_reqtfm(req); + struct cryptd_aead **ctx = crypto_aead_ctx(aead); + struct cryptd_aead *cryptd_tfm = *ctx; + + aead = &cryptd_tfm->base; + if (irq_fpu_usable() && (!in_atomic() || + !cryptd_aead_queued(cryptd_tfm))) + aead = cryptd_aead_child(cryptd_tfm); + + aead_request_set_tfm(req, aead); + + return crypto_aead_decrypt(req); +} + +static int cryptd_aegis128_aesni_init_tfm(struct crypto_aead *aead) +{ + struct cryptd_aead *cryptd_tfm; + struct cryptd_aead **ctx = crypto_aead_ctx(aead); + + cryptd_tfm = cryptd_alloc_aead("__aegis128-aesni", CRYPTO_ALG_INTERNAL, + CRYPTO_ALG_INTERNAL); + if (IS_ERR(cryptd_tfm)) + return PTR_ERR(cryptd_tfm); + + *ctx = cryptd_tfm; + crypto_aead_set_reqsize(aead, crypto_aead_reqsize(&cryptd_tfm->base)); + return 0; +} + +static void cryptd_aegis128_aesni_exit_tfm(struct crypto_aead *aead) +{ + struct cryptd_aead **ctx = crypto_aead_ctx(aead); + + cryptd_free_aead(*ctx); +} + +static struct aead_alg crypto_aegis128_aesni_alg[] = { + { + .setkey = crypto_aegis128_aesni_setkey, + .setauthsize = crypto_aegis128_aesni_setauthsize, + .encrypt = crypto_aegis128_aesni_encrypt, + .decrypt = crypto_aegis128_aesni_decrypt, + .init = crypto_aegis128_aesni_init_tfm, + .exit = crypto_aegis128_aesni_exit_tfm, + + .ivsize = AEGIS128_NONCE_SIZE, + .maxauthsize = AEGIS128_MAX_AUTH_SIZE, + .chunksize = AEGIS128_BLOCK_SIZE, + + .base = { + .cra_flags = CRYPTO_ALG_INTERNAL, + .cra_blocksize = 1, + .cra_ctxsize = sizeof(struct aegis_ctx) + + __alignof__(struct aegis_ctx), + .cra_alignmask = 0, + + .cra_name = "__aegis128", + .cra_driver_name = "__aegis128-aesni", + + .cra_module = THIS_MODULE, + } + }, { + .setkey = cryptd_aegis128_aesni_setkey, + .setauthsize = cryptd_aegis128_aesni_setauthsize, + .encrypt = cryptd_aegis128_aesni_encrypt, + .decrypt = cryptd_aegis128_aesni_decrypt, + .init = cryptd_aegis128_aesni_init_tfm, + .exit = cryptd_aegis128_aesni_exit_tfm, + + .ivsize = AEGIS128_NONCE_SIZE, + .maxauthsize = AEGIS128_MAX_AUTH_SIZE, + .chunksize = AEGIS128_BLOCK_SIZE, + + .base = { + .cra_flags = CRYPTO_ALG_ASYNC, + .cra_blocksize = 1, + .cra_ctxsize = sizeof(struct cryptd_aead *), + .cra_alignmask = 0, + + .cra_priority = 400, + + .cra_name = "aegis128", + .cra_driver_name = "aegis128-aesni", + + .cra_module = THIS_MODULE, + } + } +}; + +static const struct x86_cpu_id aesni_cpu_id[] = { + X86_FEATURE_MATCH(X86_FEATURE_AES), + X86_FEATURE_MATCH(X86_FEATURE_XMM2), + {} +}; +MODULE_DEVICE_TABLE(x86cpu, aesni_cpu_id); + +static int __init crypto_aegis128_aesni_module_init(void) +{ + if (!x86_match_cpu(aesni_cpu_id)) + return -ENODEV; + + return crypto_register_aeads(crypto_aegis128_aesni_alg, + ARRAY_SIZE(crypto_aegis128_aesni_alg)); +} + +static void __exit crypto_aegis128_aesni_module_exit(void) +{ + crypto_unregister_aeads(crypto_aegis128_aesni_alg, + ARRAY_SIZE(crypto_aegis128_aesni_alg)); +} + +module_init(crypto_aegis128_aesni_module_init); +module_exit(crypto_aegis128_aesni_module_exit); + +MODULE_LICENSE("GPL"); +MODULE_AUTHOR("Ondrej Mosnacek <omosnacek@gmail.com>"); +MODULE_DESCRIPTION("AEGIS-128 AEAD algorithm -- AESNI+SSE2 implementation"); +MODULE_ALIAS_CRYPTO("aegis128"); +MODULE_ALIAS_CRYPTO("aegis128-aesni"); diff --git a/arch/x86/crypto/aegis128l-aesni-asm.S b/arch/x86/crypto/aegis128l-aesni-asm.S new file mode 100644 index 000000000000..9263c344f2c7 --- /dev/null +++ b/arch/x86/crypto/aegis128l-aesni-asm.S @@ -0,0 +1,825 @@ +/* + * AES-NI + SSE2 implementation of AEGIS-128L + * + * Copyright (c) 2017-2018 Ondrej Mosnacek <omosnacek@gmail.com> + * Copyright (C) 2017-2018 Red Hat, Inc. All rights reserved. + * + * 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. + */ + +#include <linux/linkage.h> +#include <asm/frame.h> + +#define STATE0 %xmm0 +#define STATE1 %xmm1 +#define STATE2 %xmm2 +#define STATE3 %xmm3 +#define STATE4 %xmm4 +#define STATE5 %xmm5 +#define STATE6 %xmm6 +#define STATE7 %xmm7 +#define MSG0 %xmm8 +#define MSG1 %xmm9 +#define T0 %xmm10 +#define T1 %xmm11 +#define T2 %xmm12 +#define T3 %xmm13 + +#define STATEP %rdi +#define LEN %rsi +#define SRC %rdx +#define DST %rcx + +.section .rodata.cst16.aegis128l_const, "aM", @progbits, 32 +.align 16 +.Laegis128l_const_0: + .byte 0x00, 0x01, 0x01, 0x02, 0x03, 0x05, 0x08, 0x0d + .byte 0x15, 0x22, 0x37, 0x59, 0x90, 0xe9, 0x79, 0x62 +.Laegis128l_const_1: + .byte 0xdb, 0x3d, 0x18, 0x55, 0x6d, 0xc2, 0x2f, 0xf1 + .byte 0x20, 0x11, 0x31, 0x42, 0x73, 0xb5, 0x28, 0xdd + +.section .rodata.cst16.aegis128l_counter, "aM", @progbits, 16 +.align 16 +.Laegis128l_counter0: + .byte 0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07 + .byte 0x08, 0x09, 0x0a, 0x0b, 0x0c, 0x0d, 0x0e, 0x0f +.Laegis128l_counter1: + .byte 0x10, 0x11, 0x12, 0x13, 0x14, 0x15, 0x16, 0x17 + .byte 0x18, 0x19, 0x1a, 0x1b, 0x1c, 0x1d, 0x1e, 0x1f + +.text + +/* + * __load_partial: internal ABI + * input: + * LEN - bytes + * SRC - src + * output: + * MSG0 - first message block + * MSG1 - second message block + * changed: + * T0 + * %r8 + * %r9 + */ +__load_partial: + xor %r9, %r9 + pxor MSG0, MSG0 + pxor MSG1, MSG1 + + mov LEN, %r8 + and $0x1, %r8 + jz .Lld_partial_1 + + mov LEN, %r8 + and $0x1E, %r8 + add SRC, %r8 + mov (%r8), %r9b + +.Lld_partial_1: + mov LEN, %r8 + and $0x2, %r8 + jz .Lld_partial_2 + + mov LEN, %r8 + and $0x1C, %r8 + add SRC, %r8 + shl $0x10, %r9 + mov (%r8), %r9w + +.Lld_partial_2: + mov LEN, %r8 + and $0x4, %r8 + jz .Lld_partial_4 + + mov LEN, %r8 + and $0x18, %r8 + add SRC, %r8 + shl $32, %r9 + mov (%r8), %r8d + xor %r8, %r9 + +.Lld_partial_4: + movq %r9, MSG0 + + mov LEN, %r8 + and $0x8, %r8 + jz .Lld_partial_8 + + mov LEN, %r8 + and $0x10, %r8 + add SRC, %r8 + pslldq $8, MSG0 + movq (%r8), T0 + pxor T0, MSG0 + +.Lld_partial_8: + mov LEN, %r8 + and $0x10, %r8 + jz .Lld_partial_16 + + movdqa MSG0, MSG1 + movdqu (SRC), MSG0 + +.Lld_partial_16: + ret +ENDPROC(__load_partial) + +/* + * __store_partial: internal ABI + * input: + * LEN - bytes + * DST - dst + * output: + * T0 - first message block + * T1 - second message block + * changed: + * %r8 + * %r9 + * %r10 + */ +__store_partial: + mov LEN, %r8 + mov DST, %r9 + + cmp $16, %r8 + jl .Lst_partial_16 + + movdqu T0, (%r9) + movdqa T1, T0 + + sub $16, %r8 + add $16, %r9 + +.Lst_partial_16: + movq T0, %r10 + + cmp $8, %r8 + jl .Lst_partial_8 + + mov %r10, (%r9) + psrldq $8, T0 + movq T0, %r10 + + sub $8, %r8 + add $8, %r9 + +.Lst_partial_8: + cmp $4, %r8 + jl .Lst_partial_4 + + mov %r10d, (%r9) + shr $32, %r10 + + sub $4, %r8 + add $4, %r9 + +.Lst_partial_4: + cmp $2, %r8 + jl .Lst_partial_2 + + mov %r10w, (%r9) + shr $0x10, %r10 + + sub $2, %r8 + add $2, %r9 + +.Lst_partial_2: + cmp $1, %r8 + jl .Lst_partial_1 + + mov %r10b, (%r9) + +.Lst_partial_1: + ret +ENDPROC(__store_partial) + +.macro update + movdqa STATE7, T0 + aesenc STATE0, STATE7 + aesenc STATE1, STATE0 + aesenc STATE2, STATE1 + aesenc STATE3, STATE2 + aesenc STATE4, STATE3 + aesenc STATE5, STATE4 + aesenc STATE6, STATE5 + aesenc T0, STATE6 +.endm + +.macro update0 + update + pxor MSG0, STATE7 + pxor MSG1, STATE3 +.endm + +.macro update1 + update + pxor MSG0, STATE6 + pxor MSG1, STATE2 +.endm + +.macro update2 + update + pxor MSG0, STATE5 + pxor MSG1, STATE1 +.endm + +.macro update3 + update + pxor MSG0, STATE4 + pxor MSG1, STATE0 +.endm + +.macro update4 + update + pxor MSG0, STATE3 + pxor MSG1, STATE7 +.endm + +.macro update5 + update + pxor MSG0, STATE2 + pxor MSG1, STATE6 +.endm + +.macro update6 + update + pxor MSG0, STATE1 + pxor MSG1, STATE5 +.endm + +.macro update7 + update + pxor MSG0, STATE0 + pxor MSG1, STATE4 +.endm + +.macro state_load + movdqu 0x00(STATEP), STATE0 + movdqu 0x10(STATEP), STATE1 + movdqu 0x20(STATEP), STATE2 + movdqu 0x30(STATEP), STATE3 + movdqu 0x40(STATEP), STATE4 + movdqu 0x50(STATEP), STATE5 + movdqu 0x60(STATEP), STATE6 + movdqu 0x70(STATEP), STATE7 +.endm + +.macro state_store s0 s1 s2 s3 s4 s5 s6 s7 + movdqu \s7, 0x00(STATEP) + movdqu \s0, 0x10(STATEP) + movdqu \s1, 0x20(STATEP) + movdqu \s2, 0x30(STATEP) + movdqu \s3, 0x40(STATEP) + movdqu \s4, 0x50(STATEP) + movdqu \s5, 0x60(STATEP) + movdqu \s6, 0x70(STATEP) +.endm + +.macro state_store0 + state_store STATE0 STATE1 STATE2 STATE3 STATE4 STATE5 STATE6 STATE7 +.endm + +.macro state_store1 + state_store STATE7 STATE0 STATE1 STATE2 STATE3 STATE4 STATE5 STATE6 +.endm + +.macro state_store2 + state_store STATE6 STATE7 STATE0 STATE1 STATE2 STATE3 STATE4 STATE5 +.endm + +.macro state_store3 + state_store STATE5 STATE6 STATE7 STATE0 STATE1 STATE2 STATE3 STATE4 +.endm + +.macro state_store4 + state_store STATE4 STATE5 STATE6 STATE7 STATE0 STATE1 STATE2 STATE3 +.endm + +.macro state_store5 + state_store STATE3 STATE4 STATE5 STATE6 STATE7 STATE0 STATE1 STATE2 +.endm + +.macro state_store6 + state_store STATE2 STATE3 STATE4 STATE5 STATE6 STATE7 STATE0 STATE1 +.endm + +.macro state_store7 + state_store STATE1 STATE2 STATE3 STATE4 STATE5 STATE6 STATE7 STATE0 +.endm + +/* + * void crypto_aegis128l_aesni_init(void *state, const void *key, const void *iv); + */ +ENTRY(crypto_aegis128l_aesni_init) + FRAME_BEGIN + + /* load key: */ + movdqa (%rsi), MSG1 + movdqa MSG1, STATE0 + movdqa MSG1, STATE4 + movdqa MSG1, STATE5 + movdqa MSG1, STATE6 + movdqa MSG1, STATE7 + + /* load IV: */ + movdqu (%rdx), MSG0 + pxor MSG0, STATE0 + pxor MSG0, STATE4 + + /* load the constants: */ + movdqa .Laegis128l_const_0, STATE2 + movdqa .Laegis128l_const_1, STATE1 + movdqa STATE1, STATE3 + pxor STATE2, STATE5 + pxor STATE1, STATE6 + pxor STATE2, STATE7 + + /* update 10 times with IV and KEY: */ + update0 + update1 + update2 + update3 + update4 + update5 + update6 + update7 + update0 + update1 + + state_store1 + + FRAME_END + ret +ENDPROC(crypto_aegis128l_aesni_init) + +.macro ad_block a i + movdq\a (\i * 0x20 + 0x00)(SRC), MSG0 + movdq\a (\i * 0x20 + 0x10)(SRC), MSG1 + update\i + sub $0x20, LEN + cmp $0x20, LEN + jl .Lad_out_\i +.endm + +/* + * void crypto_aegis128l_aesni_ad(void *state, unsigned int length, + * const void *data); + */ +ENTRY(crypto_aegis128l_aesni_ad) + FRAME_BEGIN + + cmp $0x20, LEN + jb .Lad_out + + state_load + + mov SRC, %r8 + and $0xf, %r8 + jnz .Lad_u_loop + +.align 8 +.Lad_a_loop: + ad_block a 0 + ad_block a 1 + ad_block a 2 + ad_block a 3 + ad_block a 4 + ad_block a 5 + ad_block a 6 + ad_block a 7 + + add $0x100, SRC + jmp .Lad_a_loop + +.align 8 +.Lad_u_loop: + ad_block u 0 + ad_block u 1 + ad_block u 2 + ad_block u 3 + ad_block u 4 + ad_block u 5 + ad_block u 6 + ad_block u 7 + + add $0x100, SRC + jmp .Lad_u_loop + +.Lad_out_0: + state_store0 + FRAME_END + ret + +.Lad_out_1: + state_store1 + FRAME_END + ret + +.Lad_out_2: + state_store2 + FRAME_END + ret + +.Lad_out_3: + state_store3 + FRAME_END + ret + +.Lad_out_4: + state_store4 + FRAME_END + ret + +.Lad_out_5: + state_store5 + FRAME_END + ret + +.Lad_out_6: + state_store6 + FRAME_END + ret + +.Lad_out_7: + state_store7 + FRAME_END + ret + +.Lad_out: + FRAME_END + ret +ENDPROC(crypto_aegis128l_aesni_ad) + +.macro crypt m0 m1 s0 s1 s2 s3 s4 s5 s6 s7 + pxor \s1, \m0 + pxor \s6, \m0 + movdqa \s2, T3 + pand \s3, T3 + pxor T3, \m0 + + pxor \s2, \m1 + pxor \s5, \m1 + movdqa \s6, T3 + pand \s7, T3 + pxor T3, \m1 +.endm + +.macro crypt0 m0 m1 + crypt \m0 \m1 STATE0 STATE1 STATE2 STATE3 STATE4 STATE5 STATE6 STATE7 +.endm + +.macro crypt1 m0 m1 + crypt \m0 \m1 STATE7 STATE0 STATE1 STATE2 STATE3 STATE4 STATE5 STATE6 +.endm + +.macro crypt2 m0 m1 + crypt \m0 \m1 STATE6 STATE7 STATE0 STATE1 STATE2 STATE3 STATE4 STATE5 +.endm + +.macro crypt3 m0 m1 + crypt \m0 \m1 STATE5 STATE6 STATE7 STATE0 STATE1 STATE2 STATE3 STATE4 +.endm + +.macro crypt4 m0 m1 + crypt \m0 \m1 STATE4 STATE5 STATE6 STATE7 STATE0 STATE1 STATE2 STATE3 +.endm + +.macro crypt5 m0 m1 + crypt \m0 \m1 STATE3 STATE4 STATE5 STATE6 STATE7 STATE0 STATE1 STATE2 +.endm + +.macro crypt6 m0 m1 + crypt \m0 \m1 STATE2 STATE3 STATE4 STATE5 STATE6 STATE7 STATE0 STATE1 +.endm + +.macro crypt7 m0 m1 + crypt \m0 \m1 STATE1 STATE2 STATE3 STATE4 STATE5 STATE6 STATE7 STATE0 +.endm + +.macro encrypt_block a i + movdq\a (\i * 0x20 + 0x00)(SRC), MSG0 + movdq\a (\i * 0x20 + 0x10)(SRC), MSG1 + movdqa MSG0, T0 + movdqa MSG1, T1 + crypt\i T0, T1 + movdq\a T0, (\i * 0x20 + 0x00)(DST) + movdq\a T1, (\i * 0x20 + 0x10)(DST) + + update\i + + sub $0x20, LEN + cmp $0x20, LEN + jl .Lenc_out_\i +.endm + +.macro decrypt_block a i + movdq\a (\i * 0x20 + 0x00)(SRC), MSG0 + movdq\a (\i * 0x20 + 0x10)(SRC), MSG1 + crypt\i MSG0, MSG1 + movdq\a MSG0, (\i * 0x20 + 0x00)(DST) + movdq\a MSG1, (\i * 0x20 + 0x10)(DST) + + update\i + + sub $0x20, LEN + cmp $0x20, LEN + jl .Ldec_out_\i +.endm + +/* + * void crypto_aegis128l_aesni_enc(void *state, unsigned int length, + * const void *src, void *dst); + */ +ENTRY(crypto_aegis128l_aesni_enc) + FRAME_BEGIN + + cmp $0x20, LEN + jb .Lenc_out + + state_load + + mov SRC, %r8 + or DST, %r8 + and $0xf, %r8 + jnz .Lenc_u_loop + +.align 8 +.Lenc_a_loop: + encrypt_block a 0 + encrypt_block a 1 + encrypt_block a 2 + encrypt_block a 3 + encrypt_block a 4 + encrypt_block a 5 + encrypt_block a 6 + encrypt_block a 7 + + add $0x100, SRC + add $0x100, DST + jmp .Lenc_a_loop + +.align 8 +.Lenc_u_loop: + encrypt_block u 0 + encrypt_block u 1 + encrypt_block u 2 + encrypt_block u 3 + encrypt_block u 4 + encrypt_block u 5 + encrypt_block u 6 + encrypt_block u 7 + + add $0x100, SRC + add $0x100, DST + jmp .Lenc_u_loop + +.Lenc_out_0: + state_store0 + FRAME_END + ret + +.Lenc_out_1: + state_store1 + FRAME_END + ret + +.Lenc_out_2: + state_store2 + FRAME_END + ret + +.Lenc_out_3: + state_store3 + FRAME_END + ret + +.Lenc_out_4: + state_store4 + FRAME_END + ret + +.Lenc_out_5: + state_store5 + FRAME_END + ret + +.Lenc_out_6: + state_store6 + FRAME_END + ret + +.Lenc_out_7: + state_store7 + FRAME_END + ret + +.Lenc_out: + FRAME_END + ret +ENDPROC(crypto_aegis128l_aesni_enc) + +/* + * void crypto_aegis128l_aesni_enc_tail(void *state, unsigned int length, + * const void *src, void *dst); + */ +ENTRY(crypto_aegis128l_aesni_enc_tail) + FRAME_BEGIN + + state_load + + /* encrypt message: */ + call __load_partial + + movdqa MSG0, T0 + movdqa MSG1, T1 + crypt0 T0, T1 + + call __store_partial + + update0 + + state_store0 + + FRAME_END +ENDPROC(crypto_aegis128l_aesni_enc_tail) + +/* + * void crypto_aegis128l_aesni_dec(void *state, unsigned int length, + * const void *src, void *dst); + */ +ENTRY(crypto_aegis128l_aesni_dec) + FRAME_BEGIN + + cmp $0x20, LEN + jb .Ldec_out + + state_load + + mov SRC, %r8 + or DST, %r8 + and $0xF, %r8 + jnz .Ldec_u_loop + +.align 8 +.Ldec_a_loop: + decrypt_block a 0 + decrypt_block a 1 + decrypt_block a 2 + decrypt_block a 3 + decrypt_block a 4 + decrypt_block a 5 + decrypt_block a 6 + decrypt_block a 7 + + add $0x100, SRC + add $0x100, DST + jmp .Ldec_a_loop + +.align 8 +.Ldec_u_loop: + decrypt_block u 0 + decrypt_block u 1 + decrypt_block u 2 + decrypt_block u 3 + decrypt_block u 4 + decrypt_block u 5 + decrypt_block u 6 + decrypt_block u 7 + + add $0x100, SRC + add $0x100, DST + jmp .Ldec_u_loop + +.Ldec_out_0: + state_store0 + FRAME_END + ret + +.Ldec_out_1: + state_store1 + FRAME_END + ret + +.Ldec_out_2: + state_store2 + FRAME_END + ret + +.Ldec_out_3: + state_store3 + FRAME_END + ret + +.Ldec_out_4: + state_store4 + FRAME_END + ret + +.Ldec_out_5: + state_store5 + FRAME_END + ret + +.Ldec_out_6: + state_store6 + FRAME_END + ret + +.Ldec_out_7: + state_store7 + FRAME_END + ret + +.Ldec_out: + FRAME_END + ret +ENDPROC(crypto_aegis128l_aesni_dec) + +/* + * void crypto_aegis128l_aesni_dec_tail(void *state, unsigned int length, + * const void *src, void *dst); + */ +ENTRY(crypto_aegis128l_aesni_dec_tail) + FRAME_BEGIN + + state_load + + /* decrypt message: */ + call __load_partial + + crypt0 MSG0, MSG1 + + movdqa MSG0, T0 + movdqa MSG1, T1 + call __store_partial + + /* mask with byte count: */ + movq LEN, T0 + punpcklbw T0, T0 + punpcklbw T0, T0 + punpcklbw T0, T0 + punpcklbw T0, T0 + movdqa T0, T1 + movdqa .Laegis128l_counter0, T2 + movdqa .Laegis128l_counter1, T3 + pcmpgtb T2, T0 + pcmpgtb T3, T1 + pand T0, MSG0 + pand T1, MSG1 + + update0 + + state_store0 + + FRAME_END + ret +ENDPROC(crypto_aegis128l_aesni_dec_tail) + +/* + * void crypto_aegis128l_aesni_final(void *state, void *tag_xor, + * u64 assoclen, u64 cryptlen); + */ +ENTRY(crypto_aegis128l_aesni_final) + FRAME_BEGIN + + state_load + + /* prepare length block: */ + movq %rdx, MSG0 + movq %rcx, T0 + pslldq $8, T0 + pxor T0, MSG0 + psllq $3, MSG0 /* multiply by 8 (to get bit count) */ + + pxor STATE2, MSG0 + movdqa MSG0, MSG1 + + /* update state: */ + update0 + update1 + update2 + update3 + update4 + update5 + update6 + + /* xor tag: */ + movdqu (%rsi), T0 + + pxor STATE1, T0 + pxor STATE2, T0 + pxor STATE3, T0 + pxor STATE4, T0 + pxor STATE5, T0 + pxor STATE6, T0 + pxor STATE7, T0 + + movdqu T0, (%rsi) + + FRAME_END + ret +ENDPROC(crypto_aegis128l_aesni_final) diff --git a/arch/x86/crypto/aegis128l-aesni-glue.c b/arch/x86/crypto/aegis128l-aesni-glue.c new file mode 100644 index 000000000000..876e4866e633 --- /dev/null +++ b/arch/x86/crypto/aegis128l-aesni-glue.c @@ -0,0 +1,407 @@ +/* + * The AEGIS-128L Authenticated-Encryption Algorithm + * Glue for AES-NI + SSE2 implementation + * + * Copyright (c) 2017-2018 Ondrej Mosnacek <omosnacek@gmail.com> + * Copyright (C) 2017-2018 Red Hat, Inc. All rights reserved. + * + * This program is free software; you can redistribute it and/or modify it + * under the terms of the GNU General Public License as published by the Free + * Software Foundation; either version 2 of the License, or (at your option) + * any later version. + */ + +#include <crypto/cryptd.h> +#include <crypto/internal/aead.h> +#include <crypto/internal/skcipher.h> +#include <crypto/scatterwalk.h> +#include <linux/module.h> +#include <asm/fpu/api.h> +#include <asm/cpu_device_id.h> + +#define AEGIS128L_BLOCK_ALIGN 16 +#define AEGIS128L_BLOCK_SIZE 32 +#define AEGIS128L_NONCE_SIZE 16 +#define AEGIS128L_STATE_BLOCKS 8 +#define AEGIS128L_KEY_SIZE 16 +#define AEGIS128L_MIN_AUTH_SIZE 8 +#define AEGIS128L_MAX_AUTH_SIZE 16 + +asmlinkage void crypto_aegis128l_aesni_init(void *state, void *key, void *iv); + +asmlinkage void crypto_aegis128l_aesni_ad( + void *state, unsigned int length, const void *data); + +asmlinkage void crypto_aegis128l_aesni_enc( + void *state, unsigned int length, const void *src, void *dst); + +asmlinkage void crypto_aegis128l_aesni_dec( + void *state, unsigned int length, const void *src, void *dst); + +asmlinkage void crypto_aegis128l_aesni_enc_tail( + void *state, unsigned int length, const void *src, void *dst); + +asmlinkage void crypto_aegis128l_aesni_dec_tail( + void *state, unsigned int length, const void *src, void *dst); + +asmlinkage void crypto_aegis128l_aesni_final( + void *state, void *tag_xor, unsigned int cryptlen, + unsigned int assoclen); + +struct aegis_block { + u8 bytes[AEGIS128L_BLOCK_SIZE] __aligned(AEGIS128L_BLOCK_ALIGN); +}; + +struct aegis_state { + struct aegis_block blocks[AEGIS128L_STATE_BLOCKS]; +}; + +struct aegis_ctx { + struct aegis_block key; +}; + +struct aegis_crypt_ops { + int (*skcipher_walk_init)(struct skcipher_walk *walk, + struct aead_request *req, bool atomic); + + void (*crypt_blocks)(void *state, unsigned int length, const void *src, + void *dst); + void (*crypt_tail)(void *state, unsigned int length, const void *src, + void *dst); +}; + +static void crypto_aegis128l_aesni_process_ad( + struct aegis_state *state, struct scatterlist *sg_src, + unsigned int assoclen) +{ + struct scatter_walk walk; + struct aegis_block buf; + unsigned int pos = 0; + + scatterwalk_start(&walk, sg_src); + while (assoclen != 0) { + unsigned int size = scatterwalk_clamp(&walk, assoclen); + unsigned int left = size; + void *mapped = scatterwalk_map(&walk); + const u8 *src = (const u8 *)mapped; + + if (pos + size >= AEGIS128L_BLOCK_SIZE) { + if (pos > 0) { + unsigned int fill = AEGIS128L_BLOCK_SIZE - pos; + memcpy(buf.bytes + pos, src, fill); + crypto_aegis128l_aesni_ad(state, + AEGIS128L_BLOCK_SIZE, + buf.bytes); + pos = 0; + left -= fill; + src += fill; + } + + crypto_aegis128l_aesni_ad(state, left, src); + + src += left & ~(AEGIS128L_BLOCK_SIZE - 1); + left &= AEGIS128L_BLOCK_SIZE - 1; + } + + memcpy(buf.bytes + pos, src, left); + pos += left; + assoclen -= size; + + scatterwalk_unmap(mapped); + scatterwalk_advance(&walk, size); + scatterwalk_done(&walk, 0, assoclen); + } + + if (pos > 0) { + memset(buf.bytes + pos, 0, AEGIS128L_BLOCK_SIZE - pos); + crypto_aegis128l_aesni_ad(state, AEGIS128L_BLOCK_SIZE, buf.bytes); + } +} + +static void crypto_aegis128l_aesni_process_crypt( + struct aegis_state *state, struct aead_request *req, + const struct aegis_crypt_ops *ops) +{ + struct skcipher_walk walk; + u8 *src, *dst; + unsigned int chunksize, base; + + ops->skcipher_walk_init(&walk, req, false); + + while (walk.nbytes) { + src = walk.src.virt.addr; + dst = walk.dst.virt.addr; + chunksize = walk.nbytes; + + ops->crypt_blocks(state, chunksize, src, dst); + + base = chunksize & ~(AEGIS128L_BLOCK_SIZE - 1); + src += base; + dst += base; + chunksize &= AEGIS128L_BLOCK_SIZE - 1; + + if (chunksize > 0) + ops->crypt_tail(state, chunksize, src, dst); + + skcipher_walk_done(&walk, 0); + } +} + +static struct aegis_ctx *crypto_aegis128l_aesni_ctx(struct crypto_aead *aead) +{ + u8 *ctx = crypto_aead_ctx(aead); + ctx = PTR_ALIGN(ctx, __alignof__(struct aegis_ctx)); + return (void *)ctx; +} + +static int crypto_aegis128l_aesni_setkey(struct crypto_aead *aead, + const u8 *key, unsigned int keylen) +{ + struct aegis_ctx *ctx = crypto_aegis128l_aesni_ctx(aead); + + if (keylen != AEGIS128L_KEY_SIZE) { + crypto_aead_set_flags(aead, CRYPTO_TFM_RES_BAD_KEY_LEN); + return -EINVAL; + } + + memcpy(ctx->key.bytes, key, AEGIS128L_KEY_SIZE); + + return 0; +} + +static int crypto_aegis128l_aesni_setauthsize(struct crypto_aead *tfm, + unsigned int authsize) +{ + if (authsize > AEGIS128L_MAX_AUTH_SIZE) + return -EINVAL; + if (authsize < AEGIS128L_MIN_AUTH_SIZE) + return -EINVAL; + return 0; +} + +static void crypto_aegis128l_aesni_crypt(struct aead_request *req, + struct aegis_block *tag_xor, + unsigned int cryptlen, + const struct aegis_crypt_ops *ops) +{ + struct crypto_aead *tfm = crypto_aead_reqtfm(req); + struct aegis_ctx *ctx = crypto_aegis128l_aesni_ctx(tfm); + struct aegis_state state; + + kernel_fpu_begin(); + + crypto_aegis128l_aesni_init(&state, ctx->key.bytes, req->iv); + crypto_aegis128l_aesni_process_ad(&state, req->src, req->assoclen); + crypto_aegis128l_aesni_process_crypt(&state, req, ops); + crypto_aegis128l_aesni_final(&state, tag_xor, req->assoclen, cryptlen); + + kernel_fpu_end(); +} + +static int crypto_aegis128l_aesni_encrypt(struct aead_request *req) +{ + static const struct aegis_crypt_ops OPS = { + .skcipher_walk_init = skcipher_walk_aead_encrypt, + .crypt_blocks = crypto_aegis128l_aesni_enc, + .crypt_tail = crypto_aegis128l_aesni_enc_tail, + }; + + struct crypto_aead *tfm = crypto_aead_reqtfm(req); + struct aegis_block tag = {}; + unsigned int authsize = crypto_aead_authsize(tfm); + unsigned int cryptlen = req->cryptlen; + + crypto_aegis128l_aesni_crypt(req, &tag, cryptlen, &OPS); + + scatterwalk_map_and_copy(tag.bytes, req->dst, + req->assoclen + cryptlen, authsize, 1); + return 0; +} + +static int crypto_aegis128l_aesni_decrypt(struct aead_request *req) +{ + static const struct aegis_block zeros = {}; + + static const struct aegis_crypt_ops OPS = { + .skcipher_walk_init = skcipher_walk_aead_decrypt, + .crypt_blocks = crypto_aegis128l_aesni_dec, + .crypt_tail = crypto_aegis128l_aesni_dec_tail, + }; + + struct crypto_aead *tfm = crypto_aead_reqtfm(req); + struct aegis_block tag; + unsigned int authsize = crypto_aead_authsize(tfm); + unsigned int cryptlen = req->cryptlen - authsize; + + scatterwalk_map_and_copy(tag.bytes, req->src, + req->assoclen + cryptlen, authsize, 0); + + crypto_aegis128l_aesni_crypt(req, &tag, cryptlen, &OPS); + + return crypto_memneq(tag.bytes, zeros.bytes, authsize) ? -EBADMSG : 0; +} + +static int crypto_aegis128l_aesni_init_tfm(struct crypto_aead *aead) +{ + return 0; +} + +static void crypto_aegis128l_aesni_exit_tfm(struct crypto_aead *aead) +{ +} + +static int cryptd_aegis128l_aesni_setkey(struct crypto_aead *aead, + const u8 *key, unsigned int keylen) +{ + struct cryptd_aead **ctx = crypto_aead_ctx(aead); + struct cryptd_aead *cryptd_tfm = *ctx; + + return crypto_aead_setkey(&cryptd_tfm->base, key, keylen); +} + +static int cryptd_aegis128l_aesni_setauthsize(struct crypto_aead *aead, + unsigned int authsize) +{ + struct cryptd_aead **ctx = crypto_aead_ctx(aead); + struct cryptd_aead *cryptd_tfm = *ctx; + + return crypto_aead_setauthsize(&cryptd_tfm->base, authsize); +} + +static int cryptd_aegis128l_aesni_encrypt(struct aead_request *req) +{ + struct crypto_aead *aead = crypto_aead_reqtfm(req); + struct cryptd_aead **ctx = crypto_aead_ctx(aead); + struct cryptd_aead *cryptd_tfm = *ctx; + + aead = &cryptd_tfm->base; + if (irq_fpu_usable() && (!in_atomic() || + !cryptd_aead_queued(cryptd_tfm))) + aead = cryptd_aead_child(cryptd_tfm); + + aead_request_set_tfm(req, aead); + + return crypto_aead_encrypt(req); +} + +static int cryptd_aegis128l_aesni_decrypt(struct aead_request *req) +{ + struct crypto_aead *aead = crypto_aead_reqtfm(req); + struct cryptd_aead **ctx = crypto_aead_ctx(aead); + struct cryptd_aead *cryptd_tfm = *ctx; + + aead = &cryptd_tfm->base; + if (irq_fpu_usable() && (!in_atomic() || + !cryptd_aead_queued(cryptd_tfm))) + aead = cryptd_aead_child(cryptd_tfm); + + aead_request_set_tfm(req, aead); + + return crypto_aead_decrypt(req); +} + +static int cryptd_aegis128l_aesni_init_tfm(struct crypto_aead *aead) +{ + struct cryptd_aead *cryptd_tfm; + struct cryptd_aead **ctx = crypto_aead_ctx(aead); + + cryptd_tfm = cryptd_alloc_aead("__aegis128l-aesni", CRYPTO_ALG_INTERNAL, + CRYPTO_ALG_INTERNAL); + if (IS_ERR(cryptd_tfm)) + return PTR_ERR(cryptd_tfm); + + *ctx = cryptd_tfm; + crypto_aead_set_reqsize(aead, crypto_aead_reqsize(&cryptd_tfm->base)); + return 0; +} + +static void cryptd_aegis128l_aesni_exit_tfm(struct crypto_aead *aead) +{ + struct cryptd_aead **ctx = crypto_aead_ctx(aead); + + cryptd_free_aead(*ctx); +} + +static struct aead_alg crypto_aegis128l_aesni_alg[] = { + { + .setkey = crypto_aegis128l_aesni_setkey, + .setauthsize = crypto_aegis128l_aesni_setauthsize, + .encrypt = crypto_aegis128l_aesni_encrypt, + .decrypt = crypto_aegis128l_aesni_decrypt, + .init = crypto_aegis128l_aesni_init_tfm, + .exit = crypto_aegis128l_aesni_exit_tfm, + + .ivsize = AEGIS128L_NONCE_SIZE, + .maxauthsize = AEGIS128L_MAX_AUTH_SIZE, + .chunksize = AEGIS128L_BLOCK_SIZE, + + .base = { + .cra_flags = CRYPTO_ALG_INTERNAL, + .cra_blocksize = 1, + .cra_ctxsize = sizeof(struct aegis_ctx) + + __alignof__(struct aegis_ctx), + .cra_alignmask = 0, + + .cra_name = "__aegis128l", + .cra_driver_name = "__aegis128l-aesni", + + .cra_module = THIS_MODULE, + } + }, { + .setkey = cryptd_aegis128l_aesni_setkey, + .setauthsize = cryptd_aegis128l_aesni_setauthsize, + .encrypt = cryptd_aegis128l_aesni_encrypt, + .decrypt = cryptd_aegis128l_aesni_decrypt, + .init = cryptd_aegis128l_aesni_init_tfm, + .exit = cryptd_aegis128l_aesni_exit_tfm, + + .ivsize = AEGIS128L_NONCE_SIZE, + .maxauthsize = AEGIS128L_MAX_AUTH_SIZE, + .chunksize = AEGIS128L_BLOCK_SIZE, + + .base = { + .cra_flags = CRYPTO_ALG_ASYNC, + .cra_blocksize = 1, + .cra_ctxsize = sizeof(struct cryptd_aead *), + .cra_alignmask = 0, + + .cra_priority = 400, + + .cra_name = "aegis128l", + .cra_driver_name = "aegis128l-aesni", + + .cra_module = THIS_MODULE, + } + } +}; + +static const struct x86_cpu_id aesni_cpu_id[] = { + X86_FEATURE_MATCH(X86_FEATURE_AES), + X86_FEATURE_MATCH(X86_FEATURE_XMM2), + {} +}; +MODULE_DEVICE_TABLE(x86cpu, aesni_cpu_id); + +static int __init crypto_aegis128l_aesni_module_init(void) +{ + if (!x86_match_cpu(aesni_cpu_id)) + return -ENODEV; + + return crypto_register_aeads(crypto_aegis128l_aesni_alg, + ARRAY_SIZE(crypto_aegis128l_aesni_alg)); +} + +static void __exit crypto_aegis128l_aesni_module_exit(void) +{ + crypto_unregister_aeads(crypto_aegis128l_aesni_alg, + ARRAY_SIZE(crypto_aegis128l_aesni_alg)); +} + +module_init(crypto_aegis128l_aesni_module_init); +module_exit(crypto_aegis128l_aesni_module_exit); + +MODULE_LICENSE("GPL"); +MODULE_AUTHOR("Ondrej Mosnacek <omosnacek@gmail.com>"); +MODULE_DESCRIPTION("AEGIS-128L AEAD algorithm -- AESNI+SSE2 implementation"); +MODULE_ALIAS_CRYPTO("aegis128l"); +MODULE_ALIAS_CRYPTO("aegis128l-aesni"); diff --git a/arch/x86/crypto/aegis256-aesni-asm.S b/arch/x86/crypto/aegis256-aesni-asm.S new file mode 100644 index 000000000000..1d977d515bf9 --- /dev/null +++ b/arch/x86/crypto/aegis256-aesni-asm.S @@ -0,0 +1,702 @@ +/* + * AES-NI + SSE2 implementation of AEGIS-128L + * + * Copyright (c) 2017-2018 Ondrej Mosnacek <omosnacek@gmail.com> + * Copyright (C) 2017-2018 Red Hat, Inc. All rights reserved. + * + * 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. + */ + +#include <linux/linkage.h> +#include <asm/frame.h> + +#define STATE0 %xmm0 +#define STATE1 %xmm1 +#define STATE2 %xmm2 +#define STATE3 %xmm3 +#define STATE4 %xmm4 +#define STATE5 %xmm5 +#define MSG %xmm6 +#define T0 %xmm7 +#define T1 %xmm8 +#define T2 %xmm9 +#define T3 %xmm10 + +#define STATEP %rdi +#define LEN %rsi +#define SRC %rdx +#define DST %rcx + +.section .rodata.cst16.aegis256_const, "aM", @progbits, 32 +.align 16 +.Laegis256_const_0: + .byte 0x00, 0x01, 0x01, 0x02, 0x03, 0x05, 0x08, 0x0d + .byte 0x15, 0x22, 0x37, 0x59, 0x90, 0xe9, 0x79, 0x62 +.Laegis256_const_1: + .byte 0xdb, 0x3d, 0x18, 0x55, 0x6d, 0xc2, 0x2f, 0xf1 + .byte 0x20, 0x11, 0x31, 0x42, 0x73, 0xb5, 0x28, 0xdd + +.section .rodata.cst16.aegis256_counter, "aM", @progbits, 16 +.align 16 +.Laegis256_counter: + .byte 0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07 + .byte 0x08, 0x09, 0x0a, 0x0b, 0x0c, 0x0d, 0x0e, 0x0f + +.text + +/* + * __load_partial: internal ABI + * input: + * LEN - bytes + * SRC - src + * output: + * MSG - message block + * changed: + * T0 + * %r8 + * %r9 + */ +__load_partial: + xor %r9, %r9 + pxor MSG, MSG + + mov LEN, %r8 + and $0x1, %r8 + jz .Lld_partial_1 + + mov LEN, %r8 + and $0x1E, %r8 + add SRC, %r8 + mov (%r8), %r9b + +.Lld_partial_1: + mov LEN, %r8 + and $0x2, %r8 + jz .Lld_partial_2 + + mov LEN, %r8 + and $0x1C, %r8 + add SRC, %r8 + shl $0x10, %r9 + mov (%r8), %r9w + +.Lld_partial_2: + mov LEN, %r8 + and $0x4, %r8 + jz .Lld_partial_4 + + mov LEN, %r8 + and $0x18, %r8 + add SRC, %r8 + shl $32, %r9 + mov (%r8), %r8d + xor %r8, %r9 + +.Lld_partial_4: + movq %r9, MSG + + mov LEN, %r8 + and $0x8, %r8 + jz .Lld_partial_8 + + mov LEN, %r8 + and $0x10, %r8 + add SRC, %r8 + pslldq $8, MSG + movq (%r8), T0 + pxor T0, MSG + +.Lld_partial_8: + ret +ENDPROC(__load_partial) + +/* + * __store_partial: internal ABI + * input: + * LEN - bytes + * DST - dst + * output: + * T0 - message block + * changed: + * %r8 + * %r9 + * %r10 + */ +__store_partial: + mov LEN, %r8 + mov DST, %r9 + + movq T0, %r10 + + cmp $8, %r8 + jl .Lst_partial_8 + + mov %r10, (%r9) + psrldq $8, T0 + movq T0, %r10 + + sub $8, %r8 + add $8, %r9 + +.Lst_partial_8: + cmp $4, %r8 + jl .Lst_partial_4 + + mov %r10d, (%r9) + shr $32, %r10 + + sub $4, %r8 + add $4, %r9 + +.Lst_partial_4: + cmp $2, %r8 + jl .Lst_partial_2 + + mov %r10w, (%r9) + shr $0x10, %r10 + + sub $2, %r8 + add $2, %r9 + +.Lst_partial_2: + cmp $1, %r8 + jl .Lst_partial_1 + + mov %r10b, (%r9) + +.Lst_partial_1: + ret +ENDPROC(__store_partial) + +.macro update + movdqa STATE5, T0 + aesenc STATE0, STATE5 + aesenc STATE1, STATE0 + aesenc STATE2, STATE1 + aesenc STATE3, STATE2 + aesenc STATE4, STATE3 + aesenc T0, STATE4 +.endm + +.macro update0 m + update + pxor \m, STATE5 +.endm + +.macro update1 m + update + pxor \m, STATE4 +.endm + +.macro update2 m + update + pxor \m, STATE3 +.endm + +.macro update3 m + update + pxor \m, STATE2 +.endm + +.macro update4 m + update + pxor \m, STATE1 +.endm + +.macro update5 m + update + pxor \m, STATE0 +.endm + +.macro state_load + movdqu 0x00(STATEP), STATE0 + movdqu 0x10(STATEP), STATE1 + movdqu 0x20(STATEP), STATE2 + movdqu 0x30(STATEP), STATE3 + movdqu 0x40(STATEP), STATE4 + movdqu 0x50(STATEP), STATE5 +.endm + +.macro state_store s0 s1 s2 s3 s4 s5 + movdqu \s5, 0x00(STATEP) + movdqu \s0, 0x10(STATEP) + movdqu \s1, 0x20(STATEP) + movdqu \s2, 0x30(STATEP) + movdqu \s3, 0x40(STATEP) + movdqu \s4, 0x50(STATEP) +.endm + +.macro state_store0 + state_store STATE0 STATE1 STATE2 STATE3 STATE4 STATE5 +.endm + +.macro state_store1 + state_store STATE5 STATE0 STATE1 STATE2 STATE3 STATE4 +.endm + +.macro state_store2 + state_store STATE4 STATE5 STATE0 STATE1 STATE2 STATE3 +.endm + +.macro state_store3 + state_store STATE3 STATE4 STATE5 STATE0 STATE1 STATE2 +.endm + +.macro state_store4 + state_store STATE2 STATE3 STATE4 STATE5 STATE0 STATE1 +.endm + +.macro state_store5 + state_store STATE1 STATE2 STATE3 STATE4 STATE5 STATE0 +.endm + +/* + * void crypto_aegis256_aesni_init(void *state, const void *key, const void *iv); + */ +ENTRY(crypto_aegis256_aesni_init) + FRAME_BEGIN + + /* load key: */ + movdqa 0x00(%rsi), MSG + movdqa 0x10(%rsi), T1 + movdqa MSG, STATE4 + movdqa T1, STATE5 + + /* load IV: */ + movdqu 0x00(%rdx), T2 + movdqu 0x10(%rdx), T3 + pxor MSG, T2 + pxor T1, T3 + movdqa T2, STATE0 + movdqa T3, STATE1 + + /* load the constants: */ + movdqa .Laegis256_const_0, STATE3 + movdqa .Laegis256_const_1, STATE2 + pxor STATE3, STATE4 + pxor STATE2, STATE5 + + /* update 10 times with IV and KEY: */ + update0 MSG + update1 T1 + update2 T2 + update3 T3 + update4 MSG + update5 T1 + update0 T2 + update1 T3 + update2 MSG + update3 T1 + update4 T2 + update5 T3 + update0 MSG + update1 T1 + update2 T2 + update3 T3 + + state_store3 + + FRAME_END + ret +ENDPROC(crypto_aegis256_aesni_init) + +.macro ad_block a i + movdq\a (\i * 0x10)(SRC), MSG + update\i MSG + sub $0x10, LEN + cmp $0x10, LEN + jl .Lad_out_\i +.endm + +/* + * void crypto_aegis256_aesni_ad(void *state, unsigned int length, + * const void *data); + */ +ENTRY(crypto_aegis256_aesni_ad) + FRAME_BEGIN + + cmp $0x10, LEN + jb .Lad_out + + state_load + + mov SRC, %r8 + and $0xf, %r8 + jnz .Lad_u_loop + +.align 8 +.Lad_a_loop: + ad_block a 0 + ad_block a 1 + ad_block a 2 + ad_block a 3 + ad_block a 4 + ad_block a 5 + + add $0x60, SRC + jmp .Lad_a_loop + +.align 8 +.Lad_u_loop: + ad_block u 0 + ad_block u 1 + ad_block u 2 + ad_block u 3 + ad_block u 4 + ad_block u 5 + + add $0x60, SRC + jmp .Lad_u_loop + +.Lad_out_0: + state_store0 + FRAME_END + ret + +.Lad_out_1: + state_store1 + FRAME_END + ret + +.Lad_out_2: + state_store2 + FRAME_END + ret + +.Lad_out_3: + state_store3 + FRAME_END + ret + +.Lad_out_4: + state_store4 + FRAME_END + ret + +.Lad_out_5: + state_store5 + FRAME_END + ret + +.Lad_out: + FRAME_END + ret +ENDPROC(crypto_aegis256_aesni_ad) + +.macro crypt m s0 s1 s2 s3 s4 s5 + pxor \s1, \m + pxor \s4, \m + pxor \s5, \m + movdqa \s2, T3 + pand \s3, T3 + pxor T3, \m +.endm + +.macro crypt0 m + crypt \m STATE0 STATE1 STATE2 STATE3 STATE4 STATE5 +.endm + +.macro crypt1 m + crypt \m STATE5 STATE0 STATE1 STATE2 STATE3 STATE4 +.endm + +.macro crypt2 m + crypt \m STATE4 STATE5 STATE0 STATE1 STATE2 STATE3 +.endm + +.macro crypt3 m + crypt \m STATE3 STATE4 STATE5 STATE0 STATE1 STATE2 +.endm + +.macro crypt4 m + crypt \m STATE2 STATE3 STATE4 STATE5 STATE0 STATE1 +.endm + +.macro crypt5 m + crypt \m STATE1 STATE2 STATE3 STATE4 STATE5 STATE0 +.endm + +.macro encrypt_block a i + movdq\a (\i * 0x10)(SRC), MSG + movdqa MSG, T0 + crypt\i T0 + movdq\a T0, (\i * 0x10)(DST) + + update\i MSG + + sub $0x10, LEN + cmp $0x10, LEN + jl .Lenc_out_\i +.endm + +.macro decrypt_block a i + movdq\a (\i * 0x10)(SRC), MSG + crypt\i MSG + movdq\a MSG, (\i * 0x10)(DST) + + update\i MSG + + sub $0x10, LEN + cmp $0x10, LEN + jl .Ldec_out_\i +.endm + +/* + * void crypto_aegis256_aesni_enc(void *state, unsigned int length, + * const void *src, void *dst); + */ +ENTRY(crypto_aegis256_aesni_enc) + FRAME_BEGIN + + cmp $0x10, LEN + jb .Lenc_out + + state_load + + mov SRC, %r8 + or DST, %r8 + and $0xf, %r8 + jnz .Lenc_u_loop + +.align 8 +.Lenc_a_loop: + encrypt_block a 0 + encrypt_block a 1 + encrypt_block a 2 + encrypt_block a 3 + encrypt_block a 4 + encrypt_block a 5 + + add $0x60, SRC + add $0x60, DST + jmp .Lenc_a_loop + +.align 8 +.Lenc_u_loop: + encrypt_block u 0 + encrypt_block u 1 + encrypt_block u 2 + encrypt_block u 3 + encrypt_block u 4 + encrypt_block u 5 + + add $0x60, SRC + add $0x60, DST + jmp .Lenc_u_loop + +.Lenc_out_0: + state_store0 + FRAME_END + ret + +.Lenc_out_1: + state_store1 + FRAME_END + ret + +.Lenc_out_2: + state_store2 + FRAME_END + ret + +.Lenc_out_3: + state_store3 + FRAME_END + ret + +.Lenc_out_4: + state_store4 + FRAME_END + ret + +.Lenc_out_5: + state_store5 + FRAME_END + ret + +.Lenc_out: + FRAME_END + ret +ENDPROC(crypto_aegis256_aesni_enc) + +/* + * void crypto_aegis256_aesni_enc_tail(void *state, unsigned int length, + * const void *src, void *dst); + */ +ENTRY(crypto_aegis256_aesni_enc_tail) + FRAME_BEGIN + + state_load + + /* encrypt message: */ + call __load_partial + + movdqa MSG, T0 + crypt0 T0 + + call __store_partial + + update0 MSG + + state_store0 + + FRAME_END +ENDPROC(crypto_aegis256_aesni_enc_tail) + +/* + * void crypto_aegis256_aesni_dec(void *state, unsigned int length, + * const void *src, void *dst); + */ +ENTRY(crypto_aegis256_aesni_dec) + FRAME_BEGIN + + cmp $0x10, LEN + jb .Ldec_out + + state_load + + mov SRC, %r8 + or DST, %r8 + and $0xF, %r8 + jnz .Ldec_u_loop + +.align 8 +.Ldec_a_loop: + decrypt_block a 0 + decrypt_block a 1 + decrypt_block a 2 + decrypt_block a 3 + decrypt_block a 4 + decrypt_block a 5 + + add $0x60, SRC + add $0x60, DST + jmp .Ldec_a_loop + +.align 8 +.Ldec_u_loop: + decrypt_block u 0 + decrypt_block u 1 + decrypt_block u 2 + decrypt_block u 3 + decrypt_block u 4 + decrypt_block u 5 + + add $0x60, SRC + add $0x60, DST + jmp .Ldec_u_loop + +.Ldec_out_0: + state_store0 + FRAME_END + ret + +.Ldec_out_1: + state_store1 + FRAME_END + ret + +.Ldec_out_2: + state_store2 + FRAME_END + ret + +.Ldec_out_3: + state_store3 + FRAME_END + ret + +.Ldec_out_4: + state_store4 + FRAME_END + ret + +.Ldec_out_5: + state_store5 + FRAME_END + ret + +.Ldec_out: + FRAME_END + ret +ENDPROC(crypto_aegis256_aesni_dec) + +/* + * void crypto_aegis256_aesni_dec_tail(void *state, unsigned int length, + * const void *src, void *dst); + */ +ENTRY(crypto_aegis256_aesni_dec_tail) + FRAME_BEGIN + + state_load + + /* decrypt message: */ + call __load_partial + + crypt0 MSG + + movdqa MSG, T0 + call __store_partial + + /* mask with byte count: */ + movq LEN, T0 + punpcklbw T0, T0 + punpcklbw T0, T0 + punpcklbw T0, T0 + punpcklbw T0, T0 + movdqa .Laegis256_counter, T1 + pcmpgtb T1, T0 + pand T0, MSG + + update0 MSG + + state_store0 + + FRAME_END + ret +ENDPROC(crypto_aegis256_aesni_dec_tail) + +/* + * void crypto_aegis256_aesni_final(void *state, void *tag_xor, + * u64 assoclen, u64 cryptlen); + */ +ENTRY(crypto_aegis256_aesni_final) + FRAME_BEGIN + + state_load + + /* prepare length block: */ + movq %rdx, MSG + movq %rcx, T0 + pslldq $8, T0 + pxor T0, MSG + psllq $3, MSG /* multiply by 8 (to get bit count) */ + + pxor STATE3, MSG + + /* update state: */ + update0 MSG + update1 MSG + update2 MSG + update3 MSG + update4 MSG + update5 MSG + update0 MSG + + /* xor tag: */ + movdqu (%rsi), MSG + + pxor STATE0, MSG + pxor STATE1, MSG + pxor STATE2, MSG + pxor STATE3, MSG + pxor STATE4, MSG + pxor STATE5, MSG + + movdqu MSG, (%rsi) + + FRAME_END + ret +ENDPROC(crypto_aegis256_aesni_final) diff --git a/arch/x86/crypto/aegis256-aesni-glue.c b/arch/x86/crypto/aegis256-aesni-glue.c new file mode 100644 index 000000000000..2b5dd3af8f4d --- /dev/null +++ b/arch/x86/crypto/aegis256-aesni-glue.c @@ -0,0 +1,407 @@ +/* + * The AEGIS-256 Authenticated-Encryption Algorithm + * Glue for AES-NI + SSE2 implementation + * + * Copyright (c) 2017-2018 Ondrej Mosnacek <omosnacek@gmail.com> + * Copyright (C) 2017-2018 Red Hat, Inc. All rights reserved. + * + * This program is free software; you can redistribute it and/or modify it + * under the terms of the GNU General Public License as published by the Free + * Software Foundation; either version 2 of the License, or (at your option) + * any later version. + */ + +#include <crypto/cryptd.h> +#include <crypto/internal/aead.h> +#include <crypto/internal/skcipher.h> +#include <crypto/scatterwalk.h> +#include <linux/module.h> +#include <asm/fpu/api.h> +#include <asm/cpu_device_id.h> + +#define AEGIS256_BLOCK_ALIGN 16 +#define AEGIS256_BLOCK_SIZE 16 +#define AEGIS256_NONCE_SIZE 32 +#define AEGIS256_STATE_BLOCKS 6 +#define AEGIS256_KEY_SIZE 32 +#define AEGIS256_MIN_AUTH_SIZE 8 +#define AEGIS256_MAX_AUTH_SIZE 16 + +asmlinkage void crypto_aegis256_aesni_init(void *state, void *key, void *iv); + +asmlinkage void crypto_aegis256_aesni_ad( + void *state, unsigned int length, const void *data); + +asmlinkage void crypto_aegis256_aesni_enc( + void *state, unsigned int length, const void *src, void *dst); + +asmlinkage void crypto_aegis256_aesni_dec( + void *state, unsigned int length, const void *src, void *dst); + +asmlinkage void crypto_aegis256_aesni_enc_tail( + void *state, unsigned int length, const void *src, void *dst); + +asmlinkage void crypto_aegis256_aesni_dec_tail( + void *state, unsigned int length, const void *src, void *dst); + +asmlinkage void crypto_aegis256_aesni_final( + void *state, void *tag_xor, unsigned int cryptlen, + unsigned int assoclen); + +struct aegis_block { + u8 bytes[AEGIS256_BLOCK_SIZE] __aligned(AEGIS256_BLOCK_ALIGN); +}; + +struct aegis_state { + struct aegis_block blocks[AEGIS256_STATE_BLOCKS]; +}; + +struct aegis_ctx { + struct aegis_block key[AEGIS256_KEY_SIZE / AEGIS256_BLOCK_SIZE]; +}; + +struct aegis_crypt_ops { + int (*skcipher_walk_init)(struct skcipher_walk *walk, + struct aead_request *req, bool atomic); + + void (*crypt_blocks)(void *state, unsigned int length, const void *src, + void *dst); + void (*crypt_tail)(void *state, unsigned int length, const void *src, + void *dst); +}; + +static void crypto_aegis256_aesni_process_ad( + struct aegis_state *state, struct scatterlist *sg_src, + unsigned int assoclen) +{ + struct scatter_walk walk; + struct aegis_block buf; + unsigned int pos = 0; + + scatterwalk_start(&walk, sg_src); + while (assoclen != 0) { + unsigned int size = scatterwalk_clamp(&walk, assoclen); + unsigned int left = size; + void *mapped = scatterwalk_map(&walk); + const u8 *src = (const u8 *)mapped; + + if (pos + size >= AEGIS256_BLOCK_SIZE) { + if (pos > 0) { + unsigned int fill = AEGIS256_BLOCK_SIZE - pos; + memcpy(buf.bytes + pos, src, fill); + crypto_aegis256_aesni_ad(state, + AEGIS256_BLOCK_SIZE, + buf.bytes); + pos = 0; + left -= fill; + src += fill; + } + + crypto_aegis256_aesni_ad(state, left, src); + + src += left & ~(AEGIS256_BLOCK_SIZE - 1); + left &= AEGIS256_BLOCK_SIZE - 1; + } + + memcpy(buf.bytes + pos, src, left); + pos += left; + assoclen -= size; + + scatterwalk_unmap(mapped); + scatterwalk_advance(&walk, size); + scatterwalk_done(&walk, 0, assoclen); + } + + if (pos > 0) { + memset(buf.bytes + pos, 0, AEGIS256_BLOCK_SIZE - pos); + crypto_aegis256_aesni_ad(state, AEGIS256_BLOCK_SIZE, buf.bytes); + } +} + +static void crypto_aegis256_aesni_process_crypt( + struct aegis_state *state, struct aead_request *req, + const struct aegis_crypt_ops *ops) +{ + struct skcipher_walk walk; + u8 *src, *dst; + unsigned int chunksize, base; + + ops->skcipher_walk_init(&walk, req, false); + + while (walk.nbytes) { + src = walk.src.virt.addr; + dst = walk.dst.virt.addr; + chunksize = walk.nbytes; + + ops->crypt_blocks(state, chunksize, src, dst); + + base = chunksize & ~(AEGIS256_BLOCK_SIZE - 1); + src += base; + dst += base; + chunksize &= AEGIS256_BLOCK_SIZE - 1; + + if (chunksize > 0) + ops->crypt_tail(state, chunksize, src, dst); + + skcipher_walk_done(&walk, 0); + } +} + +static struct aegis_ctx *crypto_aegis256_aesni_ctx(struct crypto_aead *aead) +{ + u8 *ctx = crypto_aead_ctx(aead); + ctx = PTR_ALIGN(ctx, __alignof__(struct aegis_ctx)); + return (void *)ctx; +} + +static int crypto_aegis256_aesni_setkey(struct crypto_aead *aead, const u8 *key, + unsigned int keylen) +{ + struct aegis_ctx *ctx = crypto_aegis256_aesni_ctx(aead); + + if (keylen != AEGIS256_KEY_SIZE) { + crypto_aead_set_flags(aead, CRYPTO_TFM_RES_BAD_KEY_LEN); + return -EINVAL; + } + + memcpy(ctx->key, key, AEGIS256_KEY_SIZE); + + return 0; +} + +static int crypto_aegis256_aesni_setauthsize(struct crypto_aead *tfm, + unsigned int authsize) +{ + if (authsize > AEGIS256_MAX_AUTH_SIZE) + return -EINVAL; + if (authsize < AEGIS256_MIN_AUTH_SIZE) + return -EINVAL; + return 0; +} + +static void crypto_aegis256_aesni_crypt(struct aead_request *req, + struct aegis_block *tag_xor, + unsigned int cryptlen, + const struct aegis_crypt_ops *ops) +{ + struct crypto_aead *tfm = crypto_aead_reqtfm(req); + struct aegis_ctx *ctx = crypto_aegis256_aesni_ctx(tfm); + struct aegis_state state; + + kernel_fpu_begin(); + + crypto_aegis256_aesni_init(&state, ctx->key, req->iv); + crypto_aegis256_aesni_process_ad(&state, req->src, req->assoclen); + crypto_aegis256_aesni_process_crypt(&state, req, ops); + crypto_aegis256_aesni_final(&state, tag_xor, req->assoclen, cryptlen); + + kernel_fpu_end(); +} + +static int crypto_aegis256_aesni_encrypt(struct aead_request *req) +{ + static const struct aegis_crypt_ops OPS = { + .skcipher_walk_init = skcipher_walk_aead_encrypt, + .crypt_blocks = crypto_aegis256_aesni_enc, + .crypt_tail = crypto_aegis256_aesni_enc_tail, + }; + + struct crypto_aead *tfm = crypto_aead_reqtfm(req); + struct aegis_block tag = {}; + unsigned int authsize = crypto_aead_authsize(tfm); + unsigned int cryptlen = req->cryptlen; + + crypto_aegis256_aesni_crypt(req, &tag, cryptlen, &OPS); + + scatterwalk_map_and_copy(tag.bytes, req->dst, + req->assoclen + cryptlen, authsize, 1); + return 0; +} + +static int crypto_aegis256_aesni_decrypt(struct aead_request *req) +{ + static const struct aegis_block zeros = {}; + + static const struct aegis_crypt_ops OPS = { + .skcipher_walk_init = skcipher_walk_aead_decrypt, + .crypt_blocks = crypto_aegis256_aesni_dec, + .crypt_tail = crypto_aegis256_aesni_dec_tail, + }; + + struct crypto_aead *tfm = crypto_aead_reqtfm(req); + struct aegis_block tag; + unsigned int authsize = crypto_aead_authsize(tfm); + unsigned int cryptlen = req->cryptlen - authsize; + + scatterwalk_map_and_copy(tag.bytes, req->src, + req->assoclen + cryptlen, authsize, 0); + + crypto_aegis256_aesni_crypt(req, &tag, cryptlen, &OPS); + + return crypto_memneq(tag.bytes, zeros.bytes, authsize) ? -EBADMSG : 0; +} + +static int crypto_aegis256_aesni_init_tfm(struct crypto_aead *aead) +{ + return 0; +} + +static void crypto_aegis256_aesni_exit_tfm(struct crypto_aead *aead) +{ +} + +static int cryptd_aegis256_aesni_setkey(struct crypto_aead *aead, + const u8 *key, unsigned int keylen) +{ + struct cryptd_aead **ctx = crypto_aead_ctx(aead); + struct cryptd_aead *cryptd_tfm = *ctx; + + return crypto_aead_setkey(&cryptd_tfm->base, key, keylen); +} + +static int cryptd_aegis256_aesni_setauthsize(struct crypto_aead *aead, + unsigned int authsize) +{ + struct cryptd_aead **ctx = crypto_aead_ctx(aead); + struct cryptd_aead *cryptd_tfm = *ctx; + + return crypto_aead_setauthsize(&cryptd_tfm->base, authsize); +} + +static int cryptd_aegis256_aesni_encrypt(struct aead_request *req) +{ + struct crypto_aead *aead = crypto_aead_reqtfm(req); + struct cryptd_aead **ctx = crypto_aead_ctx(aead); + struct cryptd_aead *cryptd_tfm = *ctx; + + aead = &cryptd_tfm->base; + if (irq_fpu_usable() && (!in_atomic() || + !cryptd_aead_queued(cryptd_tfm))) + aead = cryptd_aead_child(cryptd_tfm); + + aead_request_set_tfm(req, aead); + + return crypto_aead_encrypt(req); +} + +static int cryptd_aegis256_aesni_decrypt(struct aead_request *req) +{ + struct crypto_aead *aead = crypto_aead_reqtfm(req); + struct cryptd_aead **ctx = crypto_aead_ctx(aead); + struct cryptd_aead *cryptd_tfm = *ctx; + + aead = &cryptd_tfm->base; + if (irq_fpu_usable() && (!in_atomic() || + !cryptd_aead_queued(cryptd_tfm))) + aead = cryptd_aead_child(cryptd_tfm); + + aead_request_set_tfm(req, aead); + + return crypto_aead_decrypt(req); +} + +static int cryptd_aegis256_aesni_init_tfm(struct crypto_aead *aead) +{ + struct cryptd_aead *cryptd_tfm; + struct cryptd_aead **ctx = crypto_aead_ctx(aead); + + cryptd_tfm = cryptd_alloc_aead("__aegis256-aesni", CRYPTO_ALG_INTERNAL, + CRYPTO_ALG_INTERNAL); + if (IS_ERR(cryptd_tfm)) + return PTR_ERR(cryptd_tfm); + + *ctx = cryptd_tfm; + crypto_aead_set_reqsize(aead, crypto_aead_reqsize(&cryptd_tfm->base)); + return 0; +} + +static void cryptd_aegis256_aesni_exit_tfm(struct crypto_aead *aead) +{ + struct cryptd_aead **ctx = crypto_aead_ctx(aead); + + cryptd_free_aead(*ctx); +} + +static struct aead_alg crypto_aegis256_aesni_alg[] = { + { + .setkey = crypto_aegis256_aesni_setkey, + .setauthsize = crypto_aegis256_aesni_setauthsize, + .encrypt = crypto_aegis256_aesni_encrypt, + .decrypt = crypto_aegis256_aesni_decrypt, + .init = crypto_aegis256_aesni_init_tfm, + .exit = crypto_aegis256_aesni_exit_tfm, + + .ivsize = AEGIS256_NONCE_SIZE, + .maxauthsize = AEGIS256_MAX_AUTH_SIZE, + .chunksize = AEGIS256_BLOCK_SIZE, + + .base = { + .cra_flags = CRYPTO_ALG_INTERNAL, + .cra_blocksize = 1, + .cra_ctxsize = sizeof(struct aegis_ctx) + + __alignof__(struct aegis_ctx), + .cra_alignmask = 0, + + .cra_name = "__aegis256", + .cra_driver_name = "__aegis256-aesni", + + .cra_module = THIS_MODULE, + } + }, { + .setkey = cryptd_aegis256_aesni_setkey, + .setauthsize = cryptd_aegis256_aesni_setauthsize, + .encrypt = cryptd_aegis256_aesni_encrypt, + .decrypt = cryptd_aegis256_aesni_decrypt, + .init = cryptd_aegis256_aesni_init_tfm, + .exit = cryptd_aegis256_aesni_exit_tfm, + + .ivsize = AEGIS256_NONCE_SIZE, + .maxauthsize = AEGIS256_MAX_AUTH_SIZE, + .chunksize = AEGIS256_BLOCK_SIZE, + + .base = { + .cra_flags = CRYPTO_ALG_ASYNC, + .cra_blocksize = 1, + .cra_ctxsize = sizeof(struct cryptd_aead *), + .cra_alignmask = 0, + + .cra_priority = 400, + + .cra_name = "aegis256", + .cra_driver_name = "aegis256-aesni", + + .cra_module = THIS_MODULE, + } + } +}; + +static const struct x86_cpu_id aesni_cpu_id[] = { + X86_FEATURE_MATCH(X86_FEATURE_AES), + X86_FEATURE_MATCH(X86_FEATURE_XMM2), + {} +}; +MODULE_DEVICE_TABLE(x86cpu, aesni_cpu_id); + +static int __init crypto_aegis256_aesni_module_init(void) +{ + if (!x86_match_cpu(aesni_cpu_id)) + return -ENODEV; + + return crypto_register_aeads(crypto_aegis256_aesni_alg, + ARRAY_SIZE(crypto_aegis256_aesni_alg)); +} + +static void __exit crypto_aegis256_aesni_module_exit(void) +{ + crypto_unregister_aeads(crypto_aegis256_aesni_alg, + ARRAY_SIZE(crypto_aegis256_aesni_alg)); +} + +module_init(crypto_aegis256_aesni_module_init); +module_exit(crypto_aegis256_aesni_module_exit); + +MODULE_LICENSE("GPL"); +MODULE_AUTHOR("Ondrej Mosnacek <omosnacek@gmail.com>"); +MODULE_DESCRIPTION("AEGIS-256 AEAD algorithm -- AESNI+SSE2 implementation"); +MODULE_ALIAS_CRYPTO("aegis256"); +MODULE_ALIAS_CRYPTO("aegis256-aesni"); diff --git a/arch/x86/crypto/ghash-clmulni-intel_glue.c b/arch/x86/crypto/ghash-clmulni-intel_glue.c index 0420bab19efb..2ddbe3a1868b 100644 --- a/arch/x86/crypto/ghash-clmulni-intel_glue.c +++ b/arch/x86/crypto/ghash-clmulni-intel_glue.c @@ -364,5 +364,5 @@ module_exit(ghash_pclmulqdqni_mod_exit); MODULE_LICENSE("GPL"); MODULE_DESCRIPTION("GHASH Message Digest Algorithm, " - "acclerated by PCLMULQDQ-NI"); + "accelerated by PCLMULQDQ-NI"); MODULE_ALIAS_CRYPTO("ghash"); diff --git a/arch/x86/crypto/morus1280-avx2-asm.S b/arch/x86/crypto/morus1280-avx2-asm.S new file mode 100644 index 000000000000..37d422e77931 --- /dev/null +++ b/arch/x86/crypto/morus1280-avx2-asm.S @@ -0,0 +1,621 @@ +/* + * AVX2 implementation of MORUS-1280 + * + * Copyright (c) 2017-2018 Ondrej Mosnacek <omosnacek@gmail.com> + * Copyright (C) 2017-2018 Red Hat, Inc. All rights reserved. + * + * 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. + */ + +#include <linux/linkage.h> +#include <asm/frame.h> + +#define SHUFFLE_MASK(i0, i1, i2, i3) \ + (i0 | (i1 << 2) | (i2 << 4) | (i3 << 6)) + +#define MASK1 SHUFFLE_MASK(3, 0, 1, 2) +#define MASK2 SHUFFLE_MASK(2, 3, 0, 1) +#define MASK3 SHUFFLE_MASK(1, 2, 3, 0) + +#define STATE0 %ymm0 +#define STATE0_LOW %xmm0 +#define STATE1 %ymm1 +#define STATE2 %ymm2 +#define STATE3 %ymm3 +#define STATE4 %ymm4 +#define KEY %ymm5 +#define MSG %ymm5 +#define MSG_LOW %xmm5 +#define T0 %ymm6 +#define T0_LOW %xmm6 +#define T1 %ymm7 + +.section .rodata.cst32.morus1280_const, "aM", @progbits, 32 +.align 32 +.Lmorus1280_const: + .byte 0x00, 0x01, 0x01, 0x02, 0x03, 0x05, 0x08, 0x0d + .byte 0x15, 0x22, 0x37, 0x59, 0x90, 0xe9, 0x79, 0x62 + .byte 0xdb, 0x3d, 0x18, 0x55, 0x6d, 0xc2, 0x2f, 0xf1 + .byte 0x20, 0x11, 0x31, 0x42, 0x73, 0xb5, 0x28, 0xdd + +.section .rodata.cst32.morus1280_counter, "aM", @progbits, 32 +.align 32 +.Lmorus1280_counter: + .byte 0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07 + .byte 0x08, 0x09, 0x0a, 0x0b, 0x0c, 0x0d, 0x0e, 0x0f + .byte 0x10, 0x11, 0x12, 0x13, 0x14, 0x15, 0x16, 0x17 + .byte 0x18, 0x19, 0x1a, 0x1b, 0x1c, 0x1d, 0x1e, 0x1f + +.text + +.macro morus1280_round s0, s1, s2, s3, s4, b, w + vpand \s1, \s2, T0 + vpxor T0, \s0, \s0 + vpxor \s3, \s0, \s0 + vpsllq $\b, \s0, T0 + vpsrlq $(64 - \b), \s0, \s0 + vpxor T0, \s0, \s0 + vpermq $\w, \s3, \s3 +.endm + +/* + * __morus1280_update: internal ABI + * input: + * STATE[0-4] - input state + * MSG - message block + * output: + * STATE[0-4] - output state + * changed: + * T0 + */ +__morus1280_update: + morus1280_round STATE0, STATE1, STATE2, STATE3, STATE4, 13, MASK1 + vpxor MSG, STATE1, STATE1 + morus1280_round STATE1, STATE2, STATE3, STATE4, STATE0, 46, MASK2 + vpxor MSG, STATE2, STATE2 + morus1280_round STATE2, STATE3, STATE4, STATE0, STATE1, 38, MASK3 + vpxor MSG, STATE3, STATE3 + morus1280_round STATE3, STATE4, STATE0, STATE1, STATE2, 7, MASK2 + vpxor MSG, STATE4, STATE4 + morus1280_round STATE4, STATE0, STATE1, STATE2, STATE3, 4, MASK1 + ret +ENDPROC(__morus1280_update) + +/* + * __morus1280_update_zero: internal ABI + * input: + * STATE[0-4] - input state + * output: + * STATE[0-4] - output state + * changed: + * T0 + */ +__morus1280_update_zero: + morus1280_round STATE0, STATE1, STATE2, STATE3, STATE4, 13, MASK1 + morus1280_round STATE1, STATE2, STATE3, STATE4, STATE0, 46, MASK2 + morus1280_round STATE2, STATE3, STATE4, STATE0, STATE1, 38, MASK3 + morus1280_round STATE3, STATE4, STATE0, STATE1, STATE2, 7, MASK2 + morus1280_round STATE4, STATE0, STATE1, STATE2, STATE3, 4, MASK1 + ret +ENDPROC(__morus1280_update_zero) + +/* + * __load_partial: internal ABI + * input: + * %rsi - src + * %rcx - bytes + * output: + * MSG - message block + * changed: + * %r8 + * %r9 + */ +__load_partial: + xor %r9, %r9 + vpxor MSG, MSG, MSG + + mov %rcx, %r8 + and $0x1, %r8 + jz .Lld_partial_1 + + mov %rcx, %r8 + and $0x1E, %r8 + add %rsi, %r8 + mov (%r8), %r9b + +.Lld_partial_1: + mov %rcx, %r8 + and $0x2, %r8 + jz .Lld_partial_2 + + mov %rcx, %r8 + and $0x1C, %r8 + add %rsi, %r8 + shl $16, %r9 + mov (%r8), %r9w + +.Lld_partial_2: + mov %rcx, %r8 + and $0x4, %r8 + jz .Lld_partial_4 + + mov %rcx, %r8 + and $0x18, %r8 + add %rsi, %r8 + shl $32, %r9 + mov (%r8), %r8d + xor %r8, %r9 + +.Lld_partial_4: + movq %r9, MSG_LOW + + mov %rcx, %r8 + and $0x8, %r8 + jz .Lld_partial_8 + + mov %rcx, %r8 + and $0x10, %r8 + add %rsi, %r8 + pshufd $MASK2, MSG_LOW, MSG_LOW + pinsrq $0, (%r8), MSG_LOW + +.Lld_partial_8: + mov %rcx, %r8 + and $0x10, %r8 + jz .Lld_partial_16 + + vpermq $MASK2, MSG, MSG + movdqu (%rsi), MSG_LOW + +.Lld_partial_16: + ret +ENDPROC(__load_partial) + +/* + * __store_partial: internal ABI + * input: + * %rdx - dst + * %rcx - bytes + * output: + * T0 - message block + * changed: + * %r8 + * %r9 + * %r10 + */ +__store_partial: + mov %rcx, %r8 + mov %rdx, %r9 + + cmp $16, %r8 + jl .Lst_partial_16 + + movdqu T0_LOW, (%r9) + vpermq $MASK2, T0, T0 + + sub $16, %r8 + add $16, %r9 + +.Lst_partial_16: + movq T0_LOW, %r10 + + cmp $8, %r8 + jl .Lst_partial_8 + + mov %r10, (%r9) + pextrq $1, T0_LOW, %r10 + + sub $8, %r8 + add $8, %r9 + +.Lst_partial_8: + cmp $4, %r8 + jl .Lst_partial_4 + + mov %r10d, (%r9) + shr $32, %r10 + + sub $4, %r8 + add $4, %r9 + +.Lst_partial_4: + cmp $2, %r8 + jl .Lst_partial_2 + + mov %r10w, (%r9) + shr $16, %r10 + + sub $2, %r8 + add $2, %r9 + +.Lst_partial_2: + cmp $1, %r8 + jl .Lst_partial_1 + + mov %r10b, (%r9) + +.Lst_partial_1: + ret +ENDPROC(__store_partial) + +/* + * void crypto_morus1280_avx2_init(void *state, const void *key, + * const void *iv); + */ +ENTRY(crypto_morus1280_avx2_init) + FRAME_BEGIN + + /* load IV: */ + vpxor STATE0, STATE0, STATE0 + movdqu (%rdx), STATE0_LOW + /* load key: */ + vmovdqu (%rsi), KEY + vmovdqa KEY, STATE1 + /* load all ones: */ + vpcmpeqd STATE2, STATE2, STATE2 + /* load all zeros: */ + vpxor STATE3, STATE3, STATE3 + /* load the constant: */ + vmovdqa .Lmorus1280_const, STATE4 + + /* update 16 times with zero: */ + call __morus1280_update_zero + call __morus1280_update_zero + call __morus1280_update_zero + call __morus1280_update_zero + call __morus1280_update_zero + call __morus1280_update_zero + call __morus1280_update_zero + call __morus1280_update_zero + call __morus1280_update_zero + call __morus1280_update_zero + call __morus1280_update_zero + call __morus1280_update_zero + call __morus1280_update_zero + call __morus1280_update_zero + call __morus1280_update_zero + call __morus1280_update_zero + + /* xor-in the key again after updates: */ + vpxor KEY, STATE1, STATE1 + + /* store the state: */ + vmovdqu STATE0, (0 * 32)(%rdi) + vmovdqu STATE1, (1 * 32)(%rdi) + vmovdqu STATE2, (2 * 32)(%rdi) + vmovdqu STATE3, (3 * 32)(%rdi) + vmovdqu STATE4, (4 * 32)(%rdi) + + FRAME_END + ret +ENDPROC(crypto_morus1280_avx2_init) + +/* + * void crypto_morus1280_avx2_ad(void *state, const void *data, + * unsigned int length); + */ +ENTRY(crypto_morus1280_avx2_ad) + FRAME_BEGIN + + cmp $32, %rdx + jb .Lad_out + + /* load the state: */ + vmovdqu (0 * 32)(%rdi), STATE0 + vmovdqu (1 * 32)(%rdi), STATE1 + vmovdqu (2 * 32)(%rdi), STATE2 + vmovdqu (3 * 32)(%rdi), STATE3 + vmovdqu (4 * 32)(%rdi), STATE4 + + mov %rsi, %r8 + and $0x1F, %r8 + jnz .Lad_u_loop + +.align 4 +.Lad_a_loop: + vmovdqa (%rsi), MSG + call __morus1280_update + sub $32, %rdx + add $32, %rsi + cmp $32, %rdx + jge .Lad_a_loop + + jmp .Lad_cont +.align 4 +.Lad_u_loop: + vmovdqu (%rsi), MSG + call __morus1280_update + sub $32, %rdx + add $32, %rsi + cmp $32, %rdx + jge .Lad_u_loop + +.Lad_cont: + /* store the state: */ + vmovdqu STATE0, (0 * 32)(%rdi) + vmovdqu STATE1, (1 * 32)(%rdi) + vmovdqu STATE2, (2 * 32)(%rdi) + vmovdqu STATE3, (3 * 32)(%rdi) + vmovdqu STATE4, (4 * 32)(%rdi) + +.Lad_out: + FRAME_END + ret +ENDPROC(crypto_morus1280_avx2_ad) + +/* + * void crypto_morus1280_avx2_enc(void *state, const void *src, void *dst, + * unsigned int length); + */ +ENTRY(crypto_morus1280_avx2_enc) + FRAME_BEGIN + + cmp $32, %rcx + jb .Lenc_out + + /* load the state: */ + vmovdqu (0 * 32)(%rdi), STATE0 + vmovdqu (1 * 32)(%rdi), STATE1 + vmovdqu (2 * 32)(%rdi), STATE2 + vmovdqu (3 * 32)(%rdi), STATE3 + vmovdqu (4 * 32)(%rdi), STATE4 + + mov %rsi, %r8 + or %rdx, %r8 + and $0x1F, %r8 + jnz .Lenc_u_loop + +.align 4 +.Lenc_a_loop: + vmovdqa (%rsi), MSG + vmovdqa MSG, T0 + vpxor STATE0, T0, T0 + vpermq $MASK3, STATE1, T1 + vpxor T1, T0, T0 + vpand STATE2, STATE3, T1 + vpxor T1, T0, T0 + vmovdqa T0, (%rdx) + + call __morus1280_update + sub $32, %rcx + add $32, %rsi + add $32, %rdx + cmp $32, %rcx + jge .Lenc_a_loop + + jmp .Lenc_cont +.align 4 +.Lenc_u_loop: + vmovdqu (%rsi), MSG + vmovdqa MSG, T0 + vpxor STATE0, T0, T0 + vpermq $MASK3, STATE1, T1 + vpxor T1, T0, T0 + vpand STATE2, STATE3, T1 + vpxor T1, T0, T0 + vmovdqu T0, (%rdx) + + call __morus1280_update + sub $32, %rcx + add $32, %rsi + add $32, %rdx + cmp $32, %rcx + jge .Lenc_u_loop + +.Lenc_cont: + /* store the state: */ + vmovdqu STATE0, (0 * 32)(%rdi) + vmovdqu STATE1, (1 * 32)(%rdi) + vmovdqu STATE2, (2 * 32)(%rdi) + vmovdqu STATE3, (3 * 32)(%rdi) + vmovdqu STATE4, (4 * 32)(%rdi) + +.Lenc_out: + FRAME_END + ret +ENDPROC(crypto_morus1280_avx2_enc) + +/* + * void crypto_morus1280_avx2_enc_tail(void *state, const void *src, void *dst, + * unsigned int length); + */ +ENTRY(crypto_morus1280_avx2_enc_tail) + FRAME_BEGIN + + /* load the state: */ + vmovdqu (0 * 32)(%rdi), STATE0 + vmovdqu (1 * 32)(%rdi), STATE1 + vmovdqu (2 * 32)(%rdi), STATE2 + vmovdqu (3 * 32)(%rdi), STATE3 + vmovdqu (4 * 32)(%rdi), STATE4 + + /* encrypt message: */ + call __load_partial + + vmovdqa MSG, T0 + vpxor STATE0, T0, T0 + vpermq $MASK3, STATE1, T1 + vpxor T1, T0, T0 + vpand STATE2, STATE3, T1 + vpxor T1, T0, T0 + + call __store_partial + + call __morus1280_update + + /* store the state: */ + vmovdqu STATE0, (0 * 32)(%rdi) + vmovdqu STATE1, (1 * 32)(%rdi) + vmovdqu STATE2, (2 * 32)(%rdi) + vmovdqu STATE3, (3 * 32)(%rdi) + vmovdqu STATE4, (4 * 32)(%rdi) + + FRAME_END +ENDPROC(crypto_morus1280_avx2_enc_tail) + +/* + * void crypto_morus1280_avx2_dec(void *state, const void *src, void *dst, + * unsigned int length); + */ +ENTRY(crypto_morus1280_avx2_dec) + FRAME_BEGIN + + cmp $32, %rcx + jb .Ldec_out + + /* load the state: */ + vmovdqu (0 * 32)(%rdi), STATE0 + vmovdqu (1 * 32)(%rdi), STATE1 + vmovdqu (2 * 32)(%rdi), STATE2 + vmovdqu (3 * 32)(%rdi), STATE3 + vmovdqu (4 * 32)(%rdi), STATE4 + + mov %rsi, %r8 + or %rdx, %r8 + and $0x1F, %r8 + jnz .Ldec_u_loop + +.align 4 +.Ldec_a_loop: + vmovdqa (%rsi), MSG + vpxor STATE0, MSG, MSG + vpermq $MASK3, STATE1, T0 + vpxor T0, MSG, MSG + vpand STATE2, STATE3, T0 + vpxor T0, MSG, MSG + vmovdqa MSG, (%rdx) + + call __morus1280_update + sub $32, %rcx + add $32, %rsi + add $32, %rdx + cmp $32, %rcx + jge .Ldec_a_loop + + jmp .Ldec_cont +.align 4 +.Ldec_u_loop: + vmovdqu (%rsi), MSG + vpxor STATE0, MSG, MSG + vpermq $MASK3, STATE1, T0 + vpxor T0, MSG, MSG + vpand STATE2, STATE3, T0 + vpxor T0, MSG, MSG + vmovdqu MSG, (%rdx) + + call __morus1280_update + sub $32, %rcx + add $32, %rsi + add $32, %rdx + cmp $32, %rcx + jge .Ldec_u_loop + +.Ldec_cont: + /* store the state: */ + vmovdqu STATE0, (0 * 32)(%rdi) + vmovdqu STATE1, (1 * 32)(%rdi) + vmovdqu STATE2, (2 * 32)(%rdi) + vmovdqu STATE3, (3 * 32)(%rdi) + vmovdqu STATE4, (4 * 32)(%rdi) + +.Ldec_out: + FRAME_END + ret +ENDPROC(crypto_morus1280_avx2_dec) + +/* + * void crypto_morus1280_avx2_dec_tail(void *state, const void *src, void *dst, + * unsigned int length); + */ +ENTRY(crypto_morus1280_avx2_dec_tail) + FRAME_BEGIN + + /* load the state: */ + vmovdqu (0 * 32)(%rdi), STATE0 + vmovdqu (1 * 32)(%rdi), STATE1 + vmovdqu (2 * 32)(%rdi), STATE2 + vmovdqu (3 * 32)(%rdi), STATE3 + vmovdqu (4 * 32)(%rdi), STATE4 + + /* decrypt message: */ + call __load_partial + + vpxor STATE0, MSG, MSG + vpermq $MASK3, STATE1, T0 + vpxor T0, MSG, MSG + vpand STATE2, STATE3, T0 + vpxor T0, MSG, MSG + vmovdqa MSG, T0 + + call __store_partial + + /* mask with byte count: */ + movq %rcx, T0_LOW + vpbroadcastb T0_LOW, T0 + vmovdqa .Lmorus1280_counter, T1 + vpcmpgtb T1, T0, T0 + vpand T0, MSG, MSG + + call __morus1280_update + + /* store the state: */ + vmovdqu STATE0, (0 * 32)(%rdi) + vmovdqu STATE1, (1 * 32)(%rdi) + vmovdqu STATE2, (2 * 32)(%rdi) + vmovdqu STATE3, (3 * 32)(%rdi) + vmovdqu STATE4, (4 * 32)(%rdi) + + FRAME_END + ret +ENDPROC(crypto_morus1280_avx2_dec_tail) + +/* + * void crypto_morus1280_avx2_final(void *state, void *tag_xor, + * u64 assoclen, u64 cryptlen); + */ +ENTRY(crypto_morus1280_avx2_final) + FRAME_BEGIN + + /* load the state: */ + vmovdqu (0 * 32)(%rdi), STATE0 + vmovdqu (1 * 32)(%rdi), STATE1 + vmovdqu (2 * 32)(%rdi), STATE2 + vmovdqu (3 * 32)(%rdi), STATE3 + vmovdqu (4 * 32)(%rdi), STATE4 + + /* xor state[0] into state[4]: */ + vpxor STATE0, STATE4, STATE4 + + /* prepare length block: */ + vpxor MSG, MSG, MSG + vpinsrq $0, %rdx, MSG_LOW, MSG_LOW + vpinsrq $1, %rcx, MSG_LOW, MSG_LOW + vpsllq $3, MSG, MSG /* multiply by 8 (to get bit count) */ + + /* update state: */ + call __morus1280_update + call __morus1280_update + call __morus1280_update + call __morus1280_update + call __morus1280_update + call __morus1280_update + call __morus1280_update + call __morus1280_update + call __morus1280_update + call __morus1280_update + + /* xor tag: */ + vmovdqu (%rsi), MSG + + vpxor STATE0, MSG, MSG + vpermq $MASK3, STATE1, T0 + vpxor T0, MSG, MSG + vpand STATE2, STATE3, T0 + vpxor T0, MSG, MSG + vmovdqu MSG, (%rsi) + + FRAME_END + ret +ENDPROC(crypto_morus1280_avx2_final) diff --git a/arch/x86/crypto/morus1280-avx2-glue.c b/arch/x86/crypto/morus1280-avx2-glue.c new file mode 100644 index 000000000000..f111f36d26dc --- /dev/null +++ b/arch/x86/crypto/morus1280-avx2-glue.c @@ -0,0 +1,68 @@ +/* + * The MORUS-1280 Authenticated-Encryption Algorithm + * Glue for AVX2 implementation + * + * Copyright (c) 2016-2018 Ondrej Mosnacek <omosnacek@gmail.com> + * Copyright (C) 2017-2018 Red Hat, Inc. All rights reserved. + * + * This program is free software; you can redistribute it and/or modify it + * under the terms of the GNU General Public License as published by the Free + * Software Foundation; either version 2 of the License, or (at your option) + * any later version. + */ + +#include <crypto/internal/aead.h> +#include <crypto/morus1280_glue.h> +#include <linux/module.h> +#include <asm/fpu/api.h> +#include <asm/cpu_device_id.h> + +asmlinkage void crypto_morus1280_avx2_init(void *state, const void *key, + const void *iv); +asmlinkage void crypto_morus1280_avx2_ad(void *state, const void *data, + unsigned int length); + +asmlinkage void crypto_morus1280_avx2_enc(void *state, const void *src, + void *dst, unsigned int length); +asmlinkage void crypto_morus1280_avx2_dec(void *state, const void *src, + void *dst, unsigned int length); + +asmlinkage void crypto_morus1280_avx2_enc_tail(void *state, const void *src, + void *dst, unsigned int length); +asmlinkage void crypto_morus1280_avx2_dec_tail(void *state, const void *src, + void *dst, unsigned int length); + +asmlinkage void crypto_morus1280_avx2_final(void *state, void *tag_xor, + u64 assoclen, u64 cryptlen); + +MORUS1280_DECLARE_ALGS(avx2, "morus1280-avx2", 400); + +static const struct x86_cpu_id avx2_cpu_id[] = { + X86_FEATURE_MATCH(X86_FEATURE_AVX2), + {} +}; +MODULE_DEVICE_TABLE(x86cpu, avx2_cpu_id); + +static int __init crypto_morus1280_avx2_module_init(void) +{ + if (!x86_match_cpu(avx2_cpu_id)) + return -ENODEV; + + return crypto_register_aeads(crypto_morus1280_avx2_algs, + ARRAY_SIZE(crypto_morus1280_avx2_algs)); +} + +static void __exit crypto_morus1280_avx2_module_exit(void) +{ + crypto_unregister_aeads(crypto_morus1280_avx2_algs, + ARRAY_SIZE(crypto_morus1280_avx2_algs)); +} + +module_init(crypto_morus1280_avx2_module_init); +module_exit(crypto_morus1280_avx2_module_exit); + +MODULE_LICENSE("GPL"); +MODULE_AUTHOR("Ondrej Mosnacek <omosnacek@gmail.com>"); +MODULE_DESCRIPTION("MORUS-1280 AEAD algorithm -- AVX2 implementation"); +MODULE_ALIAS_CRYPTO("morus1280"); +MODULE_ALIAS_CRYPTO("morus1280-avx2"); diff --git a/arch/x86/crypto/morus1280-sse2-asm.S b/arch/x86/crypto/morus1280-sse2-asm.S new file mode 100644 index 000000000000..1fe637c7be9d --- /dev/null +++ b/arch/x86/crypto/morus1280-sse2-asm.S @@ -0,0 +1,895 @@ +/* + * SSE2 implementation of MORUS-1280 + * + * Copyright (c) 2017-2018 Ondrej Mosnacek <omosnacek@gmail.com> + * Copyright (C) 2017-2018 Red Hat, Inc. All rights reserved. + * + * 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. + */ + +#include <linux/linkage.h> +#include <asm/frame.h> + +#define SHUFFLE_MASK(i0, i1, i2, i3) \ + (i0 | (i1 << 2) | (i2 << 4) | (i3 << 6)) + +#define MASK2 SHUFFLE_MASK(2, 3, 0, 1) + +#define STATE0_LO %xmm0 +#define STATE0_HI %xmm1 +#define STATE1_LO %xmm2 +#define STATE1_HI %xmm3 +#define STATE2_LO %xmm4 +#define STATE2_HI %xmm5 +#define STATE3_LO %xmm6 +#define STATE3_HI %xmm7 +#define STATE4_LO %xmm8 +#define STATE4_HI %xmm9 +#define KEY_LO %xmm10 +#define KEY_HI %xmm11 +#define MSG_LO %xmm10 +#define MSG_HI %xmm11 +#define T0_LO %xmm12 +#define T0_HI %xmm13 +#define T1_LO %xmm14 +#define T1_HI %xmm15 + +.section .rodata.cst16.morus640_const, "aM", @progbits, 16 +.align 16 +.Lmorus640_const_0: + .byte 0x00, 0x01, 0x01, 0x02, 0x03, 0x05, 0x08, 0x0d + .byte 0x15, 0x22, 0x37, 0x59, 0x90, 0xe9, 0x79, 0x62 +.Lmorus640_const_1: + .byte 0xdb, 0x3d, 0x18, 0x55, 0x6d, 0xc2, 0x2f, 0xf1 + .byte 0x20, 0x11, 0x31, 0x42, 0x73, 0xb5, 0x28, 0xdd + +.section .rodata.cst16.morus640_counter, "aM", @progbits, 16 +.align 16 +.Lmorus640_counter_0: + .byte 0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07 + .byte 0x08, 0x09, 0x0a, 0x0b, 0x0c, 0x0d, 0x0e, 0x0f +.Lmorus640_counter_1: + .byte 0x10, 0x11, 0x12, 0x13, 0x14, 0x15, 0x16, 0x17 + .byte 0x18, 0x19, 0x1a, 0x1b, 0x1c, 0x1d, 0x1e, 0x1f + +.text + +.macro rol1 hi, lo + /* + * HI_1 | HI_0 || LO_1 | LO_0 + * ==> + * HI_0 | HI_1 || LO_1 | LO_0 + * ==> + * HI_0 | LO_1 || LO_0 | HI_1 + */ + pshufd $MASK2, \hi, \hi + movdqa \hi, T0_LO + punpcklqdq \lo, T0_LO + punpckhqdq \hi, \lo + movdqa \lo, \hi + movdqa T0_LO, \lo +.endm + +.macro rol2 hi, lo + movdqa \lo, T0_LO + movdqa \hi, \lo + movdqa T0_LO, \hi +.endm + +.macro rol3 hi, lo + /* + * HI_1 | HI_0 || LO_1 | LO_0 + * ==> + * HI_0 | HI_1 || LO_1 | LO_0 + * ==> + * LO_0 | HI_1 || HI_0 | LO_1 + */ + pshufd $MASK2, \hi, \hi + movdqa \lo, T0_LO + punpckhqdq \hi, T0_LO + punpcklqdq \lo, \hi + movdqa T0_LO, \lo +.endm + +.macro morus1280_round s0_l, s0_h, s1_l, s1_h, s2_l, s2_h, s3_l, s3_h, s4_l, s4_h, b, w + movdqa \s1_l, T0_LO + pand \s2_l, T0_LO + pxor T0_LO, \s0_l + + movdqa \s1_h, T0_LO + pand \s2_h, T0_LO + pxor T0_LO, \s0_h + + pxor \s3_l, \s0_l + pxor \s3_h, \s0_h + + movdqa \s0_l, T0_LO + psllq $\b, T0_LO + psrlq $(64 - \b), \s0_l + pxor T0_LO, \s0_l + + movdqa \s0_h, T0_LO + psllq $\b, T0_LO + psrlq $(64 - \b), \s0_h + pxor T0_LO, \s0_h + + \w \s3_h, \s3_l +.endm + +/* + * __morus1280_update: internal ABI + * input: + * STATE[0-4] - input state + * MSG - message block + * output: + * STATE[0-4] - output state + * changed: + * T0 + */ +__morus1280_update: + morus1280_round \ + STATE0_LO, STATE0_HI, \ + STATE1_LO, STATE1_HI, \ + STATE2_LO, STATE2_HI, \ + STATE3_LO, STATE3_HI, \ + STATE4_LO, STATE4_HI, \ + 13, rol1 + pxor MSG_LO, STATE1_LO + pxor MSG_HI, STATE1_HI + morus1280_round \ + STATE1_LO, STATE1_HI, \ + STATE2_LO, STATE2_HI, \ + STATE3_LO, STATE3_HI, \ + STATE4_LO, STATE4_HI, \ + STATE0_LO, STATE0_HI, \ + 46, rol2 + pxor MSG_LO, STATE2_LO + pxor MSG_HI, STATE2_HI + morus1280_round \ + STATE2_LO, STATE2_HI, \ + STATE3_LO, STATE3_HI, \ + STATE4_LO, STATE4_HI, \ + STATE0_LO, STATE0_HI, \ + STATE1_LO, STATE1_HI, \ + 38, rol3 + pxor MSG_LO, STATE3_LO + pxor MSG_HI, STATE3_HI + morus1280_round \ + STATE3_LO, STATE3_HI, \ + STATE4_LO, STATE4_HI, \ + STATE0_LO, STATE0_HI, \ + STATE1_LO, STATE1_HI, \ + STATE2_LO, STATE2_HI, \ + 7, rol2 + pxor MSG_LO, STATE4_LO + pxor MSG_HI, STATE4_HI + morus1280_round \ + STATE4_LO, STATE4_HI, \ + STATE0_LO, STATE0_HI, \ + STATE1_LO, STATE1_HI, \ + STATE2_LO, STATE2_HI, \ + STATE3_LO, STATE3_HI, \ + 4, rol1 + ret +ENDPROC(__morus1280_update) + +/* + * __morus1280_update_zero: internal ABI + * input: + * STATE[0-4] - input state + * output: + * STATE[0-4] - output state + * changed: + * T0 + */ +__morus1280_update_zero: + morus1280_round \ + STATE0_LO, STATE0_HI, \ + STATE1_LO, STATE1_HI, \ + STATE2_LO, STATE2_HI, \ + STATE3_LO, STATE3_HI, \ + STATE4_LO, STATE4_HI, \ + 13, rol1 + morus1280_round \ + STATE1_LO, STATE1_HI, \ + STATE2_LO, STATE2_HI, \ + STATE3_LO, STATE3_HI, \ + STATE4_LO, STATE4_HI, \ + STATE0_LO, STATE0_HI, \ + 46, rol2 + morus1280_round \ + STATE2_LO, STATE2_HI, \ + STATE3_LO, STATE3_HI, \ + STATE4_LO, STATE4_HI, \ + STATE0_LO, STATE0_HI, \ + STATE1_LO, STATE1_HI, \ + 38, rol3 + morus1280_round \ + STATE3_LO, STATE3_HI, \ + STATE4_LO, STATE4_HI, \ + STATE0_LO, STATE0_HI, \ + STATE1_LO, STATE1_HI, \ + STATE2_LO, STATE2_HI, \ + 7, rol2 + morus1280_round \ + STATE4_LO, STATE4_HI, \ + STATE0_LO, STATE0_HI, \ + STATE1_LO, STATE1_HI, \ + STATE2_LO, STATE2_HI, \ + STATE3_LO, STATE3_HI, \ + 4, rol1 + ret +ENDPROC(__morus1280_update_zero) + +/* + * __load_partial: internal ABI + * input: + * %rsi - src + * %rcx - bytes + * output: + * MSG - message block + * changed: + * %r8 + * %r9 + */ +__load_partial: + xor %r9, %r9 + pxor MSG_LO, MSG_LO + pxor MSG_HI, MSG_HI + + mov %rcx, %r8 + and $0x1, %r8 + jz .Lld_partial_1 + + mov %rcx, %r8 + and $0x1E, %r8 + add %rsi, %r8 + mov (%r8), %r9b + +.Lld_partial_1: + mov %rcx, %r8 + and $0x2, %r8 + jz .Lld_partial_2 + + mov %rcx, %r8 + and $0x1C, %r8 + add %rsi, %r8 + shl $16, %r9 + mov (%r8), %r9w + +.Lld_partial_2: + mov %rcx, %r8 + and $0x4, %r8 + jz .Lld_partial_4 + + mov %rcx, %r8 + and $0x18, %r8 + add %rsi, %r8 + shl $32, %r9 + mov (%r8), %r8d + xor %r8, %r9 + +.Lld_partial_4: + movq %r9, MSG_LO + + mov %rcx, %r8 + and $0x8, %r8 + jz .Lld_partial_8 + + mov %rcx, %r8 + and $0x10, %r8 + add %rsi, %r8 + pslldq $8, MSG_LO + movq (%r8), T0_LO + pxor T0_LO, MSG_LO + +.Lld_partial_8: + mov %rcx, %r8 + and $0x10, %r8 + jz .Lld_partial_16 + + movdqa MSG_LO, MSG_HI + movdqu (%rsi), MSG_LO + +.Lld_partial_16: + ret +ENDPROC(__load_partial) + +/* + * __store_partial: internal ABI + * input: + * %rdx - dst + * %rcx - bytes + * output: + * T0 - message block + * changed: + * %r8 + * %r9 + * %r10 + */ +__store_partial: + mov %rcx, %r8 + mov %rdx, %r9 + + cmp $16, %r8 + jl .Lst_partial_16 + + movdqu T0_LO, (%r9) + movdqa T0_HI, T0_LO + + sub $16, %r8 + add $16, %r9 + +.Lst_partial_16: + movq T0_LO, %r10 + + cmp $8, %r8 + jl .Lst_partial_8 + + mov %r10, (%r9) + psrldq $8, T0_LO + movq T0_LO, %r10 + + sub $8, %r8 + add $8, %r9 + +.Lst_partial_8: + cmp $4, %r8 + jl .Lst_partial_4 + + mov %r10d, (%r9) + shr $32, %r10 + + sub $4, %r8 + add $4, %r9 + +.Lst_partial_4: + cmp $2, %r8 + jl .Lst_partial_2 + + mov %r10w, (%r9) + shr $16, %r10 + + sub $2, %r8 + add $2, %r9 + +.Lst_partial_2: + cmp $1, %r8 + jl .Lst_partial_1 + + mov %r10b, (%r9) + +.Lst_partial_1: + ret +ENDPROC(__store_partial) + +/* + * void crypto_morus1280_sse2_init(void *state, const void *key, + * const void *iv); + */ +ENTRY(crypto_morus1280_sse2_init) + FRAME_BEGIN + + /* load IV: */ + pxor STATE0_HI, STATE0_HI + movdqu (%rdx), STATE0_LO + /* load key: */ + movdqu 0(%rsi), KEY_LO + movdqu 16(%rsi), KEY_HI + movdqa KEY_LO, STATE1_LO + movdqa KEY_HI, STATE1_HI + /* load all ones: */ + pcmpeqd STATE2_LO, STATE2_LO + pcmpeqd STATE2_HI, STATE2_HI + /* load all zeros: */ + pxor STATE3_LO, STATE3_LO + pxor STATE3_HI, STATE3_HI + /* load the constant: */ + movdqa .Lmorus640_const_0, STATE4_LO + movdqa .Lmorus640_const_1, STATE4_HI + + /* update 16 times with zero: */ + call __morus1280_update_zero + call __morus1280_update_zero + call __morus1280_update_zero + call __morus1280_update_zero + call __morus1280_update_zero + call __morus1280_update_zero + call __morus1280_update_zero + call __morus1280_update_zero + call __morus1280_update_zero + call __morus1280_update_zero + call __morus1280_update_zero + call __morus1280_update_zero + call __morus1280_update_zero + call __morus1280_update_zero + call __morus1280_update_zero + call __morus1280_update_zero + + /* xor-in the key again after updates: */ + pxor KEY_LO, STATE1_LO + pxor KEY_HI, STATE1_HI + + /* store the state: */ + movdqu STATE0_LO, (0 * 16)(%rdi) + movdqu STATE0_HI, (1 * 16)(%rdi) + movdqu STATE1_LO, (2 * 16)(%rdi) + movdqu STATE1_HI, (3 * 16)(%rdi) + movdqu STATE2_LO, (4 * 16)(%rdi) + movdqu STATE2_HI, (5 * 16)(%rdi) + movdqu STATE3_LO, (6 * 16)(%rdi) + movdqu STATE3_HI, (7 * 16)(%rdi) + movdqu STATE4_LO, (8 * 16)(%rdi) + movdqu STATE4_HI, (9 * 16)(%rdi) + + FRAME_END + ret +ENDPROC(crypto_morus1280_sse2_init) + +/* + * void crypto_morus1280_sse2_ad(void *state, const void *data, + * unsigned int length); + */ +ENTRY(crypto_morus1280_sse2_ad) + FRAME_BEGIN + + cmp $32, %rdx + jb .Lad_out + + /* load the state: */ + movdqu (0 * 16)(%rdi), STATE0_LO + movdqu (1 * 16)(%rdi), STATE0_HI + movdqu (2 * 16)(%rdi), STATE1_LO + movdqu (3 * 16)(%rdi), STATE1_HI + movdqu (4 * 16)(%rdi), STATE2_LO + movdqu (5 * 16)(%rdi), STATE2_HI + movdqu (6 * 16)(%rdi), STATE3_LO + movdqu (7 * 16)(%rdi), STATE3_HI + movdqu (8 * 16)(%rdi), STATE4_LO + movdqu (9 * 16)(%rdi), STATE4_HI + + mov %rsi, %r8 + and $0xF, %r8 + jnz .Lad_u_loop + +.align 4 +.Lad_a_loop: + movdqa 0(%rsi), MSG_LO + movdqa 16(%rsi), MSG_HI + call __morus1280_update + sub $32, %rdx + add $32, %rsi + cmp $32, %rdx + jge .Lad_a_loop + + jmp .Lad_cont +.align 4 +.Lad_u_loop: + movdqu 0(%rsi), MSG_LO + movdqu 16(%rsi), MSG_HI + call __morus1280_update + sub $32, %rdx + add $32, %rsi + cmp $32, %rdx + jge .Lad_u_loop + +.Lad_cont: + /* store the state: */ + movdqu STATE0_LO, (0 * 16)(%rdi) + movdqu STATE0_HI, (1 * 16)(%rdi) + movdqu STATE1_LO, (2 * 16)(%rdi) + movdqu STATE1_HI, (3 * 16)(%rdi) + movdqu STATE2_LO, (4 * 16)(%rdi) + movdqu STATE2_HI, (5 * 16)(%rdi) + movdqu STATE3_LO, (6 * 16)(%rdi) + movdqu STATE3_HI, (7 * 16)(%rdi) + movdqu STATE4_LO, (8 * 16)(%rdi) + movdqu STATE4_HI, (9 * 16)(%rdi) + +.Lad_out: + FRAME_END + ret +ENDPROC(crypto_morus1280_sse2_ad) + +/* + * void crypto_morus1280_sse2_enc(void *state, const void *src, void *dst, + * unsigned int length); + */ +ENTRY(crypto_morus1280_sse2_enc) + FRAME_BEGIN + + cmp $32, %rcx + jb .Lenc_out + + /* load the state: */ + movdqu (0 * 16)(%rdi), STATE0_LO + movdqu (1 * 16)(%rdi), STATE0_HI + movdqu (2 * 16)(%rdi), STATE1_LO + movdqu (3 * 16)(%rdi), STATE1_HI + movdqu (4 * 16)(%rdi), STATE2_LO + movdqu (5 * 16)(%rdi), STATE2_HI + movdqu (6 * 16)(%rdi), STATE3_LO + movdqu (7 * 16)(%rdi), STATE3_HI + movdqu (8 * 16)(%rdi), STATE4_LO + movdqu (9 * 16)(%rdi), STATE4_HI + + mov %rsi, %r8 + or %rdx, %r8 + and $0xF, %r8 + jnz .Lenc_u_loop + +.align 4 +.Lenc_a_loop: + movdqa 0(%rsi), MSG_LO + movdqa 16(%rsi), MSG_HI + movdqa STATE1_LO, T1_LO + movdqa STATE1_HI, T1_HI + rol3 T1_HI, T1_LO + movdqa MSG_LO, T0_LO + movdqa MSG_HI, T0_HI + pxor T1_LO, T0_LO + pxor T1_HI, T0_HI + pxor STATE0_LO, T0_LO + pxor STATE0_HI, T0_HI + movdqa STATE2_LO, T1_LO + movdqa STATE2_HI, T1_HI + pand STATE3_LO, T1_LO + pand STATE3_HI, T1_HI + pxor T1_LO, T0_LO + pxor T1_HI, T0_HI + movdqa T0_LO, 0(%rdx) + movdqa T0_HI, 16(%rdx) + + call __morus1280_update + sub $32, %rcx + add $32, %rsi + add $32, %rdx + cmp $32, %rcx + jge .Lenc_a_loop + + jmp .Lenc_cont +.align 4 +.Lenc_u_loop: + movdqu 0(%rsi), MSG_LO + movdqu 16(%rsi), MSG_HI + movdqa STATE1_LO, T1_LO + movdqa STATE1_HI, T1_HI + rol3 T1_HI, T1_LO + movdqa MSG_LO, T0_LO + movdqa MSG_HI, T0_HI + pxor T1_LO, T0_LO + pxor T1_HI, T0_HI + pxor STATE0_LO, T0_LO + pxor STATE0_HI, T0_HI + movdqa STATE2_LO, T1_LO + movdqa STATE2_HI, T1_HI + pand STATE3_LO, T1_LO + pand STATE3_HI, T1_HI + pxor T1_LO, T0_LO + pxor T1_HI, T0_HI + movdqu T0_LO, 0(%rdx) + movdqu T0_HI, 16(%rdx) + + call __morus1280_update + sub $32, %rcx + add $32, %rsi + add $32, %rdx + cmp $32, %rcx + jge .Lenc_u_loop + +.Lenc_cont: + /* store the state: */ + movdqu STATE0_LO, (0 * 16)(%rdi) + movdqu STATE0_HI, (1 * 16)(%rdi) + movdqu STATE1_LO, (2 * 16)(%rdi) + movdqu STATE1_HI, (3 * 16)(%rdi) + movdqu STATE2_LO, (4 * 16)(%rdi) + movdqu STATE2_HI, (5 * 16)(%rdi) + movdqu STATE3_LO, (6 * 16)(%rdi) + movdqu STATE3_HI, (7 * 16)(%rdi) + movdqu STATE4_LO, (8 * 16)(%rdi) + movdqu STATE4_HI, (9 * 16)(%rdi) + +.Lenc_out: + FRAME_END + ret +ENDPROC(crypto_morus1280_sse2_enc) + +/* + * void crypto_morus1280_sse2_enc_tail(void *state, const void *src, void *dst, + * unsigned int length); + */ +ENTRY(crypto_morus1280_sse2_enc_tail) + FRAME_BEGIN + + /* load the state: */ + movdqu (0 * 16)(%rdi), STATE0_LO + movdqu (1 * 16)(%rdi), STATE0_HI + movdqu (2 * 16)(%rdi), STATE1_LO + movdqu (3 * 16)(%rdi), STATE1_HI + movdqu (4 * 16)(%rdi), STATE2_LO + movdqu (5 * 16)(%rdi), STATE2_HI + movdqu (6 * 16)(%rdi), STATE3_LO + movdqu (7 * 16)(%rdi), STATE3_HI + movdqu (8 * 16)(%rdi), STATE4_LO + movdqu (9 * 16)(%rdi), STATE4_HI + + /* encrypt message: */ + call __load_partial + + movdqa STATE1_LO, T1_LO + movdqa STATE1_HI, T1_HI + rol3 T1_HI, T1_LO + movdqa MSG_LO, T0_LO + movdqa MSG_HI, T0_HI + pxor T1_LO, T0_LO + pxor T1_HI, T0_HI + pxor STATE0_LO, T0_LO + pxor STATE0_HI, T0_HI + movdqa STATE2_LO, T1_LO + movdqa STATE2_HI, T1_HI + pand STATE3_LO, T1_LO + pand STATE3_HI, T1_HI + pxor T1_LO, T0_LO + pxor T1_HI, T0_HI + + call __store_partial + + call __morus1280_update + + /* store the state: */ + movdqu STATE0_LO, (0 * 16)(%rdi) + movdqu STATE0_HI, (1 * 16)(%rdi) + movdqu STATE1_LO, (2 * 16)(%rdi) + movdqu STATE1_HI, (3 * 16)(%rdi) + movdqu STATE2_LO, (4 * 16)(%rdi) + movdqu STATE2_HI, (5 * 16)(%rdi) + movdqu STATE3_LO, (6 * 16)(%rdi) + movdqu STATE3_HI, (7 * 16)(%rdi) + movdqu STATE4_LO, (8 * 16)(%rdi) + movdqu STATE4_HI, (9 * 16)(%rdi) + + FRAME_END +ENDPROC(crypto_morus1280_sse2_enc_tail) + +/* + * void crypto_morus1280_sse2_dec(void *state, const void *src, void *dst, + * unsigned int length); + */ +ENTRY(crypto_morus1280_sse2_dec) + FRAME_BEGIN + + cmp $32, %rcx + jb .Ldec_out + + /* load the state: */ + movdqu (0 * 16)(%rdi), STATE0_LO + movdqu (1 * 16)(%rdi), STATE0_HI + movdqu (2 * 16)(%rdi), STATE1_LO + movdqu (3 * 16)(%rdi), STATE1_HI + movdqu (4 * 16)(%rdi), STATE2_LO + movdqu (5 * 16)(%rdi), STATE2_HI + movdqu (6 * 16)(%rdi), STATE3_LO + movdqu (7 * 16)(%rdi), STATE3_HI + movdqu (8 * 16)(%rdi), STATE4_LO + movdqu (9 * 16)(%rdi), STATE4_HI + + mov %rsi, %r8 + or %rdx, %r8 + and $0xF, %r8 + jnz .Ldec_u_loop + +.align 4 +.Ldec_a_loop: + movdqa 0(%rsi), MSG_LO + movdqa 16(%rsi), MSG_HI + pxor STATE0_LO, MSG_LO + pxor STATE0_HI, MSG_HI + movdqa STATE1_LO, T1_LO + movdqa STATE1_HI, T1_HI + rol3 T1_HI, T1_LO + pxor T1_LO, MSG_LO + pxor T1_HI, MSG_HI + movdqa STATE2_LO, T1_LO + movdqa STATE2_HI, T1_HI + pand STATE3_LO, T1_LO + pand STATE3_HI, T1_HI + pxor T1_LO, MSG_LO + pxor T1_HI, MSG_HI + movdqa MSG_LO, 0(%rdx) + movdqa MSG_HI, 16(%rdx) + + call __morus1280_update + sub $32, %rcx + add $32, %rsi + add $32, %rdx + cmp $32, %rcx + jge .Ldec_a_loop + + jmp .Ldec_cont +.align 4 +.Ldec_u_loop: + movdqu 0(%rsi), MSG_LO + movdqu 16(%rsi), MSG_HI + pxor STATE0_LO, MSG_LO + pxor STATE0_HI, MSG_HI + movdqa STATE1_LO, T1_LO + movdqa STATE1_HI, T1_HI + rol3 T1_HI, T1_LO + pxor T1_LO, MSG_LO + pxor T1_HI, MSG_HI + movdqa STATE2_LO, T1_LO + movdqa STATE2_HI, T1_HI + pand STATE3_LO, T1_LO + pand STATE3_HI, T1_HI + pxor T1_LO, MSG_LO + pxor T1_HI, MSG_HI + movdqu MSG_LO, 0(%rdx) + movdqu MSG_HI, 16(%rdx) + + call __morus1280_update + sub $32, %rcx + add $32, %rsi + add $32, %rdx + cmp $32, %rcx + jge .Ldec_u_loop + +.Ldec_cont: + /* store the state: */ + movdqu STATE0_LO, (0 * 16)(%rdi) + movdqu STATE0_HI, (1 * 16)(%rdi) + movdqu STATE1_LO, (2 * 16)(%rdi) + movdqu STATE1_HI, (3 * 16)(%rdi) + movdqu STATE2_LO, (4 * 16)(%rdi) + movdqu STATE2_HI, (5 * 16)(%rdi) + movdqu STATE3_LO, (6 * 16)(%rdi) + movdqu STATE3_HI, (7 * 16)(%rdi) + movdqu STATE4_LO, (8 * 16)(%rdi) + movdqu STATE4_HI, (9 * 16)(%rdi) + +.Ldec_out: + FRAME_END + ret +ENDPROC(crypto_morus1280_sse2_dec) + +/* + * void crypto_morus1280_sse2_dec_tail(void *state, const void *src, void *dst, + * unsigned int length); + */ +ENTRY(crypto_morus1280_sse2_dec_tail) + FRAME_BEGIN + + /* load the state: */ + movdqu (0 * 16)(%rdi), STATE0_LO + movdqu (1 * 16)(%rdi), STATE0_HI + movdqu (2 * 16)(%rdi), STATE1_LO + movdqu (3 * 16)(%rdi), STATE1_HI + movdqu (4 * 16)(%rdi), STATE2_LO + movdqu (5 * 16)(%rdi), STATE2_HI + movdqu (6 * 16)(%rdi), STATE3_LO + movdqu (7 * 16)(%rdi), STATE3_HI + movdqu (8 * 16)(%rdi), STATE4_LO + movdqu (9 * 16)(%rdi), STATE4_HI + + /* decrypt message: */ + call __load_partial + + pxor STATE0_LO, MSG_LO + pxor STATE0_HI, MSG_HI + movdqa STATE1_LO, T1_LO + movdqa STATE1_HI, T1_HI + rol3 T1_HI, T1_LO + pxor T1_LO, MSG_LO + pxor T1_HI, MSG_HI + movdqa STATE2_LO, T1_LO + movdqa STATE2_HI, T1_HI + pand STATE3_LO, T1_LO + pand STATE3_HI, T1_HI + pxor T1_LO, MSG_LO + pxor T1_HI, MSG_HI + movdqa MSG_LO, T0_LO + movdqa MSG_HI, T0_HI + + call __store_partial + + /* mask with byte count: */ + movq %rcx, T0_LO + punpcklbw T0_LO, T0_LO + punpcklbw T0_LO, T0_LO + punpcklbw T0_LO, T0_LO + punpcklbw T0_LO, T0_LO + movdqa T0_LO, T0_HI + movdqa .Lmorus640_counter_0, T1_LO + movdqa .Lmorus640_counter_1, T1_HI + pcmpgtb T1_LO, T0_LO + pcmpgtb T1_HI, T0_HI + pand T0_LO, MSG_LO + pand T0_HI, MSG_HI + + call __morus1280_update + + /* store the state: */ + movdqu STATE0_LO, (0 * 16)(%rdi) + movdqu STATE0_HI, (1 * 16)(%rdi) + movdqu STATE1_LO, (2 * 16)(%rdi) + movdqu STATE1_HI, (3 * 16)(%rdi) + movdqu STATE2_LO, (4 * 16)(%rdi) + movdqu STATE2_HI, (5 * 16)(%rdi) + movdqu STATE3_LO, (6 * 16)(%rdi) + movdqu STATE3_HI, (7 * 16)(%rdi) + movdqu STATE4_LO, (8 * 16)(%rdi) + movdqu STATE4_HI, (9 * 16)(%rdi) + + FRAME_END + ret +ENDPROC(crypto_morus1280_sse2_dec_tail) + +/* + * void crypto_morus1280_sse2_final(void *state, void *tag_xor, + * u64 assoclen, u64 cryptlen); + */ +ENTRY(crypto_morus1280_sse2_final) + FRAME_BEGIN + + /* load the state: */ + movdqu (0 * 16)(%rdi), STATE0_LO + movdqu (1 * 16)(%rdi), STATE0_HI + movdqu (2 * 16)(%rdi), STATE1_LO + movdqu (3 * 16)(%rdi), STATE1_HI + movdqu (4 * 16)(%rdi), STATE2_LO + movdqu (5 * 16)(%rdi), STATE2_HI + movdqu (6 * 16)(%rdi), STATE3_LO + movdqu (7 * 16)(%rdi), STATE3_HI + movdqu (8 * 16)(%rdi), STATE4_LO + movdqu (9 * 16)(%rdi), STATE4_HI + + /* xor state[0] into state[4]: */ + pxor STATE0_LO, STATE4_LO + pxor STATE0_HI, STATE4_HI + + /* prepare length block: */ + movq %rdx, MSG_LO + movq %rcx, T0_LO + pslldq $8, T0_LO + pxor T0_LO, MSG_LO + psllq $3, MSG_LO /* multiply by 8 (to get bit count) */ + pxor MSG_HI, MSG_HI + + /* update state: */ + call __morus1280_update + call __morus1280_update + call __morus1280_update + call __morus1280_update + call __morus1280_update + call __morus1280_update + call __morus1280_update + call __morus1280_update + call __morus1280_update + call __morus1280_update + + /* xor tag: */ + movdqu 0(%rsi), MSG_LO + movdqu 16(%rsi), MSG_HI + + pxor STATE0_LO, MSG_LO + pxor STATE0_HI, MSG_HI + movdqa STATE1_LO, T0_LO + movdqa STATE1_HI, T0_HI + rol3 T0_HI, T0_LO + pxor T0_LO, MSG_LO + pxor T0_HI, MSG_HI + movdqa STATE2_LO, T0_LO + movdqa STATE2_HI, T0_HI + pand STATE3_LO, T0_LO + pand STATE3_HI, T0_HI + pxor T0_LO, MSG_LO + pxor T0_HI, MSG_HI + + movdqu MSG_LO, 0(%rsi) + movdqu MSG_HI, 16(%rsi) + + FRAME_END + ret +ENDPROC(crypto_morus1280_sse2_final) diff --git a/arch/x86/crypto/morus1280-sse2-glue.c b/arch/x86/crypto/morus1280-sse2-glue.c new file mode 100644 index 000000000000..839270aa713c --- /dev/null +++ b/arch/x86/crypto/morus1280-sse2-glue.c @@ -0,0 +1,68 @@ +/* + * The MORUS-1280 Authenticated-Encryption Algorithm + * Glue for SSE2 implementation + * + * Copyright (c) 2016-2018 Ondrej Mosnacek <omosnacek@gmail.com> + * Copyright (C) 2017-2018 Red Hat, Inc. All rights reserved. + * + * This program is free software; you can redistribute it and/or modify it + * under the terms of the GNU General Public License as published by the Free + * Software Foundation; either version 2 of the License, or (at your option) + * any later version. + */ + +#include <crypto/internal/aead.h> +#include <crypto/morus1280_glue.h> +#include <linux/module.h> +#include <asm/fpu/api.h> +#include <asm/cpu_device_id.h> + +asmlinkage void crypto_morus1280_sse2_init(void *state, const void *key, + const void *iv); +asmlinkage void crypto_morus1280_sse2_ad(void *state, const void *data, + unsigned int length); + +asmlinkage void crypto_morus1280_sse2_enc(void *state, const void *src, + void *dst, unsigned int length); +asmlinkage void crypto_morus1280_sse2_dec(void *state, const void *src, + void *dst, unsigned int length); + +asmlinkage void crypto_morus1280_sse2_enc_tail(void *state, const void *src, + void *dst, unsigned int length); +asmlinkage void crypto_morus1280_sse2_dec_tail(void *state, const void *src, + void *dst, unsigned int length); + +asmlinkage void crypto_morus1280_sse2_final(void *state, void *tag_xor, + u64 assoclen, u64 cryptlen); + +MORUS1280_DECLARE_ALGS(sse2, "morus1280-sse2", 350); + +static const struct x86_cpu_id sse2_cpu_id[] = { + X86_FEATURE_MATCH(X86_FEATURE_XMM2), + {} +}; +MODULE_DEVICE_TABLE(x86cpu, sse2_cpu_id); + +static int __init crypto_morus1280_sse2_module_init(void) +{ + if (!x86_match_cpu(sse2_cpu_id)) + return -ENODEV; + + return crypto_register_aeads(crypto_morus1280_sse2_algs, + ARRAY_SIZE(crypto_morus1280_sse2_algs)); +} + +static void __exit crypto_morus1280_sse2_module_exit(void) +{ + crypto_unregister_aeads(crypto_morus1280_sse2_algs, + ARRAY_SIZE(crypto_morus1280_sse2_algs)); +} + +module_init(crypto_morus1280_sse2_module_init); +module_exit(crypto_morus1280_sse2_module_exit); + +MODULE_LICENSE("GPL"); +MODULE_AUTHOR("Ondrej Mosnacek <omosnacek@gmail.com>"); +MODULE_DESCRIPTION("MORUS-1280 AEAD algorithm -- SSE2 implementation"); +MODULE_ALIAS_CRYPTO("morus1280"); +MODULE_ALIAS_CRYPTO("morus1280-sse2"); diff --git a/arch/x86/crypto/morus1280_glue.c b/arch/x86/crypto/morus1280_glue.c new file mode 100644 index 000000000000..0dccdda1eb3a --- /dev/null +++ b/arch/x86/crypto/morus1280_glue.c @@ -0,0 +1,302 @@ +/* + * The MORUS-1280 Authenticated-Encryption Algorithm + * Common x86 SIMD glue skeleton + * + * Copyright (c) 2016-2018 Ondrej Mosnacek <omosnacek@gmail.com> + * Copyright (C) 2017-2018 Red Hat, Inc. All rights reserved. + * + * This program is free software; you can redistribute it and/or modify it + * under the terms of the GNU General Public License as published by the Free + * Software Foundation; either version 2 of the License, or (at your option) + * any later version. + */ + +#include <crypto/cryptd.h> +#include <crypto/internal/aead.h> +#include <crypto/internal/skcipher.h> +#include <crypto/morus1280_glue.h> +#include <crypto/scatterwalk.h> +#include <linux/err.h> +#include <linux/init.h> +#include <linux/kernel.h> +#include <linux/module.h> +#include <linux/scatterlist.h> +#include <asm/fpu/api.h> + +struct morus1280_state { + struct morus1280_block s[MORUS_STATE_BLOCKS]; +}; + +struct morus1280_ops { + int (*skcipher_walk_init)(struct skcipher_walk *walk, + struct aead_request *req, bool atomic); + + void (*crypt_blocks)(void *state, const void *src, void *dst, + unsigned int length); + void (*crypt_tail)(void *state, const void *src, void *dst, + unsigned int length); +}; + +static void crypto_morus1280_glue_process_ad( + struct morus1280_state *state, + const struct morus1280_glue_ops *ops, + struct scatterlist *sg_src, unsigned int assoclen) +{ + struct scatter_walk walk; + struct morus1280_block buf; + unsigned int pos = 0; + + scatterwalk_start(&walk, sg_src); + while (assoclen != 0) { + unsigned int size = scatterwalk_clamp(&walk, assoclen); + unsigned int left = size; + void *mapped = scatterwalk_map(&walk); + const u8 *src = (const u8 *)mapped; + + if (pos + size >= MORUS1280_BLOCK_SIZE) { + if (pos > 0) { + unsigned int fill = MORUS1280_BLOCK_SIZE - pos; + memcpy(buf.bytes + pos, src, fill); + ops->ad(state, buf.bytes, MORUS1280_BLOCK_SIZE); + pos = 0; + left -= fill; + src += fill; + } + + ops->ad(state, src, left); + src += left & ~(MORUS1280_BLOCK_SIZE - 1); + left &= MORUS1280_BLOCK_SIZE - 1; + } + + memcpy(buf.bytes + pos, src, left); + + pos += left; + assoclen -= size; + scatterwalk_unmap(mapped); + scatterwalk_advance(&walk, size); + scatterwalk_done(&walk, 0, assoclen); + } + + if (pos > 0) { + memset(buf.bytes + pos, 0, MORUS1280_BLOCK_SIZE - pos); + ops->ad(state, buf.bytes, MORUS1280_BLOCK_SIZE); + } +} + +static void crypto_morus1280_glue_process_crypt(struct morus1280_state *state, + struct morus1280_ops ops, + struct aead_request *req) +{ + struct skcipher_walk walk; + u8 *cursor_src, *cursor_dst; + unsigned int chunksize, base; + + ops.skcipher_walk_init(&walk, req, false); + + while (walk.nbytes) { + cursor_src = walk.src.virt.addr; + cursor_dst = walk.dst.virt.addr; + chunksize = walk.nbytes; + + ops.crypt_blocks(state, cursor_src, cursor_dst, chunksize); + + base = chunksize & ~(MORUS1280_BLOCK_SIZE - 1); + cursor_src += base; + cursor_dst += base; + chunksize &= MORUS1280_BLOCK_SIZE - 1; + + if (chunksize > 0) + ops.crypt_tail(state, cursor_src, cursor_dst, + chunksize); + + skcipher_walk_done(&walk, 0); + } +} + +int crypto_morus1280_glue_setkey(struct crypto_aead *aead, const u8 *key, + unsigned int keylen) +{ + struct morus1280_ctx *ctx = crypto_aead_ctx(aead); + + if (keylen == MORUS1280_BLOCK_SIZE) { + memcpy(ctx->key.bytes, key, MORUS1280_BLOCK_SIZE); + } else if (keylen == MORUS1280_BLOCK_SIZE / 2) { + memcpy(ctx->key.bytes, key, keylen); + memcpy(ctx->key.bytes + keylen, key, keylen); + } else { + crypto_aead_set_flags(aead, CRYPTO_TFM_RES_BAD_KEY_LEN); + return -EINVAL; + } + + return 0; +} +EXPORT_SYMBOL_GPL(crypto_morus1280_glue_setkey); + +int crypto_morus1280_glue_setauthsize(struct crypto_aead *tfm, + unsigned int authsize) +{ + return (authsize <= MORUS_MAX_AUTH_SIZE) ? 0 : -EINVAL; +} +EXPORT_SYMBOL_GPL(crypto_morus1280_glue_setauthsize); + +static void crypto_morus1280_glue_crypt(struct aead_request *req, + struct morus1280_ops ops, + unsigned int cryptlen, + struct morus1280_block *tag_xor) +{ + struct crypto_aead *tfm = crypto_aead_reqtfm(req); + struct morus1280_ctx *ctx = crypto_aead_ctx(tfm); + struct morus1280_state state; + + kernel_fpu_begin(); + + ctx->ops->init(&state, &ctx->key, req->iv); + crypto_morus1280_glue_process_ad(&state, ctx->ops, req->src, req->assoclen); + crypto_morus1280_glue_process_crypt(&state, ops, req); + ctx->ops->final(&state, tag_xor, req->assoclen, cryptlen); + + kernel_fpu_end(); +} + +int crypto_morus1280_glue_encrypt(struct aead_request *req) +{ + struct crypto_aead *tfm = crypto_aead_reqtfm(req); + struct morus1280_ctx *ctx = crypto_aead_ctx(tfm); + struct morus1280_ops OPS = { + .skcipher_walk_init = skcipher_walk_aead_encrypt, + .crypt_blocks = ctx->ops->enc, + .crypt_tail = ctx->ops->enc_tail, + }; + + struct morus1280_block tag = {}; + unsigned int authsize = crypto_aead_authsize(tfm); + unsigned int cryptlen = req->cryptlen; + + crypto_morus1280_glue_crypt(req, OPS, cryptlen, &tag); + + scatterwalk_map_and_copy(tag.bytes, req->dst, + req->assoclen + cryptlen, authsize, 1); + return 0; +} +EXPORT_SYMBOL_GPL(crypto_morus1280_glue_encrypt); + +int crypto_morus1280_glue_decrypt(struct aead_request *req) +{ + static const u8 zeros[MORUS1280_BLOCK_SIZE] = {}; + + struct crypto_aead *tfm = crypto_aead_reqtfm(req); + struct morus1280_ctx *ctx = crypto_aead_ctx(tfm); + struct morus1280_ops OPS = { + .skcipher_walk_init = skcipher_walk_aead_decrypt, + .crypt_blocks = ctx->ops->dec, + .crypt_tail = ctx->ops->dec_tail, + }; + + struct morus1280_block tag; + unsigned int authsize = crypto_aead_authsize(tfm); + unsigned int cryptlen = req->cryptlen - authsize; + + scatterwalk_map_and_copy(tag.bytes, req->src, + req->assoclen + cryptlen, authsize, 0); + + crypto_morus1280_glue_crypt(req, OPS, cryptlen, &tag); + + return crypto_memneq(tag.bytes, zeros, authsize) ? -EBADMSG : 0; +} +EXPORT_SYMBOL_GPL(crypto_morus1280_glue_decrypt); + +void crypto_morus1280_glue_init_ops(struct crypto_aead *aead, + const struct morus1280_glue_ops *ops) +{ + struct morus1280_ctx *ctx = crypto_aead_ctx(aead); + ctx->ops = ops; +} +EXPORT_SYMBOL_GPL(crypto_morus1280_glue_init_ops); + +int cryptd_morus1280_glue_setkey(struct crypto_aead *aead, const u8 *key, + unsigned int keylen) +{ + struct cryptd_aead **ctx = crypto_aead_ctx(aead); + struct cryptd_aead *cryptd_tfm = *ctx; + + return crypto_aead_setkey(&cryptd_tfm->base, key, keylen); +} +EXPORT_SYMBOL_GPL(cryptd_morus1280_glue_setkey); + +int cryptd_morus1280_glue_setauthsize(struct crypto_aead *aead, + unsigned int authsize) +{ + struct cryptd_aead **ctx = crypto_aead_ctx(aead); + struct cryptd_aead *cryptd_tfm = *ctx; + + return crypto_aead_setauthsize(&cryptd_tfm->base, authsize); +} +EXPORT_SYMBOL_GPL(cryptd_morus1280_glue_setauthsize); + +int cryptd_morus1280_glue_encrypt(struct aead_request *req) +{ + struct crypto_aead *aead = crypto_aead_reqtfm(req); + struct cryptd_aead **ctx = crypto_aead_ctx(aead); + struct cryptd_aead *cryptd_tfm = *ctx; + + aead = &cryptd_tfm->base; + if (irq_fpu_usable() && (!in_atomic() || + !cryptd_aead_queued(cryptd_tfm))) + aead = cryptd_aead_child(cryptd_tfm); + + aead_request_set_tfm(req, aead); + + return crypto_aead_encrypt(req); +} +EXPORT_SYMBOL_GPL(cryptd_morus1280_glue_encrypt); + +int cryptd_morus1280_glue_decrypt(struct aead_request *req) +{ + struct crypto_aead *aead = crypto_aead_reqtfm(req); + struct cryptd_aead **ctx = crypto_aead_ctx(aead); + struct cryptd_aead *cryptd_tfm = *ctx; + + aead = &cryptd_tfm->base; + if (irq_fpu_usable() && (!in_atomic() || + !cryptd_aead_queued(cryptd_tfm))) + aead = cryptd_aead_child(cryptd_tfm); + + aead_request_set_tfm(req, aead); + + return crypto_aead_decrypt(req); +} +EXPORT_SYMBOL_GPL(cryptd_morus1280_glue_decrypt); + +int cryptd_morus1280_glue_init_tfm(struct crypto_aead *aead) +{ + struct cryptd_aead *cryptd_tfm; + struct cryptd_aead **ctx = crypto_aead_ctx(aead); + const char *name = crypto_aead_alg(aead)->base.cra_driver_name; + char internal_name[CRYPTO_MAX_ALG_NAME]; + + if (snprintf(internal_name, CRYPTO_MAX_ALG_NAME, "__%s", name) + >= CRYPTO_MAX_ALG_NAME) + return -ENAMETOOLONG; + + cryptd_tfm = cryptd_alloc_aead(internal_name, CRYPTO_ALG_INTERNAL, + CRYPTO_ALG_INTERNAL); + if (IS_ERR(cryptd_tfm)) + return PTR_ERR(cryptd_tfm); + + *ctx = cryptd_tfm; + crypto_aead_set_reqsize(aead, crypto_aead_reqsize(&cryptd_tfm->base)); + return 0; +} +EXPORT_SYMBOL_GPL(cryptd_morus1280_glue_init_tfm); + +void cryptd_morus1280_glue_exit_tfm(struct crypto_aead *aead) +{ + struct cryptd_aead **ctx = crypto_aead_ctx(aead); + + cryptd_free_aead(*ctx); +} +EXPORT_SYMBOL_GPL(cryptd_morus1280_glue_exit_tfm); + +MODULE_LICENSE("GPL"); +MODULE_AUTHOR("Ondrej Mosnacek <omosnacek@gmail.com>"); +MODULE_DESCRIPTION("MORUS-1280 AEAD mode -- glue for x86 optimizations"); diff --git a/arch/x86/crypto/morus640-sse2-asm.S b/arch/x86/crypto/morus640-sse2-asm.S new file mode 100644 index 000000000000..71c72a0a0862 --- /dev/null +++ b/arch/x86/crypto/morus640-sse2-asm.S @@ -0,0 +1,614 @@ +/* + * SSE2 implementation of MORUS-640 + * + * Copyright (c) 2017-2018 Ondrej Mosnacek <omosnacek@gmail.com> + * Copyright (C) 2017-2018 Red Hat, Inc. All rights reserved. + * + * 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. + */ + +#include <linux/linkage.h> +#include <asm/frame.h> + +#define SHUFFLE_MASK(i0, i1, i2, i3) \ + (i0 | (i1 << 2) | (i2 << 4) | (i3 << 6)) + +#define MASK1 SHUFFLE_MASK(3, 0, 1, 2) +#define MASK2 SHUFFLE_MASK(2, 3, 0, 1) +#define MASK3 SHUFFLE_MASK(1, 2, 3, 0) + +#define STATE0 %xmm0 +#define STATE1 %xmm1 +#define STATE2 %xmm2 +#define STATE3 %xmm3 +#define STATE4 %xmm4 +#define KEY %xmm5 +#define MSG %xmm5 +#define T0 %xmm6 +#define T1 %xmm7 + +.section .rodata.cst16.morus640_const, "aM", @progbits, 32 +.align 16 +.Lmorus640_const_0: + .byte 0x00, 0x01, 0x01, 0x02, 0x03, 0x05, 0x08, 0x0d + .byte 0x15, 0x22, 0x37, 0x59, 0x90, 0xe9, 0x79, 0x62 +.Lmorus640_const_1: + .byte 0xdb, 0x3d, 0x18, 0x55, 0x6d, 0xc2, 0x2f, 0xf1 + .byte 0x20, 0x11, 0x31, 0x42, 0x73, 0xb5, 0x28, 0xdd + +.section .rodata.cst16.morus640_counter, "aM", @progbits, 16 +.align 16 +.Lmorus640_counter: + .byte 0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07 + .byte 0x08, 0x09, 0x0a, 0x0b, 0x0c, 0x0d, 0x0e, 0x0f + +.text + +.macro morus640_round s0, s1, s2, s3, s4, b, w + movdqa \s1, T0 + pand \s2, T0 + pxor T0, \s0 + pxor \s3, \s0 + movdqa \s0, T0 + pslld $\b, T0 + psrld $(32 - \b), \s0 + pxor T0, \s0 + pshufd $\w, \s3, \s3 +.endm + +/* + * __morus640_update: internal ABI + * input: + * STATE[0-4] - input state + * MSG - message block + * output: + * STATE[0-4] - output state + * changed: + * T0 + */ +__morus640_update: + morus640_round STATE0, STATE1, STATE2, STATE3, STATE4, 5, MASK1 + pxor MSG, STATE1 + morus640_round STATE1, STATE2, STATE3, STATE4, STATE0, 31, MASK2 + pxor MSG, STATE2 + morus640_round STATE2, STATE3, STATE4, STATE0, STATE1, 7, MASK3 + pxor MSG, STATE3 + morus640_round STATE3, STATE4, STATE0, STATE1, STATE2, 22, MASK2 + pxor MSG, STATE4 + morus640_round STATE4, STATE0, STATE1, STATE2, STATE3, 13, MASK1 + ret +ENDPROC(__morus640_update) + + +/* + * __morus640_update_zero: internal ABI + * input: + * STATE[0-4] - input state + * output: + * STATE[0-4] - output state + * changed: + * T0 + */ +__morus640_update_zero: + morus640_round STATE0, STATE1, STATE2, STATE3, STATE4, 5, MASK1 + morus640_round STATE1, STATE2, STATE3, STATE4, STATE0, 31, MASK2 + morus640_round STATE2, STATE3, STATE4, STATE0, STATE1, 7, MASK3 + morus640_round STATE3, STATE4, STATE0, STATE1, STATE2, 22, MASK2 + morus640_round STATE4, STATE0, STATE1, STATE2, STATE3, 13, MASK1 + ret +ENDPROC(__morus640_update_zero) + +/* + * __load_partial: internal ABI + * input: + * %rsi - src + * %rcx - bytes + * output: + * MSG - message block + * changed: + * T0 + * %r8 + * %r9 + */ +__load_partial: + xor %r9, %r9 + pxor MSG, MSG + + mov %rcx, %r8 + and $0x1, %r8 + jz .Lld_partial_1 + + mov %rcx, %r8 + and $0x1E, %r8 + add %rsi, %r8 + mov (%r8), %r9b + +.Lld_partial_1: + mov %rcx, %r8 + and $0x2, %r8 + jz .Lld_partial_2 + + mov %rcx, %r8 + and $0x1C, %r8 + add %rsi, %r8 + shl $16, %r9 + mov (%r8), %r9w + +.Lld_partial_2: + mov %rcx, %r8 + and $0x4, %r8 + jz .Lld_partial_4 + + mov %rcx, %r8 + and $0x18, %r8 + add %rsi, %r8 + shl $32, %r9 + mov (%r8), %r8d + xor %r8, %r9 + +.Lld_partial_4: + movq %r9, MSG + + mov %rcx, %r8 + and $0x8, %r8 + jz .Lld_partial_8 + + mov %rcx, %r8 + and $0x10, %r8 + add %rsi, %r8 + pslldq $8, MSG + movq (%r8), T0 + pxor T0, MSG + +.Lld_partial_8: + ret +ENDPROC(__load_partial) + +/* + * __store_partial: internal ABI + * input: + * %rdx - dst + * %rcx - bytes + * output: + * T0 - message block + * changed: + * %r8 + * %r9 + * %r10 + */ +__store_partial: + mov %rcx, %r8 + mov %rdx, %r9 + + movq T0, %r10 + + cmp $8, %r8 + jl .Lst_partial_8 + + mov %r10, (%r9) + psrldq $8, T0 + movq T0, %r10 + + sub $8, %r8 + add $8, %r9 + +.Lst_partial_8: + cmp $4, %r8 + jl .Lst_partial_4 + + mov %r10d, (%r9) + shr $32, %r10 + + sub $4, %r8 + add $4, %r9 + +.Lst_partial_4: + cmp $2, %r8 + jl .Lst_partial_2 + + mov %r10w, (%r9) + shr $16, %r10 + + sub $2, %r8 + add $2, %r9 + +.Lst_partial_2: + cmp $1, %r8 + jl .Lst_partial_1 + + mov %r10b, (%r9) + +.Lst_partial_1: + ret +ENDPROC(__store_partial) + +/* + * void crypto_morus640_sse2_init(void *state, const void *key, const void *iv); + */ +ENTRY(crypto_morus640_sse2_init) + FRAME_BEGIN + + /* load IV: */ + movdqu (%rdx), STATE0 + /* load key: */ + movdqu (%rsi), KEY + movdqa KEY, STATE1 + /* load all ones: */ + pcmpeqd STATE2, STATE2 + /* load the constants: */ + movdqa .Lmorus640_const_0, STATE3 + movdqa .Lmorus640_const_1, STATE4 + + /* update 16 times with zero: */ + call __morus640_update_zero + call __morus640_update_zero + call __morus640_update_zero + call __morus640_update_zero + call __morus640_update_zero + call __morus640_update_zero + call __morus640_update_zero + call __morus640_update_zero + call __morus640_update_zero + call __morus640_update_zero + call __morus640_update_zero + call __morus640_update_zero + call __morus640_update_zero + call __morus640_update_zero + call __morus640_update_zero + call __morus640_update_zero + + /* xor-in the key again after updates: */ + pxor KEY, STATE1 + + /* store the state: */ + movdqu STATE0, (0 * 16)(%rdi) + movdqu STATE1, (1 * 16)(%rdi) + movdqu STATE2, (2 * 16)(%rdi) + movdqu STATE3, (3 * 16)(%rdi) + movdqu STATE4, (4 * 16)(%rdi) + + FRAME_END + ret +ENDPROC(crypto_morus640_sse2_init) + +/* + * void crypto_morus640_sse2_ad(void *state, const void *data, + * unsigned int length); + */ +ENTRY(crypto_morus640_sse2_ad) + FRAME_BEGIN + + cmp $16, %rdx + jb .Lad_out + + /* load the state: */ + movdqu (0 * 16)(%rdi), STATE0 + movdqu (1 * 16)(%rdi), STATE1 + movdqu (2 * 16)(%rdi), STATE2 + movdqu (3 * 16)(%rdi), STATE3 + movdqu (4 * 16)(%rdi), STATE4 + + mov %rsi, %r8 + and $0xF, %r8 + jnz .Lad_u_loop + +.align 4 +.Lad_a_loop: + movdqa (%rsi), MSG + call __morus640_update + sub $16, %rdx + add $16, %rsi + cmp $16, %rdx + jge .Lad_a_loop + + jmp .Lad_cont +.align 4 +.Lad_u_loop: + movdqu (%rsi), MSG + call __morus640_update + sub $16, %rdx + add $16, %rsi + cmp $16, %rdx + jge .Lad_u_loop + +.Lad_cont: + /* store the state: */ + movdqu STATE0, (0 * 16)(%rdi) + movdqu STATE1, (1 * 16)(%rdi) + movdqu STATE2, (2 * 16)(%rdi) + movdqu STATE3, (3 * 16)(%rdi) + movdqu STATE4, (4 * 16)(%rdi) + +.Lad_out: + FRAME_END + ret +ENDPROC(crypto_morus640_sse2_ad) + +/* + * void crypto_morus640_sse2_enc(void *state, const void *src, void *dst, + * unsigned int length); + */ +ENTRY(crypto_morus640_sse2_enc) + FRAME_BEGIN + + cmp $16, %rcx + jb .Lenc_out + + /* load the state: */ + movdqu (0 * 16)(%rdi), STATE0 + movdqu (1 * 16)(%rdi), STATE1 + movdqu (2 * 16)(%rdi), STATE2 + movdqu (3 * 16)(%rdi), STATE3 + movdqu (4 * 16)(%rdi), STATE4 + + mov %rsi, %r8 + or %rdx, %r8 + and $0xF, %r8 + jnz .Lenc_u_loop + +.align 4 +.Lenc_a_loop: + movdqa (%rsi), MSG + movdqa MSG, T0 + pxor STATE0, T0 + pshufd $MASK3, STATE1, T1 + pxor T1, T0 + movdqa STATE2, T1 + pand STATE3, T1 + pxor T1, T0 + movdqa T0, (%rdx) + + call __morus640_update + sub $16, %rcx + add $16, %rsi + add $16, %rdx + cmp $16, %rcx + jge .Lenc_a_loop + + jmp .Lenc_cont +.align 4 +.Lenc_u_loop: + movdqu (%rsi), MSG + movdqa MSG, T0 + pxor STATE0, T0 + pshufd $MASK3, STATE1, T1 + pxor T1, T0 + movdqa STATE2, T1 + pand STATE3, T1 + pxor T1, T0 + movdqu T0, (%rdx) + + call __morus640_update + sub $16, %rcx + add $16, %rsi + add $16, %rdx + cmp $16, %rcx + jge .Lenc_u_loop + +.Lenc_cont: + /* store the state: */ + movdqu STATE0, (0 * 16)(%rdi) + movdqu STATE1, (1 * 16)(%rdi) + movdqu STATE2, (2 * 16)(%rdi) + movdqu STATE3, (3 * 16)(%rdi) + movdqu STATE4, (4 * 16)(%rdi) + +.Lenc_out: + FRAME_END + ret +ENDPROC(crypto_morus640_sse2_enc) + +/* + * void crypto_morus640_sse2_enc_tail(void *state, const void *src, void *dst, + * unsigned int length); + */ +ENTRY(crypto_morus640_sse2_enc_tail) + FRAME_BEGIN + + /* load the state: */ + movdqu (0 * 16)(%rdi), STATE0 + movdqu (1 * 16)(%rdi), STATE1 + movdqu (2 * 16)(%rdi), STATE2 + movdqu (3 * 16)(%rdi), STATE3 + movdqu (4 * 16)(%rdi), STATE4 + + /* encrypt message: */ + call __load_partial + + movdqa MSG, T0 + pxor STATE0, T0 + pshufd $MASK3, STATE1, T1 + pxor T1, T0 + movdqa STATE2, T1 + pand STATE3, T1 + pxor T1, T0 + + call __store_partial + + call __morus640_update + + /* store the state: */ + movdqu STATE0, (0 * 16)(%rdi) + movdqu STATE1, (1 * 16)(%rdi) + movdqu STATE2, (2 * 16)(%rdi) + movdqu STATE3, (3 * 16)(%rdi) + movdqu STATE4, (4 * 16)(%rdi) + + FRAME_END +ENDPROC(crypto_morus640_sse2_enc_tail) + +/* + * void crypto_morus640_sse2_dec(void *state, const void *src, void *dst, + * unsigned int length); + */ +ENTRY(crypto_morus640_sse2_dec) + FRAME_BEGIN + + cmp $16, %rcx + jb .Ldec_out + + /* load the state: */ + movdqu (0 * 16)(%rdi), STATE0 + movdqu (1 * 16)(%rdi), STATE1 + movdqu (2 * 16)(%rdi), STATE2 + movdqu (3 * 16)(%rdi), STATE3 + movdqu (4 * 16)(%rdi), STATE4 + + mov %rsi, %r8 + or %rdx, %r8 + and $0xF, %r8 + jnz .Ldec_u_loop + +.align 4 +.Ldec_a_loop: + movdqa (%rsi), MSG + pxor STATE0, MSG + pshufd $MASK3, STATE1, T0 + pxor T0, MSG + movdqa STATE2, T0 + pand STATE3, T0 + pxor T0, MSG + movdqa MSG, (%rdx) + + call __morus640_update + sub $16, %rcx + add $16, %rsi + add $16, %rdx + cmp $16, %rcx + jge .Ldec_a_loop + + jmp .Ldec_cont +.align 4 +.Ldec_u_loop: + movdqu (%rsi), MSG + pxor STATE0, MSG + pshufd $MASK3, STATE1, T0 + pxor T0, MSG + movdqa STATE2, T0 + pand STATE3, T0 + pxor T0, MSG + movdqu MSG, (%rdx) + + call __morus640_update + sub $16, %rcx + add $16, %rsi + add $16, %rdx + cmp $16, %rcx + jge .Ldec_u_loop + +.Ldec_cont: + /* store the state: */ + movdqu STATE0, (0 * 16)(%rdi) + movdqu STATE1, (1 * 16)(%rdi) + movdqu STATE2, (2 * 16)(%rdi) + movdqu STATE3, (3 * 16)(%rdi) + movdqu STATE4, (4 * 16)(%rdi) + +.Ldec_out: + FRAME_END + ret +ENDPROC(crypto_morus640_sse2_dec) + +/* + * void crypto_morus640_sse2_dec_tail(void *state, const void *src, void *dst, + * unsigned int length); + */ +ENTRY(crypto_morus640_sse2_dec_tail) + FRAME_BEGIN + + /* load the state: */ + movdqu (0 * 16)(%rdi), STATE0 + movdqu (1 * 16)(%rdi), STATE1 + movdqu (2 * 16)(%rdi), STATE2 + movdqu (3 * 16)(%rdi), STATE3 + movdqu (4 * 16)(%rdi), STATE4 + + /* decrypt message: */ + call __load_partial + + pxor STATE0, MSG + pshufd $MASK3, STATE1, T0 + pxor T0, MSG + movdqa STATE2, T0 + pand STATE3, T0 + pxor T0, MSG + movdqa MSG, T0 + + call __store_partial + + /* mask with byte count: */ + movq %rcx, T0 + punpcklbw T0, T0 + punpcklbw T0, T0 + punpcklbw T0, T0 + punpcklbw T0, T0 + movdqa .Lmorus640_counter, T1 + pcmpgtb T1, T0 + pand T0, MSG + + call __morus640_update + + /* store the state: */ + movdqu STATE0, (0 * 16)(%rdi) + movdqu STATE1, (1 * 16)(%rdi) + movdqu STATE2, (2 * 16)(%rdi) + movdqu STATE3, (3 * 16)(%rdi) + movdqu STATE4, (4 * 16)(%rdi) + + FRAME_END + ret +ENDPROC(crypto_morus640_sse2_dec_tail) + +/* + * void crypto_morus640_sse2_final(void *state, void *tag_xor, + * u64 assoclen, u64 cryptlen); + */ +ENTRY(crypto_morus640_sse2_final) + FRAME_BEGIN + + /* load the state: */ + movdqu (0 * 16)(%rdi), STATE0 + movdqu (1 * 16)(%rdi), STATE1 + movdqu (2 * 16)(%rdi), STATE2 + movdqu (3 * 16)(%rdi), STATE3 + movdqu (4 * 16)(%rdi), STATE4 + + /* xor state[0] into state[4]: */ + pxor STATE0, STATE4 + + /* prepare length block: */ + movq %rdx, MSG + movq %rcx, T0 + pslldq $8, T0 + pxor T0, MSG + psllq $3, MSG /* multiply by 8 (to get bit count) */ + + /* update state: */ + call __morus640_update + call __morus640_update + call __morus640_update + call __morus640_update + call __morus640_update + call __morus640_update + call __morus640_update + call __morus640_update + call __morus640_update + call __morus640_update + + /* xor tag: */ + movdqu (%rsi), MSG + + pxor STATE0, MSG + pshufd $MASK3, STATE1, T0 + pxor T0, MSG + movdqa STATE2, T0 + pand STATE3, T0 + pxor T0, MSG + + movdqu MSG, (%rsi) + + FRAME_END + ret +ENDPROC(crypto_morus640_sse2_final) diff --git a/arch/x86/crypto/morus640-sse2-glue.c b/arch/x86/crypto/morus640-sse2-glue.c new file mode 100644 index 000000000000..26b47e2db8d2 --- /dev/null +++ b/arch/x86/crypto/morus640-sse2-glue.c @@ -0,0 +1,68 @@ +/* + * The MORUS-640 Authenticated-Encryption Algorithm + * Glue for SSE2 implementation + * + * Copyright (c) 2016-2018 Ondrej Mosnacek <omosnacek@gmail.com> + * Copyright (C) 2017-2018 Red Hat, Inc. All rights reserved. + * + * This program is free software; you can redistribute it and/or modify it + * under the terms of the GNU General Public License as published by the Free + * Software Foundation; either version 2 of the License, or (at your option) + * any later version. + */ + +#include <crypto/internal/aead.h> +#include <crypto/morus640_glue.h> +#include <linux/module.h> +#include <asm/fpu/api.h> +#include <asm/cpu_device_id.h> + +asmlinkage void crypto_morus640_sse2_init(void *state, const void *key, + const void *iv); +asmlinkage void crypto_morus640_sse2_ad(void *state, const void *data, + unsigned int length); + +asmlinkage void crypto_morus640_sse2_enc(void *state, const void *src, + void *dst, unsigned int length); +asmlinkage void crypto_morus640_sse2_dec(void *state, const void *src, + void *dst, unsigned int length); + +asmlinkage void crypto_morus640_sse2_enc_tail(void *state, const void *src, + void *dst, unsigned int length); +asmlinkage void crypto_morus640_sse2_dec_tail(void *state, const void *src, + void *dst, unsigned int length); + +asmlinkage void crypto_morus640_sse2_final(void *state, void *tag_xor, + u64 assoclen, u64 cryptlen); + +MORUS640_DECLARE_ALGS(sse2, "morus640-sse2", 400); + +static const struct x86_cpu_id sse2_cpu_id[] = { + X86_FEATURE_MATCH(X86_FEATURE_XMM2), + {} +}; +MODULE_DEVICE_TABLE(x86cpu, sse2_cpu_id); + +static int __init crypto_morus640_sse2_module_init(void) +{ + if (!x86_match_cpu(sse2_cpu_id)) + return -ENODEV; + + return crypto_register_aeads(crypto_morus640_sse2_algs, + ARRAY_SIZE(crypto_morus640_sse2_algs)); +} + +static void __exit crypto_morus640_sse2_module_exit(void) +{ + crypto_unregister_aeads(crypto_morus640_sse2_algs, + ARRAY_SIZE(crypto_morus640_sse2_algs)); +} + +module_init(crypto_morus640_sse2_module_init); +module_exit(crypto_morus640_sse2_module_exit); + +MODULE_LICENSE("GPL"); +MODULE_AUTHOR("Ondrej Mosnacek <omosnacek@gmail.com>"); +MODULE_DESCRIPTION("MORUS-640 AEAD algorithm -- SSE2 implementation"); +MODULE_ALIAS_CRYPTO("morus640"); +MODULE_ALIAS_CRYPTO("morus640-sse2"); diff --git a/arch/x86/crypto/morus640_glue.c b/arch/x86/crypto/morus640_glue.c new file mode 100644 index 000000000000..7b58fe4d9bd1 --- /dev/null +++ b/arch/x86/crypto/morus640_glue.c @@ -0,0 +1,298 @@ +/* + * The MORUS-640 Authenticated-Encryption Algorithm + * Common x86 SIMD glue skeleton + * + * Copyright (c) 2016-2018 Ondrej Mosnacek <omosnacek@gmail.com> + * Copyright (C) 2017-2018 Red Hat, Inc. All rights reserved. + * + * This program is free software; you can redistribute it and/or modify it + * under the terms of the GNU General Public License as published by the Free + * Software Foundation; either version 2 of the License, or (at your option) + * any later version. + */ + +#include <crypto/cryptd.h> +#include <crypto/internal/aead.h> +#include <crypto/internal/skcipher.h> +#include <crypto/morus640_glue.h> +#include <crypto/scatterwalk.h> +#include <linux/err.h> +#include <linux/init.h> +#include <linux/kernel.h> +#include <linux/module.h> +#include <linux/scatterlist.h> +#include <asm/fpu/api.h> + +struct morus640_state { + struct morus640_block s[MORUS_STATE_BLOCKS]; +}; + +struct morus640_ops { + int (*skcipher_walk_init)(struct skcipher_walk *walk, + struct aead_request *req, bool atomic); + + void (*crypt_blocks)(void *state, const void *src, void *dst, + unsigned int length); + void (*crypt_tail)(void *state, const void *src, void *dst, + unsigned int length); +}; + +static void crypto_morus640_glue_process_ad( + struct morus640_state *state, + const struct morus640_glue_ops *ops, + struct scatterlist *sg_src, unsigned int assoclen) +{ + struct scatter_walk walk; + struct morus640_block buf; + unsigned int pos = 0; + + scatterwalk_start(&walk, sg_src); + while (assoclen != 0) { + unsigned int size = scatterwalk_clamp(&walk, assoclen); + unsigned int left = size; + void *mapped = scatterwalk_map(&walk); + const u8 *src = (const u8 *)mapped; + + if (pos + size >= MORUS640_BLOCK_SIZE) { + if (pos > 0) { + unsigned int fill = MORUS640_BLOCK_SIZE - pos; + memcpy(buf.bytes + pos, src, fill); + ops->ad(state, buf.bytes, MORUS640_BLOCK_SIZE); + pos = 0; + left -= fill; + src += fill; + } + + ops->ad(state, src, left); + src += left & ~(MORUS640_BLOCK_SIZE - 1); + left &= MORUS640_BLOCK_SIZE - 1; + } + + memcpy(buf.bytes + pos, src, left); + + pos += left; + assoclen -= size; + scatterwalk_unmap(mapped); + scatterwalk_advance(&walk, size); + scatterwalk_done(&walk, 0, assoclen); + } + + if (pos > 0) { + memset(buf.bytes + pos, 0, MORUS640_BLOCK_SIZE - pos); + ops->ad(state, buf.bytes, MORUS640_BLOCK_SIZE); + } +} + +static void crypto_morus640_glue_process_crypt(struct morus640_state *state, + struct morus640_ops ops, + struct aead_request *req) +{ + struct skcipher_walk walk; + u8 *cursor_src, *cursor_dst; + unsigned int chunksize, base; + + ops.skcipher_walk_init(&walk, req, false); + + while (walk.nbytes) { + cursor_src = walk.src.virt.addr; + cursor_dst = walk.dst.virt.addr; + chunksize = walk.nbytes; + + ops.crypt_blocks(state, cursor_src, cursor_dst, chunksize); + + base = chunksize & ~(MORUS640_BLOCK_SIZE - 1); + cursor_src += base; + cursor_dst += base; + chunksize &= MORUS640_BLOCK_SIZE - 1; + + if (chunksize > 0) + ops.crypt_tail(state, cursor_src, cursor_dst, + chunksize); + + skcipher_walk_done(&walk, 0); + } +} + +int crypto_morus640_glue_setkey(struct crypto_aead *aead, const u8 *key, + unsigned int keylen) +{ + struct morus640_ctx *ctx = crypto_aead_ctx(aead); + + if (keylen != MORUS640_BLOCK_SIZE) { + crypto_aead_set_flags(aead, CRYPTO_TFM_RES_BAD_KEY_LEN); + return -EINVAL; + } + + memcpy(ctx->key.bytes, key, MORUS640_BLOCK_SIZE); + return 0; +} +EXPORT_SYMBOL_GPL(crypto_morus640_glue_setkey); + +int crypto_morus640_glue_setauthsize(struct crypto_aead *tfm, + unsigned int authsize) +{ + return (authsize <= MORUS_MAX_AUTH_SIZE) ? 0 : -EINVAL; +} +EXPORT_SYMBOL_GPL(crypto_morus640_glue_setauthsize); + +static void crypto_morus640_glue_crypt(struct aead_request *req, + struct morus640_ops ops, + unsigned int cryptlen, + struct morus640_block *tag_xor) +{ + struct crypto_aead *tfm = crypto_aead_reqtfm(req); + struct morus640_ctx *ctx = crypto_aead_ctx(tfm); + struct morus640_state state; + + kernel_fpu_begin(); + + ctx->ops->init(&state, &ctx->key, req->iv); + crypto_morus640_glue_process_ad(&state, ctx->ops, req->src, req->assoclen); + crypto_morus640_glue_process_crypt(&state, ops, req); + ctx->ops->final(&state, tag_xor, req->assoclen, cryptlen); + + kernel_fpu_end(); +} + +int crypto_morus640_glue_encrypt(struct aead_request *req) +{ + struct crypto_aead *tfm = crypto_aead_reqtfm(req); + struct morus640_ctx *ctx = crypto_aead_ctx(tfm); + struct morus640_ops OPS = { + .skcipher_walk_init = skcipher_walk_aead_encrypt, + .crypt_blocks = ctx->ops->enc, + .crypt_tail = ctx->ops->enc_tail, + }; + + struct morus640_block tag = {}; + unsigned int authsize = crypto_aead_authsize(tfm); + unsigned int cryptlen = req->cryptlen; + + crypto_morus640_glue_crypt(req, OPS, cryptlen, &tag); + + scatterwalk_map_and_copy(tag.bytes, req->dst, + req->assoclen + cryptlen, authsize, 1); + return 0; +} +EXPORT_SYMBOL_GPL(crypto_morus640_glue_encrypt); + +int crypto_morus640_glue_decrypt(struct aead_request *req) +{ + static const u8 zeros[MORUS640_BLOCK_SIZE] = {}; + + struct crypto_aead *tfm = crypto_aead_reqtfm(req); + struct morus640_ctx *ctx = crypto_aead_ctx(tfm); + struct morus640_ops OPS = { + .skcipher_walk_init = skcipher_walk_aead_decrypt, + .crypt_blocks = ctx->ops->dec, + .crypt_tail = ctx->ops->dec_tail, + }; + + struct morus640_block tag; + unsigned int authsize = crypto_aead_authsize(tfm); + unsigned int cryptlen = req->cryptlen - authsize; + + scatterwalk_map_and_copy(tag.bytes, req->src, + req->assoclen + cryptlen, authsize, 0); + + crypto_morus640_glue_crypt(req, OPS, cryptlen, &tag); + + return crypto_memneq(tag.bytes, zeros, authsize) ? -EBADMSG : 0; +} +EXPORT_SYMBOL_GPL(crypto_morus640_glue_decrypt); + +void crypto_morus640_glue_init_ops(struct crypto_aead *aead, + const struct morus640_glue_ops *ops) +{ + struct morus640_ctx *ctx = crypto_aead_ctx(aead); + ctx->ops = ops; +} +EXPORT_SYMBOL_GPL(crypto_morus640_glue_init_ops); + +int cryptd_morus640_glue_setkey(struct crypto_aead *aead, const u8 *key, + unsigned int keylen) +{ + struct cryptd_aead **ctx = crypto_aead_ctx(aead); + struct cryptd_aead *cryptd_tfm = *ctx; + + return crypto_aead_setkey(&cryptd_tfm->base, key, keylen); +} +EXPORT_SYMBOL_GPL(cryptd_morus640_glue_setkey); + +int cryptd_morus640_glue_setauthsize(struct crypto_aead *aead, + unsigned int authsize) +{ + struct cryptd_aead **ctx = crypto_aead_ctx(aead); + struct cryptd_aead *cryptd_tfm = *ctx; + + return crypto_aead_setauthsize(&cryptd_tfm->base, authsize); +} +EXPORT_SYMBOL_GPL(cryptd_morus640_glue_setauthsize); + +int cryptd_morus640_glue_encrypt(struct aead_request *req) +{ + struct crypto_aead *aead = crypto_aead_reqtfm(req); + struct cryptd_aead **ctx = crypto_aead_ctx(aead); + struct cryptd_aead *cryptd_tfm = *ctx; + + aead = &cryptd_tfm->base; + if (irq_fpu_usable() && (!in_atomic() || + !cryptd_aead_queued(cryptd_tfm))) + aead = cryptd_aead_child(cryptd_tfm); + + aead_request_set_tfm(req, aead); + + return crypto_aead_encrypt(req); +} +EXPORT_SYMBOL_GPL(cryptd_morus640_glue_encrypt); + +int cryptd_morus640_glue_decrypt(struct aead_request *req) +{ + struct crypto_aead *aead = crypto_aead_reqtfm(req); + struct cryptd_aead **ctx = crypto_aead_ctx(aead); + struct cryptd_aead *cryptd_tfm = *ctx; + + aead = &cryptd_tfm->base; + if (irq_fpu_usable() && (!in_atomic() || + !cryptd_aead_queued(cryptd_tfm))) + aead = cryptd_aead_child(cryptd_tfm); + + aead_request_set_tfm(req, aead); + + return crypto_aead_decrypt(req); +} +EXPORT_SYMBOL_GPL(cryptd_morus640_glue_decrypt); + +int cryptd_morus640_glue_init_tfm(struct crypto_aead *aead) +{ + struct cryptd_aead *cryptd_tfm; + struct cryptd_aead **ctx = crypto_aead_ctx(aead); + const char *name = crypto_aead_alg(aead)->base.cra_driver_name; + char internal_name[CRYPTO_MAX_ALG_NAME]; + + if (snprintf(internal_name, CRYPTO_MAX_ALG_NAME, "__%s", name) + >= CRYPTO_MAX_ALG_NAME) + return -ENAMETOOLONG; + + cryptd_tfm = cryptd_alloc_aead(internal_name, CRYPTO_ALG_INTERNAL, + CRYPTO_ALG_INTERNAL); + if (IS_ERR(cryptd_tfm)) + return PTR_ERR(cryptd_tfm); + + *ctx = cryptd_tfm; + crypto_aead_set_reqsize(aead, crypto_aead_reqsize(&cryptd_tfm->base)); + return 0; +} +EXPORT_SYMBOL_GPL(cryptd_morus640_glue_init_tfm); + +void cryptd_morus640_glue_exit_tfm(struct crypto_aead *aead) +{ + struct cryptd_aead **ctx = crypto_aead_ctx(aead); + + cryptd_free_aead(*ctx); +} +EXPORT_SYMBOL_GPL(cryptd_morus640_glue_exit_tfm); + +MODULE_LICENSE("GPL"); +MODULE_AUTHOR("Ondrej Mosnacek <omosnacek@gmail.com>"); +MODULE_DESCRIPTION("MORUS-640 AEAD mode -- glue for x86 optimizations"); diff --git a/arch/x86/crypto/salsa20-i586-asm_32.S b/arch/x86/crypto/salsa20-i586-asm_32.S deleted file mode 100644 index 6014b7b9e52a..000000000000 --- a/arch/x86/crypto/salsa20-i586-asm_32.S +++ /dev/null @@ -1,938 +0,0 @@ -# Derived from: -# salsa20_pm.s version 20051229 -# D. J. Bernstein -# Public domain. - -#include <linux/linkage.h> - -.text - -# enter salsa20_encrypt_bytes -ENTRY(salsa20_encrypt_bytes) - mov %esp,%eax - and $31,%eax - add $256,%eax - sub %eax,%esp - # eax_stack = eax - movl %eax,80(%esp) - # ebx_stack = ebx - movl %ebx,84(%esp) - # esi_stack = esi - movl %esi,88(%esp) - # edi_stack = edi - movl %edi,92(%esp) - # ebp_stack = ebp - movl %ebp,96(%esp) - # x = arg1 - movl 4(%esp,%eax),%edx - # m = arg2 - movl 8(%esp,%eax),%esi - # out = arg3 - movl 12(%esp,%eax),%edi - # bytes = arg4 - movl 16(%esp,%eax),%ebx - # bytes -= 0 - sub $0,%ebx - # goto done if unsigned<= - jbe ._done -._start: - # in0 = *(uint32 *) (x + 0) - movl 0(%edx),%eax - # in1 = *(uint32 *) (x + 4) - movl 4(%edx),%ecx - # in2 = *(uint32 *) (x + 8) - movl 8(%edx),%ebp - # j0 = in0 - movl %eax,164(%esp) - # in3 = *(uint32 *) (x + 12) - movl 12(%edx),%eax - # j1 = in1 - movl %ecx,168(%esp) - # in4 = *(uint32 *) (x + 16) - movl 16(%edx),%ecx - # j2 = in2 - movl %ebp,172(%esp) - # in5 = *(uint32 *) (x + 20) - movl 20(%edx),%ebp - # j3 = in3 - movl %eax,176(%esp) - # in6 = *(uint32 *) (x + 24) - movl 24(%edx),%eax - # j4 = in4 - movl %ecx,180(%esp) - # in7 = *(uint32 *) (x + 28) - movl 28(%edx),%ecx - # j5 = in5 - movl %ebp,184(%esp) - # in8 = *(uint32 *) (x + 32) - movl 32(%edx),%ebp - # j6 = in6 - movl %eax,188(%esp) - # in9 = *(uint32 *) (x + 36) - movl 36(%edx),%eax - # j7 = in7 - movl %ecx,192(%esp) - # in10 = *(uint32 *) (x + 40) - movl 40(%edx),%ecx - # j8 = in8 - movl %ebp,196(%esp) - # in11 = *(uint32 *) (x + 44) - movl 44(%edx),%ebp - # j9 = in9 - movl %eax,200(%esp) - # in12 = *(uint32 *) (x + 48) - movl 48(%edx),%eax - # j10 = in10 - movl %ecx,204(%esp) - # in13 = *(uint32 *) (x + 52) - movl 52(%edx),%ecx - # j11 = in11 - movl %ebp,208(%esp) - # in14 = *(uint32 *) (x + 56) - movl 56(%edx),%ebp - # j12 = in12 - movl %eax,212(%esp) - # in15 = *(uint32 *) (x + 60) - movl 60(%edx),%eax - # j13 = in13 - movl %ecx,216(%esp) - # j14 = in14 - movl %ebp,220(%esp) - # j15 = in15 - movl %eax,224(%esp) - # x_backup = x - movl %edx,64(%esp) -._bytesatleast1: - # bytes - 64 - cmp $64,%ebx - # goto nocopy if unsigned>= - jae ._nocopy - # ctarget = out - movl %edi,228(%esp) - # out = &tmp - leal 0(%esp),%edi - # i = bytes - mov %ebx,%ecx - # while (i) { *out++ = *m++; --i } - rep movsb - # out = &tmp - leal 0(%esp),%edi - # m = &tmp - leal 0(%esp),%esi -._nocopy: - # out_backup = out - movl %edi,72(%esp) - # m_backup = m - movl %esi,68(%esp) - # bytes_backup = bytes - movl %ebx,76(%esp) - # in0 = j0 - movl 164(%esp),%eax - # in1 = j1 - movl 168(%esp),%ecx - # in2 = j2 - movl 172(%esp),%edx - # in3 = j3 - movl 176(%esp),%ebx - # x0 = in0 - movl %eax,100(%esp) - # x1 = in1 - movl %ecx,104(%esp) - # x2 = in2 - movl %edx,108(%esp) - # x3 = in3 - movl %ebx,112(%esp) - # in4 = j4 - movl 180(%esp),%eax - # in5 = j5 - movl 184(%esp),%ecx - # in6 = j6 - movl 188(%esp),%edx - # in7 = j7 - movl 192(%esp),%ebx - # x4 = in4 - movl %eax,116(%esp) - # x5 = in5 - movl %ecx,120(%esp) - # x6 = in6 - movl %edx,124(%esp) - # x7 = in7 - movl %ebx,128(%esp) - # in8 = j8 - movl 196(%esp),%eax - # in9 = j9 - movl 200(%esp),%ecx - # in10 = j10 - movl 204(%esp),%edx - # in11 = j11 - movl 208(%esp),%ebx - # x8 = in8 - movl %eax,132(%esp) - # x9 = in9 - movl %ecx,136(%esp) - # x10 = in10 - movl %edx,140(%esp) - # x11 = in11 - movl %ebx,144(%esp) - # in12 = j12 - movl 212(%esp),%eax - # in13 = j13 - movl 216(%esp),%ecx - # in14 = j14 - movl 220(%esp),%edx - # in15 = j15 - movl 224(%esp),%ebx - # x12 = in12 - movl %eax,148(%esp) - # x13 = in13 - movl %ecx,152(%esp) - # x14 = in14 - movl %edx,156(%esp) - # x15 = in15 - movl %ebx,160(%esp) - # i = 20 - mov $20,%ebp - # p = x0 - movl 100(%esp),%eax - # s = x5 - movl 120(%esp),%ecx - # t = x10 - movl 140(%esp),%edx - # w = x15 - movl 160(%esp),%ebx -._mainloop: - # x0 = p - movl %eax,100(%esp) - # x10 = t - movl %edx,140(%esp) - # p += x12 - addl 148(%esp),%eax - # x5 = s - movl %ecx,120(%esp) - # t += x6 - addl 124(%esp),%edx - # x15 = w - movl %ebx,160(%esp) - # r = x1 - movl 104(%esp),%esi - # r += s - add %ecx,%esi - # v = x11 - movl 144(%esp),%edi - # v += w - add %ebx,%edi - # p <<<= 7 - rol $7,%eax - # p ^= x4 - xorl 116(%esp),%eax - # t <<<= 7 - rol $7,%edx - # t ^= x14 - xorl 156(%esp),%edx - # r <<<= 7 - rol $7,%esi - # r ^= x9 - xorl 136(%esp),%esi - # v <<<= 7 - rol $7,%edi - # v ^= x3 - xorl 112(%esp),%edi - # x4 = p - movl %eax,116(%esp) - # x14 = t - movl %edx,156(%esp) - # p += x0 - addl 100(%esp),%eax - # x9 = r - movl %esi,136(%esp) - # t += x10 - addl 140(%esp),%edx - # x3 = v - movl %edi,112(%esp) - # p <<<= 9 - rol $9,%eax - # p ^= x8 - xorl 132(%esp),%eax - # t <<<= 9 - rol $9,%edx - # t ^= x2 - xorl 108(%esp),%edx - # s += r - add %esi,%ecx - # s <<<= 9 - rol $9,%ecx - # s ^= x13 - xorl 152(%esp),%ecx - # w += v - add %edi,%ebx - # w <<<= 9 - rol $9,%ebx - # w ^= x7 - xorl 128(%esp),%ebx - # x8 = p - movl %eax,132(%esp) - # x2 = t - movl %edx,108(%esp) - # p += x4 - addl 116(%esp),%eax - # x13 = s - movl %ecx,152(%esp) - # t += x14 - addl 156(%esp),%edx - # x7 = w - movl %ebx,128(%esp) - # p <<<= 13 - rol $13,%eax - # p ^= x12 - xorl 148(%esp),%eax - # t <<<= 13 - rol $13,%edx - # t ^= x6 - xorl 124(%esp),%edx - # r += s - add %ecx,%esi - # r <<<= 13 - rol $13,%esi - # r ^= x1 - xorl 104(%esp),%esi - # v += w - add %ebx,%edi - # v <<<= 13 - rol $13,%edi - # v ^= x11 - xorl 144(%esp),%edi - # x12 = p - movl %eax,148(%esp) - # x6 = t - movl %edx,124(%esp) - # p += x8 - addl 132(%esp),%eax - # x1 = r - movl %esi,104(%esp) - # t += x2 - addl 108(%esp),%edx - # x11 = v - movl %edi,144(%esp) - # p <<<= 18 - rol $18,%eax - # p ^= x0 - xorl 100(%esp),%eax - # t <<<= 18 - rol $18,%edx - # t ^= x10 - xorl 140(%esp),%edx - # s += r - add %esi,%ecx - # s <<<= 18 - rol $18,%ecx - # s ^= x5 - xorl 120(%esp),%ecx - # w += v - add %edi,%ebx - # w <<<= 18 - rol $18,%ebx - # w ^= x15 - xorl 160(%esp),%ebx - # x0 = p - movl %eax,100(%esp) - # x10 = t - movl %edx,140(%esp) - # p += x3 - addl 112(%esp),%eax - # p <<<= 7 - rol $7,%eax - # x5 = s - movl %ecx,120(%esp) - # t += x9 - addl 136(%esp),%edx - # x15 = w - movl %ebx,160(%esp) - # r = x4 - movl 116(%esp),%esi - # r += s - add %ecx,%esi - # v = x14 - movl 156(%esp),%edi - # v += w - add %ebx,%edi - # p ^= x1 - xorl 104(%esp),%eax - # t <<<= 7 - rol $7,%edx - # t ^= x11 - xorl 144(%esp),%edx - # r <<<= 7 - rol $7,%esi - # r ^= x6 - xorl 124(%esp),%esi - # v <<<= 7 - rol $7,%edi - # v ^= x12 - xorl 148(%esp),%edi - # x1 = p - movl %eax,104(%esp) - # x11 = t - movl %edx,144(%esp) - # p += x0 - addl 100(%esp),%eax - # x6 = r - movl %esi,124(%esp) - # t += x10 - addl 140(%esp),%edx - # x12 = v - movl %edi,148(%esp) - # p <<<= 9 - rol $9,%eax - # p ^= x2 - xorl 108(%esp),%eax - # t <<<= 9 - rol $9,%edx - # t ^= x8 - xorl 132(%esp),%edx - # s += r - add %esi,%ecx - # s <<<= 9 - rol $9,%ecx - # s ^= x7 - xorl 128(%esp),%ecx - # w += v - add %edi,%ebx - # w <<<= 9 - rol $9,%ebx - # w ^= x13 - xorl 152(%esp),%ebx - # x2 = p - movl %eax,108(%esp) - # x8 = t - movl %edx,132(%esp) - # p += x1 - addl 104(%esp),%eax - # x7 = s - movl %ecx,128(%esp) - # t += x11 - addl 144(%esp),%edx - # x13 = w - movl %ebx,152(%esp) - # p <<<= 13 - rol $13,%eax - # p ^= x3 - xorl 112(%esp),%eax - # t <<<= 13 - rol $13,%edx - # t ^= x9 - xorl 136(%esp),%edx - # r += s - add %ecx,%esi - # r <<<= 13 - rol $13,%esi - # r ^= x4 - xorl 116(%esp),%esi - # v += w - add %ebx,%edi - # v <<<= 13 - rol $13,%edi - # v ^= x14 - xorl 156(%esp),%edi - # x3 = p - movl %eax,112(%esp) - # x9 = t - movl %edx,136(%esp) - # p += x2 - addl 108(%esp),%eax - # x4 = r - movl %esi,116(%esp) - # t += x8 - addl 132(%esp),%edx - # x14 = v - movl %edi,156(%esp) - # p <<<= 18 - rol $18,%eax - # p ^= x0 - xorl 100(%esp),%eax - # t <<<= 18 - rol $18,%edx - # t ^= x10 - xorl 140(%esp),%edx - # s += r - add %esi,%ecx - # s <<<= 18 - rol $18,%ecx - # s ^= x5 - xorl 120(%esp),%ecx - # w += v - add %edi,%ebx - # w <<<= 18 - rol $18,%ebx - # w ^= x15 - xorl 160(%esp),%ebx - # x0 = p - movl %eax,100(%esp) - # x10 = t - movl %edx,140(%esp) - # p += x12 - addl 148(%esp),%eax - # x5 = s - movl %ecx,120(%esp) - # t += x6 - addl 124(%esp),%edx - # x15 = w - movl %ebx,160(%esp) - # r = x1 - movl 104(%esp),%esi - # r += s - add %ecx,%esi - # v = x11 - movl 144(%esp),%edi - # v += w - add %ebx,%edi - # p <<<= 7 - rol $7,%eax - # p ^= x4 - xorl 116(%esp),%eax - # t <<<= 7 - rol $7,%edx - # t ^= x14 - xorl 156(%esp),%edx - # r <<<= 7 - rol $7,%esi - # r ^= x9 - xorl 136(%esp),%esi - # v <<<= 7 - rol $7,%edi - # v ^= x3 - xorl 112(%esp),%edi - # x4 = p - movl %eax,116(%esp) - # x14 = t - movl %edx,156(%esp) - # p += x0 - addl 100(%esp),%eax - # x9 = r - movl %esi,136(%esp) - # t += x10 - addl 140(%esp),%edx - # x3 = v - movl %edi,112(%esp) - # p <<<= 9 - rol $9,%eax - # p ^= x8 - xorl 132(%esp),%eax - # t <<<= 9 - rol $9,%edx - # t ^= x2 - xorl 108(%esp),%edx - # s += r - add %esi,%ecx - # s <<<= 9 - rol $9,%ecx - # s ^= x13 - xorl 152(%esp),%ecx - # w += v - add %edi,%ebx - # w <<<= 9 - rol $9,%ebx - # w ^= x7 - xorl 128(%esp),%ebx - # x8 = p - movl %eax,132(%esp) - # x2 = t - movl %edx,108(%esp) - # p += x4 - addl 116(%esp),%eax - # x13 = s - movl %ecx,152(%esp) - # t += x14 - addl 156(%esp),%edx - # x7 = w - movl %ebx,128(%esp) - # p <<<= 13 - rol $13,%eax - # p ^= x12 - xorl 148(%esp),%eax - # t <<<= 13 - rol $13,%edx - # t ^= x6 - xorl 124(%esp),%edx - # r += s - add %ecx,%esi - # r <<<= 13 - rol $13,%esi - # r ^= x1 - xorl 104(%esp),%esi - # v += w - add %ebx,%edi - # v <<<= 13 - rol $13,%edi - # v ^= x11 - xorl 144(%esp),%edi - # x12 = p - movl %eax,148(%esp) - # x6 = t - movl %edx,124(%esp) - # p += x8 - addl 132(%esp),%eax - # x1 = r - movl %esi,104(%esp) - # t += x2 - addl 108(%esp),%edx - # x11 = v - movl %edi,144(%esp) - # p <<<= 18 - rol $18,%eax - # p ^= x0 - xorl 100(%esp),%eax - # t <<<= 18 - rol $18,%edx - # t ^= x10 - xorl 140(%esp),%edx - # s += r - add %esi,%ecx - # s <<<= 18 - rol $18,%ecx - # s ^= x5 - xorl 120(%esp),%ecx - # w += v - add %edi,%ebx - # w <<<= 18 - rol $18,%ebx - # w ^= x15 - xorl 160(%esp),%ebx - # x0 = p - movl %eax,100(%esp) - # x10 = t - movl %edx,140(%esp) - # p += x3 - addl 112(%esp),%eax - # p <<<= 7 - rol $7,%eax - # x5 = s - movl %ecx,120(%esp) - # t += x9 - addl 136(%esp),%edx - # x15 = w - movl %ebx,160(%esp) - # r = x4 - movl 116(%esp),%esi - # r += s - add %ecx,%esi - # v = x14 - movl 156(%esp),%edi - # v += w - add %ebx,%edi - # p ^= x1 - xorl 104(%esp),%eax - # t <<<= 7 - rol $7,%edx - # t ^= x11 - xorl 144(%esp),%edx - # r <<<= 7 - rol $7,%esi - # r ^= x6 - xorl 124(%esp),%esi - # v <<<= 7 - rol $7,%edi - # v ^= x12 - xorl 148(%esp),%edi - # x1 = p - movl %eax,104(%esp) - # x11 = t - movl %edx,144(%esp) - # p += x0 - addl 100(%esp),%eax - # x6 = r - movl %esi,124(%esp) - # t += x10 - addl 140(%esp),%edx - # x12 = v - movl %edi,148(%esp) - # p <<<= 9 - rol $9,%eax - # p ^= x2 - xorl 108(%esp),%eax - # t <<<= 9 - rol $9,%edx - # t ^= x8 - xorl 132(%esp),%edx - # s += r - add %esi,%ecx - # s <<<= 9 - rol $9,%ecx - # s ^= x7 - xorl 128(%esp),%ecx - # w += v - add %edi,%ebx - # w <<<= 9 - rol $9,%ebx - # w ^= x13 - xorl 152(%esp),%ebx - # x2 = p - movl %eax,108(%esp) - # x8 = t - movl %edx,132(%esp) - # p += x1 - addl 104(%esp),%eax - # x7 = s - movl %ecx,128(%esp) - # t += x11 - addl 144(%esp),%edx - # x13 = w - movl %ebx,152(%esp) - # p <<<= 13 - rol $13,%eax - # p ^= x3 - xorl 112(%esp),%eax - # t <<<= 13 - rol $13,%edx - # t ^= x9 - xorl 136(%esp),%edx - # r += s - add %ecx,%esi - # r <<<= 13 - rol $13,%esi - # r ^= x4 - xorl 116(%esp),%esi - # v += w - add %ebx,%edi - # v <<<= 13 - rol $13,%edi - # v ^= x14 - xorl 156(%esp),%edi - # x3 = p - movl %eax,112(%esp) - # x9 = t - movl %edx,136(%esp) - # p += x2 - addl 108(%esp),%eax - # x4 = r - movl %esi,116(%esp) - # t += x8 - addl 132(%esp),%edx - # x14 = v - movl %edi,156(%esp) - # p <<<= 18 - rol $18,%eax - # p ^= x0 - xorl 100(%esp),%eax - # t <<<= 18 - rol $18,%edx - # t ^= x10 - xorl 140(%esp),%edx - # s += r - add %esi,%ecx - # s <<<= 18 - rol $18,%ecx - # s ^= x5 - xorl 120(%esp),%ecx - # w += v - add %edi,%ebx - # w <<<= 18 - rol $18,%ebx - # w ^= x15 - xorl 160(%esp),%ebx - # i -= 4 - sub $4,%ebp - # goto mainloop if unsigned > - ja ._mainloop - # x0 = p - movl %eax,100(%esp) - # x5 = s - movl %ecx,120(%esp) - # x10 = t - movl %edx,140(%esp) - # x15 = w - movl %ebx,160(%esp) - # out = out_backup - movl 72(%esp),%edi - # m = m_backup - movl 68(%esp),%esi - # in0 = x0 - movl 100(%esp),%eax - # in1 = x1 - movl 104(%esp),%ecx - # in0 += j0 - addl 164(%esp),%eax - # in1 += j1 - addl 168(%esp),%ecx - # in0 ^= *(uint32 *) (m + 0) - xorl 0(%esi),%eax - # in1 ^= *(uint32 *) (m + 4) - xorl 4(%esi),%ecx - # *(uint32 *) (out + 0) = in0 - movl %eax,0(%edi) - # *(uint32 *) (out + 4) = in1 - movl %ecx,4(%edi) - # in2 = x2 - movl 108(%esp),%eax - # in3 = x3 - movl 112(%esp),%ecx - # in2 += j2 - addl 172(%esp),%eax - # in3 += j3 - addl 176(%esp),%ecx - # in2 ^= *(uint32 *) (m + 8) - xorl 8(%esi),%eax - # in3 ^= *(uint32 *) (m + 12) - xorl 12(%esi),%ecx - # *(uint32 *) (out + 8) = in2 - movl %eax,8(%edi) - # *(uint32 *) (out + 12) = in3 - movl %ecx,12(%edi) - # in4 = x4 - movl 116(%esp),%eax - # in5 = x5 - movl 120(%esp),%ecx - # in4 += j4 - addl 180(%esp),%eax - # in5 += j5 - addl 184(%esp),%ecx - # in4 ^= *(uint32 *) (m + 16) - xorl 16(%esi),%eax - # in5 ^= *(uint32 *) (m + 20) - xorl 20(%esi),%ecx - # *(uint32 *) (out + 16) = in4 - movl %eax,16(%edi) - # *(uint32 *) (out + 20) = in5 - movl %ecx,20(%edi) - # in6 = x6 - movl 124(%esp),%eax - # in7 = x7 - movl 128(%esp),%ecx - # in6 += j6 - addl 188(%esp),%eax - # in7 += j7 - addl 192(%esp),%ecx - # in6 ^= *(uint32 *) (m + 24) - xorl 24(%esi),%eax - # in7 ^= *(uint32 *) (m + 28) - xorl 28(%esi),%ecx - # *(uint32 *) (out + 24) = in6 - movl %eax,24(%edi) - # *(uint32 *) (out + 28) = in7 - movl %ecx,28(%edi) - # in8 = x8 - movl 132(%esp),%eax - # in9 = x9 - movl 136(%esp),%ecx - # in8 += j8 - addl 196(%esp),%eax - # in9 += j9 - addl 200(%esp),%ecx - # in8 ^= *(uint32 *) (m + 32) - xorl 32(%esi),%eax - # in9 ^= *(uint32 *) (m + 36) - xorl 36(%esi),%ecx - # *(uint32 *) (out + 32) = in8 - movl %eax,32(%edi) - # *(uint32 *) (out + 36) = in9 - movl %ecx,36(%edi) - # in10 = x10 - movl 140(%esp),%eax - # in11 = x11 - movl 144(%esp),%ecx - # in10 += j10 - addl 204(%esp),%eax - # in11 += j11 - addl 208(%esp),%ecx - # in10 ^= *(uint32 *) (m + 40) - xorl 40(%esi),%eax - # in11 ^= *(uint32 *) (m + 44) - xorl 44(%esi),%ecx - # *(uint32 *) (out + 40) = in10 - movl %eax,40(%edi) - # *(uint32 *) (out + 44) = in11 - movl %ecx,44(%edi) - # in12 = x12 - movl 148(%esp),%eax - # in13 = x13 - movl 152(%esp),%ecx - # in12 += j12 - addl 212(%esp),%eax - # in13 += j13 - addl 216(%esp),%ecx - # in12 ^= *(uint32 *) (m + 48) - xorl 48(%esi),%eax - # in13 ^= *(uint32 *) (m + 52) - xorl 52(%esi),%ecx - # *(uint32 *) (out + 48) = in12 - movl %eax,48(%edi) - # *(uint32 *) (out + 52) = in13 - movl %ecx,52(%edi) - # in14 = x14 - movl 156(%esp),%eax - # in15 = x15 - movl 160(%esp),%ecx - # in14 += j14 - addl 220(%esp),%eax - # in15 += j15 - addl 224(%esp),%ecx - # in14 ^= *(uint32 *) (m + 56) - xorl 56(%esi),%eax - # in15 ^= *(uint32 *) (m + 60) - xorl 60(%esi),%ecx - # *(uint32 *) (out + 56) = in14 - movl %eax,56(%edi) - # *(uint32 *) (out + 60) = in15 - movl %ecx,60(%edi) - # bytes = bytes_backup - movl 76(%esp),%ebx - # in8 = j8 - movl 196(%esp),%eax - # in9 = j9 - movl 200(%esp),%ecx - # in8 += 1 - add $1,%eax - # in9 += 0 + carry - adc $0,%ecx - # j8 = in8 - movl %eax,196(%esp) - # j9 = in9 - movl %ecx,200(%esp) - # bytes - 64 - cmp $64,%ebx - # goto bytesatleast65 if unsigned> - ja ._bytesatleast65 - # goto bytesatleast64 if unsigned>= - jae ._bytesatleast64 - # m = out - mov %edi,%esi - # out = ctarget - movl 228(%esp),%edi - # i = bytes - mov %ebx,%ecx - # while (i) { *out++ = *m++; --i } - rep movsb -._bytesatleast64: - # x = x_backup - movl 64(%esp),%eax - # in8 = j8 - movl 196(%esp),%ecx - # in9 = j9 - movl 200(%esp),%edx - # *(uint32 *) (x + 32) = in8 - movl %ecx,32(%eax) - # *(uint32 *) (x + 36) = in9 - movl %edx,36(%eax) -._done: - # eax = eax_stack - movl 80(%esp),%eax - # ebx = ebx_stack - movl 84(%esp),%ebx - # esi = esi_stack - movl 88(%esp),%esi - # edi = edi_stack - movl 92(%esp),%edi - # ebp = ebp_stack - movl 96(%esp),%ebp - # leave - add %eax,%esp - ret -._bytesatleast65: - # bytes -= 64 - sub $64,%ebx - # out += 64 - add $64,%edi - # m += 64 - add $64,%esi - # goto bytesatleast1 - jmp ._bytesatleast1 -ENDPROC(salsa20_encrypt_bytes) diff --git a/arch/x86/crypto/salsa20-x86_64-asm_64.S b/arch/x86/crypto/salsa20-x86_64-asm_64.S deleted file mode 100644 index 03a4918f41ee..000000000000 --- a/arch/x86/crypto/salsa20-x86_64-asm_64.S +++ /dev/null @@ -1,805 +0,0 @@ -/* SPDX-License-Identifier: GPL-2.0 */ -#include <linux/linkage.h> - -# enter salsa20_encrypt_bytes -ENTRY(salsa20_encrypt_bytes) - mov %rsp,%r11 - and $31,%r11 - add $256,%r11 - sub %r11,%rsp - # x = arg1 - mov %rdi,%r8 - # m = arg2 - mov %rsi,%rsi - # out = arg3 - mov %rdx,%rdi - # bytes = arg4 - mov %rcx,%rdx - # unsigned>? bytes - 0 - cmp $0,%rdx - # comment:fp stack unchanged by jump - # goto done if !unsigned> - jbe ._done - # comment:fp stack unchanged by fallthrough -# start: -._start: - # r11_stack = r11 - movq %r11,0(%rsp) - # r12_stack = r12 - movq %r12,8(%rsp) - # r13_stack = r13 - movq %r13,16(%rsp) - # r14_stack = r14 - movq %r14,24(%rsp) - # r15_stack = r15 - movq %r15,32(%rsp) - # rbx_stack = rbx - movq %rbx,40(%rsp) - # rbp_stack = rbp - movq %rbp,48(%rsp) - # in0 = *(uint64 *) (x + 0) - movq 0(%r8),%rcx - # in2 = *(uint64 *) (x + 8) - movq 8(%r8),%r9 - # in4 = *(uint64 *) (x + 16) - movq 16(%r8),%rax - # in6 = *(uint64 *) (x + 24) - movq 24(%r8),%r10 - # in8 = *(uint64 *) (x + 32) - movq 32(%r8),%r11 - # in10 = *(uint64 *) (x + 40) - movq 40(%r8),%r12 - # in12 = *(uint64 *) (x + 48) - movq 48(%r8),%r13 - # in14 = *(uint64 *) (x + 56) - movq 56(%r8),%r14 - # j0 = in0 - movq %rcx,56(%rsp) - # j2 = in2 - movq %r9,64(%rsp) - # j4 = in4 - movq %rax,72(%rsp) - # j6 = in6 - movq %r10,80(%rsp) - # j8 = in8 - movq %r11,88(%rsp) - # j10 = in10 - movq %r12,96(%rsp) - # j12 = in12 - movq %r13,104(%rsp) - # j14 = in14 - movq %r14,112(%rsp) - # x_backup = x - movq %r8,120(%rsp) -# bytesatleast1: -._bytesatleast1: - # unsigned<? bytes - 64 - cmp $64,%rdx - # comment:fp stack unchanged by jump - # goto nocopy if !unsigned< - jae ._nocopy - # ctarget = out - movq %rdi,128(%rsp) - # out = &tmp - leaq 192(%rsp),%rdi - # i = bytes - mov %rdx,%rcx - # while (i) { *out++ = *m++; --i } - rep movsb - # out = &tmp - leaq 192(%rsp),%rdi - # m = &tmp - leaq 192(%rsp),%rsi - # comment:fp stack unchanged by fallthrough -# nocopy: -._nocopy: - # out_backup = out - movq %rdi,136(%rsp) - # m_backup = m - movq %rsi,144(%rsp) - # bytes_backup = bytes - movq %rdx,152(%rsp) - # x1 = j0 - movq 56(%rsp),%rdi - # x0 = x1 - mov %rdi,%rdx - # (uint64) x1 >>= 32 - shr $32,%rdi - # x3 = j2 - movq 64(%rsp),%rsi - # x2 = x3 - mov %rsi,%rcx - # (uint64) x3 >>= 32 - shr $32,%rsi - # x5 = j4 - movq 72(%rsp),%r8 - # x4 = x5 - mov %r8,%r9 - # (uint64) x5 >>= 32 - shr $32,%r8 - # x5_stack = x5 - movq %r8,160(%rsp) - # x7 = j6 - movq 80(%rsp),%r8 - # x6 = x7 - mov %r8,%rax - # (uint64) x7 >>= 32 - shr $32,%r8 - # x9 = j8 - movq 88(%rsp),%r10 - # x8 = x9 - mov %r10,%r11 - # (uint64) x9 >>= 32 - shr $32,%r10 - # x11 = j10 - movq 96(%rsp),%r12 - # x10 = x11 - mov %r12,%r13 - # x10_stack = x10 - movq %r13,168(%rsp) - # (uint64) x11 >>= 32 - shr $32,%r12 - # x13 = j12 - movq 104(%rsp),%r13 - # x12 = x13 - mov %r13,%r14 - # (uint64) x13 >>= 32 - shr $32,%r13 - # x15 = j14 - movq 112(%rsp),%r15 - # x14 = x15 - mov %r15,%rbx - # (uint64) x15 >>= 32 - shr $32,%r15 - # x15_stack = x15 - movq %r15,176(%rsp) - # i = 20 - mov $20,%r15 -# mainloop: -._mainloop: - # i_backup = i - movq %r15,184(%rsp) - # x5 = x5_stack - movq 160(%rsp),%r15 - # a = x12 + x0 - lea (%r14,%rdx),%rbp - # (uint32) a <<<= 7 - rol $7,%ebp - # x4 ^= a - xor %rbp,%r9 - # b = x1 + x5 - lea (%rdi,%r15),%rbp - # (uint32) b <<<= 7 - rol $7,%ebp - # x9 ^= b - xor %rbp,%r10 - # a = x0 + x4 - lea (%rdx,%r9),%rbp - # (uint32) a <<<= 9 - rol $9,%ebp - # x8 ^= a - xor %rbp,%r11 - # b = x5 + x9 - lea (%r15,%r10),%rbp - # (uint32) b <<<= 9 - rol $9,%ebp - # x13 ^= b - xor %rbp,%r13 - # a = x4 + x8 - lea (%r9,%r11),%rbp - # (uint32) a <<<= 13 - rol $13,%ebp - # x12 ^= a - xor %rbp,%r14 - # b = x9 + x13 - lea (%r10,%r13),%rbp - # (uint32) b <<<= 13 - rol $13,%ebp - # x1 ^= b - xor %rbp,%rdi - # a = x8 + x12 - lea (%r11,%r14),%rbp - # (uint32) a <<<= 18 - rol $18,%ebp - # x0 ^= a - xor %rbp,%rdx - # b = x13 + x1 - lea (%r13,%rdi),%rbp - # (uint32) b <<<= 18 - rol $18,%ebp - # x5 ^= b - xor %rbp,%r15 - # x10 = x10_stack - movq 168(%rsp),%rbp - # x5_stack = x5 - movq %r15,160(%rsp) - # c = x6 + x10 - lea (%rax,%rbp),%r15 - # (uint32) c <<<= 7 - rol $7,%r15d - # x14 ^= c - xor %r15,%rbx - # c = x10 + x14 - lea (%rbp,%rbx),%r15 - # (uint32) c <<<= 9 - rol $9,%r15d - # x2 ^= c - xor %r15,%rcx - # c = x14 + x2 - lea (%rbx,%rcx),%r15 - # (uint32) c <<<= 13 - rol $13,%r15d - # x6 ^= c - xor %r15,%rax - # c = x2 + x6 - lea (%rcx,%rax),%r15 - # (uint32) c <<<= 18 - rol $18,%r15d - # x10 ^= c - xor %r15,%rbp - # x15 = x15_stack - movq 176(%rsp),%r15 - # x10_stack = x10 - movq %rbp,168(%rsp) - # d = x11 + x15 - lea (%r12,%r15),%rbp - # (uint32) d <<<= 7 - rol $7,%ebp - # x3 ^= d - xor %rbp,%rsi - # d = x15 + x3 - lea (%r15,%rsi),%rbp - # (uint32) d <<<= 9 - rol $9,%ebp - # x7 ^= d - xor %rbp,%r8 - # d = x3 + x7 - lea (%rsi,%r8),%rbp - # (uint32) d <<<= 13 - rol $13,%ebp - # x11 ^= d - xor %rbp,%r12 - # d = x7 + x11 - lea (%r8,%r12),%rbp - # (uint32) d <<<= 18 - rol $18,%ebp - # x15 ^= d - xor %rbp,%r15 - # x15_stack = x15 - movq %r15,176(%rsp) - # x5 = x5_stack - movq 160(%rsp),%r15 - # a = x3 + x0 - lea (%rsi,%rdx),%rbp - # (uint32) a <<<= 7 - rol $7,%ebp - # x1 ^= a - xor %rbp,%rdi - # b = x4 + x5 - lea (%r9,%r15),%rbp - # (uint32) b <<<= 7 - rol $7,%ebp - # x6 ^= b - xor %rbp,%rax - # a = x0 + x1 - lea (%rdx,%rdi),%rbp - # (uint32) a <<<= 9 - rol $9,%ebp - # x2 ^= a - xor %rbp,%rcx - # b = x5 + x6 - lea (%r15,%rax),%rbp - # (uint32) b <<<= 9 - rol $9,%ebp - # x7 ^= b - xor %rbp,%r8 - # a = x1 + x2 - lea (%rdi,%rcx),%rbp - # (uint32) a <<<= 13 - rol $13,%ebp - # x3 ^= a - xor %rbp,%rsi - # b = x6 + x7 - lea (%rax,%r8),%rbp - # (uint32) b <<<= 13 - rol $13,%ebp - # x4 ^= b - xor %rbp,%r9 - # a = x2 + x3 - lea (%rcx,%rsi),%rbp - # (uint32) a <<<= 18 - rol $18,%ebp - # x0 ^= a - xor %rbp,%rdx - # b = x7 + x4 - lea (%r8,%r9),%rbp - # (uint32) b <<<= 18 - rol $18,%ebp - # x5 ^= b - xor %rbp,%r15 - # x10 = x10_stack - movq 168(%rsp),%rbp - # x5_stack = x5 - movq %r15,160(%rsp) - # c = x9 + x10 - lea (%r10,%rbp),%r15 - # (uint32) c <<<= 7 - rol $7,%r15d - # x11 ^= c - xor %r15,%r12 - # c = x10 + x11 - lea (%rbp,%r12),%r15 - # (uint32) c <<<= 9 - rol $9,%r15d - # x8 ^= c - xor %r15,%r11 - # c = x11 + x8 - lea (%r12,%r11),%r15 - # (uint32) c <<<= 13 - rol $13,%r15d - # x9 ^= c - xor %r15,%r10 - # c = x8 + x9 - lea (%r11,%r10),%r15 - # (uint32) c <<<= 18 - rol $18,%r15d - # x10 ^= c - xor %r15,%rbp - # x15 = x15_stack - movq 176(%rsp),%r15 - # x10_stack = x10 - movq %rbp,168(%rsp) - # d = x14 + x15 - lea (%rbx,%r15),%rbp - # (uint32) d <<<= 7 - rol $7,%ebp - # x12 ^= d - xor %rbp,%r14 - # d = x15 + x12 - lea (%r15,%r14),%rbp - # (uint32) d <<<= 9 - rol $9,%ebp - # x13 ^= d - xor %rbp,%r13 - # d = x12 + x13 - lea (%r14,%r13),%rbp - # (uint32) d <<<= 13 - rol $13,%ebp - # x14 ^= d - xor %rbp,%rbx - # d = x13 + x14 - lea (%r13,%rbx),%rbp - # (uint32) d <<<= 18 - rol $18,%ebp - # x15 ^= d - xor %rbp,%r15 - # x15_stack = x15 - movq %r15,176(%rsp) - # x5 = x5_stack - movq 160(%rsp),%r15 - # a = x12 + x0 - lea (%r14,%rdx),%rbp - # (uint32) a <<<= 7 - rol $7,%ebp - # x4 ^= a - xor %rbp,%r9 - # b = x1 + x5 - lea (%rdi,%r15),%rbp - # (uint32) b <<<= 7 - rol $7,%ebp - # x9 ^= b - xor %rbp,%r10 - # a = x0 + x4 - lea (%rdx,%r9),%rbp - # (uint32) a <<<= 9 - rol $9,%ebp - # x8 ^= a - xor %rbp,%r11 - # b = x5 + x9 - lea (%r15,%r10),%rbp - # (uint32) b <<<= 9 - rol $9,%ebp - # x13 ^= b - xor %rbp,%r13 - # a = x4 + x8 - lea (%r9,%r11),%rbp - # (uint32) a <<<= 13 - rol $13,%ebp - # x12 ^= a - xor %rbp,%r14 - # b = x9 + x13 - lea (%r10,%r13),%rbp - # (uint32) b <<<= 13 - rol $13,%ebp - # x1 ^= b - xor %rbp,%rdi - # a = x8 + x12 - lea (%r11,%r14),%rbp - # (uint32) a <<<= 18 - rol $18,%ebp - # x0 ^= a - xor %rbp,%rdx - # b = x13 + x1 - lea (%r13,%rdi),%rbp - # (uint32) b <<<= 18 - rol $18,%ebp - # x5 ^= b - xor %rbp,%r15 - # x10 = x10_stack - movq 168(%rsp),%rbp - # x5_stack = x5 - movq %r15,160(%rsp) - # c = x6 + x10 - lea (%rax,%rbp),%r15 - # (uint32) c <<<= 7 - rol $7,%r15d - # x14 ^= c - xor %r15,%rbx - # c = x10 + x14 - lea (%rbp,%rbx),%r15 - # (uint32) c <<<= 9 - rol $9,%r15d - # x2 ^= c - xor %r15,%rcx - # c = x14 + x2 - lea (%rbx,%rcx),%r15 - # (uint32) c <<<= 13 - rol $13,%r15d - # x6 ^= c - xor %r15,%rax - # c = x2 + x6 - lea (%rcx,%rax),%r15 - # (uint32) c <<<= 18 - rol $18,%r15d - # x10 ^= c - xor %r15,%rbp - # x15 = x15_stack - movq 176(%rsp),%r15 - # x10_stack = x10 - movq %rbp,168(%rsp) - # d = x11 + x15 - lea (%r12,%r15),%rbp - # (uint32) d <<<= 7 - rol $7,%ebp - # x3 ^= d - xor %rbp,%rsi - # d = x15 + x3 - lea (%r15,%rsi),%rbp - # (uint32) d <<<= 9 - rol $9,%ebp - # x7 ^= d - xor %rbp,%r8 - # d = x3 + x7 - lea (%rsi,%r8),%rbp - # (uint32) d <<<= 13 - rol $13,%ebp - # x11 ^= d - xor %rbp,%r12 - # d = x7 + x11 - lea (%r8,%r12),%rbp - # (uint32) d <<<= 18 - rol $18,%ebp - # x15 ^= d - xor %rbp,%r15 - # x15_stack = x15 - movq %r15,176(%rsp) - # x5 = x5_stack - movq 160(%rsp),%r15 - # a = x3 + x0 - lea (%rsi,%rdx),%rbp - # (uint32) a <<<= 7 - rol $7,%ebp - # x1 ^= a - xor %rbp,%rdi - # b = x4 + x5 - lea (%r9,%r15),%rbp - # (uint32) b <<<= 7 - rol $7,%ebp - # x6 ^= b - xor %rbp,%rax - # a = x0 + x1 - lea (%rdx,%rdi),%rbp - # (uint32) a <<<= 9 - rol $9,%ebp - # x2 ^= a - xor %rbp,%rcx - # b = x5 + x6 - lea (%r15,%rax),%rbp - # (uint32) b <<<= 9 - rol $9,%ebp - # x7 ^= b - xor %rbp,%r8 - # a = x1 + x2 - lea (%rdi,%rcx),%rbp - # (uint32) a <<<= 13 - rol $13,%ebp - # x3 ^= a - xor %rbp,%rsi - # b = x6 + x7 - lea (%rax,%r8),%rbp - # (uint32) b <<<= 13 - rol $13,%ebp - # x4 ^= b - xor %rbp,%r9 - # a = x2 + x3 - lea (%rcx,%rsi),%rbp - # (uint32) a <<<= 18 - rol $18,%ebp - # x0 ^= a - xor %rbp,%rdx - # b = x7 + x4 - lea (%r8,%r9),%rbp - # (uint32) b <<<= 18 - rol $18,%ebp - # x5 ^= b - xor %rbp,%r15 - # x10 = x10_stack - movq 168(%rsp),%rbp - # x5_stack = x5 - movq %r15,160(%rsp) - # c = x9 + x10 - lea (%r10,%rbp),%r15 - # (uint32) c <<<= 7 - rol $7,%r15d - # x11 ^= c - xor %r15,%r12 - # c = x10 + x11 - lea (%rbp,%r12),%r15 - # (uint32) c <<<= 9 - rol $9,%r15d - # x8 ^= c - xor %r15,%r11 - # c = x11 + x8 - lea (%r12,%r11),%r15 - # (uint32) c <<<= 13 - rol $13,%r15d - # x9 ^= c - xor %r15,%r10 - # c = x8 + x9 - lea (%r11,%r10),%r15 - # (uint32) c <<<= 18 - rol $18,%r15d - # x10 ^= c - xor %r15,%rbp - # x15 = x15_stack - movq 176(%rsp),%r15 - # x10_stack = x10 - movq %rbp,168(%rsp) - # d = x14 + x15 - lea (%rbx,%r15),%rbp - # (uint32) d <<<= 7 - rol $7,%ebp - # x12 ^= d - xor %rbp,%r14 - # d = x15 + x12 - lea (%r15,%r14),%rbp - # (uint32) d <<<= 9 - rol $9,%ebp - # x13 ^= d - xor %rbp,%r13 - # d = x12 + x13 - lea (%r14,%r13),%rbp - # (uint32) d <<<= 13 - rol $13,%ebp - # x14 ^= d - xor %rbp,%rbx - # d = x13 + x14 - lea (%r13,%rbx),%rbp - # (uint32) d <<<= 18 - rol $18,%ebp - # x15 ^= d - xor %rbp,%r15 - # x15_stack = x15 - movq %r15,176(%rsp) - # i = i_backup - movq 184(%rsp),%r15 - # unsigned>? i -= 4 - sub $4,%r15 - # comment:fp stack unchanged by jump - # goto mainloop if unsigned> - ja ._mainloop - # (uint32) x2 += j2 - addl 64(%rsp),%ecx - # x3 <<= 32 - shl $32,%rsi - # x3 += j2 - addq 64(%rsp),%rsi - # (uint64) x3 >>= 32 - shr $32,%rsi - # x3 <<= 32 - shl $32,%rsi - # x2 += x3 - add %rsi,%rcx - # (uint32) x6 += j6 - addl 80(%rsp),%eax - # x7 <<= 32 - shl $32,%r8 - # x7 += j6 - addq 80(%rsp),%r8 - # (uint64) x7 >>= 32 - shr $32,%r8 - # x7 <<= 32 - shl $32,%r8 - # x6 += x7 - add %r8,%rax - # (uint32) x8 += j8 - addl 88(%rsp),%r11d - # x9 <<= 32 - shl $32,%r10 - # x9 += j8 - addq 88(%rsp),%r10 - # (uint64) x9 >>= 32 - shr $32,%r10 - # x9 <<= 32 - shl $32,%r10 - # x8 += x9 - add %r10,%r11 - # (uint32) x12 += j12 - addl 104(%rsp),%r14d - # x13 <<= 32 - shl $32,%r13 - # x13 += j12 - addq 104(%rsp),%r13 - # (uint64) x13 >>= 32 - shr $32,%r13 - # x13 <<= 32 - shl $32,%r13 - # x12 += x13 - add %r13,%r14 - # (uint32) x0 += j0 - addl 56(%rsp),%edx - # x1 <<= 32 - shl $32,%rdi - # x1 += j0 - addq 56(%rsp),%rdi - # (uint64) x1 >>= 32 - shr $32,%rdi - # x1 <<= 32 - shl $32,%rdi - # x0 += x1 - add %rdi,%rdx - # x5 = x5_stack - movq 160(%rsp),%rdi - # (uint32) x4 += j4 - addl 72(%rsp),%r9d - # x5 <<= 32 - shl $32,%rdi - # x5 += j4 - addq 72(%rsp),%rdi - # (uint64) x5 >>= 32 - shr $32,%rdi - # x5 <<= 32 - shl $32,%rdi - # x4 += x5 - add %rdi,%r9 - # x10 = x10_stack - movq 168(%rsp),%r8 - # (uint32) x10 += j10 - addl 96(%rsp),%r8d - # x11 <<= 32 - shl $32,%r12 - # x11 += j10 - addq 96(%rsp),%r12 - # (uint64) x11 >>= 32 - shr $32,%r12 - # x11 <<= 32 - shl $32,%r12 - # x10 += x11 - add %r12,%r8 - # x15 = x15_stack - movq 176(%rsp),%rdi - # (uint32) x14 += j14 - addl 112(%rsp),%ebx - # x15 <<= 32 - shl $32,%rdi - # x15 += j14 - addq 112(%rsp),%rdi - # (uint64) x15 >>= 32 - shr $32,%rdi - # x15 <<= 32 - shl $32,%rdi - # x14 += x15 - add %rdi,%rbx - # out = out_backup - movq 136(%rsp),%rdi - # m = m_backup - movq 144(%rsp),%rsi - # x0 ^= *(uint64 *) (m + 0) - xorq 0(%rsi),%rdx - # *(uint64 *) (out + 0) = x0 - movq %rdx,0(%rdi) - # x2 ^= *(uint64 *) (m + 8) - xorq 8(%rsi),%rcx - # *(uint64 *) (out + 8) = x2 - movq %rcx,8(%rdi) - # x4 ^= *(uint64 *) (m + 16) - xorq 16(%rsi),%r9 - # *(uint64 *) (out + 16) = x4 - movq %r9,16(%rdi) - # x6 ^= *(uint64 *) (m + 24) - xorq 24(%rsi),%rax - # *(uint64 *) (out + 24) = x6 - movq %rax,24(%rdi) - # x8 ^= *(uint64 *) (m + 32) - xorq 32(%rsi),%r11 - # *(uint64 *) (out + 32) = x8 - movq %r11,32(%rdi) - # x10 ^= *(uint64 *) (m + 40) - xorq 40(%rsi),%r8 - # *(uint64 *) (out + 40) = x10 - movq %r8,40(%rdi) - # x12 ^= *(uint64 *) (m + 48) - xorq 48(%rsi),%r14 - # *(uint64 *) (out + 48) = x12 - movq %r14,48(%rdi) - # x14 ^= *(uint64 *) (m + 56) - xorq 56(%rsi),%rbx - # *(uint64 *) (out + 56) = x14 - movq %rbx,56(%rdi) - # bytes = bytes_backup - movq 152(%rsp),%rdx - # in8 = j8 - movq 88(%rsp),%rcx - # in8 += 1 - add $1,%rcx - # j8 = in8 - movq %rcx,88(%rsp) - # unsigned>? unsigned<? bytes - 64 - cmp $64,%rdx - # comment:fp stack unchanged by jump - # goto bytesatleast65 if unsigned> - ja ._bytesatleast65 - # comment:fp stack unchanged by jump - # goto bytesatleast64 if !unsigned< - jae ._bytesatleast64 - # m = out - mov %rdi,%rsi - # out = ctarget - movq 128(%rsp),%rdi - # i = bytes - mov %rdx,%rcx - # while (i) { *out++ = *m++; --i } - rep movsb - # comment:fp stack unchanged by fallthrough -# bytesatleast64: -._bytesatleast64: - # x = x_backup - movq 120(%rsp),%rdi - # in8 = j8 - movq 88(%rsp),%rsi - # *(uint64 *) (x + 32) = in8 - movq %rsi,32(%rdi) - # r11 = r11_stack - movq 0(%rsp),%r11 - # r12 = r12_stack - movq 8(%rsp),%r12 - # r13 = r13_stack - movq 16(%rsp),%r13 - # r14 = r14_stack - movq 24(%rsp),%r14 - # r15 = r15_stack - movq 32(%rsp),%r15 - # rbx = rbx_stack - movq 40(%rsp),%rbx - # rbp = rbp_stack - movq 48(%rsp),%rbp - # comment:fp stack unchanged by fallthrough -# done: -._done: - # leave - add %r11,%rsp - mov %rdi,%rax - mov %rsi,%rdx - ret -# bytesatleast65: -._bytesatleast65: - # bytes -= 64 - sub $64,%rdx - # out += 64 - add $64,%rdi - # m += 64 - add $64,%rsi - # comment:fp stack unchanged by jump - # goto bytesatleast1 - jmp ._bytesatleast1 -ENDPROC(salsa20_encrypt_bytes) diff --git a/arch/x86/crypto/salsa20_glue.c b/arch/x86/crypto/salsa20_glue.c deleted file mode 100644 index b07d7d959806..000000000000 --- a/arch/x86/crypto/salsa20_glue.c +++ /dev/null @@ -1,91 +0,0 @@ -/* - * Glue code for optimized assembly version of Salsa20. - * - * Copyright (c) 2007 Tan Swee Heng <thesweeheng@gmail.com> - * - * The assembly codes are public domain assembly codes written by Daniel. J. - * Bernstein <djb@cr.yp.to>. The codes are modified to include indentation - * and to remove extraneous comments and functions that are not needed. - * - i586 version, renamed as salsa20-i586-asm_32.S - * available from <http://cr.yp.to/snuffle/salsa20/x86-pm/salsa20.s> - * - x86-64 version, renamed as salsa20-x86_64-asm_64.S - * available from <http://cr.yp.to/snuffle/salsa20/amd64-3/salsa20.s> - * - * Also modified to set up the initial state using the generic C code rather - * than in assembly. - * - * This program is free software; you can redistribute it and/or modify it - * under the terms of the GNU General Public License as published by the Free - * Software Foundation; either version 2 of the License, or (at your option) - * any later version. - * - */ - -#include <asm/unaligned.h> -#include <crypto/internal/skcipher.h> -#include <crypto/salsa20.h> -#include <linux/module.h> - -asmlinkage void salsa20_encrypt_bytes(u32 state[16], const u8 *src, u8 *dst, - u32 bytes); - -static int salsa20_asm_crypt(struct skcipher_request *req) -{ - struct crypto_skcipher *tfm = crypto_skcipher_reqtfm(req); - const struct salsa20_ctx *ctx = crypto_skcipher_ctx(tfm); - struct skcipher_walk walk; - u32 state[16]; - int err; - - err = skcipher_walk_virt(&walk, req, true); - - crypto_salsa20_init(state, ctx, walk.iv); - - while (walk.nbytes > 0) { - unsigned int nbytes = walk.nbytes; - - if (nbytes < walk.total) - nbytes = round_down(nbytes, walk.stride); - - salsa20_encrypt_bytes(state, walk.src.virt.addr, - walk.dst.virt.addr, nbytes); - err = skcipher_walk_done(&walk, walk.nbytes - nbytes); - } - - return err; -} - -static struct skcipher_alg alg = { - .base.cra_name = "salsa20", - .base.cra_driver_name = "salsa20-asm", - .base.cra_priority = 200, - .base.cra_blocksize = 1, - .base.cra_ctxsize = sizeof(struct salsa20_ctx), - .base.cra_module = THIS_MODULE, - - .min_keysize = SALSA20_MIN_KEY_SIZE, - .max_keysize = SALSA20_MAX_KEY_SIZE, - .ivsize = SALSA20_IV_SIZE, - .chunksize = SALSA20_BLOCK_SIZE, - .setkey = crypto_salsa20_setkey, - .encrypt = salsa20_asm_crypt, - .decrypt = salsa20_asm_crypt, -}; - -static int __init init(void) -{ - return crypto_register_skcipher(&alg); -} - -static void __exit fini(void) -{ - crypto_unregister_skcipher(&alg); -} - -module_init(init); -module_exit(fini); - -MODULE_LICENSE("GPL"); -MODULE_DESCRIPTION ("Salsa20 stream cipher algorithm (optimized assembly version)"); -MODULE_ALIAS_CRYPTO("salsa20"); -MODULE_ALIAS_CRYPTO("salsa20-asm"); diff --git a/arch/x86/include/asm/apic.h b/arch/x86/include/asm/apic.h index 08acd954f00e..74a9e06b6cfd 100644 --- a/arch/x86/include/asm/apic.h +++ b/arch/x86/include/asm/apic.h @@ -436,6 +436,8 @@ static inline void apic_set_eoi_write(void (*eoi_write)(u32 reg, u32 v)) {} #endif /* CONFIG_X86_LOCAL_APIC */ +extern void apic_ack_irq(struct irq_data *data); + static inline void ack_APIC_irq(void) { /* diff --git a/arch/x86/include/asm/cpufeatures.h b/arch/x86/include/asm/cpufeatures.h index fb00a2fca990..5701f5cecd31 100644 --- a/arch/x86/include/asm/cpufeatures.h +++ b/arch/x86/include/asm/cpufeatures.h @@ -282,7 +282,9 @@ #define X86_FEATURE_AMD_IBPB (13*32+12) /* "" Indirect Branch Prediction Barrier */ #define X86_FEATURE_AMD_IBRS (13*32+14) /* "" Indirect Branch Restricted Speculation */ #define X86_FEATURE_AMD_STIBP (13*32+15) /* "" Single Thread Indirect Branch Predictors */ +#define X86_FEATURE_AMD_SSBD (13*32+24) /* "" Speculative Store Bypass Disable */ #define X86_FEATURE_VIRT_SSBD (13*32+25) /* Virtualized Speculative Store Bypass Disable */ +#define X86_FEATURE_AMD_SSB_NO (13*32+26) /* "" Speculative Store Bypass is fixed in hardware. */ /* Thermal and Power Management Leaf, CPUID level 0x00000006 (EAX), word 14 */ #define X86_FEATURE_DTHERM (14*32+ 0) /* Digital Thermal Sensor */ diff --git a/arch/x86/include/asm/mcsafe_test.h b/arch/x86/include/asm/mcsafe_test.h new file mode 100644 index 000000000000..eb59804b6201 --- /dev/null +++ b/arch/x86/include/asm/mcsafe_test.h @@ -0,0 +1,75 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +#ifndef _MCSAFE_TEST_H_ +#define _MCSAFE_TEST_H_ + +#ifndef __ASSEMBLY__ +#ifdef CONFIG_MCSAFE_TEST +extern unsigned long mcsafe_test_src; +extern unsigned long mcsafe_test_dst; + +static inline void mcsafe_inject_src(void *addr) +{ + if (addr) + mcsafe_test_src = (unsigned long) addr; + else + mcsafe_test_src = ~0UL; +} + +static inline void mcsafe_inject_dst(void *addr) +{ + if (addr) + mcsafe_test_dst = (unsigned long) addr; + else + mcsafe_test_dst = ~0UL; +} +#else /* CONFIG_MCSAFE_TEST */ +static inline void mcsafe_inject_src(void *addr) +{ +} + +static inline void mcsafe_inject_dst(void *addr) +{ +} +#endif /* CONFIG_MCSAFE_TEST */ + +#else /* __ASSEMBLY__ */ +#include <asm/export.h> + +#ifdef CONFIG_MCSAFE_TEST +.macro MCSAFE_TEST_CTL + .pushsection .data + .align 8 + .globl mcsafe_test_src + mcsafe_test_src: + .quad 0 + EXPORT_SYMBOL_GPL(mcsafe_test_src) + .globl mcsafe_test_dst + mcsafe_test_dst: + .quad 0 + EXPORT_SYMBOL_GPL(mcsafe_test_dst) + .popsection +.endm + +.macro MCSAFE_TEST_SRC reg count target + leaq \count(\reg), %r9 + cmp mcsafe_test_src, %r9 + ja \target +.endm + +.macro MCSAFE_TEST_DST reg count target + leaq \count(\reg), %r9 + cmp mcsafe_test_dst, %r9 + ja \target +.endm +#else +.macro MCSAFE_TEST_CTL +.endm + +.macro MCSAFE_TEST_SRC reg count target +.endm + +.macro MCSAFE_TEST_DST reg count target +.endm +#endif /* CONFIG_MCSAFE_TEST */ +#endif /* __ASSEMBLY__ */ +#endif /* _MCSAFE_TEST_H_ */ diff --git a/arch/x86/include/asm/mmu_context.h b/arch/x86/include/asm/mmu_context.h index cf9911b5a53c..bbc796eb0a3b 100644 --- a/arch/x86/include/asm/mmu_context.h +++ b/arch/x86/include/asm/mmu_context.h @@ -288,21 +288,6 @@ static inline void arch_unmap(struct mm_struct *mm, struct vm_area_struct *vma, mpx_notify_unmap(mm, vma, start, end); } -#ifdef CONFIG_X86_INTEL_MEMORY_PROTECTION_KEYS -static inline int vma_pkey(struct vm_area_struct *vma) -{ - unsigned long vma_pkey_mask = VM_PKEY_BIT0 | VM_PKEY_BIT1 | - VM_PKEY_BIT2 | VM_PKEY_BIT3; - - return (vma->vm_flags & vma_pkey_mask) >> VM_PKEY_SHIFT; -} -#else -static inline int vma_pkey(struct vm_area_struct *vma) -{ - return 0; -} -#endif - /* * We only want to enforce protection keys on the current process * because we effectively have no access to PKRU for other diff --git a/arch/x86/include/asm/msr-index.h b/arch/x86/include/asm/msr-index.h index fda2114197b3..68b2c3150de1 100644 --- a/arch/x86/include/asm/msr-index.h +++ b/arch/x86/include/asm/msr-index.h @@ -62,8 +62,8 @@ #define NHM_C3_AUTO_DEMOTE (1UL << 25) #define NHM_C1_AUTO_DEMOTE (1UL << 26) #define ATM_LNC_C6_AUTO_DEMOTE (1UL << 25) -#define SNB_C1_AUTO_UNDEMOTE (1UL << 27) -#define SNB_C3_AUTO_UNDEMOTE (1UL << 28) +#define SNB_C3_AUTO_UNDEMOTE (1UL << 27) +#define SNB_C1_AUTO_UNDEMOTE (1UL << 28) #define MSR_MTRRcap 0x000000fe diff --git a/arch/x86/include/asm/nospec-branch.h b/arch/x86/include/asm/nospec-branch.h index 8b38df98548e..f6f6c63da62f 100644 --- a/arch/x86/include/asm/nospec-branch.h +++ b/arch/x86/include/asm/nospec-branch.h @@ -308,16 +308,20 @@ do { \ * lfence * jmp spec_trap * do_rop: - * mov %rax,(%rsp) + * mov %rax,(%rsp) for x86_64 + * mov %edx,(%esp) for x86_32 * retq * * Without retpolines configured: * - * jmp *%rax + * jmp *%rax for x86_64 + * jmp *%edx for x86_32 */ #ifdef CONFIG_RETPOLINE -# define RETPOLINE_RAX_BPF_JIT_SIZE 17 -# define RETPOLINE_RAX_BPF_JIT() \ +# ifdef CONFIG_X86_64 +# define RETPOLINE_RAX_BPF_JIT_SIZE 17 +# define RETPOLINE_RAX_BPF_JIT() \ +do { \ EMIT1_off32(0xE8, 7); /* callq do_rop */ \ /* spec_trap: */ \ EMIT2(0xF3, 0x90); /* pause */ \ @@ -325,11 +329,30 @@ do { \ EMIT2(0xEB, 0xF9); /* jmp spec_trap */ \ /* do_rop: */ \ EMIT4(0x48, 0x89, 0x04, 0x24); /* mov %rax,(%rsp) */ \ - EMIT1(0xC3); /* retq */ -#else -# define RETPOLINE_RAX_BPF_JIT_SIZE 2 -# define RETPOLINE_RAX_BPF_JIT() \ - EMIT2(0xFF, 0xE0); /* jmp *%rax */ + EMIT1(0xC3); /* retq */ \ +} while (0) +# else /* !CONFIG_X86_64 */ +# define RETPOLINE_EDX_BPF_JIT() \ +do { \ + EMIT1_off32(0xE8, 7); /* call do_rop */ \ + /* spec_trap: */ \ + EMIT2(0xF3, 0x90); /* pause */ \ + EMIT3(0x0F, 0xAE, 0xE8); /* lfence */ \ + EMIT2(0xEB, 0xF9); /* jmp spec_trap */ \ + /* do_rop: */ \ + EMIT3(0x89, 0x14, 0x24); /* mov %edx,(%esp) */ \ + EMIT1(0xC3); /* ret */ \ +} while (0) +# endif +#else /* !CONFIG_RETPOLINE */ +# ifdef CONFIG_X86_64 +# define RETPOLINE_RAX_BPF_JIT_SIZE 2 +# define RETPOLINE_RAX_BPF_JIT() \ + EMIT2(0xFF, 0xE0); /* jmp *%rax */ +# else /* !CONFIG_X86_64 */ +# define RETPOLINE_EDX_BPF_JIT() \ + EMIT2(0xFF, 0xE2) /* jmp *%edx */ +# endif #endif #endif /* _ASM_X86_NOSPEC_BRANCH_H_ */ diff --git a/arch/x86/include/asm/page_types.h b/arch/x86/include/asm/page_types.h index 1e53560a84bb..c85e15010f48 100644 --- a/arch/x86/include/asm/page_types.h +++ b/arch/x86/include/asm/page_types.h @@ -17,7 +17,6 @@ #define PUD_PAGE_SIZE (_AC(1, UL) << PUD_SHIFT) #define PUD_PAGE_MASK (~(PUD_PAGE_SIZE-1)) -#define __PHYSICAL_MASK ((phys_addr_t)(__sme_clr((1ULL << __PHYSICAL_MASK_SHIFT) - 1))) #define __VIRTUAL_MASK ((1UL << __VIRTUAL_MASK_SHIFT) - 1) /* Cast *PAGE_MASK to a signed type so that it is sign-extended if @@ -55,6 +54,13 @@ #ifndef __ASSEMBLY__ +#ifdef CONFIG_DYNAMIC_PHYSICAL_MASK +extern phys_addr_t physical_mask; +#define __PHYSICAL_MASK physical_mask +#else +#define __PHYSICAL_MASK ((phys_addr_t)((1ULL << __PHYSICAL_MASK_SHIFT) - 1)) +#endif + extern int devmem_is_allowed(unsigned long pagenr); extern unsigned long max_low_pfn_mapped; diff --git a/arch/x86/include/asm/pgtable_64.h b/arch/x86/include/asm/pgtable_64.h index 3c5385f9a88f..0fdcd21dadbd 100644 --- a/arch/x86/include/asm/pgtable_64.h +++ b/arch/x86/include/asm/pgtable_64.h @@ -216,7 +216,7 @@ static inline pgd_t pti_set_user_pgd(pgd_t *pgdp, pgd_t pgd) } #endif -static inline void native_set_p4d(p4d_t *p4dp, p4d_t p4d) +static __always_inline void native_set_p4d(p4d_t *p4dp, p4d_t p4d) { pgd_t pgd; @@ -230,7 +230,7 @@ static inline void native_set_p4d(p4d_t *p4dp, p4d_t p4d) *p4dp = native_make_p4d(native_pgd_val(pgd)); } -static inline void native_p4d_clear(p4d_t *p4d) +static __always_inline void native_p4d_clear(p4d_t *p4d) { native_set_p4d(p4d, native_make_p4d(0)); } diff --git a/arch/x86/include/asm/pgtable_types.h b/arch/x86/include/asm/pgtable_types.h index 1e5a40673953..99fff853c944 100644 --- a/arch/x86/include/asm/pgtable_types.h +++ b/arch/x86/include/asm/pgtable_types.h @@ -65,7 +65,6 @@ #define _PAGE_PKEY_BIT2 (_AT(pteval_t, 0)) #define _PAGE_PKEY_BIT3 (_AT(pteval_t, 0)) #endif -#define __HAVE_ARCH_PTE_SPECIAL #define _PAGE_PKEY_MASK (_PAGE_PKEY_BIT0 | \ _PAGE_PKEY_BIT1 | \ diff --git a/arch/x86/include/asm/pkeys.h b/arch/x86/include/asm/pkeys.h index 851c04b7a092..19b137f1b3be 100644 --- a/arch/x86/include/asm/pkeys.h +++ b/arch/x86/include/asm/pkeys.h @@ -9,6 +9,11 @@ extern int arch_set_user_pkey_access(struct task_struct *tsk, int pkey, unsigned long init_val); +static inline bool arch_pkeys_enabled(void) +{ + return boot_cpu_has(X86_FEATURE_OSPKE); +} + /* * Try to dedicate one of the protection keys to be used as an * execute-only protection key. @@ -116,4 +121,12 @@ extern int __arch_set_user_pkey_access(struct task_struct *tsk, int pkey, unsigned long init_val); extern void copy_init_pkru_to_fpregs(void); +static inline int vma_pkey(struct vm_area_struct *vma) +{ + unsigned long vma_pkey_mask = VM_PKEY_BIT0 | VM_PKEY_BIT1 | + VM_PKEY_BIT2 | VM_PKEY_BIT3; + + return (vma->vm_flags & vma_pkey_mask) >> VM_PKEY_SHIFT; +} + #endif /*_ASM_X86_PKEYS_H */ diff --git a/arch/x86/include/asm/trace/irq_vectors.h b/arch/x86/include/asm/trace/irq_vectors.h index 22647a642e98..0af81b590a0c 100644 --- a/arch/x86/include/asm/trace/irq_vectors.h +++ b/arch/x86/include/asm/trace/irq_vectors.h @@ -236,7 +236,7 @@ TRACE_EVENT(vector_alloc, TP_PROTO(unsigned int irq, unsigned int vector, bool reserved, int ret), - TP_ARGS(irq, vector, ret, reserved), + TP_ARGS(irq, vector, reserved, ret), TP_STRUCT__entry( __field( unsigned int, irq ) diff --git a/arch/x86/include/asm/x86_init.h b/arch/x86/include/asm/x86_init.h index 2d27236c16a3..b85a7c54c6a1 100644 --- a/arch/x86/include/asm/x86_init.h +++ b/arch/x86/include/asm/x86_init.h @@ -301,5 +301,6 @@ extern struct x86_apic_ops x86_apic_ops; extern void x86_early_init_platform_quirks(void); extern void x86_init_noop(void); extern void x86_init_uint_noop(unsigned int unused); +extern bool x86_pnpbios_disabled(void); #endif diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c index 7553819c74c3..3982f79d2377 100644 --- a/arch/x86/kernel/apic/io_apic.c +++ b/arch/x86/kernel/apic/io_apic.c @@ -1851,7 +1851,7 @@ static void ioapic_ir_ack_level(struct irq_data *irq_data) * intr-remapping table entry. Hence for the io-apic * EOI we use the pin number. */ - ack_APIC_irq(); + apic_ack_irq(irq_data); eoi_ioapic_pin(data->entry.vector, data); } diff --git a/arch/x86/kernel/apic/vector.c b/arch/x86/kernel/apic/vector.c index bb6f7a2148d7..35aaee4fc028 100644 --- a/arch/x86/kernel/apic/vector.c +++ b/arch/x86/kernel/apic/vector.c @@ -235,6 +235,15 @@ static int allocate_vector(struct irq_data *irqd, const struct cpumask *dest) if (vector && cpu_online(cpu) && cpumask_test_cpu(cpu, dest)) return 0; + /* + * Careful here. @apicd might either have move_in_progress set or + * be enqueued for cleanup. Assigning a new vector would either + * leave a stale vector on some CPU around or in case of a pending + * cleanup corrupt the hlist. + */ + if (apicd->move_in_progress || !hlist_unhashed(&apicd->clist)) + return -EBUSY; + vector = irq_matrix_alloc(vector_matrix, dest, resvd, &cpu); if (vector > 0) apic_update_vector(irqd, vector, cpu); @@ -579,8 +588,7 @@ error: static void x86_vector_debug_show(struct seq_file *m, struct irq_domain *d, struct irq_data *irqd, int ind) { - unsigned int cpu, vector, prev_cpu, prev_vector; - struct apic_chip_data *apicd; + struct apic_chip_data apicd; unsigned long flags; int irq; @@ -596,24 +604,26 @@ static void x86_vector_debug_show(struct seq_file *m, struct irq_domain *d, return; } - apicd = irqd->chip_data; - if (!apicd) { + if (!irqd->chip_data) { seq_printf(m, "%*sVector: Not assigned\n", ind, ""); return; } raw_spin_lock_irqsave(&vector_lock, flags); - cpu = apicd->cpu; - vector = apicd->vector; - prev_cpu = apicd->prev_cpu; - prev_vector = apicd->prev_vector; + memcpy(&apicd, irqd->chip_data, sizeof(apicd)); raw_spin_unlock_irqrestore(&vector_lock, flags); - seq_printf(m, "%*sVector: %5u\n", ind, "", vector); - seq_printf(m, "%*sTarget: %5u\n", ind, "", cpu); - if (prev_vector) { - seq_printf(m, "%*sPrevious vector: %5u\n", ind, "", prev_vector); - seq_printf(m, "%*sPrevious target: %5u\n", ind, "", prev_cpu); + + seq_printf(m, "%*sVector: %5u\n", ind, "", apicd.vector); + seq_printf(m, "%*sTarget: %5u\n", ind, "", apicd.cpu); + if (apicd.prev_vector) { + seq_printf(m, "%*sPrevious vector: %5u\n", ind, "", apicd.prev_vector); + seq_printf(m, "%*sPrevious target: %5u\n", ind, "", apicd.prev_cpu); } + seq_printf(m, "%*smove_in_progress: %u\n", ind, "", apicd.move_in_progress ? 1 : 0); + seq_printf(m, "%*sis_managed: %u\n", ind, "", apicd.is_managed ? 1 : 0); + seq_printf(m, "%*scan_reserve: %u\n", ind, "", apicd.can_reserve ? 1 : 0); + seq_printf(m, "%*shas_reserved: %u\n", ind, "", apicd.has_reserved ? 1 : 0); + seq_printf(m, "%*scleanup_pending: %u\n", ind, "", !hlist_unhashed(&apicd.clist)); } #endif @@ -800,13 +810,18 @@ static int apic_retrigger_irq(struct irq_data *irqd) return 1; } -void apic_ack_edge(struct irq_data *irqd) +void apic_ack_irq(struct irq_data *irqd) { - irq_complete_move(irqd_cfg(irqd)); irq_move_irq(irqd); ack_APIC_irq(); } +void apic_ack_edge(struct irq_data *irqd) +{ + irq_complete_move(irqd_cfg(irqd)); + apic_ack_irq(irqd); +} + static struct irq_chip lapic_controller = { .name = "APIC", .irq_ack = apic_ack_edge, diff --git a/arch/x86/kernel/cpu/bugs.c b/arch/x86/kernel/cpu/bugs.c index 7416fc206b4a..cd0fda1fff6d 100644 --- a/arch/x86/kernel/cpu/bugs.c +++ b/arch/x86/kernel/cpu/bugs.c @@ -529,18 +529,15 @@ static enum ssb_mitigation __init __ssb_select_mitigation(void) if (mode == SPEC_STORE_BYPASS_DISABLE) { setup_force_cpu_cap(X86_FEATURE_SPEC_STORE_BYPASS_DISABLE); /* - * Intel uses the SPEC CTRL MSR Bit(2) for this, while AMD uses - * a completely different MSR and bit dependent on family. + * Intel uses the SPEC CTRL MSR Bit(2) for this, while AMD may + * use a completely different MSR and bit dependent on family. */ - switch (boot_cpu_data.x86_vendor) { - case X86_VENDOR_INTEL: + if (!static_cpu_has(X86_FEATURE_MSR_SPEC_CTRL)) + x86_amd_ssb_disable(); + else { x86_spec_ctrl_base |= SPEC_CTRL_SSBD; x86_spec_ctrl_mask |= SPEC_CTRL_SSBD; wrmsrl(MSR_IA32_SPEC_CTRL, x86_spec_ctrl_base); - break; - case X86_VENDOR_AMD: - x86_amd_ssb_disable(); - break; } } diff --git a/arch/x86/kernel/cpu/common.c b/arch/x86/kernel/cpu/common.c index 95c8e507580d..910b47ee8078 100644 --- a/arch/x86/kernel/cpu/common.c +++ b/arch/x86/kernel/cpu/common.c @@ -803,6 +803,12 @@ static void init_speculation_control(struct cpuinfo_x86 *c) set_cpu_cap(c, X86_FEATURE_STIBP); set_cpu_cap(c, X86_FEATURE_MSR_SPEC_CTRL); } + + if (cpu_has(c, X86_FEATURE_AMD_SSBD)) { + set_cpu_cap(c, X86_FEATURE_SSBD); + set_cpu_cap(c, X86_FEATURE_MSR_SPEC_CTRL); + clear_cpu_cap(c, X86_FEATURE_VIRT_SSBD); + } } void get_cpu_cap(struct cpuinfo_x86 *c) @@ -992,7 +998,8 @@ static void __init cpu_set_bug_bits(struct cpuinfo_x86 *c) rdmsrl(MSR_IA32_ARCH_CAPABILITIES, ia32_cap); if (!x86_match_cpu(cpu_no_spec_store_bypass) && - !(ia32_cap & ARCH_CAP_SSB_NO)) + !(ia32_cap & ARCH_CAP_SSB_NO) && + !cpu_has(c, X86_FEATURE_AMD_SSB_NO)) setup_force_cpu_bug(X86_BUG_SPEC_STORE_BYPASS); if (x86_match_cpu(cpu_no_meltdown)) diff --git a/arch/x86/kernel/cpu/intel_rdt.c b/arch/x86/kernel/cpu/intel_rdt.c index 24bfa63e86cf..ec4754f81cbd 100644 --- a/arch/x86/kernel/cpu/intel_rdt.c +++ b/arch/x86/kernel/cpu/intel_rdt.c @@ -845,6 +845,8 @@ static __init void rdt_quirks(void) case INTEL_FAM6_SKYLAKE_X: if (boot_cpu_data.x86_stepping <= 4) set_rdt_options("!cmt,!mbmtotal,!mbmlocal,!l3cat"); + else + set_rdt_options("!l3cat"); } } diff --git a/arch/x86/kernel/cpu/microcode/core.c b/arch/x86/kernel/cpu/microcode/core.c index 77e201301528..08286269fd24 100644 --- a/arch/x86/kernel/cpu/microcode/core.c +++ b/arch/x86/kernel/cpu/microcode/core.c @@ -70,7 +70,7 @@ static DEFINE_MUTEX(microcode_mutex); /* * Serialize late loading so that CPUs get updated one-by-one. */ -static DEFINE_SPINLOCK(update_lock); +static DEFINE_RAW_SPINLOCK(update_lock); struct ucode_cpu_info ucode_cpu_info[NR_CPUS]; @@ -560,9 +560,9 @@ static int __reload_late(void *info) if (__wait_for_cpus(&late_cpus_in, NSEC_PER_SEC)) return -1; - spin_lock(&update_lock); + raw_spin_lock(&update_lock); apply_microcode_local(&err); - spin_unlock(&update_lock); + raw_spin_unlock(&update_lock); /* siblings return UCODE_OK because their engine got updated already */ if (err > UCODE_NFOUND) { diff --git a/arch/x86/kernel/cpu/mtrr/if.c b/arch/x86/kernel/cpu/mtrr/if.c index 558444b23923..c610f47373e4 100644 --- a/arch/x86/kernel/cpu/mtrr/if.c +++ b/arch/x86/kernel/cpu/mtrr/if.c @@ -106,17 +106,9 @@ mtrr_write(struct file *file, const char __user *buf, size_t len, loff_t * ppos) memset(line, 0, LINE_SIZE); - length = len; - length--; - - if (length > LINE_SIZE - 1) - length = LINE_SIZE - 1; - + length = strncpy_from_user(line, buf, LINE_SIZE - 1); if (length < 0) - return -EINVAL; - - if (copy_from_user(line, buf, length)) - return -EFAULT; + return length; linelen = strlen(line); ptr = line + linelen - 1; @@ -149,17 +141,16 @@ mtrr_write(struct file *file, const char __user *buf, size_t len, loff_t * ppos) return -EINVAL; ptr = skip_spaces(ptr + 5); - for (i = 0; i < MTRR_NUM_TYPES; ++i) { - if (strcmp(ptr, mtrr_strings[i])) - continue; - base >>= PAGE_SHIFT; - size >>= PAGE_SHIFT; - err = mtrr_add_page((unsigned long)base, (unsigned long)size, i, true); - if (err < 0) - return err; - return len; - } - return -EINVAL; + i = match_string(mtrr_strings, MTRR_NUM_TYPES, ptr); + if (i < 0) + return i; + + base >>= PAGE_SHIFT; + size >>= PAGE_SHIFT; + err = mtrr_add_page((unsigned long)base, (unsigned long)size, i, true); + if (err < 0) + return err; + return len; } static long diff --git a/arch/x86/kernel/i8237.c b/arch/x86/kernel/i8237.c index 8eeaa81de066..0a3e70fd00d6 100644 --- a/arch/x86/kernel/i8237.c +++ b/arch/x86/kernel/i8237.c @@ -9,10 +9,12 @@ * your option) any later version. */ +#include <linux/dmi.h> #include <linux/init.h> #include <linux/syscore_ops.h> #include <asm/dma.h> +#include <asm/x86_init.h> /* * This module just handles suspend/resume issues with the @@ -49,6 +51,29 @@ static struct syscore_ops i8237_syscore_ops = { static int __init i8237A_init_ops(void) { + /* + * From SKL PCH onwards, the legacy DMA device is removed in which the + * I/O ports (81h-83h, 87h, 89h-8Bh, 8Fh) related to it are removed + * as well. All removed ports must return 0xff for a inb() request. + * + * Note: DMA_PAGE_2 (port 0x81) should not be checked for detecting + * the presence of DMA device since it may be used by BIOS to decode + * LPC traffic for POST codes. Original LPC only decodes one byte of + * port 0x80 but some BIOS may choose to enhance PCH LPC port 0x8x + * decoding. + */ + if (dma_inb(DMA_PAGE_0) == 0xFF) + return -ENODEV; + + /* + * It is not required to load this driver as newer SoC may not + * support 8237 DMA or bus mastering from LPC. Platform firmware + * must announce the support for such legacy devices via + * ACPI_FADT_LEGACY_DEVICES field in FADT table. + */ + if (x86_pnpbios_disabled() && dmi_get_bios_year() >= 2017) + return -ENODEV; + register_syscore_ops(&i8237_syscore_ops); return 0; } diff --git a/arch/x86/kernel/idt.c b/arch/x86/kernel/idt.c index 2c3a1b4294eb..74383a3780dc 100644 --- a/arch/x86/kernel/idt.c +++ b/arch/x86/kernel/idt.c @@ -317,15 +317,12 @@ void __init idt_setup_apic_and_irq_gates(void) set_intr_gate(i, entry); } - for_each_clear_bit_from(i, system_vectors, NR_VECTORS) { #ifdef CONFIG_X86_LOCAL_APIC + for_each_clear_bit_from(i, system_vectors, NR_VECTORS) { set_bit(i, system_vectors); set_intr_gate(i, spurious_interrupt); -#else - entry = irq_entries_start + 8 * (i - FIRST_EXTERNAL_VECTOR); - set_intr_gate(i, entry); -#endif } +#endif } /** diff --git a/arch/x86/kernel/platform-quirks.c b/arch/x86/kernel/platform-quirks.c index 235fe6008ac8..b348a672f71d 100644 --- a/arch/x86/kernel/platform-quirks.c +++ b/arch/x86/kernel/platform-quirks.c @@ -33,9 +33,14 @@ void __init x86_early_init_platform_quirks(void) x86_platform.set_legacy_features(); } +bool __init x86_pnpbios_disabled(void) +{ + return x86_platform.legacy.devices.pnpbios == 0; +} + #if defined(CONFIG_PNPBIOS) bool __init arch_pnpbios_disabled(void) { - return x86_platform.legacy.devices.pnpbios == 0; + return x86_pnpbios_disabled(); } #endif diff --git a/arch/x86/kernel/setup.c b/arch/x86/kernel/setup.c index 5c623dfe39d1..2f86d883dd95 100644 --- a/arch/x86/kernel/setup.c +++ b/arch/x86/kernel/setup.c @@ -1312,11 +1312,3 @@ static int __init register_kernel_offset_dumper(void) return 0; } __initcall(register_kernel_offset_dumper); - -void arch_show_smap(struct seq_file *m, struct vm_area_struct *vma) -{ - if (!boot_cpu_has(X86_FEATURE_OSPKE)) - return; - - seq_printf(m, "ProtectionKey: %8u\n", vma_pkey(vma)); -} diff --git a/arch/x86/kvm/cpuid.c b/arch/x86/kvm/cpuid.c index 92bf2f2e7cdd..f4f30d0c25c4 100644 --- a/arch/x86/kvm/cpuid.c +++ b/arch/x86/kvm/cpuid.c @@ -379,7 +379,8 @@ static inline int __do_cpuid_ent(struct kvm_cpuid_entry2 *entry, u32 function, /* cpuid 0x80000008.ebx */ const u32 kvm_cpuid_8000_0008_ebx_x86_features = - F(AMD_IBPB) | F(AMD_IBRS) | F(VIRT_SSBD); + F(AMD_IBPB) | F(AMD_IBRS) | F(AMD_SSBD) | F(VIRT_SSBD) | + F(AMD_SSB_NO); /* cpuid 0xC0000001.edx */ const u32 kvm_cpuid_C000_0001_edx_x86_features = @@ -664,7 +665,12 @@ static inline int __do_cpuid_ent(struct kvm_cpuid_entry2 *entry, u32 function, entry->ebx |= F(VIRT_SSBD); entry->ebx &= kvm_cpuid_8000_0008_ebx_x86_features; cpuid_mask(&entry->ebx, CPUID_8000_0008_EBX); - if (boot_cpu_has(X86_FEATURE_LS_CFG_SSBD)) + /* + * The preference is to use SPEC CTRL MSR instead of the + * VIRT_SPEC MSR. + */ + if (boot_cpu_has(X86_FEATURE_LS_CFG_SSBD) && + !boot_cpu_has(X86_FEATURE_AMD_SSBD)) entry->ebx |= F(VIRT_SSBD); break; } diff --git a/arch/x86/kvm/svm.c b/arch/x86/kvm/svm.c index 26110c202b19..950ec50f77c3 100644 --- a/arch/x86/kvm/svm.c +++ b/arch/x86/kvm/svm.c @@ -4115,7 +4115,8 @@ static int svm_get_msr(struct kvm_vcpu *vcpu, struct msr_data *msr_info) break; case MSR_IA32_SPEC_CTRL: if (!msr_info->host_initiated && - !guest_cpuid_has(vcpu, X86_FEATURE_AMD_IBRS)) + !guest_cpuid_has(vcpu, X86_FEATURE_AMD_IBRS) && + !guest_cpuid_has(vcpu, X86_FEATURE_AMD_SSBD)) return 1; msr_info->data = svm->spec_ctrl; @@ -4217,11 +4218,12 @@ static int svm_set_msr(struct kvm_vcpu *vcpu, struct msr_data *msr) break; case MSR_IA32_SPEC_CTRL: if (!msr->host_initiated && - !guest_cpuid_has(vcpu, X86_FEATURE_AMD_IBRS)) + !guest_cpuid_has(vcpu, X86_FEATURE_AMD_IBRS) && + !guest_cpuid_has(vcpu, X86_FEATURE_AMD_SSBD)) return 1; /* The STIBP bit doesn't fault even if it's not advertised */ - if (data & ~(SPEC_CTRL_IBRS | SPEC_CTRL_STIBP)) + if (data & ~(SPEC_CTRL_IBRS | SPEC_CTRL_STIBP | SPEC_CTRL_SSBD)) return 1; svm->spec_ctrl = data; diff --git a/arch/x86/lib/memcpy_64.S b/arch/x86/lib/memcpy_64.S index c3b527a9f95d..298ef1479240 100644 --- a/arch/x86/lib/memcpy_64.S +++ b/arch/x86/lib/memcpy_64.S @@ -3,6 +3,7 @@ #include <linux/linkage.h> #include <asm/errno.h> #include <asm/cpufeatures.h> +#include <asm/mcsafe_test.h> #include <asm/alternative-asm.h> #include <asm/export.h> @@ -183,6 +184,9 @@ ENTRY(memcpy_orig) ENDPROC(memcpy_orig) #ifndef CONFIG_UML + +MCSAFE_TEST_CTL + /* * __memcpy_mcsafe - memory copy with machine check exception handling * Note that we only catch machine checks when reading the source addresses. @@ -206,6 +210,8 @@ ENTRY(__memcpy_mcsafe) subl %ecx, %edx .L_read_leading_bytes: movb (%rsi), %al + MCSAFE_TEST_SRC %rsi 1 .E_leading_bytes + MCSAFE_TEST_DST %rdi 1 .E_leading_bytes .L_write_leading_bytes: movb %al, (%rdi) incq %rsi @@ -221,6 +227,8 @@ ENTRY(__memcpy_mcsafe) .L_read_words: movq (%rsi), %r8 + MCSAFE_TEST_SRC %rsi 8 .E_read_words + MCSAFE_TEST_DST %rdi 8 .E_write_words .L_write_words: movq %r8, (%rdi) addq $8, %rsi @@ -237,6 +245,8 @@ ENTRY(__memcpy_mcsafe) movl %edx, %ecx .L_read_trailing_bytes: movb (%rsi), %al + MCSAFE_TEST_SRC %rsi 1 .E_trailing_bytes + MCSAFE_TEST_DST %rdi 1 .E_trailing_bytes .L_write_trailing_bytes: movb %al, (%rdi) incq %rsi diff --git a/arch/x86/mm/mem_encrypt_identity.c b/arch/x86/mm/mem_encrypt_identity.c index 1b2197d13832..7ae36868aed2 100644 --- a/arch/x86/mm/mem_encrypt_identity.c +++ b/arch/x86/mm/mem_encrypt_identity.c @@ -527,6 +527,7 @@ void __init sme_enable(struct boot_params *bp) /* SEV state cannot be controlled by a command line option */ sme_me_mask = me_mask; sev_enabled = true; + physical_mask &= ~sme_me_mask; return; } @@ -561,4 +562,6 @@ void __init sme_enable(struct boot_params *bp) sme_me_mask = 0; else sme_me_mask = active_by_default ? me_mask : 0; + + physical_mask &= ~sme_me_mask; } diff --git a/arch/x86/mm/pgtable.c b/arch/x86/mm/pgtable.c index ffc8c13c50e4..47b5951e592b 100644 --- a/arch/x86/mm/pgtable.c +++ b/arch/x86/mm/pgtable.c @@ -8,6 +8,11 @@ #include <asm/fixmap.h> #include <asm/mtrr.h> +#ifdef CONFIG_DYNAMIC_PHYSICAL_MASK +phys_addr_t physical_mask __ro_after_init = (1ULL << __PHYSICAL_MASK_SHIFT) - 1; +EXPORT_SYMBOL(physical_mask); +#endif + #define PGALLOC_GFP (GFP_KERNEL_ACCOUNT | __GFP_ZERO) #ifdef CONFIG_HIGHPTE @@ -114,13 +119,12 @@ static inline void pgd_list_del(pgd_t *pgd) static void pgd_set_mm(pgd_t *pgd, struct mm_struct *mm) { - BUILD_BUG_ON(sizeof(virt_to_page(pgd)->index) < sizeof(mm)); - virt_to_page(pgd)->index = (pgoff_t)mm; + virt_to_page(pgd)->pt_mm = mm; } struct mm_struct *pgd_page_get_mm(struct page *page) { - return (struct mm_struct *)page->index; + return page->pt_mm; } static void pgd_ctor(struct mm_struct *mm, pgd_t *pgd) diff --git a/arch/x86/net/Makefile b/arch/x86/net/Makefile index fefb4b619598..59e123da580c 100644 --- a/arch/x86/net/Makefile +++ b/arch/x86/net/Makefile @@ -1,6 +1,9 @@ # # Arch-specific network modules # -OBJECT_FILES_NON_STANDARD_bpf_jit.o += y -obj-$(CONFIG_BPF_JIT) += bpf_jit.o bpf_jit_comp.o +ifeq ($(CONFIG_X86_32),y) + obj-$(CONFIG_BPF_JIT) += bpf_jit_comp32.o +else + obj-$(CONFIG_BPF_JIT) += bpf_jit_comp.o +endif diff --git a/arch/x86/net/bpf_jit.S b/arch/x86/net/bpf_jit.S deleted file mode 100644 index b33093f84528..000000000000 --- a/arch/x86/net/bpf_jit.S +++ /dev/null @@ -1,154 +0,0 @@ -/* bpf_jit.S : BPF JIT helper functions - * - * Copyright (C) 2011 Eric Dumazet (eric.dumazet@gmail.com) - * - * This program is free software; you can redistribute it and/or - * modify it under the terms of the GNU General Public License - * as published by the Free Software Foundation; version 2 - * of the License. - */ -#include <linux/linkage.h> -#include <asm/frame.h> - -/* - * Calling convention : - * rbx : skb pointer (callee saved) - * esi : offset of byte(s) to fetch in skb (can be scratched) - * r10 : copy of skb->data - * r9d : hlen = skb->len - skb->data_len - */ -#define SKBDATA %r10 -#define SKF_MAX_NEG_OFF $(-0x200000) /* SKF_LL_OFF from filter.h */ - -#define FUNC(name) \ - .globl name; \ - .type name, @function; \ - name: - -FUNC(sk_load_word) - test %esi,%esi - js bpf_slow_path_word_neg - -FUNC(sk_load_word_positive_offset) - mov %r9d,%eax # hlen - sub %esi,%eax # hlen - offset - cmp $3,%eax - jle bpf_slow_path_word - mov (SKBDATA,%rsi),%eax - bswap %eax /* ntohl() */ - ret - -FUNC(sk_load_half) - test %esi,%esi - js bpf_slow_path_half_neg - -FUNC(sk_load_half_positive_offset) - mov %r9d,%eax - sub %esi,%eax # hlen - offset - cmp $1,%eax - jle bpf_slow_path_half - movzwl (SKBDATA,%rsi),%eax - rol $8,%ax # ntohs() - ret - -FUNC(sk_load_byte) - test %esi,%esi - js bpf_slow_path_byte_neg - -FUNC(sk_load_byte_positive_offset) - cmp %esi,%r9d /* if (offset >= hlen) goto bpf_slow_path_byte */ - jle bpf_slow_path_byte - movzbl (SKBDATA,%rsi),%eax - ret - -/* rsi contains offset and can be scratched */ -#define bpf_slow_path_common(LEN) \ - lea 32(%rbp), %rdx;\ - FRAME_BEGIN; \ - mov %rbx, %rdi; /* arg1 == skb */ \ - push %r9; \ - push SKBDATA; \ -/* rsi already has offset */ \ - mov $LEN,%ecx; /* len */ \ - call skb_copy_bits; \ - test %eax,%eax; \ - pop SKBDATA; \ - pop %r9; \ - FRAME_END - - -bpf_slow_path_word: - bpf_slow_path_common(4) - js bpf_error - mov 32(%rbp),%eax - bswap %eax - ret - -bpf_slow_path_half: - bpf_slow_path_common(2) - js bpf_error - mov 32(%rbp),%ax - rol $8,%ax - movzwl %ax,%eax - ret - -bpf_slow_path_byte: - bpf_slow_path_common(1) - js bpf_error - movzbl 32(%rbp),%eax - ret - -#define sk_negative_common(SIZE) \ - FRAME_BEGIN; \ - mov %rbx, %rdi; /* arg1 == skb */ \ - push %r9; \ - push SKBDATA; \ -/* rsi already has offset */ \ - mov $SIZE,%edx; /* size */ \ - call bpf_internal_load_pointer_neg_helper; \ - test %rax,%rax; \ - pop SKBDATA; \ - pop %r9; \ - FRAME_END; \ - jz bpf_error - -bpf_slow_path_word_neg: - cmp SKF_MAX_NEG_OFF, %esi /* test range */ - jl bpf_error /* offset lower -> error */ - -FUNC(sk_load_word_negative_offset) - sk_negative_common(4) - mov (%rax), %eax - bswap %eax - ret - -bpf_slow_path_half_neg: - cmp SKF_MAX_NEG_OFF, %esi - jl bpf_error - -FUNC(sk_load_half_negative_offset) - sk_negative_common(2) - mov (%rax),%ax - rol $8,%ax - movzwl %ax,%eax - ret - -bpf_slow_path_byte_neg: - cmp SKF_MAX_NEG_OFF, %esi - jl bpf_error - -FUNC(sk_load_byte_negative_offset) - sk_negative_common(1) - movzbl (%rax), %eax - ret - -bpf_error: -# force a return 0 from jit handler - xor %eax,%eax - mov (%rbp),%rbx - mov 8(%rbp),%r13 - mov 16(%rbp),%r14 - mov 24(%rbp),%r15 - add $40, %rbp - leaveq - ret diff --git a/arch/x86/net/bpf_jit_comp.c b/arch/x86/net/bpf_jit_comp.c index d765acedc05c..8fca446aaef6 100644 --- a/arch/x86/net/bpf_jit_comp.c +++ b/arch/x86/net/bpf_jit_comp.c @@ -17,15 +17,6 @@ #include <asm/set_memory.h> #include <asm/nospec-branch.h> -/* - * Assembly code in arch/x86/net/bpf_jit.S - */ -extern u8 sk_load_word[], sk_load_half[], sk_load_byte[]; -extern u8 sk_load_word_positive_offset[], sk_load_half_positive_offset[]; -extern u8 sk_load_byte_positive_offset[]; -extern u8 sk_load_word_negative_offset[], sk_load_half_negative_offset[]; -extern u8 sk_load_byte_negative_offset[]; - static u8 *emit_code(u8 *ptr, u32 bytes, unsigned int len) { if (len == 1) @@ -107,9 +98,6 @@ static int bpf_size_to_x86_bytes(int bpf_size) #define X86_JLE 0x7E #define X86_JG 0x7F -#define CHOOSE_LOAD_FUNC(K, func) \ - ((int)K < 0 ? ((int)K >= SKF_LL_OFF ? func##_negative_offset : func) : func##_positive_offset) - /* Pick a register outside of BPF range for JIT internal work */ #define AUX_REG (MAX_BPF_JIT_REG + 1) @@ -120,8 +108,8 @@ static int bpf_size_to_x86_bytes(int bpf_size) * register in load/store instructions, it always needs an * extra byte of encoding and is callee saved. * - * R9 caches skb->len - skb->data_len - * R10 caches skb->data, and used for blinding (if enabled) + * Also x86-64 register R9 is unused. x86-64 register R10 is + * used for blinding (if enabled). */ static const int reg2hex[] = { [BPF_REG_0] = 0, /* RAX */ @@ -196,19 +184,15 @@ static void jit_fill_hole(void *area, unsigned int size) struct jit_context { int cleanup_addr; /* Epilogue code offset */ - bool seen_ld_abs; - bool seen_ax_reg; }; /* Maximum number of bytes emitted while JITing one eBPF insn */ #define BPF_MAX_INSN_SIZE 128 #define BPF_INSN_SAFETY 64 -#define AUX_STACK_SPACE \ - (32 /* Space for RBX, R13, R14, R15 */ + \ - 8 /* Space for skb_copy_bits() buffer */) +#define AUX_STACK_SPACE 40 /* Space for RBX, R13, R14, R15, tailcnt */ -#define PROLOGUE_SIZE 37 +#define PROLOGUE_SIZE 37 /* * Emit x86-64 prologue code for BPF program and check its size. @@ -232,20 +216,8 @@ static void emit_prologue(u8 **pprog, u32 stack_depth, bool ebpf_from_cbpf) /* sub rbp, AUX_STACK_SPACE */ EMIT4(0x48, 0x83, 0xED, AUX_STACK_SPACE); - /* All classic BPF filters use R6(rbx) save it */ - /* mov qword ptr [rbp+0],rbx */ EMIT4(0x48, 0x89, 0x5D, 0); - - /* - * bpf_convert_filter() maps classic BPF register X to R7 and uses R8 - * as temporary, so all tcpdump filters need to spill/fill R7(R13) and - * R8(R14). R9(R15) spill could be made conditional, but there is only - * one 'bpf_error' return path out of helper functions inside bpf_jit.S - * The overhead of extra spill is negligible for any filter other - * than synthetic ones. Therefore not worth adding complexity. - */ - /* mov qword ptr [rbp+8],r13 */ EMIT4(0x4C, 0x89, 0x6D, 8); /* mov qword ptr [rbp+16],r14 */ @@ -353,27 +325,6 @@ static void emit_bpf_tail_call(u8 **pprog) *pprog = prog; } - -static void emit_load_skb_data_hlen(u8 **pprog) -{ - u8 *prog = *pprog; - int cnt = 0; - - /* - * r9d = skb->len - skb->data_len (headlen) - * r10 = skb->data - */ - /* mov %r9d, off32(%rdi) */ - EMIT3_off32(0x44, 0x8b, 0x8f, offsetof(struct sk_buff, len)); - - /* sub %r9d, off32(%rdi) */ - EMIT3_off32(0x44, 0x2b, 0x8f, offsetof(struct sk_buff, data_len)); - - /* mov %r10, off32(%rdi) */ - EMIT3_off32(0x4c, 0x8b, 0x97, offsetof(struct sk_buff, data)); - *pprog = prog; -} - static void emit_mov_imm32(u8 **pprog, bool sign_propagate, u32 dst_reg, const u32 imm32) { @@ -462,8 +413,6 @@ static int do_jit(struct bpf_prog *bpf_prog, int *addrs, u8 *image, { struct bpf_insn *insn = bpf_prog->insnsi; int insn_cnt = bpf_prog->len; - bool seen_ld_abs = ctx->seen_ld_abs | (oldproglen == 0); - bool seen_ax_reg = ctx->seen_ax_reg | (oldproglen == 0); bool seen_exit = false; u8 temp[BPF_MAX_INSN_SIZE + BPF_INSN_SAFETY]; int i, cnt = 0; @@ -473,9 +422,6 @@ static int do_jit(struct bpf_prog *bpf_prog, int *addrs, u8 *image, emit_prologue(&prog, bpf_prog->aux->stack_depth, bpf_prog_was_classic(bpf_prog)); - if (seen_ld_abs) - emit_load_skb_data_hlen(&prog); - for (i = 0; i < insn_cnt; i++, insn++) { const s32 imm32 = insn->imm; u32 dst_reg = insn->dst_reg; @@ -483,13 +429,9 @@ static int do_jit(struct bpf_prog *bpf_prog, int *addrs, u8 *image, u8 b2 = 0, b3 = 0; s64 jmp_offset; u8 jmp_cond; - bool reload_skb_data; int ilen; u8 *func; - if (dst_reg == BPF_REG_AX || src_reg == BPF_REG_AX) - ctx->seen_ax_reg = seen_ax_reg = true; - switch (insn->code) { /* ALU */ case BPF_ALU | BPF_ADD | BPF_X: @@ -916,36 +858,12 @@ xadd: if (is_imm8(insn->off)) case BPF_JMP | BPF_CALL: func = (u8 *) __bpf_call_base + imm32; jmp_offset = func - (image + addrs[i]); - if (seen_ld_abs) { - reload_skb_data = bpf_helper_changes_pkt_data(func); - if (reload_skb_data) { - EMIT1(0x57); /* push %rdi */ - jmp_offset += 22; /* pop, mov, sub, mov */ - } else { - EMIT2(0x41, 0x52); /* push %r10 */ - EMIT2(0x41, 0x51); /* push %r9 */ - /* - * We need to adjust jmp offset, since - * pop %r9, pop %r10 take 4 bytes after call insn - */ - jmp_offset += 4; - } - } if (!imm32 || !is_simm32(jmp_offset)) { pr_err("unsupported BPF func %d addr %p image %p\n", imm32, func, image); return -EINVAL; } EMIT1_off32(0xE8, jmp_offset); - if (seen_ld_abs) { - if (reload_skb_data) { - EMIT1(0x5F); /* pop %rdi */ - emit_load_skb_data_hlen(&prog); - } else { - EMIT2(0x41, 0x59); /* pop %r9 */ - EMIT2(0x41, 0x5A); /* pop %r10 */ - } - } break; case BPF_JMP | BPF_TAIL_CALL: @@ -1080,60 +998,6 @@ emit_jmp: } break; - case BPF_LD | BPF_IND | BPF_W: - func = sk_load_word; - goto common_load; - case BPF_LD | BPF_ABS | BPF_W: - func = CHOOSE_LOAD_FUNC(imm32, sk_load_word); -common_load: - ctx->seen_ld_abs = seen_ld_abs = true; - jmp_offset = func - (image + addrs[i]); - if (!func || !is_simm32(jmp_offset)) { - pr_err("unsupported BPF func %d addr %p image %p\n", - imm32, func, image); - return -EINVAL; - } - if (BPF_MODE(insn->code) == BPF_ABS) { - /* mov %esi, imm32 */ - EMIT1_off32(0xBE, imm32); - } else { - /* mov %rsi, src_reg */ - EMIT_mov(BPF_REG_2, src_reg); - if (imm32) { - if (is_imm8(imm32)) - /* add %esi, imm8 */ - EMIT3(0x83, 0xC6, imm32); - else - /* add %esi, imm32 */ - EMIT2_off32(0x81, 0xC6, imm32); - } - } - /* - * skb pointer is in R6 (%rbx), it will be copied into - * %rdi if skb_copy_bits() call is necessary. - * sk_load_* helpers also use %r10 and %r9d. - * See bpf_jit.S - */ - if (seen_ax_reg) - /* r10 = skb->data, mov %r10, off32(%rbx) */ - EMIT3_off32(0x4c, 0x8b, 0x93, - offsetof(struct sk_buff, data)); - EMIT1_off32(0xE8, jmp_offset); /* call */ - break; - - case BPF_LD | BPF_IND | BPF_H: - func = sk_load_half; - goto common_load; - case BPF_LD | BPF_ABS | BPF_H: - func = CHOOSE_LOAD_FUNC(imm32, sk_load_half); - goto common_load; - case BPF_LD | BPF_IND | BPF_B: - func = sk_load_byte; - goto common_load; - case BPF_LD | BPF_ABS | BPF_B: - func = CHOOSE_LOAD_FUNC(imm32, sk_load_byte); - goto common_load; - case BPF_JMP | BPF_EXIT: if (seen_exit) { jmp_offset = ctx->cleanup_addr - addrs[i]; diff --git a/arch/x86/net/bpf_jit_comp32.c b/arch/x86/net/bpf_jit_comp32.c new file mode 100644 index 000000000000..0cc04e30adc1 --- /dev/null +++ b/arch/x86/net/bpf_jit_comp32.c @@ -0,0 +1,2419 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Just-In-Time compiler for eBPF filters on IA32 (32bit x86) + * + * Author: Wang YanQing (udknight@gmail.com) + * The code based on code and ideas from: + * Eric Dumazet (eric.dumazet@gmail.com) + * and from: + * Shubham Bansal <illusionist.neo@gmail.com> + */ + +#include <linux/netdevice.h> +#include <linux/filter.h> +#include <linux/if_vlan.h> +#include <asm/cacheflush.h> +#include <asm/set_memory.h> +#include <asm/nospec-branch.h> +#include <linux/bpf.h> + +/* + * eBPF prog stack layout: + * + * high + * original ESP => +-----+ + * | | callee saved registers + * +-----+ + * | ... | eBPF JIT scratch space + * BPF_FP,IA32_EBP => +-----+ + * | ... | eBPF prog stack + * +-----+ + * |RSVD | JIT scratchpad + * current ESP => +-----+ + * | | + * | ... | Function call stack + * | | + * +-----+ + * low + * + * The callee saved registers: + * + * high + * original ESP => +------------------+ \ + * | ebp | | + * current EBP => +------------------+ } callee saved registers + * | ebx,esi,edi | | + * +------------------+ / + * low + */ + +static u8 *emit_code(u8 *ptr, u32 bytes, unsigned int len) +{ + if (len == 1) + *ptr = bytes; + else if (len == 2) + *(u16 *)ptr = bytes; + else { + *(u32 *)ptr = bytes; + barrier(); + } + return ptr + len; +} + +#define EMIT(bytes, len) \ + do { prog = emit_code(prog, bytes, len); cnt += len; } while (0) + +#define EMIT1(b1) EMIT(b1, 1) +#define EMIT2(b1, b2) EMIT((b1) + ((b2) << 8), 2) +#define EMIT3(b1, b2, b3) EMIT((b1) + ((b2) << 8) + ((b3) << 16), 3) +#define EMIT4(b1, b2, b3, b4) \ + EMIT((b1) + ((b2) << 8) + ((b3) << 16) + ((b4) << 24), 4) + +#define EMIT1_off32(b1, off) \ + do { EMIT1(b1); EMIT(off, 4); } while (0) +#define EMIT2_off32(b1, b2, off) \ + do { EMIT2(b1, b2); EMIT(off, 4); } while (0) +#define EMIT3_off32(b1, b2, b3, off) \ + do { EMIT3(b1, b2, b3); EMIT(off, 4); } while (0) +#define EMIT4_off32(b1, b2, b3, b4, off) \ + do { EMIT4(b1, b2, b3, b4); EMIT(off, 4); } while (0) + +#define jmp_label(label, jmp_insn_len) (label - cnt - jmp_insn_len) + +static bool is_imm8(int value) +{ + return value <= 127 && value >= -128; +} + +static bool is_simm32(s64 value) +{ + return value == (s64) (s32) value; +} + +#define STACK_OFFSET(k) (k) +#define TCALL_CNT (MAX_BPF_JIT_REG + 0) /* Tail Call Count */ + +#define IA32_EAX (0x0) +#define IA32_EBX (0x3) +#define IA32_ECX (0x1) +#define IA32_EDX (0x2) +#define IA32_ESI (0x6) +#define IA32_EDI (0x7) +#define IA32_EBP (0x5) +#define IA32_ESP (0x4) + +/* + * List of x86 cond jumps opcodes (. + s8) + * Add 0x10 (and an extra 0x0f) to generate far jumps (. + s32) + */ +#define IA32_JB 0x72 +#define IA32_JAE 0x73 +#define IA32_JE 0x74 +#define IA32_JNE 0x75 +#define IA32_JBE 0x76 +#define IA32_JA 0x77 +#define IA32_JL 0x7C +#define IA32_JGE 0x7D +#define IA32_JLE 0x7E +#define IA32_JG 0x7F + +/* + * Map eBPF registers to IA32 32bit registers or stack scratch space. + * + * 1. All the registers, R0-R10, are mapped to scratch space on stack. + * 2. We need two 64 bit temp registers to do complex operations on eBPF + * registers. + * 3. For performance reason, the BPF_REG_AX for blinding constant, is + * mapped to real hardware register pair, IA32_ESI and IA32_EDI. + * + * As the eBPF registers are all 64 bit registers and IA32 has only 32 bit + * registers, we have to map each eBPF registers with two IA32 32 bit regs + * or scratch memory space and we have to build eBPF 64 bit register from those. + * + * We use IA32_EAX, IA32_EDX, IA32_ECX, IA32_EBX as temporary registers. + */ +static const u8 bpf2ia32[][2] = { + /* Return value from in-kernel function, and exit value from eBPF */ + [BPF_REG_0] = {STACK_OFFSET(0), STACK_OFFSET(4)}, + + /* The arguments from eBPF program to in-kernel function */ + /* Stored on stack scratch space */ + [BPF_REG_1] = {STACK_OFFSET(8), STACK_OFFSET(12)}, + [BPF_REG_2] = {STACK_OFFSET(16), STACK_OFFSET(20)}, + [BPF_REG_3] = {STACK_OFFSET(24), STACK_OFFSET(28)}, + [BPF_REG_4] = {STACK_OFFSET(32), STACK_OFFSET(36)}, + [BPF_REG_5] = {STACK_OFFSET(40), STACK_OFFSET(44)}, + + /* Callee saved registers that in-kernel function will preserve */ + /* Stored on stack scratch space */ + [BPF_REG_6] = {STACK_OFFSET(48), STACK_OFFSET(52)}, + [BPF_REG_7] = {STACK_OFFSET(56), STACK_OFFSET(60)}, + [BPF_REG_8] = {STACK_OFFSET(64), STACK_OFFSET(68)}, + [BPF_REG_9] = {STACK_OFFSET(72), STACK_OFFSET(76)}, + + /* Read only Frame Pointer to access Stack */ + [BPF_REG_FP] = {STACK_OFFSET(80), STACK_OFFSET(84)}, + + /* Temporary register for blinding constants. */ + [BPF_REG_AX] = {IA32_ESI, IA32_EDI}, + + /* Tail call count. Stored on stack scratch space. */ + [TCALL_CNT] = {STACK_OFFSET(88), STACK_OFFSET(92)}, +}; + +#define dst_lo dst[0] +#define dst_hi dst[1] +#define src_lo src[0] +#define src_hi src[1] + +#define STACK_ALIGNMENT 8 +/* + * Stack space for BPF_REG_1, BPF_REG_2, BPF_REG_3, BPF_REG_4, + * BPF_REG_5, BPF_REG_6, BPF_REG_7, BPF_REG_8, BPF_REG_9, + * BPF_REG_FP, BPF_REG_AX and Tail call counts. + */ +#define SCRATCH_SIZE 96 + +/* Total stack size used in JITed code */ +#define _STACK_SIZE (stack_depth + SCRATCH_SIZE) + +#define STACK_SIZE ALIGN(_STACK_SIZE, STACK_ALIGNMENT) + +/* Get the offset of eBPF REGISTERs stored on scratch space. */ +#define STACK_VAR(off) (off) + +/* Encode 'dst_reg' register into IA32 opcode 'byte' */ +static u8 add_1reg(u8 byte, u32 dst_reg) +{ + return byte + dst_reg; +} + +/* Encode 'dst_reg' and 'src_reg' registers into IA32 opcode 'byte' */ +static u8 add_2reg(u8 byte, u32 dst_reg, u32 src_reg) +{ + return byte + dst_reg + (src_reg << 3); +} + +static void jit_fill_hole(void *area, unsigned int size) +{ + /* Fill whole space with int3 instructions */ + memset(area, 0xcc, size); +} + +static inline void emit_ia32_mov_i(const u8 dst, const u32 val, bool dstk, + u8 **pprog) +{ + u8 *prog = *pprog; + int cnt = 0; + + if (dstk) { + if (val == 0) { + /* xor eax,eax */ + EMIT2(0x33, add_2reg(0xC0, IA32_EAX, IA32_EAX)); + /* mov dword ptr [ebp+off],eax */ + EMIT3(0x89, add_2reg(0x40, IA32_EBP, IA32_EAX), + STACK_VAR(dst)); + } else { + EMIT3_off32(0xC7, add_1reg(0x40, IA32_EBP), + STACK_VAR(dst), val); + } + } else { + if (val == 0) + EMIT2(0x33, add_2reg(0xC0, dst, dst)); + else + EMIT2_off32(0xC7, add_1reg(0xC0, dst), + val); + } + *pprog = prog; +} + +/* dst = imm (4 bytes)*/ +static inline void emit_ia32_mov_r(const u8 dst, const u8 src, bool dstk, + bool sstk, u8 **pprog) +{ + u8 *prog = *pprog; + int cnt = 0; + u8 sreg = sstk ? IA32_EAX : src; + + if (sstk) + /* mov eax,dword ptr [ebp+off] */ + EMIT3(0x8B, add_2reg(0x40, IA32_EBP, IA32_EAX), STACK_VAR(src)); + if (dstk) + /* mov dword ptr [ebp+off],eax */ + EMIT3(0x89, add_2reg(0x40, IA32_EBP, sreg), STACK_VAR(dst)); + else + /* mov dst,sreg */ + EMIT2(0x89, add_2reg(0xC0, dst, sreg)); + + *pprog = prog; +} + +/* dst = src */ +static inline void emit_ia32_mov_r64(const bool is64, const u8 dst[], + const u8 src[], bool dstk, + bool sstk, u8 **pprog) +{ + emit_ia32_mov_r(dst_lo, src_lo, dstk, sstk, pprog); + if (is64) + /* complete 8 byte move */ + emit_ia32_mov_r(dst_hi, src_hi, dstk, sstk, pprog); + else + /* zero out high 4 bytes */ + emit_ia32_mov_i(dst_hi, 0, dstk, pprog); +} + +/* Sign extended move */ +static inline void emit_ia32_mov_i64(const bool is64, const u8 dst[], + const u32 val, bool dstk, u8 **pprog) +{ + u32 hi = 0; + + if (is64 && (val & (1<<31))) + hi = (u32)~0; + emit_ia32_mov_i(dst_lo, val, dstk, pprog); + emit_ia32_mov_i(dst_hi, hi, dstk, pprog); +} + +/* + * ALU operation (32 bit) + * dst = dst * src + */ +static inline void emit_ia32_mul_r(const u8 dst, const u8 src, bool dstk, + bool sstk, u8 **pprog) +{ + u8 *prog = *pprog; + int cnt = 0; + u8 sreg = sstk ? IA32_ECX : src; + + if (sstk) + /* mov ecx,dword ptr [ebp+off] */ + EMIT3(0x8B, add_2reg(0x40, IA32_EBP, IA32_ECX), STACK_VAR(src)); + + if (dstk) + /* mov eax,dword ptr [ebp+off] */ + EMIT3(0x8B, add_2reg(0x40, IA32_EBP, IA32_EAX), STACK_VAR(dst)); + else + /* mov eax,dst */ + EMIT2(0x8B, add_2reg(0xC0, dst, IA32_EAX)); + + + EMIT2(0xF7, add_1reg(0xE0, sreg)); + + if (dstk) + /* mov dword ptr [ebp+off],eax */ + EMIT3(0x89, add_2reg(0x40, IA32_EBP, IA32_EAX), + STACK_VAR(dst)); + else + /* mov dst,eax */ + EMIT2(0x89, add_2reg(0xC0, dst, IA32_EAX)); + + *pprog = prog; +} + +static inline void emit_ia32_to_le_r64(const u8 dst[], s32 val, + bool dstk, u8 **pprog) +{ + u8 *prog = *pprog; + int cnt = 0; + u8 dreg_lo = dstk ? IA32_EAX : dst_lo; + u8 dreg_hi = dstk ? IA32_EDX : dst_hi; + + if (dstk && val != 64) { + EMIT3(0x8B, add_2reg(0x40, IA32_EBP, IA32_EAX), + STACK_VAR(dst_lo)); + EMIT3(0x8B, add_2reg(0x40, IA32_EBP, IA32_EDX), + STACK_VAR(dst_hi)); + } + switch (val) { + case 16: + /* + * Emit 'movzwl eax,ax' to zero extend 16-bit + * into 64 bit + */ + EMIT2(0x0F, 0xB7); + EMIT1(add_2reg(0xC0, dreg_lo, dreg_lo)); + /* xor dreg_hi,dreg_hi */ + EMIT2(0x33, add_2reg(0xC0, dreg_hi, dreg_hi)); + break; + case 32: + /* xor dreg_hi,dreg_hi */ + EMIT2(0x33, add_2reg(0xC0, dreg_hi, dreg_hi)); + break; + case 64: + /* nop */ + break; + } + + if (dstk && val != 64) { + /* mov dword ptr [ebp+off],dreg_lo */ + EMIT3(0x89, add_2reg(0x40, IA32_EBP, dreg_lo), + STACK_VAR(dst_lo)); + /* mov dword ptr [ebp+off],dreg_hi */ + EMIT3(0x89, add_2reg(0x40, IA32_EBP, dreg_hi), + STACK_VAR(dst_hi)); + } + *pprog = prog; +} + +static inline void emit_ia32_to_be_r64(const u8 dst[], s32 val, + bool dstk, u8 **pprog) +{ + u8 *prog = *pprog; + int cnt = 0; + u8 dreg_lo = dstk ? IA32_EAX : dst_lo; + u8 dreg_hi = dstk ? IA32_EDX : dst_hi; + + if (dstk) { + EMIT3(0x8B, add_2reg(0x40, IA32_EBP, IA32_EAX), + STACK_VAR(dst_lo)); + EMIT3(0x8B, add_2reg(0x40, IA32_EBP, IA32_EDX), + STACK_VAR(dst_hi)); + } + switch (val) { + case 16: + /* Emit 'ror %ax, 8' to swap lower 2 bytes */ + EMIT1(0x66); + EMIT3(0xC1, add_1reg(0xC8, dreg_lo), 8); + + EMIT2(0x0F, 0xB7); + EMIT1(add_2reg(0xC0, dreg_lo, dreg_lo)); + + /* xor dreg_hi,dreg_hi */ + EMIT2(0x33, add_2reg(0xC0, dreg_hi, dreg_hi)); + break; + case 32: + /* Emit 'bswap eax' to swap lower 4 bytes */ + EMIT1(0x0F); + EMIT1(add_1reg(0xC8, dreg_lo)); + + /* xor dreg_hi,dreg_hi */ + EMIT2(0x33, add_2reg(0xC0, dreg_hi, dreg_hi)); + break; + case 64: + /* Emit 'bswap eax' to swap lower 4 bytes */ + EMIT1(0x0F); + EMIT1(add_1reg(0xC8, dreg_lo)); + + /* Emit 'bswap edx' to swap lower 4 bytes */ + EMIT1(0x0F); + EMIT1(add_1reg(0xC8, dreg_hi)); + + /* mov ecx,dreg_hi */ + EMIT2(0x89, add_2reg(0xC0, IA32_ECX, dreg_hi)); + /* mov dreg_hi,dreg_lo */ + EMIT2(0x89, add_2reg(0xC0, dreg_hi, dreg_lo)); + /* mov dreg_lo,ecx */ + EMIT2(0x89, add_2reg(0xC0, dreg_lo, IA32_ECX)); + + break; + } + if (dstk) { + /* mov dword ptr [ebp+off],dreg_lo */ + EMIT3(0x89, add_2reg(0x40, IA32_EBP, dreg_lo), + STACK_VAR(dst_lo)); + /* mov dword ptr [ebp+off],dreg_hi */ + EMIT3(0x89, add_2reg(0x40, IA32_EBP, dreg_hi), + STACK_VAR(dst_hi)); + } + *pprog = prog; +} + +/* + * ALU operation (32 bit) + * dst = dst (div|mod) src + */ +static inline void emit_ia32_div_mod_r(const u8 op, const u8 dst, const u8 src, + bool dstk, bool sstk, u8 **pprog) +{ + u8 *prog = *pprog; + int cnt = 0; + + if (sstk) + /* mov ecx,dword ptr [ebp+off] */ + EMIT3(0x8B, add_2reg(0x40, IA32_EBP, IA32_ECX), + STACK_VAR(src)); + else if (src != IA32_ECX) + /* mov ecx,src */ + EMIT2(0x8B, add_2reg(0xC0, src, IA32_ECX)); + + if (dstk) + /* mov eax,dword ptr [ebp+off] */ + EMIT3(0x8B, add_2reg(0x40, IA32_EBP, IA32_EAX), + STACK_VAR(dst)); + else + /* mov eax,dst */ + EMIT2(0x8B, add_2reg(0xC0, dst, IA32_EAX)); + + /* xor edx,edx */ + EMIT2(0x31, add_2reg(0xC0, IA32_EDX, IA32_EDX)); + /* div ecx */ + EMIT2(0xF7, add_1reg(0xF0, IA32_ECX)); + + if (op == BPF_MOD) { + if (dstk) + EMIT3(0x89, add_2reg(0x40, IA32_EBP, IA32_EDX), + STACK_VAR(dst)); + else + EMIT2(0x89, add_2reg(0xC0, dst, IA32_EDX)); + } else { + if (dstk) + EMIT3(0x89, add_2reg(0x40, IA32_EBP, IA32_EAX), + STACK_VAR(dst)); + else + EMIT2(0x89, add_2reg(0xC0, dst, IA32_EAX)); + } + *pprog = prog; +} + +/* + * ALU operation (32 bit) + * dst = dst (shift) src + */ +static inline void emit_ia32_shift_r(const u8 op, const u8 dst, const u8 src, + bool dstk, bool sstk, u8 **pprog) +{ + u8 *prog = *pprog; + int cnt = 0; + u8 dreg = dstk ? IA32_EAX : dst; + u8 b2; + + if (dstk) + /* mov eax,dword ptr [ebp+off] */ + EMIT3(0x8B, add_2reg(0x40, IA32_EBP, IA32_EAX), STACK_VAR(dst)); + + if (sstk) + /* mov ecx,dword ptr [ebp+off] */ + EMIT3(0x8B, add_2reg(0x40, IA32_EBP, IA32_ECX), STACK_VAR(src)); + else if (src != IA32_ECX) + /* mov ecx,src */ + EMIT2(0x8B, add_2reg(0xC0, src, IA32_ECX)); + + switch (op) { + case BPF_LSH: + b2 = 0xE0; break; + case BPF_RSH: + b2 = 0xE8; break; + case BPF_ARSH: + b2 = 0xF8; break; + default: + return; + } + EMIT2(0xD3, add_1reg(b2, dreg)); + + if (dstk) + /* mov dword ptr [ebp+off],dreg */ + EMIT3(0x89, add_2reg(0x40, IA32_EBP, dreg), STACK_VAR(dst)); + *pprog = prog; +} + +/* + * ALU operation (32 bit) + * dst = dst (op) src + */ +static inline void emit_ia32_alu_r(const bool is64, const bool hi, const u8 op, + const u8 dst, const u8 src, bool dstk, + bool sstk, u8 **pprog) +{ + u8 *prog = *pprog; + int cnt = 0; + u8 sreg = sstk ? IA32_EAX : src; + u8 dreg = dstk ? IA32_EDX : dst; + + if (sstk) + /* mov eax,dword ptr [ebp+off] */ + EMIT3(0x8B, add_2reg(0x40, IA32_EBP, IA32_EAX), STACK_VAR(src)); + + if (dstk) + /* mov eax,dword ptr [ebp+off] */ + EMIT3(0x8B, add_2reg(0x40, IA32_EBP, IA32_EDX), STACK_VAR(dst)); + + switch (BPF_OP(op)) { + /* dst = dst + src */ + case BPF_ADD: + if (hi && is64) + EMIT2(0x11, add_2reg(0xC0, dreg, sreg)); + else + EMIT2(0x01, add_2reg(0xC0, dreg, sreg)); + break; + /* dst = dst - src */ + case BPF_SUB: + if (hi && is64) + EMIT2(0x19, add_2reg(0xC0, dreg, sreg)); + else + EMIT2(0x29, add_2reg(0xC0, dreg, sreg)); + break; + /* dst = dst | src */ + case BPF_OR: + EMIT2(0x09, add_2reg(0xC0, dreg, sreg)); + break; + /* dst = dst & src */ + case BPF_AND: + EMIT2(0x21, add_2reg(0xC0, dreg, sreg)); + break; + /* dst = dst ^ src */ + case BPF_XOR: + EMIT2(0x31, add_2reg(0xC0, dreg, sreg)); + break; + } + + if (dstk) + /* mov dword ptr [ebp+off],dreg */ + EMIT3(0x89, add_2reg(0x40, IA32_EBP, dreg), + STACK_VAR(dst)); + *pprog = prog; +} + +/* ALU operation (64 bit) */ +static inline void emit_ia32_alu_r64(const bool is64, const u8 op, + const u8 dst[], const u8 src[], + bool dstk, bool sstk, + u8 **pprog) +{ + u8 *prog = *pprog; + + emit_ia32_alu_r(is64, false, op, dst_lo, src_lo, dstk, sstk, &prog); + if (is64) + emit_ia32_alu_r(is64, true, op, dst_hi, src_hi, dstk, sstk, + &prog); + else + emit_ia32_mov_i(dst_hi, 0, dstk, &prog); + *pprog = prog; +} + +/* + * ALU operation (32 bit) + * dst = dst (op) val + */ +static inline void emit_ia32_alu_i(const bool is64, const bool hi, const u8 op, + const u8 dst, const s32 val, bool dstk, + u8 **pprog) +{ + u8 *prog = *pprog; + int cnt = 0; + u8 dreg = dstk ? IA32_EAX : dst; + u8 sreg = IA32_EDX; + + if (dstk) + /* mov eax,dword ptr [ebp+off] */ + EMIT3(0x8B, add_2reg(0x40, IA32_EBP, IA32_EAX), STACK_VAR(dst)); + + if (!is_imm8(val)) + /* mov edx,imm32*/ + EMIT2_off32(0xC7, add_1reg(0xC0, IA32_EDX), val); + + switch (op) { + /* dst = dst + val */ + case BPF_ADD: + if (hi && is64) { + if (is_imm8(val)) + EMIT3(0x83, add_1reg(0xD0, dreg), val); + else + EMIT2(0x11, add_2reg(0xC0, dreg, sreg)); + } else { + if (is_imm8(val)) + EMIT3(0x83, add_1reg(0xC0, dreg), val); + else + EMIT2(0x01, add_2reg(0xC0, dreg, sreg)); + } + break; + /* dst = dst - val */ + case BPF_SUB: + if (hi && is64) { + if (is_imm8(val)) + EMIT3(0x83, add_1reg(0xD8, dreg), val); + else + EMIT2(0x19, add_2reg(0xC0, dreg, sreg)); + } else { + if (is_imm8(val)) + EMIT3(0x83, add_1reg(0xE8, dreg), val); + else + EMIT2(0x29, add_2reg(0xC0, dreg, sreg)); + } + break; + /* dst = dst | val */ + case BPF_OR: + if (is_imm8(val)) + EMIT3(0x83, add_1reg(0xC8, dreg), val); + else + EMIT2(0x09, add_2reg(0xC0, dreg, sreg)); + break; + /* dst = dst & val */ + case BPF_AND: + if (is_imm8(val)) + EMIT3(0x83, add_1reg(0xE0, dreg), val); + else + EMIT2(0x21, add_2reg(0xC0, dreg, sreg)); + break; + /* dst = dst ^ val */ + case BPF_XOR: + if (is_imm8(val)) + EMIT3(0x83, add_1reg(0xF0, dreg), val); + else + EMIT2(0x31, add_2reg(0xC0, dreg, sreg)); + break; + case BPF_NEG: + EMIT2(0xF7, add_1reg(0xD8, dreg)); + break; + } + + if (dstk) + /* mov dword ptr [ebp+off],dreg */ + EMIT3(0x89, add_2reg(0x40, IA32_EBP, dreg), + STACK_VAR(dst)); + *pprog = prog; +} + +/* ALU operation (64 bit) */ +static inline void emit_ia32_alu_i64(const bool is64, const u8 op, + const u8 dst[], const u32 val, + bool dstk, u8 **pprog) +{ + u8 *prog = *pprog; + u32 hi = 0; + + if (is64 && (val & (1<<31))) + hi = (u32)~0; + + emit_ia32_alu_i(is64, false, op, dst_lo, val, dstk, &prog); + if (is64) + emit_ia32_alu_i(is64, true, op, dst_hi, hi, dstk, &prog); + else + emit_ia32_mov_i(dst_hi, 0, dstk, &prog); + + *pprog = prog; +} + +/* dst = ~dst (64 bit) */ +static inline void emit_ia32_neg64(const u8 dst[], bool dstk, u8 **pprog) +{ + u8 *prog = *pprog; + int cnt = 0; + u8 dreg_lo = dstk ? IA32_EAX : dst_lo; + u8 dreg_hi = dstk ? IA32_EDX : dst_hi; + + if (dstk) { + EMIT3(0x8B, add_2reg(0x40, IA32_EBP, IA32_EAX), + STACK_VAR(dst_lo)); + EMIT3(0x8B, add_2reg(0x40, IA32_EBP, IA32_EDX), + STACK_VAR(dst_hi)); + } + + /* xor ecx,ecx */ + EMIT2(0x31, add_2reg(0xC0, IA32_ECX, IA32_ECX)); + /* sub dreg_lo,ecx */ + EMIT2(0x2B, add_2reg(0xC0, dreg_lo, IA32_ECX)); + /* mov dreg_lo,ecx */ + EMIT2(0x89, add_2reg(0xC0, dreg_lo, IA32_ECX)); + + /* xor ecx,ecx */ + EMIT2(0x31, add_2reg(0xC0, IA32_ECX, IA32_ECX)); + /* sbb dreg_hi,ecx */ + EMIT2(0x19, add_2reg(0xC0, dreg_hi, IA32_ECX)); + /* mov dreg_hi,ecx */ + EMIT2(0x89, add_2reg(0xC0, dreg_hi, IA32_ECX)); + + if (dstk) { + /* mov dword ptr [ebp+off],dreg_lo */ + EMIT3(0x89, add_2reg(0x40, IA32_EBP, dreg_lo), + STACK_VAR(dst_lo)); + /* mov dword ptr [ebp+off],dreg_hi */ + EMIT3(0x89, add_2reg(0x40, IA32_EBP, dreg_hi), + STACK_VAR(dst_hi)); + } + *pprog = prog; +} + +/* dst = dst << src */ +static inline void emit_ia32_lsh_r64(const u8 dst[], const u8 src[], + bool dstk, bool sstk, u8 **pprog) +{ + u8 *prog = *pprog; + int cnt = 0; + static int jmp_label1 = -1; + static int jmp_label2 = -1; + static int jmp_label3 = -1; + u8 dreg_lo = dstk ? IA32_EAX : dst_lo; + u8 dreg_hi = dstk ? IA32_EDX : dst_hi; + + if (dstk) { + EMIT3(0x8B, add_2reg(0x40, IA32_EBP, IA32_EAX), + STACK_VAR(dst_lo)); + EMIT3(0x8B, add_2reg(0x40, IA32_EBP, IA32_EDX), + STACK_VAR(dst_hi)); + } + + if (sstk) + /* mov ecx,dword ptr [ebp+off] */ + EMIT3(0x8B, add_2reg(0x40, IA32_EBP, IA32_ECX), + STACK_VAR(src_lo)); + else + /* mov ecx,src_lo */ + EMIT2(0x8B, add_2reg(0xC0, src_lo, IA32_ECX)); + + /* cmp ecx,32 */ + EMIT3(0x83, add_1reg(0xF8, IA32_ECX), 32); + /* Jumps when >= 32 */ + if (is_imm8(jmp_label(jmp_label1, 2))) + EMIT2(IA32_JAE, jmp_label(jmp_label1, 2)); + else + EMIT2_off32(0x0F, IA32_JAE + 0x10, jmp_label(jmp_label1, 6)); + + /* < 32 */ + /* shl dreg_hi,cl */ + EMIT2(0xD3, add_1reg(0xE0, dreg_hi)); + /* mov ebx,dreg_lo */ + EMIT2(0x8B, add_2reg(0xC0, dreg_lo, IA32_EBX)); + /* shl dreg_lo,cl */ + EMIT2(0xD3, add_1reg(0xE0, dreg_lo)); + + /* IA32_ECX = -IA32_ECX + 32 */ + /* neg ecx */ + EMIT2(0xF7, add_1reg(0xD8, IA32_ECX)); + /* add ecx,32 */ + EMIT3(0x83, add_1reg(0xC0, IA32_ECX), 32); + + /* shr ebx,cl */ + EMIT2(0xD3, add_1reg(0xE8, IA32_EBX)); + /* or dreg_hi,ebx */ + EMIT2(0x09, add_2reg(0xC0, dreg_hi, IA32_EBX)); + + /* goto out; */ + if (is_imm8(jmp_label(jmp_label3, 2))) + EMIT2(0xEB, jmp_label(jmp_label3, 2)); + else + EMIT1_off32(0xE9, jmp_label(jmp_label3, 5)); + + /* >= 32 */ + if (jmp_label1 == -1) + jmp_label1 = cnt; + + /* cmp ecx,64 */ + EMIT3(0x83, add_1reg(0xF8, IA32_ECX), 64); + /* Jumps when >= 64 */ + if (is_imm8(jmp_label(jmp_label2, 2))) + EMIT2(IA32_JAE, jmp_label(jmp_label2, 2)); + else + EMIT2_off32(0x0F, IA32_JAE + 0x10, jmp_label(jmp_label2, 6)); + + /* >= 32 && < 64 */ + /* sub ecx,32 */ + EMIT3(0x83, add_1reg(0xE8, IA32_ECX), 32); + /* shl dreg_lo,cl */ + EMIT2(0xD3, add_1reg(0xE0, dreg_lo)); + /* mov dreg_hi,dreg_lo */ + EMIT2(0x89, add_2reg(0xC0, dreg_hi, dreg_lo)); + + /* xor dreg_lo,dreg_lo */ + EMIT2(0x33, add_2reg(0xC0, dreg_lo, dreg_lo)); + + /* goto out; */ + if (is_imm8(jmp_label(jmp_label3, 2))) + EMIT2(0xEB, jmp_label(jmp_label3, 2)); + else + EMIT1_off32(0xE9, jmp_label(jmp_label3, 5)); + + /* >= 64 */ + if (jmp_label2 == -1) + jmp_label2 = cnt; + /* xor dreg_lo,dreg_lo */ + EMIT2(0x33, add_2reg(0xC0, dreg_lo, dreg_lo)); + /* xor dreg_hi,dreg_hi */ + EMIT2(0x33, add_2reg(0xC0, dreg_hi, dreg_hi)); + + if (jmp_label3 == -1) + jmp_label3 = cnt; + + if (dstk) { + /* mov dword ptr [ebp+off],dreg_lo */ + EMIT3(0x89, add_2reg(0x40, IA32_EBP, dreg_lo), + STACK_VAR(dst_lo)); + /* mov dword ptr [ebp+off],dreg_hi */ + EMIT3(0x89, add_2reg(0x40, IA32_EBP, dreg_hi), + STACK_VAR(dst_hi)); + } + /* out: */ + *pprog = prog; +} + +/* dst = dst >> src (signed)*/ +static inline void emit_ia32_arsh_r64(const u8 dst[], const u8 src[], + bool dstk, bool sstk, u8 **pprog) +{ + u8 *prog = *pprog; + int cnt = 0; + static int jmp_label1 = -1; + static int jmp_label2 = -1; + static int jmp_label3 = -1; + u8 dreg_lo = dstk ? IA32_EAX : dst_lo; + u8 dreg_hi = dstk ? IA32_EDX : dst_hi; + + if (dstk) { + EMIT3(0x8B, add_2reg(0x40, IA32_EBP, IA32_EAX), + STACK_VAR(dst_lo)); + EMIT3(0x8B, add_2reg(0x40, IA32_EBP, IA32_EDX), + STACK_VAR(dst_hi)); + } + + if (sstk) + /* mov ecx,dword ptr [ebp+off] */ + EMIT3(0x8B, add_2reg(0x40, IA32_EBP, IA32_ECX), + STACK_VAR(src_lo)); + else + /* mov ecx,src_lo */ + EMIT2(0x8B, add_2reg(0xC0, src_lo, IA32_ECX)); + + /* cmp ecx,32 */ + EMIT3(0x83, add_1reg(0xF8, IA32_ECX), 32); + /* Jumps when >= 32 */ + if (is_imm8(jmp_label(jmp_label1, 2))) + EMIT2(IA32_JAE, jmp_label(jmp_label1, 2)); + else + EMIT2_off32(0x0F, IA32_JAE + 0x10, jmp_label(jmp_label1, 6)); + + /* < 32 */ + /* lshr dreg_lo,cl */ + EMIT2(0xD3, add_1reg(0xE8, dreg_lo)); + /* mov ebx,dreg_hi */ + EMIT2(0x8B, add_2reg(0xC0, dreg_hi, IA32_EBX)); + /* ashr dreg_hi,cl */ + EMIT2(0xD3, add_1reg(0xF8, dreg_hi)); + + /* IA32_ECX = -IA32_ECX + 32 */ + /* neg ecx */ + EMIT2(0xF7, add_1reg(0xD8, IA32_ECX)); + /* add ecx,32 */ + EMIT3(0x83, add_1reg(0xC0, IA32_ECX), 32); + + /* shl ebx,cl */ + EMIT2(0xD3, add_1reg(0xE0, IA32_EBX)); + /* or dreg_lo,ebx */ + EMIT2(0x09, add_2reg(0xC0, dreg_lo, IA32_EBX)); + + /* goto out; */ + if (is_imm8(jmp_label(jmp_label3, 2))) + EMIT2(0xEB, jmp_label(jmp_label3, 2)); + else + EMIT1_off32(0xE9, jmp_label(jmp_label3, 5)); + + /* >= 32 */ + if (jmp_label1 == -1) + jmp_label1 = cnt; + + /* cmp ecx,64 */ + EMIT3(0x83, add_1reg(0xF8, IA32_ECX), 64); + /* Jumps when >= 64 */ + if (is_imm8(jmp_label(jmp_label2, 2))) + EMIT2(IA32_JAE, jmp_label(jmp_label2, 2)); + else + EMIT2_off32(0x0F, IA32_JAE + 0x10, jmp_label(jmp_label2, 6)); + + /* >= 32 && < 64 */ + /* sub ecx,32 */ + EMIT3(0x83, add_1reg(0xE8, IA32_ECX), 32); + /* ashr dreg_hi,cl */ + EMIT2(0xD3, add_1reg(0xF8, dreg_hi)); + /* mov dreg_lo,dreg_hi */ + EMIT2(0x89, add_2reg(0xC0, dreg_lo, dreg_hi)); + + /* ashr dreg_hi,imm8 */ + EMIT3(0xC1, add_1reg(0xF8, dreg_hi), 31); + + /* goto out; */ + if (is_imm8(jmp_label(jmp_label3, 2))) + EMIT2(0xEB, jmp_label(jmp_label3, 2)); + else + EMIT1_off32(0xE9, jmp_label(jmp_label3, 5)); + + /* >= 64 */ + if (jmp_label2 == -1) + jmp_label2 = cnt; + /* ashr dreg_hi,imm8 */ + EMIT3(0xC1, add_1reg(0xF8, dreg_hi), 31); + /* mov dreg_lo,dreg_hi */ + EMIT2(0x89, add_2reg(0xC0, dreg_lo, dreg_hi)); + + if (jmp_label3 == -1) + jmp_label3 = cnt; + + if (dstk) { + /* mov dword ptr [ebp+off],dreg_lo */ + EMIT3(0x89, add_2reg(0x40, IA32_EBP, dreg_lo), + STACK_VAR(dst_lo)); + /* mov dword ptr [ebp+off],dreg_hi */ + EMIT3(0x89, add_2reg(0x40, IA32_EBP, dreg_hi), + STACK_VAR(dst_hi)); + } + /* out: */ + *pprog = prog; +} + +/* dst = dst >> src */ +static inline void emit_ia32_rsh_r64(const u8 dst[], const u8 src[], bool dstk, + bool sstk, u8 **pprog) +{ + u8 *prog = *pprog; + int cnt = 0; + static int jmp_label1 = -1; + static int jmp_label2 = -1; + static int jmp_label3 = -1; + u8 dreg_lo = dstk ? IA32_EAX : dst_lo; + u8 dreg_hi = dstk ? IA32_EDX : dst_hi; + + if (dstk) { + EMIT3(0x8B, add_2reg(0x40, IA32_EBP, IA32_EAX), + STACK_VAR(dst_lo)); + EMIT3(0x8B, add_2reg(0x40, IA32_EBP, IA32_EDX), + STACK_VAR(dst_hi)); + } + + if (sstk) + /* mov ecx,dword ptr [ebp+off] */ + EMIT3(0x8B, add_2reg(0x40, IA32_EBP, IA32_ECX), + STACK_VAR(src_lo)); + else + /* mov ecx,src_lo */ + EMIT2(0x8B, add_2reg(0xC0, src_lo, IA32_ECX)); + + /* cmp ecx,32 */ + EMIT3(0x83, add_1reg(0xF8, IA32_ECX), 32); + /* Jumps when >= 32 */ + if (is_imm8(jmp_label(jmp_label1, 2))) + EMIT2(IA32_JAE, jmp_label(jmp_label1, 2)); + else + EMIT2_off32(0x0F, IA32_JAE + 0x10, jmp_label(jmp_label1, 6)); + + /* < 32 */ + /* lshr dreg_lo,cl */ + EMIT2(0xD3, add_1reg(0xE8, dreg_lo)); + /* mov ebx,dreg_hi */ + EMIT2(0x8B, add_2reg(0xC0, dreg_hi, IA32_EBX)); + /* shr dreg_hi,cl */ + EMIT2(0xD3, add_1reg(0xE8, dreg_hi)); + + /* IA32_ECX = -IA32_ECX + 32 */ + /* neg ecx */ + EMIT2(0xF7, add_1reg(0xD8, IA32_ECX)); + /* add ecx,32 */ + EMIT3(0x83, add_1reg(0xC0, IA32_ECX), 32); + + /* shl ebx,cl */ + EMIT2(0xD3, add_1reg(0xE0, IA32_EBX)); + /* or dreg_lo,ebx */ + EMIT2(0x09, add_2reg(0xC0, dreg_lo, IA32_EBX)); + + /* goto out; */ + if (is_imm8(jmp_label(jmp_label3, 2))) + EMIT2(0xEB, jmp_label(jmp_label3, 2)); + else + EMIT1_off32(0xE9, jmp_label(jmp_label3, 5)); + + /* >= 32 */ + if (jmp_label1 == -1) + jmp_label1 = cnt; + /* cmp ecx,64 */ + EMIT3(0x83, add_1reg(0xF8, IA32_ECX), 64); + /* Jumps when >= 64 */ + if (is_imm8(jmp_label(jmp_label2, 2))) + EMIT2(IA32_JAE, jmp_label(jmp_label2, 2)); + else + EMIT2_off32(0x0F, IA32_JAE + 0x10, jmp_label(jmp_label2, 6)); + + /* >= 32 && < 64 */ + /* sub ecx,32 */ + EMIT3(0x83, add_1reg(0xE8, IA32_ECX), 32); + /* shr dreg_hi,cl */ + EMIT2(0xD3, add_1reg(0xE8, dreg_hi)); + /* mov dreg_lo,dreg_hi */ + EMIT2(0x89, add_2reg(0xC0, dreg_lo, dreg_hi)); + /* xor dreg_hi,dreg_hi */ + EMIT2(0x33, add_2reg(0xC0, dreg_hi, dreg_hi)); + + /* goto out; */ + if (is_imm8(jmp_label(jmp_label3, 2))) + EMIT2(0xEB, jmp_label(jmp_label3, 2)); + else + EMIT1_off32(0xE9, jmp_label(jmp_label3, 5)); + + /* >= 64 */ + if (jmp_label2 == -1) + jmp_label2 = cnt; + /* xor dreg_lo,dreg_lo */ + EMIT2(0x33, add_2reg(0xC0, dreg_lo, dreg_lo)); + /* xor dreg_hi,dreg_hi */ + EMIT2(0x33, add_2reg(0xC0, dreg_hi, dreg_hi)); + + if (jmp_label3 == -1) + jmp_label3 = cnt; + + if (dstk) { + /* mov dword ptr [ebp+off],dreg_lo */ + EMIT3(0x89, add_2reg(0x40, IA32_EBP, dreg_lo), + STACK_VAR(dst_lo)); + /* mov dword ptr [ebp+off],dreg_hi */ + EMIT3(0x89, add_2reg(0x40, IA32_EBP, dreg_hi), + STACK_VAR(dst_hi)); + } + /* out: */ + *pprog = prog; +} + +/* dst = dst << val */ +static inline void emit_ia32_lsh_i64(const u8 dst[], const u32 val, + bool dstk, u8 **pprog) +{ + u8 *prog = *pprog; + int cnt = 0; + u8 dreg_lo = dstk ? IA32_EAX : dst_lo; + u8 dreg_hi = dstk ? IA32_EDX : dst_hi; + + if (dstk) { + EMIT3(0x8B, add_2reg(0x40, IA32_EBP, IA32_EAX), + STACK_VAR(dst_lo)); + EMIT3(0x8B, add_2reg(0x40, IA32_EBP, IA32_EDX), + STACK_VAR(dst_hi)); + } + /* Do LSH operation */ + if (val < 32) { + /* shl dreg_hi,imm8 */ + EMIT3(0xC1, add_1reg(0xE0, dreg_hi), val); + /* mov ebx,dreg_lo */ + EMIT2(0x8B, add_2reg(0xC0, dreg_lo, IA32_EBX)); + /* shl dreg_lo,imm8 */ + EMIT3(0xC1, add_1reg(0xE0, dreg_lo), val); + + /* IA32_ECX = 32 - val */ + /* mov ecx,val */ + EMIT2(0xB1, val); + /* movzx ecx,ecx */ + EMIT3(0x0F, 0xB6, add_2reg(0xC0, IA32_ECX, IA32_ECX)); + /* neg ecx */ + EMIT2(0xF7, add_1reg(0xD8, IA32_ECX)); + /* add ecx,32 */ + EMIT3(0x83, add_1reg(0xC0, IA32_ECX), 32); + + /* shr ebx,cl */ + EMIT2(0xD3, add_1reg(0xE8, IA32_EBX)); + /* or dreg_hi,ebx */ + EMIT2(0x09, add_2reg(0xC0, dreg_hi, IA32_EBX)); + } else if (val >= 32 && val < 64) { + u32 value = val - 32; + + /* shl dreg_lo,imm8 */ + EMIT3(0xC1, add_1reg(0xE0, dreg_lo), value); + /* mov dreg_hi,dreg_lo */ + EMIT2(0x89, add_2reg(0xC0, dreg_hi, dreg_lo)); + /* xor dreg_lo,dreg_lo */ + EMIT2(0x33, add_2reg(0xC0, dreg_lo, dreg_lo)); + } else { + /* xor dreg_lo,dreg_lo */ + EMIT2(0x33, add_2reg(0xC0, dreg_lo, dreg_lo)); + /* xor dreg_hi,dreg_hi */ + EMIT2(0x33, add_2reg(0xC0, dreg_hi, dreg_hi)); + } + + if (dstk) { + /* mov dword ptr [ebp+off],dreg_lo */ + EMIT3(0x89, add_2reg(0x40, IA32_EBP, dreg_lo), + STACK_VAR(dst_lo)); + /* mov dword ptr [ebp+off],dreg_hi */ + EMIT3(0x89, add_2reg(0x40, IA32_EBP, dreg_hi), + STACK_VAR(dst_hi)); + } + *pprog = prog; +} + +/* dst = dst >> val */ +static inline void emit_ia32_rsh_i64(const u8 dst[], const u32 val, + bool dstk, u8 **pprog) +{ + u8 *prog = *pprog; + int cnt = 0; + u8 dreg_lo = dstk ? IA32_EAX : dst_lo; + u8 dreg_hi = dstk ? IA32_EDX : dst_hi; + + if (dstk) { + EMIT3(0x8B, add_2reg(0x40, IA32_EBP, IA32_EAX), + STACK_VAR(dst_lo)); + EMIT3(0x8B, add_2reg(0x40, IA32_EBP, IA32_EDX), + STACK_VAR(dst_hi)); + } + + /* Do RSH operation */ + if (val < 32) { + /* shr dreg_lo,imm8 */ + EMIT3(0xC1, add_1reg(0xE8, dreg_lo), val); + /* mov ebx,dreg_hi */ + EMIT2(0x8B, add_2reg(0xC0, dreg_hi, IA32_EBX)); + /* shr dreg_hi,imm8 */ + EMIT3(0xC1, add_1reg(0xE8, dreg_hi), val); + + /* IA32_ECX = 32 - val */ + /* mov ecx,val */ + EMIT2(0xB1, val); + /* movzx ecx,ecx */ + EMIT3(0x0F, 0xB6, add_2reg(0xC0, IA32_ECX, IA32_ECX)); + /* neg ecx */ + EMIT2(0xF7, add_1reg(0xD8, IA32_ECX)); + /* add ecx,32 */ + EMIT3(0x83, add_1reg(0xC0, IA32_ECX), 32); + + /* shl ebx,cl */ + EMIT2(0xD3, add_1reg(0xE0, IA32_EBX)); + /* or dreg_lo,ebx */ + EMIT2(0x09, add_2reg(0xC0, dreg_lo, IA32_EBX)); + } else if (val >= 32 && val < 64) { + u32 value = val - 32; + + /* shr dreg_hi,imm8 */ + EMIT3(0xC1, add_1reg(0xE8, dreg_hi), value); + /* mov dreg_lo,dreg_hi */ + EMIT2(0x89, add_2reg(0xC0, dreg_lo, dreg_hi)); + /* xor dreg_hi,dreg_hi */ + EMIT2(0x33, add_2reg(0xC0, dreg_hi, dreg_hi)); + } else { + /* xor dreg_lo,dreg_lo */ + EMIT2(0x33, add_2reg(0xC0, dreg_lo, dreg_lo)); + /* xor dreg_hi,dreg_hi */ + EMIT2(0x33, add_2reg(0xC0, dreg_hi, dreg_hi)); + } + + if (dstk) { + /* mov dword ptr [ebp+off],dreg_lo */ + EMIT3(0x89, add_2reg(0x40, IA32_EBP, dreg_lo), + STACK_VAR(dst_lo)); + /* mov dword ptr [ebp+off],dreg_hi */ + EMIT3(0x89, add_2reg(0x40, IA32_EBP, dreg_hi), + STACK_VAR(dst_hi)); + } + *pprog = prog; +} + +/* dst = dst >> val (signed) */ +static inline void emit_ia32_arsh_i64(const u8 dst[], const u32 val, + bool dstk, u8 **pprog) +{ + u8 *prog = *pprog; + int cnt = 0; + u8 dreg_lo = dstk ? IA32_EAX : dst_lo; + u8 dreg_hi = dstk ? IA32_EDX : dst_hi; + + if (dstk) { + EMIT3(0x8B, add_2reg(0x40, IA32_EBP, IA32_EAX), + STACK_VAR(dst_lo)); + EMIT3(0x8B, add_2reg(0x40, IA32_EBP, IA32_EDX), + STACK_VAR(dst_hi)); + } + /* Do RSH operation */ + if (val < 32) { + /* shr dreg_lo,imm8 */ + EMIT3(0xC1, add_1reg(0xE8, dreg_lo), val); + /* mov ebx,dreg_hi */ + EMIT2(0x8B, add_2reg(0xC0, dreg_hi, IA32_EBX)); + /* ashr dreg_hi,imm8 */ + EMIT3(0xC1, add_1reg(0xF8, dreg_hi), val); + + /* IA32_ECX = 32 - val */ + /* mov ecx,val */ + EMIT2(0xB1, val); + /* movzx ecx,ecx */ + EMIT3(0x0F, 0xB6, add_2reg(0xC0, IA32_ECX, IA32_ECX)); + /* neg ecx */ + EMIT2(0xF7, add_1reg(0xD8, IA32_ECX)); + /* add ecx,32 */ + EMIT3(0x83, add_1reg(0xC0, IA32_ECX), 32); + + /* shl ebx,cl */ + EMIT2(0xD3, add_1reg(0xE0, IA32_EBX)); + /* or dreg_lo,ebx */ + EMIT2(0x09, add_2reg(0xC0, dreg_lo, IA32_EBX)); + } else if (val >= 32 && val < 64) { + u32 value = val - 32; + + /* ashr dreg_hi,imm8 */ + EMIT3(0xC1, add_1reg(0xF8, dreg_hi), value); + /* mov dreg_lo,dreg_hi */ + EMIT2(0x89, add_2reg(0xC0, dreg_lo, dreg_hi)); + + /* ashr dreg_hi,imm8 */ + EMIT3(0xC1, add_1reg(0xF8, dreg_hi), 31); + } else { + /* ashr dreg_hi,imm8 */ + EMIT3(0xC1, add_1reg(0xF8, dreg_hi), 31); + /* mov dreg_lo,dreg_hi */ + EMIT2(0x89, add_2reg(0xC0, dreg_lo, dreg_hi)); + } + + if (dstk) { + /* mov dword ptr [ebp+off],dreg_lo */ + EMIT3(0x89, add_2reg(0x40, IA32_EBP, dreg_lo), + STACK_VAR(dst_lo)); + /* mov dword ptr [ebp+off],dreg_hi */ + EMIT3(0x89, add_2reg(0x40, IA32_EBP, dreg_hi), + STACK_VAR(dst_hi)); + } + *pprog = prog; +} + +static inline void emit_ia32_mul_r64(const u8 dst[], const u8 src[], bool dstk, + bool sstk, u8 **pprog) +{ + u8 *prog = *pprog; + int cnt = 0; + + if (dstk) + /* mov eax,dword ptr [ebp+off] */ + EMIT3(0x8B, add_2reg(0x40, IA32_EBP, IA32_EAX), + STACK_VAR(dst_hi)); + else + /* mov eax,dst_hi */ + EMIT2(0x8B, add_2reg(0xC0, dst_hi, IA32_EAX)); + + if (sstk) + /* mul dword ptr [ebp+off] */ + EMIT3(0xF7, add_1reg(0x60, IA32_EBP), STACK_VAR(src_lo)); + else + /* mul src_lo */ + EMIT2(0xF7, add_1reg(0xE0, src_lo)); + + /* mov ecx,eax */ + EMIT2(0x89, add_2reg(0xC0, IA32_ECX, IA32_EAX)); + + if (dstk) + /* mov eax,dword ptr [ebp+off] */ + EMIT3(0x8B, add_2reg(0x40, IA32_EBP, IA32_EAX), + STACK_VAR(dst_lo)); + else + /* mov eax,dst_lo */ + EMIT2(0x8B, add_2reg(0xC0, dst_lo, IA32_EAX)); + + if (sstk) + /* mul dword ptr [ebp+off] */ + EMIT3(0xF7, add_1reg(0x60, IA32_EBP), STACK_VAR(src_hi)); + else + /* mul src_hi */ + EMIT2(0xF7, add_1reg(0xE0, src_hi)); + + /* add eax,eax */ + EMIT2(0x01, add_2reg(0xC0, IA32_ECX, IA32_EAX)); + + if (dstk) + /* mov eax,dword ptr [ebp+off] */ + EMIT3(0x8B, add_2reg(0x40, IA32_EBP, IA32_EAX), + STACK_VAR(dst_lo)); + else + /* mov eax,dst_lo */ + EMIT2(0x8B, add_2reg(0xC0, dst_lo, IA32_EAX)); + + if (sstk) + /* mul dword ptr [ebp+off] */ + EMIT3(0xF7, add_1reg(0x60, IA32_EBP), STACK_VAR(src_lo)); + else + /* mul src_lo */ + EMIT2(0xF7, add_1reg(0xE0, src_lo)); + + /* add ecx,edx */ + EMIT2(0x01, add_2reg(0xC0, IA32_ECX, IA32_EDX)); + + if (dstk) { + /* mov dword ptr [ebp+off],eax */ + EMIT3(0x89, add_2reg(0x40, IA32_EBP, IA32_EAX), + STACK_VAR(dst_lo)); + /* mov dword ptr [ebp+off],ecx */ + EMIT3(0x89, add_2reg(0x40, IA32_EBP, IA32_ECX), + STACK_VAR(dst_hi)); + } else { + /* mov dst_lo,eax */ + EMIT2(0x89, add_2reg(0xC0, dst_lo, IA32_EAX)); + /* mov dst_hi,ecx */ + EMIT2(0x89, add_2reg(0xC0, dst_hi, IA32_ECX)); + } + + *pprog = prog; +} + +static inline void emit_ia32_mul_i64(const u8 dst[], const u32 val, + bool dstk, u8 **pprog) +{ + u8 *prog = *pprog; + int cnt = 0; + u32 hi; + + hi = val & (1<<31) ? (u32)~0 : 0; + /* movl eax,imm32 */ + EMIT2_off32(0xC7, add_1reg(0xC0, IA32_EAX), val); + if (dstk) + /* mul dword ptr [ebp+off] */ + EMIT3(0xF7, add_1reg(0x60, IA32_EBP), STACK_VAR(dst_hi)); + else + /* mul dst_hi */ + EMIT2(0xF7, add_1reg(0xE0, dst_hi)); + + /* mov ecx,eax */ + EMIT2(0x89, add_2reg(0xC0, IA32_ECX, IA32_EAX)); + + /* movl eax,imm32 */ + EMIT2_off32(0xC7, add_1reg(0xC0, IA32_EAX), hi); + if (dstk) + /* mul dword ptr [ebp+off] */ + EMIT3(0xF7, add_1reg(0x60, IA32_EBP), STACK_VAR(dst_lo)); + else + /* mul dst_lo */ + EMIT2(0xF7, add_1reg(0xE0, dst_lo)); + /* add ecx,eax */ + EMIT2(0x01, add_2reg(0xC0, IA32_ECX, IA32_EAX)); + + /* movl eax,imm32 */ + EMIT2_off32(0xC7, add_1reg(0xC0, IA32_EAX), val); + if (dstk) + /* mul dword ptr [ebp+off] */ + EMIT3(0xF7, add_1reg(0x60, IA32_EBP), STACK_VAR(dst_lo)); + else + /* mul dst_lo */ + EMIT2(0xF7, add_1reg(0xE0, dst_lo)); + + /* add ecx,edx */ + EMIT2(0x01, add_2reg(0xC0, IA32_ECX, IA32_EDX)); + + if (dstk) { + /* mov dword ptr [ebp+off],eax */ + EMIT3(0x89, add_2reg(0x40, IA32_EBP, IA32_EAX), + STACK_VAR(dst_lo)); + /* mov dword ptr [ebp+off],ecx */ + EMIT3(0x89, add_2reg(0x40, IA32_EBP, IA32_ECX), + STACK_VAR(dst_hi)); + } else { + /* mov dword ptr [ebp+off],eax */ + EMIT2(0x89, add_2reg(0xC0, dst_lo, IA32_EAX)); + /* mov dword ptr [ebp+off],ecx */ + EMIT2(0x89, add_2reg(0xC0, dst_hi, IA32_ECX)); + } + + *pprog = prog; +} + +static int bpf_size_to_x86_bytes(int bpf_size) +{ + if (bpf_size == BPF_W) + return 4; + else if (bpf_size == BPF_H) + return 2; + else if (bpf_size == BPF_B) + return 1; + else if (bpf_size == BPF_DW) + return 4; /* imm32 */ + else + return 0; +} + +struct jit_context { + int cleanup_addr; /* Epilogue code offset */ +}; + +/* Maximum number of bytes emitted while JITing one eBPF insn */ +#define BPF_MAX_INSN_SIZE 128 +#define BPF_INSN_SAFETY 64 + +#define PROLOGUE_SIZE 35 + +/* + * Emit prologue code for BPF program and check it's size. + * bpf_tail_call helper will skip it while jumping into another program. + */ +static void emit_prologue(u8 **pprog, u32 stack_depth) +{ + u8 *prog = *pprog; + int cnt = 0; + const u8 *r1 = bpf2ia32[BPF_REG_1]; + const u8 fplo = bpf2ia32[BPF_REG_FP][0]; + const u8 fphi = bpf2ia32[BPF_REG_FP][1]; + const u8 *tcc = bpf2ia32[TCALL_CNT]; + + /* push ebp */ + EMIT1(0x55); + /* mov ebp,esp */ + EMIT2(0x89, 0xE5); + /* push edi */ + EMIT1(0x57); + /* push esi */ + EMIT1(0x56); + /* push ebx */ + EMIT1(0x53); + + /* sub esp,STACK_SIZE */ + EMIT2_off32(0x81, 0xEC, STACK_SIZE); + /* sub ebp,SCRATCH_SIZE+4+12*/ + EMIT3(0x83, add_1reg(0xE8, IA32_EBP), SCRATCH_SIZE + 16); + /* xor ebx,ebx */ + EMIT2(0x31, add_2reg(0xC0, IA32_EBX, IA32_EBX)); + + /* Set up BPF prog stack base register */ + EMIT3(0x89, add_2reg(0x40, IA32_EBP, IA32_EBP), STACK_VAR(fplo)); + EMIT3(0x89, add_2reg(0x40, IA32_EBP, IA32_EBX), STACK_VAR(fphi)); + + /* Move BPF_CTX (EAX) to BPF_REG_R1 */ + /* mov dword ptr [ebp+off],eax */ + EMIT3(0x89, add_2reg(0x40, IA32_EBP, IA32_EAX), STACK_VAR(r1[0])); + EMIT3(0x89, add_2reg(0x40, IA32_EBP, IA32_EBX), STACK_VAR(r1[1])); + + /* Initialize Tail Count */ + EMIT3(0x89, add_2reg(0x40, IA32_EBP, IA32_EBX), STACK_VAR(tcc[0])); + EMIT3(0x89, add_2reg(0x40, IA32_EBP, IA32_EBX), STACK_VAR(tcc[1])); + + BUILD_BUG_ON(cnt != PROLOGUE_SIZE); + *pprog = prog; +} + +/* Emit epilogue code for BPF program */ +static void emit_epilogue(u8 **pprog, u32 stack_depth) +{ + u8 *prog = *pprog; + const u8 *r0 = bpf2ia32[BPF_REG_0]; + int cnt = 0; + + /* mov eax,dword ptr [ebp+off]*/ + EMIT3(0x8B, add_2reg(0x40, IA32_EBP, IA32_EAX), STACK_VAR(r0[0])); + /* mov edx,dword ptr [ebp+off]*/ + EMIT3(0x8B, add_2reg(0x40, IA32_EBP, IA32_EDX), STACK_VAR(r0[1])); + + /* add ebp,SCRATCH_SIZE+4+12*/ + EMIT3(0x83, add_1reg(0xC0, IA32_EBP), SCRATCH_SIZE + 16); + + /* mov ebx,dword ptr [ebp-12]*/ + EMIT3(0x8B, add_2reg(0x40, IA32_EBP, IA32_EBX), -12); + /* mov esi,dword ptr [ebp-8]*/ + EMIT3(0x8B, add_2reg(0x40, IA32_EBP, IA32_ESI), -8); + /* mov edi,dword ptr [ebp-4]*/ + EMIT3(0x8B, add_2reg(0x40, IA32_EBP, IA32_EDI), -4); + + EMIT1(0xC9); /* leave */ + EMIT1(0xC3); /* ret */ + *pprog = prog; +} + +/* + * Generate the following code: + * ... bpf_tail_call(void *ctx, struct bpf_array *array, u64 index) ... + * if (index >= array->map.max_entries) + * goto out; + * if (++tail_call_cnt > MAX_TAIL_CALL_CNT) + * goto out; + * prog = array->ptrs[index]; + * if (prog == NULL) + * goto out; + * goto *(prog->bpf_func + prologue_size); + * out: + */ +static void emit_bpf_tail_call(u8 **pprog) +{ + u8 *prog = *pprog; + int cnt = 0; + const u8 *r1 = bpf2ia32[BPF_REG_1]; + const u8 *r2 = bpf2ia32[BPF_REG_2]; + const u8 *r3 = bpf2ia32[BPF_REG_3]; + const u8 *tcc = bpf2ia32[TCALL_CNT]; + u32 lo, hi; + static int jmp_label1 = -1; + + /* + * if (index >= array->map.max_entries) + * goto out; + */ + /* mov eax,dword ptr [ebp+off] */ + EMIT3(0x8B, add_2reg(0x40, IA32_EBP, IA32_EAX), STACK_VAR(r2[0])); + /* mov edx,dword ptr [ebp+off] */ + EMIT3(0x8B, add_2reg(0x40, IA32_EBP, IA32_EDX), STACK_VAR(r3[0])); + + /* cmp dword ptr [eax+off],edx */ + EMIT3(0x39, add_2reg(0x40, IA32_EAX, IA32_EDX), + offsetof(struct bpf_array, map.max_entries)); + /* jbe out */ + EMIT2(IA32_JBE, jmp_label(jmp_label1, 2)); + + /* + * if (tail_call_cnt > MAX_TAIL_CALL_CNT) + * goto out; + */ + lo = (u32)MAX_TAIL_CALL_CNT; + hi = (u32)((u64)MAX_TAIL_CALL_CNT >> 32); + EMIT3(0x8B, add_2reg(0x40, IA32_EBP, IA32_ECX), STACK_VAR(tcc[0])); + EMIT3(0x8B, add_2reg(0x40, IA32_EBP, IA32_EBX), STACK_VAR(tcc[1])); + + /* cmp edx,hi */ + EMIT3(0x83, add_1reg(0xF8, IA32_EBX), hi); + EMIT2(IA32_JNE, 3); + /* cmp ecx,lo */ + EMIT3(0x83, add_1reg(0xF8, IA32_ECX), lo); + + /* ja out */ + EMIT2(IA32_JAE, jmp_label(jmp_label1, 2)); + + /* add eax,0x1 */ + EMIT3(0x83, add_1reg(0xC0, IA32_ECX), 0x01); + /* adc ebx,0x0 */ + EMIT3(0x83, add_1reg(0xD0, IA32_EBX), 0x00); + + /* mov dword ptr [ebp+off],eax */ + EMIT3(0x89, add_2reg(0x40, IA32_EBP, IA32_ECX), STACK_VAR(tcc[0])); + /* mov dword ptr [ebp+off],edx */ + EMIT3(0x89, add_2reg(0x40, IA32_EBP, IA32_EBX), STACK_VAR(tcc[1])); + + /* prog = array->ptrs[index]; */ + /* mov edx, [eax + edx * 4 + offsetof(...)] */ + EMIT3_off32(0x8B, 0x94, 0x90, offsetof(struct bpf_array, ptrs)); + + /* + * if (prog == NULL) + * goto out; + */ + /* test edx,edx */ + EMIT2(0x85, add_2reg(0xC0, IA32_EDX, IA32_EDX)); + /* je out */ + EMIT2(IA32_JE, jmp_label(jmp_label1, 2)); + + /* goto *(prog->bpf_func + prologue_size); */ + /* mov edx, dword ptr [edx + 32] */ + EMIT3(0x8B, add_2reg(0x40, IA32_EDX, IA32_EDX), + offsetof(struct bpf_prog, bpf_func)); + /* add edx,prologue_size */ + EMIT3(0x83, add_1reg(0xC0, IA32_EDX), PROLOGUE_SIZE); + + /* mov eax,dword ptr [ebp+off] */ + EMIT3(0x8B, add_2reg(0x40, IA32_EBP, IA32_EAX), STACK_VAR(r1[0])); + + /* + * Now we're ready to jump into next BPF program: + * eax == ctx (1st arg) + * edx == prog->bpf_func + prologue_size + */ + RETPOLINE_EDX_BPF_JIT(); + + if (jmp_label1 == -1) + jmp_label1 = cnt; + + /* out: */ + *pprog = prog; +} + +/* Push the scratch stack register on top of the stack. */ +static inline void emit_push_r64(const u8 src[], u8 **pprog) +{ + u8 *prog = *pprog; + int cnt = 0; + + /* mov ecx,dword ptr [ebp+off] */ + EMIT3(0x8B, add_2reg(0x40, IA32_EBP, IA32_ECX), STACK_VAR(src_hi)); + /* push ecx */ + EMIT1(0x51); + + /* mov ecx,dword ptr [ebp+off] */ + EMIT3(0x8B, add_2reg(0x40, IA32_EBP, IA32_ECX), STACK_VAR(src_lo)); + /* push ecx */ + EMIT1(0x51); + + *pprog = prog; +} + +static int do_jit(struct bpf_prog *bpf_prog, int *addrs, u8 *image, + int oldproglen, struct jit_context *ctx) +{ + struct bpf_insn *insn = bpf_prog->insnsi; + int insn_cnt = bpf_prog->len; + bool seen_exit = false; + u8 temp[BPF_MAX_INSN_SIZE + BPF_INSN_SAFETY]; + int i, cnt = 0; + int proglen = 0; + u8 *prog = temp; + + emit_prologue(&prog, bpf_prog->aux->stack_depth); + + for (i = 0; i < insn_cnt; i++, insn++) { + const s32 imm32 = insn->imm; + const bool is64 = BPF_CLASS(insn->code) == BPF_ALU64; + const bool dstk = insn->dst_reg == BPF_REG_AX ? false : true; + const bool sstk = insn->src_reg == BPF_REG_AX ? false : true; + const u8 code = insn->code; + const u8 *dst = bpf2ia32[insn->dst_reg]; + const u8 *src = bpf2ia32[insn->src_reg]; + const u8 *r0 = bpf2ia32[BPF_REG_0]; + s64 jmp_offset; + u8 jmp_cond; + int ilen; + u8 *func; + + switch (code) { + /* ALU operations */ + /* dst = src */ + case BPF_ALU | BPF_MOV | BPF_K: + case BPF_ALU | BPF_MOV | BPF_X: + case BPF_ALU64 | BPF_MOV | BPF_K: + case BPF_ALU64 | BPF_MOV | BPF_X: + switch (BPF_SRC(code)) { + case BPF_X: + emit_ia32_mov_r64(is64, dst, src, dstk, + sstk, &prog); + break; + case BPF_K: + /* Sign-extend immediate value to dst reg */ + emit_ia32_mov_i64(is64, dst, imm32, + dstk, &prog); + break; + } + break; + /* dst = dst + src/imm */ + /* dst = dst - src/imm */ + /* dst = dst | src/imm */ + /* dst = dst & src/imm */ + /* dst = dst ^ src/imm */ + /* dst = dst * src/imm */ + /* dst = dst << src */ + /* dst = dst >> src */ + case BPF_ALU | BPF_ADD | BPF_K: + case BPF_ALU | BPF_ADD | BPF_X: + case BPF_ALU | BPF_SUB | BPF_K: + case BPF_ALU | BPF_SUB | BPF_X: + case BPF_ALU | BPF_OR | BPF_K: + case BPF_ALU | BPF_OR | BPF_X: + case BPF_ALU | BPF_AND | BPF_K: + case BPF_ALU | BPF_AND | BPF_X: + case BPF_ALU | BPF_XOR | BPF_K: + case BPF_ALU | BPF_XOR | BPF_X: + case BPF_ALU64 | BPF_ADD | BPF_K: + case BPF_ALU64 | BPF_ADD | BPF_X: + case BPF_ALU64 | BPF_SUB | BPF_K: + case BPF_ALU64 | BPF_SUB | BPF_X: + case BPF_ALU64 | BPF_OR | BPF_K: + case BPF_ALU64 | BPF_OR | BPF_X: + case BPF_ALU64 | BPF_AND | BPF_K: + case BPF_ALU64 | BPF_AND | BPF_X: + case BPF_ALU64 | BPF_XOR | BPF_K: + case BPF_ALU64 | BPF_XOR | BPF_X: + switch (BPF_SRC(code)) { + case BPF_X: + emit_ia32_alu_r64(is64, BPF_OP(code), dst, + src, dstk, sstk, &prog); + break; + case BPF_K: + emit_ia32_alu_i64(is64, BPF_OP(code), dst, + imm32, dstk, &prog); + break; + } + break; + case BPF_ALU | BPF_MUL | BPF_K: + case BPF_ALU | BPF_MUL | BPF_X: + switch (BPF_SRC(code)) { + case BPF_X: + emit_ia32_mul_r(dst_lo, src_lo, dstk, + sstk, &prog); + break; + case BPF_K: + /* mov ecx,imm32*/ + EMIT2_off32(0xC7, add_1reg(0xC0, IA32_ECX), + imm32); + emit_ia32_mul_r(dst_lo, IA32_ECX, dstk, + false, &prog); + break; + } + emit_ia32_mov_i(dst_hi, 0, dstk, &prog); + break; + case BPF_ALU | BPF_LSH | BPF_X: + case BPF_ALU | BPF_RSH | BPF_X: + case BPF_ALU | BPF_ARSH | BPF_K: + case BPF_ALU | BPF_ARSH | BPF_X: + switch (BPF_SRC(code)) { + case BPF_X: + emit_ia32_shift_r(BPF_OP(code), dst_lo, src_lo, + dstk, sstk, &prog); + break; + case BPF_K: + /* mov ecx,imm32*/ + EMIT2_off32(0xC7, add_1reg(0xC0, IA32_ECX), + imm32); + emit_ia32_shift_r(BPF_OP(code), dst_lo, + IA32_ECX, dstk, false, + &prog); + break; + } + emit_ia32_mov_i(dst_hi, 0, dstk, &prog); + break; + /* dst = dst / src(imm) */ + /* dst = dst % src(imm) */ + case BPF_ALU | BPF_DIV | BPF_K: + case BPF_ALU | BPF_DIV | BPF_X: + case BPF_ALU | BPF_MOD | BPF_K: + case BPF_ALU | BPF_MOD | BPF_X: + switch (BPF_SRC(code)) { + case BPF_X: + emit_ia32_div_mod_r(BPF_OP(code), dst_lo, + src_lo, dstk, sstk, &prog); + break; + case BPF_K: + /* mov ecx,imm32*/ + EMIT2_off32(0xC7, add_1reg(0xC0, IA32_ECX), + imm32); + emit_ia32_div_mod_r(BPF_OP(code), dst_lo, + IA32_ECX, dstk, false, + &prog); + break; + } + emit_ia32_mov_i(dst_hi, 0, dstk, &prog); + break; + case BPF_ALU64 | BPF_DIV | BPF_K: + case BPF_ALU64 | BPF_DIV | BPF_X: + case BPF_ALU64 | BPF_MOD | BPF_K: + case BPF_ALU64 | BPF_MOD | BPF_X: + goto notyet; + /* dst = dst >> imm */ + /* dst = dst << imm */ + case BPF_ALU | BPF_RSH | BPF_K: + case BPF_ALU | BPF_LSH | BPF_K: + if (unlikely(imm32 > 31)) + return -EINVAL; + /* mov ecx,imm32*/ + EMIT2_off32(0xC7, add_1reg(0xC0, IA32_ECX), imm32); + emit_ia32_shift_r(BPF_OP(code), dst_lo, IA32_ECX, dstk, + false, &prog); + emit_ia32_mov_i(dst_hi, 0, dstk, &prog); + break; + /* dst = dst << imm */ + case BPF_ALU64 | BPF_LSH | BPF_K: + if (unlikely(imm32 > 63)) + return -EINVAL; + emit_ia32_lsh_i64(dst, imm32, dstk, &prog); + break; + /* dst = dst >> imm */ + case BPF_ALU64 | BPF_RSH | BPF_K: + if (unlikely(imm32 > 63)) + return -EINVAL; + emit_ia32_rsh_i64(dst, imm32, dstk, &prog); + break; + /* dst = dst << src */ + case BPF_ALU64 | BPF_LSH | BPF_X: + emit_ia32_lsh_r64(dst, src, dstk, sstk, &prog); + break; + /* dst = dst >> src */ + case BPF_ALU64 | BPF_RSH | BPF_X: + emit_ia32_rsh_r64(dst, src, dstk, sstk, &prog); + break; + /* dst = dst >> src (signed) */ + case BPF_ALU64 | BPF_ARSH | BPF_X: + emit_ia32_arsh_r64(dst, src, dstk, sstk, &prog); + break; + /* dst = dst >> imm (signed) */ + case BPF_ALU64 | BPF_ARSH | BPF_K: + if (unlikely(imm32 > 63)) + return -EINVAL; + emit_ia32_arsh_i64(dst, imm32, dstk, &prog); + break; + /* dst = ~dst */ + case BPF_ALU | BPF_NEG: + emit_ia32_alu_i(is64, false, BPF_OP(code), + dst_lo, 0, dstk, &prog); + emit_ia32_mov_i(dst_hi, 0, dstk, &prog); + break; + /* dst = ~dst (64 bit) */ + case BPF_ALU64 | BPF_NEG: + emit_ia32_neg64(dst, dstk, &prog); + break; + /* dst = dst * src/imm */ + case BPF_ALU64 | BPF_MUL | BPF_X: + case BPF_ALU64 | BPF_MUL | BPF_K: + switch (BPF_SRC(code)) { + case BPF_X: + emit_ia32_mul_r64(dst, src, dstk, sstk, &prog); + break; + case BPF_K: + emit_ia32_mul_i64(dst, imm32, dstk, &prog); + break; + } + break; + /* dst = htole(dst) */ + case BPF_ALU | BPF_END | BPF_FROM_LE: + emit_ia32_to_le_r64(dst, imm32, dstk, &prog); + break; + /* dst = htobe(dst) */ + case BPF_ALU | BPF_END | BPF_FROM_BE: + emit_ia32_to_be_r64(dst, imm32, dstk, &prog); + break; + /* dst = imm64 */ + case BPF_LD | BPF_IMM | BPF_DW: { + s32 hi, lo = imm32; + + hi = insn[1].imm; + emit_ia32_mov_i(dst_lo, lo, dstk, &prog); + emit_ia32_mov_i(dst_hi, hi, dstk, &prog); + insn++; + i++; + break; + } + /* ST: *(u8*)(dst_reg + off) = imm */ + case BPF_ST | BPF_MEM | BPF_H: + case BPF_ST | BPF_MEM | BPF_B: + case BPF_ST | BPF_MEM | BPF_W: + case BPF_ST | BPF_MEM | BPF_DW: + if (dstk) + /* mov eax,dword ptr [ebp+off] */ + EMIT3(0x8B, add_2reg(0x40, IA32_EBP, IA32_EAX), + STACK_VAR(dst_lo)); + else + /* mov eax,dst_lo */ + EMIT2(0x8B, add_2reg(0xC0, dst_lo, IA32_EAX)); + + switch (BPF_SIZE(code)) { + case BPF_B: + EMIT(0xC6, 1); break; + case BPF_H: + EMIT2(0x66, 0xC7); break; + case BPF_W: + case BPF_DW: + EMIT(0xC7, 1); break; + } + + if (is_imm8(insn->off)) + EMIT2(add_1reg(0x40, IA32_EAX), insn->off); + else + EMIT1_off32(add_1reg(0x80, IA32_EAX), + insn->off); + EMIT(imm32, bpf_size_to_x86_bytes(BPF_SIZE(code))); + + if (BPF_SIZE(code) == BPF_DW) { + u32 hi; + + hi = imm32 & (1<<31) ? (u32)~0 : 0; + EMIT2_off32(0xC7, add_1reg(0x80, IA32_EAX), + insn->off + 4); + EMIT(hi, 4); + } + break; + + /* STX: *(u8*)(dst_reg + off) = src_reg */ + case BPF_STX | BPF_MEM | BPF_B: + case BPF_STX | BPF_MEM | BPF_H: + case BPF_STX | BPF_MEM | BPF_W: + case BPF_STX | BPF_MEM | BPF_DW: + if (dstk) + /* mov eax,dword ptr [ebp+off] */ + EMIT3(0x8B, add_2reg(0x40, IA32_EBP, IA32_EAX), + STACK_VAR(dst_lo)); + else + /* mov eax,dst_lo */ + EMIT2(0x8B, add_2reg(0xC0, dst_lo, IA32_EAX)); + + if (sstk) + /* mov edx,dword ptr [ebp+off] */ + EMIT3(0x8B, add_2reg(0x40, IA32_EBP, IA32_EDX), + STACK_VAR(src_lo)); + else + /* mov edx,src_lo */ + EMIT2(0x8B, add_2reg(0xC0, src_lo, IA32_EDX)); + + switch (BPF_SIZE(code)) { + case BPF_B: + EMIT(0x88, 1); break; + case BPF_H: + EMIT2(0x66, 0x89); break; + case BPF_W: + case BPF_DW: + EMIT(0x89, 1); break; + } + + if (is_imm8(insn->off)) + EMIT2(add_2reg(0x40, IA32_EAX, IA32_EDX), + insn->off); + else + EMIT1_off32(add_2reg(0x80, IA32_EAX, IA32_EDX), + insn->off); + + if (BPF_SIZE(code) == BPF_DW) { + if (sstk) + /* mov edi,dword ptr [ebp+off] */ + EMIT3(0x8B, add_2reg(0x40, IA32_EBP, + IA32_EDX), + STACK_VAR(src_hi)); + else + /* mov edi,src_hi */ + EMIT2(0x8B, add_2reg(0xC0, src_hi, + IA32_EDX)); + EMIT1(0x89); + if (is_imm8(insn->off + 4)) { + EMIT2(add_2reg(0x40, IA32_EAX, + IA32_EDX), + insn->off + 4); + } else { + EMIT1(add_2reg(0x80, IA32_EAX, + IA32_EDX)); + EMIT(insn->off + 4, 4); + } + } + break; + + /* LDX: dst_reg = *(u8*)(src_reg + off) */ + case BPF_LDX | BPF_MEM | BPF_B: + case BPF_LDX | BPF_MEM | BPF_H: + case BPF_LDX | BPF_MEM | BPF_W: + case BPF_LDX | BPF_MEM | BPF_DW: + if (sstk) + /* mov eax,dword ptr [ebp+off] */ + EMIT3(0x8B, add_2reg(0x40, IA32_EBP, IA32_EAX), + STACK_VAR(src_lo)); + else + /* mov eax,dword ptr [ebp+off] */ + EMIT2(0x8B, add_2reg(0xC0, src_lo, IA32_EAX)); + + switch (BPF_SIZE(code)) { + case BPF_B: + EMIT2(0x0F, 0xB6); break; + case BPF_H: + EMIT2(0x0F, 0xB7); break; + case BPF_W: + case BPF_DW: + EMIT(0x8B, 1); break; + } + + if (is_imm8(insn->off)) + EMIT2(add_2reg(0x40, IA32_EAX, IA32_EDX), + insn->off); + else + EMIT1_off32(add_2reg(0x80, IA32_EAX, IA32_EDX), + insn->off); + + if (dstk) + /* mov dword ptr [ebp+off],edx */ + EMIT3(0x89, add_2reg(0x40, IA32_EBP, IA32_EDX), + STACK_VAR(dst_lo)); + else + /* mov dst_lo,edx */ + EMIT2(0x89, add_2reg(0xC0, dst_lo, IA32_EDX)); + switch (BPF_SIZE(code)) { + case BPF_B: + case BPF_H: + case BPF_W: + if (dstk) { + EMIT3(0xC7, add_1reg(0x40, IA32_EBP), + STACK_VAR(dst_hi)); + EMIT(0x0, 4); + } else { + EMIT3(0xC7, add_1reg(0xC0, dst_hi), 0); + } + break; + case BPF_DW: + EMIT2_off32(0x8B, + add_2reg(0x80, IA32_EAX, IA32_EDX), + insn->off + 4); + if (dstk) + EMIT3(0x89, + add_2reg(0x40, IA32_EBP, + IA32_EDX), + STACK_VAR(dst_hi)); + else + EMIT2(0x89, + add_2reg(0xC0, dst_hi, IA32_EDX)); + break; + default: + break; + } + break; + /* call */ + case BPF_JMP | BPF_CALL: + { + const u8 *r1 = bpf2ia32[BPF_REG_1]; + const u8 *r2 = bpf2ia32[BPF_REG_2]; + const u8 *r3 = bpf2ia32[BPF_REG_3]; + const u8 *r4 = bpf2ia32[BPF_REG_4]; + const u8 *r5 = bpf2ia32[BPF_REG_5]; + + if (insn->src_reg == BPF_PSEUDO_CALL) + goto notyet; + + func = (u8 *) __bpf_call_base + imm32; + jmp_offset = func - (image + addrs[i]); + + if (!imm32 || !is_simm32(jmp_offset)) { + pr_err("unsupported BPF func %d addr %p image %p\n", + imm32, func, image); + return -EINVAL; + } + + /* mov eax,dword ptr [ebp+off] */ + EMIT3(0x8B, add_2reg(0x40, IA32_EBP, IA32_EAX), + STACK_VAR(r1[0])); + /* mov edx,dword ptr [ebp+off] */ + EMIT3(0x8B, add_2reg(0x40, IA32_EBP, IA32_EDX), + STACK_VAR(r1[1])); + + emit_push_r64(r5, &prog); + emit_push_r64(r4, &prog); + emit_push_r64(r3, &prog); + emit_push_r64(r2, &prog); + + EMIT1_off32(0xE8, jmp_offset + 9); + + /* mov dword ptr [ebp+off],eax */ + EMIT3(0x89, add_2reg(0x40, IA32_EBP, IA32_EAX), + STACK_VAR(r0[0])); + /* mov dword ptr [ebp+off],edx */ + EMIT3(0x89, add_2reg(0x40, IA32_EBP, IA32_EDX), + STACK_VAR(r0[1])); + + /* add esp,32 */ + EMIT3(0x83, add_1reg(0xC0, IA32_ESP), 32); + break; + } + case BPF_JMP | BPF_TAIL_CALL: + emit_bpf_tail_call(&prog); + break; + + /* cond jump */ + case BPF_JMP | BPF_JEQ | BPF_X: + case BPF_JMP | BPF_JNE | BPF_X: + case BPF_JMP | BPF_JGT | BPF_X: + case BPF_JMP | BPF_JLT | BPF_X: + case BPF_JMP | BPF_JGE | BPF_X: + case BPF_JMP | BPF_JLE | BPF_X: + case BPF_JMP | BPF_JSGT | BPF_X: + case BPF_JMP | BPF_JSLE | BPF_X: + case BPF_JMP | BPF_JSLT | BPF_X: + case BPF_JMP | BPF_JSGE | BPF_X: { + u8 dreg_lo = dstk ? IA32_EAX : dst_lo; + u8 dreg_hi = dstk ? IA32_EDX : dst_hi; + u8 sreg_lo = sstk ? IA32_ECX : src_lo; + u8 sreg_hi = sstk ? IA32_EBX : src_hi; + + if (dstk) { + EMIT3(0x8B, add_2reg(0x40, IA32_EBP, IA32_EAX), + STACK_VAR(dst_lo)); + EMIT3(0x8B, add_2reg(0x40, IA32_EBP, IA32_EDX), + STACK_VAR(dst_hi)); + } + + if (sstk) { + EMIT3(0x8B, add_2reg(0x40, IA32_EBP, IA32_ECX), + STACK_VAR(src_lo)); + EMIT3(0x8B, add_2reg(0x40, IA32_EBP, IA32_EBX), + STACK_VAR(src_hi)); + } + + /* cmp dreg_hi,sreg_hi */ + EMIT2(0x39, add_2reg(0xC0, dreg_hi, sreg_hi)); + EMIT2(IA32_JNE, 2); + /* cmp dreg_lo,sreg_lo */ + EMIT2(0x39, add_2reg(0xC0, dreg_lo, sreg_lo)); + goto emit_cond_jmp; + } + case BPF_JMP | BPF_JSET | BPF_X: { + u8 dreg_lo = dstk ? IA32_EAX : dst_lo; + u8 dreg_hi = dstk ? IA32_EDX : dst_hi; + u8 sreg_lo = sstk ? IA32_ECX : src_lo; + u8 sreg_hi = sstk ? IA32_EBX : src_hi; + + if (dstk) { + EMIT3(0x8B, add_2reg(0x40, IA32_EBP, IA32_EAX), + STACK_VAR(dst_lo)); + EMIT3(0x8B, add_2reg(0x40, IA32_EBP, IA32_EDX), + STACK_VAR(dst_hi)); + } + + if (sstk) { + EMIT3(0x8B, add_2reg(0x40, IA32_EBP, IA32_ECX), + STACK_VAR(src_lo)); + EMIT3(0x8B, add_2reg(0x40, IA32_EBP, IA32_EBX), + STACK_VAR(src_hi)); + } + /* and dreg_lo,sreg_lo */ + EMIT2(0x23, add_2reg(0xC0, sreg_lo, dreg_lo)); + /* and dreg_hi,sreg_hi */ + EMIT2(0x23, add_2reg(0xC0, sreg_hi, dreg_hi)); + /* or dreg_lo,dreg_hi */ + EMIT2(0x09, add_2reg(0xC0, dreg_lo, dreg_hi)); + goto emit_cond_jmp; + } + case BPF_JMP | BPF_JSET | BPF_K: { + u32 hi; + u8 dreg_lo = dstk ? IA32_EAX : dst_lo; + u8 dreg_hi = dstk ? IA32_EDX : dst_hi; + u8 sreg_lo = IA32_ECX; + u8 sreg_hi = IA32_EBX; + + if (dstk) { + EMIT3(0x8B, add_2reg(0x40, IA32_EBP, IA32_EAX), + STACK_VAR(dst_lo)); + EMIT3(0x8B, add_2reg(0x40, IA32_EBP, IA32_EDX), + STACK_VAR(dst_hi)); + } + hi = imm32 & (1<<31) ? (u32)~0 : 0; + + /* mov ecx,imm32 */ + EMIT2_off32(0xC7, add_1reg(0xC0, IA32_ECX), imm32); + /* mov ebx,imm32 */ + EMIT2_off32(0xC7, add_1reg(0xC0, IA32_EBX), hi); + + /* and dreg_lo,sreg_lo */ + EMIT2(0x23, add_2reg(0xC0, sreg_lo, dreg_lo)); + /* and dreg_hi,sreg_hi */ + EMIT2(0x23, add_2reg(0xC0, sreg_hi, dreg_hi)); + /* or dreg_lo,dreg_hi */ + EMIT2(0x09, add_2reg(0xC0, dreg_lo, dreg_hi)); + goto emit_cond_jmp; + } + case BPF_JMP | BPF_JEQ | BPF_K: + case BPF_JMP | BPF_JNE | BPF_K: + case BPF_JMP | BPF_JGT | BPF_K: + case BPF_JMP | BPF_JLT | BPF_K: + case BPF_JMP | BPF_JGE | BPF_K: + case BPF_JMP | BPF_JLE | BPF_K: + case BPF_JMP | BPF_JSGT | BPF_K: + case BPF_JMP | BPF_JSLE | BPF_K: + case BPF_JMP | BPF_JSLT | BPF_K: + case BPF_JMP | BPF_JSGE | BPF_K: { + u32 hi; + u8 dreg_lo = dstk ? IA32_EAX : dst_lo; + u8 dreg_hi = dstk ? IA32_EDX : dst_hi; + u8 sreg_lo = IA32_ECX; + u8 sreg_hi = IA32_EBX; + + if (dstk) { + EMIT3(0x8B, add_2reg(0x40, IA32_EBP, IA32_EAX), + STACK_VAR(dst_lo)); + EMIT3(0x8B, add_2reg(0x40, IA32_EBP, IA32_EDX), + STACK_VAR(dst_hi)); + } + + hi = imm32 & (1<<31) ? (u32)~0 : 0; + /* mov ecx,imm32 */ + EMIT2_off32(0xC7, add_1reg(0xC0, IA32_ECX), imm32); + /* mov ebx,imm32 */ + EMIT2_off32(0xC7, add_1reg(0xC0, IA32_EBX), hi); + + /* cmp dreg_hi,sreg_hi */ + EMIT2(0x39, add_2reg(0xC0, dreg_hi, sreg_hi)); + EMIT2(IA32_JNE, 2); + /* cmp dreg_lo,sreg_lo */ + EMIT2(0x39, add_2reg(0xC0, dreg_lo, sreg_lo)); + +emit_cond_jmp: /* Convert BPF opcode to x86 */ + switch (BPF_OP(code)) { + case BPF_JEQ: + jmp_cond = IA32_JE; + break; + case BPF_JSET: + case BPF_JNE: + jmp_cond = IA32_JNE; + break; + case BPF_JGT: + /* GT is unsigned '>', JA in x86 */ + jmp_cond = IA32_JA; + break; + case BPF_JLT: + /* LT is unsigned '<', JB in x86 */ + jmp_cond = IA32_JB; + break; + case BPF_JGE: + /* GE is unsigned '>=', JAE in x86 */ + jmp_cond = IA32_JAE; + break; + case BPF_JLE: + /* LE is unsigned '<=', JBE in x86 */ + jmp_cond = IA32_JBE; + break; + case BPF_JSGT: + /* Signed '>', GT in x86 */ + jmp_cond = IA32_JG; + break; + case BPF_JSLT: + /* Signed '<', LT in x86 */ + jmp_cond = IA32_JL; + break; + case BPF_JSGE: + /* Signed '>=', GE in x86 */ + jmp_cond = IA32_JGE; + break; + case BPF_JSLE: + /* Signed '<=', LE in x86 */ + jmp_cond = IA32_JLE; + break; + default: /* to silence GCC warning */ + return -EFAULT; + } + jmp_offset = addrs[i + insn->off] - addrs[i]; + if (is_imm8(jmp_offset)) { + EMIT2(jmp_cond, jmp_offset); + } else if (is_simm32(jmp_offset)) { + EMIT2_off32(0x0F, jmp_cond + 0x10, jmp_offset); + } else { + pr_err("cond_jmp gen bug %llx\n", jmp_offset); + return -EFAULT; + } + + break; + } + case BPF_JMP | BPF_JA: + if (insn->off == -1) + /* -1 jmp instructions will always jump + * backwards two bytes. Explicitly handling + * this case avoids wasting too many passes + * when there are long sequences of replaced + * dead code. + */ + jmp_offset = -2; + else + jmp_offset = addrs[i + insn->off] - addrs[i]; + + if (!jmp_offset) + /* Optimize out nop jumps */ + break; +emit_jmp: + if (is_imm8(jmp_offset)) { + EMIT2(0xEB, jmp_offset); + } else if (is_simm32(jmp_offset)) { + EMIT1_off32(0xE9, jmp_offset); + } else { + pr_err("jmp gen bug %llx\n", jmp_offset); + return -EFAULT; + } + break; + /* STX XADD: lock *(u32 *)(dst + off) += src */ + case BPF_STX | BPF_XADD | BPF_W: + /* STX XADD: lock *(u64 *)(dst + off) += src */ + case BPF_STX | BPF_XADD | BPF_DW: + goto notyet; + case BPF_JMP | BPF_EXIT: + if (seen_exit) { + jmp_offset = ctx->cleanup_addr - addrs[i]; + goto emit_jmp; + } + seen_exit = true; + /* Update cleanup_addr */ + ctx->cleanup_addr = proglen; + emit_epilogue(&prog, bpf_prog->aux->stack_depth); + break; +notyet: + pr_info_once("*** NOT YET: opcode %02x ***\n", code); + return -EFAULT; + default: + /* + * This error will be seen if new instruction was added + * to interpreter, but not to JIT or if there is junk in + * bpf_prog + */ + pr_err("bpf_jit: unknown opcode %02x\n", code); + return -EINVAL; + } + + ilen = prog - temp; + if (ilen > BPF_MAX_INSN_SIZE) { + pr_err("bpf_jit: fatal insn size error\n"); + return -EFAULT; + } + + if (image) { + if (unlikely(proglen + ilen > oldproglen)) { + pr_err("bpf_jit: fatal error\n"); + return -EFAULT; + } + memcpy(image + proglen, temp, ilen); + } + proglen += ilen; + addrs[i] = proglen; + prog = temp; + } + return proglen; +} + +struct bpf_prog *bpf_int_jit_compile(struct bpf_prog *prog) +{ + struct bpf_binary_header *header = NULL; + struct bpf_prog *tmp, *orig_prog = prog; + int proglen, oldproglen = 0; + struct jit_context ctx = {}; + bool tmp_blinded = false; + u8 *image = NULL; + int *addrs; + int pass; + int i; + + if (!prog->jit_requested) + return orig_prog; + + tmp = bpf_jit_blind_constants(prog); + /* + * If blinding was requested and we failed during blinding, + * we must fall back to the interpreter. + */ + if (IS_ERR(tmp)) + return orig_prog; + if (tmp != prog) { + tmp_blinded = true; + prog = tmp; + } + + addrs = kmalloc(prog->len * sizeof(*addrs), GFP_KERNEL); + if (!addrs) { + prog = orig_prog; + goto out; + } + + /* + * Before first pass, make a rough estimation of addrs[] + * each BPF instruction is translated to less than 64 bytes + */ + for (proglen = 0, i = 0; i < prog->len; i++) { + proglen += 64; + addrs[i] = proglen; + } + ctx.cleanup_addr = proglen; + + /* + * JITed image shrinks with every pass and the loop iterates + * until the image stops shrinking. Very large BPF programs + * may converge on the last pass. In such case do one more + * pass to emit the final image. + */ + for (pass = 0; pass < 20 || image; pass++) { + proglen = do_jit(prog, addrs, image, oldproglen, &ctx); + if (proglen <= 0) { +out_image: + image = NULL; + if (header) + bpf_jit_binary_free(header); + prog = orig_prog; + goto out_addrs; + } + if (image) { + if (proglen != oldproglen) { + pr_err("bpf_jit: proglen=%d != oldproglen=%d\n", + proglen, oldproglen); + goto out_image; + } + break; + } + if (proglen == oldproglen) { + header = bpf_jit_binary_alloc(proglen, &image, + 1, jit_fill_hole); + if (!header) { + prog = orig_prog; + goto out_addrs; + } + } + oldproglen = proglen; + cond_resched(); + } + + if (bpf_jit_enable > 1) + bpf_jit_dump(prog->len, proglen, pass + 1, image); + + if (image) { + bpf_jit_binary_lock_ro(header); + prog->bpf_func = (void *)image; + prog->jited = 1; + prog->jited_len = proglen; + } else { + prog = orig_prog; + } + +out_addrs: + kfree(addrs); +out: + if (tmp_blinded) + bpf_jit_prog_release_other(prog, prog == orig_prog ? + tmp : orig_prog); + return prog; +} diff --git a/arch/x86/pci/early.c b/arch/x86/pci/early.c index f0114007e915..e5f753cbb1c3 100644 --- a/arch/x86/pci/early.c +++ b/arch/x86/pci/early.c @@ -59,24 +59,15 @@ int early_pci_allowed(void) void early_dump_pci_device(u8 bus, u8 slot, u8 func) { + u32 value[256 / 4]; int i; - int j; - u32 val; - printk(KERN_INFO "pci 0000:%02x:%02x.%d config space:", - bus, slot, func); + pr_info("pci 0000:%02x:%02x.%d config space:\n", bus, slot, func); - for (i = 0; i < 256; i += 4) { - if (!(i & 0x0f)) - printk("\n %02x:",i); + for (i = 0; i < 256; i += 4) + value[i / 4] = read_pci_config(bus, slot, func, i); - val = read_pci_config(bus, slot, func, i); - for (j = 0; j < 4; j++) { - printk(" %02x", val & 0xff); - val >>= 8; - } - } - printk("\n"); + print_hex_dump(KERN_INFO, "", DUMP_PREFIX_OFFSET, 16, 1, value, 256, false); } void early_dump_pci_devices(void) diff --git a/arch/x86/pci/fixup.c b/arch/x86/pci/fixup.c index 54ef19e90705..13f4485ca388 100644 --- a/arch/x86/pci/fixup.c +++ b/arch/x86/pci/fixup.c @@ -636,6 +636,10 @@ DECLARE_PCI_FIXUP_EARLY(PCI_VENDOR_ID_INTEL, 0x2030, quirk_no_aersid); DECLARE_PCI_FIXUP_EARLY(PCI_VENDOR_ID_INTEL, 0x2031, quirk_no_aersid); DECLARE_PCI_FIXUP_EARLY(PCI_VENDOR_ID_INTEL, 0x2032, quirk_no_aersid); DECLARE_PCI_FIXUP_EARLY(PCI_VENDOR_ID_INTEL, 0x2033, quirk_no_aersid); +DECLARE_PCI_FIXUP_EARLY(PCI_VENDOR_ID_INTEL, 0x334a, quirk_no_aersid); +DECLARE_PCI_FIXUP_EARLY(PCI_VENDOR_ID_INTEL, 0x334b, quirk_no_aersid); +DECLARE_PCI_FIXUP_EARLY(PCI_VENDOR_ID_INTEL, 0x334c, quirk_no_aersid); +DECLARE_PCI_FIXUP_EARLY(PCI_VENDOR_ID_INTEL, 0x334d, quirk_no_aersid); #ifdef CONFIG_PHYS_ADDR_T_64BIT diff --git a/arch/x86/platform/uv/tlb_uv.c b/arch/x86/platform/uv/tlb_uv.c index b36caae0fb2f..b96d38288c60 100644 --- a/arch/x86/platform/uv/tlb_uv.c +++ b/arch/x86/platform/uv/tlb_uv.c @@ -615,7 +615,7 @@ static int uv2_3_wait_completion(struct bau_desc *bau_desc, /* spin on the status MMR, waiting for it to go idle */ while (descriptor_stat != UV2H_DESC_IDLE) { - if ((descriptor_stat == UV2H_DESC_SOURCE_TIMEOUT)) { + if (descriptor_stat == UV2H_DESC_SOURCE_TIMEOUT) { /* * A h/w bug on the destination side may * have prevented the message being marked diff --git a/arch/x86/platform/uv/uv_irq.c b/arch/x86/platform/uv/uv_irq.c index e4cb9f4cde8a..fc13cbbb2dce 100644 --- a/arch/x86/platform/uv/uv_irq.c +++ b/arch/x86/platform/uv/uv_irq.c @@ -47,11 +47,6 @@ static void uv_program_mmr(struct irq_cfg *cfg, struct uv_irq_2_mmr_pnode *info) static void uv_noop(struct irq_data *data) { } -static void uv_ack_apic(struct irq_data *data) -{ - ack_APIC_irq(); -} - static int uv_set_irq_affinity(struct irq_data *data, const struct cpumask *mask, bool force) @@ -73,7 +68,7 @@ static struct irq_chip uv_irq_chip = { .name = "UV-CORE", .irq_mask = uv_noop, .irq_unmask = uv_noop, - .irq_eoi = uv_ack_apic, + .irq_eoi = apic_ack_irq, .irq_set_affinity = uv_set_irq_affinity, }; diff --git a/arch/x86/um/Kconfig b/arch/x86/um/Kconfig index 13ed827c7c66..9d529f22fd9d 100644 --- a/arch/x86/um/Kconfig +++ b/arch/x86/um/Kconfig @@ -1,5 +1,9 @@ # SPDX-License-Identifier: GPL-2.0 -mainmenu "User Mode Linux/$SUBARCH $KERNELVERSION Kernel Configuration" +mainmenu "User Mode Linux/$(SUBARCH) $(KERNELVERSION) Kernel Configuration" + +comment "Compiler: $(CC_VERSION_TEXT)" + +source "scripts/Kconfig.include" source "arch/um/Kconfig.common" @@ -16,8 +20,8 @@ config UML_X86 select GENERIC_FIND_FIRST_BIT config 64BIT - bool "64-bit kernel" if SUBARCH = "x86" - default SUBARCH != "i386" + bool "64-bit kernel" if "$(SUBARCH)" = "x86" + default "$(SUBARCH)" != "i386" config X86_32 def_bool !64BIT diff --git a/arch/x86/xen/mmu.c b/arch/x86/xen/mmu.c index 2d76106788a3..96fc2f0fdbfe 100644 --- a/arch/x86/xen/mmu.c +++ b/arch/x86/xen/mmu.c @@ -63,37 +63,44 @@ static noinline void xen_flush_tlb_all(void) #define REMAP_BATCH_SIZE 16 struct remap_data { - xen_pfn_t *mfn; + xen_pfn_t *pfn; bool contiguous; + bool no_translate; pgprot_t prot; struct mmu_update *mmu_update; }; -static int remap_area_mfn_pte_fn(pte_t *ptep, pgtable_t token, +static int remap_area_pfn_pte_fn(pte_t *ptep, pgtable_t token, unsigned long addr, void *data) { struct remap_data *rmd = data; - pte_t pte = pte_mkspecial(mfn_pte(*rmd->mfn, rmd->prot)); + pte_t pte = pte_mkspecial(mfn_pte(*rmd->pfn, rmd->prot)); - /* If we have a contiguous range, just update the mfn itself, - else update pointer to be "next mfn". */ + /* + * If we have a contiguous range, just update the pfn itself, + * else update pointer to be "next pfn". + */ if (rmd->contiguous) - (*rmd->mfn)++; + (*rmd->pfn)++; else - rmd->mfn++; + rmd->pfn++; - rmd->mmu_update->ptr = virt_to_machine(ptep).maddr | MMU_NORMAL_PT_UPDATE; + rmd->mmu_update->ptr = virt_to_machine(ptep).maddr; + rmd->mmu_update->ptr |= rmd->no_translate ? + MMU_PT_UPDATE_NO_TRANSLATE : + MMU_NORMAL_PT_UPDATE; rmd->mmu_update->val = pte_val_ma(pte); rmd->mmu_update++; return 0; } -static int do_remap_gfn(struct vm_area_struct *vma, +static int do_remap_pfn(struct vm_area_struct *vma, unsigned long addr, - xen_pfn_t *gfn, int nr, + xen_pfn_t *pfn, int nr, int *err_ptr, pgprot_t prot, - unsigned domid, + unsigned int domid, + bool no_translate, struct page **pages) { int err = 0; @@ -104,11 +111,14 @@ static int do_remap_gfn(struct vm_area_struct *vma, BUG_ON(!((vma->vm_flags & (VM_PFNMAP | VM_IO)) == (VM_PFNMAP | VM_IO))); - rmd.mfn = gfn; + rmd.pfn = pfn; rmd.prot = prot; - /* We use the err_ptr to indicate if there we are doing a contiguous - * mapping or a discontigious mapping. */ + /* + * We use the err_ptr to indicate if there we are doing a contiguous + * mapping or a discontigious mapping. + */ rmd.contiguous = !err_ptr; + rmd.no_translate = no_translate; while (nr) { int index = 0; @@ -119,7 +129,7 @@ static int do_remap_gfn(struct vm_area_struct *vma, rmd.mmu_update = mmu_update; err = apply_to_page_range(vma->vm_mm, addr, range, - remap_area_mfn_pte_fn, &rmd); + remap_area_pfn_pte_fn, &rmd); if (err) goto out; @@ -173,7 +183,8 @@ int xen_remap_domain_gfn_range(struct vm_area_struct *vma, if (xen_feature(XENFEAT_auto_translated_physmap)) return -EOPNOTSUPP; - return do_remap_gfn(vma, addr, &gfn, nr, NULL, prot, domid, pages); + return do_remap_pfn(vma, addr, &gfn, nr, NULL, prot, domid, false, + pages); } EXPORT_SYMBOL_GPL(xen_remap_domain_gfn_range); @@ -192,10 +203,25 @@ int xen_remap_domain_gfn_array(struct vm_area_struct *vma, * cause of "wrong memory was mapped in". */ BUG_ON(err_ptr == NULL); - return do_remap_gfn(vma, addr, gfn, nr, err_ptr, prot, domid, pages); + return do_remap_pfn(vma, addr, gfn, nr, err_ptr, prot, domid, + false, pages); } EXPORT_SYMBOL_GPL(xen_remap_domain_gfn_array); +int xen_remap_domain_mfn_array(struct vm_area_struct *vma, + unsigned long addr, + xen_pfn_t *mfn, int nr, + int *err_ptr, pgprot_t prot, + unsigned int domid, struct page **pages) +{ + if (xen_feature(XENFEAT_auto_translated_physmap)) + return -EOPNOTSUPP; + + return do_remap_pfn(vma, addr, mfn, nr, err_ptr, prot, domid, + true, pages); +} +EXPORT_SYMBOL_GPL(xen_remap_domain_mfn_array); + /* Returns: 0 success */ int xen_unmap_domain_gfn_range(struct vm_area_struct *vma, int nr, struct page **pages) diff --git a/arch/x86/xen/xen-pvh.S b/arch/x86/xen/xen-pvh.S index e1a5fbeae08d..ca2d3b2bf2af 100644 --- a/arch/x86/xen/xen-pvh.S +++ b/arch/x86/xen/xen-pvh.S @@ -54,12 +54,19 @@ * charge of setting up it's own stack, GDT and IDT. */ +#define PVH_GDT_ENTRY_CS 1 +#define PVH_GDT_ENTRY_DS 2 +#define PVH_GDT_ENTRY_CANARY 3 +#define PVH_CS_SEL (PVH_GDT_ENTRY_CS * 8) +#define PVH_DS_SEL (PVH_GDT_ENTRY_DS * 8) +#define PVH_CANARY_SEL (PVH_GDT_ENTRY_CANARY * 8) + ENTRY(pvh_start_xen) cld lgdt (_pa(gdt)) - mov $(__BOOT_DS),%eax + mov $PVH_DS_SEL,%eax mov %eax,%ds mov %eax,%es mov %eax,%ss @@ -93,11 +100,17 @@ ENTRY(pvh_start_xen) mov %eax, %cr0 /* Jump to 64-bit mode. */ - ljmp $__KERNEL_CS, $_pa(1f) + ljmp $PVH_CS_SEL, $_pa(1f) /* 64-bit entry point. */ .code64 1: + /* Set base address in stack canary descriptor. */ + mov $MSR_GS_BASE,%ecx + mov $_pa(canary), %eax + xor %edx, %edx + wrmsr + call xen_prepare_pvh /* startup_64 expects boot_params in %rsi. */ @@ -107,6 +120,17 @@ ENTRY(pvh_start_xen) #else /* CONFIG_X86_64 */ + /* Set base address in stack canary descriptor. */ + movl $_pa(gdt_start),%eax + movl $_pa(canary),%ecx + movw %cx, (PVH_GDT_ENTRY_CANARY * 8) + 2(%eax) + shrl $16, %ecx + movb %cl, (PVH_GDT_ENTRY_CANARY * 8) + 4(%eax) + movb %ch, (PVH_GDT_ENTRY_CANARY * 8) + 7(%eax) + + mov $PVH_CANARY_SEL,%eax + mov %eax,%gs + call mk_early_pgtbl_32 mov $_pa(initial_page_table), %eax @@ -116,13 +140,13 @@ ENTRY(pvh_start_xen) or $(X86_CR0_PG | X86_CR0_PE), %eax mov %eax, %cr0 - ljmp $__BOOT_CS, $1f + ljmp $PVH_CS_SEL, $1f 1: call xen_prepare_pvh mov $_pa(pvh_bootparams), %esi /* startup_32 doesn't expect paging and PAE to be on. */ - ljmp $__BOOT_CS, $_pa(2f) + ljmp $PVH_CS_SEL, $_pa(2f) 2: mov %cr0, %eax and $~X86_CR0_PG, %eax @@ -131,7 +155,7 @@ ENTRY(pvh_start_xen) and $~X86_CR4_PAE, %eax mov %eax, %cr4 - ljmp $__BOOT_CS, $_pa(startup_32) + ljmp $PVH_CS_SEL, $_pa(startup_32) #endif END(pvh_start_xen) @@ -143,16 +167,19 @@ gdt: .word 0 gdt_start: .quad 0x0000000000000000 /* NULL descriptor */ - .quad 0x0000000000000000 /* reserved */ #ifdef CONFIG_X86_64 - .quad GDT_ENTRY(0xa09a, 0, 0xfffff) /* __KERNEL_CS */ + .quad GDT_ENTRY(0xa09a, 0, 0xfffff) /* PVH_CS_SEL */ #else - .quad GDT_ENTRY(0xc09a, 0, 0xfffff) /* __KERNEL_CS */ + .quad GDT_ENTRY(0xc09a, 0, 0xfffff) /* PVH_CS_SEL */ #endif - .quad GDT_ENTRY(0xc092, 0, 0xfffff) /* __KERNEL_DS */ + .quad GDT_ENTRY(0xc092, 0, 0xfffff) /* PVH_DS_SEL */ + .quad GDT_ENTRY(0x4090, 0, 0x18) /* PVH_CANARY_SEL */ gdt_end: - .balign 4 + .balign 16 +canary: + .fill 48, 1, 0 + early_stack: .fill 256, 1, 0 early_stack_end: |