hsblst 0.0.2 → 0.0.3

raw patch · 84 files changed

+2712/−840 lines, 84 filesdep +base16-bytestringdep −hex-textdep ~basedep ~bytestringdep ~deepseqPVP: major bump suggested

API removals or changes: PVP suggests a major version bump

Dependencies added: base16-bytestring

Dependencies removed: hex-text

Dependency ranges changed: base, bytestring, deepseq, tasty, text

API changes (from Hackage documentation)

Files

• CHANGELOG.md +6/−0
• README.md +1/−1
• c-source/README.md +4/−2
• c-source/bindings/blst.h +14/−11
• c-source/bindings/blst_aux.h +20/−5
• c-source/build/assembly.S +21/−28
• c-source/build/coff/add_mod_256-x86_64.s +59/−46
• c-source/build/coff/add_mod_384-armv8.S +75/−0
• c-source/build/coff/add_mod_384-x86_64.s +200/−79
• c-source/build/coff/add_mod_384x384-x86_64.s +16/−12
• c-source/build/coff/ct_inverse_mod_256-armv8.S +3/−2
• c-source/build/coff/ct_inverse_mod_256-x86_64.s +9/−5
• c-source/build/coff/ct_inverse_mod_384-armv8.S +3/−2
• c-source/build/coff/ct_is_square_mod_384-armv8.S +1/−0
• c-source/build/coff/ct_is_square_mod_384-x86_64.s +7/−3
• c-source/build/coff/ctq_inverse_mod_384-x86_64.s +12/−3
• c-source/build/coff/ctx_inverse_mod_384-x86_64.s +13/−8
• c-source/build/coff/div3w-armv8.S +1/−1
• c-source/build/coff/div3w-x86_64.s +117/−9
• c-source/build/coff/mulq_mont_256-x86_64.s +41/−16
• c-source/build/coff/mulq_mont_384-x86_64.s +174/−77
• c-source/build/coff/mulx_mont_256-x86_64.s +28/−16
• c-source/build/coff/mulx_mont_384-x86_64.s +128/−79
• c-source/build/coff/sha256-armv8.S +7/−1
• c-source/build/coff/sha256-portable-x86_64.s +31/−23
• c-source/build/coff/sha256-x86_64.s +91/−89
• c-source/build/elf/add_mod_384-armv8.S +69/−0
• c-source/build/elf/add_mod_384-x86_64.s +98/−0
• c-source/build/elf/ct_inverse_mod_256-armv8.S +3/−2
• c-source/build/elf/ct_inverse_mod_256-x86_64.s +1/−0
• c-source/build/elf/ct_inverse_mod_384-armv8.S +3/−2
• c-source/build/elf/ct_is_square_mod_384-armv8.S +1/−0
• c-source/build/elf/ct_is_square_mod_384-x86_64.s +1/−0
• c-source/build/elf/ctq_inverse_mod_384-x86_64.s +6/−0
• c-source/build/elf/ctx_inverse_mod_384-x86_64.s +8/−6
• c-source/build/elf/div3w-armv8.S +1/−1
• c-source/build/elf/div3w-x86_64.s +12/−3
• c-source/build/elf/mulq_mont_256-x86_64.s +17/−0
• c-source/build/elf/mulq_mont_384-x86_64.s +90/−29
• c-source/build/elf/mulx_mont_256-x86_64.s +4/−0
• c-source/build/elf/mulx_mont_384-x86_64.s +42/−27
• c-source/build/elf/sha256-armv8.S +7/−1
• c-source/build/elf/sha256-portable-x86_64.s +20/−16
• c-source/build/elf/sha256-x86_64.s +38/−29
• c-source/build/mach-o/add_mod_384-armv8.S +69/−0
• c-source/build/mach-o/add_mod_384-x86_64.s +98/−0
• c-source/build/mach-o/ct_inverse_mod_256-armv8.S +3/−2
• c-source/build/mach-o/ct_inverse_mod_256-x86_64.s +1/−0
• c-source/build/mach-o/ct_inverse_mod_384-armv8.S +3/−2
• c-source/build/mach-o/ct_is_square_mod_384-armv8.S +1/−0
• c-source/build/mach-o/ct_is_square_mod_384-x86_64.s +1/−0
• c-source/build/mach-o/ctq_inverse_mod_384-x86_64.s +6/−0
• c-source/build/mach-o/ctx_inverse_mod_384-x86_64.s +6/−4
• c-source/build/mach-o/div3w-armv8.S +1/−1
• c-source/build/mach-o/div3w-x86_64.s +12/−3
• c-source/build/mach-o/mulq_mont_256-x86_64.s +17/−0
• c-source/build/mach-o/mulq_mont_384-x86_64.s +82/−21
• c-source/build/mach-o/mulx_mont_256-x86_64.s +4/−0
• c-source/build/mach-o/mulx_mont_384-x86_64.s +34/−19
• c-source/build/mach-o/sha256-armv8.S +7/−1
• c-source/build/mach-o/sha256-portable-x86_64.s +20/−16
• c-source/build/mach-o/sha256-x86_64.s +38/−29
• c-source/src/aggregate.c +1/−1
• c-source/src/bulk_addition.c +10/−2
• c-source/src/bytes.h +2/−2
• c-source/src/cpuid.c +85/−0
• c-source/src/e1.c +4/−4
• c-source/src/e2.c +4/−4
• c-source/src/ec_mult.h +3/−2
• c-source/src/exports.c +71/−47
• c-source/src/fields.h +2/−2
• c-source/src/fp12_tower.c +2/−2
• c-source/src/multi_scalar.c +14/−1
• c-source/src/no_asm.h +59/−5
• c-source/src/pairing.c +49/−0
• c-source/src/pentaroot-addchain.h +333/−0
• c-source/src/pentaroot.c +76/−0
• c-source/src/server.c +7/−1
• c-source/src/vect.h +47/−13
• hsblst.cabal +24/−17
• src/Crypto/BLST/Internal/Bindings.chs +1/−0
• src/Crypto/BLST/Internal/Bindings/Types.hs +3/−0
• src/Crypto/BLST/Internal/Types.hs +3/−0
• test/Test/BLST/Util.hs +6/−5

@@ -3,6 +3,12 @@    - SPDX-License-Identifier: MPL-2.0    --> +## v0.0.3++* Minor tweaks for GHC 9.8 compatibility.+* Types `Point` and `Affine` now have `PointKind` parameter be role `nominal`, as coercing those between point types will break invariants.+* Bundled blst version bumped to v0.3.11.+ ## v0.0.2  * Prevent inlining of foreign calls. This fixes a potential efficiency issue,

@@ -8,7 +8,7 @@  HsBLST is low-level Haskell bindings and a high-level interface to [BLST](https://github.com/supranational/blst) -- a multilingual BLS12-381 signature library. -This package includes sources from v0.3.10 version of BLST, licensed under+This package includes sources from v0.3.11 version of BLST, licensed under [Apache License 2.0](https://github.com/supranational/blst/blob/master/LICENSE)  ## Build Instructions [↑](#hsblst)

@@ -1,4 +1,4 @@-[![Build Status](https://travis-ci.com/supranational/blst.svg?branch=master)](https://travis-ci.com/github/supranational/blst) [![Actions status](https://github.com/supranational/blst/workflows/build/badge.svg)](https://github.com/supranational/blst/actions) [![Total alerts](https://img.shields.io/lgtm/alerts/g/supranational/blst.svg?logo=lgtm&logoWidth=18)](https://lgtm.com/projects/g/supranational/blst/alerts/)+[![Build Status](https://travis-ci.com/supranational/blst.svg?branch=master)](https://travis-ci.com/github/supranational/blst) [![Actions status](https://github.com/supranational/blst/workflows/build/badge.svg)](https://github.com/supranational/blst/actions) <div align="left">   <img src=blst_logo_small.png> </div>@@ -56,6 +56,8 @@ - [Python](bindings/python) - [Java](bindings/java) - [Node.js](bindings/node.js)+- [Emscripten](bindings/emscripten)+- [C#](bindings/c%23)  ## API @@ -134,7 +136,7 @@ * The third-most significant bit is set if (and only if) this point is in compressed form _and_ it is not the point at infinity _and_ its y-coordinate is the lexicographically largest of the two associated with the encoded x-coordinate.  ## Build-The build process is very simple and only requires a C complier. It's integrated into the Go and Rust ecosystems, so that respective users would go about as they would with any other external module. Otherwise, a binary library would have to be compiled.+The build process is very simple and only requires a C compiler. It's integrated into the Go and Rust ecosystems, so that respective users would go about as they would with any other external module. Otherwise, a binary library would have to be compiled.  ### C static library A static library called libblst.a can be built in the current working directory of the user's choice:

@@ -83,7 +83,7 @@  #ifndef SWIG /*- * BLS12-381-specifc Fr operations.+ * BLS12-381-specific Fr operations.  */ void blst_fr_add(blst_fr *ret, const blst_fr *a, const blst_fr *b); void blst_fr_sub(blst_fr *ret, const blst_fr *a, const blst_fr *b);@@ -102,7 +102,7 @@ void blst_scalar_from_fr(blst_scalar *ret, const blst_fr *a);  /*- * BLS12-381-specifc Fp operations.+ * BLS12-381-specific Fp operations.  */ void blst_fp_add(blst_fp *ret, const blst_fp *a, const blst_fp *b); void blst_fp_sub(blst_fp *ret, const blst_fp *a, const blst_fp *b);@@ -126,7 +126,7 @@ void blst_lendian_from_fp(byte ret[48], const blst_fp *a);  /*- * BLS12-381-specifc Fp2 operations.+ * BLS12-381-specific Fp2 operations.  */ void blst_fp2_add(blst_fp2 *ret, const blst_fp2 *a, const blst_fp2 *b); void blst_fp2_sub(blst_fp2 *ret, const blst_fp2 *a, const blst_fp2 *b);@@ -141,7 +141,7 @@ bool blst_fp2_sqrt(blst_fp2 *ret, const blst_fp2 *a);  /*- * BLS12-381-specifc Fp12 operations.+ * BLS12-381-specific Fp12 operations.  */ void blst_fp12_sqr(blst_fp12 *ret, const blst_fp12 *a); void blst_fp12_cyclotomic_sqr(blst_fp12 *ret, const blst_fp12 *a);@@ -155,11 +155,11 @@ bool blst_fp12_is_equal(const blst_fp12 *a, const blst_fp12 *b); bool blst_fp12_is_one(const blst_fp12 *a); bool blst_fp12_in_group(const blst_fp12 *a);-const blst_fp12 *blst_fp12_one();+const blst_fp12 *blst_fp12_one(void); #endif  // SWIG  /*- * BLS12-381-specifc point operations.+ * BLS12-381-specific point operations.  */ typedef struct { blst_fp x, y, z; } blst_p1; typedef struct { blst_fp x, y; } blst_p1_affine;@@ -180,13 +180,13 @@ bool blst_p1_in_g1(const blst_p1 *p); bool blst_p1_is_equal(const blst_p1 *a, const blst_p1 *b); bool blst_p1_is_inf(const blst_p1 *a);-const blst_p1 *blst_p1_generator();+const blst_p1 *blst_p1_generator(void);  bool blst_p1_affine_on_curve(const blst_p1_affine *p); bool blst_p1_affine_in_g1(const blst_p1_affine *p); bool blst_p1_affine_is_equal(const blst_p1_affine *a, const blst_p1_affine *b); bool blst_p1_affine_is_inf(const blst_p1_affine *a);-const blst_p1_affine *blst_p1_affine_generator();+const blst_p1_affine *blst_p1_affine_generator(void);  typedef struct { blst_fp2 x, y, z; } blst_p2; typedef struct { blst_fp2 x, y; } blst_p2_affine;@@ -207,13 +207,13 @@ bool blst_p2_in_g2(const blst_p2 *p); bool blst_p2_is_equal(const blst_p2 *a, const blst_p2 *b); bool blst_p2_is_inf(const blst_p2 *a);-const blst_p2 *blst_p2_generator();+const blst_p2 *blst_p2_generator(void);  bool blst_p2_affine_on_curve(const blst_p2_affine *p); bool blst_p2_affine_in_g2(const blst_p2_affine *p); bool blst_p2_affine_is_equal(const blst_p2_affine *a, const blst_p2_affine *b); bool blst_p2_affine_is_inf(const blst_p2_affine *a);-const blst_p2_affine *blst_p2_affine_generator();+const blst_p2_affine *blst_p2_affine_generator(void);  /*  * Multi-scalar multiplications and other multi-point operations.@@ -337,6 +337,9 @@ #ifndef SWIG void blst_miller_loop(blst_fp12 *ret, const blst_p2_affine *Q,                                       const blst_p1_affine *P);+void blst_miller_loop_n(blst_fp12 *ret, const blst_p2_affine *const Qs[],+                                        const blst_p1_affine *const Ps[],+                                        size_t n); void blst_final_exp(blst_fp12 *ret, const blst_fp12 *f); void blst_precompute_lines(blst_fp6 Qlines[68], const blst_p2_affine *Q); void blst_miller_loop_lines(blst_fp12 *ret, const blst_fp6 Qlines[68],@@ -352,7 +355,7 @@ typedef struct blst_opaque blst_pairing; #endif -size_t blst_pairing_sizeof();+size_t blst_pairing_sizeof(void); void blst_pairing_init(blst_pairing *new_ctx, bool hash_or_encode,                        const byte *DST DEFNULL, size_t DST_len DEFNULL); const byte *blst_pairing_get_dst(const blst_pairing *ctx);

@@ -10,8 +10,14 @@  * depending on their proven/unproven worthiness.  */ +void blst_fr_ct_bfly(blst_fr *x0, blst_fr *x1, const blst_fr *twiddle);+void blst_fr_gs_bfly(blst_fr *x0, blst_fr *x1, const blst_fr *twiddle); void blst_fr_to(blst_fr *ret, const blst_fr *a); void blst_fr_from(blst_fr *ret, const blst_fr *a);+#ifdef BLST_FR_PENTAROOT+void blst_fr_pentaroot(blst_fr *ret, const blst_fr *a);+void blst_fr_pentapow(blst_fr *ret, const blst_fr *a);+#endif  void blst_fp_to(blst_fp *ret, const blst_fp *a); void blst_fp_from(blst_fp *ret, const blst_fp *a);@@ -50,7 +56,11 @@ void blst_sign_pk2_in_g2(byte out[96], blst_p1_affine *out_sig,                          const blst_p1 *hash, const blst_scalar *SK); +#ifdef __BLST_RUST_BINDGEN__ typedef struct {} blst_uniq;+#else+typedef struct blst_opaque blst_uniq;+#endif  size_t blst_uniq_sizeof(size_t n_nodes); void blst_uniq_init(blst_uniq *tree);@@ -94,9 +104,14 @@ void blst_fr_from_hexascii(blst_fr *ret, const byte *hex); void blst_fp_from_hexascii(blst_fp *ret, const byte *hex); -size_t blst_p1_sizeof();-size_t blst_p1_affine_sizeof();-size_t blst_p2_sizeof();-size_t blst_p2_affine_sizeof();-size_t blst_fp12_sizeof();+size_t blst_p1_sizeof(void);+size_t blst_p1_affine_sizeof(void);+size_t blst_p2_sizeof(void);+size_t blst_p2_affine_sizeof(void);+size_t blst_fp12_sizeof(void);++/*+ * Single-shot SHA-256 hash function.+ */+void blst_sha256(byte out[32], const byte *msg, size_t msg_len); #endif

@@ -2,23 +2,22 @@ # if defined(__ELF__) #  if defined(__BLST_PORTABLE__) #   include "elf/sha256-portable-x86_64.s"-#  else-#   include "elf/sha256-x86_64.s"+#   define blst_sha256_block_data_order blst_sha256_block_ssse3 #  endif-#  if defined(__ADX__) && !defined(__BLST_PORTABLE__)+#  include "elf/sha256-x86_64.s"+#  if defined(__ADX__) || defined(__BLST_PORTABLE__) #   include "elf/ctx_inverse_mod_384-x86_64.s"-#  else+#  endif+#  if !defined(__ADX__) || defined(__BLST_PORTABLE__) #   include "elf/ctq_inverse_mod_384-x86_64.s" #  endif #  include "elf/add_mod_384-x86_64.s" #  include "elf/add_mod_384x384-x86_64.s"-#  define __add_mod_384     __add_mont_384-#  define __sub_mod_384     __sub_mont_384-#  define __sub_mod_384x384 __sub_mont_384x384-#  if defined(__ADX__) && !defined(__BLST_PORTABLE__)+#  if defined(__ADX__) || defined(__BLST_PORTABLE__) #   include "elf/mulx_mont_384-x86_64.s" #   include "elf/mulx_mont_256-x86_64.s"-#  else+#  endif+#  if !defined(__ADX__) || defined(__BLST_PORTABLE__) #   include "elf/mulq_mont_384-x86_64.s" #   include "elf/mulq_mont_256-x86_64.s" #  endif@@ -27,25 +26,20 @@ #  include "elf/div3w-x86_64.s" #  include "elf/ct_is_square_mod_384-x86_64.s" # elif defined(_WIN64) || defined(__CYGWIN__)-#  if defined(__BLST_PORTABLE__)-#   include "coff/sha256-portable-x86_64.s"-#  else-#   include "coff/sha256-x86_64.s"-#  endif-#  if defined(__ADX__) && !defined(__BLST_PORTABLE__)+#  include "coff/sha256-x86_64.s"+#  if defined(__ADX__) || defined(__BLST_PORTABLE__) #   include "coff/ctx_inverse_mod_384-x86_64.s"-#  else+#  endif+#  if !defined(__ADX__) || defined(__BLST_PORTABLE__) #   include "coff/ctq_inverse_mod_384-x86_64.s" #  endif #  include "coff/add_mod_384-x86_64.s" #  include "coff/add_mod_384x384-x86_64.s"-#  define __add_mod_384     __add_mont_384-#  define __sub_mod_384     __sub_mont_384-#  define __sub_mod_384x384 __sub_mont_384x384-#  if defined(__ADX__) && !defined(__BLST_PORTABLE__)+#  if defined(__ADX__) || defined(__BLST_PORTABLE__) #   include "coff/mulx_mont_384-x86_64.s" #   include "coff/mulx_mont_256-x86_64.s"-#  else+#  endif+#  if !defined(__ADX__) || defined(__BLST_PORTABLE__) #   include "coff/mulq_mont_384-x86_64.s" #   include "coff/mulq_mont_256-x86_64.s" #  endif@@ -55,20 +49,19 @@ #  include "coff/ct_is_square_mod_384-x86_64.s" # elif defined(__APPLE__) #  include "mach-o/sha256-x86_64.s"-#  if defined(__ADX__) && !defined(__BLST_PORTABLE__)+#  if defined(__ADX__) || defined(__BLST_PORTABLE__) #   include "mach-o/ctx_inverse_mod_384-x86_64.s"-#  else+#  endif+#  if !defined(__ADX__) || defined(__BLST_PORTABLE__) #   include "mach-o/ctq_inverse_mod_384-x86_64.s" #  endif #  include "mach-o/add_mod_384-x86_64.s" #  include "mach-o/add_mod_384x384-x86_64.s"-#  define __add_mod_384     __add_mont_384-#  define __sub_mod_384     __sub_mont_384-#  define __sub_mod_384x384 __sub_mont_384x384-#  if defined(__ADX__) && !defined(__BLST_PORTABLE__)+#  if defined(__ADX__) || defined(__BLST_PORTABLE__) #   include "mach-o/mulx_mont_384-x86_64.s" #   include "mach-o/mulx_mont_256-x86_64.s"-#  else+#  endif+#  if !defined(__ADX__) || defined(__BLST_PORTABLE__) #   include "mach-o/mulq_mont_384-x86_64.s" #   include "mach-o/mulq_mont_256-x86_64.s" #  endif

@@ -10,14 +10,14 @@ 	movq	%rsi,16(%rsp) 	movq	%rsp,%r11 .LSEH_begin_add_mod_256:-	movq	%rcx,%rdi-	movq	%rdx,%rsi-	movq	%r8,%rdx-	movq	%r9,%rcx   	pushq	%rbp +	movq	%rcx,%rdi+	movq	%rdx,%rsi+	movq	%r8,%rdx+	movq	%r9,%rcx 	pushq	%rbx  	subq	$8,%rsp@@ -81,13 +81,13 @@ 	movq	%rsi,16(%rsp) 	movq	%rsp,%r11 .LSEH_begin_mul_by_3_mod_256:-	movq	%rcx,%rdi-	movq	%rdx,%rsi-	movq	%r8,%rdx   	pushq	%rbp +	movq	%rcx,%rdi+	movq	%rdx,%rsi+	movq	%r8,%rdx 	pushq	%rbx  	pushq	%r12@@ -161,14 +161,14 @@ 	movq	%rsi,16(%rsp) 	movq	%rsp,%r11 .LSEH_begin_lshift_mod_256:-	movq	%rcx,%rdi-	movq	%rdx,%rsi-	movq	%r8,%rdx-	movq	%r9,%rcx   	pushq	%rbp +	movq	%rcx,%rdi+	movq	%rdx,%rsi+	movq	%r8,%rdx+	movq	%r9,%rcx 	pushq	%rbx  	pushq	%r12@@ -218,14 +218,14 @@ 	movq	%rsi,16(%rsp) 	movq	%rsp,%r11 .LSEH_begin_rshift_mod_256:-	movq	%rcx,%rdi-	movq	%rdx,%rsi-	movq	%r8,%rdx-	movq	%r9,%rcx   	pushq	%rbp +	movq	%rcx,%rdi+	movq	%rdx,%rsi+	movq	%r8,%rdx+	movq	%r9,%rcx 	pushq	%rbx  	subq	$8,%rsp@@ -307,14 +307,14 @@ 	movq	%rsi,16(%rsp) 	movq	%rsp,%r11 .LSEH_begin_cneg_mod_256:-	movq	%rcx,%rdi-	movq	%rdx,%rsi-	movq	%r8,%rdx-	movq	%r9,%rcx   	pushq	%rbp +	movq	%rcx,%rdi+	movq	%rdx,%rsi+	movq	%r8,%rdx+	movq	%r9,%rcx 	pushq	%rbx  	pushq	%r12@@ -385,14 +385,14 @@ 	movq	%rsi,16(%rsp) 	movq	%rsp,%r11 .LSEH_begin_sub_mod_256:-	movq	%rcx,%rdi-	movq	%rdx,%rsi-	movq	%r8,%rdx-	movq	%r9,%rcx   	pushq	%rbp +	movq	%rcx,%rdi+	movq	%rdx,%rsi+	movq	%r8,%rdx+	movq	%r9,%rcx 	pushq	%rbx  	subq	$8,%rsp@@ -454,10 +454,10 @@ 	movq	%rsi,16(%rsp) 	movq	%rsp,%r11 .LSEH_begin_check_mod_256:-	movq	%rcx,%rdi-	movq	%rdx,%rsi  +	movq	%rcx,%rdi+	movq	%rdx,%rsi 	movq	0(%rdi),%rax 	movq	8(%rdi),%r9 	movq	16(%rdi),%r10@@ -497,14 +497,14 @@ 	movq	%rsi,16(%rsp) 	movq	%rsp,%r11 .LSEH_begin_add_n_check_mod_256:-	movq	%rcx,%rdi-	movq	%rdx,%rsi-	movq	%r8,%rdx-	movq	%r9,%rcx   	pushq	%rbp +	movq	%rcx,%rdi+	movq	%rdx,%rsi+	movq	%r8,%rdx+	movq	%r9,%rcx 	pushq	%rbx  	subq	$8,%rsp@@ -573,14 +573,14 @@ 	movq	%rsi,16(%rsp) 	movq	%rsp,%r11 .LSEH_begin_sub_n_check_mod_256:-	movq	%rcx,%rdi-	movq	%rdx,%rsi-	movq	%r8,%rdx-	movq	%r9,%rcx   	pushq	%rbp +	movq	%rcx,%rdi+	movq	%rdx,%rsi+	movq	%r8,%rdx+	movq	%r9,%rcx 	pushq	%rbx  	subq	$8,%rsp@@ -744,8 +744,9 @@ .byte	1,0,5,0x0b .byte	0,0x74,1,0 .byte	0,0x64,2,0-.byte	0,0x03+.byte	0,0xb3 .byte	0,0+.long	0,0 .LSEH_info_add_mod_256_body: .byte	1,0,9,0 .byte	0x00,0x34,0x01,0x00@@ -753,7 +754,8 @@ .byte	0x00,0x74,0x04,0x00 .byte	0x00,0x64,0x05,0x00 .byte	0x00,0x22-.byte	0x00,0x00+.byte	0x00,0x00,0x00,0x00,0x00,0x00+.byte	0x00,0x00,0x00,0x00 .LSEH_info_add_mod_256_epilogue: .byte	1,0,4,0 .byte	0x00,0x74,0x01,0x00@@ -764,8 +766,9 @@ .byte	1,0,5,0x0b .byte	0,0x74,1,0 .byte	0,0x64,2,0-.byte	0,0x03+.byte	0,0xb3 .byte	0,0+.long	0,0 .LSEH_info_mul_by_3_mod_256_body: .byte	1,0,11,0 .byte	0x00,0xc4,0x00,0x00@@ -785,8 +788,9 @@ .byte	1,0,5,0x0b .byte	0,0x74,1,0 .byte	0,0x64,2,0-.byte	0,0x03+.byte	0,0xb3 .byte	0,0+.long	0,0 .LSEH_info_lshift_mod_256_body: .byte	1,0,11,0 .byte	0x00,0xc4,0x00,0x00@@ -806,8 +810,9 @@ .byte	1,0,5,0x0b .byte	0,0x74,1,0 .byte	0,0x64,2,0-.byte	0,0x03+.byte	0,0xb3 .byte	0,0+.long	0,0 .LSEH_info_rshift_mod_256_body: .byte	1,0,9,0 .byte	0x00,0x34,0x01,0x00@@ -815,7 +820,8 @@ .byte	0x00,0x74,0x04,0x00 .byte	0x00,0x64,0x05,0x00 .byte	0x00,0x22-.byte	0x00,0x00+.byte	0x00,0x00,0x00,0x00,0x00,0x00+.byte	0x00,0x00,0x00,0x00 .LSEH_info_rshift_mod_256_epilogue: .byte	1,0,4,0 .byte	0x00,0x74,0x01,0x00@@ -826,8 +832,9 @@ .byte	1,0,5,0x0b .byte	0,0x74,1,0 .byte	0,0x64,2,0-.byte	0,0x03+.byte	0,0xb3 .byte	0,0+.long	0,0 .LSEH_info_cneg_mod_256_body: .byte	1,0,11,0 .byte	0x00,0xc4,0x00,0x00@@ -847,8 +854,9 @@ .byte	1,0,5,0x0b .byte	0,0x74,1,0 .byte	0,0x64,2,0-.byte	0,0x03+.byte	0,0xb3 .byte	0,0+.long	0,0 .LSEH_info_sub_mod_256_body: .byte	1,0,9,0 .byte	0x00,0x34,0x01,0x00@@ -856,7 +864,8 @@ .byte	0x00,0x74,0x04,0x00 .byte	0x00,0x64,0x05,0x00 .byte	0x00,0x22-.byte	0x00,0x00+.byte	0x00,0x00,0x00,0x00,0x00,0x00+.byte	0x00,0x00,0x00,0x00 .LSEH_info_sub_mod_256_epilogue: .byte	1,0,4,0 .byte	0x00,0x74,0x01,0x00@@ -873,8 +882,9 @@ .byte	1,0,5,0x0b .byte	0,0x74,1,0 .byte	0,0x64,2,0-.byte	0,0x03+.byte	0,0xb3 .byte	0,0+.long	0,0 .LSEH_info_add_n_check_mod_256_body: .byte	1,0,9,0 .byte	0x00,0x34,0x01,0x00@@ -882,7 +892,8 @@ .byte	0x00,0x74,0x04,0x00 .byte	0x00,0x64,0x05,0x00 .byte	0x00,0x22-.byte	0x00,0x00+.byte	0x00,0x00,0x00,0x00,0x00,0x00+.byte	0x00,0x00,0x00,0x00 .LSEH_info_add_n_check_mod_256_epilogue: .byte	1,0,4,0 .byte	0x00,0x74,0x01,0x00@@ -893,8 +904,9 @@ .byte	1,0,5,0x0b .byte	0,0x74,1,0 .byte	0,0x64,2,0-.byte	0,0x03+.byte	0,0xb3 .byte	0,0+.long	0,0 .LSEH_info_sub_n_check_mod_256_body: .byte	1,0,9,0 .byte	0x00,0x34,0x01,0x00@@ -902,7 +914,8 @@ .byte	0x00,0x74,0x04,0x00 .byte	0x00,0x64,0x05,0x00 .byte	0x00,0x22-.byte	0x00,0x00+.byte	0x00,0x00,0x00,0x00,0x00,0x00+.byte	0x00,0x00,0x00,0x00 .LSEH_info_sub_n_check_mod_256_epilogue: .byte	1,0,4,0 .byte	0x00,0x74,0x01,0x00

@@ -786,6 +786,23 @@  	ret +.globl	vec_select_32++.def	vec_select_32;+.type	32;+.endef+.p2align	5+vec_select_32:+	dup	v6.2d, x3+	ld1	{v0.2d, v1.2d, v2.2d}, [x1],#48+	cmeq	v6.2d, v6.2d, #0+	ld1	{v3.2d, v4.2d, v5.2d}, [x2],#48+	bit	v0.16b, v3.16b, v6.16b+	bit	v1.16b, v4.16b, v6.16b+	bit	v2.16b, v5.16b, v6.16b+	st1	{v0.2d, v1.2d, v2.2d}, [x0]+	ret+ .globl	vec_select_48  .def	vec_select_48;@@ -977,5 +994,63 @@ 	cmp	x0, x1 	csel	x0, x1, x0, hi 	prfm	pldl1keep, [x0]+	ret++.globl	vec_is_zero_16x++.def	vec_is_zero_16x;+.type	32;+.endef+.p2align	5+vec_is_zero_16x:+	ld1	{v0.2d}, [x0], #16+	lsr	x1, x1, #4+	sub	x1, x1, #1+	cbz	x1, .Loop_is_zero_done++.Loop_is_zero:+	ld1	{v1.2d}, [x0], #16+	orr	v0.16b, v0.16b, v1.16b+	sub	x1, x1, #1+	cbnz	x1, .Loop_is_zero++.Loop_is_zero_done:+	dup	v1.2d, v0.d[1]+	orr	v0.16b, v0.16b, v1.16b+	mov	x1, v0.d[0]+	mov	x0, #1+	cmp	x1, #0+	csel	x0, x0, xzr, eq+	ret++.globl	vec_is_equal_16x++.def	vec_is_equal_16x;+.type	32;+.endef+.p2align	5+vec_is_equal_16x:+	ld1	{v0.2d}, [x0], #16+	ld1	{v1.2d}, [x1], #16+	lsr	x2, x2, #4+	eor	v0.16b, v0.16b, v1.16b++.Loop_is_equal:+	sub	x2, x2, #1+	cbz	x2, .Loop_is_equal_done+	ld1	{v1.2d}, [x0], #16+	ld1	{v2.2d}, [x1], #16+	eor	v1.16b, v1.16b, v2.16b+	orr	v0.16b, v0.16b, v1.16b+	b	.Loop_is_equal+	nop++.Loop_is_equal_done:+	dup	v1.2d, v0.d[1]+	orr	v0.16b, v0.16b, v1.16b+	mov	x1, v0.d[0]+	mov	x0, #1+	cmp	x1, #0+	csel	x0, x0, xzr, eq 	ret 

@@ -10,14 +10,14 @@ 	movq	%rsi,16(%rsp) 	movq	%rsp,%r11 .LSEH_begin_add_mod_384:-	movq	%rcx,%rdi-	movq	%rdx,%rsi-	movq	%r8,%rdx-	movq	%r9,%rcx   	pushq	%rbp +	movq	%rcx,%rdi+	movq	%rdx,%rsi+	movq	%r8,%rdx+	movq	%r9,%rcx 	pushq	%rbx  	pushq	%r12@@ -118,14 +118,14 @@ 	movq	%rsi,16(%rsp) 	movq	%rsp,%r11 .LSEH_begin_add_mod_384x:-	movq	%rcx,%rdi-	movq	%rdx,%rsi-	movq	%r8,%rdx-	movq	%r9,%rcx   	pushq	%rbp +	movq	%rcx,%rdi+	movq	%rdx,%rsi+	movq	%r8,%rdx+	movq	%r9,%rcx 	pushq	%rbx  	pushq	%r12@@ -186,14 +186,14 @@ 	movq	%rsi,16(%rsp) 	movq	%rsp,%r11 .LSEH_begin_rshift_mod_384:-	movq	%rcx,%rdi-	movq	%rdx,%rsi-	movq	%r8,%rdx-	movq	%r9,%rcx   	pushq	%rbp +	movq	%rcx,%rdi+	movq	%rdx,%rsi+	movq	%r8,%rdx+	movq	%r9,%rcx 	pushq	%rbx  	pushq	%r12@@ -315,13 +315,13 @@ 	movq	%rsi,16(%rsp) 	movq	%rsp,%r11 .LSEH_begin_div_by_2_mod_384:-	movq	%rcx,%rdi-	movq	%rdx,%rsi-	movq	%r8,%rdx   	pushq	%rbp +	movq	%rcx,%rdi+	movq	%rdx,%rsi+	movq	%r8,%rdx 	pushq	%rbx  	pushq	%r12@@ -387,14 +387,14 @@ 	movq	%rsi,16(%rsp) 	movq	%rsp,%r11 .LSEH_begin_lshift_mod_384:-	movq	%rcx,%rdi-	movq	%rdx,%rsi-	movq	%r8,%rdx-	movq	%r9,%rcx   	pushq	%rbp +	movq	%rcx,%rdi+	movq	%rdx,%rsi+	movq	%r8,%rdx+	movq	%r9,%rcx 	pushq	%rbx  	pushq	%r12@@ -528,13 +528,13 @@ 	movq	%rsi,16(%rsp) 	movq	%rsp,%r11 .LSEH_begin_mul_by_3_mod_384:-	movq	%rcx,%rdi-	movq	%rdx,%rsi-	movq	%r8,%rdx   	pushq	%rbp +	movq	%rcx,%rdi+	movq	%rdx,%rsi+	movq	%r8,%rdx 	pushq	%rbx  	pushq	%r12@@ -595,13 +595,13 @@ 	movq	%rsi,16(%rsp) 	movq	%rsp,%r11 .LSEH_begin_mul_by_8_mod_384:-	movq	%rcx,%rdi-	movq	%rdx,%rsi-	movq	%r8,%rdx   	pushq	%rbp +	movq	%rcx,%rdi+	movq	%rdx,%rsi+	movq	%r8,%rdx 	pushq	%rbx  	pushq	%r12@@ -669,13 +669,13 @@ 	movq	%rsi,16(%rsp) 	movq	%rsp,%r11 .LSEH_begin_mul_by_3_mod_384x:-	movq	%rcx,%rdi-	movq	%rdx,%rsi-	movq	%r8,%rdx   	pushq	%rbp +	movq	%rcx,%rdi+	movq	%rdx,%rsi+	movq	%r8,%rdx 	pushq	%rbx  	pushq	%r12@@ -752,13 +752,13 @@ 	movq	%rsi,16(%rsp) 	movq	%rsp,%r11 .LSEH_begin_mul_by_8_mod_384x:-	movq	%rcx,%rdi-	movq	%rdx,%rsi-	movq	%r8,%rdx   	pushq	%rbp +	movq	%rcx,%rdi+	movq	%rdx,%rsi+	movq	%r8,%rdx 	pushq	%rbx  	pushq	%r12@@ -845,14 +845,14 @@ 	movq	%rsi,16(%rsp) 	movq	%rsp,%r11 .LSEH_begin_cneg_mod_384:-	movq	%rcx,%rdi-	movq	%rdx,%rsi-	movq	%r8,%rdx-	movq	%r9,%rcx   	pushq	%rbp +	movq	%rcx,%rdi+	movq	%rdx,%rsi+	movq	%r8,%rdx+	movq	%r9,%rcx 	pushq	%rbx  	pushq	%r12@@ -952,14 +952,14 @@ 	movq	%rsi,16(%rsp) 	movq	%rsp,%r11 .LSEH_begin_sub_mod_384:-	movq	%rcx,%rdi-	movq	%rdx,%rsi-	movq	%r8,%rdx-	movq	%r9,%rcx   	pushq	%rbp +	movq	%rcx,%rdi+	movq	%rdx,%rsi+	movq	%r8,%rdx+	movq	%r9,%rcx 	pushq	%rbx  	pushq	%r12@@ -1058,14 +1058,14 @@ 	movq	%rsi,16(%rsp) 	movq	%rsp,%r11 .LSEH_begin_sub_mod_384x:-	movq	%rcx,%rdi-	movq	%rdx,%rsi-	movq	%r8,%rdx-	movq	%r9,%rcx   	pushq	%rbp +	movq	%rcx,%rdi+	movq	%rdx,%rsi+	movq	%r8,%rdx+	movq	%r9,%rcx 	pushq	%rbx  	pushq	%r12@@ -1124,13 +1124,13 @@ 	movq	%rsi,16(%rsp) 	movq	%rsp,%r11 .LSEH_begin_mul_by_1_plus_i_mod_384x:-	movq	%rcx,%rdi-	movq	%rdx,%rsi-	movq	%r8,%rdx   	pushq	%rbp +	movq	%rcx,%rdi+	movq	%rdx,%rsi+	movq	%r8,%rdx 	pushq	%rbx  	pushq	%r12@@ -1274,10 +1274,10 @@ 	movq	%rsi,16(%rsp) 	movq	%rsp,%r11 .LSEH_begin_sgn0_pty_mod_384:-	movq	%rcx,%rdi-	movq	%rdx,%rsi  +	movq	%rcx,%rdi+	movq	%rdx,%rsi .LSEH_body_sgn0_pty_mod_384:  	movq	0(%rdi),%r8@@ -1328,12 +1328,12 @@ 	movq	%rsi,16(%rsp) 	movq	%rsp,%r11 .LSEH_begin_sgn0_pty_mod_384x:-	movq	%rcx,%rdi-	movq	%rdx,%rsi   	pushq	%rbp +	movq	%rcx,%rdi+	movq	%rdx,%rsi 	pushq	%rbx  	subq	$8,%rsp@@ -1439,6 +1439,35 @@ 	.byte	0xf3,0xc3  .LSEH_end_sgn0_pty_mod_384x:+.globl	vec_select_32++.def	vec_select_32;	.scl 2;	.type 32;	.endef+.p2align	5+vec_select_32:+	.byte	0xf3,0x0f,0x1e,0xfa++	movd	%r9d,%xmm5+	pxor	%xmm4,%xmm4+	pshufd	$0,%xmm5,%xmm5+	movdqu	(%rdx),%xmm0+	leaq	16(%rdx),%rdx+	pcmpeqd	%xmm4,%xmm5+	movdqu	(%r8),%xmm1+	leaq	16(%r8),%r8+	pcmpeqd	%xmm5,%xmm4+	leaq	16(%rcx),%rcx+	pand	%xmm4,%xmm0+	movdqu	0+16-16(%rdx),%xmm2+	pand	%xmm5,%xmm1+	movdqu	0+16-16(%r8),%xmm3+	por	%xmm1,%xmm0+	movdqu	%xmm0,0-16(%rcx)+	pand	%xmm4,%xmm2+	pand	%xmm5,%xmm3+	por	%xmm3,%xmm2+	movdqu	%xmm2,16-16(%rcx)+	.byte	0xf3,0xc3+ .globl	vec_select_48  .def	vec_select_48;	.scl 2;	.type 32;	.endef@@ -1854,6 +1883,69 @@ 	prefetchnta	(%rcx) 	.byte	0xf3,0xc3 +.globl	vec_is_zero_16x++.def	vec_is_zero_16x;	.scl 2;	.type 32;	.endef+.p2align	5+vec_is_zero_16x:+	.byte	0xf3,0x0f,0x1e,0xfa++	shrl	$4,%edx+	movdqu	(%rcx),%xmm0+	leaq	16(%rcx),%rcx++.Loop_is_zero:+	decl	%edx+	jz	.Loop_is_zero_done+	movdqu	(%rcx),%xmm1+	leaq	16(%rcx),%rcx+	por	%xmm1,%xmm0+	jmp	.Loop_is_zero++.Loop_is_zero_done:+	pshufd	$0x4e,%xmm0,%xmm1+	por	%xmm1,%xmm0+.byte	102,72,15,126,192+	incl	%edx+	testq	%rax,%rax+	cmovnzl	%edx,%eax+	xorl	$1,%eax+	.byte	0xf3,0xc3++.globl	vec_is_equal_16x++.def	vec_is_equal_16x;	.scl 2;	.type 32;	.endef+.p2align	5+vec_is_equal_16x:+	.byte	0xf3,0x0f,0x1e,0xfa++	shrl	$4,%r8d+	movdqu	(%rcx),%xmm0+	movdqu	(%rdx),%xmm1+	subq	%rcx,%rdx+	leaq	16(%rcx),%rcx+	pxor	%xmm1,%xmm0++.Loop_is_equal:+	decl	%r8d+	jz	.Loop_is_equal_done+	movdqu	(%rcx),%xmm1+	movdqu	(%rcx,%rdx,1),%xmm2+	leaq	16(%rcx),%rcx+	pxor	%xmm2,%xmm1+	por	%xmm1,%xmm0+	jmp	.Loop_is_equal++.Loop_is_equal_done:+	pshufd	$0x4e,%xmm0,%xmm1+	por	%xmm1,%xmm0+.byte	102,72,15,126,192+	incl	%r8d+	testq	%rax,%rax+	cmovnzl	%r8d,%eax+	xorl	$1,%eax+	.byte	0xf3,0xc3+ .section	.pdata .p2align	2 .rva	.LSEH_begin_add_mod_384@@ -2042,8 +2134,9 @@ .byte	1,0,5,0x0b .byte	0,0x74,1,0 .byte	0,0x64,2,0-.byte	0,0x03+.byte	0,0xb3 .byte	0,0+.long	0,0 .LSEH_info_add_mod_384_body: .byte	1,0,17,0 .byte	0x00,0xf4,0x01,0x00@@ -2055,7 +2148,8 @@ .byte	0x00,0x74,0x08,0x00 .byte	0x00,0x64,0x09,0x00 .byte	0x00,0x62-.byte	0x00,0x00+.byte	0x00,0x00,0x00,0x00,0x00,0x00+.byte	0x00,0x00,0x00,0x00 .LSEH_info_add_mod_384_epilogue: .byte	1,0,4,0 .byte	0x00,0x74,0x01,0x00@@ -2066,8 +2160,9 @@ .byte	1,0,5,0x0b .byte	0,0x74,1,0 .byte	0,0x64,2,0-.byte	0,0x03+.byte	0,0xb3 .byte	0,0+.long	0,0 .LSEH_info_add_mod_384x_body: .byte	1,0,17,0 .byte	0x00,0xf4,0x03,0x00@@ -2079,7 +2174,8 @@ .byte	0x00,0x74,0x0a,0x00 .byte	0x00,0x64,0x0b,0x00 .byte	0x00,0x82-.byte	0x00,0x00+.byte	0x00,0x00,0x00,0x00,0x00,0x00+.byte	0x00,0x00,0x00,0x00 .LSEH_info_add_mod_384x_epilogue: .byte	1,0,4,0 .byte	0x00,0x74,0x01,0x00@@ -2090,8 +2186,9 @@ .byte	1,0,5,0x0b .byte	0,0x74,1,0 .byte	0,0x64,2,0-.byte	0,0x03+.byte	0,0xb3 .byte	0,0+.long	0,0 .LSEH_info_rshift_mod_384_body: .byte	1,0,17,0 .byte	0x00,0xf4,0x01,0x00@@ -2103,7 +2200,8 @@ .byte	0x00,0x74,0x08,0x00 .byte	0x00,0x64,0x09,0x00 .byte	0x00,0x62-.byte	0x00,0x00+.byte	0x00,0x00,0x00,0x00,0x00,0x00+.byte	0x00,0x00,0x00,0x00 .LSEH_info_rshift_mod_384_epilogue: .byte	1,0,4,0 .byte	0x00,0x74,0x01,0x00@@ -2114,8 +2212,9 @@ .byte	1,0,5,0x0b .byte	0,0x74,1,0 .byte	0,0x64,2,0-.byte	0,0x03+.byte	0,0xb3 .byte	0,0+.long	0,0 .LSEH_info_div_by_2_mod_384_body: .byte	1,0,17,0 .byte	0x00,0xf4,0x01,0x00@@ -2127,7 +2226,8 @@ .byte	0x00,0x74,0x08,0x00 .byte	0x00,0x64,0x09,0x00 .byte	0x00,0x62-.byte	0x00,0x00+.byte	0x00,0x00,0x00,0x00,0x00,0x00+.byte	0x00,0x00,0x00,0x00 .LSEH_info_div_by_2_mod_384_epilogue: .byte	1,0,4,0 .byte	0x00,0x74,0x01,0x00@@ -2138,8 +2238,9 @@ .byte	1,0,5,0x0b .byte	0,0x74,1,0 .byte	0,0x64,2,0-.byte	0,0x03+.byte	0,0xb3 .byte	0,0+.long	0,0 .LSEH_info_lshift_mod_384_body: .byte	1,0,17,0 .byte	0x00,0xf4,0x01,0x00@@ -2151,7 +2252,8 @@ .byte	0x00,0x74,0x08,0x00 .byte	0x00,0x64,0x09,0x00 .byte	0x00,0x62-.byte	0x00,0x00+.byte	0x00,0x00,0x00,0x00,0x00,0x00+.byte	0x00,0x00,0x00,0x00 .LSEH_info_lshift_mod_384_epilogue: .byte	1,0,4,0 .byte	0x00,0x74,0x01,0x00@@ -2162,8 +2264,9 @@ .byte	1,0,5,0x0b .byte	0,0x74,1,0 .byte	0,0x64,2,0-.byte	0,0x03+.byte	0,0xb3 .byte	0,0+.long	0,0 .LSEH_info_mul_by_3_mod_384_body: .byte	1,0,17,0 .byte	0x00,0xf4,0x01,0x00@@ -2175,7 +2278,8 @@ .byte	0x00,0x74,0x08,0x00 .byte	0x00,0x64,0x09,0x00 .byte	0x00,0x62-.byte	0x00,0x00+.byte	0x00,0x00,0x00,0x00,0x00,0x00+.byte	0x00,0x00,0x00,0x00 .LSEH_info_mul_by_3_mod_384_epilogue: .byte	1,0,4,0 .byte	0x00,0x74,0x01,0x00@@ -2186,8 +2290,9 @@ .byte	1,0,5,0x0b .byte	0,0x74,1,0 .byte	0,0x64,2,0-.byte	0,0x03+.byte	0,0xb3 .byte	0,0+.long	0,0 .LSEH_info_mul_by_8_mod_384_body: .byte	1,0,17,0 .byte	0x00,0xf4,0x01,0x00@@ -2199,7 +2304,8 @@ .byte	0x00,0x74,0x08,0x00 .byte	0x00,0x64,0x09,0x00 .byte	0x00,0x62-.byte	0x00,0x00+.byte	0x00,0x00,0x00,0x00,0x00,0x00+.byte	0x00,0x00,0x00,0x00 .LSEH_info_mul_by_8_mod_384_epilogue: .byte	1,0,4,0 .byte	0x00,0x74,0x01,0x00@@ -2210,8 +2316,9 @@ .byte	1,0,5,0x0b .byte	0,0x74,1,0 .byte	0,0x64,2,0-.byte	0,0x03+.byte	0,0xb3 .byte	0,0+.long	0,0 .LSEH_info_mul_by_3_mod_384x_body: .byte	1,0,17,0 .byte	0x00,0xf4,0x01,0x00@@ -2223,7 +2330,8 @@ .byte	0x00,0x74,0x08,0x00 .byte	0x00,0x64,0x09,0x00 .byte	0x00,0x62-.byte	0x00,0x00+.byte	0x00,0x00,0x00,0x00,0x00,0x00+.byte	0x00,0x00,0x00,0x00 .LSEH_info_mul_by_3_mod_384x_epilogue: .byte	1,0,4,0 .byte	0x00,0x74,0x01,0x00@@ -2234,8 +2342,9 @@ .byte	1,0,5,0x0b .byte	0,0x74,1,0 .byte	0,0x64,2,0-.byte	0,0x03+.byte	0,0xb3 .byte	0,0+.long	0,0 .LSEH_info_mul_by_8_mod_384x_body: .byte	1,0,17,0 .byte	0x00,0xf4,0x01,0x00@@ -2247,7 +2356,8 @@ .byte	0x00,0x74,0x08,0x00 .byte	0x00,0x64,0x09,0x00 .byte	0x00,0x62-.byte	0x00,0x00+.byte	0x00,0x00,0x00,0x00,0x00,0x00+.byte	0x00,0x00,0x00,0x00 .LSEH_info_mul_by_8_mod_384x_epilogue: .byte	1,0,4,0 .byte	0x00,0x74,0x01,0x00@@ -2258,8 +2368,9 @@ .byte	1,0,5,0x0b .byte	0,0x74,1,0 .byte	0,0x64,2,0-.byte	0,0x03+.byte	0,0xb3 .byte	0,0+.long	0,0 .LSEH_info_cneg_mod_384_body: .byte	1,0,17,0 .byte	0x00,0xf4,0x01,0x00@@ -2271,7 +2382,8 @@ .byte	0x00,0x74,0x08,0x00 .byte	0x00,0x64,0x09,0x00 .byte	0x00,0x62-.byte	0x00,0x00+.byte	0x00,0x00,0x00,0x00,0x00,0x00+.byte	0x00,0x00,0x00,0x00 .LSEH_info_cneg_mod_384_epilogue: .byte	1,0,4,0 .byte	0x00,0x74,0x01,0x00@@ -2282,8 +2394,9 @@ .byte	1,0,5,0x0b .byte	0,0x74,1,0 .byte	0,0x64,2,0-.byte	0,0x03+.byte	0,0xb3 .byte	0,0+.long	0,0 .LSEH_info_sub_mod_384_body: .byte	1,0,17,0 .byte	0x00,0xf4,0x01,0x00@@ -2295,7 +2408,8 @@ .byte	0x00,0x74,0x08,0x00 .byte	0x00,0x64,0x09,0x00 .byte	0x00,0x62-.byte	0x00,0x00+.byte	0x00,0x00,0x00,0x00,0x00,0x00+.byte	0x00,0x00,0x00,0x00 .LSEH_info_sub_mod_384_epilogue: .byte	1,0,4,0 .byte	0x00,0x74,0x01,0x00@@ -2306,8 +2420,9 @@ .byte	1,0,5,0x0b .byte	0,0x74,1,0 .byte	0,0x64,2,0-.byte	0,0x03+.byte	0,0xb3 .byte	0,0+.long	0,0 .LSEH_info_sub_mod_384x_body: .byte	1,0,17,0 .byte	0x00,0xf4,0x03,0x00@@ -2319,7 +2434,8 @@ .byte	0x00,0x74,0x0a,0x00 .byte	0x00,0x64,0x0b,0x00 .byte	0x00,0x82-.byte	0x00,0x00+.byte	0x00,0x00,0x00,0x00,0x00,0x00+.byte	0x00,0x00,0x00,0x00 .LSEH_info_sub_mod_384x_epilogue: .byte	1,0,4,0 .byte	0x00,0x74,0x01,0x00@@ -2330,8 +2446,9 @@ .byte	1,0,5,0x0b .byte	0,0x74,1,0 .byte	0,0x64,2,0-.byte	0,0x03+.byte	0,0xb3 .byte	0,0+.long	0,0 .LSEH_info_mul_by_1_plus_i_mod_384x_body: .byte	1,0,17,0 .byte	0x00,0xf4,0x07,0x00@@ -2343,7 +2460,8 @@ .byte	0x00,0x74,0x0e,0x00 .byte	0x00,0x64,0x0f,0x00 .byte	0x00,0xc2-.byte	0x00,0x00+.byte	0x00,0x00,0x00,0x00,0x00,0x00+.byte	0x00,0x00,0x00,0x00 .LSEH_info_mul_by_1_plus_i_mod_384x_epilogue: .byte	1,0,4,0 .byte	0x00,0x74,0x01,0x00@@ -2354,8 +2472,9 @@ .byte	1,0,5,0x0b .byte	0,0x74,1,0 .byte	0,0x64,2,0-.byte	0,0x03+.byte	0,0xb3 .byte	0,0+.long	0,0 .LSEH_info_sgn0_pty_mod_384_body: .byte	1,0,4,0 .byte	0x00,0x74,0x01,0x00@@ -2371,8 +2490,9 @@ .byte	1,0,5,0x0b .byte	0,0x74,1,0 .byte	0,0x64,2,0-.byte	0,0x03+.byte	0,0xb3 .byte	0,0+.long	0,0 .LSEH_info_sgn0_pty_mod_384x_body: .byte	1,0,9,0 .byte	0x00,0x34,0x01,0x00@@ -2380,7 +2500,8 @@ .byte	0x00,0x74,0x04,0x00 .byte	0x00,0x64,0x05,0x00 .byte	0x00,0x22-.byte	0x00,0x00+.byte	0x00,0x00,0x00,0x00,0x00,0x00+.byte	0x00,0x00,0x00,0x00 .LSEH_info_sgn0_pty_mod_384x_epilogue: .byte	1,0,4,0 .byte	0x00,0x74,0x01,0x00

@@ -145,14 +145,14 @@ 	movq	%rsi,16(%rsp) 	movq	%rsp,%r11 .LSEH_begin_add_mod_384x384:-	movq	%rcx,%rdi-	movq	%rdx,%rsi-	movq	%r8,%rdx-	movq	%r9,%rcx   	pushq	%rbp +	movq	%rcx,%rdi+	movq	%rdx,%rsi+	movq	%r8,%rdx+	movq	%r9,%rcx 	pushq	%rbx  	pushq	%r12@@ -202,14 +202,14 @@ 	movq	%rsi,16(%rsp) 	movq	%rsp,%r11 .LSEH_begin_sub_mod_384x384:-	movq	%rcx,%rdi-	movq	%rdx,%rsi-	movq	%r8,%rdx-	movq	%r9,%rcx   	pushq	%rbp +	movq	%rcx,%rdi+	movq	%rdx,%rsi+	movq	%r8,%rdx+	movq	%r9,%rcx 	pushq	%rbx  	pushq	%r12@@ -280,8 +280,9 @@ .byte	1,0,5,0x0b .byte	0,0x74,1,0 .byte	0,0x64,2,0-.byte	0,0x03+.byte	0,0xb3 .byte	0,0+.long	0,0 .LSEH_info_add_mod_384x384_body: .byte	1,0,17,0 .byte	0x00,0xf4,0x01,0x00@@ -293,7 +294,8 @@ .byte	0x00,0x74,0x08,0x00 .byte	0x00,0x64,0x09,0x00 .byte	0x00,0x62-.byte	0x00,0x00+.byte	0x00,0x00,0x00,0x00,0x00,0x00+.byte	0x00,0x00,0x00,0x00 .LSEH_info_add_mod_384x384_epilogue: .byte	1,0,4,0 .byte	0x00,0x74,0x01,0x00@@ -304,8 +306,9 @@ .byte	1,0,5,0x0b .byte	0,0x74,1,0 .byte	0,0x64,2,0-.byte	0,0x03+.byte	0,0xb3 .byte	0,0+.long	0,0 .LSEH_info_sub_mod_384x384_body: .byte	1,0,17,0 .byte	0x00,0xf4,0x01,0x00@@ -317,7 +320,8 @@ .byte	0x00,0x74,0x08,0x00 .byte	0x00,0x64,0x09,0x00 .byte	0x00,0x62-.byte	0x00,0x00+.byte	0x00,0x00,0x00,0x00,0x00,0x00+.byte	0x00,0x00,0x00,0x00 .LSEH_info_sub_mod_384x384_epilogue: .byte	1,0,4,0 .byte	0x00,0x74,0x01,0x00

@@ -1,6 +1,7 @@ .text  .globl	ct_inverse_mod_256+ .def	ct_inverse_mod_256; .type	32; .endef@@ -62,14 +63,14 @@ 	madd	x4, x16, x8, xzr	// |u|*|f0| 	madd	x4, x17, x9, x4	// |v|*|g0| 	str	x4, [x0,#8*4]-	asr	x5, x4, #63		// sign extenstion+	asr	x5, x4, #63		// sign extension 	stp	x5, x5, [x0,#8*5] 	stp	x5, x5, [x0,#8*7]  	madd	x4, x12, x8, xzr	// |u|*|f1| 	madd	x4, x13, x9, x4	// |v|*|g1| 	str	x4, [x0,#8*9]-	asr	x5, x4, #63		// sign extenstion+	asr	x5, x4, #63		// sign extension 	stp	x5, x5, [x0,#8*10] 	stp	x5, x5, [x0,#8*12] 	eor	x1, x1, #256		// flip-flop src |a|b|u|v|

@@ -1,6 +1,7 @@ .text	  .globl	ct_inverse_mod_256+ .def	ct_inverse_mod_256;	.scl 2;	.type 32;	.endef .p2align	5 ct_inverse_mod_256:@@ -9,14 +10,14 @@ 	movq	%rsi,16(%rsp) 	movq	%rsp,%r11 .LSEH_begin_ct_inverse_mod_256:-	movq	%rcx,%rdi-	movq	%rdx,%rsi-	movq	%r8,%rdx-	movq	%r9,%rcx   	pushq	%rbp +	movq	%rcx,%rdi+	movq	%rdx,%rsi+	movq	%r8,%rdx+	movq	%r9,%rcx 	pushq	%rbx  	pushq	%r12@@ -1188,8 +1189,9 @@ .byte	1,0,5,0x0b .byte	0,0x74,1,0 .byte	0,0x64,2,0-.byte	0,0x03+.byte	0,0xb3 .byte	0,0+.long	0,0 .LSEH_info_ct_inverse_mod_256_body: .byte	1,0,18,0 .byte	0x00,0xf4,0x86,0x00@@ -1201,6 +1203,8 @@ .byte	0x00,0x74,0x8d,0x00 .byte	0x00,0x64,0x8e,0x00 .byte	0x00,0x01,0x8c,0x00+.byte	0x00,0x00,0x00,0x00+.byte	0x00,0x00,0x00,0x00 .LSEH_info_ct_inverse_mod_256_epilogue: .byte	1,0,4,0 .byte	0x00,0x74,0x01,0x00

@@ -1,6 +1,7 @@ .text  .globl	ct_inverse_mod_383+ .def	ct_inverse_mod_383; .type	32; .endef@@ -73,7 +74,7 @@ 	adds	x3, x3, x5 	adc	x4, x4, x6 	stp	x3, x4, [x0,#8*6]-	asr	x5, x4, #63		// sign extenstion+	asr	x5, x4, #63		// sign extension 	stp	x5, x5, [x0,#8*8] 	stp	x5, x5, [x0,#8*10] @@ -84,7 +85,7 @@ 	adds	x3, x3, x5 	adc	x4, x4, x6 	stp	x3, x4, [x0,#8*12]-	asr	x5, x4, #63		// sign extenstion+	asr	x5, x4, #63		// sign extension 	stp	x5, x5, [x0,#8*14] 	stp	x5, x5, [x0,#8*16] 	eor	x1, x1, #256		// flip-flop src |a|b|u|v|

@@ -1,6 +1,7 @@ .text  .globl	ct_is_square_mod_384+ .def	ct_is_square_mod_384; .type	32; .endef

@@ -1,6 +1,7 @@ .text	  .globl	ct_is_square_mod_384+ .def	ct_is_square_mod_384;	.scl 2;	.type 32;	.endef .p2align	5 ct_is_square_mod_384:@@ -9,12 +10,12 @@ 	movq	%rsi,16(%rsp) 	movq	%rsp,%r11 .LSEH_begin_ct_is_square_mod_384:-	movq	%rcx,%rdi-	movq	%rdx,%rsi   	pushq	%rbp +	movq	%rcx,%rdi+	movq	%rdx,%rsi 	pushq	%rbx  	pushq	%r12@@ -484,8 +485,9 @@ .byte	1,0,5,0x0b .byte	0,0x74,1,0 .byte	0,0x64,2,0-.byte	0,0x03+.byte	0,0xb3 .byte	0,0+.long	0,0 .LSEH_info_ct_is_square_mod_384_body: .byte	1,0,18,0 .byte	0x00,0xf4,0x43,0x00@@ -497,6 +499,8 @@ .byte	0x00,0x74,0x4a,0x00 .byte	0x00,0x64,0x4b,0x00 .byte	0x00,0x01,0x49,0x00+.byte	0x00,0x00,0x00,0x00+.byte	0x00,0x00,0x00,0x00 .LSEH_info_ct_is_square_mod_384_epilogue: .byte	1,0,4,0 .byte	0x00,0x74,0x01,0x00

@@ -1,6 +1,8 @@+.comm	__blst_platform_cap,4 .text	  .globl	ct_inverse_mod_383+ .def	ct_inverse_mod_383;	.scl 2;	.type 32;	.endef .p2align	5 ct_inverse_mod_383:@@ -9,12 +11,16 @@ 	movq	%rsi,16(%rsp) 	movq	%rsp,%r11 .LSEH_begin_ct_inverse_mod_383:++ 	movq	%rcx,%rdi 	movq	%rdx,%rsi 	movq	%r8,%rdx 	movq	%r9,%rcx--+#ifdef __BLST_PORTABLE__+	testl	$1,__blst_platform_cap(%rip)+	jnz	ct_inverse_mod_383$1+#endif 	pushq	%rbp  	pushq	%rbx@@ -1200,8 +1206,9 @@ .byte	1,0,5,0x0b .byte	0,0x74,1,0 .byte	0,0x64,2,0-.byte	0,0x03+.byte	0,0xb3 .byte	0,0+.long	0,0 .LSEH_info_ct_inverse_mod_383_body: .byte	1,0,18,0 .byte	0x00,0xf4,0x8b,0x00@@ -1213,6 +1220,8 @@ .byte	0x00,0x74,0x92,0x00 .byte	0x00,0x64,0x93,0x00 .byte	0x00,0x01,0x91,0x00+.byte	0x00,0x00,0x00,0x00+.byte	0x00,0x00,0x00,0x00 .LSEH_info_ct_inverse_mod_383_epilogue: .byte	1,0,4,0 .byte	0x00,0x74,0x01,0x00

@@ -1,6 +1,7 @@ .text	  .globl	ctx_inverse_mod_383+ .def	ctx_inverse_mod_383;	.scl 2;	.type 32;	.endef .p2align	5 ctx_inverse_mod_383:@@ -9,12 +10,13 @@ 	movq	%rsi,16(%rsp) 	movq	%rsp,%r11 .LSEH_begin_ctx_inverse_mod_383:++ 	movq	%rcx,%rdi 	movq	%rdx,%rsi 	movq	%r8,%rdx 	movq	%r9,%rcx--+ct_inverse_mod_383$1: 	pushq	%rbp  	pushq	%rbx@@ -812,7 +814,7 @@  	movq	48(%rsi),%r10 -	call	__inner_loop_62+	call	__tail_loop_53   @@ -1514,9 +1516,9 @@ 	.byte	0xf3,0xc3  -.def	__inner_loop_62;	.scl 3;	.type 32;	.endef+.def	__tail_loop_53;	.scl 3;	.type 32;	.endef .p2align	5-__inner_loop_62:+__tail_loop_53: 	.byte	0xf3,0x0f,0x1e,0xfa  	movq	$1,%rdx@@ -1524,7 +1526,7 @@ 	xorq	%r12,%r12 	movq	$1,%r13 -.Loop_62:+.Loop_53: 	xorq	%rax,%rax 	testq	$1,%r8 	movq	%r10,%rbx@@ -1551,7 +1553,7 @@ 	subq	%rax,%rdx 	subq	%rbx,%rcx 	subl	$1,%edi-	jnz	.Loop_62+	jnz	.Loop_53  	.byte	0xf3,0xc3 @@ -1575,8 +1577,9 @@ .byte	1,0,5,0x0b .byte	0,0x74,1,0 .byte	0,0x64,2,0-.byte	0,0x03+.byte	0,0xb3 .byte	0,0+.long	0,0 .LSEH_info_ctx_inverse_mod_383_body: .byte	1,0,18,0 .byte	0x00,0xf4,0x8b,0x00@@ -1588,6 +1591,8 @@ .byte	0x00,0x74,0x92,0x00 .byte	0x00,0x64,0x93,0x00 .byte	0x00,0x01,0x91,0x00+.byte	0x00,0x00,0x00,0x00+.byte	0x00,0x00,0x00,0x00 .LSEH_info_ctx_inverse_mod_383_epilogue: .byte	1,0,4,0 .byte	0x00,0x74,0x01,0x00

@@ -27,7 +27,7 @@ 	asr	x3,x0,#63	// top bit -> mask 	add	x0,x0,x0	// Q <<= 1 	subs	x6,x4,x1	// R - D-	add	x0,x0,#1	// Q + specilative bit+	add	x0,x0,#1	// Q + speculative bit 	sbcs	x7,x5,x2 	sbc	x0,x0,xzr	// subtract speculative bit 

@@ -8,11 +8,14 @@ 	.byte	0xf3,0x0f,0x1e,0xfa 	movq	%rdi,8(%rsp) 	movq	%rsi,16(%rsp)-	movq	%rsp,%rax+	movq	%rsp,%r11 .LSEH_begin_div_3_limbs:++ 	movq	%rcx,%rdi 	movq	%rdx,%rsi 	movq	%r8,%rdx+.LSEH_body_div_3_limbs:  	movq	(%rdi),%r8 	movq	8(%rdi),%r9@@ -45,9 +48,12 @@  	orq	%rcx,%rax -	movq	8(%rsp),%rdi-	movq	16(%rsp),%rsi+.LSEH_epilogue_div_3_limbs:+	mov	8(%rsp),%rdi+	mov	16(%rsp),%rsi+ 	.byte	0xf3,0xc3+ .LSEH_end_div_3_limbs: .globl	quot_rem_128 @@ -57,11 +63,14 @@ 	.byte	0xf3,0x0f,0x1e,0xfa 	movq	%rdi,8(%rsp) 	movq	%rsi,16(%rsp)-	movq	%rsp,%rax+	movq	%rsp,%r11 .LSEH_begin_quot_rem_128:++ 	movq	%rcx,%rdi 	movq	%rdx,%rsi 	movq	%r8,%rdx+.LSEH_body_quot_rem_128:  	movq	%rdx,%rax 	movq	%rdx,%rcx@@ -97,9 +106,12 @@  	movq	%rcx,%rax -	movq	8(%rsp),%rdi-	movq	16(%rsp),%rsi+.LSEH_epilogue_quot_rem_128:+	mov	8(%rsp),%rdi+	mov	16(%rsp),%rsi+ 	.byte	0xf3,0xc3+ .LSEH_end_quot_rem_128:  @@ -114,11 +126,14 @@ 	.byte	0xf3,0x0f,0x1e,0xfa 	movq	%rdi,8(%rsp) 	movq	%rsi,16(%rsp)-	movq	%rsp,%rax+	movq	%rsp,%r11 .LSEH_begin_quot_rem_64:++ 	movq	%rcx,%rdi 	movq	%rdx,%rsi 	movq	%r8,%rdx+.LSEH_body_quot_rem_64:  	movq	%rdx,%rax 	imulq	0(%rsi),%rdx@@ -130,11 +145,104 @@ 	movq	%r10,0(%rdi) 	movq	%rax,8(%rdi) -	movq	8(%rsp),%rdi-	movq	16(%rsp),%rsi+.LSEH_epilogue_quot_rem_64:+	mov	8(%rsp),%rdi+	mov	16(%rsp),%rsi+ 	.byte	0xf3,0xc3+ .LSEH_end_quot_rem_64: .section	.pdata .p2align	2+.rva	.LSEH_begin_div_3_limbs+.rva	.LSEH_body_div_3_limbs+.rva	.LSEH_info_div_3_limbs_prologue++.rva	.LSEH_body_div_3_limbs+.rva	.LSEH_epilogue_div_3_limbs+.rva	.LSEH_info_div_3_limbs_body++.rva	.LSEH_epilogue_div_3_limbs+.rva	.LSEH_end_div_3_limbs+.rva	.LSEH_info_div_3_limbs_epilogue++.rva	.LSEH_begin_quot_rem_128+.rva	.LSEH_body_quot_rem_128+.rva	.LSEH_info_quot_rem_128_prologue++.rva	.LSEH_body_quot_rem_128+.rva	.LSEH_epilogue_quot_rem_128+.rva	.LSEH_info_quot_rem_128_body++.rva	.LSEH_epilogue_quot_rem_128+.rva	.LSEH_end_quot_rem_128+.rva	.LSEH_info_quot_rem_128_epilogue++.rva	.LSEH_begin_quot_rem_64+.rva	.LSEH_body_quot_rem_64+.rva	.LSEH_info_quot_rem_64_prologue++.rva	.LSEH_body_quot_rem_64+.rva	.LSEH_epilogue_quot_rem_64+.rva	.LSEH_info_quot_rem_64_body++.rva	.LSEH_epilogue_quot_rem_64+.rva	.LSEH_end_quot_rem_64+.rva	.LSEH_info_quot_rem_64_epilogue+ .section	.xdata .p2align	3+.LSEH_info_div_3_limbs_prologue:+.byte	1,0,5,0x0b+.byte	0,0x74,1,0+.byte	0,0x64,2,0+.byte	0,0xb3+.byte	0,0+.long	0,0+.LSEH_info_div_3_limbs_body:+.byte	1,0,4,0+.byte	0x00,0x74,0x01,0x00+.byte	0x00,0x64,0x02,0x00+.byte	0x00,0x00,0x00,0x00+.LSEH_info_div_3_limbs_epilogue:+.byte	1,0,4,0+.byte	0x00,0x74,0x01,0x00+.byte	0x00,0x64,0x02,0x00+.byte	0x00,0x00,0x00,0x00++.LSEH_info_quot_rem_128_prologue:+.byte	1,0,5,0x0b+.byte	0,0x74,1,0+.byte	0,0x64,2,0+.byte	0,0xb3+.byte	0,0+.long	0,0+.LSEH_info_quot_rem_128_body:+.byte	1,0,4,0+.byte	0x00,0x74,0x01,0x00+.byte	0x00,0x64,0x02,0x00+.byte	0x00,0x00,0x00,0x00+.LSEH_info_quot_rem_128_epilogue:+.byte	1,0,4,0+.byte	0x00,0x74,0x01,0x00+.byte	0x00,0x64,0x02,0x00+.byte	0x00,0x00,0x00,0x00++.LSEH_info_quot_rem_64_prologue:+.byte	1,0,5,0x0b+.byte	0,0x74,1,0+.byte	0,0x64,2,0+.byte	0,0xb3+.byte	0,0+.long	0,0+.LSEH_info_quot_rem_64_body:+.byte	1,0,4,0+.byte	0x00,0x74,0x01,0x00+.byte	0x00,0x64,0x02,0x00+.byte	0x00,0x00,0x00,0x00+.LSEH_info_quot_rem_64_epilogue:+.byte	1,0,4,0+.byte	0x00,0x74,0x01,0x00+.byte	0x00,0x64,0x02,0x00+.byte	0x00,0x00,0x00,0x00+

@@ -1,3 +1,4 @@+.comm	__blst_platform_cap,4 .text	  .globl	mul_mont_sparse_256@@ -10,13 +11,17 @@ 	movq	%rsi,16(%rsp) 	movq	%rsp,%r11 .LSEH_begin_mul_mont_sparse_256:++ 	movq	%rcx,%rdi 	movq	%rdx,%rsi 	movq	%r8,%rdx 	movq	%r9,%rcx 	movq	40(%rsp),%r8--+#ifdef __BLST_PORTABLE__+	testl	$1,__blst_platform_cap(%rip)+	jnz	mul_mont_sparse_256$1+#endif 	pushq	%rbp  	pushq	%rbx@@ -80,12 +85,16 @@ 	movq	%rsi,16(%rsp) 	movq	%rsp,%r11 .LSEH_begin_sqr_mont_sparse_256:++ 	movq	%rcx,%rdi 	movq	%rdx,%rsi 	movq	%r8,%rdx 	movq	%r9,%rcx--+#ifdef __BLST_PORTABLE__+	testl	$1,__blst_platform_cap(%rip)+	jnz	sqr_mont_sparse_256$1+#endif 	pushq	%rbp  	pushq	%rbx@@ -430,12 +439,16 @@ 	movq	%rsi,16(%rsp) 	movq	%rsp,%r11 .LSEH_begin_from_mont_256:++ 	movq	%rcx,%rdi 	movq	%rdx,%rsi 	movq	%r8,%rdx 	movq	%r9,%rcx--+#ifdef __BLST_PORTABLE__+	testl	$1,__blst_platform_cap(%rip)+	jnz	from_mont_256$1+#endif 	pushq	%rbp  	pushq	%rbx@@ -510,12 +523,16 @@ 	movq	%rsi,16(%rsp) 	movq	%rsp,%r11 .LSEH_begin_redc_mont_256:++ 	movq	%rcx,%rdi 	movq	%rdx,%rsi 	movq	%r8,%rdx 	movq	%r9,%rcx--+#ifdef __BLST_PORTABLE__+	testl	$1,__blst_platform_cap(%rip)+	jnz	redc_mont_256$1+#endif 	pushq	%rbp  	pushq	%rbx@@ -778,8 +795,9 @@ .byte	1,0,5,0x0b .byte	0,0x74,1,0 .byte	0,0x64,2,0-.byte	0,0x03+.byte	0,0xb3 .byte	0,0+.long	0,0 .LSEH_info_mul_mont_sparse_256_body: .byte	1,0,17,0 .byte	0x00,0xf4,0x01,0x00@@ -791,7 +809,8 @@ .byte	0x00,0x74,0x08,0x00 .byte	0x00,0x64,0x09,0x00 .byte	0x00,0x62-.byte	0x00,0x00+.byte	0x00,0x00,0x00,0x00,0x00,0x00+.byte	0x00,0x00,0x00,0x00 .LSEH_info_mul_mont_sparse_256_epilogue: .byte	1,0,4,0 .byte	0x00,0x74,0x01,0x00@@ -802,8 +821,9 @@ .byte	1,0,5,0x0b .byte	0,0x74,1,0 .byte	0,0x64,2,0-.byte	0,0x03+.byte	0,0xb3 .byte	0,0+.long	0,0 .LSEH_info_sqr_mont_sparse_256_body: .byte	1,0,17,0 .byte	0x00,0xf4,0x01,0x00@@ -815,7 +835,8 @@ .byte	0x00,0x74,0x08,0x00 .byte	0x00,0x64,0x09,0x00 .byte	0x00,0x62-.byte	0x00,0x00+.byte	0x00,0x00,0x00,0x00,0x00,0x00+.byte	0x00,0x00,0x00,0x00 .LSEH_info_sqr_mont_sparse_256_epilogue: .byte	1,0,4,0 .byte	0x00,0x74,0x01,0x00@@ -826,8 +847,9 @@ .byte	1,0,5,0x0b .byte	0,0x74,1,0 .byte	0,0x64,2,0-.byte	0,0x03+.byte	0,0xb3 .byte	0,0+.long	0,0 .LSEH_info_from_mont_256_body: .byte	1,0,17,0 .byte	0x00,0xf4,0x01,0x00@@ -839,7 +861,8 @@ .byte	0x00,0x74,0x08,0x00 .byte	0x00,0x64,0x09,0x00 .byte	0x00,0x62-.byte	0x00,0x00+.byte	0x00,0x00,0x00,0x00,0x00,0x00+.byte	0x00,0x00,0x00,0x00 .LSEH_info_from_mont_256_epilogue: .byte	1,0,4,0 .byte	0x00,0x74,0x01,0x00@@ -850,8 +873,9 @@ .byte	1,0,5,0x0b .byte	0,0x74,1,0 .byte	0,0x64,2,0-.byte	0,0x03+.byte	0,0xb3 .byte	0,0+.long	0,0 .LSEH_info_redc_mont_256_body: .byte	1,0,17,0 .byte	0x00,0xf4,0x01,0x00@@ -863,7 +887,8 @@ .byte	0x00,0x74,0x08,0x00 .byte	0x00,0x64,0x09,0x00 .byte	0x00,0x62-.byte	0x00,0x00+.byte	0x00,0x00,0x00,0x00,0x00,0x00+.byte	0x00,0x00,0x00,0x00 .LSEH_info_redc_mont_256_epilogue: .byte	1,0,4,0 .byte	0x00,0x74,0x01,0x00

@@ -1,3 +1,4 @@+.comm	__blst_platform_cap,4 .text	  @@ -6,9 +7,9 @@   -.def	__sub_mod_384x384;	.scl 3;	.type 32;	.endef+.def	__subq_mod_384x384;	.scl 3;	.type 32;	.endef .p2align	5-__sub_mod_384x384:+__subq_mod_384x384: 	.byte	0xf3,0x0f,0x1e,0xfa  	movq	0(%rsi),%r8@@ -73,9 +74,9 @@ 	.byte	0xf3,0xc3  -.def	__add_mod_384;	.scl 3;	.type 32;	.endef+.def	__addq_mod_384;	.scl 3;	.type 32;	.endef .p2align	5-__add_mod_384:+__addq_mod_384: 	.byte	0xf3,0x0f,0x1e,0xfa  	movq	0(%rsi),%r8@@ -123,9 +124,9 @@ 	.byte	0xf3,0xc3  -.def	__sub_mod_384;	.scl 3;	.type 32;	.endef+.def	__subq_mod_384;	.scl 3;	.type 32;	.endef .p2align	5-__sub_mod_384:+__subq_mod_384: 	.byte	0xf3,0x0f,0x1e,0xfa  	movq	0(%rsi),%r8@@ -135,7 +136,7 @@ 	movq	32(%rsi),%r12 	movq	40(%rsi),%r13 -__sub_mod_384_a_is_loaded:+__subq_mod_384_a_is_loaded: 	subq	0(%rdx),%r8 	movq	0(%rcx),%r14 	sbbq	8(%rdx),%r9@@ -182,13 +183,17 @@ 	movq	%rsi,16(%rsp) 	movq	%rsp,%r11 .LSEH_begin_mul_mont_384x:++ 	movq	%rcx,%rdi 	movq	%rdx,%rsi 	movq	%r8,%rdx 	movq	%r9,%rcx 	movq	40(%rsp),%r8--+#ifdef __BLST_PORTABLE__+	testl	$1,__blst_platform_cap(%rip)+	jnz	mul_mont_384x$1+#endif 	pushq	%rbp  	pushq	%rbx@@ -229,12 +234,12 @@ 	movq	8(%rsp),%rcx 	leaq	-48(%rsi),%rdx 	leaq	40+192+48(%rsp),%rdi-	call	__add_mod_384+	call	__addq_mod_384  	movq	16(%rsp),%rsi 	leaq	48(%rsi),%rdx 	leaq	-48(%rdi),%rdi-	call	__add_mod_384+	call	__addq_mod_384  	leaq	(%rdi),%rbx 	leaq	48(%rdi),%rsi@@ -244,17 +249,17 @@ 	leaq	(%rdi),%rsi 	leaq	40(%rsp),%rdx 	movq	8(%rsp),%rcx-	call	__sub_mod_384x384+	call	__subq_mod_384x384  	leaq	(%rdi),%rsi 	leaq	-96(%rdi),%rdx-	call	__sub_mod_384x384+	call	__subq_mod_384x384   	leaq	40(%rsp),%rsi 	leaq	40+96(%rsp),%rdx 	leaq	40(%rsp),%rdi-	call	__sub_mod_384x384+	call	__subq_mod_384x384  	movq	%rcx,%rbx @@ -263,14 +268,14 @@ 	movq	0(%rsp),%rcx 	movq	32(%rsp),%rdi 	call	__mulq_by_1_mont_384-	call	__redc_tail_mont_384+	call	__redq_tail_mont_384   	leaq	40+192(%rsp),%rsi 	movq	0(%rsp),%rcx 	leaq	48(%rdi),%rdi 	call	__mulq_by_1_mont_384-	call	__redc_tail_mont_384+	call	__redq_tail_mont_384  	leaq	328(%rsp),%r8 	movq	0(%r8),%r15@@ -304,12 +309,16 @@ 	movq	%rsi,16(%rsp) 	movq	%rsp,%r11 .LSEH_begin_sqr_mont_384x:++ 	movq	%rcx,%rdi 	movq	%rdx,%rsi 	movq	%r8,%rdx 	movq	%r9,%rcx--+#ifdef __BLST_PORTABLE__+	testl	$1,__blst_platform_cap(%rip)+	jnz	sqr_mont_384x$1+#endif 	pushq	%rbp  	pushq	%rbx@@ -335,13 +344,13 @@  	leaq	48(%rsi),%rdx 	leaq	32(%rsp),%rdi-	call	__add_mod_384+	call	__addq_mod_384   	movq	16(%rsp),%rsi 	leaq	48(%rsi),%rdx 	leaq	32+48(%rsp),%rdi-	call	__sub_mod_384+	call	__subq_mod_384   	movq	16(%rsp),%rsi@@ -433,12 +442,16 @@ 	movq	%rsi,16(%rsp) 	movq	%rsp,%r11 .LSEH_begin_mul_382x:++ 	movq	%rcx,%rdi 	movq	%rdx,%rsi 	movq	%r8,%rdx 	movq	%r9,%rcx--+#ifdef __BLST_PORTABLE__+	testl	$1,__blst_platform_cap(%rip)+	jnz	mul_382x$1+#endif 	pushq	%rbp  	pushq	%rbx@@ -528,18 +541,18 @@ 	leaq	32(%rsp),%rdx 	movq	24(%rsp),%rcx 	movq	%rsi,%rdi-	call	__sub_mod_384x384+	call	__subq_mod_384x384   	leaq	0(%rdi),%rsi 	leaq	-96(%rdi),%rdx-	call	__sub_mod_384x384+	call	__subq_mod_384x384   	leaq	-96(%rdi),%rsi 	leaq	32(%rsp),%rdx 	leaq	-96(%rdi),%rdi-	call	__sub_mod_384x384+	call	__subq_mod_384x384  	leaq	136(%rsp),%r8 	movq	0(%r8),%r15@@ -573,11 +586,15 @@ 	movq	%rsi,16(%rsp) 	movq	%rsp,%r11 .LSEH_begin_sqr_382x:++ 	movq	%rcx,%rdi 	movq	%rdx,%rsi 	movq	%r8,%rdx--+#ifdef __BLST_PORTABLE__+	testl	$1,__blst_platform_cap(%rip)+	jnz	sqr_382x$1+#endif 	pushq	%rbp  	pushq	%rbx@@ -628,7 +645,7 @@  	leaq	48(%rsi),%rdx 	leaq	48(%rdi),%rdi-	call	__sub_mod_384_a_is_loaded+	call	__subq_mod_384_a_is_loaded   	leaq	(%rdi),%rsi@@ -710,11 +727,15 @@ 	movq	%rsi,16(%rsp) 	movq	%rsp,%r11 .LSEH_begin_mul_384:++ 	movq	%rcx,%rdi 	movq	%rdx,%rsi 	movq	%r8,%rdx--+#ifdef __BLST_PORTABLE__+	testl	$1,__blst_platform_cap(%rip)+	jnz	mul_384$1+#endif 	pushq	%rbp  	pushq	%rbx@@ -1039,10 +1060,14 @@ 	movq	%rsi,16(%rsp) 	movq	%rsp,%r11 .LSEH_begin_sqr_384:-	movq	%rcx,%rdi-	movq	%rdx,%rsi  +	movq	%rcx,%rdi+	movq	%rdx,%rsi+#ifdef __BLST_PORTABLE__+	testl	$1,__blst_platform_cap(%rip)+	jnz	sqr_384$1+#endif 	pushq	%rbp  	pushq	%rbx@@ -1286,12 +1311,16 @@ 	movq	%rsi,16(%rsp) 	movq	%rsp,%r11 .LSEH_begin_sqr_mont_384:++ 	movq	%rcx,%rdi 	movq	%rdx,%rsi 	movq	%r8,%rdx 	movq	%r9,%rcx--+#ifdef __BLST_PORTABLE__+	testl	$1,__blst_platform_cap(%rip)+	jnz	sqr_mont_384$1+#endif 	pushq	%rbp  	pushq	%rbx@@ -1321,7 +1350,7 @@ 	movq	104(%rsp),%rbx 	movq	112(%rsp),%rdi 	call	__mulq_by_1_mont_384-	call	__redc_tail_mont_384+	call	__redq_tail_mont_384  	leaq	120(%rsp),%r8 	movq	120(%rsp),%r15@@ -1358,12 +1387,16 @@ 	movq	%rsi,16(%rsp) 	movq	%rsp,%r11 .LSEH_begin_redc_mont_384:++ 	movq	%rcx,%rdi 	movq	%rdx,%rsi 	movq	%r8,%rdx 	movq	%r9,%rcx--+#ifdef __BLST_PORTABLE__+	testl	$1,__blst_platform_cap(%rip)+	jnz	redc_mont_384$1+#endif 	pushq	%rbp  	pushq	%rbx@@ -1383,7 +1416,7 @@  	movq	%rdx,%rbx 	call	__mulq_by_1_mont_384-	call	__redc_tail_mont_384+	call	__redq_tail_mont_384  	movq	8(%rsp),%r15 @@ -1420,12 +1453,16 @@ 	movq	%rsi,16(%rsp) 	movq	%rsp,%r11 .LSEH_begin_from_mont_384:++ 	movq	%rcx,%rdi 	movq	%rdx,%rsi 	movq	%r8,%rdx 	movq	%r9,%rcx--+#ifdef __BLST_PORTABLE__+	testl	$1,__blst_platform_cap(%rip)+	jnz	from_mont_384$1+#endif 	pushq	%rbp  	pushq	%rbx@@ -1795,9 +1832,9 @@ 	.byte	0xf3,0xc3  -.def	__redc_tail_mont_384;	.scl 3;	.type 32;	.endef+.def	__redq_tail_mont_384;	.scl 3;	.type 32;	.endef .p2align	5-__redc_tail_mont_384:+__redq_tail_mont_384: 	.byte	0xf3,0x0f,0x1e,0xfa  	addq	48(%rsi),%r14@@ -1852,11 +1889,15 @@ 	movq	%rsi,16(%rsp) 	movq	%rsp,%r11 .LSEH_begin_sgn0_pty_mont_384:++ 	movq	%rcx,%rdi 	movq	%rdx,%rsi 	movq	%r8,%rdx--+#ifdef __BLST_PORTABLE__+	testl	$1,__blst_platform_cap(%rip)+	jnz	sgn0_pty_mont_384$1+#endif 	pushq	%rbp  	pushq	%rbx@@ -1934,11 +1975,15 @@ 	movq	%rsi,16(%rsp) 	movq	%rsp,%r11 .LSEH_begin_sgn0_pty_mont_384x:++ 	movq	%rcx,%rdi 	movq	%rdx,%rsi 	movq	%r8,%rdx--+#ifdef __BLST_PORTABLE__+	testl	$1,__blst_platform_cap(%rip)+	jnz	sgn0_pty_mont_384x$1+#endif 	pushq	%rbp  	pushq	%rbx@@ -2065,13 +2110,17 @@ 	movq	%rsi,16(%rsp) 	movq	%rsp,%r11 .LSEH_begin_mul_mont_384:++ 	movq	%rcx,%rdi 	movq	%rdx,%rsi 	movq	%r8,%rdx 	movq	%r9,%rcx 	movq	40(%rsp),%r8--+#ifdef __BLST_PORTABLE__+	testl	$1,__blst_platform_cap(%rip)+	jnz	mul_mont_384$1+#endif 	pushq	%rbp  	pushq	%rbx@@ -2733,14 +2782,18 @@ 	movq	%rsi,16(%rsp) 	movq	%rsp,%r11 .LSEH_begin_sqr_n_mul_mont_384:++ 	movq	%rcx,%rdi 	movq	%rdx,%rsi 	movq	%r8,%rdx 	movq	%r9,%rcx 	movq	40(%rsp),%r8 	movq	48(%rsp),%r9--+#ifdef __BLST_PORTABLE__+	testl	$1,__blst_platform_cap(%rip)+	jnz	sqr_n_mul_mont_384$1+#endif 	pushq	%rbp  	pushq	%rbx@@ -2774,7 +2827,7 @@ 	movq	0(%rsp),%rcx 	movq	16(%rsp),%rbx 	call	__mulq_by_1_mont_384-	call	__redc_tail_mont_384+	call	__redq_tail_mont_384  	movd	%xmm1,%edx 	leaq	0(%rdi),%rsi@@ -2828,14 +2881,18 @@ 	movq	%rsi,16(%rsp) 	movq	%rsp,%r11 .LSEH_begin_sqr_n_mul_mont_383:++ 	movq	%rcx,%rdi 	movq	%rdx,%rsi 	movq	%r8,%rdx 	movq	%r9,%rcx 	movq	40(%rsp),%r8 	movq	48(%rsp),%r9--+#ifdef __BLST_PORTABLE__+	testl	$1,__blst_platform_cap(%rip)+	jnz	sqr_n_mul_mont_383$1+#endif 	pushq	%rbp  	pushq	%rbx@@ -3494,12 +3551,16 @@ 	movq	%rsi,16(%rsp) 	movq	%rsp,%r11 .LSEH_begin_sqr_mont_382x:++ 	movq	%rcx,%rdi 	movq	%rdx,%rsi 	movq	%r8,%rdx 	movq	%r9,%rcx--+#ifdef __BLST_PORTABLE__+	testl	$1,__blst_platform_cap(%rip)+	jnz	sqr_mont_382x$1+#endif 	pushq	%rbp  	pushq	%rbx@@ -3858,8 +3919,9 @@ .byte	1,0,5,0x0b .byte	0,0x74,1,0 .byte	0,0x64,2,0-.byte	0,0x03+.byte	0,0xb3 .byte	0,0+.long	0,0 .LSEH_info_mul_mont_384x_body: .byte	1,0,18,0 .byte	0x00,0xf4,0x29,0x00@@ -3871,6 +3933,8 @@ .byte	0x00,0x74,0x30,0x00 .byte	0x00,0x64,0x31,0x00 .byte	0x00,0x01,0x2f,0x00+.byte	0x00,0x00,0x00,0x00+.byte	0x00,0x00,0x00,0x00 .LSEH_info_mul_mont_384x_epilogue: .byte	1,0,4,0 .byte	0x00,0x74,0x01,0x00@@ -3881,8 +3945,9 @@ .byte	1,0,5,0x0b .byte	0,0x74,1,0 .byte	0,0x64,2,0-.byte	0,0x03+.byte	0,0xb3 .byte	0,0+.long	0,0 .LSEH_info_sqr_mont_384x_body: .byte	1,0,18,0 .byte	0x00,0xf4,0x11,0x00@@ -3894,6 +3959,8 @@ .byte	0x00,0x74,0x18,0x00 .byte	0x00,0x64,0x19,0x00 .byte	0x00,0x01,0x17,0x00+.byte	0x00,0x00,0x00,0x00+.byte	0x00,0x00,0x00,0x00 .LSEH_info_sqr_mont_384x_epilogue: .byte	1,0,4,0 .byte	0x00,0x74,0x01,0x00@@ -3904,8 +3971,9 @@ .byte	1,0,5,0x0b .byte	0,0x74,1,0 .byte	0,0x64,2,0-.byte	0,0x03+.byte	0,0xb3 .byte	0,0+.long	0,0 .LSEH_info_mul_382x_body: .byte	1,0,18,0 .byte	0x00,0xf4,0x11,0x00@@ -3917,6 +3985,8 @@ .byte	0x00,0x74,0x18,0x00 .byte	0x00,0x64,0x19,0x00 .byte	0x00,0x01,0x17,0x00+.byte	0x00,0x00,0x00,0x00+.byte	0x00,0x00,0x00,0x00 .LSEH_info_mul_382x_epilogue: .byte	1,0,4,0 .byte	0x00,0x74,0x01,0x00@@ -3927,8 +3997,9 @@ .byte	1,0,5,0x0b .byte	0,0x74,1,0 .byte	0,0x64,2,0-.byte	0,0x03+.byte	0,0xb3 .byte	0,0+.long	0,0 .LSEH_info_sqr_382x_body: .byte	1,0,17,0 .byte	0x00,0xf4,0x01,0x00@@ -3940,7 +4011,8 @@ .byte	0x00,0x74,0x08,0x00 .byte	0x00,0x64,0x09,0x00 .byte	0x00,0x62-.byte	0x00,0x00+.byte	0x00,0x00,0x00,0x00,0x00,0x00+.byte	0x00,0x00,0x00,0x00 .LSEH_info_sqr_382x_epilogue: .byte	1,0,4,0 .byte	0x00,0x74,0x01,0x00@@ -3951,8 +4023,9 @@ .byte	1,0,5,0x0b .byte	0,0x74,1,0 .byte	0,0x64,2,0-.byte	0,0x03+.byte	0,0xb3 .byte	0,0+.long	0,0 .LSEH_info_mul_384_body: .byte	1,0,11,0 .byte	0x00,0xc4,0x00,0x00@@ -3972,8 +4045,9 @@ .byte	1,0,5,0x0b .byte	0,0x74,1,0 .byte	0,0x64,2,0-.byte	0,0x03+.byte	0,0xb3 .byte	0,0+.long	0,0 .LSEH_info_sqr_384_body: .byte	1,0,17,0 .byte	0x00,0xf4,0x01,0x00@@ -3985,7 +4059,8 @@ .byte	0x00,0x74,0x08,0x00 .byte	0x00,0x64,0x09,0x00 .byte	0x00,0x62-.byte	0x00,0x00+.byte	0x00,0x00,0x00,0x00,0x00,0x00+.byte	0x00,0x00,0x00,0x00 .LSEH_info_sqr_384_epilogue: .byte	1,0,4,0 .byte	0x00,0x74,0x01,0x00@@ -3996,8 +4071,9 @@ .byte	1,0,5,0x0b .byte	0,0x74,1,0 .byte	0,0x64,2,0-.byte	0,0x03+.byte	0,0xb3 .byte	0,0+.long	0,0 .LSEH_info_sqr_mont_384_body: .byte	1,0,18,0 .byte	0x00,0xf4,0x0f,0x00@@ -4009,6 +4085,8 @@ .byte	0x00,0x74,0x16,0x00 .byte	0x00,0x64,0x17,0x00 .byte	0x00,0x01,0x15,0x00+.byte	0x00,0x00,0x00,0x00+.byte	0x00,0x00,0x00,0x00 .LSEH_info_sqr_mont_384_epilogue: .byte	1,0,4,0 .byte	0x00,0x74,0x01,0x00@@ -4019,8 +4097,9 @@ .byte	1,0,5,0x0b .byte	0,0x74,1,0 .byte	0,0x64,2,0-.byte	0,0x03+.byte	0,0xb3 .byte	0,0+.long	0,0 .LSEH_info_redc_mont_384_body: .byte	1,0,17,0 .byte	0x00,0xf4,0x01,0x00@@ -4032,7 +4111,8 @@ .byte	0x00,0x74,0x08,0x00 .byte	0x00,0x64,0x09,0x00 .byte	0x00,0x62-.byte	0x00,0x00+.byte	0x00,0x00,0x00,0x00,0x00,0x00+.byte	0x00,0x00,0x00,0x00 .LSEH_info_redc_mont_384_epilogue: .byte	1,0,4,0 .byte	0x00,0x74,0x01,0x00@@ -4043,8 +4123,9 @@ .byte	1,0,5,0x0b .byte	0,0x74,1,0 .byte	0,0x64,2,0-.byte	0,0x03+.byte	0,0xb3 .byte	0,0+.long	0,0 .LSEH_info_from_mont_384_body: .byte	1,0,17,0 .byte	0x00,0xf4,0x01,0x00@@ -4056,7 +4137,8 @@ .byte	0x00,0x74,0x08,0x00 .byte	0x00,0x64,0x09,0x00 .byte	0x00,0x62-.byte	0x00,0x00+.byte	0x00,0x00,0x00,0x00,0x00,0x00+.byte	0x00,0x00,0x00,0x00 .LSEH_info_from_mont_384_epilogue: .byte	1,0,4,0 .byte	0x00,0x74,0x01,0x00@@ -4067,8 +4149,9 @@ .byte	1,0,5,0x0b .byte	0,0x74,1,0 .byte	0,0x64,2,0-.byte	0,0x03+.byte	0,0xb3 .byte	0,0+.long	0,0 .LSEH_info_sgn0_pty_mont_384_body: .byte	1,0,17,0 .byte	0x00,0xf4,0x01,0x00@@ -4080,7 +4163,8 @@ .byte	0x00,0x74,0x08,0x00 .byte	0x00,0x64,0x09,0x00 .byte	0x00,0x62-.byte	0x00,0x00+.byte	0x00,0x00,0x00,0x00,0x00,0x00+.byte	0x00,0x00,0x00,0x00 .LSEH_info_sgn0_pty_mont_384_epilogue: .byte	1,0,4,0 .byte	0x00,0x74,0x01,0x00@@ -4091,8 +4175,9 @@ .byte	1,0,5,0x0b .byte	0,0x74,1,0 .byte	0,0x64,2,0-.byte	0,0x03+.byte	0,0xb3 .byte	0,0+.long	0,0 .LSEH_info_sgn0_pty_mont_384x_body: .byte	1,0,17,0 .byte	0x00,0xf4,0x01,0x00@@ -4104,7 +4189,8 @@ .byte	0x00,0x74,0x08,0x00 .byte	0x00,0x64,0x09,0x00 .byte	0x00,0x62-.byte	0x00,0x00+.byte	0x00,0x00,0x00,0x00,0x00,0x00+.byte	0x00,0x00,0x00,0x00 .LSEH_info_sgn0_pty_mont_384x_epilogue: .byte	1,0,4,0 .byte	0x00,0x74,0x01,0x00@@ -4115,8 +4201,9 @@ .byte	1,0,5,0x0b .byte	0,0x74,1,0 .byte	0,0x64,2,0-.byte	0,0x03+.byte	0,0xb3 .byte	0,0+.long	0,0 .LSEH_info_mul_mont_384_body: .byte	1,0,17,0 .byte	0x00,0xf4,0x03,0x00@@ -4128,7 +4215,8 @@ .byte	0x00,0x74,0x0a,0x00 .byte	0x00,0x64,0x0b,0x00 .byte	0x00,0x82-.byte	0x00,0x00+.byte	0x00,0x00,0x00,0x00,0x00,0x00+.byte	0x00,0x00,0x00,0x00 .LSEH_info_mul_mont_384_epilogue: .byte	1,0,4,0 .byte	0x00,0x74,0x01,0x00@@ -4139,8 +4227,9 @@ .byte	1,0,5,0x0b .byte	0,0x74,1,0 .byte	0,0x64,2,0-.byte	0,0x03+.byte	0,0xb3 .byte	0,0+.long	0,0 .LSEH_info_sqr_n_mul_mont_384_body: .byte	1,0,18,0 .byte	0x00,0xf4,0x11,0x00@@ -4152,6 +4241,8 @@ .byte	0x00,0x74,0x18,0x00 .byte	0x00,0x64,0x19,0x00 .byte	0x00,0x01,0x17,0x00+.byte	0x00,0x00,0x00,0x00+.byte	0x00,0x00,0x00,0x00 .LSEH_info_sqr_n_mul_mont_384_epilogue: .byte	1,0,4,0 .byte	0x00,0x74,0x01,0x00@@ -4162,8 +4253,9 @@ .byte	1,0,5,0x0b .byte	0,0x74,1,0 .byte	0,0x64,2,0-.byte	0,0x03+.byte	0,0xb3 .byte	0,0+.long	0,0 .LSEH_info_sqr_n_mul_mont_383_body: .byte	1,0,18,0 .byte	0x00,0xf4,0x11,0x00@@ -4175,6 +4267,8 @@ .byte	0x00,0x74,0x18,0x00 .byte	0x00,0x64,0x19,0x00 .byte	0x00,0x01,0x17,0x00+.byte	0x00,0x00,0x00,0x00+.byte	0x00,0x00,0x00,0x00 .LSEH_info_sqr_n_mul_mont_383_epilogue: .byte	1,0,4,0 .byte	0x00,0x74,0x01,0x00@@ -4185,8 +4279,9 @@ .byte	1,0,5,0x0b .byte	0,0x74,1,0 .byte	0,0x64,2,0-.byte	0,0x03+.byte	0,0xb3 .byte	0,0+.long	0,0 .LSEH_info_sqr_mont_382x_body: .byte	1,0,18,0 .byte	0x00,0xf4,0x11,0x00@@ -4198,6 +4293,8 @@ .byte	0x00,0x74,0x18,0x00 .byte	0x00,0x64,0x19,0x00 .byte	0x00,0x01,0x17,0x00+.byte	0x00,0x00,0x00,0x00+.byte	0x00,0x00,0x00,0x00 .LSEH_info_sqr_mont_382x_epilogue: .byte	1,0,4,0 .byte	0x00,0x74,0x01,0x00

@@ -10,13 +10,14 @@ 	movq	%rsi,16(%rsp) 	movq	%rsp,%r11 .LSEH_begin_mulx_mont_sparse_256:++ 	movq	%rcx,%rdi 	movq	%rdx,%rsi 	movq	%r8,%rdx 	movq	%r9,%rcx 	movq	40(%rsp),%r8--+mul_mont_sparse_256$1: 	pushq	%rbp  	pushq	%rbx@@ -78,12 +79,13 @@ 	movq	%rsi,16(%rsp) 	movq	%rsp,%r11 .LSEH_begin_sqrx_mont_sparse_256:++ 	movq	%rcx,%rdi 	movq	%rdx,%rsi 	movq	%r8,%rdx 	movq	%r9,%rcx--+sqr_mont_sparse_256$1: 	pushq	%rbp  	pushq	%rbx@@ -342,12 +344,13 @@ 	movq	%rsi,16(%rsp) 	movq	%rsp,%r11 .LSEH_begin_fromx_mont_256:++ 	movq	%rcx,%rdi 	movq	%rdx,%rsi 	movq	%r8,%rdx 	movq	%r9,%rcx--+from_mont_256$1: 	pushq	%rbp  	pushq	%rbx@@ -422,12 +425,13 @@ 	movq	%rsi,16(%rsp) 	movq	%rsp,%r11 .LSEH_begin_redcx_mont_256:++ 	movq	%rcx,%rdi 	movq	%rdx,%rsi 	movq	%r8,%rdx 	movq	%r9,%rcx--+redc_mont_256$1: 	pushq	%rbp  	pushq	%rbx@@ -690,8 +694,9 @@ .byte	1,0,5,0x0b .byte	0,0x74,1,0 .byte	0,0x64,2,0-.byte	0,0x03+.byte	0,0xb3 .byte	0,0+.long	0,0 .LSEH_info_mulx_mont_sparse_256_body: .byte	1,0,17,0 .byte	0x00,0xf4,0x01,0x00@@ -703,7 +708,8 @@ .byte	0x00,0x74,0x08,0x00 .byte	0x00,0x64,0x09,0x00 .byte	0x00,0x62-.byte	0x00,0x00+.byte	0x00,0x00,0x00,0x00,0x00,0x00+.byte	0x00,0x00,0x00,0x00 .LSEH_info_mulx_mont_sparse_256_epilogue: .byte	1,0,4,0 .byte	0x00,0x74,0x01,0x00@@ -714,8 +720,9 @@ .byte	1,0,5,0x0b .byte	0,0x74,1,0 .byte	0,0x64,2,0-.byte	0,0x03+.byte	0,0xb3 .byte	0,0+.long	0,0 .LSEH_info_sqrx_mont_sparse_256_body: .byte	1,0,17,0 .byte	0x00,0xf4,0x01,0x00@@ -727,7 +734,8 @@ .byte	0x00,0x74,0x08,0x00 .byte	0x00,0x64,0x09,0x00 .byte	0x00,0x62-.byte	0x00,0x00+.byte	0x00,0x00,0x00,0x00,0x00,0x00+.byte	0x00,0x00,0x00,0x00 .LSEH_info_sqrx_mont_sparse_256_epilogue: .byte	1,0,4,0 .byte	0x00,0x74,0x01,0x00@@ -738,8 +746,9 @@ .byte	1,0,5,0x0b .byte	0,0x74,1,0 .byte	0,0x64,2,0-.byte	0,0x03+.byte	0,0xb3 .byte	0,0+.long	0,0 .LSEH_info_fromx_mont_256_body: .byte	1,0,17,0 .byte	0x00,0xf4,0x01,0x00@@ -751,7 +760,8 @@ .byte	0x00,0x74,0x08,0x00 .byte	0x00,0x64,0x09,0x00 .byte	0x00,0x62-.byte	0x00,0x00+.byte	0x00,0x00,0x00,0x00,0x00,0x00+.byte	0x00,0x00,0x00,0x00 .LSEH_info_fromx_mont_256_epilogue: .byte	1,0,4,0 .byte	0x00,0x74,0x01,0x00@@ -762,8 +772,9 @@ .byte	1,0,5,0x0b .byte	0,0x74,1,0 .byte	0,0x64,2,0-.byte	0,0x03+.byte	0,0xb3 .byte	0,0+.long	0,0 .LSEH_info_redcx_mont_256_body: .byte	1,0,17,0 .byte	0x00,0xf4,0x01,0x00@@ -775,7 +786,8 @@ .byte	0x00,0x74,0x08,0x00 .byte	0x00,0x64,0x09,0x00 .byte	0x00,0x62-.byte	0x00,0x00+.byte	0x00,0x00,0x00,0x00,0x00,0x00+.byte	0x00,0x00,0x00,0x00 .LSEH_info_redcx_mont_256_epilogue: .byte	1,0,4,0 .byte	0x00,0x74,0x01,0x00

@@ -6,9 +6,9 @@   -.def	__sub_mod_384x384;	.scl 3;	.type 32;	.endef+.def	__subx_mod_384x384;	.scl 3;	.type 32;	.endef .p2align	5-__sub_mod_384x384:+__subx_mod_384x384: 	.byte	0xf3,0x0f,0x1e,0xfa  	movq	0(%rsi),%r8@@ -73,9 +73,9 @@ 	.byte	0xf3,0xc3  -.def	__add_mod_384;	.scl 3;	.type 32;	.endef+.def	__addx_mod_384;	.scl 3;	.type 32;	.endef .p2align	5-__add_mod_384:+__addx_mod_384: 	.byte	0xf3,0x0f,0x1e,0xfa  	movq	0(%rsi),%r8@@ -123,9 +123,9 @@ 	.byte	0xf3,0xc3  -.def	__sub_mod_384;	.scl 3;	.type 32;	.endef+.def	__subx_mod_384;	.scl 3;	.type 32;	.endef .p2align	5-__sub_mod_384:+__subx_mod_384: 	.byte	0xf3,0x0f,0x1e,0xfa  	movq	0(%rsi),%r8@@ -135,7 +135,7 @@ 	movq	32(%rsi),%r12 	movq	40(%rsi),%r13 -__sub_mod_384_a_is_loaded:+__subx_mod_384_a_is_loaded: 	subq	0(%rdx),%r8 	movq	0(%rcx),%r14 	sbbq	8(%rdx),%r9@@ -182,13 +182,14 @@ 	movq	%rsi,16(%rsp) 	movq	%rsp,%r11 .LSEH_begin_mulx_mont_384x:++ 	movq	%rcx,%rdi 	movq	%rdx,%rsi 	movq	%r8,%rdx 	movq	%r9,%rcx 	movq	40(%rsp),%r8--+mul_mont_384x$1: 	pushq	%rbp  	pushq	%rbx@@ -230,12 +231,12 @@ 	leaq	(%rbx),%rsi 	leaq	-48(%rbx),%rdx 	leaq	40+192+48(%rsp),%rdi-	call	__add_mod_384+	call	__addx_mod_384  	movq	24(%rsp),%rsi 	leaq	48(%rsi),%rdx 	leaq	-48(%rdi),%rdi-	call	__add_mod_384+	call	__addx_mod_384  	leaq	(%rdi),%rbx 	leaq	48(%rdi),%rsi@@ -245,17 +246,17 @@ 	leaq	(%rdi),%rsi 	leaq	40(%rsp),%rdx 	movq	8(%rsp),%rcx-	call	__sub_mod_384x384+	call	__subx_mod_384x384  	leaq	(%rdi),%rsi 	leaq	-96(%rdi),%rdx-	call	__sub_mod_384x384+	call	__subx_mod_384x384   	leaq	40(%rsp),%rsi 	leaq	40+96(%rsp),%rdx 	leaq	40(%rsp),%rdi-	call	__sub_mod_384x384+	call	__subx_mod_384x384  	leaq	(%rcx),%rbx @@ -264,14 +265,14 @@ 	movq	0(%rsp),%rcx 	movq	32(%rsp),%rdi 	call	__mulx_by_1_mont_384-	call	__redc_tail_mont_384+	call	__redx_tail_mont_384   	leaq	40+192(%rsp),%rsi 	movq	0(%rsp),%rcx 	leaq	48(%rdi),%rdi 	call	__mulx_by_1_mont_384-	call	__redc_tail_mont_384+	call	__redx_tail_mont_384  	leaq	328(%rsp),%r8 	movq	0(%r8),%r15@@ -305,12 +306,13 @@ 	movq	%rsi,16(%rsp) 	movq	%rsp,%r11 .LSEH_begin_sqrx_mont_384x:++ 	movq	%rcx,%rdi 	movq	%rdx,%rsi 	movq	%r8,%rdx 	movq	%r9,%rcx--+sqr_mont_384x$1: 	pushq	%rbp  	pushq	%rbx@@ -337,13 +339,13 @@  	leaq	48(%rsi),%rdx 	leaq	32(%rsp),%rdi-	call	__add_mod_384+	call	__addx_mod_384   	movq	24(%rsp),%rsi 	leaq	48(%rsi),%rdx 	leaq	32+48(%rsp),%rdi-	call	__sub_mod_384+	call	__subx_mod_384   	movq	24(%rsp),%rsi@@ -445,12 +447,13 @@ 	movq	%rsi,16(%rsp) 	movq	%rsp,%r11 .LSEH_begin_mulx_382x:++ 	movq	%rcx,%rdi 	movq	%rdx,%rsi 	movq	%r8,%rdx 	movq	%r9,%rcx--+mul_382x$1: 	pushq	%rbp  	pushq	%rbx@@ -540,18 +543,18 @@ 	leaq	32(%rsp),%rdx 	movq	24(%rsp),%rcx 	movq	%rsi,%rdi-	call	__sub_mod_384x384+	call	__subx_mod_384x384   	leaq	0(%rdi),%rsi 	leaq	-96(%rdi),%rdx-	call	__sub_mod_384x384+	call	__subx_mod_384x384   	leaq	-96(%rdi),%rsi 	leaq	32(%rsp),%rdx 	leaq	-96(%rdi),%rdi-	call	__sub_mod_384x384+	call	__subx_mod_384x384  	leaq	136(%rsp),%r8 	movq	0(%r8),%r15@@ -585,11 +588,12 @@ 	movq	%rsi,16(%rsp) 	movq	%rsp,%r11 .LSEH_begin_sqrx_382x:++ 	movq	%rcx,%rdi 	movq	%rdx,%rsi 	movq	%r8,%rdx--+sqr_382x$1: 	pushq	%rbp  	pushq	%rbx@@ -640,7 +644,7 @@  	leaq	48(%rsi),%rdx 	leaq	48(%rdi),%rdi-	call	__sub_mod_384_a_is_loaded+	call	__subx_mod_384_a_is_loaded   	leaq	(%rdi),%rsi@@ -722,11 +726,12 @@ 	movq	%rsi,16(%rsp) 	movq	%rsp,%r11 .LSEH_begin_mulx_384:++ 	movq	%rcx,%rdi 	movq	%rdx,%rsi 	movq	%r8,%rdx--+mul_384$1: 	pushq	%rbp  	pushq	%rbx@@ -950,10 +955,11 @@ 	movq	%rsi,16(%rsp) 	movq	%rsp,%r11 .LSEH_begin_sqrx_384:-	movq	%rcx,%rdi-	movq	%rdx,%rsi  +	movq	%rcx,%rdi+	movq	%rdx,%rsi+sqr_384$1: 	pushq	%rbp  	pushq	%rbx@@ -1145,12 +1151,13 @@ 	movq	%rsi,16(%rsp) 	movq	%rsp,%r11 .LSEH_begin_redcx_mont_384:++ 	movq	%rcx,%rdi 	movq	%rdx,%rsi 	movq	%r8,%rdx 	movq	%r9,%rcx--+redc_mont_384$1: 	pushq	%rbp  	pushq	%rbx@@ -1170,7 +1177,7 @@  	movq	%rdx,%rbx 	call	__mulx_by_1_mont_384-	call	__redc_tail_mont_384+	call	__redx_tail_mont_384  	movq	8(%rsp),%r15 @@ -1207,12 +1214,13 @@ 	movq	%rsi,16(%rsp) 	movq	%rsp,%r11 .LSEH_begin_fromx_mont_384:++ 	movq	%rcx,%rdi 	movq	%rdx,%rsi 	movq	%r8,%rdx 	movq	%r9,%rcx--+from_mont_384$1: 	pushq	%rbp  	pushq	%rbx@@ -1473,9 +1481,9 @@ 	.byte	0xf3,0xc3  -.def	__redc_tail_mont_384;	.scl 3;	.type 32;	.endef+.def	__redx_tail_mont_384;	.scl 3;	.type 32;	.endef .p2align	5-__redc_tail_mont_384:+__redx_tail_mont_384: 	.byte	0xf3,0x0f,0x1e,0xfa  	addq	48(%rsi),%r14@@ -1530,11 +1538,12 @@ 	movq	%rsi,16(%rsp) 	movq	%rsp,%r11 .LSEH_begin_sgn0x_pty_mont_384:++ 	movq	%rcx,%rdi 	movq	%rdx,%rsi 	movq	%r8,%rdx--+sgn0_pty_mont_384$1: 	pushq	%rbp  	pushq	%rbx@@ -1612,11 +1621,12 @@ 	movq	%rsi,16(%rsp) 	movq	%rsp,%r11 .LSEH_begin_sgn0x_pty_mont_384x:++ 	movq	%rcx,%rdi 	movq	%rdx,%rsi 	movq	%r8,%rdx--+sgn0_pty_mont_384x$1: 	pushq	%rbp  	pushq	%rbx@@ -1743,13 +1753,14 @@ 	movq	%rsi,16(%rsp) 	movq	%rsp,%r11 .LSEH_begin_mulx_mont_384:++ 	movq	%rcx,%rdi 	movq	%rdx,%rsi 	movq	%r8,%rdx 	movq	%r9,%rcx 	movq	40(%rsp),%r8--+mul_mont_384$1: 	pushq	%rbp  	pushq	%rbx@@ -2215,12 +2226,13 @@ 	movq	%rsi,16(%rsp) 	movq	%rsp,%r11 .LSEH_begin_sqrx_mont_384:++ 	movq	%rcx,%rdi 	movq	%rdx,%rsi 	movq	%r8,%rdx 	movq	%r9,%rcx--+sqr_mont_384$1: 	pushq	%rbp  	pushq	%rbx@@ -2287,14 +2299,15 @@ 	movq	%rsi,16(%rsp) 	movq	%rsp,%r11 .LSEH_begin_sqrx_n_mul_mont_384:++ 	movq	%rcx,%rdi 	movq	%rdx,%rsi 	movq	%r8,%rdx 	movq	%r9,%rcx 	movq	40(%rsp),%r8 	movq	48(%rsp),%r9--+sqr_n_mul_mont_384$1: 	pushq	%rbp  	pushq	%rbx@@ -2379,14 +2392,15 @@ 	movq	%rsi,16(%rsp) 	movq	%rsp,%r11 .LSEH_begin_sqrx_n_mul_mont_383:++ 	movq	%rcx,%rdi 	movq	%rdx,%rsi 	movq	%r8,%rdx 	movq	%r9,%rcx 	movq	40(%rsp),%r8 	movq	48(%rsp),%r9--+sqr_n_mul_mont_383$1: 	pushq	%rbp  	pushq	%rbx@@ -2831,12 +2845,13 @@ 	movq	%rsi,16(%rsp) 	movq	%rsp,%r11 .LSEH_begin_sqrx_mont_382x:++ 	movq	%rcx,%rdi 	movq	%rdx,%rsi 	movq	%r8,%rdx 	movq	%r9,%rcx--+sqr_mont_382x$1: 	pushq	%rbp  	pushq	%rbx@@ -3205,8 +3220,9 @@ .byte	1,0,5,0x0b .byte	0,0x74,1,0 .byte	0,0x64,2,0-.byte	0,0x03+.byte	0,0xb3 .byte	0,0+.long	0,0 .LSEH_info_mulx_mont_384x_body: .byte	1,0,18,0 .byte	0x00,0xf4,0x29,0x00@@ -3218,6 +3234,8 @@ .byte	0x00,0x74,0x30,0x00 .byte	0x00,0x64,0x31,0x00 .byte	0x00,0x01,0x2f,0x00+.byte	0x00,0x00,0x00,0x00+.byte	0x00,0x00,0x00,0x00 .LSEH_info_mulx_mont_384x_epilogue: .byte	1,0,4,0 .byte	0x00,0x74,0x01,0x00@@ -3228,8 +3246,9 @@ .byte	1,0,5,0x0b .byte	0,0x74,1,0 .byte	0,0x64,2,0-.byte	0,0x03+.byte	0,0xb3 .byte	0,0+.long	0,0 .LSEH_info_sqrx_mont_384x_body: .byte	1,0,18,0 .byte	0x00,0xf4,0x11,0x00@@ -3241,6 +3260,8 @@ .byte	0x00,0x74,0x18,0x00 .byte	0x00,0x64,0x19,0x00 .byte	0x00,0x01,0x17,0x00+.byte	0x00,0x00,0x00,0x00+.byte	0x00,0x00,0x00,0x00 .LSEH_info_sqrx_mont_384x_epilogue: .byte	1,0,4,0 .byte	0x00,0x74,0x01,0x00@@ -3251,8 +3272,9 @@ .byte	1,0,5,0x0b .byte	0,0x74,1,0 .byte	0,0x64,2,0-.byte	0,0x03+.byte	0,0xb3 .byte	0,0+.long	0,0 .LSEH_info_mulx_382x_body: .byte	1,0,18,0 .byte	0x00,0xf4,0x11,0x00@@ -3264,6 +3286,8 @@ .byte	0x00,0x74,0x18,0x00 .byte	0x00,0x64,0x19,0x00 .byte	0x00,0x01,0x17,0x00+.byte	0x00,0x00,0x00,0x00+.byte	0x00,0x00,0x00,0x00 .LSEH_info_mulx_382x_epilogue: .byte	1,0,4,0 .byte	0x00,0x74,0x01,0x00@@ -3274,8 +3298,9 @@ .byte	1,0,5,0x0b .byte	0,0x74,1,0 .byte	0,0x64,2,0-.byte	0,0x03+.byte	0,0xb3 .byte	0,0+.long	0,0 .LSEH_info_sqrx_382x_body: .byte	1,0,17,0 .byte	0x00,0xf4,0x01,0x00@@ -3287,7 +3312,8 @@ .byte	0x00,0x74,0x08,0x00 .byte	0x00,0x64,0x09,0x00 .byte	0x00,0x62-.byte	0x00,0x00+.byte	0x00,0x00,0x00,0x00,0x00,0x00+.byte	0x00,0x00,0x00,0x00 .LSEH_info_sqrx_382x_epilogue: .byte	1,0,4,0 .byte	0x00,0x74,0x01,0x00@@ -3298,8 +3324,9 @@ .byte	1,0,5,0x0b .byte	0,0x74,1,0 .byte	0,0x64,2,0-.byte	0,0x03+.byte	0,0xb3 .byte	0,0+.long	0,0 .LSEH_info_mulx_384_body: .byte	1,0,17,0 .byte	0x00,0xf4,0x00,0x00@@ -3311,7 +3338,8 @@ .byte	0x00,0x74,0x07,0x00 .byte	0x00,0x64,0x08,0x00 .byte	0x00,0x52-.byte	0x00,0x00+.byte	0x00,0x00,0x00,0x00,0x00,0x00+.byte	0x00,0x00,0x00,0x00 .LSEH_info_mulx_384_epilogue: .byte	1,0,4,0 .byte	0x00,0x74,0x01,0x00@@ -3322,8 +3350,9 @@ .byte	1,0,5,0x0b .byte	0,0x74,1,0 .byte	0,0x64,2,0-.byte	0,0x03+.byte	0,0xb3 .byte	0,0+.long	0,0 .LSEH_info_sqrx_384_body: .byte	1,0,17,0 .byte	0x00,0xf4,0x01,0x00@@ -3335,7 +3364,8 @@ .byte	0x00,0x74,0x08,0x00 .byte	0x00,0x64,0x09,0x00 .byte	0x00,0x62-.byte	0x00,0x00+.byte	0x00,0x00,0x00,0x00,0x00,0x00+.byte	0x00,0x00,0x00,0x00 .LSEH_info_sqrx_384_epilogue: .byte	1,0,4,0 .byte	0x00,0x74,0x01,0x00@@ -3346,8 +3376,9 @@ .byte	1,0,5,0x0b .byte	0,0x74,1,0 .byte	0,0x64,2,0-.byte	0,0x03+.byte	0,0xb3 .byte	0,0+.long	0,0 .LSEH_info_redcx_mont_384_body: .byte	1,0,17,0 .byte	0x00,0xf4,0x01,0x00@@ -3359,7 +3390,8 @@ .byte	0x00,0x74,0x08,0x00 .byte	0x00,0x64,0x09,0x00 .byte	0x00,0x62-.byte	0x00,0x00+.byte	0x00,0x00,0x00,0x00,0x00,0x00+.byte	0x00,0x00,0x00,0x00 .LSEH_info_redcx_mont_384_epilogue: .byte	1,0,4,0 .byte	0x00,0x74,0x01,0x00@@ -3370,8 +3402,9 @@ .byte	1,0,5,0x0b .byte	0,0x74,1,0 .byte	0,0x64,2,0-.byte	0,0x03+.byte	0,0xb3 .byte	0,0+.long	0,0 .LSEH_info_fromx_mont_384_body: .byte	1,0,17,0 .byte	0x00,0xf4,0x01,0x00@@ -3383,7 +3416,8 @@ .byte	0x00,0x74,0x08,0x00 .byte	0x00,0x64,0x09,0x00 .byte	0x00,0x62-.byte	0x00,0x00+.byte	0x00,0x00,0x00,0x00,0x00,0x00+.byte	0x00,0x00,0x00,0x00 .LSEH_info_fromx_mont_384_epilogue: .byte	1,0,4,0 .byte	0x00,0x74,0x01,0x00@@ -3394,8 +3428,9 @@ .byte	1,0,5,0x0b .byte	0,0x74,1,0 .byte	0,0x64,2,0-.byte	0,0x03+.byte	0,0xb3 .byte	0,0+.long	0,0 .LSEH_info_sgn0x_pty_mont_384_body: .byte	1,0,17,0 .byte	0x00,0xf4,0x01,0x00@@ -3407,7 +3442,8 @@ .byte	0x00,0x74,0x08,0x00 .byte	0x00,0x64,0x09,0x00 .byte	0x00,0x62-.byte	0x00,0x00+.byte	0x00,0x00,0x00,0x00,0x00,0x00+.byte	0x00,0x00,0x00,0x00 .LSEH_info_sgn0x_pty_mont_384_epilogue: .byte	1,0,4,0 .byte	0x00,0x74,0x01,0x00@@ -3418,8 +3454,9 @@ .byte	1,0,5,0x0b .byte	0,0x74,1,0 .byte	0,0x64,2,0-.byte	0,0x03+.byte	0,0xb3 .byte	0,0+.long	0,0 .LSEH_info_sgn0x_pty_mont_384x_body: .byte	1,0,17,0 .byte	0x00,0xf4,0x01,0x00@@ -3431,7 +3468,8 @@ .byte	0x00,0x74,0x08,0x00 .byte	0x00,0x64,0x09,0x00 .byte	0x00,0x62-.byte	0x00,0x00+.byte	0x00,0x00,0x00,0x00,0x00,0x00+.byte	0x00,0x00,0x00,0x00 .LSEH_info_sgn0x_pty_mont_384x_epilogue: .byte	1,0,4,0 .byte	0x00,0x74,0x01,0x00@@ -3442,8 +3480,9 @@ .byte	1,0,5,0x0b .byte	0,0x74,1,0 .byte	0,0x64,2,0-.byte	0,0x03+.byte	0,0xb3 .byte	0,0+.long	0,0 .LSEH_info_mulx_mont_384_body: .byte	1,0,17,0 .byte	0x00,0xf4,0x03,0x00@@ -3455,7 +3494,8 @@ .byte	0x00,0x74,0x0a,0x00 .byte	0x00,0x64,0x0b,0x00 .byte	0x00,0x82-.byte	0x00,0x00+.byte	0x00,0x00,0x00,0x00,0x00,0x00+.byte	0x00,0x00,0x00,0x00 .LSEH_info_mulx_mont_384_epilogue: .byte	1,0,4,0 .byte	0x00,0x74,0x01,0x00@@ -3466,8 +3506,9 @@ .byte	1,0,5,0x0b .byte	0,0x74,1,0 .byte	0,0x64,2,0-.byte	0,0x03+.byte	0,0xb3 .byte	0,0+.long	0,0 .LSEH_info_sqrx_mont_384_body: .byte	1,0,17,0 .byte	0x00,0xf4,0x03,0x00@@ -3479,7 +3520,8 @@ .byte	0x00,0x74,0x0a,0x00 .byte	0x00,0x64,0x0b,0x00 .byte	0x00,0x82-.byte	0x00,0x00+.byte	0x00,0x00,0x00,0x00,0x00,0x00+.byte	0x00,0x00,0x00,0x00 .LSEH_info_sqrx_mont_384_epilogue: .byte	1,0,4,0 .byte	0x00,0x74,0x01,0x00@@ -3490,8 +3532,9 @@ .byte	1,0,5,0x0b .byte	0,0x74,1,0 .byte	0,0x64,2,0-.byte	0,0x03+.byte	0,0xb3 .byte	0,0+.long	0,0 .LSEH_info_sqrx_n_mul_mont_384_body: .byte	1,0,17,0 .byte	0x00,0xf4,0x05,0x00@@ -3503,7 +3546,8 @@ .byte	0x00,0x74,0x0c,0x00 .byte	0x00,0x64,0x0d,0x00 .byte	0x00,0xa2-.byte	0x00,0x00+.byte	0x00,0x00,0x00,0x00,0x00,0x00+.byte	0x00,0x00,0x00,0x00 .LSEH_info_sqrx_n_mul_mont_384_epilogue: .byte	1,0,4,0 .byte	0x00,0x74,0x01,0x00@@ -3514,8 +3558,9 @@ .byte	1,0,5,0x0b .byte	0,0x74,1,0 .byte	0,0x64,2,0-.byte	0,0x03+.byte	0,0xb3 .byte	0,0+.long	0,0 .LSEH_info_sqrx_n_mul_mont_383_body: .byte	1,0,17,0 .byte	0x00,0xf4,0x05,0x00@@ -3527,7 +3572,8 @@ .byte	0x00,0x74,0x0c,0x00 .byte	0x00,0x64,0x0d,0x00 .byte	0x00,0xa2-.byte	0x00,0x00+.byte	0x00,0x00,0x00,0x00,0x00,0x00+.byte	0x00,0x00,0x00,0x00 .LSEH_info_sqrx_n_mul_mont_383_epilogue: .byte	1,0,4,0 .byte	0x00,0x74,0x01,0x00@@ -3538,8 +3584,9 @@ .byte	1,0,5,0x0b .byte	0,0x74,1,0 .byte	0,0x64,2,0-.byte	0,0x03+.byte	0,0xb3 .byte	0,0+.long	0,0 .LSEH_info_sqrx_mont_382x_body: .byte	1,0,18,0 .byte	0x00,0xf4,0x11,0x00@@ -3551,6 +3598,8 @@ .byte	0x00,0x74,0x18,0x00 .byte	0x00,0x64,0x19,0x00 .byte	0x00,0x01,0x17,0x00+.byte	0x00,0x00,0x00,0x00+.byte	0x00,0x00,0x00,0x00 .LSEH_info_sqrx_mont_382x_epilogue: .byte	1,0,4,0 .byte	0x00,0x74,0x01,0x00

@@ -10,11 +10,12 @@ // // sha256_block procedure for ARMv8. //-// This module is stripped of scalar code paths, with raionale that all+// This module is stripped of scalar code paths, with rationale that all // known processors are NEON-capable. // // See original module at CRYPTOGAMS for further details. +.comm	__blst_platform_cap,4 .text  .p2align	6@@ -188,6 +189,11 @@ .endef .p2align	4 blst_sha256_block_data_order:+	adrp	x16,__blst_platform_cap+	ldr	w16,[x16,#:lo12:__blst_platform_cap]+	tst	w16,#1+	b.ne	.Lv8_entry+ 	stp	x29, x30, [sp, #-16]! 	mov	x29, sp 	sub	sp,sp,#16*4

@@ -1,3 +1,4 @@+.comm	__blst_platform_cap,4 .text	  .globl	blst_sha256_block_data_order@@ -9,15 +10,21 @@ 	movq	%rsi,16(%rsp) 	movq	%rsp,%r11 .LSEH_begin_blst_sha256_block_data_order:+++	pushq	%rbp++	movq	%rsp,%rbp+ 	movq	%rcx,%rdi 	movq	%rdx,%rsi 	movq	%r8,%rdx--+#ifdef __BLST_PORTABLE__+	testl	$2,__blst_platform_cap(%rip)+	jnz	.Lblst_sha256_block_data_order$2+#endif 	pushq	%rbx -	pushq	%rbp- 	pushq	%r12  	pushq	%r13@@ -29,13 +36,14 @@ 	shlq	$4,%rdx 	subq	$64+24,%rsp ++.LSEH_body_blst_sha256_block_data_order:+ 	leaq	(%rsi,%rdx,4),%rdx 	movq	%rdi,64+0(%rsp) 	movq	%rsi,64+8(%rsp) 	movq	%rdx,64+16(%rsp)-.LSEH_body_blst_sha256_block_data_order: - 	movl	0(%rdi),%eax 	movl	4(%rdi),%ebx 	movl	8(%rdi),%ecx@@ -1637,17 +1645,11 @@ 	leaq	64+24+48(%rsp),%r11  	movq	64+24(%rsp),%r15- 	movq	-40(%r11),%r14- 	movq	-32(%r11),%r13- 	movq	-24(%r11),%r12--	movq	-16(%r11),%rbp--	movq	-8(%r11),%rbx-+	movq	-16(%r11),%rbx+	movq	-8(%r11),%rbp .LSEH_epilogue_blst_sha256_block_data_order: 	mov	8(%r11),%rdi 	mov	16(%r11),%rsi@@ -1657,6 +1659,7 @@  .LSEH_end_blst_sha256_block_data_order: +#ifndef __BLST_PORTABLE__ .p2align	6  K256:@@ -1742,6 +1745,7 @@ 	movq	%r11,24(%rcx) 	.byte	0xf3,0xc3 +#endif .section	.pdata .p2align	2 .rva	.LSEH_begin_blst_sha256_block_data_order@@ -1759,26 +1763,30 @@ .section	.xdata .p2align	3 .LSEH_info_blst_sha256_block_data_order_prologue:-.byte	1,0,5,0x0b-.byte	0,0x74,1,0-.byte	0,0x64,2,0-.byte	0,0x03-.byte	0,0+.byte	1,4,6,0x05+.byte	4,0x74,2,0+.byte	4,0x64,3,0+.byte	4,0x53+.byte	1,0x50+.long	0,0 .LSEH_info_blst_sha256_block_data_order_body: .byte	1,0,18,0 .byte	0x00,0xf4,0x0b,0x00 .byte	0x00,0xe4,0x0c,0x00 .byte	0x00,0xd4,0x0d,0x00 .byte	0x00,0xc4,0x0e,0x00-.byte	0x00,0x54,0x0f,0x00-.byte	0x00,0x34,0x10,0x00+.byte	0x00,0x34,0x0f,0x00+.byte	0x00,0x54,0x10,0x00 .byte	0x00,0x74,0x12,0x00 .byte	0x00,0x64,0x13,0x00 .byte	0x00,0x01,0x11,0x00+.byte	0x00,0x00,0x00,0x00+.byte	0x00,0x00,0x00,0x00 .LSEH_info_blst_sha256_block_data_order_epilogue: .byte	1,0,5,11 .byte	0x00,0x74,0x01,0x00 .byte	0x00,0x64,0x02,0x00-.byte	0x00,0x03-.byte	0x00,0x00+.byte	0x00,0xb3+.byte	0x00,0x00,0x00,0x00,0x00,0x00+.byte	0x00,0x00,0x00,0x00 

@@ -1,3 +1,4 @@+.comm	__blst_platform_cap,4 .text	  .p2align	6@@ -34,22 +35,23 @@ 	movq	%rsi,16(%rsp) 	movq	%rsp,%r11 .LSEH_begin_blst_sha256_block_data_order_shaext:-	movq	%rcx,%rdi-	movq	%rdx,%rsi-	movq	%r8,%rdx  -	subq	$0x58,%rsp--	movaps	%xmm6,-88(%r11)--	movaps	%xmm7,-72(%r11)+	pushq	%rbp -	movaps	%xmm8,-56(%r11)+	movq	%rsp,%rbp -	movaps	%xmm9,-40(%r11)+	movq	%rcx,%rdi+	movq	%rdx,%rsi+	movq	%r8,%rdx+.Lblst_sha256_block_data_order$2:+	subq	$0x50,%rsp -	movaps	%xmm10,-24(%r11)+	movaps	%xmm6,-80(%rbp)+	movaps	%xmm7,-64(%rbp)+	movaps	%xmm8,-48(%rbp)+	movaps	%xmm9,-32(%rbp)+	movaps	%xmm10,-16(%rbp)  .LSEH_body_blst_sha256_block_data_order_shaext: @@ -254,16 +256,18 @@  	movdqu	%xmm1,(%rdi) 	movdqu	%xmm2,16(%rdi)-	movaps	-88(%r11),%xmm6-	movaps	-72(%r11),%xmm7-	movaps	-56(%r11),%xmm8-	movaps	-40(%r11),%xmm9-	movaps	-24(%r11),%xmm10-	movq	%r11,%rsp+	movaps	-80(%rbp),%xmm6+	movaps	-64(%rbp),%xmm7+	movaps	-48(%rbp),%xmm8+	movaps	-32(%rbp),%xmm9+	movaps	-16(%rbp),%xmm10+	movq	%rbp,%rsp +	popq	%rbp+ .LSEH_epilogue_blst_sha256_block_data_order_shaext:-	mov	8(%r11),%rdi-	mov	16(%r11),%rsi+	mov	8(%rsp),%rdi+	mov	16(%rsp),%rsi  	.byte	0xf3,0xc3 @@ -278,13 +282,17 @@ 	movq	%rsi,16(%rsp) 	movq	%rsp,%r11 .LSEH_begin_blst_sha256_block_data_order:-	movq	%rcx,%rdi-	movq	%rdx,%rsi-	movq	%r8,%rdx   	pushq	%rbp +	movq	%rsp,%rbp++	movq	%rcx,%rdi+	movq	%rdx,%rsi+	movq	%r8,%rdx+	testl	$2,__blst_platform_cap(%rip)+	jnz	.Lblst_sha256_block_data_order$2 	pushq	%rbx  	pushq	%r12@@ -296,21 +304,16 @@ 	pushq	%r15  	shlq	$4,%rdx-	subq	$104,%rsp+	subq	$88,%rsp  	leaq	(%rsi,%rdx,4),%rdx-	movq	%rdi,0(%rsp)--	movq	%rdx,16(%rsp)-	movaps	%xmm6,32(%rsp)--	movaps	%xmm7,48(%rsp)--	movaps	%xmm8,64(%rsp)--	movaps	%xmm9,80(%rsp)+	movq	%rdi,-64(%rbp) -	movq	%rsp,%rbp+	movq	%rdx,-48(%rbp)+	movaps	%xmm6,-128(%rbp)+	movaps	%xmm7,-112(%rbp)+	movaps	%xmm8,-96(%rbp)+	movaps	%xmm9,-80(%rbp)  .LSEH_body_blst_sha256_block_data_order: @@ -331,7 +334,7 @@ .p2align	4 .Lloop_ssse3: 	movdqa	K256+256(%rip),%xmm7-	movq	%rsi,8(%rbp)+	movq	%rsi,-56(%rbp) 	movdqu	0(%rsi),%xmm0 	movdqu	16(%rsi),%xmm1 	movdqu	32(%rsi),%xmm2@@ -1356,9 +1359,9 @@ 	addl	%r15d,%eax 	movl	%r8d,%r13d 	addl	%eax,%r14d-	movq	0(%rbp),%rdi+	movq	-64(%rbp),%rdi 	movl	%r14d,%eax-	movq	8(%rbp),%rsi+	movq	-56(%rbp),%rsi  	addl	0(%rdi),%eax 	addl	4(%rdi),%ebx@@ -1370,7 +1373,7 @@ 	addl	28(%rdi),%r11d  	leaq	64(%rsi),%rsi-	cmpq	16(%rbp),%rsi+	cmpq	-48(%rbp),%rsi  	movl	%eax,0(%rdi) 	movl	%ebx,4(%rdi)@@ -1383,33 +1386,27 @@ 	jb	.Lloop_ssse3  	xorps	%xmm0,%xmm0-	leaq	104+48(%rbp),%r11- 	movaps	%xmm0,0(%rsp) 	movaps	%xmm0,16(%rsp) 	movaps	%xmm0,32(%rsp) 	movaps	%xmm0,48(%rsp)-	movaps	32(%rbp),%xmm6-	movaps	48(%rbp),%xmm7-	movaps	64(%rbp),%xmm8-	movaps	80(%rbp),%xmm9-	movq	104(%rbp),%r15--	movq	-40(%r11),%r14--	movq	-32(%r11),%r13--	movq	-24(%r11),%r12--	movq	-16(%r11),%rbx+	movaps	-128(%rbp),%xmm6+	movaps	-112(%rbp),%xmm7+	movaps	-96(%rbp),%xmm8+	movaps	-80(%rbp),%xmm9+	movq	-40(%rbp),%r15+	movq	-32(%rbp),%r14+	movq	-24(%rbp),%r13+	movq	-16(%rbp),%r12+	movq	-8(%rbp),%rbx+	movq	%rbp,%rsp -	movq	-8(%r11),%rbp+	popq	%rbp  .LSEH_epilogue_blst_sha256_block_data_order:-	mov	8(%r11),%rdi-	mov	16(%r11),%rsi+	mov	8(%rsp),%rdi+	mov	16(%rsp),%rsi -	leaq	(%r11),%rsp 	.byte	0xf3,0xc3  .LSEH_end_blst_sha256_block_data_order:@@ -1506,13 +1503,14 @@ .section	.xdata .p2align	3 .LSEH_info_blst_sha256_block_data_order_shaext_prologue:-.byte	1,0,5,0x0b-.byte	0,0x74,1,0-.byte	0,0x64,2,0-.byte	0,0x03-.byte	0,0+.byte	1,4,6,0x05+.byte	4,0x74,2,0+.byte	4,0x64,3,0+.byte	4,0x53+.byte	1,0x50+.long	0,0 .LSEH_info_blst_sha256_block_data_order_shaext_body:-.byte	1,0,15,0+.byte	1,0,17,85 .byte	0x00,0x68,0x00,0x00 .byte	0x00,0x78,0x01,0x00 .byte	0x00,0x88,0x02,0x00@@ -1520,41 +1518,45 @@ .byte	0x00,0xa8,0x04,0x00 .byte	0x00,0x74,0x0c,0x00 .byte	0x00,0x64,0x0d,0x00-.byte	0x00,0xa2+.byte	0x00,0x53+.byte	0x00,0x92+.byte	0x00,0x50 .byte	0x00,0x00,0x00,0x00,0x00,0x00+.byte	0x00,0x00,0x00,0x00 .LSEH_info_blst_sha256_block_data_order_shaext_epilogue:-.byte	1,0,5,11+.byte	1,0,4,0 .byte	0x00,0x74,0x01,0x00 .byte	0x00,0x64,0x02,0x00-.byte	0x00,0x03-.byte	0x00,0x00+.byte	0x00,0x00,0x00,0x00  .LSEH_info_blst_sha256_block_data_order_prologue:-.byte	1,0,5,0x0b-.byte	0,0x74,1,0-.byte	0,0x64,2,0-.byte	0,0x03-.byte	0,0+.byte	1,4,6,0x05+.byte	4,0x74,2,0+.byte	4,0x64,3,0+.byte	4,0x53+.byte	1,0x50+.long	0,0 .LSEH_info_blst_sha256_block_data_order_body:-.byte	1,0,26,5-.byte	0x00,0x68,0x02,0x00-.byte	0x00,0x78,0x03,0x00-.byte	0x00,0x88,0x04,0x00-.byte	0x00,0x98,0x05,0x00-.byte	0x00,0xf4,0x0d,0x00-.byte	0x00,0xe4,0x0e,0x00-.byte	0x00,0xd4,0x0f,0x00-.byte	0x00,0xc4,0x10,0x00-.byte	0x00,0x34,0x11,0x00-.byte	0x00,0x74,0x14,0x00-.byte	0x00,0x64,0x15,0x00-.byte	0x00,0x03-.byte	0x00,0x01,0x12,0x00+.byte	1,0,25,133+.byte	0x00,0x68,0x00,0x00+.byte	0x00,0x78,0x01,0x00+.byte	0x00,0x88,0x02,0x00+.byte	0x00,0x98,0x03,0x00+.byte	0x00,0xf4,0x0b,0x00+.byte	0x00,0xe4,0x0c,0x00+.byte	0x00,0xd4,0x0d,0x00+.byte	0x00,0xc4,0x0e,0x00+.byte	0x00,0x34,0x0f,0x00+.byte	0x00,0x74,0x12,0x00+.byte	0x00,0x64,0x13,0x00+.byte	0x00,0x53+.byte	0x00,0xf2 .byte	0x00,0x50+.byte	0x00,0x00,0x00,0x00,0x00,0x00+.byte	0x00,0x00,0x00,0x00 .LSEH_info_blst_sha256_block_data_order_epilogue:-.byte	1,0,5,11+.byte	1,0,4,0 .byte	0x00,0x74,0x01,0x00 .byte	0x00,0x64,0x02,0x00-.byte	0x00,0x03-.byte	0x00,0x00+.byte	0x00,0x00,0x00,0x00 

@@ -748,6 +748,21 @@  	ret .size	sgn0_pty_mod_384x,.-sgn0_pty_mod_384x+.globl	vec_select_32+.hidden	vec_select_32+.type	vec_select_32,%function+.align	5+vec_select_32:+	dup	v6.2d, x3+	ld1	{v0.2d, v1.2d, v2.2d}, [x1],#48+	cmeq	v6.2d, v6.2d, #0+	ld1	{v3.2d, v4.2d, v5.2d}, [x2],#48+	bit	v0.16b, v3.16b, v6.16b+	bit	v1.16b, v4.16b, v6.16b+	bit	v2.16b, v5.16b, v6.16b+	st1	{v0.2d, v1.2d, v2.2d}, [x0]+	ret+.size	vec_select_32,.-vec_select_32 .globl	vec_select_48 .hidden	vec_select_48 .type	vec_select_48,%function@@ -929,3 +944,57 @@ 	prfm	pldl1keep, [x0] 	ret .size	vec_prefetch,.-vec_prefetch+.globl	vec_is_zero_16x+.hidden	vec_is_zero_16x+.type	vec_is_zero_16x,%function+.align	5+vec_is_zero_16x:+	ld1	{v0.2d}, [x0], #16+	lsr	x1, x1, #4+	sub	x1, x1, #1+	cbz	x1, .Loop_is_zero_done++.Loop_is_zero:+	ld1	{v1.2d}, [x0], #16+	orr	v0.16b, v0.16b, v1.16b+	sub	x1, x1, #1+	cbnz	x1, .Loop_is_zero++.Loop_is_zero_done:+	dup	v1.2d, v0.d[1]+	orr	v0.16b, v0.16b, v1.16b+	mov	x1, v0.d[0]+	mov	x0, #1+	cmp	x1, #0+	csel	x0, x0, xzr, eq+	ret+.size	vec_is_zero_16x,.-vec_is_zero_16x+.globl	vec_is_equal_16x+.hidden	vec_is_equal_16x+.type	vec_is_equal_16x,%function+.align	5+vec_is_equal_16x:+	ld1	{v0.2d}, [x0], #16+	ld1	{v1.2d}, [x1], #16+	lsr	x2, x2, #4+	eor	v0.16b, v0.16b, v1.16b++.Loop_is_equal:+	sub	x2, x2, #1+	cbz	x2, .Loop_is_equal_done+	ld1	{v1.2d}, [x0], #16+	ld1	{v2.2d}, [x1], #16+	eor	v1.16b, v1.16b, v2.16b+	orr	v0.16b, v0.16b, v1.16b+	b	.Loop_is_equal+	nop++.Loop_is_equal_done:+	dup	v1.2d, v0.d[1]+	orr	v0.16b, v0.16b, v1.16b+	mov	x1, v0.d[0]+	mov	x0, #1+	cmp	x1, #0+	csel	x0, x0, xzr, eq+	ret+.size	vec_is_equal_16x,.-vec_is_equal_16x

@@ -1372,6 +1372,37 @@ 	.byte	0xf3,0xc3 .cfi_endproc	 .size	sgn0_pty_mod_384x,.-sgn0_pty_mod_384x+.globl	vec_select_32+.hidden	vec_select_32+.type	vec_select_32,@function+.align	32+vec_select_32:+.cfi_startproc+	.byte	0xf3,0x0f,0x1e,0xfa++	movd	%ecx,%xmm5+	pxor	%xmm4,%xmm4+	pshufd	$0,%xmm5,%xmm5+	movdqu	(%rsi),%xmm0+	leaq	16(%rsi),%rsi+	pcmpeqd	%xmm4,%xmm5+	movdqu	(%rdx),%xmm1+	leaq	16(%rdx),%rdx+	pcmpeqd	%xmm5,%xmm4+	leaq	16(%rdi),%rdi+	pand	%xmm4,%xmm0+	movdqu	0+16-16(%rsi),%xmm2+	pand	%xmm5,%xmm1+	movdqu	0+16-16(%rdx),%xmm3+	por	%xmm1,%xmm0+	movdqu	%xmm0,0-16(%rdi)+	pand	%xmm4,%xmm2+	pand	%xmm5,%xmm3+	por	%xmm3,%xmm2+	movdqu	%xmm2,16-16(%rdi)+	.byte	0xf3,0xc3+.cfi_endproc+.size	vec_select_32,.-vec_select_32 .globl	vec_select_48 .hidden	vec_select_48 .type	vec_select_48,@function@@ -1799,6 +1830,73 @@ 	.byte	0xf3,0xc3 .cfi_endproc .size	vec_prefetch,.-vec_prefetch+.globl	vec_is_zero_16x+.hidden	vec_is_zero_16x+.type	vec_is_zero_16x,@function+.align	32+vec_is_zero_16x:+.cfi_startproc+	.byte	0xf3,0x0f,0x1e,0xfa++	shrl	$4,%esi+	movdqu	(%rdi),%xmm0+	leaq	16(%rdi),%rdi++.Loop_is_zero:+	decl	%esi+	jz	.Loop_is_zero_done+	movdqu	(%rdi),%xmm1+	leaq	16(%rdi),%rdi+	por	%xmm1,%xmm0+	jmp	.Loop_is_zero++.Loop_is_zero_done:+	pshufd	$0x4e,%xmm0,%xmm1+	por	%xmm1,%xmm0+.byte	102,72,15,126,192+	incl	%esi+	testq	%rax,%rax+	cmovnzl	%esi,%eax+	xorl	$1,%eax+	.byte	0xf3,0xc3+.cfi_endproc+.size	vec_is_zero_16x,.-vec_is_zero_16x+.globl	vec_is_equal_16x+.hidden	vec_is_equal_16x+.type	vec_is_equal_16x,@function+.align	32+vec_is_equal_16x:+.cfi_startproc+	.byte	0xf3,0x0f,0x1e,0xfa++	shrl	$4,%edx+	movdqu	(%rdi),%xmm0+	movdqu	(%rsi),%xmm1+	subq	%rdi,%rsi+	leaq	16(%rdi),%rdi+	pxor	%xmm1,%xmm0++.Loop_is_equal:+	decl	%edx+	jz	.Loop_is_equal_done+	movdqu	(%rdi),%xmm1+	movdqu	(%rdi,%rsi,1),%xmm2+	leaq	16(%rdi),%rdi+	pxor	%xmm2,%xmm1+	por	%xmm1,%xmm0+	jmp	.Loop_is_equal++.Loop_is_equal_done:+	pshufd	$0x4e,%xmm0,%xmm1+	por	%xmm1,%xmm0+.byte	102,72,15,126,192+	incl	%edx+	testq	%rax,%rax+	cmovnzl	%edx,%eax+	xorl	$1,%eax+	.byte	0xf3,0xc3+.cfi_endproc+.size	vec_is_equal_16x,.-vec_is_equal_16x  .section	.note.GNU-stack,"",@progbits .section	.note.gnu.property,"a",@note

@@ -1,6 +1,7 @@ .text  .globl	ct_inverse_mod_256+.hidden	ct_inverse_mod_256 .type	ct_inverse_mod_256, %function .align	5 ct_inverse_mod_256:@@ -60,14 +61,14 @@ 	madd	x4, x16, x8, xzr	// |u|*|f0| 	madd	x4, x17, x9, x4	// |v|*|g0| 	str	x4, [x0,#8*4]-	asr	x5, x4, #63		// sign extenstion+	asr	x5, x4, #63		// sign extension 	stp	x5, x5, [x0,#8*5] 	stp	x5, x5, [x0,#8*7]  	madd	x4, x12, x8, xzr	// |u|*|f1| 	madd	x4, x13, x9, x4	// |v|*|g1| 	str	x4, [x0,#8*9]-	asr	x5, x4, #63		// sign extenstion+	asr	x5, x4, #63		// sign extension 	stp	x5, x5, [x0,#8*10] 	stp	x5, x5, [x0,#8*12] 	eor	x1, x1, #256		// flip-flop src |a|b|u|v|

@@ -1,6 +1,7 @@ .text	  .globl	ct_inverse_mod_256+.hidden	ct_inverse_mod_256 .type	ct_inverse_mod_256,@function .align	32 ct_inverse_mod_256:

@@ -1,6 +1,7 @@ .text  .globl	ct_inverse_mod_383+.hidden	ct_inverse_mod_383 .type	ct_inverse_mod_383, %function .align	5 ct_inverse_mod_383:@@ -71,7 +72,7 @@ 	adds	x3, x3, x5 	adc	x4, x4, x6 	stp	x3, x4, [x0,#8*6]-	asr	x5, x4, #63		// sign extenstion+	asr	x5, x4, #63		// sign extension 	stp	x5, x5, [x0,#8*8] 	stp	x5, x5, [x0,#8*10] @@ -82,7 +83,7 @@ 	adds	x3, x3, x5 	adc	x4, x4, x6 	stp	x3, x4, [x0,#8*12]-	asr	x5, x4, #63		// sign extenstion+	asr	x5, x4, #63		// sign extension 	stp	x5, x5, [x0,#8*14] 	stp	x5, x5, [x0,#8*16] 	eor	x1, x1, #256		// flip-flop src |a|b|u|v|

@@ -1,6 +1,7 @@ .text  .globl	ct_is_square_mod_384+.hidden	ct_is_square_mod_384 .type	ct_is_square_mod_384, %function .align	5 ct_is_square_mod_384:

@@ -1,6 +1,7 @@ .text	  .globl	ct_is_square_mod_384+.hidden	ct_is_square_mod_384 .type	ct_is_square_mod_384,@function .align	32 ct_is_square_mod_384:

@@ -1,6 +1,8 @@+.comm	__blst_platform_cap,4 .text	  .globl	ct_inverse_mod_383+.hidden	ct_inverse_mod_383 .type	ct_inverse_mod_383,@function .align	32 ct_inverse_mod_383:@@ -8,6 +10,10 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +#ifdef __BLST_PORTABLE__+	testl	$1,__blst_platform_cap(%rip)+	jnz	ct_inverse_mod_383$1+#endif 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16

@@ -1,6 +1,7 @@ .text	  .globl	ctx_inverse_mod_383+.hidden	ctx_inverse_mod_383 .type	ctx_inverse_mod_383,@function .align	32 ctx_inverse_mod_383:@@ -8,6 +9,7 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +ct_inverse_mod_383$1: 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16@@ -810,7 +812,7 @@  	movq	48(%rsi),%r10 -	call	__inner_loop_62+	call	__tail_loop_53   @@ -1521,9 +1523,9 @@ .cfi_endproc .size	__inner_loop_31,.-__inner_loop_31 -.type	__inner_loop_62,@function+.type	__tail_loop_53,@function .align	32-__inner_loop_62:+__tail_loop_53: .cfi_startproc 	.byte	0xf3,0x0f,0x1e,0xfa @@ -1532,7 +1534,7 @@ 	xorq	%r12,%r12 	movq	$1,%r13 -.Loop_62:+.Loop_53: 	xorq	%rax,%rax 	testq	$1,%r8 	movq	%r10,%rbx@@ -1559,11 +1561,11 @@ 	subq	%rax,%rdx 	subq	%rbx,%rcx 	subl	$1,%edi-	jnz	.Loop_62+	jnz	.Loop_53  	.byte	0xf3,0xc3 .cfi_endproc-.size	__inner_loop_62,.-__inner_loop_62+.size	__tail_loop_53,.-__tail_loop_53  .section	.note.GNU-stack,"",@progbits .section	.note.gnu.property,"a",@note

@@ -25,7 +25,7 @@ 	asr	x3,x0,#63	// top bit -> mask 	add	x0,x0,x0	// Q <<= 1 	subs	x6,x4,x1	// R - D-	add	x0,x0,#1	// Q + specilative bit+	add	x0,x0,#1	// Q + speculative bit 	sbcs	x7,x5,x2 	sbc	x0,x0,xzr	// subtract speculative bit 

@@ -8,6 +8,8 @@ .cfi_startproc 	.byte	0xf3,0x0f,0x1e,0xfa ++ 	movq	(%rdi),%r8 	movq	8(%rdi),%r9 	xorq	%rax,%rax@@ -39,8 +41,9 @@  	orq	%rcx,%rax + 	.byte	0xf3,0xc3-.cfi_endproc+.cfi_endproc	 .size	div_3_limbs,.-div_3_limbs .globl	quot_rem_128 .hidden	quot_rem_128@@ -50,6 +53,8 @@ .cfi_startproc 	.byte	0xf3,0x0f,0x1e,0xfa ++ 	movq	%rdx,%rax 	movq	%rdx,%rcx @@ -84,8 +89,9 @@  	movq	%rcx,%rax + 	.byte	0xf3,0xc3-.cfi_endproc+.cfi_endproc	 .size	quot_rem_128,.-quot_rem_128  @@ -100,6 +106,8 @@ .cfi_startproc 	.byte	0xf3,0x0f,0x1e,0xfa ++ 	movq	%rdx,%rax 	imulq	0(%rsi),%rdx @@ -110,8 +118,9 @@ 	movq	%r10,0(%rdi) 	movq	%rax,8(%rdi) + 	.byte	0xf3,0xc3-.cfi_endproc+.cfi_endproc	 .size	quot_rem_64,.-quot_rem_64  .section	.note.GNU-stack,"",@progbits

@@ -1,3 +1,4 @@+.comm	__blst_platform_cap,4 .text	  .globl	mul_mont_sparse_256@@ -9,6 +10,10 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +#ifdef __BLST_PORTABLE__+	testl	$1,__blst_platform_cap(%rip)+	jnz	mul_mont_sparse_256$1+#endif 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16@@ -73,6 +78,10 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +#ifdef __BLST_PORTABLE__+	testl	$1,__blst_platform_cap(%rip)+	jnz	sqr_mont_sparse_256$1+#endif 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16@@ -419,6 +428,10 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +#ifdef __BLST_PORTABLE__+	testl	$1,__blst_platform_cap(%rip)+	jnz	from_mont_256$1+#endif 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16@@ -494,6 +507,10 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +#ifdef __BLST_PORTABLE__+	testl	$1,__blst_platform_cap(%rip)+	jnz	redc_mont_256$1+#endif 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16

@@ -1,3 +1,4 @@+.comm	__blst_platform_cap,4 .text	  @@ -6,9 +7,9 @@   -.type	__sub_mod_384x384,@function+.type	__subq_mod_384x384,@function .align	32-__sub_mod_384x384:+__subq_mod_384x384: .cfi_startproc 	.byte	0xf3,0x0f,0x1e,0xfa @@ -73,11 +74,11 @@  	.byte	0xf3,0xc3 .cfi_endproc-.size	__sub_mod_384x384,.-__sub_mod_384x384+.size	__subq_mod_384x384,.-__subq_mod_384x384 -.type	__add_mod_384,@function+.type	__addq_mod_384,@function .align	32-__add_mod_384:+__addq_mod_384: .cfi_startproc 	.byte	0xf3,0x0f,0x1e,0xfa @@ -125,11 +126,11 @@  	.byte	0xf3,0xc3 .cfi_endproc-.size	__add_mod_384,.-__add_mod_384+.size	__addq_mod_384,.-__addq_mod_384 -.type	__sub_mod_384,@function+.type	__subq_mod_384,@function .align	32-__sub_mod_384:+__subq_mod_384: .cfi_startproc 	.byte	0xf3,0x0f,0x1e,0xfa @@ -140,7 +141,7 @@ 	movq	32(%rsi),%r12 	movq	40(%rsi),%r13 -__sub_mod_384_a_is_loaded:+__subq_mod_384_a_is_loaded: 	subq	0(%rdx),%r8 	movq	0(%rcx),%r14 	sbbq	8(%rdx),%r9@@ -177,7 +178,7 @@  	.byte	0xf3,0xc3 .cfi_endproc-.size	__sub_mod_384,.-__sub_mod_384+.size	__subq_mod_384,.-__subq_mod_384 .globl	mul_mont_384x .hidden	mul_mont_384x .type	mul_mont_384x,@function@@ -187,6 +188,10 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +#ifdef __BLST_PORTABLE__+	testl	$1,__blst_platform_cap(%rip)+	jnz	mul_mont_384x$1+#endif 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16@@ -232,12 +237,12 @@ 	movq	8(%rsp),%rcx 	leaq	-48(%rsi),%rdx 	leaq	40+192+48(%rsp),%rdi-	call	__add_mod_384+	call	__addq_mod_384  	movq	16(%rsp),%rsi 	leaq	48(%rsi),%rdx 	leaq	-48(%rdi),%rdi-	call	__add_mod_384+	call	__addq_mod_384  	leaq	(%rdi),%rbx 	leaq	48(%rdi),%rsi@@ -247,17 +252,17 @@ 	leaq	(%rdi),%rsi 	leaq	40(%rsp),%rdx 	movq	8(%rsp),%rcx-	call	__sub_mod_384x384+	call	__subq_mod_384x384  	leaq	(%rdi),%rsi 	leaq	-96(%rdi),%rdx-	call	__sub_mod_384x384+	call	__subq_mod_384x384   	leaq	40(%rsp),%rsi 	leaq	40+96(%rsp),%rdx 	leaq	40(%rsp),%rdi-	call	__sub_mod_384x384+	call	__subq_mod_384x384  	movq	%rcx,%rbx @@ -266,14 +271,14 @@ 	movq	0(%rsp),%rcx 	movq	32(%rsp),%rdi 	call	__mulq_by_1_mont_384-	call	__redc_tail_mont_384+	call	__redq_tail_mont_384   	leaq	40+192(%rsp),%rsi 	movq	0(%rsp),%rcx 	leaq	48(%rdi),%rdi 	call	__mulq_by_1_mont_384-	call	__redc_tail_mont_384+	call	__redq_tail_mont_384  	leaq	328(%rsp),%r8 	movq	0(%r8),%r15@@ -303,6 +308,10 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +#ifdef __BLST_PORTABLE__+	testl	$1,__blst_platform_cap(%rip)+	jnz	sqr_mont_384x$1+#endif 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16@@ -333,13 +342,13 @@  	leaq	48(%rsi),%rdx 	leaq	32(%rsp),%rdi-	call	__add_mod_384+	call	__addq_mod_384   	movq	16(%rsp),%rsi 	leaq	48(%rsi),%rdx 	leaq	32+48(%rsp),%rdi-	call	__sub_mod_384+	call	__subq_mod_384   	movq	16(%rsp),%rsi@@ -427,6 +436,10 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +#ifdef __BLST_PORTABLE__+	testl	$1,__blst_platform_cap(%rip)+	jnz	mul_382x$1+#endif 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16@@ -521,18 +534,18 @@ 	leaq	32(%rsp),%rdx 	movq	24(%rsp),%rcx 	movq	%rsi,%rdi-	call	__sub_mod_384x384+	call	__subq_mod_384x384   	leaq	0(%rdi),%rsi 	leaq	-96(%rdi),%rdx-	call	__sub_mod_384x384+	call	__subq_mod_384x384   	leaq	-96(%rdi),%rsi 	leaq	32(%rsp),%rdx 	leaq	-96(%rdi),%rdi-	call	__sub_mod_384x384+	call	__subq_mod_384x384  	leaq	136(%rsp),%r8 	movq	0(%r8),%r15@@ -562,6 +575,10 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +#ifdef __BLST_PORTABLE__+	testl	$1,__blst_platform_cap(%rip)+	jnz	sqr_382x$1+#endif 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16@@ -617,7 +634,7 @@  	leaq	48(%rsi),%rdx 	leaq	48(%rdi),%rdi-	call	__sub_mod_384_a_is_loaded+	call	__subq_mod_384_a_is_loaded   	leaq	(%rdi),%rsi@@ -695,6 +712,10 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +#ifdef __BLST_PORTABLE__+	testl	$1,__blst_platform_cap(%rip)+	jnz	mul_384$1+#endif 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16@@ -1019,6 +1040,10 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +#ifdef __BLST_PORTABLE__+	testl	$1,__blst_platform_cap(%rip)+	jnz	sqr_384$1+#endif 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16@@ -1265,6 +1290,10 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +#ifdef __BLST_PORTABLE__+	testl	$1,__blst_platform_cap(%rip)+	jnz	sqr_mont_384$1+#endif 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16@@ -1299,7 +1328,7 @@ 	movq	104(%rsp),%rbx 	movq	112(%rsp),%rdi 	call	__mulq_by_1_mont_384-	call	__redc_tail_mont_384+	call	__redq_tail_mont_384  	leaq	120(%rsp),%r8 	movq	120(%rsp),%r15@@ -1332,6 +1361,10 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +#ifdef __BLST_PORTABLE__+	testl	$1,__blst_platform_cap(%rip)+	jnz	redc_mont_384$1+#endif 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16@@ -1356,7 +1389,7 @@  	movq	%rdx,%rbx 	call	__mulq_by_1_mont_384-	call	__redc_tail_mont_384+	call	__redq_tail_mont_384  	movq	8(%rsp),%r15 .cfi_restore	%r15@@ -1389,6 +1422,10 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +#ifdef __BLST_PORTABLE__+	testl	$1,__blst_platform_cap(%rip)+	jnz	from_mont_384$1+#endif 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16@@ -1762,9 +1799,9 @@ .cfi_endproc .size	__mulq_by_1_mont_384,.-__mulq_by_1_mont_384 -.type	__redc_tail_mont_384,@function+.type	__redq_tail_mont_384,@function .align	32-__redc_tail_mont_384:+__redq_tail_mont_384: .cfi_startproc 	.byte	0xf3,0x0f,0x1e,0xfa @@ -1809,7 +1846,7 @@  	.byte	0xf3,0xc3 .cfi_endproc-.size	__redc_tail_mont_384,.-__redc_tail_mont_384+.size	__redq_tail_mont_384,.-__redq_tail_mont_384  .globl	sgn0_pty_mont_384 .hidden	sgn0_pty_mont_384@@ -1820,6 +1857,10 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +#ifdef __BLST_PORTABLE__+	testl	$1,__blst_platform_cap(%rip)+	jnz	sgn0_pty_mont_384$1+#endif 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16@@ -1898,6 +1939,10 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +#ifdef __BLST_PORTABLE__+	testl	$1,__blst_platform_cap(%rip)+	jnz	sgn0_pty_mont_384x$1+#endif 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16@@ -2025,6 +2070,10 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +#ifdef __BLST_PORTABLE__+	testl	$1,__blst_platform_cap(%rip)+	jnz	mul_mont_384$1+#endif 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16@@ -2689,6 +2738,10 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +#ifdef __BLST_PORTABLE__+	testl	$1,__blst_platform_cap(%rip)+	jnz	sqr_n_mul_mont_384$1+#endif 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16@@ -2727,7 +2780,7 @@ 	movq	0(%rsp),%rcx 	movq	16(%rsp),%rbx 	call	__mulq_by_1_mont_384-	call	__redc_tail_mont_384+	call	__redq_tail_mont_384  	movd	%xmm1,%edx 	leaq	0(%rdi),%rsi@@ -2777,6 +2830,10 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +#ifdef __BLST_PORTABLE__+	testl	$1,__blst_platform_cap(%rip)+	jnz	sqr_n_mul_mont_383$1+#endif 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16@@ -3438,6 +3495,10 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +#ifdef __BLST_PORTABLE__+	testl	$1,__blst_platform_cap(%rip)+	jnz	sqr_mont_382x$1+#endif 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16

@@ -9,6 +9,7 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +mul_mont_sparse_256$1: 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16@@ -71,6 +72,7 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +sqr_mont_sparse_256$1: 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16@@ -332,6 +334,7 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +from_mont_256$1: 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16@@ -407,6 +410,7 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +redc_mont_256$1: 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16

@@ -6,9 +6,9 @@   -.type	__sub_mod_384x384,@function+.type	__subx_mod_384x384,@function .align	32-__sub_mod_384x384:+__subx_mod_384x384: .cfi_startproc 	.byte	0xf3,0x0f,0x1e,0xfa @@ -73,11 +73,11 @@  	.byte	0xf3,0xc3 .cfi_endproc-.size	__sub_mod_384x384,.-__sub_mod_384x384+.size	__subx_mod_384x384,.-__subx_mod_384x384 -.type	__add_mod_384,@function+.type	__addx_mod_384,@function .align	32-__add_mod_384:+__addx_mod_384: .cfi_startproc 	.byte	0xf3,0x0f,0x1e,0xfa @@ -125,11 +125,11 @@  	.byte	0xf3,0xc3 .cfi_endproc-.size	__add_mod_384,.-__add_mod_384+.size	__addx_mod_384,.-__addx_mod_384 -.type	__sub_mod_384,@function+.type	__subx_mod_384,@function .align	32-__sub_mod_384:+__subx_mod_384: .cfi_startproc 	.byte	0xf3,0x0f,0x1e,0xfa @@ -140,7 +140,7 @@ 	movq	32(%rsi),%r12 	movq	40(%rsi),%r13 -__sub_mod_384_a_is_loaded:+__subx_mod_384_a_is_loaded: 	subq	0(%rdx),%r8 	movq	0(%rcx),%r14 	sbbq	8(%rdx),%r9@@ -177,7 +177,7 @@  	.byte	0xf3,0xc3 .cfi_endproc-.size	__sub_mod_384,.-__sub_mod_384+.size	__subx_mod_384,.-__subx_mod_384 .globl	mulx_mont_384x .hidden	mulx_mont_384x .type	mulx_mont_384x,@function@@ -187,6 +187,7 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +mul_mont_384x$1: 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16@@ -233,12 +234,12 @@ 	leaq	(%rbx),%rsi 	leaq	-48(%rbx),%rdx 	leaq	40+192+48(%rsp),%rdi-	call	__add_mod_384+	call	__addx_mod_384  	movq	24(%rsp),%rsi 	leaq	48(%rsi),%rdx 	leaq	-48(%rdi),%rdi-	call	__add_mod_384+	call	__addx_mod_384  	leaq	(%rdi),%rbx 	leaq	48(%rdi),%rsi@@ -248,17 +249,17 @@ 	leaq	(%rdi),%rsi 	leaq	40(%rsp),%rdx 	movq	8(%rsp),%rcx-	call	__sub_mod_384x384+	call	__subx_mod_384x384  	leaq	(%rdi),%rsi 	leaq	-96(%rdi),%rdx-	call	__sub_mod_384x384+	call	__subx_mod_384x384   	leaq	40(%rsp),%rsi 	leaq	40+96(%rsp),%rdx 	leaq	40(%rsp),%rdi-	call	__sub_mod_384x384+	call	__subx_mod_384x384  	leaq	(%rcx),%rbx @@ -267,14 +268,14 @@ 	movq	0(%rsp),%rcx 	movq	32(%rsp),%rdi 	call	__mulx_by_1_mont_384-	call	__redc_tail_mont_384+	call	__redx_tail_mont_384   	leaq	40+192(%rsp),%rsi 	movq	0(%rsp),%rcx 	leaq	48(%rdi),%rdi 	call	__mulx_by_1_mont_384-	call	__redc_tail_mont_384+	call	__redx_tail_mont_384  	leaq	328(%rsp),%r8 	movq	0(%r8),%r15@@ -304,6 +305,7 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +sqr_mont_384x$1: 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16@@ -335,13 +337,13 @@  	leaq	48(%rsi),%rdx 	leaq	32(%rsp),%rdi-	call	__add_mod_384+	call	__addx_mod_384   	movq	24(%rsp),%rsi 	leaq	48(%rsi),%rdx 	leaq	32+48(%rsp),%rdi-	call	__sub_mod_384+	call	__subx_mod_384   	movq	24(%rsp),%rsi@@ -439,6 +441,7 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +mul_382x$1: 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16@@ -533,18 +536,18 @@ 	leaq	32(%rsp),%rdx 	movq	24(%rsp),%rcx 	movq	%rsi,%rdi-	call	__sub_mod_384x384+	call	__subx_mod_384x384   	leaq	0(%rdi),%rsi 	leaq	-96(%rdi),%rdx-	call	__sub_mod_384x384+	call	__subx_mod_384x384   	leaq	-96(%rdi),%rsi 	leaq	32(%rsp),%rdx 	leaq	-96(%rdi),%rdi-	call	__sub_mod_384x384+	call	__subx_mod_384x384  	leaq	136(%rsp),%r8 	movq	0(%r8),%r15@@ -574,6 +577,7 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +sqr_382x$1: 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16@@ -629,7 +633,7 @@  	leaq	48(%rsi),%rdx 	leaq	48(%rdi),%rdi-	call	__sub_mod_384_a_is_loaded+	call	__subx_mod_384_a_is_loaded   	leaq	(%rdi),%rsi@@ -707,6 +711,7 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +mul_384$1: 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16@@ -933,6 +938,7 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +sqr_384$1: 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16@@ -1127,6 +1133,7 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +redc_mont_384$1: 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16@@ -1151,7 +1158,7 @@  	movq	%rdx,%rbx 	call	__mulx_by_1_mont_384-	call	__redc_tail_mont_384+	call	__redx_tail_mont_384  	movq	8(%rsp),%r15 .cfi_restore	%r15@@ -1184,6 +1191,7 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +from_mont_384$1: 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16@@ -1448,9 +1456,9 @@ .cfi_endproc .size	__mulx_by_1_mont_384,.-__mulx_by_1_mont_384 -.type	__redc_tail_mont_384,@function+.type	__redx_tail_mont_384,@function .align	32-__redc_tail_mont_384:+__redx_tail_mont_384: .cfi_startproc 	.byte	0xf3,0x0f,0x1e,0xfa @@ -1495,7 +1503,7 @@  	.byte	0xf3,0xc3 .cfi_endproc-.size	__redc_tail_mont_384,.-__redc_tail_mont_384+.size	__redx_tail_mont_384,.-__redx_tail_mont_384  .globl	sgn0x_pty_mont_384 .hidden	sgn0x_pty_mont_384@@ -1506,6 +1514,7 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +sgn0_pty_mont_384$1: 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16@@ -1584,6 +1593,7 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +sgn0_pty_mont_384x$1: 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16@@ -1711,6 +1721,7 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +mul_mont_384$1: 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16@@ -2178,6 +2189,7 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +sqr_mont_384$1: 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16@@ -2245,6 +2257,7 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +sqr_n_mul_mont_384$1: 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16@@ -2330,6 +2343,7 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +sqr_n_mul_mont_383$1: 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16@@ -2776,6 +2790,7 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +sqr_mont_382x$1: 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16

@@ -10,11 +10,12 @@ // // sha256_block procedure for ARMv8. //-// This module is stripped of scalar code paths, with raionale that all+// This module is stripped of scalar code paths, with rationale that all // known processors are NEON-capable. // // See original module at CRYPTOGAMS for further details. +.comm	__blst_platform_cap,4 .text  .align	6@@ -184,6 +185,11 @@ .type	blst_sha256_block_data_order,%function .align	4 blst_sha256_block_data_order:+	adrp	x16,__blst_platform_cap+	ldr	w16,[x16,#:lo12:__blst_platform_cap]+	tst	w16,#1+	b.ne	.Lv8_entry+ 	stp	x29, x30, [sp, #-16]! 	mov	x29, sp 	sub	sp,sp,#16*4

@@ -1,3 +1,4 @@+.comm	__blst_platform_cap,4 .text	  .globl	blst_sha256_block_data_order@@ -8,33 +9,35 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  -	pushq	%rbx-.cfi_adjust_cfa_offset	8-.cfi_offset	%rbx,-16 	pushq	%rbp .cfi_adjust_cfa_offset	8-.cfi_offset	%rbp,-24+.cfi_offset	%rbp,-16+	movq	%rsp,%rbp+.cfi_def_cfa_register	%rbp+#ifdef __BLST_PORTABLE__+	testl	$2,__blst_platform_cap(%rip)+	jnz	.Lblst_sha256_block_data_order$2+#endif+	pushq	%rbx+.cfi_offset	%rbx,-24 	pushq	%r12-.cfi_adjust_cfa_offset	8 .cfi_offset	%r12,-32 	pushq	%r13-.cfi_adjust_cfa_offset	8 .cfi_offset	%r13,-40 	pushq	%r14-.cfi_adjust_cfa_offset	8 .cfi_offset	%r14,-48 	pushq	%r15-.cfi_adjust_cfa_offset	8 .cfi_offset	%r15,-56 	shlq	$4,%rdx 	subq	$64+24,%rsp-.cfi_adjust_cfa_offset	16*4+3*8++.cfi_def_cfa	%rsp,144+ 	leaq	(%rsi,%rdx,4),%rdx 	movq	%rdi,64+0(%rsp) 	movq	%rsi,64+8(%rsp) 	movq	%rdx,64+16(%rsp) - 	movl	0(%rdi),%eax 	movl	4(%rdi),%ebx 	movl	8(%rdi),%ecx@@ -1636,23 +1639,23 @@ 	leaq	64+24+48(%rsp),%r11 .cfi_def_cfa	%r11,8 	movq	64+24(%rsp),%r15-.cfi_restore	%r15 	movq	-40(%r11),%r14-.cfi_restore	%r14 	movq	-32(%r11),%r13-.cfi_restore	%r13 	movq	-24(%r11),%r12+	movq	-16(%r11),%rbx+	movq	-8(%r11),%rbp .cfi_restore	%r12-	movq	-16(%r11),%rbp+.cfi_restore	%r13+.cfi_restore	%r14+.cfi_restore	%r15 .cfi_restore	%rbp-	movq	-8(%r11),%rbx .cfi_restore	%rbx- 	leaq	(%r11),%rsp 	.byte	0xf3,0xc3 .cfi_endproc	 .size	blst_sha256_block_data_order,.-blst_sha256_block_data_order +#ifndef __BLST_PORTABLE__ .align	64 .type	K256,@object K256:@@ -1744,6 +1747,7 @@ 	.byte	0xf3,0xc3 .cfi_endproc .size	blst_sha256_hcopy,.-blst_sha256_hcopy+#endif  .section	.note.GNU-stack,"",@progbits .section	.note.gnu.property,"a",@note

@@ -1,3 +1,4 @@+.comm	__blst_platform_cap,4 .text	  .align	64@@ -33,6 +34,13 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +	pushq	%rbp+.cfi_adjust_cfa_offset	8+.cfi_offset	%rbp,-16+	movq	%rsp,%rbp+.cfi_def_cfa_register	%rbp+.Lblst_sha256_block_data_order$2:+ 	leaq	K256+128(%rip),%rcx 	movdqu	(%rdi),%xmm1 	movdqu	16(%rdi),%xmm2@@ -234,6 +242,11 @@  	movdqu	%xmm1,(%rdi) 	movdqu	%xmm2,16(%rdi)+.cfi_def_cfa_register	%rsp+	popq	%rbp+.cfi_adjust_cfa_offset	-8+.cfi_restore	%rbp+ 	.byte	0xf3,0xc3 .cfi_endproc	 .size	blst_sha256_block_data_order_shaext,.-blst_sha256_block_data_order_shaext@@ -249,30 +262,27 @@ 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16+	movq	%rsp,%rbp+.cfi_def_cfa_register	%rbp+	testl	$2,__blst_platform_cap(%rip)+	jnz	.Lblst_sha256_block_data_order$2 	pushq	%rbx-.cfi_adjust_cfa_offset	8 .cfi_offset	%rbx,-24 	pushq	%r12-.cfi_adjust_cfa_offset	8 .cfi_offset	%r12,-32 	pushq	%r13-.cfi_adjust_cfa_offset	8 .cfi_offset	%r13,-40 	pushq	%r14-.cfi_adjust_cfa_offset	8 .cfi_offset	%r14,-48 	pushq	%r15-.cfi_adjust_cfa_offset	8 .cfi_offset	%r15,-56 	shlq	$4,%rdx-	subq	$40,%rsp-.cfi_adjust_cfa_offset	40+	subq	$24,%rsp+ 	leaq	(%rsi,%rdx,4),%rdx-	movq	%rdi,0(%rsp)+	movq	%rdi,-64(%rbp) -	movq	%rdx,16(%rsp)-	movq	%rsp,%rbp-.cfi_def_cfa_register	%rbp+	movq	%rdx,-48(%rbp)   	leaq	-64(%rsp),%rsp@@ -291,7 +301,7 @@ .align	16 .Lloop_ssse3: 	movdqa	K256+256(%rip),%xmm7-	movq	%rsi,8(%rbp)+	movq	%rsi,-56(%rbp) 	movdqu	0(%rsi),%xmm0 	movdqu	16(%rsi),%xmm1 	movdqu	32(%rsi),%xmm2@@ -1316,9 +1326,9 @@ 	addl	%r15d,%eax 	movl	%r8d,%r13d 	addl	%eax,%r14d-	movq	0(%rbp),%rdi+	movq	-64(%rbp),%rdi 	movl	%r14d,%eax-	movq	8(%rbp),%rsi+	movq	-56(%rbp),%rsi  	addl	0(%rdi),%eax 	addl	4(%rdi),%ebx@@ -1330,7 +1340,7 @@ 	addl	28(%rdi),%r11d  	leaq	64(%rsi),%rsi-	cmpq	16(%rbp),%rsi+	cmpq	-48(%rbp),%rsi  	movl	%eax,0(%rdi) 	movl	%ebx,4(%rdi)@@ -1343,26 +1353,25 @@ 	jb	.Lloop_ssse3  	xorps	%xmm0,%xmm0-	leaq	40+48(%rbp),%r11-.cfi_def_cfa	%r11,8 	movaps	%xmm0,0(%rsp) 	movaps	%xmm0,16(%rsp) 	movaps	%xmm0,32(%rsp) 	movaps	%xmm0,48(%rsp)-	movq	40(%rbp),%r15-.cfi_restore	%r15-	movq	-40(%r11),%r14-.cfi_restore	%r14-	movq	-32(%r11),%r13-.cfi_restore	%r13-	movq	-24(%r11),%r12+	movq	-40(%rbp),%r15+	movq	-32(%rbp),%r14+	movq	-24(%rbp),%r13+	movq	-16(%rbp),%r12+	movq	-8(%rbp),%rbx+	movq	%rbp,%rsp+.cfi_def_cfa_register	%rsp+	popq	%rbp+.cfi_adjust_cfa_offset	-8+.cfi_restore	%rbp .cfi_restore	%r12-	movq	-16(%r11),%rbx+.cfi_restore	%r13+.cfi_restore	%r14+.cfi_restore	%r15 .cfi_restore	%rbx-	movq	-8(%r11),%rbp-.cfi_restore	%rbp--	leaq	(%r11),%rsp 	.byte	0xf3,0xc3 .cfi_endproc	 .size	blst_sha256_block_data_order,.-blst_sha256_block_data_order

@@ -748,6 +748,21 @@  	ret +.globl	_vec_select_32+.private_extern	_vec_select_32++.align	5+_vec_select_32:+	dup	v6.2d, x3+	ld1	{v0.2d, v1.2d, v2.2d}, [x1],#48+	cmeq	v6.2d, v6.2d, #0+	ld1	{v3.2d, v4.2d, v5.2d}, [x2],#48+	bit	v0.16b, v3.16b, v6.16b+	bit	v1.16b, v4.16b, v6.16b+	bit	v2.16b, v5.16b, v6.16b+	st1	{v0.2d, v1.2d, v2.2d}, [x0]+	ret+ .globl	_vec_select_48 .private_extern	_vec_select_48 @@ -927,5 +942,59 @@ 	cmp	x0, x1 	csel	x0, x1, x0, hi 	prfm	pldl1keep, [x0]+	ret++.globl	_vec_is_zero_16x+.private_extern	_vec_is_zero_16x++.align	5+_vec_is_zero_16x:+	ld1	{v0.2d}, [x0], #16+	lsr	x1, x1, #4+	sub	x1, x1, #1+	cbz	x1, Loop_is_zero_done++Loop_is_zero:+	ld1	{v1.2d}, [x0], #16+	orr	v0.16b, v0.16b, v1.16b+	sub	x1, x1, #1+	cbnz	x1, Loop_is_zero++Loop_is_zero_done:+	dup	v1.2d, v0.d[1]+	orr	v0.16b, v0.16b, v1.16b+	mov	x1, v0.d[0]+	mov	x0, #1+	cmp	x1, #0+	csel	x0, x0, xzr, eq+	ret++.globl	_vec_is_equal_16x+.private_extern	_vec_is_equal_16x++.align	5+_vec_is_equal_16x:+	ld1	{v0.2d}, [x0], #16+	ld1	{v1.2d}, [x1], #16+	lsr	x2, x2, #4+	eor	v0.16b, v0.16b, v1.16b++Loop_is_equal:+	sub	x2, x2, #1+	cbz	x2, Loop_is_equal_done+	ld1	{v1.2d}, [x0], #16+	ld1	{v2.2d}, [x1], #16+	eor	v1.16b, v1.16b, v2.16b+	orr	v0.16b, v0.16b, v1.16b+	b	Loop_is_equal+	nop++Loop_is_equal_done:+	dup	v1.2d, v0.d[1]+	orr	v0.16b, v0.16b, v1.16b+	mov	x1, v0.d[0]+	mov	x0, #1+	cmp	x1, #0+	csel	x0, x0, xzr, eq 	ret 

@@ -1372,6 +1372,37 @@ 	.byte	0xf3,0xc3 .cfi_endproc	 +.globl	_vec_select_32+.private_extern	_vec_select_32++.p2align	5+_vec_select_32:+.cfi_startproc+	.byte	0xf3,0x0f,0x1e,0xfa++	movd	%ecx,%xmm5+	pxor	%xmm4,%xmm4+	pshufd	$0,%xmm5,%xmm5+	movdqu	(%rsi),%xmm0+	leaq	16(%rsi),%rsi+	pcmpeqd	%xmm4,%xmm5+	movdqu	(%rdx),%xmm1+	leaq	16(%rdx),%rdx+	pcmpeqd	%xmm5,%xmm4+	leaq	16(%rdi),%rdi+	pand	%xmm4,%xmm0+	movdqu	0+16-16(%rsi),%xmm2+	pand	%xmm5,%xmm1+	movdqu	0+16-16(%rdx),%xmm3+	por	%xmm1,%xmm0+	movdqu	%xmm0,0-16(%rdi)+	pand	%xmm4,%xmm2+	pand	%xmm5,%xmm3+	por	%xmm3,%xmm2+	movdqu	%xmm2,16-16(%rdi)+	.byte	0xf3,0xc3+.cfi_endproc+ .globl	_vec_select_48 .private_extern	_vec_select_48 @@ -1796,6 +1827,73 @@ 	cmpq	%rsi,%rdi 	cmovaq	%rsi,%rdi 	prefetchnta	(%rdi)+	.byte	0xf3,0xc3+.cfi_endproc++.globl	_vec_is_zero_16x+.private_extern	_vec_is_zero_16x++.p2align	5+_vec_is_zero_16x:+.cfi_startproc+	.byte	0xf3,0x0f,0x1e,0xfa++	shrl	$4,%esi+	movdqu	(%rdi),%xmm0+	leaq	16(%rdi),%rdi++L$oop_is_zero:+	decl	%esi+	jz	L$oop_is_zero_done+	movdqu	(%rdi),%xmm1+	leaq	16(%rdi),%rdi+	por	%xmm1,%xmm0+	jmp	L$oop_is_zero++L$oop_is_zero_done:+	pshufd	$0x4e,%xmm0,%xmm1+	por	%xmm1,%xmm0+.byte	102,72,15,126,192+	incl	%esi+	testq	%rax,%rax+	cmovnzl	%esi,%eax+	xorl	$1,%eax+	.byte	0xf3,0xc3+.cfi_endproc++.globl	_vec_is_equal_16x+.private_extern	_vec_is_equal_16x++.p2align	5+_vec_is_equal_16x:+.cfi_startproc+	.byte	0xf3,0x0f,0x1e,0xfa++	shrl	$4,%edx+	movdqu	(%rdi),%xmm0+	movdqu	(%rsi),%xmm1+	subq	%rdi,%rsi+	leaq	16(%rdi),%rdi+	pxor	%xmm1,%xmm0++L$oop_is_equal:+	decl	%edx+	jz	L$oop_is_equal_done+	movdqu	(%rdi),%xmm1+	movdqu	(%rdi,%rsi,1),%xmm2+	leaq	16(%rdi),%rdi+	pxor	%xmm2,%xmm1+	por	%xmm1,%xmm0+	jmp	L$oop_is_equal++L$oop_is_equal_done:+	pshufd	$0x4e,%xmm0,%xmm1+	por	%xmm1,%xmm0+.byte	102,72,15,126,192+	incl	%edx+	testq	%rax,%rax+	cmovnzl	%edx,%eax+	xorl	$1,%eax 	.byte	0xf3,0xc3 .cfi_endproc 

@@ -1,6 +1,7 @@ .text  .globl	_ct_inverse_mod_256+.private_extern	_ct_inverse_mod_256  .align	5 _ct_inverse_mod_256:@@ -60,14 +61,14 @@ 	madd	x4, x16, x8, xzr	// |u|*|f0| 	madd	x4, x17, x9, x4	// |v|*|g0| 	str	x4, [x0,#8*4]-	asr	x5, x4, #63		// sign extenstion+	asr	x5, x4, #63		// sign extension 	stp	x5, x5, [x0,#8*5] 	stp	x5, x5, [x0,#8*7]  	madd	x4, x12, x8, xzr	// |u|*|f1| 	madd	x4, x13, x9, x4	// |v|*|g1| 	str	x4, [x0,#8*9]-	asr	x5, x4, #63		// sign extenstion+	asr	x5, x4, #63		// sign extension 	stp	x5, x5, [x0,#8*10] 	stp	x5, x5, [x0,#8*12] 	eor	x1, x1, #256		// flip-flop src |a|b|u|v|

@@ -1,6 +1,7 @@ .text	  .globl	_ct_inverse_mod_256+.private_extern	_ct_inverse_mod_256  .p2align	5 _ct_inverse_mod_256:

@@ -1,6 +1,7 @@ .text  .globl	_ct_inverse_mod_383+.private_extern	_ct_inverse_mod_383  .align	5 _ct_inverse_mod_383:@@ -71,7 +72,7 @@ 	adds	x3, x3, x5 	adc	x4, x4, x6 	stp	x3, x4, [x0,#8*6]-	asr	x5, x4, #63		// sign extenstion+	asr	x5, x4, #63		// sign extension 	stp	x5, x5, [x0,#8*8] 	stp	x5, x5, [x0,#8*10] @@ -82,7 +83,7 @@ 	adds	x3, x3, x5 	adc	x4, x4, x6 	stp	x3, x4, [x0,#8*12]-	asr	x5, x4, #63		// sign extenstion+	asr	x5, x4, #63		// sign extension 	stp	x5, x5, [x0,#8*14] 	stp	x5, x5, [x0,#8*16] 	eor	x1, x1, #256		// flip-flop src |a|b|u|v|

@@ -1,6 +1,7 @@ .text  .globl	_ct_is_square_mod_384+.private_extern	_ct_is_square_mod_384  .align	5 _ct_is_square_mod_384:

@@ -1,6 +1,7 @@ .text	  .globl	_ct_is_square_mod_384+.private_extern	_ct_is_square_mod_384  .p2align	5 _ct_is_square_mod_384:

@@ -1,6 +1,8 @@+.comm	___blst_platform_cap,4 .text	  .globl	_ct_inverse_mod_383+.private_extern	_ct_inverse_mod_383  .p2align	5 _ct_inverse_mod_383:@@ -8,6 +10,10 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +#ifdef __BLST_PORTABLE__+	testl	$1,___blst_platform_cap(%rip)+	jnz	ct_inverse_mod_383$1+#endif 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16

@@ -1,6 +1,7 @@ .text	  .globl	_ctx_inverse_mod_383+.private_extern	_ctx_inverse_mod_383  .p2align	5 _ctx_inverse_mod_383:@@ -8,6 +9,7 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +ct_inverse_mod_383$1: 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16@@ -810,7 +812,7 @@  	movq	48(%rsi),%r10 -	call	__inner_loop_62+	call	__tail_loop_53   @@ -1523,7 +1525,7 @@   .p2align	5-__inner_loop_62:+__tail_loop_53: .cfi_startproc 	.byte	0xf3,0x0f,0x1e,0xfa @@ -1532,7 +1534,7 @@ 	xorq	%r12,%r12 	movq	$1,%r13 -L$oop_62:+L$oop_53: 	xorq	%rax,%rax 	testq	$1,%r8 	movq	%r10,%rbx@@ -1559,7 +1561,7 @@ 	subq	%rax,%rdx 	subq	%rbx,%rcx 	subl	$1,%edi-	jnz	L$oop_62+	jnz	L$oop_53  	.byte	0xf3,0xc3 .cfi_endproc

@@ -25,7 +25,7 @@ 	asr	x3,x0,#63	// top bit -> mask 	add	x0,x0,x0	// Q <<= 1 	subs	x6,x4,x1	// R - D-	add	x0,x0,#1	// Q + specilative bit+	add	x0,x0,#1	// Q + speculative bit 	sbcs	x7,x5,x2 	sbc	x0,x0,xzr	// subtract speculative bit 

@@ -8,6 +8,8 @@ .cfi_startproc 	.byte	0xf3,0x0f,0x1e,0xfa ++ 	movq	(%rdi),%r8 	movq	8(%rdi),%r9 	xorq	%rax,%rax@@ -39,8 +41,9 @@  	orq	%rcx,%rax + 	.byte	0xf3,0xc3-.cfi_endproc+.cfi_endproc	  .globl	_quot_rem_128 .private_extern	_quot_rem_128@@ -50,6 +53,8 @@ .cfi_startproc 	.byte	0xf3,0x0f,0x1e,0xfa ++ 	movq	%rdx,%rax 	movq	%rdx,%rcx @@ -84,8 +89,9 @@  	movq	%rcx,%rax + 	.byte	0xf3,0xc3-.cfi_endproc+.cfi_endproc	   @@ -100,6 +106,8 @@ .cfi_startproc 	.byte	0xf3,0x0f,0x1e,0xfa ++ 	movq	%rdx,%rax 	imulq	0(%rsi),%rdx @@ -110,6 +118,7 @@ 	movq	%r10,0(%rdi) 	movq	%rax,8(%rdi) + 	.byte	0xf3,0xc3-.cfi_endproc+.cfi_endproc	 

@@ -1,3 +1,4 @@+.comm	___blst_platform_cap,4 .text	  .globl	_mul_mont_sparse_256@@ -9,6 +10,10 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +#ifdef __BLST_PORTABLE__+	testl	$1,___blst_platform_cap(%rip)+	jnz	mul_mont_sparse_256$1+#endif 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16@@ -73,6 +78,10 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +#ifdef __BLST_PORTABLE__+	testl	$1,___blst_platform_cap(%rip)+	jnz	sqr_mont_sparse_256$1+#endif 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16@@ -419,6 +428,10 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +#ifdef __BLST_PORTABLE__+	testl	$1,___blst_platform_cap(%rip)+	jnz	from_mont_256$1+#endif 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16@@ -494,6 +507,10 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +#ifdef __BLST_PORTABLE__+	testl	$1,___blst_platform_cap(%rip)+	jnz	redc_mont_256$1+#endif 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16

@@ -1,3 +1,4 @@+.comm	___blst_platform_cap,4 .text	  @@ -8,7 +9,7 @@   .p2align	5-__sub_mod_384x384:+__subq_mod_384x384: .cfi_startproc 	.byte	0xf3,0x0f,0x1e,0xfa @@ -77,7 +78,7 @@   .p2align	5-__add_mod_384:+__addq_mod_384: .cfi_startproc 	.byte	0xf3,0x0f,0x1e,0xfa @@ -129,7 +130,7 @@   .p2align	5-__sub_mod_384:+__subq_mod_384: .cfi_startproc 	.byte	0xf3,0x0f,0x1e,0xfa @@ -140,7 +141,7 @@ 	movq	32(%rsi),%r12 	movq	40(%rsi),%r13 -__sub_mod_384_a_is_loaded:+__subq_mod_384_a_is_loaded: 	subq	0(%rdx),%r8 	movq	0(%rcx),%r14 	sbbq	8(%rdx),%r9@@ -187,6 +188,10 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +#ifdef __BLST_PORTABLE__+	testl	$1,___blst_platform_cap(%rip)+	jnz	mul_mont_384x$1+#endif 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16@@ -232,12 +237,12 @@ 	movq	8(%rsp),%rcx 	leaq	-48(%rsi),%rdx 	leaq	40+192+48(%rsp),%rdi-	call	__add_mod_384+	call	__addq_mod_384  	movq	16(%rsp),%rsi 	leaq	48(%rsi),%rdx 	leaq	-48(%rdi),%rdi-	call	__add_mod_384+	call	__addq_mod_384  	leaq	(%rdi),%rbx 	leaq	48(%rdi),%rsi@@ -247,17 +252,17 @@ 	leaq	(%rdi),%rsi 	leaq	40(%rsp),%rdx 	movq	8(%rsp),%rcx-	call	__sub_mod_384x384+	call	__subq_mod_384x384  	leaq	(%rdi),%rsi 	leaq	-96(%rdi),%rdx-	call	__sub_mod_384x384+	call	__subq_mod_384x384   	leaq	40(%rsp),%rsi 	leaq	40+96(%rsp),%rdx 	leaq	40(%rsp),%rdi-	call	__sub_mod_384x384+	call	__subq_mod_384x384  	movq	%rcx,%rbx @@ -266,14 +271,14 @@ 	movq	0(%rsp),%rcx 	movq	32(%rsp),%rdi 	call	__mulq_by_1_mont_384-	call	__redc_tail_mont_384+	call	__redq_tail_mont_384   	leaq	40+192(%rsp),%rsi 	movq	0(%rsp),%rcx 	leaq	48(%rdi),%rdi 	call	__mulq_by_1_mont_384-	call	__redc_tail_mont_384+	call	__redq_tail_mont_384  	leaq	328(%rsp),%r8 	movq	0(%r8),%r15@@ -303,6 +308,10 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +#ifdef __BLST_PORTABLE__+	testl	$1,___blst_platform_cap(%rip)+	jnz	sqr_mont_384x$1+#endif 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16@@ -333,13 +342,13 @@  	leaq	48(%rsi),%rdx 	leaq	32(%rsp),%rdi-	call	__add_mod_384+	call	__addq_mod_384   	movq	16(%rsp),%rsi 	leaq	48(%rsi),%rdx 	leaq	32+48(%rsp),%rdi-	call	__sub_mod_384+	call	__subq_mod_384   	movq	16(%rsp),%rsi@@ -427,6 +436,10 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +#ifdef __BLST_PORTABLE__+	testl	$1,___blst_platform_cap(%rip)+	jnz	mul_382x$1+#endif 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16@@ -521,18 +534,18 @@ 	leaq	32(%rsp),%rdx 	movq	24(%rsp),%rcx 	movq	%rsi,%rdi-	call	__sub_mod_384x384+	call	__subq_mod_384x384   	leaq	0(%rdi),%rsi 	leaq	-96(%rdi),%rdx-	call	__sub_mod_384x384+	call	__subq_mod_384x384   	leaq	-96(%rdi),%rsi 	leaq	32(%rsp),%rdx 	leaq	-96(%rdi),%rdi-	call	__sub_mod_384x384+	call	__subq_mod_384x384  	leaq	136(%rsp),%r8 	movq	0(%r8),%r15@@ -562,6 +575,10 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +#ifdef __BLST_PORTABLE__+	testl	$1,___blst_platform_cap(%rip)+	jnz	sqr_382x$1+#endif 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16@@ -617,7 +634,7 @@  	leaq	48(%rsi),%rdx 	leaq	48(%rdi),%rdi-	call	__sub_mod_384_a_is_loaded+	call	__subq_mod_384_a_is_loaded   	leaq	(%rdi),%rsi@@ -695,6 +712,10 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +#ifdef __BLST_PORTABLE__+	testl	$1,___blst_platform_cap(%rip)+	jnz	mul_384$1+#endif 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16@@ -1019,6 +1040,10 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +#ifdef __BLST_PORTABLE__+	testl	$1,___blst_platform_cap(%rip)+	jnz	sqr_384$1+#endif 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16@@ -1265,6 +1290,10 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +#ifdef __BLST_PORTABLE__+	testl	$1,___blst_platform_cap(%rip)+	jnz	sqr_mont_384$1+#endif 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16@@ -1299,7 +1328,7 @@ 	movq	104(%rsp),%rbx 	movq	112(%rsp),%rdi 	call	__mulq_by_1_mont_384-	call	__redc_tail_mont_384+	call	__redq_tail_mont_384  	leaq	120(%rsp),%r8 	movq	120(%rsp),%r15@@ -1332,6 +1361,10 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +#ifdef __BLST_PORTABLE__+	testl	$1,___blst_platform_cap(%rip)+	jnz	redc_mont_384$1+#endif 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16@@ -1356,7 +1389,7 @@  	movq	%rdx,%rbx 	call	__mulq_by_1_mont_384-	call	__redc_tail_mont_384+	call	__redq_tail_mont_384  	movq	8(%rsp),%r15 .cfi_restore	%r15@@ -1389,6 +1422,10 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +#ifdef __BLST_PORTABLE__+	testl	$1,___blst_platform_cap(%rip)+	jnz	from_mont_384$1+#endif 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16@@ -1764,7 +1801,7 @@   .p2align	5-__redc_tail_mont_384:+__redq_tail_mont_384: .cfi_startproc 	.byte	0xf3,0x0f,0x1e,0xfa @@ -1820,6 +1857,10 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +#ifdef __BLST_PORTABLE__+	testl	$1,___blst_platform_cap(%rip)+	jnz	sgn0_pty_mont_384$1+#endif 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16@@ -1898,6 +1939,10 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +#ifdef __BLST_PORTABLE__+	testl	$1,___blst_platform_cap(%rip)+	jnz	sgn0_pty_mont_384x$1+#endif 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16@@ -2025,6 +2070,10 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +#ifdef __BLST_PORTABLE__+	testl	$1,___blst_platform_cap(%rip)+	jnz	mul_mont_384$1+#endif 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16@@ -2689,6 +2738,10 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +#ifdef __BLST_PORTABLE__+	testl	$1,___blst_platform_cap(%rip)+	jnz	sqr_n_mul_mont_384$1+#endif 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16@@ -2727,7 +2780,7 @@ 	movq	0(%rsp),%rcx 	movq	16(%rsp),%rbx 	call	__mulq_by_1_mont_384-	call	__redc_tail_mont_384+	call	__redq_tail_mont_384  	movd	%xmm1,%edx 	leaq	0(%rdi),%rsi@@ -2777,6 +2830,10 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +#ifdef __BLST_PORTABLE__+	testl	$1,___blst_platform_cap(%rip)+	jnz	sqr_n_mul_mont_383$1+#endif 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16@@ -3438,6 +3495,10 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +#ifdef __BLST_PORTABLE__+	testl	$1,___blst_platform_cap(%rip)+	jnz	sqr_mont_382x$1+#endif 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16

@@ -9,6 +9,7 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +mul_mont_sparse_256$1: 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16@@ -71,6 +72,7 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +sqr_mont_sparse_256$1: 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16@@ -332,6 +334,7 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +from_mont_256$1: 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16@@ -407,6 +410,7 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +redc_mont_256$1: 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16

@@ -8,7 +8,7 @@   .p2align	5-__sub_mod_384x384:+__subx_mod_384x384: .cfi_startproc 	.byte	0xf3,0x0f,0x1e,0xfa @@ -77,7 +77,7 @@   .p2align	5-__add_mod_384:+__addx_mod_384: .cfi_startproc 	.byte	0xf3,0x0f,0x1e,0xfa @@ -129,7 +129,7 @@   .p2align	5-__sub_mod_384:+__subx_mod_384: .cfi_startproc 	.byte	0xf3,0x0f,0x1e,0xfa @@ -140,7 +140,7 @@ 	movq	32(%rsi),%r12 	movq	40(%rsi),%r13 -__sub_mod_384_a_is_loaded:+__subx_mod_384_a_is_loaded: 	subq	0(%rdx),%r8 	movq	0(%rcx),%r14 	sbbq	8(%rdx),%r9@@ -187,6 +187,7 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +mul_mont_384x$1: 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16@@ -233,12 +234,12 @@ 	leaq	(%rbx),%rsi 	leaq	-48(%rbx),%rdx 	leaq	40+192+48(%rsp),%rdi-	call	__add_mod_384+	call	__addx_mod_384  	movq	24(%rsp),%rsi 	leaq	48(%rsi),%rdx 	leaq	-48(%rdi),%rdi-	call	__add_mod_384+	call	__addx_mod_384  	leaq	(%rdi),%rbx 	leaq	48(%rdi),%rsi@@ -248,17 +249,17 @@ 	leaq	(%rdi),%rsi 	leaq	40(%rsp),%rdx 	movq	8(%rsp),%rcx-	call	__sub_mod_384x384+	call	__subx_mod_384x384  	leaq	(%rdi),%rsi 	leaq	-96(%rdi),%rdx-	call	__sub_mod_384x384+	call	__subx_mod_384x384   	leaq	40(%rsp),%rsi 	leaq	40+96(%rsp),%rdx 	leaq	40(%rsp),%rdi-	call	__sub_mod_384x384+	call	__subx_mod_384x384  	leaq	(%rcx),%rbx @@ -267,14 +268,14 @@ 	movq	0(%rsp),%rcx 	movq	32(%rsp),%rdi 	call	__mulx_by_1_mont_384-	call	__redc_tail_mont_384+	call	__redx_tail_mont_384   	leaq	40+192(%rsp),%rsi 	movq	0(%rsp),%rcx 	leaq	48(%rdi),%rdi 	call	__mulx_by_1_mont_384-	call	__redc_tail_mont_384+	call	__redx_tail_mont_384  	leaq	328(%rsp),%r8 	movq	0(%r8),%r15@@ -304,6 +305,7 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +sqr_mont_384x$1: 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16@@ -335,13 +337,13 @@  	leaq	48(%rsi),%rdx 	leaq	32(%rsp),%rdi-	call	__add_mod_384+	call	__addx_mod_384   	movq	24(%rsp),%rsi 	leaq	48(%rsi),%rdx 	leaq	32+48(%rsp),%rdi-	call	__sub_mod_384+	call	__subx_mod_384   	movq	24(%rsp),%rsi@@ -439,6 +441,7 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +mul_382x$1: 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16@@ -533,18 +536,18 @@ 	leaq	32(%rsp),%rdx 	movq	24(%rsp),%rcx 	movq	%rsi,%rdi-	call	__sub_mod_384x384+	call	__subx_mod_384x384   	leaq	0(%rdi),%rsi 	leaq	-96(%rdi),%rdx-	call	__sub_mod_384x384+	call	__subx_mod_384x384   	leaq	-96(%rdi),%rsi 	leaq	32(%rsp),%rdx 	leaq	-96(%rdi),%rdi-	call	__sub_mod_384x384+	call	__subx_mod_384x384  	leaq	136(%rsp),%r8 	movq	0(%r8),%r15@@ -574,6 +577,7 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +sqr_382x$1: 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16@@ -629,7 +633,7 @@  	leaq	48(%rsi),%rdx 	leaq	48(%rdi),%rdi-	call	__sub_mod_384_a_is_loaded+	call	__subx_mod_384_a_is_loaded   	leaq	(%rdi),%rsi@@ -707,6 +711,7 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +mul_384$1: 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16@@ -933,6 +938,7 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +sqr_384$1: 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16@@ -1127,6 +1133,7 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +redc_mont_384$1: 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16@@ -1151,7 +1158,7 @@  	movq	%rdx,%rbx 	call	__mulx_by_1_mont_384-	call	__redc_tail_mont_384+	call	__redx_tail_mont_384  	movq	8(%rsp),%r15 .cfi_restore	%r15@@ -1184,6 +1191,7 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +from_mont_384$1: 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16@@ -1450,7 +1458,7 @@   .p2align	5-__redc_tail_mont_384:+__redx_tail_mont_384: .cfi_startproc 	.byte	0xf3,0x0f,0x1e,0xfa @@ -1506,6 +1514,7 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +sgn0_pty_mont_384$1: 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16@@ -1584,6 +1593,7 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +sgn0_pty_mont_384x$1: 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16@@ -1711,6 +1721,7 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +mul_mont_384$1: 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16@@ -2178,6 +2189,7 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +sqr_mont_384$1: 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16@@ -2245,6 +2257,7 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +sqr_n_mul_mont_384$1: 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16@@ -2330,6 +2343,7 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +sqr_n_mul_mont_383$1: 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16@@ -2776,6 +2790,7 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +sqr_mont_382x$1: 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16

@@ -10,11 +10,12 @@ // // sha256_block procedure for ARMv8. //-// This module is stripped of scalar code paths, with raionale that all+// This module is stripped of scalar code paths, with rationale that all // known processors are NEON-capable. // // See original module at CRYPTOGAMS for further details. +.comm	___blst_platform_cap,4 .text  .align	6@@ -184,6 +185,11 @@  .align	4 _blst_sha256_block_data_order:+	adrp	x16,___blst_platform_cap@PAGE+	ldr	w16,[x16,___blst_platform_cap@PAGEOFF]+	tst	w16,#1+	b.ne	Lv8_entry+ 	stp	x29, x30, [sp, #-16]! 	mov	x29, sp 	sub	sp,sp,#16*4

@@ -1,3 +1,4 @@+.comm	___blst_platform_cap,4 .text	  .globl	_blst_sha256_block_data_order@@ -8,33 +9,35 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  -	pushq	%rbx-.cfi_adjust_cfa_offset	8-.cfi_offset	%rbx,-16 	pushq	%rbp .cfi_adjust_cfa_offset	8-.cfi_offset	%rbp,-24+.cfi_offset	%rbp,-16+	movq	%rsp,%rbp+.cfi_def_cfa_register	%rbp+#ifdef __BLST_PORTABLE__+	testl	$2,___blst_platform_cap(%rip)+	jnz	L$blst_sha256_block_data_order$2+#endif+	pushq	%rbx+.cfi_offset	%rbx,-24 	pushq	%r12-.cfi_adjust_cfa_offset	8 .cfi_offset	%r12,-32 	pushq	%r13-.cfi_adjust_cfa_offset	8 .cfi_offset	%r13,-40 	pushq	%r14-.cfi_adjust_cfa_offset	8 .cfi_offset	%r14,-48 	pushq	%r15-.cfi_adjust_cfa_offset	8 .cfi_offset	%r15,-56 	shlq	$4,%rdx 	subq	$64+24,%rsp-.cfi_adjust_cfa_offset	16*4+3*8++.cfi_def_cfa	%rsp,144+ 	leaq	(%rsi,%rdx,4),%rdx 	movq	%rdi,64+0(%rsp) 	movq	%rsi,64+8(%rsp) 	movq	%rdx,64+16(%rsp) - 	movl	0(%rdi),%eax 	movl	4(%rdi),%ebx 	movl	8(%rdi),%ecx@@ -1636,23 +1639,23 @@ 	leaq	64+24+48(%rsp),%r11 .cfi_def_cfa	%r11,8 	movq	64+24(%rsp),%r15-.cfi_restore	%r15 	movq	-40(%r11),%r14-.cfi_restore	%r14 	movq	-32(%r11),%r13-.cfi_restore	%r13 	movq	-24(%r11),%r12+	movq	-16(%r11),%rbx+	movq	-8(%r11),%rbp .cfi_restore	%r12-	movq	-16(%r11),%rbp+.cfi_restore	%r13+.cfi_restore	%r14+.cfi_restore	%r15 .cfi_restore	%rbp-	movq	-8(%r11),%rbx .cfi_restore	%rbx- 	leaq	(%r11),%rsp 	.byte	0xf3,0xc3 .cfi_endproc	  +#ifndef __BLST_PORTABLE__ .p2align	6  K256:@@ -1744,3 +1747,4 @@ 	.byte	0xf3,0xc3 .cfi_endproc +#endif

@@ -1,3 +1,4 @@+.comm	___blst_platform_cap,4 .text	  .p2align	6@@ -33,6 +34,13 @@ 	.byte	0xf3,0x0f,0x1e,0xfa  +	pushq	%rbp+.cfi_adjust_cfa_offset	8+.cfi_offset	%rbp,-16+	movq	%rsp,%rbp+.cfi_def_cfa_register	%rbp+L$blst_sha256_block_data_order$2:+ 	leaq	K256+128(%rip),%rcx 	movdqu	(%rdi),%xmm1 	movdqu	16(%rdi),%xmm2@@ -234,6 +242,11 @@  	movdqu	%xmm1,(%rdi) 	movdqu	%xmm2,16(%rdi)+.cfi_def_cfa_register	%rsp+	popq	%rbp+.cfi_adjust_cfa_offset	-8+.cfi_restore	%rbp+ 	.byte	0xf3,0xc3 .cfi_endproc	 @@ -249,30 +262,27 @@ 	pushq	%rbp .cfi_adjust_cfa_offset	8 .cfi_offset	%rbp,-16+	movq	%rsp,%rbp+.cfi_def_cfa_register	%rbp+	testl	$2,___blst_platform_cap(%rip)+	jnz	L$blst_sha256_block_data_order$2 	pushq	%rbx-.cfi_adjust_cfa_offset	8 .cfi_offset	%rbx,-24 	pushq	%r12-.cfi_adjust_cfa_offset	8 .cfi_offset	%r12,-32 	pushq	%r13-.cfi_adjust_cfa_offset	8 .cfi_offset	%r13,-40 	pushq	%r14-.cfi_adjust_cfa_offset	8 .cfi_offset	%r14,-48 	pushq	%r15-.cfi_adjust_cfa_offset	8 .cfi_offset	%r15,-56 	shlq	$4,%rdx-	subq	$40,%rsp-.cfi_adjust_cfa_offset	40+	subq	$24,%rsp+ 	leaq	(%rsi,%rdx,4),%rdx-	movq	%rdi,0(%rsp)+	movq	%rdi,-64(%rbp) -	movq	%rdx,16(%rsp)-	movq	%rsp,%rbp-.cfi_def_cfa_register	%rbp+	movq	%rdx,-48(%rbp)   	leaq	-64(%rsp),%rsp@@ -291,7 +301,7 @@ .p2align	4 L$loop_ssse3: 	movdqa	K256+256(%rip),%xmm7-	movq	%rsi,8(%rbp)+	movq	%rsi,-56(%rbp) 	movdqu	0(%rsi),%xmm0 	movdqu	16(%rsi),%xmm1 	movdqu	32(%rsi),%xmm2@@ -1316,9 +1326,9 @@ 	addl	%r15d,%eax 	movl	%r8d,%r13d 	addl	%eax,%r14d-	movq	0(%rbp),%rdi+	movq	-64(%rbp),%rdi 	movl	%r14d,%eax-	movq	8(%rbp),%rsi+	movq	-56(%rbp),%rsi  	addl	0(%rdi),%eax 	addl	4(%rdi),%ebx@@ -1330,7 +1340,7 @@ 	addl	28(%rdi),%r11d  	leaq	64(%rsi),%rsi-	cmpq	16(%rbp),%rsi+	cmpq	-48(%rbp),%rsi  	movl	%eax,0(%rdi) 	movl	%ebx,4(%rdi)@@ -1343,26 +1353,25 @@ 	jb	L$loop_ssse3  	xorps	%xmm0,%xmm0-	leaq	40+48(%rbp),%r11-.cfi_def_cfa	%r11,8 	movaps	%xmm0,0(%rsp) 	movaps	%xmm0,16(%rsp) 	movaps	%xmm0,32(%rsp) 	movaps	%xmm0,48(%rsp)-	movq	40(%rbp),%r15-.cfi_restore	%r15-	movq	-40(%r11),%r14-.cfi_restore	%r14-	movq	-32(%r11),%r13-.cfi_restore	%r13-	movq	-24(%r11),%r12+	movq	-40(%rbp),%r15+	movq	-32(%rbp),%r14+	movq	-24(%rbp),%r13+	movq	-16(%rbp),%r12+	movq	-8(%rbp),%rbx+	movq	%rbp,%rsp+.cfi_def_cfa_register	%rsp+	popq	%rbp+.cfi_adjust_cfa_offset	-8+.cfi_restore	%rbp .cfi_restore	%r12-	movq	-16(%r11),%rbx+.cfi_restore	%r13+.cfi_restore	%r14+.cfi_restore	%r15 .cfi_restore	%rbx-	movq	-8(%r11),%rbp-.cfi_restore	%rbp--	leaq	(%r11),%rsp 	.byte	0xf3,0xc3 .cfi_endproc	 

@@ -90,7 +90,7 @@  /*  * Optional |nbits|-wide |scalar| is used to facilitate multiple aggregated- * signature vetification as discussed at+ * signature verification as discussed at  * https://ethresear.ch/t/fast-verification-of-multiple-bls-signatures/5407.  * Usage pattern is not finalized yet, because (sig != NULL) is better and  * will be handled separately...

@@ -145,8 +145,7 @@ void prefix##s_add(ptype *sum, const ptype##_affine *const points[], \                                size_t npoints) \ { \-    /* Performance with 288K scratch is within 1-2-3% from optimal */ \-    const size_t stride = sizeof(ptype)==sizeof(POINTonE1) ? 2048 : 1024; \+    const size_t stride = SCRATCH_LIMIT / sizeof(ptype); \     ptype *scratch = alloca((npoints > stride ? stride : npoints) * \                             sizeof(ptype)); \     const ptype##_affine *point = NULL; \@@ -162,6 +161,15 @@         npoints -= j; \     } \ }++#ifndef SCRATCH_LIMIT+# ifdef __wasm__+#  define SCRATCH_LIMIT (45 * 1024)+# else+   /* Performance with 144K scratch is within 1-2-3% from optimal */+#  define SCRATCH_LIMIT (144 * 1024)+# endif+#endif  ADDITION_BTREE(blst_p1, POINTonE1, 384, fp, BLS12_381_Rx.p2) 

@@ -26,7 +26,7 @@          * 'if (n % sizeof(limb_t) == 0)' is omitted because it's cheaper          * to perform redundant stores than to pay penalty for          * mispredicted branch. Besides, some compilers unroll the-         * loop and remove redundant stores to 'restict'-ed storage...+         * loop and remove redundant stores to 'restrict'-ed storage...          */         ret[n / sizeof(limb_t)] = limb;     }@@ -55,7 +55,7 @@          * 'if (n % sizeof(limb_t) == 0)' is omitted because it's cheaper          * to perform redundant stores than to pay penalty for          * mispredicted branch. Besides, some compilers unroll the-         * loop and remove redundant stores to 'restict'-ed storage...+         * loop and remove redundant stores to 'restrict'-ed storage...          */         ret[n / sizeof(limb_t)] = limb;     }

@@ -0,0 +1,85 @@+/*+ * Copyright Supranational LLC+ * Licensed under the Apache License, Version 2.0, see LICENSE for details.+ * SPDX-License-Identifier: Apache-2.0+ */++#if (defined(__GNUC__) || defined(__clang__) || defined(__SUNPRO_C)) && !defined(_WIN32)+__attribute__((visibility("hidden")))+#endif+int __blst_platform_cap = 0;++#if defined(__x86_64__) || defined(__x86_64) || defined(_M_X64)++# if defined(__GNUC__) || defined(__clang__) || defined(__SUNPRO_C)+static void __cpuidex(int info[4], int func, int sub)+{+    int eax, ebx, ecx, edx;++    __asm__("cpuid" : "=a"(eax), "=b"(ebx), "=c"(ecx), "=d"(edx)+                    : "a"(func), "c"(sub));++    info[0] = eax;+    info[1] = ebx;+    info[2] = ecx;+    info[3] = edx;+}+# else+#  include <intrin.h>+# endif++# if defined(__GNUC__) || defined(__clang__)+__attribute__((constructor))+# endif+static int __blst_cpuid(void)+{+    int info[4], cap = 0;++    __cpuidex(info, 0, 0);+    if (info[0] > 6) {+        __cpuidex(info, 7, 0);+        cap |= (info[1]>>19) & 1; /* ADX */+        cap |= (info[1]>>28) & 2; /* SHA */+    }++    __blst_platform_cap = cap;++    return 0;+}++# if defined(_MSC_VER) && !defined(__clang__)+#  pragma section(".CRT$XCU",read)+__declspec(allocate(".CRT$XCU")) static int (*p)(void) = __blst_cpuid;+# elif defined(__SUNPRO_C)+#  pragma init(__blst_cpuid)+# endif++#elif defined(__aarch64__) || defined(__aarch64)++# if defined(__linux__) && (defined(__GNUC__) || defined(__clang__))+extern unsigned long getauxval(unsigned long type) __attribute__ ((weak));++__attribute__((constructor))+static int __blst_cpuid(void)+{+    int cap = 0;++    if (getauxval) {+        unsigned long hwcap_ce = getauxval(16);+        cap = (hwcap_ce>>6) & 1; /* SHA256 */+    }++    __blst_platform_cap = cap;++    return 0;+}+# elif defined(__APPLE__) && (defined(__GNUC__) || defined(__clang__))+__attribute__((constructor))+static int __blst_cpuid()+{+    __blst_platform_cap = 1; /* SHA256 */+    return 0;+}+# endif++#endif

@@ -155,7 +155,7 @@ {     if (vec_is_zero(in->X, 2*sizeof(in->X))) {         bytes_zero(out, 96);-        out[0] = 0x40;    /* infinitiy bit */+        out[0] = 0x40;    /* infinity bit */     } else {         (void)POINTonE1_affine_Serialize_BE(out, in);     }@@ -178,7 +178,7 @@ {     if (vec_is_zero(in->Z, sizeof(in->Z))) {         bytes_zero(out, 96);-        out[0] = 0x40;    /* infinitiy bit */+        out[0] = 0x40;    /* infinity bit */     } else {         (void)POINTonE1_Serialize_BE(out, in);     }@@ -202,7 +202,7 @@ {     if (vec_is_zero(in->X, 2*sizeof(in->X))) {         bytes_zero(out, 48);-        out[0] = 0xc0;    /* compressed and infinitiy bits */+        out[0] = 0xc0;    /* compressed and infinity bits */     } else {         limb_t sign = POINTonE1_affine_Compress_BE(out, in);         out[0] |= (unsigned char)(0x80 | ((sign & 2) << 4));@@ -226,7 +226,7 @@ {     if (vec_is_zero(in->Z, sizeof(in->Z))) {         bytes_zero(out, 48);-        out[0] = 0xc0;    /* compressed and infinitiy bits */+        out[0] = 0xc0;    /* compressed and infinity bits */     } else {         limb_t sign = POINTonE1_Compress_BE(out, in);         out[0] |= (unsigned char)(0x80 | ((sign & 2) << 4));

@@ -196,7 +196,7 @@ {     if (vec_is_zero(in->X, 2*sizeof(in->X))) {         bytes_zero(out, 192);-        out[0] = 0x40;    /* infinitiy bit */+        out[0] = 0x40;    /* infinity bit */     } else {         (void)POINTonE2_affine_Serialize_BE(out, in);     }@@ -219,7 +219,7 @@ {     if (vec_is_zero(in->Z, sizeof(in->Z))) {         bytes_zero(out, 192);-        out[0] = 0x40;    /* infinitiy bit */+        out[0] = 0x40;    /* infinity bit */     } else {         (void)POINTonE2_Serialize_BE(out, in);     }@@ -245,7 +245,7 @@ {     if (vec_is_zero(in->X, 2*sizeof(in->X))) {         bytes_zero(out, 96);-        out[0] = 0xc0;    /* compressed and infinitiy bits */+        out[0] = 0xc0;    /* compressed and infinity bits */     } else {         limb_t sign = POINTonE2_affine_Compress_BE(out, in);         out[0] |= (unsigned char)(0x80 | ((sign & 2) << 4));@@ -269,7 +269,7 @@ {     if (vec_is_zero(in->Z, sizeof(in->Z))) {         bytes_zero(out, 96);-        out[0] = 0xc0;    /* compressed and infinitiy bits */+        out[0] = 0xc0;    /* compressed and infinity bits */     } else {         limb_t sign = POINTonE2_Compress_BE(out, in);         out[0] |= (unsigned char)(0x80 | ((sign & 2) << 4));

@@ -46,9 +46,10 @@ static limb_t booth_encode(limb_t wval, size_t sz) {     limb_t mask = 0 - (wval >> sz);     /* "sign" bit -> mask */+    launder(mask);      wval = (wval + 1) >> 1;-    wval = (wval & ~mask) | ((0-wval) & mask);+    wval = (wval ^ mask) - mask;      /* &0x1f, but <=0x10, is index in table, rest is extended "sign" bit */     return wval;@@ -61,7 +62,7 @@  * pass order's bit-length, which is customarily publicly known, instead  * of the factual scalars' bit-lengths. This is facilitated by point  * addition subroutines implemented to handle points at infinity, which- * are encoded as Z==0. [Doubling agorithms handle such points at+ * are encoded as Z==0. [Doubling algorithms handle such points at  * infinity "naturally," since resulting Z is product of original Z.]  */ #define POINT_MULT_SCALAR_WX_IMPL(ptype, SZ) \

@@ -19,7 +19,7 @@ #include "bytes.h"  /*- * BLS12-381-specifc Fr shortcuts to assembly.+ * BLS12-381-specific Fr shortcuts to assembly.  */ void blst_fr_add(vec256 ret, const vec256 a, const vec256 b) {   add_mod_256(ret, a, b, BLS12_381_r);   }@@ -39,6 +39,24 @@ void blst_fr_mul(vec256 ret, const vec256 a, const vec256 b) {   mul_mont_sparse_256(ret, a, b, BLS12_381_r, r0);   } +void blst_fr_ct_bfly(vec256 x0, vec256 x1, const vec256 twiddle)+{+    vec256 x2;++    mul_mont_sparse_256(x2, x1, twiddle, BLS12_381_r, r0);+    sub_mod_256(x1, x0, x2, BLS12_381_r);+    add_mod_256(x0, x0, x2, BLS12_381_r);+}++void blst_fr_gs_bfly(vec256 x0, vec256 x1, const vec256 twiddle)+{+    vec256 x2;++    sub_mod_256(x2, x0, x1, BLS12_381_r);+    add_mod_256(x0, x0, x1, BLS12_381_r);+    mul_mont_sparse_256(x1, x2, twiddle, BLS12_381_r, r0);+}+ void blst_fr_sqr(vec256 ret, const vec256 a) {   sqr_mont_sparse_256(ret, a, BLS12_381_r, r0);   } @@ -102,27 +120,26 @@  int blst_sk_mul_n_check(pow256 ret, const pow256 a, const pow256 b) {-    vec256 a_fr, b_fr;+    vec256 t[2];     const union {         long one;         char little;     } is_endian = { 1 };+    bool_t is_zero;      if (((size_t)a|(size_t)b)%sizeof(limb_t) != 0 || !is_endian.little) {-        limbs_from_le_bytes(a_fr, a, sizeof(a_fr));-        limbs_from_le_bytes(b_fr, b, sizeof(a_fr));-        a = (const byte *)a_fr;-        b = (const byte *)b_fr;+        limbs_from_le_bytes(t[0], a, sizeof(pow256));+        limbs_from_le_bytes(t[1], b, sizeof(pow256));+        a = (const byte *)t[0];+        b = (const byte *)t[1];     }-    mul_mont_sparse_256(a_fr, (const limb_t *)a, BLS12_381_rRR,-                                                 BLS12_381_r, r0);-    mul_mont_sparse_256(b_fr, (const limb_t *)b, BLS12_381_rRR,-                                                 BLS12_381_r, r0);-    mul_mont_sparse_256(a_fr, a_fr, b_fr, BLS12_381_r, r0);-    from_mont_256(a_fr, a_fr, BLS12_381_r, r0);-    le_bytes_from_limbs(ret, a_fr, sizeof(a_fr));+    mul_mont_sparse_256(t[0], BLS12_381_rRR, (const limb_t *)a, BLS12_381_r, r0);+    mul_mont_sparse_256(t[0], t[0], (const limb_t *)b, BLS12_381_r, r0);+    le_bytes_from_limbs(ret, t[0], sizeof(pow256));+    is_zero = vec_is_zero(t[0], sizeof(vec256));+    vec_zero(t, sizeof(t)); -    return (int)(vec_is_zero(a_fr, sizeof(a_fr)) ^ 1);+    return (int)(is_zero^1); }  void blst_sk_inverse(pow256 ret, const pow256 a)@@ -150,7 +167,7 @@ }  /*- * BLS12-381-specifc Fp shortcuts to assembly.+ * BLS12-381-specific Fp shortcuts to assembly.  */ void blst_fp_add(vec384 ret, const vec384 a, const vec384 b) {   add_fp(ret, a, b);   }@@ -284,7 +301,7 @@ }  /*- * BLS12-381-specifc Fp2 shortcuts to assembly.+ * BLS12-381-specific Fp2 shortcuts to assembly.  */ void blst_fp2_add(vec384x ret, const vec384x a, const vec384x b) {   add_fp2(ret, a, b);   }@@ -311,7 +328,7 @@ {   cneg_fp2(ret, a, is_zero(flag) ^ 1);   }  /*- * Scalar serialization/deseriazation+ * Scalar serialization/deserialization.  */ void blst_scalar_from_uint32(pow256 ret, const unsigned int a[8]) {@@ -480,68 +497,75 @@  int blst_scalar_from_le_bytes(pow256 out, const unsigned char *bytes, size_t n) {-    struct { vec256 out, digit, radix; } t;+    size_t rem = (n - 1) % 32 + 1;+    struct { vec256 out, digit; } t;     limb_t ret;      vec_zero(t.out, sizeof(t.out));-    vec_copy(t.radix, BLS12_381_rRR, sizeof(t.radix)); -    while (n > 32) {-        limbs_from_le_bytes(t.digit, bytes, 32);-        from_mont_256(t.digit, t.digit, BLS12_381_r, r0);-        mul_mont_sparse_256(t.digit, t.digit, t.radix, BLS12_381_r, r0);+    n -= rem;+    limbs_from_le_bytes(t.out, bytes += n, rem);+    mul_mont_sparse_256(t.out, BLS12_381_rRR, t.out, BLS12_381_r, r0);++    while (n) {+        limbs_from_le_bytes(t.digit, bytes -= 32, 32);         add_mod_256(t.out, t.out, t.digit, BLS12_381_r);-        mul_mont_sparse_256(t.radix, t.radix, BLS12_381_rRR, BLS12_381_r, r0);-        bytes += 32;+        mul_mont_sparse_256(t.out, BLS12_381_rRR, t.out, BLS12_381_r, r0);         n -= 32;     } -    vec_zero(t.digit, sizeof(t.digit));-    limbs_from_le_bytes(t.digit, bytes, n);-    from_mont_256(t.digit, t.digit, BLS12_381_r, r0);-    mul_mont_sparse_256(t.digit, t.digit, t.radix, BLS12_381_r, r0);-    add_mod_256(t.out, t.out, t.digit, BLS12_381_r);+    from_mont_256(t.out, t.out, BLS12_381_r, r0);      ret = vec_is_zero(t.out, sizeof(t.out));     le_bytes_from_limbs(out, t.out, 32);-    vec_zero(t.out, 2*sizeof(t.out));+    vec_zero(&t, sizeof(t));      return (int)(ret^1); }  int blst_scalar_from_be_bytes(pow256 out, const unsigned char *bytes, size_t n) {-    struct { vec256 out, digit, radix; } t;+    size_t rem = (n - 1) % 32 + 1;+    struct { vec256 out, digit; } t;     limb_t ret;      vec_zero(t.out, sizeof(t.out));-    vec_copy(t.radix, BLS12_381_rRR, sizeof(t.radix)); -    bytes += n;-    while (n > 32) {-        limbs_from_be_bytes(t.digit, bytes -= 32, 32);-        from_mont_256(t.digit, t.digit, BLS12_381_r, r0);-        mul_mont_sparse_256(t.digit, t.digit, t.radix, BLS12_381_r, r0);+    limbs_from_be_bytes(t.out, bytes, rem);+    mul_mont_sparse_256(t.out, BLS12_381_rRR, t.out, BLS12_381_r, r0);++    while (n -= rem) {+        limbs_from_be_bytes(t.digit, bytes += rem, 32);         add_mod_256(t.out, t.out, t.digit, BLS12_381_r);-        mul_mont_sparse_256(t.radix, t.radix, BLS12_381_rRR, BLS12_381_r, r0);-        n -= 32;+        mul_mont_sparse_256(t.out, BLS12_381_rRR, t.out, BLS12_381_r, r0);+        rem = 32;     } -    vec_zero(t.digit, sizeof(t.digit));-    limbs_from_be_bytes(t.digit, bytes -= n, n);-    from_mont_256(t.digit, t.digit, BLS12_381_r, r0);-    mul_mont_sparse_256(t.digit, t.digit, t.radix, BLS12_381_r, r0);-    add_mod_256(t.out, t.out, t.digit, BLS12_381_r);+    from_mont_256(t.out, t.out, BLS12_381_r, r0);      ret = vec_is_zero(t.out, sizeof(t.out));     le_bytes_from_limbs(out, t.out, 32);-    vec_zero(t.out, 2*sizeof(t.out));+    vec_zero(&t, sizeof(t));      return (int)(ret^1); }  /*- * Test facilitator+ * Single-short SHA-256 hash function.+ */+#include "sha256.h"++void blst_sha256(unsigned char md[32], const void *msg, size_t len)+{+    SHA256_CTX ctx;++    sha256_init(&ctx);+    sha256_update(&ctx, msg, len);+    sha256_final(md, &ctx);+}++/*+ * Test facilitator.  */ void blst_scalar_from_hexascii(pow256 ret, const char *hex) {   bytes_from_hexascii(ret, sizeof(pow256), hex);   }

@@ -10,7 +10,7 @@ #include "consts.h"  /*- * BLS12-381-specifc Fp shortcuts to assembly.+ * BLS12-381-specific Fp shortcuts to assembly.  */ static inline void add_fp(vec384 ret, const vec384 a, const vec384 b) {   add_mod_384(ret, a, b, BLS12_381_P);   }@@ -49,7 +49,7 @@ {   redc_mont_384(ret, a, BLS12_381_P, p0);   }  /*- * BLS12-381-specifc Fp2 shortcuts to assembly.+ * BLS12-381-specific Fp2 shortcuts to assembly.  */ static inline void add_fp2(vec384x ret, const vec384x a, const vec384x b) {   add_mod_384x(ret, a, b, BLS12_381_P);   }