fix libsnark build with optimisations on gcc/g++ 11.x

It always was default behavior to build KomodoOcean with `-DUSE_ASM=1`
which is set by default through `config/bitcoin-config.h`. This flag
used not only by `libsnark`, but also by the hardrware SHA256 optimisations
in other sources, etc. However, it turns out that on g++ 11.3.0 enabled
optimisations for libsnark makes ibzcash::PHGRProof (SproutProofVerifier)
not working, and as a result when users tried to sync old chains which
had `sprout` transactions in history, like `SUPERNET`, the verification
of valid transactions throw an error:

```
2023-07-11 15:01:07 ERROR: CheckTransaction(): joinsplit does not verify
2023-07-11 15:01:07 ERROR: CheckBlock: CheckTransaction failed
```

While on gcc/g++ 10.x there was no such error. That's why we decided to
disable `libsnark` optimisations at all, by "shadowing" `USE_ASM` in
internal libnark implementation and rename it to `USE_ASM_SNARK`, to
avoid build with ASM optimisations, which are "unstable" on modern compilers.

More details here:

KomodoPlatform/komodo#591

src/snark/Makefile

@@ -18,9 +18,9 @@ CURVE = BN128
 OPTFLAGS = -O2 -march=x86-64 -g -mtune=x86-64
 
 ifneq ($(PLATFORM),darwin)
-	FEATUREFLAGS = -DUSE_ASM -DMONTGOMERY_OUTPUT
+	FEATUREFLAGS = -DUSE_ASM_SNARK -DMONTGOMERY_OUTPUT
 else
-	FEATUREFLAGS = -DUSE_ASM -DMONTGOMERY_OUTPUT -D__SIZE_TYPE__="unsigned long long"
+	FEATUREFLAGS = -DUSE_ASM_SNARK -DMONTGOMERY_OUTPUT -D__SIZE_TYPE__="unsigned long long"
 endif
 
 # Initialize this using "CXXFLAGS=... make". The makefile appends to that.

src/snark/README.md

@@ -426,7 +426,7 @@ The following flags change the behavior of the compiled code.
     of the corresponding algebraic objects. This option works for all
     curves except bn128.
 
-*   define `USE_ASM` (on by default)
+*   define `USE_ASM_SNARK` (on by default)
 
     Use unrolled assembly routines for F[p] arithmetic and faster heap in
     multi-exponentiation. (When not set, use GMP's `mpn_*` routines instead.)
@@ -472,7 +472,7 @@ with respect to portability. Specifically:
    tested with g++ 4.7, g++ 4.8, and clang 3.4.
 
 6. On x86-64, we by default use highly optimized assembly implementations for some
-   operations (see `USE_ASM` above). On other architectures we fall back to a
+   operations (see `USE_ASM_SNARK` above). On other architectures we fall back to a
    portable C++ implementation, which is slower.
 
 Tested configurations include:

src/snark/libsnark/algebra/fields/fp.tcc

@@ -23,7 +23,7 @@ template<mp_size_t n, const bigint<n>& modulus>
 void Fp_model<n,modulus>::mul_reduce(const bigint<n> &other)
 {
     /* stupid pre-processor tricks; beware */
-#if defined(__x86_64__) && defined(USE_ASM)
+#if defined(__x86_64__) && defined(USE_ASM_SNARK)
     if (n == 3)
     { // Use asm-optimized Comba multiplication and reduction
         mp_limb_t res[2*n];
@@ -293,7 +293,7 @@ Fp_model<n,modulus>& Fp_model<n,modulus>::operator+=(const Fp_model<n,modulus>&
 #ifdef PROFILE_OP_COUNTS
     this->add_cnt++;
 #endif
-#if defined(__x86_64__) && defined(USE_ASM)
+#if defined(__x86_64__) && defined(USE_ASM_SNARK)
     if (n == 3)
     {
         __asm__
@@ -406,7 +406,7 @@ Fp_model<n,modulus>& Fp_model<n,modulus>::operator-=(const Fp_model<n,modulus>&
 #ifdef PROFILE_OP_COUNTS
     this->sub_cnt++;
 #endif
-#if defined(__x86_64__) && defined(USE_ASM)
+#if defined(__x86_64__) && defined(USE_ASM_SNARK)
     if (n == 3)
     {
         __asm__
@@ -579,7 +579,7 @@ Fp_model<n,modulus> Fp_model<n,modulus>::squared() const
     this->mul_cnt--; // zero out the upcoming mul
 #endif
     /* stupid pre-processor tricks; beware */
-#if defined(__x86_64__) && defined(USE_ASM)
+#if defined(__x86_64__) && defined(USE_ASM_SNARK)
     if (n == 3)
     { // use asm-optimized Comba squaring
         mp_limb_t res[2*n];

src/snark/libsnark/algebra/fields/fp_aux.tcc

@@ -1,8 +1,8 @@
 /** @file
  *****************************************************************************
  Assembly code snippets for F[p] finite field arithmetic, used by fp.tcc .
- Specific to x86-64, and used only if USE_ASM is defined.
- On other architectures or without USE_ASM, fp.tcc uses a portable
+ Specific to x86-64, and used only if USE_ASM_SNARK is defined.
+ On other architectures or without USE_ASM_SNARK, fp.tcc uses a portable
  C++ implementation instead.
  *****************************************************************************
  * @author     This file is part of libsnark, developed by SCIPR Lab

src/snark/libsnark/algebra/scalar_multiplication/multiexp.tcc

@@ -38,7 +38,7 @@ public:
 
     bool operator<(const ordered_exponent<n> &other) const
     {
-#if defined(__x86_64__) && defined(USE_ASM)
+#if defined(__x86_64__) && defined(USE_ASM_SNARK)
         if (n == 3)
         {
             int64_t res;