Merge bitcoin-core/secp256k1#1307: Mark more assembly outputs as early clobber

8c9ae37a5a26cdeb6365624fee43f41b238830e4 Add release note (Pieter Wuille)
350b4bd6e6efd3c62875820fdeb2740738937922 Mark stack variables as early clobber for technical correctness (Pieter Wuille)
0c729ba70d963f2798184b0b8524d7de2f3ced9f Bugfix: mark outputs as early clobber in scalar x86_64 asm (Pieter Wuille)

Pull request description:

ACKs for top commit:
  real-or-random:
    ACK 8c9ae37a5a26cdeb6365624fee43f41b238830e4
  jonasnick:
    ACK 8c9ae37a5a26cdeb6365624fee43f41b238830e4

Tree-SHA512: 874d01f5540d14b5188aec25f6441dbc6631f8d3980416040a3e250f1aef75150068415e7a458a9a3fb0d7cbdeb97f5c7e089b187d6d3dd79aa6e45274c241b6

CHANGELOG.md

@@ -7,6 +7,9 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [Unreleased]
 
+#### Fixed
+ - Fixed an old bug that permitted compilers to potentially output bad assembly code on x86_64. In theory, it could lead to a crash or a read of unrelated memory, but this has never been observed on any compilers so far.
+
 ## [0.3.1] - 2023-04-10
 We strongly recommend updating to 0.3.1 if you use or plan to use Clang >=14 to compile libsecp256k1, e.g., Xcode >=14 on macOS has Clang >=14. When in doubt, check the Clang version using `clang -v`.
 

src/field_5x52_asm_impl.h

@@ -280,7 +280,7 @@ __asm__ __volatile__(
     "addq %%rsi,%%r8\n"
     /* r[4] = c */
     "movq %%r8,32(%%rdi)\n"
-: "+S"(a), "=m"(tmp1), "=m"(tmp2), "=m"(tmp3)
+: "+S"(a), "=&m"(tmp1), "=&m"(tmp2), "=&m"(tmp3)
 : "b"(b), "D"(r)
 : "%rax", "%rcx", "%rdx", "%r8", "%r9", "%r10", "%r11", "%r12", "%r13", "%r14", "%r15", "cc", "memory"
 );
@@ -495,7 +495,7 @@ __asm__ __volatile__(
     "addq %%rsi,%%r8\n"
     /* r[4] = c */
     "movq %%r8,32(%%rdi)\n"
-: "+S"(a), "=m"(tmp1), "=m"(tmp2), "=m"(tmp3)
+: "+S"(a), "=&m"(tmp1), "=&m"(tmp2), "=&m"(tmp3)
 : "D"(r)
 : "%rax", "%rbx", "%rcx", "%rdx", "%r8", "%r9", "%r10", "%r11", "%r12", "%r13", "%r14", "%r15", "cc", "memory"
 );

src/scalar_4x64_impl.h

@@ -383,7 +383,7 @@ static void secp256k1_scalar_reduce_512(secp256k1_scalar *r, const uint64_t *l)
     "movq %%r10, %q5\n"
     /* extract m6 */
     "movq %%r8, %q6\n"
-    : "=g"(m0), "=g"(m1), "=g"(m2), "=g"(m3), "=g"(m4), "=g"(m5), "=g"(m6)
+    : "=&g"(m0), "=&g"(m1), "=&g"(m2), "=g"(m3), "=g"(m4), "=g"(m5), "=g"(m6)
     : "S"(l), "i"(SECP256K1_N_C_0), "i"(SECP256K1_N_C_1)
     : "rax", "rdx", "r8", "r9", "r10", "r11", "r12", "r13", "r14", "cc");