mimalloc: prevent crashes in the post-command hook handling

Similar to 1475d9f (mimalloc: use "weak" random seed when
statically linked, 2023-05-12) (which was actually already applied
upstream in microsoft/mimalloc@3e1d800e
(potential fix for windows static linking with thread creation in dll's,
2022-11-07), this ensures that another code path uses "weak" random
seeds when initializing mimalloc.

The scenario when this happens is likely the exact same as back when
the original fix was implemented (but I can only hypothesize about this,
as I was unable to find a reproducer in my setup, and even the reporters
could not find a reliable reproducer, either): an `atexit()` handler
is called after mimalloc's own `atexit()` handler deinitializes
mimalloc. This `atexit()` handler (namely `post_command_hook_atexit()`)
then calls `getenv()`, which internally resolves to `mingw_getenv()`,
which wants to `calloc()` the result so that it is not overwritten
while in use. This, in turn, uses mimalloc, which dutifully initializes
the random seed. It apparently takes a different code path than back in
2023, though, where it now tries to initialize a strong random generator
implemented in bcrypt.dll. However, that initialization fails because
it happens during the phase when only `atexit()` handlers are called,
and bcrypt has obviously already been deinitialized, in a way that
causes a crash.

The stack trace in question looks like this:

 ntdll!ZwWaitForMultipleObjects+0x140 ...
 KERNELBASE!WaitForMultipleObjectsEx+0x123 ...
 KERNELBASE!WaitForMultipleObjects+0x11 ...
 kernel32!WerpReportFaultInternal+0x62c ...
 kernel32!WerpReportFault+0xc5 ...
 KERNELBASE!UnhandledExceptionFilter+0x34c ...
 ntdll!RtlpThreadExceptionFilter+0x2e ...
 ntdll!RtlUserThreadStart$filt$0+0x3f0 ...
 ntdll !_ C_specific_handler+0x93 ...
 ntdll!RtlpExecuteHandlerForException+0xf ...
 ntdll!RtlDispatchException+0x437 ...
 ntdll!KiUserExceptionDispatch+0x2e() ...
 bcryptPrimitives!AesRNGState_generate+0x79 ...
 bcryptPrimitives!GenRandomAes+0xf70 ...
 bcryptPrimitives!MSCryptGenRandom+0x15f ...
 bcrypt!BCryptGenRandom+0x7a ...
 bcrypt!BCryptGenSysternPreferredRandom+0x340 ....
 bcrypt!BCryptGenRandom+0x119 ...
 git!_ mi_prim_random_buf+0x23() [compat\mimalloc\prim\windows\prim.c @ 648]
 git!mi_random_init_ex+0x61() [compat\mimalloc\random.c @ 179]
 git!_ mi_heap_init+0xe8() [compat\mimalloc\heap.c @ 226]
 git!mi_thread_init+0x21b() [compat\mimalloc\init.c @ 396]
 git!mi_heap_get_default+0x9() [compat\mimalloc\mimalloc\prim.h @ 415]
 git!_ mi_malloc_generic+0x9d() [compat\mimalloc\page.c @ 993]
 git!mingw_getenv+0x4f() [compat\mingw.c @ 2446]
 git!run_post_command_hook+0x84() [git.c @ 515]

Let's treat this code path in the exact same way as the code path in
`mi_heap_main_init()`.

Signed-off-by: Johannes Schindelin <[email protected]>

Author: dscho

compat/mimalloc/heap.c

@@ -223,7 +223,11 @@ void _mi_heap_init(mi_heap_t* heap, mi_tld_t* tld, mi_arena_id_t arena_id, bool
   heap->no_reclaim = noreclaim;
   heap->tag        = tag;
   if (heap == tld->heap_backing) {
+    #if defined(_WIN32) && !defined(MI_SHARED_LIB)
+    _mi_random_init_weak(&heap->random); // prevent allocation failure during bcrypt dll initialization with static linking
+    #else
     _mi_random_init(&heap->random);
+    #endif
   }
   else {
     _mi_random_split(&tld->heap_backing->random, &heap->random);