[clang] [Thread Analysis] Fix a bug in context update in alias-analysis (PR #178952)

[Marco Elver via cfe-commits]

melver wrote:

I tried to test this on Linux -next (which is enabling -Wthread-safety): https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git/

But when applying this patch I get tons of false positives:
```
  CC      lib/test_context-analysis.o
lib/test_context-analysis.c:84:1: error: releasing raw_spinlock '__UNIQUE_ID_unlock_245' that was not held [-Werror,-Wthread-safety-analysis]
   84 | TEST_SPINLOCK_COMMON(raw_spinlock,
      | ^
lib/test_context-analysis.c:76:5: note: expanded from macro 'TEST_SPINLOCK_COMMON'
   76 |                 { guard(class)(&d->lock);               op(d->counter); }                       \
      |                   ^
./include/linux/cleanup.h:419:2: note: expanded from macro 'guard'
  419 |         CLASS(_name, __UNIQUE_ID(guard))
      |         ^
./include/linux/cleanup.h:300:3: note: expanded from macro 'CLASS'
  300 |                 class_##_name##_constructor
      |                 ^
note: (skipping 4 expansions in backtrace; use -fmacro-backtrace-limit=0 to see all)
././include/linux/compiler_types.h:16:23: note: expanded from macro '__PASTE'
   16 | #define __PASTE(a, b) ___PASTE(a, b)
      |                       ^
././include/linux/compiler_types.h:15:24: note: expanded from macro '___PASTE'
   15 | #define ___PASTE(a, b) a##b
      |                        ^
<scratch space>:29:1: note: expanded from here
   29 | __UNIQUE_ID_unlock_245
      | ^
lib/test_context-analysis.c:84:1: error: acquiring raw_spinlock 'd->lock' that is already held [-Werror,-Wthread-safety-analysis]
lib/test_context-analysis.c:77:5: note: expanded from macro 'TEST_SPINLOCK_COMMON'
   77 |                 { guard(class##_irq)(&d->lock);         op(d->counter); }                       \
      |                   ^
./include/linux/cleanup.h:419:2: note: expanded from macro 'guard'
  419 |         CLASS(_name, __UNIQUE_ID(guard))
      |         ^
./include/linux/cleanup.h:300:3: note: expanded from macro 'CLASS'
  300 |                 class_##_name##_constructor
      |                 ^
```

I haven't been able to isolate a minimal reproducer yet that we can use as a test case.

https://github.com/llvm/llvm-project/pull/178952