Enable -Wshadow=local

[kees]

This would protect the kernel from silent corruptions in overlapping macro variables, for example:

#define something(output, len) \
({ \
    size_t __len = (len); \
    if (__len == 5) \
        do_something(output, __len); \
    else \
         do_other_thing(output, __len); \
    result(output); \
})
...
int __len = 5;
...
something(buffer, __len);

https://godbolt.org/z/P1ox1xWPo

While GCC 4.8 improved global collisions, -Wshadow is still currently unusable:

./arch/x86/include/asm/bitops.h:283:28: warning: declaration of 'ffs' shadows a built-in function [-
Wshadow]                                                                                            
  283 | static __always_inline int ffs(int x)                                                       
      |                            ^~~                                                              

But instead, -Wshadow=local handles the cases that are the most interesting.

This is wider coverage than what might accidentally get caught by -Wunused-but-set-variable, which also has a lot of false positives:
https://lore.kernel.org/r/[email protected]/
https://lore.kernel.org/r/[email protected]/

[kees]

Dealing with the explicit shadowing in the __wait_event() macro family looking extremely difficult to deal with.

[kees]

Potential __wait_event() cleanup has been proposed: https://lore.kernel.org/lkml/[email protected]/