stdlib: Fix stdbit.h with -Wconversion for clang

stdlib: Fix stdbit.h with -Wconversion for clang

With clang 14 and also with main the tst-stdbit-Wconversion
issues the warnings:

  ../stdlib/stdbit.h:701:40: error: implicit conversion loses integer
  precision: 'int' to 'uint16_t' (aka 'unsigned short')
  [-Werror,-Wimplicit-int-conversion]
    return __x == 0 ? 0 : ((uint16_t) 1) << (__bw16_inline (__x) - 1);
    ~~~~~~                ~~~~~~~~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~~~~~
  ../stdlib/stdbit.h:707:39: error: implicit conversion loses integer
  precision: 'int' to 'uint8_t' (aka 'unsigned char')
  [-Werror,-Wimplicit-int-conversion]
    return __x == 0 ? 0 : ((uint8_t) 1) << (__bw8_inline (__x) - 1);
    ~~~~~~                ~~~~~~~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~~~~
  ../stdlib/stdbit.h:751:40: error: implicit conversion loses integer
  precision: 'int' to 'uint16_t' (aka 'unsigned short')
  [-Werror,-Wimplicit-int-conversion]
    return __x <= 1 ? 1 : ((uint16_t) 2) << (__bw16_inline (__x - 1) - 1);
    ~~~~~~                ~~~~~~~~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
  ../stdlib/stdbit.h:757:39: error: implicit conversion loses integer
  precision: 'int' to 'uint8_t' (aka 'unsigned char')
  [-Werror,-Wimplicit-int-conversion]
    return __x <= 1 ? 1 : ((uint8_t) 2) << (__bw8_inline (__x - 1) - 1);
    ~~~~~~                ~~~~~~~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
  tst-stdbit-Wconversion.c:45:31: error: implicit conversion loses integer
  precision: 'unsigned short' to 'uint8_t' (aka 'unsigned char')
  [-Werror,-Wimplicit-int-conversion]
    (void) stdc_trailing_zeros (us);
           ~~~~~~~~~~~~~~~~~~~~~^~~
  ../stdlib/stdbit.h:164:30: note: expanded from macro
  'stdc_trailing_zeros'
     : stdc_trailing_zeros_uc (x))
       ~~~~~~~~~~~~~~~~~~~~~~~~^~
  ../stdlib/stdbit.h:191:52: note: expanded from macro
  'stdc_trailing_zeros_uc'
  # define stdc_trailing_zeros_uc(x) (__ctz8_inline (x))
                                      ~~~~~~~~~~~~~  ^
  tst-stdbit-Wconversion.c:46:31: error: implicit conversion loses integer
  precision: 'unsigned int' to 'uint16_t' (aka 'unsigned short')
  [-Werror,-Wimplicit-int-conversion]
    (void) stdc_trailing_zeros (ui);
           ~~~~~~~~~~~~~~~~~~~~~^~~
  ../stdlib/stdbit.h:163:48: note: expanded from macro
  'stdc_trailing_zeros'
     : sizeof (x) == 2 ? stdc_trailing_zeros_us (x)       \
                         ~~~~~~~~~~~~~~~~~~~~~~~~^~
  ../stdlib/stdbit.h:192:53: note: expanded from macro
  'stdc_trailing_zeros_us'
  # define stdc_trailing_zeros_us(x) (__ctz16_inline (x))
                                      ~~~~~~~~~~~~~~  ^
  tst-stdbit-Wconversion.c:46:31: error: implicit conversion loses integer
  precision: 'unsigned int' to 'uint8_t' (aka 'unsigned char')
  [-Werror,-Wimplicit-int-conversion]
    (void) stdc_trailing_zeros (ui);
           ~~~~~~~~~~~~~~~~~~~~~^~~
  ../stdlib/stdbit.h:164:30: note: expanded from macro
  'stdc_trailing_zeros'
     : stdc_trailing_zeros_uc (x))
       ~~~~~~~~~~~~~~~~~~~~~~~~^~
  ../stdlib/stdbit.h:191:52: note: expanded from macro
  'stdc_trailing_zeros_uc'
  # define stdc_trailing_zeros_uc(x) (__ctz8_inline (x))
                                      ~~~~~~~~~~~~~  ^
  tst-stdbit-Wconversion.c:47:31: error: implicit conversion loses integer
  precision: 'unsigned long' to 'uint16_t' (aka 'unsigned short')
  [-Werror,-Wimplicit-int-conversion]
    (void) stdc_trailing_zeros (ul);
           ~~~~~~~~~~~~~~~~~~~~~^~~
  ../stdlib/stdbit.h:163:48: note: expanded from macro
  'stdc_trailing_zeros'
     : sizeof (x) == 2 ? stdc_trailing_zeros_us (x)       \
                         ~~~~~~~~~~~~~~~~~~~~~~~~^~
  ../stdlib/stdbit.h:192:53: note: expanded from macro
  'stdc_trailing_zeros_us'
  # define stdc_trailing_zeros_us(x) (__ctz16_inline (x))
                                      ~~~~~~~~~~~~~~  ^
  [...]

It seems to boiler down to __builtin_clz not having a variant for 8 or
16 bits.  Fix it by explicit casting to the expected types.  Although
not strickly required for older gcc, using the same __pacify macro
simpify the required code.

Checked on x86_64-linux-gnu and i686-linux-gnu.