diff options
| author | Ivan Maidanski <ivmai@mail.ru> | 2020-03-31 09:45:21 +0300 |
|---|---|---|
| committer | Ivan Maidanski <ivmai@mail.ru> | 2021-09-09 08:32:46 +0300 |
| commit | ebaad4a88c9dc3e6286288c12bc525838e316e85 (patch) | |
| tree | 72a9e4f7a57cf3c7248b99757e037319a3665540 | |
| parent | 64e514cfe74e6f369fbc0f4a30a686656a8c3a60 (diff) | |
| download | libatomic_ops-ebaad4a88c9dc3e6286288c12bc525838e316e85.tar.gz | |
Remove comments that gcc/powerpc.h and gcc/s390.h are untested
* src/atomic_ops/sysdeps/gcc/powerpc.h [AO_DISABLE_GCC_ATOMICS
&& !AO_PREFER_GENERALIZED] (AO_test_and_set): Remove comment that the
function is untested.
* src/atomic_ops/sysdeps/gcc/s390.h: Remove comment that the all
functions are untested; replace FIXME with TODO for
test_and_set_t_is_ao_t.h.
| -rw-r--r-- | src/atomic_ops/sysdeps/gcc/powerpc.h | 2 | ||||
| -rw-r--r-- | src/atomic_ops/sysdeps/gcc/s390.h | 3 |
2 files changed, 2 insertions, 3 deletions
diff --git a/src/atomic_ops/sysdeps/gcc/powerpc.h b/src/atomic_ops/sysdeps/gcc/powerpc.h index eb594a0..e70b76d 100644 --- a/src/atomic_ops/sysdeps/gcc/powerpc.h +++ b/src/atomic_ops/sysdeps/gcc/powerpc.h @@ -138,7 +138,7 @@ AO_store_release(volatile AO_t *addr, AO_t value) /* only cost us a load immediate instruction. */ AO_INLINE AO_TS_VAL_t AO_test_and_set(volatile AO_TS_t *addr) { -/* Completely untested. And we should be using smaller objects anyway. */ + /* TODO: And we should be using smaller objects anyway. */ AO_t oldval; AO_t temp = 1; /* locked value */ diff --git a/src/atomic_ops/sysdeps/gcc/s390.h b/src/atomic_ops/sysdeps/gcc/s390.h index fe41898..91bec4e 100644 --- a/src/atomic_ops/sysdeps/gcc/s390.h +++ b/src/atomic_ops/sysdeps/gcc/s390.h @@ -15,7 +15,6 @@ * */ -/* FIXME: untested. */ /* The relevant documentation appears to be at */ /* http://publibz.boulder.ibm.com/epubs/pdf/dz9zr003.pdf */ /* around page 5-96. Apparently: */ @@ -35,7 +34,7 @@ #include "../ordered_except_wr.h" #include "../test_and_set_t_is_ao_t.h" -/* FIXME: Is there a way to do byte-sized test-and-set? */ +/* TODO: Is there a way to do byte-sized test-and-set? */ /* TODO: AO_nop_full should probably be implemented directly. */ /* It appears that certain BCR instructions have that effect. */ |