On 29.09.26 11:18, Chao Li wrote:
>> On Sep 29, 2026, at 13:47, Peter Eisentraut <peter@eisentraut.org> wrote:
>>
>> In clang, -fsanitize=undefined includes -fsanitize=function, which reports every call made through a function
pointerwhose type does not exactly match the called function, so it fires all over the place on expression tree walkers
andmutators, as well as a few other places. So -fsanitize=undefined hasn't been working cleanly under clang for a
while. (Before clang 17, it only applied to C++.)
>>
>> This is the same issue that caused us to use -Wno-cast-function-type-strict with clang. That warning applies at the
placewhere the mismatching function pointer is passed, so there are potentially hundreds of sites. Therefore, a global
disablingis appropriate. The sanitizer, on the other hand, triggers where the function is called, which are only about
twodozen places, so it seems possible to silence these checks individually and still main the check for accidental
violationselsewhere.
>>
>> I propose to add pg_attribute_no_sanitize_function() and place it on the functions that make such calls. This is
similarto some existing pg_attribute_no_sanitize_xxx attributes.
>> <0001-Silence-fsanitize-function-where-we-cast-function-po.patch>
>
> Overall looks good to me.
>
> Just one comment, in dynahash.c, hash_search_with_hash_value() is marked with the new annotation, feels like
hash_update_hash_key()also needs to be annotated, because it also invokes match and keycopy etc callbacks.
Yes, I think you are right. My patch was based on what is required to
get the test suite to pass. It appears that there are currently no
callers of hash_update_hash_key() with a string-based hash table, so
this case isn't triggered, but we should add the annotation there as
well for completeness.
