Thanks for the clarification, and sorry for misunderstanding your
intention earlier.
I’ve updated the implementation to use malloc/aligned_alloc() directly
in running_alloc()
as you suggested. Please check whether the new version matches your
expectation.
On 12/9/2025 6:59 PM, Konstantin Ananyev wrote:
Sorry, I didn't express myself clear enough.
My thought was to use malloc/aligned_alloc() directly in running_alloc()
function.
To be more specific here:
static void *running_alloc(char *name, size_t size, size_t align, int32_t
socket_id, void *udata)
{
...
/* just for the API correctness check */
If (udata != <expected_value>) {
/* report error */
return NULL;
}
addr = aligned_alloc(align, size);
if (addr == NULL) {
/* report error*/
}
return addr;
}
