On Fri, 23 Jan 2026 11:47:22 GMT, Yasumasa Suenaga <[email protected]> wrote:
>> src/jdk.hotspot.agent/linux/native/libsaproc/symtab.c line 560:
>>
>>> 558: struct elf_symbol* sym = &(symtab->symbols[n]);
>>> 559: if (sym->name != NULL &&
>>> 560: offset >= sym->offset && offset <= sym->offset + sym->size) {
>>
>> Good to have this issue not affecting the symtab file.
>> Is this change part of reverting the file, or is it needed?
>
> It is required because `sym->size` is zero on my environment
> (glibc-2.42-5.fc43.x86_64).
> Maybe it should be handled as special case.
Right, we have been here before with recognising symbols when PC is off the end
of the defined symbol size in a different issue. I think then we (you!) tried
hard not to simply recognise all symbols as being a byte larger than they
really are.
sym->size for the specific symbol we care about here is zero? Yes I'm seeing 0
size for __restore_rt also.
Comparing offset == sym->offset as an alternative condition, without using
sym->size would work?
-------------
PR Review Comment: https://git.openjdk.org/jdk/pull/29023#discussion_r2720916096
