https://gcc.gnu.org/bugzilla/show_bug.cgi?id=120002
Thomas Weißschuh <g...@t-8ch.de> changed:
What |Removed |Added
----------------------------------------------------------------------------
Resolution|INVALID |FIXED
--- Comment #7 from Thomas Weißschuh <g...@t-8ch.de> ---
(In reply to Andrew Pinski from comment #5)
> But the way GCC is doing it is valid for normal DSOs but in this case vDSO
> is not a normal DSO and needs hacks like this in general to get around vDSO
> limitations.
Ok, that is unfortunate.
I also added a compile-time check to reject absolute relocations:
https://lore.kernel.org/lkml/20250429-vdso-absolute-reloc-v1-2-987a0afd1...@linutronix.de/
Do you think this is sufficient?
(In reply to Andrew Pinski from comment #6)
> It does, arm 32bit is affected which is why the compat vdso does that :).
The comment for the compat vdso does not talk about correctness, only
performance.
If it was a problem for 32bit arm before I would have expected it to have been
a problem for 64bit at the same time. But so far it wasn't an issue there.
