[PATCH v2 0/2] malloc: introduce ifuncs for malloc functions

[Florian Weimer]

* Yury Khrustalev:

>> Even with an implicit tag verification architecture, I don't see how
>> user-tagged chunks in tcache cause problems.  The pointer comparisons in
>> the main allocator are problematic if the tags don't match as expected,
>> but I don't think that exists in tcache.
>> 
>
> I agree that tagging chunks would be the most elegant solution.
> Unfortunately, in practice it causes problems: the existing
> implementation of memory tagging has become broken because changes to
> the internal malloc handling of chunks has not respected tagging that
> had been embedded into the malloc code. Not having tests (and hardware)
> didn't help.

This doesn't matter to tcache because it's the outer allocator.  I'm
pretty sure that this works, after verifying outer allocator behavior
with software tag checking.  We must be careful not to pass tagged
pointers (or use tagged memory) in the inner allocator.

> Also, tagging the entire chunk means that the chunk metadata can be
> accessed via the user pointer, which is not ideal.

Oh, I'm still reserving tag zero as the metadata tag.  The previous user
pointer cannot be used to access the tcache metadata because the tag has
changed.

> Implementing memory tagging as a layer is the only choice for the Glibc
> malloc.

I have come to a similar conclusion, except for tcache.  It's itself
layered on top of the inner allocator, so we can make sure that we feed
only untagged pointers to allocator below.

Thanks,
Florian