[REGRESSION] rseq: refactoring in v6.19 broke everyone on arm64 and tcmalloc everywhere

[Florian Weimer]

* Linus Torvalds:

>>     // Note: this makes __rseq_abi.cpu_id_start unusable for its original purpose.
>>
>>     Do you seriously want to proliferate that?
>
> Absolutely.
>
> That's how clever hacks work - they take advantage of things past
> their design parameters. "If it works, it's not stupid".
>
> We don't then turn around and say "you were clever, and we did
> something stupid, so now we'll hurt you".
>
> This is all 100% on the RSEQ kernel code, not on users who took
> advantage of it.

RSEQ was intended to be modular, with more than one library using it
within a process, without coordination (beyond sticking to the RSEQ
protocol).  The tcmalloc approach is incompatible with that.  Once
tcmalloc starts using RSEQ in its peculiar way, nothing else in the
process can, and vice versa.  This is far from ideal because the
particular descheduling notification that tcmalloc uses could be
implemented in a much simpler way than full RSEQ, given its non-modular
nature.

Thanks,
Florian