Re: [PATCH bpf-next] bpf,x86: do RSB balance for trampoline

[Peter Zijlstra]

On Wed, Nov 05, 2025 at 03:46:47PM +0800, Menglong Dong wrote:

> The "return POS" will miss the RSB, but the later return
> will hit it.

Right this. The moment you have an asymmetry in the RSB all future RETs
will miss and that sucks. Missing one but then hitting all the others is
much better.

Anyway, I see this thread came to a good conclusion. Just one more note,
people are working on supervisor shadow stack support, which is going to
make all these things even more 'fun' :-)