On Mon 31 Jan 2022 at 21:31, Eric Dumazet [off-list ref] wrote:

On Mon, Jan 31, 2022 at 11:28 AM Vlad Buslov [off-list ref] wrote:

quoted

Yeah, but I also didn't get how the "chain" variable can get reused in
that other function (tcf_new_tfilter()). It seems to always be
unconditionally assigned with return value of tcf_chain_get().

This is why I removed the chain=NULL initialization which is not needed.

The potential issue about replay was really about @q variable.

Got it. Initially it wasn't clear because I got the impression from
commit message that reuse of both variables can lead to use-after-free:
"we need to make sure @q and @chain local variables are cleared again,
or risk use-after-free".