On Thu, 11 Dec 2025 14:21:42 +0900, Jakub Kicinski wrote:

quoted

quoted

Have you investigated the other callers? Given problems with previous
version of this patch I'm worried you have not. If you did please extend
the commit message with the appropriate explanation.

Are you asking if I investigated other zerocopy tests? NO.

I said callers. You're changing behavior of a function, is it going
to break any of the callers.

I investigated the relevant callers and found no similar restore/revert
operations. Furthermore, if such a restore action existed, the ZC
related test programs would have detected it, just as the tests
detected the revert operation in skb_zerocopy_iter_stream().