On Tue, Aug 06, 2024 at 01:07:47PM +0300, Tariq Toukan wrote: > Adding Maxim Mikityanskiy, he might have some insights. I think the important part to find out is if the in-kernel TLS API has a limitation to PAGE_SIZE buffer segments, or not. If it does we need to disable large folios for the in-kernel consumers of the API, which is easy (but annoying) for NFS that trigger this, but quite painful for NVMe. Or the API is supposed to handle it and just this one offload gets it wrong, in which case we need to fix it.
