On Tue, Jun 18, 2024 at 6:13 PM Christoph Hellwig <hch@xxxxxx> wrote: > > On Tue, Jun 18, 2024 at 06:05:33PM +1200, Barry Song wrote: > > Yes, that was exactly what I missed. I then figured it out, reproduced > > the issue, > > and discovered that the root cause was unrelated to large folios. It > > was actually > > due to a batched bio plugging optimization from 2022. You can find the new patch > > here: > > > > https://lore.kernel.org/linux-mm/20240617220135.43563-1-21cnbao@xxxxxxxxx/ > > I don't really see any point in keeping the VM_BUG_ON. The underlying > direct I/O code doesn't really care about the size at all. I am perfectly fine with dropping the VM_BUG/WARN. I was keeping it for debugging purposes while we add mTHP swapout support for the swapfile. Large folios might be in bio vec, this can help verify things are still normal after having large folios there. But I agree that this is not important for the NFS code itself. I can add this warning locally for debugging when needed.
