On 2022/9/27 15:27, Christoph Hellwig wrote: > On Wed, Sep 21, 2022 at 09:42:41AM +0800, Liu Shixin wrote: >> This revert makes code complicated, but I think it's necessary. When >> enable zswap, I expect it to work for all swap devices as much as >> possible. > But why would expect something to work on a device that has been > swapped on before? That's not usually how things work. > > . If not do this, while some user enable swap device first and then enable zswap, zswap will not take effect. The user need to re-enable the swap device which is inconvenient. >
