On Thu, 21 Feb 2019 15:02:24 +0100 Miquel Raynal <miquel.raynal@xxxxxxxxxxx> wrote: > Hi Adam, > > Adam Ford <aford173@xxxxxxxxx> wrote on Thu, 21 Feb 2019 07:48:46 -0600: > > > On Thu, Feb 21, 2019 at 6:58 AM Miquel Raynal <miquel.raynal@xxxxxxxxxxx> wrote: > > > > > > There is no reason to prevent the software BCH ECC engine > > > implementation to be compiled as a module, so change the 'bool' into a > > > 'tristate'. > > > > If you're booting from nand and need the BCH engine, it seems to me > > like you'd need it to be compiled into the kernel because you'd have > > to mount the filesystem before loading the module unless you have a > > initramfs. If you need the BCH engine to mount the filesystem, it > > creates a dependency issue. > > That's true, I could mention this use case in the commit log. But the > action of choosing to compile it as a module is done by the user, so I > guess preventing any user to use it as a module is nonetheless too > restrictive? I agree. ______________________________________________________ Linux MTD discussion mailing list http://lists.infradead.org/mailman/listinfo/linux-mtd/