On Wed, Jun 22, 2016 at 09:36:23PM +0200, Maxime Ripard wrote: > On Thu, Jun 16, 2016 at 10:29:39AM -0700, Steve Rae wrote: > > On Wed, Jun 15, 2016 at 1:18 AM, Maxime Ripard > > <[email protected]> wrote: > > > > > > On Tue, Jun 07, 2016 at 11:19:36AM -0700, Steve Rae wrote: > > > > This file originally came from upstream code. > > > > > > > > While retaining the storage abstraction feature, this is the first > > > > set of the changes required to resync with the > > > > cmd_flash_mmc_sparse_img() > > > > in the file > > > > aboot.c > > > > from > > > > > > > > https://us.codeaurora.org/cgit/quic/la/kernel/lk/plain/app/aboot/aboot.c?h=LE.BR.1.2.1 > > > > > > > > Signed-off-by: Steve Rae <[email protected]> > > > > > > Again, please split that in several patches to have one patch > > > per-change you're doing. > > > > > > This is just impossible to review. > > > > And I think you just reinforced the point: > > this code was so far away from the original upstream code that it > > is not even recognizable anymore.... > > I think the only point that was made is that a different bootloader > has a different implementation of the same protocol, just like for any > other protocol. > > An implementation relying on a 120 lines switch statement, and a 250 > lines functions, that hardcodes the backing storage device. > > I'm not sure this is such a good inspiration.
True. But do we want to have a compatible implementation, or match the canonical implementation? Especially since there's many existing 3rd party tools that will test against the upstream version of things. -- Tom
signature.asc
Description: Digital signature
_______________________________________________ U-Boot mailing list [email protected] http://lists.denx.de/mailman/listinfo/u-boot
