Re: [PATCH 0/3] Revert panel fixes and original buggy patch

[Greg Kroah-Hartman]

On Tue, Nov 28, 2023 at 01:52:44AM +0200, Laurent Pinchart wrote:
> On Tue, Nov 28, 2023 at 12:36:15AM +0100, Linus Walleij wrote:
> > On Tue, Nov 28, 2023 at 12:25 AM Laurent Pinchart wrote:
> > > On Tue, Nov 28, 2023 at 12:10:18AM +0100, Linus Walleij wrote:
> > > > This series reverts the attempts to fix the bug that went
> > > > into v6.7-rc1 in commit 199cf07ebd2b
> > > > "drm/bridge: panel: Add a device link between drm device and panel device"
> > > > and then it reverts that patch as well.
> > > >
> > > > Signed-off-by: Linus Walleij <linus.walleij@xxxxxxxxxx>
> > > > ---
> > > > Linus Walleij (3):
> > > >       Revert "driver core: Export device_is_dependent() to modules"
> > > >       Revert "drm/bridge: panel: Check device dependency before managing device link"
> > > >       Revert "drm/bridge: panel: Add a device link between drm device and panel device"
> > >
> > > To preserve bisectability, you should revert in the opposite order.
> > 
> > You mean apply patch 2, then 1, then 3 so the kernel builds after each
> > revert?
> > 
> > Yeah that's a good idea, I don't know if I should apply these though, better
> > someone else do it since I screwed up too much.
> > 
> > Another option is to just squash the reverts into one, that bisects too :/
> 
> I thought the commits have been applied to drm-misc in a bisectable
> order in the first place, but that doesn't seem to be the case :-(
> Reverting "driver core: Export device_is_dependent() to modules" last
> seems to be the best option in this case. I wouldn't squash them.

Agreed, don't squash, just revert in the opposite order they were
applied in originally, that way the tree can always build.

thanks,

greg k-h