Re: backport of patches 8238b4579866b7c1bb99883cfe102a43db5506ff and d6ffe6067a54972564552ea45d320fb98db1ac5e

[Greg KH <gregkh@linuxfoundation.org>]

On Thu, Oct 27, 2022 at 07:45:52AM -0400, Mikulas Patocka wrote:
> 
> 
> On Wed, 26 Oct 2022, Greg KH wrote:
> 
> > On Tue, Oct 18, 2022 at 07:36:22AM -0400, Mikulas Patocka wrote:
> > > 
> > > 
> > > On Tue, 11 Oct 2022, Greg KH wrote:
> > > 
> > > > On Tue, Oct 11, 2022 at 05:48:26AM -0400, Mikulas Patocka wrote:
> > > > > 
> > > > > 
> > > > > On Mon, 10 Oct 2022, Greg KH wrote:
> > > > > 
> > > > > > Nope, these cause loads of breakages.  See
> > > > > > https://lore.kernel.org/r/09eca44e-4d91-a060-d48c-d0aa41ac5045@roeck-us.net
> > > > > > for one such example, and I know kbuild sent you other build problems.
> > > > > > I'll drop all of these from the stable trees now.  Please feel free to
> > > > > > resend them when you have the build issues worked out.
> > > > > > 
> > > > > > thanks,
> > > > > > 
> > > > > > greg k-h
> > > > > 
> > > > > I don't have cross compilers for all the architectures that Linux 
> > > > > supports. Is there some way how to have the patch compile-tested before I 
> > > > > send it to you?
> > > > 
> > > > You can download those compilers from kernel.org, they are all available
> > > > there.
> > > 
> > > OK. I downloaded cross compilers from 
> > > https://mirrors.edge.kernel.org/pub/tools/crosstool/ and compile-tested 
> > > the patches with all possible architectures.
> > > 
> > > Here I'm sending new versions.
> > 
> > But don't you need 2 patches, not just 1, to be applied?
> 
> Just one patch is sufficient.
> 
> The upstream patch 8238b4579866b7c1bb99883cfe102a43db5506ff fixes a bug 
> and the patch d6ffe6067a54972564552ea45d320fb98db1ac5e fixes compile 
> failures triggered by 8238b4579866b7c1bb99883cfe102a43db5506ff on some 
> architectures.
> 
> For simplicity of making and testing the stable branch patches I folded 
> these changes into just one patch - that fixes the bug and fixes compile 
> failures as well.

No, please do not do that.  We want both commits at once, not a "fixed
up" change, right?  Otherwise our tools will want to apply the second
one as it is insisting that a fix is still needed.

> > Please resend a set of series, one series per stable kernel branch, to
> > make it more obvious what to do.  Your thread here is very confusing.
> 
> I'll resend it, but except for the subject line I don't know what have I 
> done wrong.

Subject line is everything :)

As is the text in the body, I would have had to remove that from your
last one.  See the examples on the list for how to make this easy for us
to apply.

thanks,

greg k-h