Re: [PATCH v2 1/3] dt-bindings: PCI: brcmstb: Add two optional props

[Rob Herring]

On Wed, Apr 12, 2023 at 09:12:07AM -0700, Florian Fainelli wrote:
> On 4/12/23 08:37, Rob Herring wrote:
> > On Wed, Apr 12, 2023 at 10:14:46AM -0400, Jim Quinlan wrote:
> > > On Wed, Apr 12, 2023 at 7:56 AM Krzysztof Kozlowski
> > > <krzysztof.kozlowski@xxxxxxxxxx> wrote:
> > > > 
> > > > On 12/04/2023 13:49, Florian Fainelli wrote:
> > > > > 
> > > > > 
> > > > > On 4/12/2023 1:09 AM, Krzysztof Kozlowski wrote:
> > > > > > On 11/04/2023 18:59, Jim Quinlan wrote:
> > > > > > > Regarding "brcm,enable-l1ss":
> > > > > > > 
> > > > > > >     The Broadcom STB/CM PCIe HW -- a core that is also used by RPi SOCs --
> > > > > > >     requires the driver probe() to deliberately place the HW one of three
> > > > > > >     CLKREQ# modes:
> > > > > > > 
> > > > > > >     (a) CLKREQ# driven by the RC unconditionally
> > > > > > >     (b) CLKREQ# driven by the EP for ASPM L0s, L1
> > > > > > >     (c) Bidirectional CLKREQ#, as used for L1 Substates (L1SS).
> > > > > > > 
> > > > > > >     The HW+driver can tell the difference between downstream devices that
> > > > > > >     need (a) and (b), but does not know when to configure (c).  Further, the
> > > > > > >     HW may cause a CPU abort on boot if guesses wrong regarding the need for
> > > > > > >     (c).  So we introduce the boolean "brcm,enable-l1ss" property to indicate
> > > > > > >     that (c) is desired.  Setting this property only makes sense when the
> > > > > > >     downstream device is L1SS-capable and the OS is configured to activate
> > > > > > >     this mode (e.g. policy==superpowersave).
> > > > > > > 
> > > > > > >     This property is already present in the Raspian version of Linux, but the
> > > > > > >     upstream driver implementaion that will follow adds more details and
> > > > > > 
> > > > > > typo, implementation
> > > > > > 
> > > > > > >     discerns between (a) and (b).
> > > > > > > 
> > > > > > > Regarding "brcm,completion-timeout-us"
> > > > > > > 
> > > > > > >     Our HW will cause a CPU abort if the L1SS exit time is longer than the
> > > > > > >     PCIe transaction completion abort timeout.  We've been asked to make this
> > > > > > >     configurable, so we are introducing "brcm,completion-timeout-us".
> > > > > > > 
> > > > > > > Signed-off-by: Jim Quinlan <jim2101024@xxxxxxxxx>
> > > > > > 
> > > > > > What happened here? Where is the changelog?
> > > > > 
> > > > > It is in the cover letter:
> > > > > 
> > > > > https://lore.kernel.org/all/20230411165919.23955-1-jim2101024@xxxxxxxxx/
> > > > > 
> > > > > but it does not look like the cover letter was copied to you or Rob.
> > > > 
> > > > As you said, I did not get it.
> > > 
> > > Yes, sorry about that; I use a wrapper over the "cocci_cc" script and
> > > I need to modify one or both scripts to send the cover to the
> > > superset of recipients in the constituent commits.
> > 
> > Try out 'b4'. It's much easier.
> > 
> > In any case, I don't read cover letters. Changes to a patch belong with
> > the patch.
> 
> This is not what most other maintainers do, and there does not appear to be
> a general consensus amongst maintainers that the changes belong in the
> individual patches, or in the cover letter. 

Well, I stole that phrase from someone else (gregkh). 

> Some trees like the networking
> tree do merge commits of patch sets where the cover letter is used as part
> of the merge commit message. Other maintainers don't, and some want the
> change log after the '---' and some do not.

I'm not aware of anyone except for DRM wanting the changelog in the 
final commits, but that's really a different issue.

I'm pretty sure no one will complain about a changelog in the patches. I 
guess you just have to duplicate it if you think it should be in both. 
b4 could be taught to do that I suppose. IMO, the cover letter should 
have a higher level changelog than the individual patches.

Rob