[PATCH v4 1/5] dt-bindings: PCI: brcmstb: brcm,{enable-l1ss,completion-timeout-us} props

[Bjorn Helgaas]

On Fri, May 05, 2023 at 08:39:52AM -0400, Jim Quinlan wrote:
> On Wed, May 3, 2023 at 6:18 PM Bjorn Helgaas <helgaas at kernel.org> wrote:
> > On Wed, May 03, 2023 at 05:38:15PM -0400, Jim Quinlan wrote:
> > > On Wed, May 3, 2023 at 2:07 PM Bjorn Helgaas <helgaas at kernel.org> wrote:
> > > > On Wed, May 03, 2023 at 10:38:57AM -0400, Jim Quinlan wrote:
> > > > > On Sun, Apr 30, 2023 at 3:10 PM Bjorn Helgaas <helgaas at kernel.org> wrote:
> > > > > > On Fri, Apr 28, 2023 at 06:34:55PM -0400, Jim Quinlan wrote:
> > > > > > > brcm,enable-l1ss (bool):
> > > > > > >
> > > > > > >   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).  All devices
> > > > > > >   should work fine when the driver chooses (a) or (b), but (c) may be
> > > > > > >   desired to realize the extra power savings that L1SS offers.  So we
> > > > > > >   introduce the boolean "brcm,enable-l1ss" property to inform the driver
> > > > > > >   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).

Just noticed that this should be "policy==powersupersave"

> > > > > > What bad things would happen if the driver always configured (c)?
> > > > >
> > > > > Well, our driver has traditionally only supported (b) and our
> > > > > existing boards have been designed with this in mind.  I would not
> > > > > want to switch modes w'o the user/customer/engineer opting-in to do
> > > > > so.  Further, the PCIe HW engineer told me defaulting to (c) was a
> > > > > bad idea and was "asking for trouble".  Note that the commit's
> > > > > comment has that warning about L1SS mode not meeting this 400ns
> > > > > spec, and I suspect that many of our existing designs have bumped
> > > > > into that.
> > > > >
> > > > > But to answer your question, I haven't found a scenario that did not
> > > > > work by setting mode (c).  That doesn't mean they are not out there.
> > > > >
> > > > > > Other platforms don't require this, and having to edit the DT
> > > > > > based on what PCIe device is plugged in seems wrong.  If brcmstb
> > > > > > does need it, that suggests a hardware defect.  If we need this to
> > > > > > work around a defect, that's OK, but we should acknowledge the
> > > > > > defect so we can stop using this for future hardware that doesn't
> > > > > > need it.
> > > > >
> > > > > All devices should work w/o the user having to change the DT.  Only
> > > > > if they desire L1SS must they add the "brcm,enable-l1ss" property.
> > > >
> > > > I thought the DT was supposed to describe properties of the
> > > > *hardware*, but this seems more like "use this untested clkreq
> > > > configuration," which maybe could be done via a module parameter?
> > >
> > > Electrically, it has been tested, but  specifically for L1SS capable
> > > devices.  What is untested AFAICT are platforms using this mode on
> > > non-L1SS capable devices.
> >
> > Non-L1SS behavior is a subset of L1SS, so if you've tested with L1SS
> > enabled, I would think you'd be covered.

I think this point is still worth considering.  Maybe your hardware
folks have an opinion here?

> > But I'm not a hardware engineer, so maybe there's some subtlety there.
> > The "asking for trouble" comment from your engineer is definitely
> > concerning, but I have no idea what's behind that.
> >
> > And obviously even if we have "brcm,enable-l1ss", the user may decide
> > to disable L1SS administratively, so even if the Root Port and the
> > device both support L1SS, it may be never be enabled.
> >
> > > WRT bootline param
> > > pci=[<domain>:]<bus>:<dev>.<func>[/<dev>.<func>]*pci:<vendor>:<device>[:<subvendor>:<subdevice>]:
> > > this does not look compatible for vendor specific DT options like
> > > "brcm,enable-l1ss".  I observe that pci_dev_str_match_path() is a
> > > static function and I don't see a single option in pci.c  that is
> > > vendor specific.  FWIW, moving something like this to the bootline
> > > would not be popular with our customers; for some reason they really
> > > don't like changes to the bootline.
> >
> > They prefer editing the DT?
> >
> > I agree the "pci=B:D.F" stuff is a bit ugly.  Do you have multiple
> > slots such that you would have to apply this parameter to some but not
> > others?  I guess I was imagining a single-slot system where you
> > wouldn't need to identify the specific device because there *is* only
> > one.
> 
> We typically have a single device per controller.  Occasionally,
> there is a mismatch in needs, and the customer adds a switch to
> their board until we can add another controller to the next rev of
> the SOC.

If you add a switch, it sounds like there's still only a single link
between the brcmstb controller and the switch.  I'm assuming
"brcm,enable-l1ss" only affects CLKREQ# on that link and it has
nothing to do with links below the switch.

(c) must be the standard PCIe situation because no other systems
require the user to configure CLKREQ# based on the type of card
plugged in.  And we don't know about any actual problems that happen
in (c) with any cards.

That makes me think the ideal end state would be to use (c) by
default so everything just works like every other platform with no
fuss.  If there's some situation that requires (a) or (b), there could
be a property or parameter to select *that* because that would be the
unusual case.

But obviously the comment from the hardware engineer:

> > > > > Further, the PCIe HW engineer told me defaulting to (c) was
> > > > > a bad idea and was "asking for trouble".

would need to be understood before doing that.

Bjorn