Re: [PATCH v2 16/21] ipmi: kcs_bmc: Add a "raw" character device interface

[Andrew Jeffery]

On Tue, 13 Apr 2021, at 17:52, Arnd Bergmann wrote:
> On Tue, Apr 13, 2021 at 1:45 AM Andrew Jeffery <andrew@xxxxxxxx> wrote:
> > On Mon, 12 Apr 2021, at 18:18, Arnd Bergmann wrote:
> > > On Mon, Apr 12, 2021 at 3:33 AM Andrew Jeffery <andrew@xxxxxxxx> wrote:
> > > > On Fri, 9 Apr 2021, at 17:25, Arnd Bergmann wrote:
> > > > > On Fri, Mar 19, 2021 at 7:31 AM Andrew Jeffery <andrew@xxxxxxxx> wrote:
> > > > > >
> > > > > > The existing IPMI chardev encodes IPMI behaviours as the name suggests.
> > > > > > However, KCS devices are useful beyond IPMI (or keyboards), as they
> > > > > > provide a means to generate IRQs and exchange arbitrary data between a
> > > > > > BMC and its host system.
> > > > >
> > > > > I only noticed the series after Joel asked about the DT changes on the arm
> > > > > side. One question though:
> > > > >
> > > > > How does this related to the drivers/input/serio/ framework that also talks
> > > > > to the keyboard controller for things that are not keyboards?
> > > >
> > > > I've taken a brief look and I feel they're somewhat closely related.
> > > >
> > > > It's plausible that we could wrangle the code so the Aspeed and Nuvoton
> > > > KCS drivers move under drivers/input/serio. If you squint, the i8042
> > > > serio device driver has similarities with what the Aspeed and Nuvoton
> > > > device drivers are providing to the KCS IPMI stack.
> > >
> > > After looking some more into it, I finally understood that the two are
> > > rather complementary. While the  drivers/char/ipmi/kcs_bmc.c
> > > is the other (bmc) end of drivers/char/ipmi/ipmi_kcs_sm.c, it seems
> > > that the proposed kcs_bmc_cdev_raw.c interface would be
> > > what corresponds to the other side of
> > > drivers/input/serio/i8042.c+userio.c.
> >
> > Right. I guess the question is should we be splitting kernel subsystems
> > along host/bmc lines? Doesn't feel intuitive, it's all Linux, but maybe
> > we can consolidate in the future if it makes sense?
> 
> We actually have a number of subsystems with somewhat overlapping
> functionality. I brought up serio, because it has an abstraction for multiple
> things that communicate over the keyboard controller and I thought
> the problem you were trying to solve was also related to the keyboard
> controller.
> It is also one of multiple abstractions that allow you to connect a device
> to a uart (along with serdev and tty_ldisc, probably at least one more that
> you can nest above or below these).
> 
> Consolidating the kcs_bmc.c interface into something that already
> exists would obviously be best, but it's not clear which of these that
> should be, that depends on the fundamental properties of the hardware
> interface.
> 
> > > Then again, these are also on
> > > separate ports (0x60 for the keyboard controller, 0xca2 for the BMC
> > > KCS), so they would never actually talk to one another.
> >
> > Well, sort of I guess. On Power systems we don't use the keyboard
> > controller for IPMI or keyboards, so we're just kinda exploiting the
> > hardware for our own purposes.
> 
> Can you describe in an abstract form what the hardware interface
> can do here and what you want from it? I wonder if it could be
> part of a higher-level interface such as drivers/mailbox/ instead.

It gives us interrupts each way between the host and BMC when we send 
some (small amount of) data/metadata. Mailbox is possibly a fit for 
this? We're (ab)using the keyboard controllers to implement a vendor 
MCTP binding over LPC[1] and also a simple protocol for the (Power) 
host to trigger BMC debug data capture in the event of issues with 
other (more complex) in-band communication stacks. The MCTP binding is 
what requires access to STR.

It's feasible that we could implement the debug capture protocol with 
the serio_raw interface now that I think about it (as it only makes use 
of data and not status). What's unclear to me right now is what impact 
that has on the Aspeed/Nuvoton KCS drivers we have in the IPMI 
subsystem. If we can do something sensible to service both serio and 
IPMI with the one driver implementation then I can put together a PoC 
for the debug data stuff using serio_raw.

Regarding the MCTP binding, Jeremy Kerr is working in an in-kernel, 
socket-based implementation of MCTP. Eventually this will allow us to 
bury the KCS details in the MCTP subsystem, which removes some of the 
motivation for the raw interface here.

Andrew

[1] https://github.com/openbmc/libmctp/blob/master/docs/bindings/vendor-ibm-astlpc.md