Hi Heikki,
Good to see things are happening with Type-C!
On 2017-02-21 15:24, Heikki Krogerus wrote:
> ...
> +When connected, the partner will be presented also as its own device under
> +/sys/class/typec/. The parent of the partner device will always be the port it
> +is attached to. The partner attached to port "port0" will be named
> +"port0-partner". Full path to the device would be
> +/sys/class/typec/port0/port0-partner/.
A "/port0" too much?
> +
> +The cable and the two plugs on it may also be optionally presented as their own
> +devices under /sys/class/typec/. The cable attached to the port "port0" port
> +will be named port0-cable and the plug on the SOP Prime end (see USB Power
> +Delivery Specification ch. 2.4) will be named "port0-plug0" and on the SOP
> +Double Prime end "port0-plug1". The parent of a cable will always be the port,
> +and the parent of the cable plugs will always be the cable.
> +
> +If the port, partner or cable plug support Alternate Modes, every supported
> +Alternate Mode SVID will have their own device describing them. The Alternate
> +Modes will not be attached to the typec class. The parent of an alternate mode
> +will be the device that supports it, so for example an alternate mode of
> +port0-partner will bees presented under /sys/class/typec/port0-partner/. Every
bees?
> +mode that is supported will have its own group under the Alternate Mode device
> +named "mode<index>", for example /sys/class/typec/port0/<alternate mode>/mode1/.
> +The requests for entering/exiting a mode can be done with "active" attribute
> +file in that group.
> +
> ...
I'm hoping to find time to upgrade the kernel and try these patches in my system.
Looking forward, one thing I have run into is how to connect the typec driver with a
driver for an alternate mode. E.g. the DisplayPort Alternate Mode specification
includes the HPD (hot plug) and HPD-INT (hot plug interrupt) signals as bits in the
Attention message. These signals are needed by the DisplayPort driver to know when to
start negotiation etc.
Have you got any thoughts on how to standardize such interfaces?
BR // Mats
On 03/02/2017 07:22 AM, Mats Karrman wrote:
> Hi Heikki,
>
> Good to see things are happening with Type-C!
>
> On 2017-02-21 15:24, Heikki Krogerus wrote:
>
>> ...
>> +When connected, the partner will be presented also as its own device under
>> +/sys/class/typec/. The parent of the partner device will always be the port it
>> +is attached to. The partner attached to port "port0" will be named
>> +"port0-partner". Full path to the device would be
>> +/sys/class/typec/port0/port0-partner/.
>
> A "/port0" too much?
>
>> +
>> +The cable and the two plugs on it may also be optionally presented as their own
>> +devices under /sys/class/typec/. The cable attached to the port "port0" port
>> +will be named port0-cable and the plug on the SOP Prime end (see USB Power
>> +Delivery Specification ch. 2.4) will be named "port0-plug0" and on the SOP
>> +Double Prime end "port0-plug1". The parent of a cable will always be the port,
>> +and the parent of the cable plugs will always be the cable.
>> +
>> +If the port, partner or cable plug support Alternate Modes, every supported
>> +Alternate Mode SVID will have their own device describing them. The Alternate
>> +Modes will not be attached to the typec class. The parent of an alternate mode
>> +will be the device that supports it, so for example an alternate mode of
>> +port0-partner will bees presented under /sys/class/typec/port0-partner/. Every
>
> bees?
>
>> +mode that is supported will have its own group under the Alternate Mode device
>> +named "mode<index>", for example /sys/class/typec/port0/<alternate mode>/mode1/.
>> +The requests for entering/exiting a mode can be done with "active" attribute
>> +file in that group.
>> +
>> ...
>
> I'm hoping to find time to upgrade the kernel and try these patches in my system.
>
> Looking forward, one thing I have run into is how to connect the typec driver with a
> driver for an alternate mode. E.g. the DisplayPort Alternate Mode specification
> includes the HPD (hot plug) and HPD-INT (hot plug interrupt) signals as bits in the
> Attention message. These signals are needed by the DisplayPort driver to know when to
> start negotiation etc.
> Have you got any thoughts on how to standardize such interfaces?
>
That really depends on the lower level driver. For Chromebooks, where the Type-C
Protocol Manager runs on the EC, we have an extcon driver which reports the pin states
to the graphics drivers and connects to the Type-C class code using the Type-C class
API. I still need to update, re-test, and publish that code. The published code in
https://chromium.googlesource.com/chromiumos/third_party/kernel/, branch chromeos-4.4,
shows how it can be done, though that code currently still uses the Android Type-C
infrastructure.
Guenter
On 2017-03-03 04:13, Guenter Roeck wrote:
> On 03/02/2017 07:22 AM, Mats Karrman wrote:
>> ....
>> Looking forward, one thing I have run into is how to connect the typec driver with a
>> driver for an alternate mode. E.g. the DisplayPort Alternate Mode specification
>> includes the HPD (hot plug) and HPD-INT (hot plug interrupt) signals as bits in the
>> Attention message. These signals are needed by the DisplayPort driver to know when to
>> start negotiation etc.
>> Have you got any thoughts on how to standardize such interfaces?
> That really depends on the lower level driver. For Chromebooks, where the Type-C
> Protocol Manager runs on the EC, we have an extcon driver which reports the pin states
> to the graphics drivers and connects to the Type-C class code using the Type-C class
> API. I still need to update, re-test, and publish that code. The published code in
> https://chromium.googlesource.com/chromiumos/third_party/kernel/, branch chromeos-4.4,
> shows how it can be done, though that code currently still uses the Android Type-C
> infrastructure.
OK, thanks!
My system is a bit different. It's an i.MX6 SoC with the typec phy and DP controller connected
directly to the SoC and it's using DTB/OF.
Using extcon I would have a driver that is both typec class and extcon driver at the same time
since I can't share the access to the typec phy. Is this done elsewhere in the kernel?
I don't know much about the wcove PMIC and what alternate modes it might support but I
guess that driver would end up in the same place.
Do we need to further standardize attributes under (each) specific alternate mode to
include things such as HPD for the DP mode?
BR // Mats
2017-03-03 8:29 GMT+01:00 Mats Karrman <[email protected]>:
> On 2017-03-03 04:13, Guenter Roeck wrote:
>
>> On 03/02/2017 07:22 AM, Mats Karrman wrote:
>>>
>>> ....
>>> Looking forward, one thing I have run into is how to connect the typec
>>> driver with a
>>> driver for an alternate mode. E.g. the DisplayPort Alternate Mode
>>> specification
>>> includes the HPD (hot plug) and HPD-INT (hot plug interrupt) signals as
>>> bits in the
>>> Attention message. These signals are needed by the DisplayPort driver to
>>> know when to
>>> start negotiation etc.
>>> Have you got any thoughts on how to standardize such interfaces?
>>
>> That really depends on the lower level driver. For Chromebooks, where the
>> Type-C
>> Protocol Manager runs on the EC, we have an extcon driver which reports
>> the pin states
>> to the graphics drivers and connects to the Type-C class code using the
>> Type-C class
>> API. I still need to update, re-test, and publish that code. The published
>> code in
>> https://chromium.googlesource.com/chromiumos/third_party/kernel/, branch
>> chromeos-4.4,
>> shows how it can be done, though that code currently still uses the
>> Android Type-C
>> infrastructure.
>
On a side note, the cros-ec extcon part for DP is currently discussed
here [1]. This driver works together with the cdn-dp driver that just
landed in mainline. So comments and feedback are welcome
[1] https://lkml.org/lkml/2017/3/1/301
>
> OK, thanks!
>
> My system is a bit different. It's an i.MX6 SoC with the typec phy and DP
> controller connected
>
> directly to the SoC and it's using DTB/OF.
>
> Using extcon I would have a driver that is both typec class and extcon
> driver at the same time
>
> since I can't share the access to the typec phy. Is this done elsewhere in
> the kernel?
>
> I don't know much about the wcove PMIC and what alternate modes it might
> support but I
>
> guess that driver would end up in the same place.
>
> Do we need to further standardize attributes under (each) specific alternate
> mode to
>
> include things such as HPD for the DP mode?
>
> BR // Mats
>
Hi,
On Fri, Mar 03, 2017 at 08:29:18AM +0100, Mats Karrman wrote:
> On 2017-03-03 04:13, Guenter Roeck wrote:
>
> > On 03/02/2017 07:22 AM, Mats Karrman wrote:
> > > ....
> > > Looking forward, one thing I have run into is how to connect the typec driver with a
> > > driver for an alternate mode. E.g. the DisplayPort Alternate Mode specification
> > > includes the HPD (hot plug) and HPD-INT (hot plug interrupt) signals as bits in the
> > > Attention message. These signals are needed by the DisplayPort driver to know when to
> > > start negotiation etc.
> > > Have you got any thoughts on how to standardize such interfaces?
My idea was to have something like the altmode "bus" at one point.
We create a device for every alternate mode registered in typec class,
so the alternate modes registered for the ports and partners would
simply be attached to the altmode bus. There would be a bus per port
of course.
The drivers for the port alternate modes would take care of things
like muxing and other platform specific stuff as needed, and they
would be tied to the underlying subsystems and drivers, graphics in
case of DisplayPort. The drivers for the partner alternate modes would
take care of the actual communication with the alternate mode with
VDMs if needed (but not necessarily), and they would need to be tied
to the port alternate modes. In practice the driver for both the port
and the partner alternate modes will be the same (in the same
location) obviously, at least in most cases.
I think a bus would allow us to support several ways of handling the
alternate modes on different platforms. It would work fine also on
platforms that had no use for it of course, like platforms where
firmware or EC takes care of most things related to Type-C.
But please note that since this is just a high level idea still, we
wouldn't for example need to create an actual bus if there is no use
for it, but since we have the SVIDs that can be used for matching,
then why not try take advantage of them, right.
How would something like that sound to you guys?
> > That really depends on the lower level driver. For Chromebooks, where the Type-C
> > Protocol Manager runs on the EC, we have an extcon driver which reports the pin states
> > to the graphics drivers and connects to the Type-C class code using the Type-C class
> > API. I still need to update, re-test, and publish that code. The published code in
> > https://chromium.googlesource.com/chromiumos/third_party/kernel/, branch chromeos-4.4,
> > shows how it can be done, though that code currently still uses the Android Type-C
> > infrastructure.
In this case I think you would only need to register a driver with the
bus in case you want the handle to the device for the alternate mode.
> OK, thanks!
>
> My system is a bit different. It's an i.MX6 SoC with the typec phy and DP controller connected
> directly to the SoC and it's using DTB/OF.
Is this "DP controller" a controller that is capable of taking care of
the USB Power Delivery communication with the partner regarding
DisplayPort alternate mode?
> Using extcon I would have a driver that is both typec class and extcon driver at the same time
> since I can't share the access to the typec phy. Is this done elsewhere in the kernel?
> I don't know much about the wcove PMIC and what alternate modes it might support but I
> guess that driver would end up in the same place.
What alternate modes systems with WhiskeyCove supports depends on
the platform. WhiskeyCove PMIC (as in Power Management IC) is
available on a few Intel Atom platforms. The USB Type-C PHY in it
provides a simple USB PD transceiver that does not touch the actual
communication with the partners. The communication needs to be done in
software, including dealing with alternate modes.
I'm not planning on using extcon for anything with WhiskeyCove. I
don't have any use for it. It looks to me that extcon is used just as
a tool to create software couplings in many cases, and I'm not
completely comfortable with that.
In my case with DP altmode, if we had for example muxes to take care
of, I don't think it would be a problem to tie the driver for the mux
to the graphics directly, so basically make it part of the graphics
stack. That driver would be the port (and partner) altmode driver.
> Do we need to further standardize attributes under (each) specific alternate mode to
> include things such as HPD for the DP mode?
I'm not completely sure what kind of system you have, but I would
imagine that if we had the bus, your DP controller driver would be the
port (and partner) alternate mode driver. The bus would bind you to
the typec phy.
Thanks,
--
heikki
On 03/02/2017 11:29 PM, Mats Karrman wrote:
> On 2017-03-03 04:13, Guenter Roeck wrote:
>
>> On 03/02/2017 07:22 AM, Mats Karrman wrote:
>>> ....
>>> Looking forward, one thing I have run into is how to connect the typec driver with a
>>> driver for an alternate mode. E.g. the DisplayPort Alternate Mode specification
>>> includes the HPD (hot plug) and HPD-INT (hot plug interrupt) signals as bits in the
>>> Attention message. These signals are needed by the DisplayPort driver to know when to
>>> start negotiation etc.
>>> Have you got any thoughts on how to standardize such interfaces?
>> That really depends on the lower level driver. For Chromebooks, where the Type-C
>> Protocol Manager runs on the EC, we have an extcon driver which reports the pin states
>> to the graphics drivers and connects to the Type-C class code using the Type-C class
>> API. I still need to update, re-test, and publish that code. The published code in
>> https://chromium.googlesource.com/chromiumos/third_party/kernel/, branch chromeos-4.4,
>> shows how it can be done, though that code currently still uses the Android Type-C
>> infrastructure.
>
> OK, thanks!
>
> My system is a bit different. It's an i.MX6 SoC with the typec phy and DP controller connected
>
> directly to the SoC and it's using DTB/OF.
>
Sounds quite similar to rk3399 (Samsung Chromebook Plus). I have not ported the Type-C class code
to that system yet, but it does use the extcon cros_ec driver, and DP connectivity is signaled
to the DP code through extcon events.
> Using extcon I would have a driver that is both typec class and extcon driver at the same time
>
> since I can't share the access to the typec phy. Is this done elsewhere in the kernel?
>
See above. The code is available in the repository mentioned above.
Guenter
> I don't know much about the wcove PMIC and what alternate modes it might support but I
>
> guess that driver would end up in the same place.
>
> Do we need to further standardize attributes under (each) specific alternate mode to
>
> include things such as HPD for the DP mode?
>
> BR // Mats
>
>
On 03/03/2017 04:59 AM, Heikki Krogerus wrote:
> Hi,
>
> On Fri, Mar 03, 2017 at 08:29:18AM +0100, Mats Karrman wrote:
>> On 2017-03-03 04:13, Guenter Roeck wrote:
>>
>>> On 03/02/2017 07:22 AM, Mats Karrman wrote:
>>>> ....
>>>> Looking forward, one thing I have run into is how to connect the typec driver with a
>>>> driver for an alternate mode. E.g. the DisplayPort Alternate Mode specification
>>>> includes the HPD (hot plug) and HPD-INT (hot plug interrupt) signals as bits in the
>>>> Attention message. These signals are needed by the DisplayPort driver to know when to
>>>> start negotiation etc.
>>>> Have you got any thoughts on how to standardize such interfaces?
>
> My idea was to have something like the altmode "bus" at one point.
> We create a device for every alternate mode registered in typec class,
> so the alternate modes registered for the ports and partners would
> simply be attached to the altmode bus. There would be a bus per port
> of course.
>
> The drivers for the port alternate modes would take care of things
> like muxing and other platform specific stuff as needed, and they
> would be tied to the underlying subsystems and drivers, graphics in
> case of DisplayPort. The drivers for the partner alternate modes would
> take care of the actual communication with the alternate mode with
> VDMs if needed (but not necessarily), and they would need to be tied
> to the port alternate modes. In practice the driver for both the port
> and the partner alternate modes will be the same (in the same
> location) obviously, at least in most cases.
>
> I think a bus would allow us to support several ways of handling the
> alternate modes on different platforms. It would work fine also on
> platforms that had no use for it of course, like platforms where
> firmware or EC takes care of most things related to Type-C.
>
> But please note that since this is just a high level idea still, we
> wouldn't for example need to create an actual bus if there is no use
> for it, but since we have the SVIDs that can be used for matching,
> then why not try take advantage of them, right.
>
> How would something like that sound to you guys?
>
>>> That really depends on the lower level driver. For Chromebooks, where the Type-C
>>> Protocol Manager runs on the EC, we have an extcon driver which reports the pin states
>>> to the graphics drivers and connects to the Type-C class code using the Type-C class
>>> API. I still need to update, re-test, and publish that code. The published code in
>>> https://chromium.googlesource.com/chromiumos/third_party/kernel/, branch chromeos-4.4,
>>> shows how it can be done, though that code currently still uses the Android Type-C
>>> infrastructure.
>
> In this case I think you would only need to register a driver with the
> bus in case you want the handle to the device for the alternate mode.
>
>> OK, thanks!
>>
>> My system is a bit different. It's an i.MX6 SoC with the typec phy and DP controller connected
>> directly to the SoC and it's using DTB/OF.
>
> Is this "DP controller" a controller that is capable of taking care of
> the USB Power Delivery communication with the partner regarding
> DisplayPort alternate mode?
>
>> Using extcon I would have a driver that is both typec class and extcon driver at the same time
>> since I can't share the access to the typec phy. Is this done elsewhere in the kernel?
>> I don't know much about the wcove PMIC and what alternate modes it might support but I
>> guess that driver would end up in the same place.
>
> What alternate modes systems with WhiskeyCove supports depends on
> the platform. WhiskeyCove PMIC (as in Power Management IC) is
> available on a few Intel Atom platforms. The USB Type-C PHY in it
> provides a simple USB PD transceiver that does not touch the actual
> communication with the partners. The communication needs to be done in
> software, including dealing with alternate modes.
>
> I'm not planning on using extcon for anything with WhiskeyCove. I
> don't have any use for it. It looks to me that extcon is used just as
> a tool to create software couplings in many cases, and I'm not
> completely comfortable with that.
>
For rk3399 we went back and forth trying other approaches, but using extcon
turned out to be the most straightforward mechanism; everything else would
just have replicated what extcon already provides. Having said that,
I am open to other approaches. The rk3399 implementation may be seen as an
example on what is required, specifically when it comes to the interaction
between cable, DP, and the Type-C phy.
Guenter
> In my case with DP altmode, if we had for example muxes to take care
> of, I don't think it would be a problem to tie the driver for the mux
> to the graphics directly, so basically make it part of the graphics
> stack. That driver would be the port (and partner) altmode driver.
>
>> Do we need to further standardize attributes under (each) specific alternate mode to
>> include things such as HPD for the DP mode?
>
> I'm not completely sure what kind of system you have, but I would
> imagine that if we had the bus, your DP controller driver would be the
> port (and partner) alternate mode driver. The bus would bind you to
> the typec phy.
>
>
> Thanks,
>
On 2017-03-03 13:59, Heikki Krogerus wrote:
> On Fri, Mar 03, 2017 at 08:29:18AM +0100, Mats Karrman wrote:
> ....
> How would something like that sound to you guys?
Complicated... Need to marinate on that for a while ;)
>> My system is a bit different. It's an i.MX6 SoC with the typec phy and DP controller connected
>> directly to the SoC and it's using DTB/OF.
> Is this "DP controller" a controller that is capable of taking care of
> the USB Power Delivery communication with the partner regarding
> DisplayPort alternate mode?
No, the "DP controller" just talks DP and knows nothing about Type-C or USB PD.
It takes a video stream from the SoC and turns it into a DP link, set up and orchestrated
by the corresponding driver. And all the driver needs from Type-C is the plugged in / interrupt /
plugged out events.
The analog switching between USB / safe / DP signal levels in the Type-C connector is, I think,
best handled by the software doing the USB PD negotiation / Altmode handling (using some GPIOs).
>> Do we need to further standardize attributes under (each) specific alternate mode to
>> include things such as HPD for the DP mode?
> I'm not completely sure what kind of system you have, but I would
> imagine that if we had the bus, your DP controller driver would be the
> port (and partner) alternate mode driver. The bus would bind you to
> the typec phy.
So, both the DP controller and the USB PD phy are I2C devices, and now I have to make them both
attach to the AM bus as well?
BR,
Mats
Am Freitag, den 03.03.2017, 20:27 +0100 schrieb Mats Karrman:
> On 2017-03-03 13:59, Heikki Krogerus wrote:
>
> >
> > On Fri, Mar 03, 2017 at 08:29:18AM +0100, Mats Karrman wrote:
> > ....
>
> >
> > How would something like that sound to you guys?
>
> Complicated... Need to marinate on that for a while ;)
The thing is that a type C plug has operations that affect every
device on it and need to be synchronized. Reset and power
requirements would be the obvious examples.
It seems to me that the natural way to represent this is a bus.
Like SCSI where you have a bus reset operation.
Regards
Oliver
Hi Mats,
On Fri, Mar 03, 2017 at 08:27:08PM +0100, Mats Karrman wrote:
> On 2017-03-03 13:59, Heikki Krogerus wrote:
>
> > On Fri, Mar 03, 2017 at 08:29:18AM +0100, Mats Karrman wrote:
> > ....
>
> > How would something like that sound to you guys?
>
> Complicated... Need to marinate on that for a while ;)
Sorry about the bad explanation :-). Let me try again.. I'm simply
looking for a method that is as scalable as possible to handle the
alternate modes, basically how to couple the different components
involved. Bus would feel like the best approach at the moment.
> > > My system is a bit different. It's an i.MX6 SoC with the typec phy and DP controller connected
> > > directly to the SoC and it's using DTB/OF.
> > Is this "DP controller" a controller that is capable of taking care of
> > the USB Power Delivery communication with the partner regarding
> > DisplayPort alternate mode?
>
> No, the "DP controller" just talks DP and knows nothing about Type-C or USB PD.
> It takes a video stream from the SoC and turns it into a DP link, set up and orchestrated
> by the corresponding driver. And all the driver needs from Type-C is the plugged in / interrupt /
> plugged out events.
Got it.
> The analog switching between USB / safe / DP signal levels in the Type-C connector is, I think,
> best handled by the software doing the USB PD negotiation / Altmode handling (using some GPIOs).
>
> > > Do we need to further standardize attributes under (each) specific alternate mode to
> > > include things such as HPD for the DP mode?
> > I'm not completely sure what kind of system you have, but I would
> > imagine that if we had the bus, your DP controller driver would be the
> > port (and partner) alternate mode driver. The bus would bind you to
> > the typec phy.
>
> So, both the DP controller and the USB PD phy are I2C devices, and now I have to make them both
> attach to the AM bus as well?
The DP controller would provide the driver and the USB PD phy
(actually, the typec class) the device.
Would it be a problem to register these I2C devices with some other
subsystem, was it extcon or something like AM bus? It really would not
be that uncommon. Or have I misunderstood your question?
Thanks,
--
heikki
On 2017-03-06 14:14, Heikki Krogerus wrote:
> Hi Mats,
>
> On Fri, Mar 03, 2017 at 08:27:08PM +0100, Mats Karrman wrote:
>>>> My system is a bit different. It's an i.MX6 SoC with the typec phy and DP controller connected
>>>> directly to the SoC and it's using DTB/OF.
>>> Is this "DP controller" a controller that is capable of taking care of
>>> the USB Power Delivery communication with the partner regarding
>>> DisplayPort alternate mode?
>> No, the "DP controller" just talks DP and knows nothing about Type-C or USB PD.
>> It takes a video stream from the SoC and turns it into a DP link, set up and orchestrated
>> by the corresponding driver. And all the driver needs from Type-C is the plugged in / interrupt /
>> plugged out events.
> Got it.
>
>> The analog switching between USB / safe / DP signal levels in the Type-C connector is, I think,
>> best handled by the software doing the USB PD negotiation / Altmode handling (using some GPIOs).
>>
>>>> Do we need to further standardize attributes under (each) specific alternate mode to
>>>> include things such as HPD for the DP mode?
>>> I'm not completely sure what kind of system you have, but I would
>>> imagine that if we had the bus, your DP controller driver would be the
>>> port (and partner) alternate mode driver. The bus would bind you to
>>> the typec phy.
>> So, both the DP controller and the USB PD phy are I2C devices, and now I have to make them both
>> attach to the AM bus as well?
> The DP controller would provide the driver and the USB PD phy
> (actually, the typec class) the device.
>
> Would it be a problem to register these I2C devices with some other
> subsystem, was it extcon or something like AM bus? It really would not
> be that uncommon. Or have I misunderstood your question?
OK, so a bus could be used for drivers to find each other but it still does not say
anything about how those drivers are supposed to communicate so that must be prescribed
separately, right?
If I read Heikki's original suggestion I understand it like the DP driver would be
responsible for AM specific USB PD/VDM communication. But wouldn't that lead
to a lot of code duplication since the AM protocol is the same for all drivers of
a kind?
I'm still struggling to catch up on what you guys have been up to during the
last year or so :-) and came across some patches of Guenter from last October:
http://www.mail-archive.com/[email protected]/msg1243527.html
What happened to them? Has there been any progress since then?
BR // Mats
On 03/07/2017 02:30 PM, Mats Karrman wrote:
[ ... ]
>
> I'm still struggling to catch up on what you guys have been up to during the
> last year or so :-) and came across some patches of Guenter from last October:
>
> http://www.mail-archive.com/[email protected]/msg1243527.html
>
> What happened to them? Has there been any progress since then?
>
Updates to keep in sync with API changes, bug fixes, and minor improvements,
for the most part. I can post a current version if there is interest.
The latest version is also available from
https://chromium-review.googlesource.com/#/c/389917/
Guenter
On Tue, Mar 07, 2017 at 11:30:54PM +0100, Mats Karrman wrote:
> > > So, both the DP controller and the USB PD phy are I2C devices, and now I have to make them both
> > > attach to the AM bus as well?
> > The DP controller would provide the driver and the USB PD phy
> > (actually, the typec class) the device.
> >
> > Would it be a problem to register these I2C devices with some other
> > subsystem, was it extcon or something like AM bus? It really would not
> > be that uncommon. Or have I misunderstood your question?
>
> OK, so a bus could be used for drivers to find each other but it still does not say
> anything about how those drivers are supposed to communicate so that must be prescribed
> separately, right?
Yes.
> If I read Heikki's original suggestion I understand it like the DP driver would be
> responsible for AM specific USB PD/VDM communication. But wouldn't that lead
> to a lot of code duplication since the AM protocol is the same for all drivers of
> a kind?
No that's not what I mean. I'm still mixing your PD controller with
something else above, sorry about that. Your PD controller driver
should not ideally even need to be aware of Type-C connector, right?
It definitely does not need to do any USB PD communication.
I would imagine you have on top of the DP controller, a mux (which
could be a DP/USB3 PHY like on Rockchip RK3399, discrete mux like
Pericom PI3USB30532, or something else), and a USB Type-C PHY or USB
PD controller. The bus would be tying the mux to the Type-C port (PHY
or PD controller) and its partner (note that it does not tie the mux
to the DP controller). Please correct me if I'm wrong about your
hardware.
Assuming that is how your board roughly looks like, the driver for the
mux would be the driver for the DP altmode devices. That driver would
be the one converting things like the Attention messages notifying
about HPD into toggling of GPIOs, or what ever is needed on your
board, etc.
The actual PD communication with VDMs should be considered as just the
protocol, so we probable should have "protocol drivers". For example
DP alternate mode VDMs and communication will always be the same
despite of the hardware. The DP alternate mode "protocol driver" would
then be tied to the alternate mode device for the partner, and that
driver could have its own hooks for what ever is needed, like HPD
signal handling, configuration changes, whatever. In any case,
hopefully making things easy and straightforward for the "mux driver",
_so that it does not need to care about the actual PD communication_.
Thanks,
--
heikki
On 2017-03-08 14:58, Heikki Krogerus wrote:
> On Tue, Mar 07, 2017 at 11:30:54PM +0100, Mats Karrman wrote:
>> If I read Heikki's original suggestion I understand it like the DP driver would be
>> responsible for AM specific USB PD/VDM communication. But wouldn't that lead
>> to a lot of code duplication since the AM protocol is the same for all drivers of
>> a kind?
> No that's not what I mean. I'm still mixing your PD controller with
> something else above, sorry about that. Your PD controller driver
> should not ideally even need to be aware of Type-C connector, right?
> It definitely does not need to do any USB PD communication.
Right.
> I would imagine you have on top of the DP controller, a mux (which
> could be a DP/USB3 PHY like on Rockchip RK3399, discrete mux like
> Pericom PI3USB30532, or something else), and a USB Type-C PHY or USB
> PD controller. The bus would be tying the mux to the Type-C port (PHY
> or PD controller) and its partner (note that it does not tie the mux
> to the DP controller). Please correct me if I'm wrong about your
> hardware.
No, you're correct, a discrete mux and a fusb302.
> Assuming that is how your board roughly looks like, the driver for the
> mux would be the driver for the DP altmode devices. That driver would
> be the one converting things like the Attention messages notifying
> about HPD into toggling of GPIOs, or what ever is needed on your
> board, etc.
OK.
> The actual PD communication with VDMs should be considered as just the
> protocol, so we probable should have "protocol drivers". For example
> DP alternate mode VDMs and communication will always be the same
> despite of the hardware. The DP alternate mode "protocol driver" would
> then be tied to the alternate mode device for the partner, and that
> driver could have its own hooks for what ever is needed, like HPD
> signal handling, configuration changes, whatever. In any case,
> hopefully making things easy and straightforward for the "mux driver",
> _so that it does not need to care about the actual PD communication_.
I'm digesting your and Guenter's replies and patches.
I will try getting something up and running too soon and hopefully the foggy parts will
dissolve. As for now I find it a lot easier to grok Guenter's drivers than to see the
advantages and/or disadvantages of an altmode bus :-)
@Guenter: There _is_ interest for your fusb302 driver, thank you
BR // Mats
On 03/10/2017 02:22 PM, Mats Karrman wrote:
> On 2017-03-08 14:58, Heikki Krogerus wrote:
>
>> On Tue, Mar 07, 2017 at 11:30:54PM +0100, Mats Karrman wrote:
>>> If I read Heikki's original suggestion I understand it like the DP driver would be
>>> responsible for AM specific USB PD/VDM communication. But wouldn't that lead
>>> to a lot of code duplication since the AM protocol is the same for all drivers of
>>> a kind?
>> No that's not what I mean. I'm still mixing your PD controller with
>> something else above, sorry about that. Your PD controller driver
>> should not ideally even need to be aware of Type-C connector, right?
>> It definitely does not need to do any USB PD communication.
>
> Right.
>
>> I would imagine you have on top of the DP controller, a mux (which
>> could be a DP/USB3 PHY like on Rockchip RK3399, discrete mux like
>> Pericom PI3USB30532, or something else), and a USB Type-C PHY or USB
>> PD controller. The bus would be tying the mux to the Type-C port (PHY
>> or PD controller) and its partner (note that it does not tie the mux
>> to the DP controller). Please correct me if I'm wrong about your
>> hardware.
>
> No, you're correct, a discrete mux and a fusb302.
>
>> Assuming that is how your board roughly looks like, the driver for the
>> mux would be the driver for the DP altmode devices. That driver would
>> be the one converting things like the Attention messages notifying
>> about HPD into toggling of GPIOs, or what ever is needed on your
>> board, etc.
>
> OK.
>
>> The actual PD communication with VDMs should be considered as just the
>> protocol, so we probable should have "protocol drivers". For example
>> DP alternate mode VDMs and communication will always be the same
>> despite of the hardware. The DP alternate mode "protocol driver" would
>> then be tied to the alternate mode device for the partner, and that
>> driver could have its own hooks for what ever is needed, like HPD
>> signal handling, configuration changes, whatever. In any case,
>> hopefully making things easy and straightforward for the "mux driver",
>> _so that it does not need to care about the actual PD communication_.
>
> I'm digesting your and Guenter's replies and patches.
> I will try getting something up and running too soon and hopefully the foggy parts will
> dissolve. As for now I find it a lot easier to grok Guenter's drivers than to see the
> advantages and/or disadvantages of an altmode bus :-)
>
> @Guenter: There _is_ interest for your fusb302 driver, thank you
>
Ok, I'll see what I need to do to publish it.
Guenter
Hi Guenter,
From a previous thread:
On 03/08/2017 02:38 AM, Guenter Roeck wrote:
> On 03/07/2017 02:30 PM, Mats Karrman wrote:
> [ ... ]
>
>>
>> I'm still struggling to catch up on what you guys have been up to
>> during the
>> last year or so :-) and came across some patches of Guenter from last
>> October:
>>
>> http://www.mail-archive.com/[email protected]/msg1243527.html
>>
>> What happened to them? Has there been any progress since then?
>>
>
> Updates to keep in sync with API changes, bug fixes, and minor
> improvements,
> for the most part. I can post a current version if there is interest.
> The latest version is also available from
> https://chromium-review.googlesource.com/#/c/389917/
I'm working on a tcpi driver and have some concern about the tcpm api.
The tcpm_register_port() is typically called from the probe function of
tcpi driver where the tcpm_port reference returned is stored in the
driver private data. The problem I ran into is that tcpm_register_port()
calls back to the not yet fully initialized tcpi driver, causing null-
pointer dereferences. This could of course be solved by extra logic in
the tcpi driver but I think it would be more elegant if the registration
of a port could be free of premature callbacks. E.g. it could be required
that the tcpi driver probe called tcpm_tcpc_reset() once it's done
initializing or the necessary data could be supplied in the call to
tcpm_register_port().
What do you think?
BR,
Mats
Hi Mats,
On Sun, Apr 09, 2017 at 01:09:57AM +0200, Mats Karrman wrote:
> Hi Guenter,
>
> From a previous thread:
>
> On 03/08/2017 02:38 AM, Guenter Roeck wrote:
> >On 03/07/2017 02:30 PM, Mats Karrman wrote:
> >[ ... ]
> >
> >>
> >>I'm still struggling to catch up on what you guys have been up to during
> >>the
> >>last year or so :-) and came across some patches of Guenter from last
> >>October:
> >>
> >>http://www.mail-archive.com/[email protected]/msg1243527.html
> >>
> >>What happened to them? Has there been any progress since then?
> >>
> >
> >Updates to keep in sync with API changes, bug fixes, and minor
> >improvements,
> >for the most part. I can post a current version if there is interest.
> >The latest version is also available from
> >https://chromium-review.googlesource.com/#/c/389917/
>
> I'm working on a tcpi driver and have some concern about the tcpm api.
> The tcpm_register_port() is typically called from the probe function of
> tcpi driver where the tcpm_port reference returned is stored in the
> driver private data. The problem I ran into is that tcpm_register_port()
> calls back to the not yet fully initialized tcpi driver, causing null-
> pointer dereferences. This could of course be solved by extra logic in
> the tcpi driver but I think it would be more elegant if the registration
> of a port could be free of premature callbacks. E.g. it could be required
> that the tcpi driver probe called tcpm_tcpc_reset() once it's done
> initializing or the necessary data could be supplied in the call to
> tcpm_register_port().
> What do you think?
Let me think about it. In theory it should be possible to avoid callbacks into
the underlying driver until after the return from the registration code, but
that would still be racy if the underlying driver is not ready.
Basic problem seems to be that an API in general assumes that the caller is
ready to serve it once it registers itself. It is kind of unusual to have two
calls, one to register the driver and one to tell the infrastructure that it
is ready (which I assume your reset call would do). Ultimately this means
that the tcpm driver would have to have additional logic to identify if the
underlying driver is ready to handle callbacks.
Can you delay tcpm registration until after the underlying driver is ready,
ie typically to the end of its probe function ? Or am I misunderstanding
your problem ?
Thanks,
Guenter
On 04/09/2017 05:16 PM, Guenter Roeck wrote:
> Hi Mats,
>
> On Sun, Apr 09, 2017 at 01:09:57AM +0200, Mats Karrman wrote:
>> I'm working on a tcpi driver and have some concern about the tcpm api.
>> The tcpm_register_port() is typically called from the probe function of
>> tcpi driver where the tcpm_port reference returned is stored in the
>> driver private data. The problem I ran into is that tcpm_register_port()
>> calls back to the not yet fully initialized tcpi driver, causing null-
>> pointer dereferences. This could of course be solved by extra logic in
>> the tcpi driver but I think it would be more elegant if the registration
>> of a port could be free of premature callbacks. E.g. it could be required
>> that the tcpi driver probe called tcpm_tcpc_reset() once it's done
>> initializing or the necessary data could be supplied in the call to
>> tcpm_register_port().
>> What do you think?
> Let me think about it. In theory it should be possible to avoid callbacks into
> the underlying driver until after the return from the registration code, but
> that would still be racy if the underlying driver is not ready.
>
> Basic problem seems to be that an API in general assumes that the caller is
> ready to serve it once it registers itself. It is kind of unusual to have two
> calls, one to register the driver and one to tell the infrastructure that it
> is ready (which I assume your reset call would do). Ultimately this means
> that the tcpm driver would have to have additional logic to identify if the
> underlying driver is ready to handle callbacks.
>
> Can you delay tcpm registration until after the underlying driver is ready,
> ie typically to the end of its probe function ? Or am I misunderstanding
> your problem ?
The problem I had was that I was trying to pull the same trick that you do ;)
I.e. the probe function calls tcpm_register_port() that is calling back
to the driver that was trying to call back to tcpm, just that the call to
tcpm_register_port() has not yet returned so the pointer to tcpm_port in the
driver data structure was still null.
I was able to fix the issue by commenting out the call to tcpm_init() at the
end of tcpm_register_port() and instead call ("your") tcpm_tcpc_reset(), that
currently does nothing but calling tcpm_init(), after I'm done.
Even though I'm not overly excited about the tcpm register function calling back
to the driver I don't think my fix is much better. I should live by my own words
and refrain from calling back to tcpm until registration has finished...
// Mats
On 04/09/2017 02:05 PM, Mats Karrman wrote:
> On 04/09/2017 05:16 PM, Guenter Roeck wrote:
>
>> Hi Mats,
>>
>> On Sun, Apr 09, 2017 at 01:09:57AM +0200, Mats Karrman wrote:
>>> I'm working on a tcpi driver and have some concern about the tcpm api.
>>> The tcpm_register_port() is typically called from the probe function of
>>> tcpi driver where the tcpm_port reference returned is stored in the
>>> driver private data. The problem I ran into is that tcpm_register_port()
>>> calls back to the not yet fully initialized tcpi driver, causing null-
>>> pointer dereferences. This could of course be solved by extra logic in
>>> the tcpi driver but I think it would be more elegant if the registration
>>> of a port could be free of premature callbacks. E.g. it could be required
>>> that the tcpi driver probe called tcpm_tcpc_reset() once it's done
>>> initializing or the necessary data could be supplied in the call to
>>> tcpm_register_port().
>>> What do you think?
>> Let me think about it. In theory it should be possible to avoid callbacks into
>> the underlying driver until after the return from the registration code, but
>> that would still be racy if the underlying driver is not ready.
>>
>> Basic problem seems to be that an API in general assumes that the caller is
>> ready to serve it once it registers itself. It is kind of unusual to have two
>> calls, one to register the driver and one to tell the infrastructure that it
>> is ready (which I assume your reset call would do). Ultimately this means
>> that the tcpm driver would have to have additional logic to identify if the
>> underlying driver is ready to handle callbacks.
>>
>> Can you delay tcpm registration until after the underlying driver is ready,
>> ie typically to the end of its probe function ? Or am I misunderstanding
>> your problem ?
>
> The problem I had was that I was trying to pull the same trick that you do ;)
> I.e. the probe function calls tcpm_register_port() that is calling back
> to the driver that was trying to call back to tcpm, just that the call to
> tcpm_register_port() has not yet returned so the pointer to tcpm_port in the
> driver data structure was still null.
>
> I was able to fix the issue by commenting out the call to tcpm_init() at the
> end of tcpm_register_port() and instead call ("your") tcpm_tcpc_reset(), that
> currently does nothing but calling tcpm_init(), after I'm done.
>
> Even though I'm not overly excited about the tcpm register function calling back
> to the driver I don't think my fix is much better. I should live by my own words
> and refrain from calling back to tcpm until registration has finished...
>
Problem is that I can't even defer the call to tcpm_init(), or for that matter any
other call into the low level driver code, into the worker, since there is still
no guarantee that the low level driver is "done" with its initialization. The only
real solution I can think of is that the low level driver should not use the pointer
to tcpm_port in any of its callback functions.
Overall I think there is an assumption in any API that any callback functions provided
in a registration call can immediately be called. Otherwise any API would be in trouble.
Can you modify your code to not require the port pointer in its callback functions ?
Thanks,
Guenter
On 04/14/2017 04:57 AM, Guenter Roeck wrote:
> ... Overall I think there is an assumption in any API that any
> callback functions provided in a registration call can immediately be
> called. Otherwise any API would be in trouble. Can you modify your
> code to not require the port pointer in its callback functions ?
Yes, I have deferred my calls until after the registration function has
returned which seems only logical. Thanks, Mats
Hi Heikki,
I have a question regarding the preferred_role node.
+What: /sys/class/typec/<port>/preferred_role
+Date: March 2017
+Contact: Heikki Krogerus <[email protected]>
+Description:
+ The user space can notify the driver about the preferred role.
+ It should be handled as enabling of Try.SRC or Try.SNK, as
+ defined in USB Type-C specification, in the port drivers. By
+ default the preferred role should come from the platform.
+
+ Valid values: source, sink, none (to remove preference)
What is the expected behavior when the userspace changes the
preferred_role node when the port is in connected state ?
1. the state machine re-resolves the port roles right away based on
the new state machine in place ? (or)
2. Wait till the subsequent connect for resolving port roles based on the
new state machine.
For #1 to happen the policy_engine layer would have to reset the port
to resolve the port roles based on the (Try.SRC /Try.SNK/ Default)
new state machine preference.
Say for example when two non-PD devices following none (default state
machine) are connected, the port role resolution is going to be random.
But, if the userspace in one of the devices later changes the
preferred_role to source, then that device is most likely to become source
if the Try.SRC state-machine is re-run.
Does the above question fall under a policy decision ? If so, should there
be another node to say if the port roles have to re-resolved based on the
new state machine right away ?
Thanks !!
On Fri, Mar 10, 2017 at 3:41 PM, Guenter Roeck <[email protected]> wrote:
> On 03/10/2017 02:22 PM, Mats Karrman wrote:
>>
>> On 2017-03-08 14:58, Heikki Krogerus wrote:
>>
>>> On Tue, Mar 07, 2017 at 11:30:54PM +0100, Mats Karrman wrote:
>>>>
>>>> If I read Heikki's original suggestion I understand it like the DP
>>>> driver would be
>>>> responsible for AM specific USB PD/VDM communication. But wouldn't that
>>>> lead
>>>> to a lot of code duplication since the AM protocol is the same for all
>>>> drivers of
>>>> a kind?
>>>
>>> No that's not what I mean. I'm still mixing your PD controller with
>>> something else above, sorry about that. Your PD controller driver
>>> should not ideally even need to be aware of Type-C connector, right?
>>> It definitely does not need to do any USB PD communication.
>>
>>
>> Right.
>>
>>> I would imagine you have on top of the DP controller, a mux (which
>>> could be a DP/USB3 PHY like on Rockchip RK3399, discrete mux like
>>> Pericom PI3USB30532, or something else), and a USB Type-C PHY or USB
>>> PD controller. The bus would be tying the mux to the Type-C port (PHY
>>> or PD controller) and its partner (note that it does not tie the mux
>>> to the DP controller). Please correct me if I'm wrong about your
>>> hardware.
>>
>>
>> No, you're correct, a discrete mux and a fusb302.
>>
>>> Assuming that is how your board roughly looks like, the driver for the
>>> mux would be the driver for the DP altmode devices. That driver would
>>> be the one converting things like the Attention messages notifying
>>> about HPD into toggling of GPIOs, or what ever is needed on your
>>> board, etc.
>>
>>
>> OK.
>>
>>> The actual PD communication with VDMs should be considered as just the
>>> protocol, so we probable should have "protocol drivers". For example
>>> DP alternate mode VDMs and communication will always be the same
>>> despite of the hardware. The DP alternate mode "protocol driver" would
>>> then be tied to the alternate mode device for the partner, and that
>>> driver could have its own hooks for what ever is needed, like HPD
>>> signal handling, configuration changes, whatever. In any case,
>>> hopefully making things easy and straightforward for the "mux driver",
>>> _so that it does not need to care about the actual PD communication_.
>>
>>
>> I'm digesting your and Guenter's replies and patches.
>> I will try getting something up and running too soon and hopefully the
>> foggy parts will
>> dissolve. As for now I find it a lot easier to grok Guenter's drivers than
>> to see the
>> advantages and/or disadvantages of an altmode bus :-)
>>
>> @Guenter: There _is_ interest for your fusb302 driver, thank you
>>
>
> Ok, I'll see what I need to do to publish it.
>
> Guenter
>
>
> --
> To unsubscribe from this list: send the line "unsubscribe linux-usb" in
> the body of a message to [email protected]
> More majordomo info at http://vger.kernel.org/majordomo-info.html
Hi,
On Tue, Apr 18, 2017 at 11:52:33AM -0700, Badhri Jagan Sridharan wrote:
> Hi Heikki,
>
> I have a question regarding the preferred_role node.
>
> +What: /sys/class/typec/<port>/preferred_role
> +Date: March 2017
> +Contact: Heikki Krogerus <[email protected]>
> +Description:
> + The user space can notify the driver about the preferred role.
> + It should be handled as enabling of Try.SRC or Try.SNK, as
> + defined in USB Type-C specification, in the port drivers. By
> + default the preferred role should come from the platform.
> +
> + Valid values: source, sink, none (to remove preference)
>
> What is the expected behavior when the userspace changes the
> preferred_role node when the port is in connected state ?
>
> 1. the state machine re-resolves the port roles right away based on
> the new state machine in place ? (or)
No! There are separate attributes for sending role swap requests.
The attribute will "enable" Try.SRC/SNK states, i.e. next time the
state machine is executed, those states need to be considered.
Changing the value of this attribute must not affect the current
connection.
> 2. Wait till the subsequent connect for resolving port roles based on the
> new state machine.
Yes.
> For #1 to happen the policy_engine layer would have to reset the port
> to resolve the port roles based on the (Try.SRC /Try.SNK/ Default)
> new state machine preference.
>
> Say for example when two non-PD devices following none (default state
> machine) are connected, the port role resolution is going to be random.
> But, if the userspace in one of the devices later changes the
> preferred_role to source, then that device is most likely to become source
> if the Try.SRC state-machine is re-run.
>
> Does the above question fall under a policy decision ? If so, should there
> be another node to say if the port roles have to re-resolved based on the
> new state machine right away ?
I don't think we should even consider option #1, but just to be sure,
Oliver, what do you say?
I guess we need to say in the documentation explicitly that changing
the value will not affect the current connection.
Thanks,
--
heikki
On Wed, Apr 19, 2017 at 4:23 AM, Heikki Krogerus
<[email protected]> wrote:
> Hi,
>
> On Tue, Apr 18, 2017 at 11:52:33AM -0700, Badhri Jagan Sridharan wrote:
>> Hi Heikki,
>>
>> I have a question regarding the preferred_role node.
>>
>> +What: /sys/class/typec/<port>/preferred_role
>> +Date: March 2017
>> +Contact: Heikki Krogerus <[email protected]>
>> +Description:
>> + The user space can notify the driver about the preferred role.
>> + It should be handled as enabling of Try.SRC or Try.SNK, as
>> + defined in USB Type-C specification, in the port drivers. By
>> + default the preferred role should come from the platform.
>> +
>> + Valid values: source, sink, none (to remove preference)
>>
>> What is the expected behavior when the userspace changes the
>> preferred_role node when the port is in connected state ?
>>
>> 1. the state machine re-resolves the port roles right away based on
>> the new state machine in place ? (or)
>
> No! There are separate attributes for sending role swap requests.
Right. But, that might not be helpful in cases when PD is not implemented.
and Implementing PD is not mandatory according the spec :/
FYI quoting from the Type-C specification release(page 24),
role swaps are not limited to devices that only support PD.
"Two independent set of mechanisms are defined to allow a USB Type-C
DRP to functionally swap power and data roles. When USB PD is
supported, power and data role swapping is performed as a subsequent
step following the initial connection process. For non-PD implementations,
power/data role swapping can optionally be dealt with as part of the initial
connection process."
But, the current interface definition actually prevents current/data role
swaps for non-pd devices.
>
> The attribute will "enable" Try.SRC/SNK states, i.e. next time the
> state machine is executed, those states need to be considered.
> Changing the value of this attribute must not affect the current
> connection.
>
>> 2. Wait till the subsequent connect for resolving port roles based on the
>> new state machine.
>
> Yes.
>
>> For #1 to happen the policy_engine layer would have to reset the port
>> to resolve the port roles based on the (Try.SRC /Try.SNK/ Default)
>> new state machine preference.
>>
>> Say for example when two non-PD devices following none (default state
>> machine) are connected, the port role resolution is going to be random.
>> But, if the userspace in one of the devices later changes the
>> preferred_role to source, then that device is most likely to become source
>> if the Try.SRC state-machine is re-run.
>>
>> Does the above question fall under a policy decision ? If so, should there
>> be another node to say if the port roles have to re-resolved based on the
>> new state machine right away ?
>
> I don't think we should even consider option #1, but just to be sure,
> Oliver, what do you say?
Can we at least consider exposing a port_reset field so that the userspace
at least has an option to make the state machine to kick in right away with
a hard reset ?
Please do consider. We can't expect all low-end phones and devices with
smaller form factors then phones to implement PD as it might be an overkill
for them.
>
> I guess we need to say in the documentation explicitly that changing
> the value will not affect the current connection.
>
>
> Thanks,
>
> --
> heikki
On Wed, Apr 19, 2017 at 07:45:00AM -0700, Badhri Jagan Sridharan wrote:
> On Wed, Apr 19, 2017 at 4:23 AM, Heikki Krogerus
> <[email protected]> wrote:
> > Hi,
> >
> > On Tue, Apr 18, 2017 at 11:52:33AM -0700, Badhri Jagan Sridharan wrote:
> >> Hi Heikki,
> >>
> >> I have a question regarding the preferred_role node.
> >>
> >> +What: /sys/class/typec/<port>/preferred_role
> >> +Date: March 2017
> >> +Contact: Heikki Krogerus <[email protected]>
> >> +Description:
> >> + The user space can notify the driver about the preferred role.
> >> + It should be handled as enabling of Try.SRC or Try.SNK, as
> >> + defined in USB Type-C specification, in the port drivers. By
> >> + default the preferred role should come from the platform.
> >> +
> >> + Valid values: source, sink, none (to remove preference)
> >>
> >> What is the expected behavior when the userspace changes the
> >> preferred_role node when the port is in connected state ?
> >>
> >> 1. the state machine re-resolves the port roles right away based on
> >> the new state machine in place ? (or)
> >
> > No! There are separate attributes for sending role swap requests.
>
> Right. But, that might not be helpful in cases when PD is not implemented.
> and Implementing PD is not mandatory according the spec :/
>
> FYI quoting from the Type-C specification release(page 24),
> role swaps are not limited to devices that only support PD.
>
> "Two independent set of mechanisms are defined to allow a USB Type-C
> DRP to functionally swap power and data roles. When USB PD is
> supported, power and data role swapping is performed as a subsequent
> step following the initial connection process. For non-PD implementations,
> power/data role swapping can optionally be dealt with as part of the initial
> connection process."
>
> But, the current interface definition actually prevents current/data role
> swaps for non-pd devices.
>
This is correct for the attribute definition, but it is not implemented
that way. Writing the attribute is only read-only for non-DRP ports.
Given the standard, I would consider that to be intentional; it might
make sense to update the description accordingly.
How about implementing a mechanism in the dr_set and pr_set code in tcpm
which would handle that situation ? Something along the line of
if (!port->pd_capable && connected && current role != desired role) {
reset_port();
goto done;
}
My current code doesn't handle the !pd_capable state, so I'll need to do
something anyway.
Thanks,
Guenter
> >
> > The attribute will "enable" Try.SRC/SNK states, i.e. next time the
> > state machine is executed, those states need to be considered.
> > Changing the value of this attribute must not affect the current
> > connection.
> >
> >> 2. Wait till the subsequent connect for resolving port roles based on the
> >> new state machine.
> >
> > Yes.
> >
> >> For #1 to happen the policy_engine layer would have to reset the port
> >> to resolve the port roles based on the (Try.SRC /Try.SNK/ Default)
> >> new state machine preference.
> >>
> >> Say for example when two non-PD devices following none (default state
> >> machine) are connected, the port role resolution is going to be random.
> >> But, if the userspace in one of the devices later changes the
> >> preferred_role to source, then that device is most likely to become source
> >> if the Try.SRC state-machine is re-run.
> >>
> >> Does the above question fall under a policy decision ? If so, should there
> >> be another node to say if the port roles have to re-resolved based on the
> >> new state machine right away ?
> >
> > I don't think we should even consider option #1, but just to be sure,
> > Oliver, what do you say?
>
> Can we at least consider exposing a port_reset field so that the userspace
> at least has an option to make the state machine to kick in right away with
> a hard reset ?
>
> Please do consider. We can't expect all low-end phones and devices with
> smaller form factors then phones to implement PD as it might be an overkill
> for them.
>
> >
> > I guess we need to say in the documentation explicitly that changing
> > the value will not affect the current connection.
> >
> >
> > Thanks,
> >
> > --
> > heikki
On Wed, Apr 19, 2017 at 8:14 AM, Guenter Roeck <[email protected]> wrote:
> On Wed, Apr 19, 2017 at 07:45:00AM -0700, Badhri Jagan Sridharan wrote:
>> On Wed, Apr 19, 2017 at 4:23 AM, Heikki Krogerus
>> <[email protected]> wrote:
>> > Hi,
>> >
>> > On Tue, Apr 18, 2017 at 11:52:33AM -0700, Badhri Jagan Sridharan wrote:
>> >> Hi Heikki,
>> >>
>> >> I have a question regarding the preferred_role node.
>> >>
>> >> +What: /sys/class/typec/<port>/preferred_role
>> >> +Date: March 2017
>> >> +Contact: Heikki Krogerus <[email protected]>
>> >> +Description:
>> >> + The user space can notify the driver about the preferred role.
>> >> + It should be handled as enabling of Try.SRC or Try.SNK, as
>> >> + defined in USB Type-C specification, in the port drivers. By
>> >> + default the preferred role should come from the platform.
>> >> +
>> >> + Valid values: source, sink, none (to remove preference)
>> >>
>> >> What is the expected behavior when the userspace changes the
>> >> preferred_role node when the port is in connected state ?
>> >>
>> >> 1. the state machine re-resolves the port roles right away based on
>> >> the new state machine in place ? (or)
>> >
>> > No! There are separate attributes for sending role swap requests.
>>
>> Right. But, that might not be helpful in cases when PD is not implemented.
>> and Implementing PD is not mandatory according the spec :/
>>
>> FYI quoting from the Type-C specification release(page 24),
>> role swaps are not limited to devices that only support PD.
>>
>> "Two independent set of mechanisms are defined to allow a USB Type-C
>> DRP to functionally swap power and data roles. When USB PD is
>> supported, power and data role swapping is performed as a subsequent
>> step following the initial connection process. For non-PD implementations,
>> power/data role swapping can optionally be dealt with as part of the initial
>> connection process."
>>
>> But, the current interface definition actually prevents current/data role
>> swaps for non-pd devices.
>>
> This is correct for the attribute definition, but it is not implemented
> that way. Writing the attribute is only read-only for non-DRP ports.
i.e. tcpm_dr_set/tcpm_pr_set at tcpm.c would return EINVAL when type
is not TYPEC_PORT_DRP, is that what you are referring to ?
if (port->typec_caps.type != TYPEC_PORT_DRP) {
ret = -EINVAL;
goto port_unlock;
}
I do agree that this is actually correct. I am referring to the case
where port is
dual-role-power and dual-role-data but NOT PD capable.
> Given the standard, I would consider that to be intentional; it might
> make sense to update the description accordingly.
>
> How about implementing a mechanism in the dr_set and pr_set code in tcpm
> which would handle that situation ? Something along the line of
>
> if (!port->pd_capable && connected && current role != desired role) {
> reset_port();
> goto done;
> }
By "desired role" you are referring to preferred_role right ?
If so yes, That's a good idea as well and it might work as long as
type-c connector
class allows the call to reach tcpm code :) But the current connector
class code does
not allow that because the power_role and data_role nodes are defined that way.
port->cap->pd_revision and the port->pwr_opmode check in the below code
stub have to removed/refactored to make current_role/data_role writes to
reach the tcpm code.
+static ssize_t power_role_store(struct device *dev,
+ struct device_attribute *attr,
+ const char *buf, size_t size)
+{
+ struct typec_port *port = to_typec_port(dev);
+ int ret = size;
+
+ if (!port->cap->pd_revision) {
+ dev_dbg(dev, "USB Power Delivery not supported\n");
+ return -EOPNOTSUPP;
+ }
+
+ if (!port->cap->pr_set) {
+ dev_dbg(dev, "power role swapping not supported\n");
+ return -EOPNOTSUPP;
+ }
+
+ if (port->pwr_opmode != TYPEC_PWR_MODE_PD) {
+ dev_dbg(dev, "partner unable to swap power role\n");
+ return -EIO;
+ }
+
+ ret = sysfs_match_string(typec_roles, buf);
+ if (ret < 0)
+ return ret;
+
+ ret = port->cap->pr_set(port->cap, ret);
+ if (ret)
+ return ret;
+
+ return size;
+}
Thanks,
Badhri
>
> My current code doesn't handle the !pd_capable state, so I'll need to do
> something anyway.
>
> Thanks,
> Guenter
>
>> >
>> > The attribute will "enable" Try.SRC/SNK states, i.e. next time the
>> > state machine is executed, those states need to be considered.
>> > Changing the value of this attribute must not affect the current
>> > connection.
>> >
>> >> 2. Wait till the subsequent connect for resolving port roles based on the
>> >> new state machine.
>> >
>> > Yes.
>> >
>> >> For #1 to happen the policy_engine layer would have to reset the port
>> >> to resolve the port roles based on the (Try.SRC /Try.SNK/ Default)
>> >> new state machine preference.
>> >>
>> >> Say for example when two non-PD devices following none (default state
>> >> machine) are connected, the port role resolution is going to be random.
>> >> But, if the userspace in one of the devices later changes the
>> >> preferred_role to source, then that device is most likely to become source
>> >> if the Try.SRC state-machine is re-run.
>> >>
>> >> Does the above question fall under a policy decision ? If so, should there
>> >> be another node to say if the port roles have to re-resolved based on the
>> >> new state machine right away ?
>> >
>> > I don't think we should even consider option #1, but just to be sure,
>> > Oliver, what do you say?
>>
>> Can we at least consider exposing a port_reset field so that the userspace
>> at least has an option to make the state machine to kick in right away with
>> a hard reset ?
>>
>> Please do consider. We can't expect all low-end phones and devices with
>> smaller form factors then phones to implement PD as it might be an overkill
>> for them.
>>
>> >
>> > I guess we need to say in the documentation explicitly that changing
>> > the value will not affect the current connection.
>> >
>> >
>> > Thanks,
>> >
>> > --
>> > heikki
On Wed, Apr 19, 2017 at 10:22:47AM -0700, Badhri Jagan Sridharan wrote:
> On Wed, Apr 19, 2017 at 8:14 AM, Guenter Roeck <[email protected]> wrote:
> > On Wed, Apr 19, 2017 at 07:45:00AM -0700, Badhri Jagan Sridharan wrote:
> >> On Wed, Apr 19, 2017 at 4:23 AM, Heikki Krogerus
> >> <[email protected]> wrote:
> >> > Hi,
> >> >
> >> > On Tue, Apr 18, 2017 at 11:52:33AM -0700, Badhri Jagan Sridharan wrote:
> >> >> Hi Heikki,
> >> >>
> >> >> I have a question regarding the preferred_role node.
> >> >>
> >> >> +What: /sys/class/typec/<port>/preferred_role
> >> >> +Date: March 2017
> >> >> +Contact: Heikki Krogerus <[email protected]>
> >> >> +Description:
> >> >> + The user space can notify the driver about the preferred role.
> >> >> + It should be handled as enabling of Try.SRC or Try.SNK, as
> >> >> + defined in USB Type-C specification, in the port drivers. By
> >> >> + default the preferred role should come from the platform.
> >> >> +
> >> >> + Valid values: source, sink, none (to remove preference)
> >> >>
> >> >> What is the expected behavior when the userspace changes the
> >> >> preferred_role node when the port is in connected state ?
> >> >>
> >> >> 1. the state machine re-resolves the port roles right away based on
> >> >> the new state machine in place ? (or)
> >> >
> >> > No! There are separate attributes for sending role swap requests.
> >>
> >> Right. But, that might not be helpful in cases when PD is not implemented.
> >> and Implementing PD is not mandatory according the spec :/
> >>
> >> FYI quoting from the Type-C specification release(page 24),
> >> role swaps are not limited to devices that only support PD.
> >>
> >> "Two independent set of mechanisms are defined to allow a USB Type-C
> >> DRP to functionally swap power and data roles. When USB PD is
> >> supported, power and data role swapping is performed as a subsequent
> >> step following the initial connection process. For non-PD implementations,
> >> power/data role swapping can optionally be dealt with as part of the initial
> >> connection process."
> >>
> >> But, the current interface definition actually prevents current/data role
> >> swaps for non-pd devices.
> >>
>
> > This is correct for the attribute definition, but it is not implemented
> > that way. Writing the attribute is only read-only for non-DRP ports.
>
> i.e. tcpm_dr_set/tcpm_pr_set at tcpm.c would return EINVAL when type
> is not TYPEC_PORT_DRP, is that what you are referring to ?
>
> if (port->typec_caps.type != TYPEC_PORT_DRP) {
> ret = -EINVAL;
> goto port_unlock;
> }
>
> I do agree that this is actually correct. I am referring to the case
> where port is
> dual-role-power and dual-role-data but NOT PD capable.
>
At least in the current implementation of tcpm, the port type is reported
from low level drivers and does not change if the partner is not PD
capable. Port capabilities don't change if the partner is not PD capable,
and both are reported separately to the typec infrastructure.
> > Given the standard, I would consider that to be intentional; it might
> > make sense to update the description accordingly.
> >
> > How about implementing a mechanism in the dr_set and pr_set code in tcpm
> > which would handle that situation ? Something along the line of
> >
> > if (!port->pd_capable && connected && current role != desired role) {
> > reset_port();
> > goto done;
> > }
>
> By "desired role" you are referring to preferred_role right ?
>
That is what it would boil down to, but that is really up to user space
to decide. User space would check if the reported role is the desired role,
and request a role change if it is not.
> If so yes, That's a good idea as well and it might work as long as
> type-c connector
> class allows the call to reach tcpm code :) But the current connector
> class code does
> not allow that because the power_role and data_role nodes are defined that way.
>
> port->cap->pd_revision and the port->pwr_opmode check in the below code
> stub have to removed/refactored to make current_role/data_role writes to
> reach the tcpm code.
>
> +static ssize_t power_role_store(struct device *dev,
> + struct device_attribute *attr,
> + const char *buf, size_t size)
> +{
> + struct typec_port *port = to_typec_port(dev);
> + int ret = size;
> +
> + if (!port->cap->pd_revision) {
> + dev_dbg(dev, "USB Power Delivery not supported\n");
> + return -EOPNOTSUPP;
> + }
> +
> + if (!port->cap->pr_set) {
> + dev_dbg(dev, "power role swapping not supported\n");
> + return -EOPNOTSUPP;
> + }
> +
> + if (port->pwr_opmode != TYPEC_PWR_MODE_PD) {
> + dev_dbg(dev, "partner unable to swap power role\n");
> + return -EIO;
> + }
> +
Ah yes, you are right. Of course, you could request a data role change,
which doesn't have that restriction, at least not right now, but that
would not really be clean. Maybe it should be left to the low level
driver to do that check, or the check could be augmented to only
require pwr_opmode == TYPEC_PWR_MODE_PD if a preferred role is not set.
Thanks,
Guenter
> + ret = sysfs_match_string(typec_roles, buf);
> + if (ret < 0)
> + return ret;
> +
> + ret = port->cap->pr_set(port->cap, ret);
> + if (ret)
> + return ret;
> +
> + return size;
> +}
>
> Thanks,
> Badhri
>
> >
> > My current code doesn't handle the !pd_capable state, so I'll need to do
> > something anyway.
> >
> > Thanks,
> > Guenter
> >
> >> >
> >> > The attribute will "enable" Try.SRC/SNK states, i.e. next time the
> >> > state machine is executed, those states need to be considered.
> >> > Changing the value of this attribute must not affect the current
> >> > connection.
> >> >
> >> >> 2. Wait till the subsequent connect for resolving port roles based on the
> >> >> new state machine.
> >> >
> >> > Yes.
> >> >
> >> >> For #1 to happen the policy_engine layer would have to reset the port
> >> >> to resolve the port roles based on the (Try.SRC /Try.SNK/ Default)
> >> >> new state machine preference.
> >> >>
> >> >> Say for example when two non-PD devices following none (default state
> >> >> machine) are connected, the port role resolution is going to be random.
> >> >> But, if the userspace in one of the devices later changes the
> >> >> preferred_role to source, then that device is most likely to become source
> >> >> if the Try.SRC state-machine is re-run.
> >> >>
> >> >> Does the above question fall under a policy decision ? If so, should there
> >> >> be another node to say if the port roles have to re-resolved based on the
> >> >> new state machine right away ?
> >> >
> >> > I don't think we should even consider option #1, but just to be sure,
> >> > Oliver, what do you say?
> >>
> >> Can we at least consider exposing a port_reset field so that the userspace
> >> at least has an option to make the state machine to kick in right away with
> >> a hard reset ?
> >>
> >> Please do consider. We can't expect all low-end phones and devices with
> >> smaller form factors then phones to implement PD as it might be an overkill
> >> for them.
> >>
> >> >
> >> > I guess we need to say in the documentation explicitly that changing
> >> > the value will not affect the current connection.
> >> >
> >> >
> >> > Thanks,
> >> >
> >> > --
> >> > heikki
Hi,
On Wed, Apr 19, 2017 at 07:45:00AM -0700, Badhri Jagan Sridharan wrote:
> >> What is the expected behavior when the userspace changes the
> >> preferred_role node when the port is in connected state ?
> >>
> >> 1. the state machine re-resolves the port roles right away based on
> >> the new state machine in place ? (or)
> >
> > No! There are separate attributes for sending role swap requests.
>
> Right. But, that might not be helpful in cases when PD is not implemented.
> and Implementing PD is not mandatory according the spec :/
>
> FYI quoting from the Type-C specification release(page 24),
> role swaps are not limited to devices that only support PD.
>
> "Two independent set of mechanisms are defined to allow a USB Type-C
> DRP to functionally swap power and data roles. When USB PD is
> supported, power and data role swapping is performed as a subsequent
> step following the initial connection process. For non-PD implementations,
> power/data role swapping can optionally be dealt with as part of the initial
> connection process."
>
> But, the current interface definition actually prevents current/data role
> swaps for non-pd devices.
No it doesn't. When USB PD is not supported, you can still use
data_role to swap the role. Of course, the lower level driver may
not support any role swapping in that case.
Thanks,
--
heikki
On Wed, Apr 19, 2017 at 10:22:47AM -0700, Badhri Jagan Sridharan wrote:
> On Wed, Apr 19, 2017 at 8:14 AM, Guenter Roeck <[email protected]> wrote:
> > On Wed, Apr 19, 2017 at 07:45:00AM -0700, Badhri Jagan Sridharan wrote:
> >> On Wed, Apr 19, 2017 at 4:23 AM, Heikki Krogerus
> >> <[email protected]> wrote:
> >> > Hi,
> >> >
> >> > On Tue, Apr 18, 2017 at 11:52:33AM -0700, Badhri Jagan Sridharan wrote:
> >> >> Hi Heikki,
> >> >>
> >> >> I have a question regarding the preferred_role node.
> >> >>
> >> >> +What: /sys/class/typec/<port>/preferred_role
> >> >> +Date: March 2017
> >> >> +Contact: Heikki Krogerus <[email protected]>
> >> >> +Description:
> >> >> + The user space can notify the driver about the preferred role.
> >> >> + It should be handled as enabling of Try.SRC or Try.SNK, as
> >> >> + defined in USB Type-C specification, in the port drivers. By
> >> >> + default the preferred role should come from the platform.
> >> >> +
> >> >> + Valid values: source, sink, none (to remove preference)
> >> >>
> >> >> What is the expected behavior when the userspace changes the
> >> >> preferred_role node when the port is in connected state ?
> >> >>
> >> >> 1. the state machine re-resolves the port roles right away based on
> >> >> the new state machine in place ? (or)
> >> >
> >> > No! There are separate attributes for sending role swap requests.
> >>
> >> Right. But, that might not be helpful in cases when PD is not implemented.
> >> and Implementing PD is not mandatory according the spec :/
> >>
> >> FYI quoting from the Type-C specification release(page 24),
> >> role swaps are not limited to devices that only support PD.
> >>
> >> "Two independent set of mechanisms are defined to allow a USB Type-C
> >> DRP to functionally swap power and data roles. When USB PD is
> >> supported, power and data role swapping is performed as a subsequent
> >> step following the initial connection process. For non-PD implementations,
> >> power/data role swapping can optionally be dealt with as part of the initial
> >> connection process."
> >>
> >> But, the current interface definition actually prevents current/data role
> >> swaps for non-pd devices.
> >>
>
> > This is correct for the attribute definition, but it is not implemented
> > that way. Writing the attribute is only read-only for non-DRP ports.
>
> i.e. tcpm_dr_set/tcpm_pr_set at tcpm.c would return EINVAL when type
> is not TYPEC_PORT_DRP, is that what you are referring to ?
>
> if (port->typec_caps.type != TYPEC_PORT_DRP) {
> ret = -EINVAL;
> goto port_unlock;
> }
>
> I do agree that this is actually correct. I am referring to the case
> where port is
> dual-role-power and dual-role-data but NOT PD capable.
>
> > Given the standard, I would consider that to be intentional; it might
> > make sense to update the description accordingly.
> >
> > How about implementing a mechanism in the dr_set and pr_set code in tcpm
> > which would handle that situation ? Something along the line of
> >
> > if (!port->pd_capable && connected && current role != desired role) {
> > reset_port();
> > goto done;
> > }
>
> By "desired role" you are referring to preferred_role right ?
>
> If so yes, That's a good idea as well and it might work as long as
> type-c connector
> class allows the call to reach tcpm code :) But the current connector
> class code does
> not allow that because the power_role and data_role nodes are defined that way.
Well, the data_role does not limit the requests from reaching the low
level drivers, but..
> port->cap->pd_revision and the port->pwr_opmode check in the below code
> stub have to removed/refactored to make current_role/data_role writes to
> reach the tcpm code.
>
> +static ssize_t power_role_store(struct device *dev,
> + struct device_attribute *attr,
> + const char *buf, size_t size)
> +{
> + struct typec_port *port = to_typec_port(dev);
> + int ret = size;
> +
> + if (!port->cap->pd_revision) {
> + dev_dbg(dev, "USB Power Delivery not supported\n");
> + return -EOPNOTSUPP;
> + }
> +
> + if (!port->cap->pr_set) {
> + dev_dbg(dev, "power role swapping not supported\n");
> + return -EOPNOTSUPP;
> + }
> +
> + if (port->pwr_opmode != TYPEC_PWR_MODE_PD) {
> + dev_dbg(dev, "partner unable to swap power role\n");
> + return -EIO;
> + }
> +
> + ret = sysfs_match_string(typec_roles, buf);
> + if (ret < 0)
> + return ret;
> +
> + ret = port->cap->pr_set(port->cap, ret);
> + if (ret)
> + return ret;
> +
> + return size;
> +}
.. yes. The power_role_store() does indeed need to be refactored. The
PD requirement should only be applied to Type-C spec versions < 1.2,
or removed completely. I would be happy to leave the checks to the low
level drivers.
Thanks,
--
heikki
Thanks for the responses :)
So seems like we have a plan.
In Type-C connector class the checks for TYPEC_PWR_MODE_PD
and pd_revision for both the port and the partner will be removed in
power_role_store and the data_role_store and will be delegated
to the low level drivers.
TCPM code will issue hard reset in tcpm_dr_set and tcpm_pr_set if
current_role is not same as the preferred_role.
I am going to make changes in my local kernel code base to start
making the corresponding changes in userspace.
Should I post-back the local kernel changes or Heikki and Geunter
you are planning to upload them ?
Thanks for the support !!
Badhri.
On Thu, Apr 20, 2017 at 5:24 AM, Heikki Krogerus
<[email protected]> wrote:
> On Wed, Apr 19, 2017 at 10:22:47AM -0700, Badhri Jagan Sridharan wrote:
>> On Wed, Apr 19, 2017 at 8:14 AM, Guenter Roeck <[email protected]> wrote:
>> > On Wed, Apr 19, 2017 at 07:45:00AM -0700, Badhri Jagan Sridharan wrote:
>> >> On Wed, Apr 19, 2017 at 4:23 AM, Heikki Krogerus
>> >> <[email protected]> wrote:
>> >> > Hi,
>> >> >
>> >> > On Tue, Apr 18, 2017 at 11:52:33AM -0700, Badhri Jagan Sridharan wrote:
>> >> >> Hi Heikki,
>> >> >>
>> >> >> I have a question regarding the preferred_role node.
>> >> >>
>> >> >> +What: /sys/class/typec/<port>/preferred_role
>> >> >> +Date: March 2017
>> >> >> +Contact: Heikki Krogerus <[email protected]>
>> >> >> +Description:
>> >> >> + The user space can notify the driver about the preferred role.
>> >> >> + It should be handled as enabling of Try.SRC or Try.SNK, as
>> >> >> + defined in USB Type-C specification, in the port drivers. By
>> >> >> + default the preferred role should come from the platform.
>> >> >> +
>> >> >> + Valid values: source, sink, none (to remove preference)
>> >> >>
>> >> >> What is the expected behavior when the userspace changes the
>> >> >> preferred_role node when the port is in connected state ?
>> >> >>
>> >> >> 1. the state machine re-resolves the port roles right away based on
>> >> >> the new state machine in place ? (or)
>> >> >
>> >> > No! There are separate attributes for sending role swap requests.
>> >>
>> >> Right. But, that might not be helpful in cases when PD is not implemented.
>> >> and Implementing PD is not mandatory according the spec :/
>> >>
>> >> FYI quoting from the Type-C specification release(page 24),
>> >> role swaps are not limited to devices that only support PD.
>> >>
>> >> "Two independent set of mechanisms are defined to allow a USB Type-C
>> >> DRP to functionally swap power and data roles. When USB PD is
>> >> supported, power and data role swapping is performed as a subsequent
>> >> step following the initial connection process. For non-PD implementations,
>> >> power/data role swapping can optionally be dealt with as part of the initial
>> >> connection process."
>> >>
>> >> But, the current interface definition actually prevents current/data role
>> >> swaps for non-pd devices.
>> >>
>>
>> > This is correct for the attribute definition, but it is not implemented
>> > that way. Writing the attribute is only read-only for non-DRP ports.
>>
>> i.e. tcpm_dr_set/tcpm_pr_set at tcpm.c would return EINVAL when type
>> is not TYPEC_PORT_DRP, is that what you are referring to ?
>>
>> if (port->typec_caps.type != TYPEC_PORT_DRP) {
>> ret = -EINVAL;
>> goto port_unlock;
>> }
>>
>> I do agree that this is actually correct. I am referring to the case
>> where port is
>> dual-role-power and dual-role-data but NOT PD capable.
>>
>> > Given the standard, I would consider that to be intentional; it might
>> > make sense to update the description accordingly.
>> >
>> > How about implementing a mechanism in the dr_set and pr_set code in tcpm
>> > which would handle that situation ? Something along the line of
>> >
>> > if (!port->pd_capable && connected && current role != desired role) {
>> > reset_port();
>> > goto done;
>> > }
>>
>> By "desired role" you are referring to preferred_role right ?
>>
>> If so yes, That's a good idea as well and it might work as long as
>> type-c connector
>> class allows the call to reach tcpm code :) But the current connector
>> class code does
>> not allow that because the power_role and data_role nodes are defined that way.
>
> Well, the data_role does not limit the requests from reaching the low
> level drivers, but..
>
>> port->cap->pd_revision and the port->pwr_opmode check in the below code
>> stub have to removed/refactored to make current_role/data_role writes to
>> reach the tcpm code.
>>
>> +static ssize_t power_role_store(struct device *dev,
>> + struct device_attribute *attr,
>> + const char *buf, size_t size)
>> +{
>> + struct typec_port *port = to_typec_port(dev);
>> + int ret = size;
>> +
>> + if (!port->cap->pd_revision) {
>> + dev_dbg(dev, "USB Power Delivery not supported\n");
>> + return -EOPNOTSUPP;
>> + }
>> +
>> + if (!port->cap->pr_set) {
>> + dev_dbg(dev, "power role swapping not supported\n");
>> + return -EOPNOTSUPP;
>> + }
>> +
>> + if (port->pwr_opmode != TYPEC_PWR_MODE_PD) {
>> + dev_dbg(dev, "partner unable to swap power role\n");
>> + return -EIO;
>> + }
>> +
>> + ret = sysfs_match_string(typec_roles, buf);
>> + if (ret < 0)
>> + return ret;
>> +
>> + ret = port->cap->pr_set(port->cap, ret);
>> + if (ret)
>> + return ret;
>> +
>> + return size;
>> +}
>
> .. yes. The power_role_store() does indeed need to be refactored. The
> PD requirement should only be applied to Type-C spec versions < 1.2,
> or removed completely. I would be happy to leave the checks to the low
> level drivers.
>
>
> Thanks,
>
> --
> heikki
Hi,
On Thu, Apr 20, 2017 at 12:46:59PM -0700, Badhri Jagan Sridharan wrote:
> Thanks for the responses :)
>
> So seems like we have a plan.
>
> In Type-C connector class the checks for TYPEC_PWR_MODE_PD
> and pd_revision for both the port and the partner will be removed in
> power_role_store and the data_role_store and will be delegated
> to the low level drivers.
>
> TCPM code will issue hard reset in tcpm_dr_set and tcpm_pr_set if
> current_role is not same as the preferred_role.
>
> I am going to make changes in my local kernel code base to start
> making the corresponding changes in userspace.
> Should I post-back the local kernel changes or Heikki and Geunter
> you are planning to upload them ?
Please send a patch for the class at least. I think the changes to the
tcpm can just be amended still at this point. That is of course up to
Guenter.
Thanks,
--
heikki
On 04/21/2017 05:12 AM, Heikki Krogerus wrote:
> Hi,
>
> On Thu, Apr 20, 2017 at 12:46:59PM -0700, Badhri Jagan Sridharan wrote:
>> Thanks for the responses :)
>>
>> So seems like we have a plan.
>>
>> In Type-C connector class the checks for TYPEC_PWR_MODE_PD
>> and pd_revision for both the port and the partner will be removed in
>> power_role_store and the data_role_store and will be delegated
>> to the low level drivers.
>>
>> TCPM code will issue hard reset in tcpm_dr_set and tcpm_pr_set if
>> current_role is not same as the preferred_role.
>>
>> I am going to make changes in my local kernel code base to start
>> making the corresponding changes in userspace.
>> Should I post-back the local kernel changes or Heikki and Geunter
>> you are planning to upload them ?
>
> Please send a patch for the class at least. I think the changes to the
> tcpm can just be amended still at this point. That is of course up to
> Guenter.
>
I'll try to make the changes and post a new version today.
Guenter
On Fri, Apr 21, 2017 at 1:16 AM, Badhri Jagan Sridharan
<[email protected]> wrote:
> Thanks for the responses :)
>
> So seems like we have a plan.
>
> In Type-C connector class the checks for TYPEC_PWR_MODE_PD
> and pd_revision for both the port and the partner will be removed in
> power_role_store and the data_role_store and will be delegated
> to the low level drivers.
It is important to remember what USB Type-C provide is mechanisms for
"TRYing" to become a particular role and not guaranteeing.
With what device combination do you fore see we could get the desired
role with this change ?
>
> TCPM code will issue hard reset in tcpm_dr_set and tcpm_pr_set if
> current_role is not same as the preferred_role.
>
> I am going to make changes in my local kernel code base to start
> making the corresponding changes in userspace.
> Should I post-back the local kernel changes or Heikki and Geunter
> you are planning to upload them ?
>
> Thanks for the support !!
> Badhri.
>
> On Thu, Apr 20, 2017 at 5:24 AM, Heikki Krogerus
> <[email protected]> wrote:
>> On Wed, Apr 19, 2017 at 10:22:47AM -0700, Badhri Jagan Sridharan wrote:
>>> On Wed, Apr 19, 2017 at 8:14 AM, Guenter Roeck <[email protected]> wrote:
>>> > On Wed, Apr 19, 2017 at 07:45:00AM -0700, Badhri Jagan Sridharan wrote:
>>> >> On Wed, Apr 19, 2017 at 4:23 AM, Heikki Krogerus
>>> >> <[email protected]> wrote:
>>> >> > Hi,
>>> >> >
>>> >> > On Tue, Apr 18, 2017 at 11:52:33AM -0700, Badhri Jagan Sridharan wrote:
>>> >> >> Hi Heikki,
>>> >> >>
>>> >> >> I have a question regarding the preferred_role node.
>>> >> >>
>>> >> >> +What: /sys/class/typec/<port>/preferred_role
>>> >> >> +Date: March 2017
>>> >> >> +Contact: Heikki Krogerus <[email protected]>
>>> >> >> +Description:
>>> >> >> + The user space can notify the driver about the preferred role.
>>> >> >> + It should be handled as enabling of Try.SRC or Try.SNK, as
>>> >> >> + defined in USB Type-C specification, in the port drivers. By
>>> >> >> + default the preferred role should come from the platform.
>>> >> >> +
>>> >> >> + Valid values: source, sink, none (to remove preference)
>>> >> >>
>>> >> >> What is the expected behavior when the userspace changes the
>>> >> >> preferred_role node when the port is in connected state ?
>>> >> >>
>>> >> >> 1. the state machine re-resolves the port roles right away based on
>>> >> >> the new state machine in place ? (or)
>>> >> >
>>> >> > No! There are separate attributes for sending role swap requests.
>>> >>
>>> >> Right. But, that might not be helpful in cases when PD is not implemented.
>>> >> and Implementing PD is not mandatory according the spec :/
>>> >>
>>> >> FYI quoting from the Type-C specification release(page 24),
>>> >> role swaps are not limited to devices that only support PD.
>>> >>
>>> >> "Two independent set of mechanisms are defined to allow a USB Type-C
>>> >> DRP to functionally swap power and data roles. When USB PD is
>>> >> supported, power and data role swapping is performed as a subsequent
>>> >> step following the initial connection process. For non-PD implementations,
>>> >> power/data role swapping can optionally be dealt with as part of the initial
>>> >> connection process."
>>> >>
>>> >> But, the current interface definition actually prevents current/data role
>>> >> swaps for non-pd devices.
>>> >>
>>>
>>> > This is correct for the attribute definition, but it is not implemented
>>> > that way. Writing the attribute is only read-only for non-DRP ports.
>>>
>>> i.e. tcpm_dr_set/tcpm_pr_set at tcpm.c would return EINVAL when type
>>> is not TYPEC_PORT_DRP, is that what you are referring to ?
>>>
>>> if (port->typec_caps.type != TYPEC_PORT_DRP) {
>>> ret = -EINVAL;
>>> goto port_unlock;
>>> }
>>>
>>> I do agree that this is actually correct. I am referring to the case
>>> where port is
>>> dual-role-power and dual-role-data but NOT PD capable.
>>>
>>> > Given the standard, I would consider that to be intentional; it might
>>> > make sense to update the description accordingly.
>>> >
>>> > How about implementing a mechanism in the dr_set and pr_set code in tcpm
>>> > which would handle that situation ? Something along the line of
>>> >
>>> > if (!port->pd_capable && connected && current role != desired role) {
>>> > reset_port();
>>> > goto done;
>>> > }
>>>
>>> By "desired role" you are referring to preferred_role right ?
>>>
>>> If so yes, That's a good idea as well and it might work as long as
>>> type-c connector
>>> class allows the call to reach tcpm code :) But the current connector
>>> class code does
>>> not allow that because the power_role and data_role nodes are defined that way.
>>
>> Well, the data_role does not limit the requests from reaching the low
>> level drivers, but..
>>
>>> port->cap->pd_revision and the port->pwr_opmode check in the below code
>>> stub have to removed/refactored to make current_role/data_role writes to
>>> reach the tcpm code.
>>>
>>> +static ssize_t power_role_store(struct device *dev,
>>> + struct device_attribute *attr,
>>> + const char *buf, size_t size)
>>> +{
>>> + struct typec_port *port = to_typec_port(dev);
>>> + int ret = size;
>>> +
>>> + if (!port->cap->pd_revision) {
>>> + dev_dbg(dev, "USB Power Delivery not supported\n");
>>> + return -EOPNOTSUPP;
>>> + }
>>> +
>>> + if (!port->cap->pr_set) {
>>> + dev_dbg(dev, "power role swapping not supported\n");
>>> + return -EOPNOTSUPP;
>>> + }
>>> +
>>> + if (port->pwr_opmode != TYPEC_PWR_MODE_PD) {
>>> + dev_dbg(dev, "partner unable to swap power role\n");
>>> + return -EIO;
>>> + }
>>> +
>>> + ret = sysfs_match_string(typec_roles, buf);
>>> + if (ret < 0)
>>> + return ret;
>>> +
>>> + ret = port->cap->pr_set(port->cap, ret);
>>> + if (ret)
>>> + return ret;
>>> +
>>> + return size;
>>> +}
>>
>> .. yes. The power_role_store() does indeed need to be refactored. The
>> PD requirement should only be applied to Type-C spec versions < 1.2,
>> or removed completely. I would be happy to leave the checks to the low
>> level drivers.
>>
>>
>> Thanks,
>>
>> --
>> heikki
> --
> To unsubscribe from this list: send the line "unsubscribe linux-usb" in
> the body of a message to [email protected]
> More majordomo info at http://vger.kernel.org/majordomo-info.html
On Fri, Apr 21, 2017 at 07:57:52PM +0530, Rajaram R wrote:
> On Fri, Apr 21, 2017 at 1:16 AM, Badhri Jagan Sridharan
> <[email protected]> wrote:
> > Thanks for the responses :)
> >
> > So seems like we have a plan.
> >
> > In Type-C connector class the checks for TYPEC_PWR_MODE_PD
> > and pd_revision for both the port and the partner will be removed in
> > power_role_store and the data_role_store and will be delegated
> > to the low level drivers.
>
> It is important to remember what USB Type-C provide is mechanisms for
> "TRYing" to become a particular role and not guaranteeing.
>
> With what device combination do you fore see we could get the desired
> role with this change ?
>
If the partner is not PD capable, if a preferred role is specified,
if the current cole does not match the preferred role, and if the request
is to set the role to match the preferred role, I think it is reasonable
to expect that re-establishing the connection would accomplish that if the
partner supports it.
Of course, that won't change anything if the partner does not support the
desired role, but it is better than doing nothing. This is also comparable
to requesting a role change from the partner if it does support PD.
Do you have a better idea ?
Thanks,
Guenter
>
> >
> > TCPM code will issue hard reset in tcpm_dr_set and tcpm_pr_set if
> > current_role is not same as the preferred_role.
> >
... if the partner is not PD capable.
> > I am going to make changes in my local kernel code base to start
> > making the corresponding changes in userspace.
> > Should I post-back the local kernel changes or Heikki and Geunter
> > you are planning to upload them ?
> >
> > Thanks for the support !!
> > Badhri.
> >
> > On Thu, Apr 20, 2017 at 5:24 AM, Heikki Krogerus
> > <[email protected]> wrote:
> >> On Wed, Apr 19, 2017 at 10:22:47AM -0700, Badhri Jagan Sridharan wrote:
> >>> On Wed, Apr 19, 2017 at 8:14 AM, Guenter Roeck <[email protected]> wrote:
> >>> > On Wed, Apr 19, 2017 at 07:45:00AM -0700, Badhri Jagan Sridharan wrote:
> >>> >> On Wed, Apr 19, 2017 at 4:23 AM, Heikki Krogerus
> >>> >> <[email protected]> wrote:
> >>> >> > Hi,
> >>> >> >
> >>> >> > On Tue, Apr 18, 2017 at 11:52:33AM -0700, Badhri Jagan Sridharan wrote:
> >>> >> >> Hi Heikki,
> >>> >> >>
> >>> >> >> I have a question regarding the preferred_role node.
> >>> >> >>
> >>> >> >> +What: /sys/class/typec/<port>/preferred_role
> >>> >> >> +Date: March 2017
> >>> >> >> +Contact: Heikki Krogerus <[email protected]>
> >>> >> >> +Description:
> >>> >> >> + The user space can notify the driver about the preferred role.
> >>> >> >> + It should be handled as enabling of Try.SRC or Try.SNK, as
> >>> >> >> + defined in USB Type-C specification, in the port drivers. By
> >>> >> >> + default the preferred role should come from the platform.
> >>> >> >> +
> >>> >> >> + Valid values: source, sink, none (to remove preference)
> >>> >> >>
> >>> >> >> What is the expected behavior when the userspace changes the
> >>> >> >> preferred_role node when the port is in connected state ?
> >>> >> >>
> >>> >> >> 1. the state machine re-resolves the port roles right away based on
> >>> >> >> the new state machine in place ? (or)
> >>> >> >
> >>> >> > No! There are separate attributes for sending role swap requests.
> >>> >>
> >>> >> Right. But, that might not be helpful in cases when PD is not implemented.
> >>> >> and Implementing PD is not mandatory according the spec :/
> >>> >>
> >>> >> FYI quoting from the Type-C specification release(page 24),
> >>> >> role swaps are not limited to devices that only support PD.
> >>> >>
> >>> >> "Two independent set of mechanisms are defined to allow a USB Type-C
> >>> >> DRP to functionally swap power and data roles. When USB PD is
> >>> >> supported, power and data role swapping is performed as a subsequent
> >>> >> step following the initial connection process. For non-PD implementations,
> >>> >> power/data role swapping can optionally be dealt with as part of the initial
> >>> >> connection process."
> >>> >>
> >>> >> But, the current interface definition actually prevents current/data role
> >>> >> swaps for non-pd devices.
> >>> >>
> >>>
> >>> > This is correct for the attribute definition, but it is not implemented
> >>> > that way. Writing the attribute is only read-only for non-DRP ports.
> >>>
> >>> i.e. tcpm_dr_set/tcpm_pr_set at tcpm.c would return EINVAL when type
> >>> is not TYPEC_PORT_DRP, is that what you are referring to ?
> >>>
> >>> if (port->typec_caps.type != TYPEC_PORT_DRP) {
> >>> ret = -EINVAL;
> >>> goto port_unlock;
> >>> }
> >>>
> >>> I do agree that this is actually correct. I am referring to the case
> >>> where port is
> >>> dual-role-power and dual-role-data but NOT PD capable.
> >>>
> >>> > Given the standard, I would consider that to be intentional; it might
> >>> > make sense to update the description accordingly.
> >>> >
> >>> > How about implementing a mechanism in the dr_set and pr_set code in tcpm
> >>> > which would handle that situation ? Something along the line of
> >>> >
> >>> > if (!port->pd_capable && connected && current role != desired role) {
> >>> > reset_port();
> >>> > goto done;
> >>> > }
> >>>
> >>> By "desired role" you are referring to preferred_role right ?
> >>>
> >>> If so yes, That's a good idea as well and it might work as long as
> >>> type-c connector
> >>> class allows the call to reach tcpm code :) But the current connector
> >>> class code does
> >>> not allow that because the power_role and data_role nodes are defined that way.
> >>
> >> Well, the data_role does not limit the requests from reaching the low
> >> level drivers, but..
> >>
> >>> port->cap->pd_revision and the port->pwr_opmode check in the below code
> >>> stub have to removed/refactored to make current_role/data_role writes to
> >>> reach the tcpm code.
> >>>
> >>> +static ssize_t power_role_store(struct device *dev,
> >>> + struct device_attribute *attr,
> >>> + const char *buf, size_t size)
> >>> +{
> >>> + struct typec_port *port = to_typec_port(dev);
> >>> + int ret = size;
> >>> +
> >>> + if (!port->cap->pd_revision) {
> >>> + dev_dbg(dev, "USB Power Delivery not supported\n");
> >>> + return -EOPNOTSUPP;
> >>> + }
> >>> +
> >>> + if (!port->cap->pr_set) {
> >>> + dev_dbg(dev, "power role swapping not supported\n");
> >>> + return -EOPNOTSUPP;
> >>> + }
> >>> +
> >>> + if (port->pwr_opmode != TYPEC_PWR_MODE_PD) {
> >>> + dev_dbg(dev, "partner unable to swap power role\n");
> >>> + return -EIO;
> >>> + }
> >>> +
> >>> + ret = sysfs_match_string(typec_roles, buf);
> >>> + if (ret < 0)
> >>> + return ret;
> >>> +
> >>> + ret = port->cap->pr_set(port->cap, ret);
> >>> + if (ret)
> >>> + return ret;
> >>> +
> >>> + return size;
> >>> +}
> >>
> >> .. yes. The power_role_store() does indeed need to be refactored. The
> >> PD requirement should only be applied to Type-C spec versions < 1.2,
> >> or removed completely. I would be happy to leave the checks to the low
> >> level drivers.
> >>
> >>
> >> Thanks,
> >>
> >> --
> >> heikki
> > --
> > To unsubscribe from this list: send the line "unsubscribe linux-usb" in
> > the body of a message to [email protected]
> > More majordomo info at http://vger.kernel.org/majordomo-info.html
On Fri, Apr 21, 2017 at 10:13 PM, Guenter Roeck <[email protected]> wrote:
> On Fri, Apr 21, 2017 at 07:57:52PM +0530, Rajaram R wrote:
>> On Fri, Apr 21, 2017 at 1:16 AM, Badhri Jagan Sridharan
>> <[email protected]> wrote:
>> > Thanks for the responses :)
>> >
>> > So seems like we have a plan.
>> >
>> > In Type-C connector class the checks for TYPEC_PWR_MODE_PD
>> > and pd_revision for both the port and the partner will be removed in
>> > power_role_store and the data_role_store and will be delegated
>> > to the low level drivers.
>>
>> It is important to remember what USB Type-C provide is mechanisms for
>> "TRYing" to become a particular role and not guaranteeing.
>>
>> With what device combination do you fore see we could get the desired
>> role with this change ?
>>
>
> If the partner is not PD capable, if a preferred role is specified,
> if the current cole does not match the preferred role, and if the request
> is to set the role to match the preferred role, I think it is reasonable
> to expect that re-establishing the connection would accomplish that if the
> partner supports it.
>
In this context I believe we have two different inputs as follows:
/sys/class/typec/<port>/supported_power_roles
/sys/class/typec/<port>/preferred_role
The need of preferred role is required when DRP is set in
supported_power_roles option.
Ideally a battery powered device will TRY to be SNK and a a/c plugged
device will TRY to be SRC
We need to understand which non-PD device will set to DRP? In the
current ecosystem all legacy devices
will sit behind adapters which either present an Rp or Rd.
If it is a power adapter in 5V range can either present Rp or DRP with
TRY.SRC and there is no role swap requirement.
If it is a laptop port or similar with non-PD (??) DRP there is no
guaranteed role swap in a non-PD mode.
So we need to understand what non PD device will fit into this scenario ?
> Of course, that won't change anything if the partner does not support the
> desired role, but it is better than doing nothing. This is also comparable
> to requesting a role change from the partner if it does support PD.
All I am highlighting is that we can only TRY and there is no
guaranteed role swap with Type-C
> Do you have a better idea ?
>
If need a guaranteed role in a non-PD mode we need to set the required
role in supported_power_roles.
An understanding of scenario will help take better approach.
> Thanks,
> Guenter
>
>>
>> >
>> > TCPM code will issue hard reset in tcpm_dr_set and tcpm_pr_set if
>> > current_role is not same as the preferred_role.
>> >
>
> ... if the partner is not PD capable.
>
>> > I am going to make changes in my local kernel code base to start
>> > making the corresponding changes in userspace.
>> > Should I post-back the local kernel changes or Heikki and Geunter
>> > you are planning to upload them ?
>> >
>> > Thanks for the support !!
>> > Badhri.
>> >
>> > On Thu, Apr 20, 2017 at 5:24 AM, Heikki Krogerus
>> > <[email protected]> wrote:
>> >> On Wed, Apr 19, 2017 at 10:22:47AM -0700, Badhri Jagan Sridharan wrote:
>> >>> On Wed, Apr 19, 2017 at 8:14 AM, Guenter Roeck <[email protected]> wrote:
>> >>> > On Wed, Apr 19, 2017 at 07:45:00AM -0700, Badhri Jagan Sridharan wrote:
>> >>> >> On Wed, Apr 19, 2017 at 4:23 AM, Heikki Krogerus
>> >>> >> <[email protected]> wrote:
>> >>> >> > Hi,
>> >>> >> >
>> >>> >> > On Tue, Apr 18, 2017 at 11:52:33AM -0700, Badhri Jagan Sridharan wrote:
>> >>> >> >> Hi Heikki,
>> >>> >> >>
>> >>> >> >> I have a question regarding the preferred_role node.
>> >>> >> >>
>> >>> >> >> +What: /sys/class/typec/<port>/preferred_role
>> >>> >> >> +Date: March 2017
>> >>> >> >> +Contact: Heikki Krogerus <[email protected]>
>> >>> >> >> +Description:
>> >>> >> >> + The user space can notify the driver about the preferred role.
>> >>> >> >> + It should be handled as enabling of Try.SRC or Try.SNK, as
>> >>> >> >> + defined in USB Type-C specification, in the port drivers. By
>> >>> >> >> + default the preferred role should come from the platform.
>> >>> >> >> +
>> >>> >> >> + Valid values: source, sink, none (to remove preference)
>> >>> >> >>
>> >>> >> >> What is the expected behavior when the userspace changes the
>> >>> >> >> preferred_role node when the port is in connected state ?
>> >>> >> >>
>> >>> >> >> 1. the state machine re-resolves the port roles right away based on
>> >>> >> >> the new state machine in place ? (or)
>> >>> >> >
>> >>> >> > No! There are separate attributes for sending role swap requests.
>> >>> >>
>> >>> >> Right. But, that might not be helpful in cases when PD is not implemented.
>> >>> >> and Implementing PD is not mandatory according the spec :/
>> >>> >>
>> >>> >> FYI quoting from the Type-C specification release(page 24),
>> >>> >> role swaps are not limited to devices that only support PD.
>> >>> >>
>> >>> >> "Two independent set of mechanisms are defined to allow a USB Type-C
>> >>> >> DRP to functionally swap power and data roles. When USB PD is
>> >>> >> supported, power and data role swapping is performed as a subsequent
>> >>> >> step following the initial connection process. For non-PD implementations,
>> >>> >> power/data role swapping can optionally be dealt with as part of the initial
>> >>> >> connection process."
>> >>> >>
>> >>> >> But, the current interface definition actually prevents current/data role
>> >>> >> swaps for non-pd devices.
>> >>> >>
>> >>>
>> >>> > This is correct for the attribute definition, but it is not implemented
>> >>> > that way. Writing the attribute is only read-only for non-DRP ports.
>> >>>
>> >>> i.e. tcpm_dr_set/tcpm_pr_set at tcpm.c would return EINVAL when type
>> >>> is not TYPEC_PORT_DRP, is that what you are referring to ?
>> >>>
>> >>> if (port->typec_caps.type != TYPEC_PORT_DRP) {
>> >>> ret = -EINVAL;
>> >>> goto port_unlock;
>> >>> }
>> >>>
>> >>> I do agree that this is actually correct. I am referring to the case
>> >>> where port is
>> >>> dual-role-power and dual-role-data but NOT PD capable.
>> >>>
>> >>> > Given the standard, I would consider that to be intentional; it might
>> >>> > make sense to update the description accordingly.
>> >>> >
>> >>> > How about implementing a mechanism in the dr_set and pr_set code in tcpm
>> >>> > which would handle that situation ? Something along the line of
>> >>> >
>> >>> > if (!port->pd_capable && connected && current role != desired role) {
>> >>> > reset_port();
>> >>> > goto done;
>> >>> > }
>> >>>
>> >>> By "desired role" you are referring to preferred_role right ?
>> >>>
>> >>> If so yes, That's a good idea as well and it might work as long as
>> >>> type-c connector
>> >>> class allows the call to reach tcpm code :) But the current connector
>> >>> class code does
>> >>> not allow that because the power_role and data_role nodes are defined that way.
>> >>
>> >> Well, the data_role does not limit the requests from reaching the low
>> >> level drivers, but..
>> >>
>> >>> port->cap->pd_revision and the port->pwr_opmode check in the below code
>> >>> stub have to removed/refactored to make current_role/data_role writes to
>> >>> reach the tcpm code.
>> >>>
>> >>> +static ssize_t power_role_store(struct device *dev,
>> >>> + struct device_attribute *attr,
>> >>> + const char *buf, size_t size)
>> >>> +{
>> >>> + struct typec_port *port = to_typec_port(dev);
>> >>> + int ret = size;
>> >>> +
>> >>> + if (!port->cap->pd_revision) {
>> >>> + dev_dbg(dev, "USB Power Delivery not supported\n");
>> >>> + return -EOPNOTSUPP;
>> >>> + }
>> >>> +
>> >>> + if (!port->cap->pr_set) {
>> >>> + dev_dbg(dev, "power role swapping not supported\n");
>> >>> + return -EOPNOTSUPP;
>> >>> + }
>> >>> +
>> >>> + if (port->pwr_opmode != TYPEC_PWR_MODE_PD) {
>> >>> + dev_dbg(dev, "partner unable to swap power role\n");
>> >>> + return -EIO;
>> >>> + }
>> >>> +
>> >>> + ret = sysfs_match_string(typec_roles, buf);
>> >>> + if (ret < 0)
>> >>> + return ret;
>> >>> +
>> >>> + ret = port->cap->pr_set(port->cap, ret);
>> >>> + if (ret)
>> >>> + return ret;
>> >>> +
>> >>> + return size;
>> >>> +}
>> >>
>> >> .. yes. The power_role_store() does indeed need to be refactored. The
>> >> PD requirement should only be applied to Type-C spec versions < 1.2,
>> >> or removed completely. I would be happy to leave the checks to the low
>> >> level drivers.
>> >>
>> >>
>> >> Thanks,
>> >>
>> >> --
>> >> heikki
>> > --
>> > To unsubscribe from this list: send the line "unsubscribe linux-usb" in
>> > the body of a message to [email protected]
>> > More majordomo info at http://vger.kernel.org/majordomo-info.html
On Sat, Apr 22, 2017 at 2:23 AM, Rajaram R <[email protected]> wrote:
> On Fri, Apr 21, 2017 at 10:13 PM, Guenter Roeck <[email protected]> wrote:
>> On Fri, Apr 21, 2017 at 07:57:52PM +0530, Rajaram R wrote:
>>> On Fri, Apr 21, 2017 at 1:16 AM, Badhri Jagan Sridharan
>>> <[email protected]> wrote:
>>> > Thanks for the responses :)
>>> >
>>> > So seems like we have a plan.
>>> >
>>> > In Type-C connector class the checks for TYPEC_PWR_MODE_PD
>>> > and pd_revision for both the port and the partner will be removed in
>>> > power_role_store and the data_role_store and will be delegated
>>> > to the low level drivers.
>>>
>>> It is important to remember what USB Type-C provide is mechanisms for
>>> "TRYing" to become a particular role and not guaranteeing.
>>>
>>> With what device combination do you fore see we could get the desired
>>> role with this change ?
>>>
>>
>> If the partner is not PD capable, if a preferred role is specified,
>> if the current cole does not match the preferred role, and if the request
>> is to set the role to match the preferred role, I think it is reasonable
>> to expect that re-establishing the connection would accomplish that if the
>> partner supports it.
>>
> In this context I believe we have two different inputs as follows:
>
> /sys/class/typec/<port>/supported_power_roles
> /sys/class/typec/<port>/preferred_role
>
> The need of preferred role is required when DRP is set in
> supported_power_roles option.
> Ideally a battery powered device will TRY to be SNK and a a/c plugged
> device will TRY to be SRC
>
> We need to understand which non-PD device will set to DRP? In the
Android Phones (actually it could be any phone which has a type-c port)
since it can act as usb gadget (when connected to PC) or Usb host
when connected to peripherals such as thumb drives, keyboard etc.
Phones with smaller form factors might be thermally limited to charge
above 15W, therefore supporting PD might be an overkill for them.
> current ecosystem all legacy devices
> will sit behind adapters which either present an Rp or Rd.
>
> If it is a power adapter in 5V range can either present Rp or DRP with
> TRY.SRC and there is no role swap requirement.
>
> If it is a laptop port or similar with non-PD (??) DRP there is no
> guaranteed role swap in a non-PD mode.
This is true, but following a Try.SRC or Try.SNK state machine can
increase the chances of landing in the desired role/preferred role.
> So we need to understand what non PD device will fit into this scenario ?
Answered above.
>
>> Of course, that won't change anything if the partner does not support the
>> desired role, but it is better than doing nothing. This is also comparable
>> to requesting a role change from the partner if it does support PD.
>
> All I am highlighting is that we can only TRY and there is no
> guaranteed role swap with Type-C
>
>> Do you have a better idea ?
>>
> If need a guaranteed role in a non-PD mode we need to set the required
> role in supported_power_roles.
> An understanding of scenario will help take better approach.
The current Type-c connector class interface defines the support_*_roles as
read-only nodes. Leaving that apart, I think what you are trying to say is that
instead of running through the state machine again by switching to
Try.SRC or Try.SNK, you are suggesting that switch from DRP to source/host
(or) sink/device to make sure that CC is either pulled up through Rp or
grounded through Rd so that it increases the chances of settling in the desired
role. I do agree this, but, there is a pitfall here. Say when a DRP is
connected to
a pure sink/device, when the DRP switches to being a pure sink as well, then
the port roles would not resolve at all as both would be asserting Rd on CC and
therefore it might not be possible to detect a disconnect unless we have
a VCONN powered cable. Following Try.SRC, Try.SNK state machine actually
takes care of this for you. When in Try.SRC or Try.SNK state, CC would either
be pulled up or down for a specific amount of time (tCCDebounce) to check if the
port partner is capable of switching to another role. If no port
resolution happens
within the timer expiry, the state machine forces the port into the
other role and
port resolution would eventually happen. IMHO So in short it is more safer to
switch to between Try.SRC and Try.SNK state machine to land in a preferred role
rather than switching a DRP to source or sink.
>
>> Thanks,
>> Guenter
>>
>>>
>>> >
>>> > TCPM code will issue hard reset in tcpm_dr_set and tcpm_pr_set if
>>> > current_role is not same as the preferred_role.
>>> >
>>
>> ... if the partner is not PD capable.
>>
>>> > I am going to make changes in my local kernel code base to start
>>> > making the corresponding changes in userspace.
>>> > Should I post-back the local kernel changes or Heikki and Geunter
>>> > you are planning to upload them ?
>>> >
>>> > Thanks for the support !!
>>> > Badhri.
>>> >
>>> > On Thu, Apr 20, 2017 at 5:24 AM, Heikki Krogerus
>>> > <[email protected]> wrote:
>>> >> On Wed, Apr 19, 2017 at 10:22:47AM -0700, Badhri Jagan Sridharan wrote:
>>> >>> On Wed, Apr 19, 2017 at 8:14 AM, Guenter Roeck <[email protected]> wrote:
>>> >>> > On Wed, Apr 19, 2017 at 07:45:00AM -0700, Badhri Jagan Sridharan wrote:
>>> >>> >> On Wed, Apr 19, 2017 at 4:23 AM, Heikki Krogerus
>>> >>> >> <[email protected]> wrote:
>>> >>> >> > Hi,
>>> >>> >> >
>>> >>> >> > On Tue, Apr 18, 2017 at 11:52:33AM -0700, Badhri Jagan Sridharan wrote:
>>> >>> >> >> Hi Heikki,
>>> >>> >> >>
>>> >>> >> >> I have a question regarding the preferred_role node.
>>> >>> >> >>
>>> >>> >> >> +What: /sys/class/typec/<port>/preferred_role
>>> >>> >> >> +Date: March 2017
>>> >>> >> >> +Contact: Heikki Krogerus <[email protected]>
>>> >>> >> >> +Description:
>>> >>> >> >> + The user space can notify the driver about the preferred role.
>>> >>> >> >> + It should be handled as enabling of Try.SRC or Try.SNK, as
>>> >>> >> >> + defined in USB Type-C specification, in the port drivers. By
>>> >>> >> >> + default the preferred role should come from the platform.
>>> >>> >> >> +
>>> >>> >> >> + Valid values: source, sink, none (to remove preference)
>>> >>> >> >>
>>> >>> >> >> What is the expected behavior when the userspace changes the
>>> >>> >> >> preferred_role node when the port is in connected state ?
>>> >>> >> >>
>>> >>> >> >> 1. the state machine re-resolves the port roles right away based on
>>> >>> >> >> the new state machine in place ? (or)
>>> >>> >> >
>>> >>> >> > No! There are separate attributes for sending role swap requests.
>>> >>> >>
>>> >>> >> Right. But, that might not be helpful in cases when PD is not implemented.
>>> >>> >> and Implementing PD is not mandatory according the spec :/
>>> >>> >>
>>> >>> >> FYI quoting from the Type-C specification release(page 24),
>>> >>> >> role swaps are not limited to devices that only support PD.
>>> >>> >>
>>> >>> >> "Two independent set of mechanisms are defined to allow a USB Type-C
>>> >>> >> DRP to functionally swap power and data roles. When USB PD is
>>> >>> >> supported, power and data role swapping is performed as a subsequent
>>> >>> >> step following the initial connection process. For non-PD implementations,
>>> >>> >> power/data role swapping can optionally be dealt with as part of the initial
>>> >>> >> connection process."
>>> >>> >>
>>> >>> >> But, the current interface definition actually prevents current/data role
>>> >>> >> swaps for non-pd devices.
>>> >>> >>
>>> >>>
>>> >>> > This is correct for the attribute definition, but it is not implemented
>>> >>> > that way. Writing the attribute is only read-only for non-DRP ports.
>>> >>>
>>> >>> i.e. tcpm_dr_set/tcpm_pr_set at tcpm.c would return EINVAL when type
>>> >>> is not TYPEC_PORT_DRP, is that what you are referring to ?
>>> >>>
>>> >>> if (port->typec_caps.type != TYPEC_PORT_DRP) {
>>> >>> ret = -EINVAL;
>>> >>> goto port_unlock;
>>> >>> }
>>> >>>
>>> >>> I do agree that this is actually correct. I am referring to the case
>>> >>> where port is
>>> >>> dual-role-power and dual-role-data but NOT PD capable.
>>> >>>
>>> >>> > Given the standard, I would consider that to be intentional; it might
>>> >>> > make sense to update the description accordingly.
>>> >>> >
>>> >>> > How about implementing a mechanism in the dr_set and pr_set code in tcpm
>>> >>> > which would handle that situation ? Something along the line of
>>> >>> >
>>> >>> > if (!port->pd_capable && connected && current role != desired role) {
>>> >>> > reset_port();
>>> >>> > goto done;
>>> >>> > }
>>> >>>
>>> >>> By "desired role" you are referring to preferred_role right ?
>>> >>>
>>> >>> If so yes, That's a good idea as well and it might work as long as
>>> >>> type-c connector
>>> >>> class allows the call to reach tcpm code :) But the current connector
>>> >>> class code does
>>> >>> not allow that because the power_role and data_role nodes are defined that way.
>>> >>
>>> >> Well, the data_role does not limit the requests from reaching the low
>>> >> level drivers, but..
>>> >>
>>> >>> port->cap->pd_revision and the port->pwr_opmode check in the below code
>>> >>> stub have to removed/refactored to make current_role/data_role writes to
>>> >>> reach the tcpm code.
>>> >>>
>>> >>> +static ssize_t power_role_store(struct device *dev,
>>> >>> + struct device_attribute *attr,
>>> >>> + const char *buf, size_t size)
>>> >>> +{
>>> >>> + struct typec_port *port = to_typec_port(dev);
>>> >>> + int ret = size;
>>> >>> +
>>> >>> + if (!port->cap->pd_revision) {
>>> >>> + dev_dbg(dev, "USB Power Delivery not supported\n");
>>> >>> + return -EOPNOTSUPP;
>>> >>> + }
>>> >>> +
>>> >>> + if (!port->cap->pr_set) {
>>> >>> + dev_dbg(dev, "power role swapping not supported\n");
>>> >>> + return -EOPNOTSUPP;
>>> >>> + }
>>> >>> +
>>> >>> + if (port->pwr_opmode != TYPEC_PWR_MODE_PD) {
>>> >>> + dev_dbg(dev, "partner unable to swap power role\n");
>>> >>> + return -EIO;
>>> >>> + }
>>> >>> +
>>> >>> + ret = sysfs_match_string(typec_roles, buf);
>>> >>> + if (ret < 0)
>>> >>> + return ret;
>>> >>> +
>>> >>> + ret = port->cap->pr_set(port->cap, ret);
>>> >>> + if (ret)
>>> >>> + return ret;
>>> >>> +
>>> >>> + return size;
>>> >>> +}
>>> >>
>>> >> .. yes. The power_role_store() does indeed need to be refactored. The
>>> >> PD requirement should only be applied to Type-C spec versions < 1.2,
>>> >> or removed completely. I would be happy to leave the checks to the low
>>> >> level drivers.
>>> >>
>>> >>
>>> >> Thanks,
>>> >>
>>> >> --
>>> >> heikki
>>> > --
>>> > To unsubscribe from this list: send the line "unsubscribe linux-usb" in
>>> > the body of a message to [email protected]
>>> > More majordomo info at http://vger.kernel.org/majordomo-info.html
On Mon, Apr 24, 2017 at 11:20 PM, Badhri Jagan Sridharan
<[email protected]> wrote:
> On Sat, Apr 22, 2017 at 2:23 AM, Rajaram R <[email protected]> wrote:
>> On Fri, Apr 21, 2017 at 10:13 PM, Guenter Roeck <[email protected]> wrote:
>>> On Fri, Apr 21, 2017 at 07:57:52PM +0530, Rajaram R wrote:
>>>> On Fri, Apr 21, 2017 at 1:16 AM, Badhri Jagan Sridharan
>>>> <[email protected]> wrote:
>>>> > Thanks for the responses :)
>>>> >
>>>> > So seems like we have a plan.
>>>> >
>>>> > In Type-C connector class the checks for TYPEC_PWR_MODE_PD
>>>> > and pd_revision for both the port and the partner will be removed in
>>>> > power_role_store and the data_role_store and will be delegated
>>>> > to the low level drivers.
>>>>
>>>> It is important to remember what USB Type-C provide is mechanisms for
>>>> "TRYing" to become a particular role and not guaranteeing.
>>>>
>>>> With what device combination do you fore see we could get the desired
>>>> role with this change ?
>>>>
>>>
>>> If the partner is not PD capable, if a preferred role is specified,
>>> if the current cole does not match the preferred role, and if the request
>>> is to set the role to match the preferred role, I think it is reasonable
>>> to expect that re-establishing the connection would accomplish that if the
>>> partner supports it.
>>>
>> In this context I believe we have two different inputs as follows:
>>
>> /sys/class/typec/<port>/supported_power_roles
>> /sys/class/typec/<port>/preferred_role
>>
>> The need of preferred role is required when DRP is set in
>> supported_power_roles option.
>> Ideally a battery powered device will TRY to be SNK and a a/c plugged
>> device will TRY to be SRC
>>
>> We need to understand which non-PD device will set to DRP? In the
>
> Android Phones (actually it could be any phone which has a type-c port)
> since it can act as usb gadget (when connected to PC) or Usb host
> when connected to peripherals such as thumb drives, keyboard etc.
> Phones with smaller form factors might be thermally limited to charge
> above 15W, therefore supporting PD might be an overkill for them.
>
>> current ecosystem all legacy devices
>> will sit behind adapters which either present an Rp or Rd.
>>
>> If it is a power adapter in 5V range can either present Rp or DRP with
>> TRY.SRC and there is no role swap requirement.
>>
>> If it is a laptop port or similar with non-PD (??) DRP there is no
>> guaranteed role swap in a non-PD mode.
>
> This is true, but following a Try.SRC or Try.SNK state machine can
> increase the chances of landing in the desired role/preferred role.
Agree and as indicated it increases only chances.
>
>> So we need to understand what non PD device will fit into this scenario ?
> Answered above.
>
>>
>>> Of course, that won't change anything if the partner does not support the
>>> desired role, but it is better than doing nothing. This is also comparable
>>> to requesting a role change from the partner if it does support PD.
>>
>> All I am highlighting is that we can only TRY and there is no
>> guaranteed role swap with Type-C
>>
>>> Do you have a better idea ?
>>>
>> If need a guaranteed role in a non-PD mode we need to set the required
>> role in supported_power_roles.
>> An understanding of scenario will help take better approach.
>
> The current Type-c connector class interface defines the support_*_roles as
> read-only nodes. Leaving that apart, I think what you are trying to say is that
> instead of running through the state machine again by switching to
> Try.SRC or Try.SNK, you are suggesting that switch from DRP to source/host
> (or) sink/device to make sure that CC is either pulled up through Rp or
> grounded through Rd so that it increases the chances of settling in the desired
> role. I do agree this, but, there is a pitfall here. Say when a DRP is
> connected to
> a pure sink/device, when the DRP switches to being a pure sink as well, then
> the port roles would not resolve at all as both would be asserting Rd on CC and
If it is a pure SNK(presenting Rd), there is no conflict and a DRP be
it TRY SRC or TRY SNK
will endup as SRC.
> therefore it might not be possible to detect a disconnect unless we have
> a VCONN powered cable. Following Try.SRC, Try.SNK state machine actually
VConn does not play any role in detection.
> takes care of this for you. When in Try.SRC or Try.SNK state, CC would either
> be pulled up or down for a specific amount of time (tCCDebounce) to check if the
> port partner is capable of switching to another role. If no port
> resolution happens
> within the timer expiry, the state machine forces the port into the
> other role and
> port resolution would eventually happen. IMHO So in short it is more safer to
> switch to between Try.SRC and Try.SNK state machine to land in a preferred role
> rather than switching a DRP to source or sink.
The only scenario i can think of now is when connecting two mobiles
back to back. In this case if both devices
uses the same logic of switch to between Try.SRC and Try.SNK you will
not get desired result.
I guess the way we choose USB Classes(MTP or PTP or Tethering) could
be an approach where user know what he/she need.
>
>>
>>> Thanks,
>>> Guenter
>>>
>>>>
>>>> >
>>>> > TCPM code will issue hard reset in tcpm_dr_set and tcpm_pr_set if
>>>> > current_role is not same as the preferred_role.
>>>> >
>>>
>>> ... if the partner is not PD capable.
>>>
>>>> > I am going to make changes in my local kernel code base to start
>>>> > making the corresponding changes in userspace.
>>>> > Should I post-back the local kernel changes or Heikki and Geunter
>>>> > you are planning to upload them ?
>>>> >
>>>> > Thanks for the support !!
>>>> > Badhri.
>>>> >
>>>> > On Thu, Apr 20, 2017 at 5:24 AM, Heikki Krogerus
>>>> > <[email protected]> wrote:
>>>> >> On Wed, Apr 19, 2017 at 10:22:47AM -0700, Badhri Jagan Sridharan wrote:
>>>> >>> On Wed, Apr 19, 2017 at 8:14 AM, Guenter Roeck <[email protected]> wrote:
>>>> >>> > On Wed, Apr 19, 2017 at 07:45:00AM -0700, Badhri Jagan Sridharan wrote:
>>>> >>> >> On Wed, Apr 19, 2017 at 4:23 AM, Heikki Krogerus
>>>> >>> >> <[email protected]> wrote:
>>>> >>> >> > Hi,
>>>> >>> >> >
>>>> >>> >> > On Tue, Apr 18, 2017 at 11:52:33AM -0700, Badhri Jagan Sridharan wrote:
>>>> >>> >> >> Hi Heikki,
>>>> >>> >> >>
>>>> >>> >> >> I have a question regarding the preferred_role node.
>>>> >>> >> >>
>>>> >>> >> >> +What: /sys/class/typec/<port>/preferred_role
>>>> >>> >> >> +Date: March 2017
>>>> >>> >> >> +Contact: Heikki Krogerus <[email protected]>
>>>> >>> >> >> +Description:
>>>> >>> >> >> + The user space can notify the driver about the preferred role.
>>>> >>> >> >> + It should be handled as enabling of Try.SRC or Try.SNK, as
>>>> >>> >> >> + defined in USB Type-C specification, in the port drivers. By
>>>> >>> >> >> + default the preferred role should come from the platform.
>>>> >>> >> >> +
>>>> >>> >> >> + Valid values: source, sink, none (to remove preference)
>>>> >>> >> >>
>>>> >>> >> >> What is the expected behavior when the userspace changes the
>>>> >>> >> >> preferred_role node when the port is in connected state ?
>>>> >>> >> >>
>>>> >>> >> >> 1. the state machine re-resolves the port roles right away based on
>>>> >>> >> >> the new state machine in place ? (or)
>>>> >>> >> >
>>>> >>> >> > No! There are separate attributes for sending role swap requests.
>>>> >>> >>
>>>> >>> >> Right. But, that might not be helpful in cases when PD is not implemented.
>>>> >>> >> and Implementing PD is not mandatory according the spec :/
>>>> >>> >>
>>>> >>> >> FYI quoting from the Type-C specification release(page 24),
>>>> >>> >> role swaps are not limited to devices that only support PD.
>>>> >>> >>
>>>> >>> >> "Two independent set of mechanisms are defined to allow a USB Type-C
>>>> >>> >> DRP to functionally swap power and data roles. When USB PD is
>>>> >>> >> supported, power and data role swapping is performed as a subsequent
>>>> >>> >> step following the initial connection process. For non-PD implementations,
>>>> >>> >> power/data role swapping can optionally be dealt with as part of the initial
>>>> >>> >> connection process."
>>>> >>> >>
>>>> >>> >> But, the current interface definition actually prevents current/data role
>>>> >>> >> swaps for non-pd devices.
>>>> >>> >>
>>>> >>>
>>>> >>> > This is correct for the attribute definition, but it is not implemented
>>>> >>> > that way. Writing the attribute is only read-only for non-DRP ports.
>>>> >>>
>>>> >>> i.e. tcpm_dr_set/tcpm_pr_set at tcpm.c would return EINVAL when type
>>>> >>> is not TYPEC_PORT_DRP, is that what you are referring to ?
>>>> >>>
>>>> >>> if (port->typec_caps.type != TYPEC_PORT_DRP) {
>>>> >>> ret = -EINVAL;
>>>> >>> goto port_unlock;
>>>> >>> }
>>>> >>>
>>>> >>> I do agree that this is actually correct. I am referring to the case
>>>> >>> where port is
>>>> >>> dual-role-power and dual-role-data but NOT PD capable.
>>>> >>>
>>>> >>> > Given the standard, I would consider that to be intentional; it might
>>>> >>> > make sense to update the description accordingly.
>>>> >>> >
>>>> >>> > How about implementing a mechanism in the dr_set and pr_set code in tcpm
>>>> >>> > which would handle that situation ? Something along the line of
>>>> >>> >
>>>> >>> > if (!port->pd_capable && connected && current role != desired role) {
>>>> >>> > reset_port();
>>>> >>> > goto done;
>>>> >>> > }
>>>> >>>
>>>> >>> By "desired role" you are referring to preferred_role right ?
>>>> >>>
>>>> >>> If so yes, That's a good idea as well and it might work as long as
>>>> >>> type-c connector
>>>> >>> class allows the call to reach tcpm code :) But the current connector
>>>> >>> class code does
>>>> >>> not allow that because the power_role and data_role nodes are defined that way.
>>>> >>
>>>> >> Well, the data_role does not limit the requests from reaching the low
>>>> >> level drivers, but..
>>>> >>
>>>> >>> port->cap->pd_revision and the port->pwr_opmode check in the below code
>>>> >>> stub have to removed/refactored to make current_role/data_role writes to
>>>> >>> reach the tcpm code.
>>>> >>>
>>>> >>> +static ssize_t power_role_store(struct device *dev,
>>>> >>> + struct device_attribute *attr,
>>>> >>> + const char *buf, size_t size)
>>>> >>> +{
>>>> >>> + struct typec_port *port = to_typec_port(dev);
>>>> >>> + int ret = size;
>>>> >>> +
>>>> >>> + if (!port->cap->pd_revision) {
>>>> >>> + dev_dbg(dev, "USB Power Delivery not supported\n");
>>>> >>> + return -EOPNOTSUPP;
>>>> >>> + }
>>>> >>> +
>>>> >>> + if (!port->cap->pr_set) {
>>>> >>> + dev_dbg(dev, "power role swapping not supported\n");
>>>> >>> + return -EOPNOTSUPP;
>>>> >>> + }
>>>> >>> +
>>>> >>> + if (port->pwr_opmode != TYPEC_PWR_MODE_PD) {
>>>> >>> + dev_dbg(dev, "partner unable to swap power role\n");
>>>> >>> + return -EIO;
>>>> >>> + }
>>>> >>> +
>>>> >>> + ret = sysfs_match_string(typec_roles, buf);
>>>> >>> + if (ret < 0)
>>>> >>> + return ret;
>>>> >>> +
>>>> >>> + ret = port->cap->pr_set(port->cap, ret);
>>>> >>> + if (ret)
>>>> >>> + return ret;
>>>> >>> +
>>>> >>> + return size;
>>>> >>> +}
>>>> >>
>>>> >> .. yes. The power_role_store() does indeed need to be refactored. The
>>>> >> PD requirement should only be applied to Type-C spec versions < 1.2,
>>>> >> or removed completely. I would be happy to leave the checks to the low
>>>> >> level drivers.
>>>> >>
>>>> >>
>>>> >> Thanks,
>>>> >>
>>>> >> --
>>>> >> heikki
>>>> > --
>>>> > To unsubscribe from this list: send the line "unsubscribe linux-usb" in
>>>> > the body of a message to [email protected]
>>>> > More majordomo info at http://vger.kernel.org/majordomo-info.html
On 04/25/2017 01:26 AM, Rajaram R wrote:
> On Mon, Apr 24, 2017 at 11:20 PM, Badhri Jagan Sridharan
> <[email protected]> wrote:
>> On Sat, Apr 22, 2017 at 2:23 AM, Rajaram R <[email protected]> wrote:
>>> On Fri, Apr 21, 2017 at 10:13 PM, Guenter Roeck <[email protected]> wrote:
>>>> On Fri, Apr 21, 2017 at 07:57:52PM +0530, Rajaram R wrote:
>>>>> On Fri, Apr 21, 2017 at 1:16 AM, Badhri Jagan Sridharan
>>>>> <[email protected]> wrote:
>>>>>> Thanks for the responses :)
>>>>>>
>>>>>> So seems like we have a plan.
>>>>>>
>>>>>> In Type-C connector class the checks for TYPEC_PWR_MODE_PD
>>>>>> and pd_revision for both the port and the partner will be removed in
>>>>>> power_role_store and the data_role_store and will be delegated
>>>>>> to the low level drivers.
>>>>>
>>>>> It is important to remember what USB Type-C provide is mechanisms for
>>>>> "TRYing" to become a particular role and not guaranteeing.
>>>>>
>>>>> With what device combination do you fore see we could get the desired
>>>>> role with this change ?
>>>>>
>>>>
>>>> If the partner is not PD capable, if a preferred role is specified,
>>>> if the current cole does not match the preferred role, and if the request
>>>> is to set the role to match the preferred role, I think it is reasonable
>>>> to expect that re-establishing the connection would accomplish that if the
>>>> partner supports it.
>>>>
>>> In this context I believe we have two different inputs as follows:
>>>
>>> /sys/class/typec/<port>/supported_power_roles
>>> /sys/class/typec/<port>/preferred_role
>>>
>>> The need of preferred role is required when DRP is set in
>>> supported_power_roles option.
>>> Ideally a battery powered device will TRY to be SNK and a a/c plugged
>>> device will TRY to be SRC
>>>
>>> We need to understand which non-PD device will set to DRP? In the
>>
>> Android Phones (actually it could be any phone which has a type-c port)
>> since it can act as usb gadget (when connected to PC) or Usb host
>> when connected to peripherals such as thumb drives, keyboard etc.
>> Phones with smaller form factors might be thermally limited to charge
>> above 15W, therefore supporting PD might be an overkill for them.
>>
>>> current ecosystem all legacy devices
>>> will sit behind adapters which either present an Rp or Rd.
>>>
>>> If it is a power adapter in 5V range can either present Rp or DRP with
>>> TRY.SRC and there is no role swap requirement.
>>>
>>> If it is a laptop port or similar with non-PD (??) DRP there is no
>>> guaranteed role swap in a non-PD mode.
>>
>> This is true, but following a Try.SRC or Try.SNK state machine can
>> increase the chances of landing in the desired role/preferred role.
>
> Agree and as indicated it increases only chances.
>
FWIW, this is pretty much the same as requesting a role change using PD.
Based on my experience with various devices, the chance for that to succeed
isn't really that high either.
I am not really sure I understand your problem with using Try.{SRC,SNK}
to trigger (or attempt to trigger) a role change. Can we take a step back,
and can you explain ?
Thanks,
Guenter
On Tue, Apr 25, 2017 at 7:40 PM, Guenter Roeck <[email protected]> wrote:
> On 04/25/2017 01:26 AM, Rajaram R wrote:
>>
>> On Mon, Apr 24, 2017 at 11:20 PM, Badhri Jagan Sridharan
>> <[email protected]> wrote:
>>>
>>> On Sat, Apr 22, 2017 at 2:23 AM, Rajaram R <[email protected]>
>>> wrote:
>>>>
>>>> On Fri, Apr 21, 2017 at 10:13 PM, Guenter Roeck <[email protected]>
>>>> wrote:
>>>>>
>>>>> On Fri, Apr 21, 2017 at 07:57:52PM +0530, Rajaram R wrote:
>>>>>>
>>>>>> On Fri, Apr 21, 2017 at 1:16 AM, Badhri Jagan Sridharan
>>>>>> <[email protected]> wrote:
>>>>>>>
>>>>>>> Thanks for the responses :)
>>>>>>>
>>>>>>> So seems like we have a plan.
>>>>>>>
>>>>>>> In Type-C connector class the checks for TYPEC_PWR_MODE_PD
>>>>>>> and pd_revision for both the port and the partner will be removed in
>>>>>>> power_role_store and the data_role_store and will be delegated
>>>>>>> to the low level drivers.
>>>>>>
>>>>>>
>>>>>> It is important to remember what USB Type-C provide is mechanisms for
>>>>>> "TRYing" to become a particular role and not guaranteeing.
>>>>>>
>>>>>> With what device combination do you fore see we could get the desired
>>>>>> role with this change ?
>>>>>>
>>>>>
>>>>> If the partner is not PD capable, if a preferred role is specified,
>>>>> if the current cole does not match the preferred role, and if the
>>>>> request
>>>>> is to set the role to match the preferred role, I think it is
>>>>> reasonable
>>>>> to expect that re-establishing the connection would accomplish that if
>>>>> the
>>>>> partner supports it.
>>>>>
>>>> In this context I believe we have two different inputs as follows:
>>>>
>>>> /sys/class/typec/<port>/supported_power_roles
>>>> /sys/class/typec/<port>/preferred_role
>>>>
>>>> The need of preferred role is required when DRP is set in
>>>> supported_power_roles option.
>>>> Ideally a battery powered device will TRY to be SNK and a a/c plugged
>>>> device will TRY to be SRC
>>>>
>>>> We need to understand which non-PD device will set to DRP? In the
>>>
>>>
>>> Android Phones (actually it could be any phone which has a type-c port)
>>> since it can act as usb gadget (when connected to PC) or Usb host
>>> when connected to peripherals such as thumb drives, keyboard etc.
>>> Phones with smaller form factors might be thermally limited to charge
>>> above 15W, therefore supporting PD might be an overkill for them.
>>>
>>>> current ecosystem all legacy devices
>>>> will sit behind adapters which either present an Rp or Rd.
>>>>
>>>> If it is a power adapter in 5V range can either present Rp or DRP with
>>>> TRY.SRC and there is no role swap requirement.
>>>>
>>>> If it is a laptop port or similar with non-PD (??) DRP there is no
>>>> guaranteed role swap in a non-PD mode.
>>>
>>>
>>> This is true, but following a Try.SRC or Try.SNK state machine can
>>> increase the chances of landing in the desired role/preferred role.
>>
>>
>> Agree and as indicated it increases only chances.
>>
>
> FWIW, this is pretty much the same as requesting a role change using PD.
> Based on my experience with various devices, the chance for that to succeed
> isn't really that high either.
>
> I am not really sure I understand your problem with using Try.{SRC,SNK}
> to trigger (or attempt to trigger) a role change. Can we take a step back,
> and can you explain ?
The parameters required for a type-c connection is defined as follows
and will have a default value.
/sys/class/typec/<port>/supported_power_roles
/sys/class/typec/<port>/preferred_role
When two DRP devices are connected and for which we have
preferred_role which provides input on the preference, In a DRP mode
we have randomness in the mode of connection and hence we require role
swap mechanisms. A Type-C only device cannot role swap as this is
valid only in PD operation.
# Question was how to choose a particular role in non-PD mode. Only
way to have a deterministic role in a non-PD mode is to set expected
supported_role of choice rather than DRP.
# As part of the solution suggested, checking of roles and triggering
role swaps has to be done by the policy manager(PM) and delinked from
Policy Engine. I guess the policy manager is at user space?.
I do not have the complete git repo and may be i could be missing
something. If this is in any public git please let me know
>
> Thanks,
> Guenter
>
On Thu, Apr 27, 2017 at 11:50:12AM +0530, Rajaram R wrote:
> On Tue, Apr 25, 2017 at 7:40 PM, Guenter Roeck <[email protected]> wrote:
> > On 04/25/2017 01:26 AM, Rajaram R wrote:
> >>
> >> On Mon, Apr 24, 2017 at 11:20 PM, Badhri Jagan Sridharan
> >> <[email protected]> wrote:
> >>>
> >>> On Sat, Apr 22, 2017 at 2:23 AM, Rajaram R <[email protected]>
> >>> wrote:
> >>>>
> >>>> On Fri, Apr 21, 2017 at 10:13 PM, Guenter Roeck <[email protected]>
> >>>> wrote:
> >>>>>
> >>>>> On Fri, Apr 21, 2017 at 07:57:52PM +0530, Rajaram R wrote:
> >>>>>>
> >>>>>> On Fri, Apr 21, 2017 at 1:16 AM, Badhri Jagan Sridharan
> >>>>>> <[email protected]> wrote:
> >>>>>>>
> >>>>>>> Thanks for the responses :)
> >>>>>>>
> >>>>>>> So seems like we have a plan.
> >>>>>>>
> >>>>>>> In Type-C connector class the checks for TYPEC_PWR_MODE_PD
> >>>>>>> and pd_revision for both the port and the partner will be removed in
> >>>>>>> power_role_store and the data_role_store and will be delegated
> >>>>>>> to the low level drivers.
> >>>>>>
> >>>>>>
> >>>>>> It is important to remember what USB Type-C provide is mechanisms for
> >>>>>> "TRYing" to become a particular role and not guaranteeing.
> >>>>>>
> >>>>>> With what device combination do you fore see we could get the desired
> >>>>>> role with this change ?
> >>>>>>
> >>>>>
> >>>>> If the partner is not PD capable, if a preferred role is specified,
> >>>>> if the current cole does not match the preferred role, and if the
> >>>>> request
> >>>>> is to set the role to match the preferred role, I think it is
> >>>>> reasonable
> >>>>> to expect that re-establishing the connection would accomplish that if
> >>>>> the
> >>>>> partner supports it.
> >>>>>
> >>>> In this context I believe we have two different inputs as follows:
> >>>>
> >>>> /sys/class/typec/<port>/supported_power_roles
> >>>> /sys/class/typec/<port>/preferred_role
> >>>>
> >>>> The need of preferred role is required when DRP is set in
> >>>> supported_power_roles option.
> >>>> Ideally a battery powered device will TRY to be SNK and a a/c plugged
> >>>> device will TRY to be SRC
> >>>>
> >>>> We need to understand which non-PD device will set to DRP? In the
> >>>
> >>>
> >>> Android Phones (actually it could be any phone which has a type-c port)
> >>> since it can act as usb gadget (when connected to PC) or Usb host
> >>> when connected to peripherals such as thumb drives, keyboard etc.
> >>> Phones with smaller form factors might be thermally limited to charge
> >>> above 15W, therefore supporting PD might be an overkill for them.
> >>>
> >>>> current ecosystem all legacy devices
> >>>> will sit behind adapters which either present an Rp or Rd.
> >>>>
> >>>> If it is a power adapter in 5V range can either present Rp or DRP with
> >>>> TRY.SRC and there is no role swap requirement.
> >>>>
> >>>> If it is a laptop port or similar with non-PD (??) DRP there is no
> >>>> guaranteed role swap in a non-PD mode.
> >>>
> >>>
> >>> This is true, but following a Try.SRC or Try.SNK state machine can
> >>> increase the chances of landing in the desired role/preferred role.
> >>
> >>
> >> Agree and as indicated it increases only chances.
> >>
> >
> > FWIW, this is pretty much the same as requesting a role change using PD.
> > Based on my experience with various devices, the chance for that to succeed
> > isn't really that high either.
> >
> > I am not really sure I understand your problem with using Try.{SRC,SNK}
> > to trigger (or attempt to trigger) a role change. Can we take a step back,
> > and can you explain ?
>
> The parameters required for a type-c connection is defined as follows
> and will have a default value.
>
> /sys/class/typec/<port>/supported_power_roles
I don't see that attribute (it is implied in the power_role attribute).
It only existed in an earlier version of the driver.
Also, the attribute (when it existed) listed the roles supported
by the port, not an administrative restriction of supported roles.
> /sys/class/typec/<port>/preferred_role
>
> When two DRP devices are connected and for which we have
> preferred_role which provides input on the preference, In a DRP mode
> we have randomness in the mode of connection and hence we require role
> swap mechanisms. A Type-C only device cannot role swap as this is
> valid only in PD operation.
>
The same is true for non-PD connections. Also, PD doesn't solve the problem
if both ends have the same source/sink preference. The result is still
randomness. To resolve that randomness, one of the connection partners
would have to give up its preference. That is quite independent of PD
support.
> # Question was how to choose a particular role in non-PD mode. Only
> way to have a deterministic role in a non-PD mode is to set expected
> supported_role of choice rather than DRP.
>
> # As part of the solution suggested, checking of roles and triggering
> role swaps has to be done by the policy manager(PM) and delinked from
> Policy Engine. I guess the policy manager is at user space?.
>
> I do not have the complete git repo and may be i could be missing
> something. If this is in any public git please let me know
>
Even if a device does support PD, there is no guarantee that a power role
swap is successful. That depends on the policy set by the partner, and
for all practical purposes it also depends on the quality and stability
of the PD protocol implementation on both ends. As mentioned above, if
both ends have the same preferred role, the result is still randomness
and requires manual intervention by the user if the "other" role is
desired.
It seems to me that we are discussing if we should do as good as we can
or if we should only implement what can be guaranteed. I strongly favor
the first approach. I also think we should not preclude that approach in
the kernel implementation. It should be left to a specific policy manager
implementation to decide if a given device should be a DRP or source-only
or sink-only, independent of the question if PD is supported on either
end of a connection.
If a given implementation wants to restrict available port modes, it
can do so by providing acceptable port capabilities (ie DRP or source
or sink) when registering the port with the infrastructure.
I think you are suggesting that there should be a means to override port
capabilities (more specifically, permitted roles) after port registration.
That may be a matter for discussion, but it is a separate issue.
Also, I am not sure if it was already discussed. Does anyone remember ?
Thanks,
Guenter
On Thu, Apr 27, 2017 at 11:10:55AM -0700, Guenter Roeck wrote:
> On Thu, Apr 27, 2017 at 11:50:12AM +0530, Rajaram R wrote:
> > On Tue, Apr 25, 2017 at 7:40 PM, Guenter Roeck <[email protected]> wrote:
> > > On 04/25/2017 01:26 AM, Rajaram R wrote:
> > >>
> > >> On Mon, Apr 24, 2017 at 11:20 PM, Badhri Jagan Sridharan
> > >> <[email protected]> wrote:
> > >>>
> > >>> On Sat, Apr 22, 2017 at 2:23 AM, Rajaram R <[email protected]>
> > >>> wrote:
> > >>>>
> > >>>> On Fri, Apr 21, 2017 at 10:13 PM, Guenter Roeck <[email protected]>
> > >>>> wrote:
> > >>>>>
> > >>>>> On Fri, Apr 21, 2017 at 07:57:52PM +0530, Rajaram R wrote:
> > >>>>>>
> > >>>>>> On Fri, Apr 21, 2017 at 1:16 AM, Badhri Jagan Sridharan
> > >>>>>> <[email protected]> wrote:
> > >>>>>>>
> > >>>>>>> Thanks for the responses :)
> > >>>>>>>
> > >>>>>>> So seems like we have a plan.
> > >>>>>>>
> > >>>>>>> In Type-C connector class the checks for TYPEC_PWR_MODE_PD
> > >>>>>>> and pd_revision for both the port and the partner will be removed in
> > >>>>>>> power_role_store and the data_role_store and will be delegated
> > >>>>>>> to the low level drivers.
> > >>>>>>
> > >>>>>>
> > >>>>>> It is important to remember what USB Type-C provide is mechanisms for
> > >>>>>> "TRYing" to become a particular role and not guaranteeing.
> > >>>>>>
> > >>>>>> With what device combination do you fore see we could get the desired
> > >>>>>> role with this change ?
> > >>>>>>
> > >>>>>
> > >>>>> If the partner is not PD capable, if a preferred role is specified,
> > >>>>> if the current cole does not match the preferred role, and if the
> > >>>>> request
> > >>>>> is to set the role to match the preferred role, I think it is
> > >>>>> reasonable
> > >>>>> to expect that re-establishing the connection would accomplish that if
> > >>>>> the
> > >>>>> partner supports it.
> > >>>>>
> > >>>> In this context I believe we have two different inputs as follows:
> > >>>>
> > >>>> /sys/class/typec/<port>/supported_power_roles
> > >>>> /sys/class/typec/<port>/preferred_role
> > >>>>
> > >>>> The need of preferred role is required when DRP is set in
> > >>>> supported_power_roles option.
> > >>>> Ideally a battery powered device will TRY to be SNK and a a/c plugged
> > >>>> device will TRY to be SRC
> > >>>>
> > >>>> We need to understand which non-PD device will set to DRP? In the
> > >>>
> > >>>
> > >>> Android Phones (actually it could be any phone which has a type-c port)
> > >>> since it can act as usb gadget (when connected to PC) or Usb host
> > >>> when connected to peripherals such as thumb drives, keyboard etc.
> > >>> Phones with smaller form factors might be thermally limited to charge
> > >>> above 15W, therefore supporting PD might be an overkill for them.
> > >>>
> > >>>> current ecosystem all legacy devices
> > >>>> will sit behind adapters which either present an Rp or Rd.
> > >>>>
> > >>>> If it is a power adapter in 5V range can either present Rp or DRP with
> > >>>> TRY.SRC and there is no role swap requirement.
> > >>>>
> > >>>> If it is a laptop port or similar with non-PD (??) DRP there is no
> > >>>> guaranteed role swap in a non-PD mode.
> > >>>
> > >>>
> > >>> This is true, but following a Try.SRC or Try.SNK state machine can
> > >>> increase the chances of landing in the desired role/preferred role.
> > >>
> > >>
> > >> Agree and as indicated it increases only chances.
> > >>
> > >
> > > FWIW, this is pretty much the same as requesting a role change using PD.
> > > Based on my experience with various devices, the chance for that to succeed
> > > isn't really that high either.
> > >
> > > I am not really sure I understand your problem with using Try.{SRC,SNK}
> > > to trigger (or attempt to trigger) a role change. Can we take a step back,
> > > and can you explain ?
> >
> > The parameters required for a type-c connection is defined as follows
> > and will have a default value.
> >
> > /sys/class/typec/<port>/supported_power_roles
>
> I don't see that attribute (it is implied in the power_role attribute).
> It only existed in an earlier version of the driver.
>
> Also, the attribute (when it existed) listed the roles supported
> by the port, not an administrative restriction of supported roles.
>
> > /sys/class/typec/<port>/preferred_role
> >
> > When two DRP devices are connected and for which we have
> > preferred_role which provides input on the preference, In a DRP mode
> > we have randomness in the mode of connection and hence we require role
> > swap mechanisms. A Type-C only device cannot role swap as this is
> > valid only in PD operation.
> >
>
> The same is true for non-PD connections. Also, PD doesn't solve the problem
> if both ends have the same source/sink preference. The result is still
> randomness. To resolve that randomness, one of the connection partners
> would have to give up its preference. That is quite independent of PD
> support.
>
> > # Question was how to choose a particular role in non-PD mode. Only
> > way to have a deterministic role in a non-PD mode is to set expected
> > supported_role of choice rather than DRP.
> >
> > # As part of the solution suggested, checking of roles and triggering
> > role swaps has to be done by the policy manager(PM) and delinked from
> > Policy Engine. I guess the policy manager is at user space?.
> >
> > I do not have the complete git repo and may be i could be missing
> > something. If this is in any public git please let me know
> >
>
> Even if a device does support PD, there is no guarantee that a power role
> swap is successful. That depends on the policy set by the partner, and
> for all practical purposes it also depends on the quality and stability
> of the PD protocol implementation on both ends. As mentioned above, if
> both ends have the same preferred role, the result is still randomness
> and requires manual intervention by the user if the "other" role is
> desired.
>
> It seems to me that we are discussing if we should do as good as we can
> or if we should only implement what can be guaranteed. I strongly favor
> the first approach. I also think we should not preclude that approach in
> the kernel implementation. It should be left to a specific policy manager
> implementation to decide if a given device should be a DRP or source-only
> or sink-only, independent of the question if PD is supported on either
> end of a connection.
>
> If a given implementation wants to restrict available port modes, it
> can do so by providing acceptable port capabilities (ie DRP or source
> or sink) when registering the port with the infrastructure.
>
> I think you are suggesting that there should be a means to override port
> capabilities (more specifically, permitted roles) after port registration.
> That may be a matter for discussion, but it is a separate issue.
That is how also I see it.
> Also, I am not sure if it was already discussed. Does anyone remember ?
Yes, I did propose an attribute that would allow fixing the role of a
port to either sink or source at one point. I don't remember what was
the reason I ended up dropping that proposal - I need to go through my
archives - but I think it was seen as a feature that, if needed, can be
added later by adding a new attribute file for it.
So if that is the functionality you want, please start a new thread
and propose the new attribute. At least don't try to modify the
behaviour of the existing ones because of it.
Thank you,
--
heikki