On Wed, Nov 29, 2017 at 3:50 AM, <Mario.Limonciello@xxxxxxxx> wrote: >> -----Original Message----- >> From: platform-driver-x86-owner@xxxxxxxxxxxxxxx [mailto:platform-driver-x86- >> owner@xxxxxxxxxxxxxxx] On Behalf Of Andy Shevchenko >> Sent: Tuesday, November 28, 2017 1:28 PM >> To: Alex Hung <alex.hung@xxxxxxxxxxxxx> >> Cc: Jason Gerecke <killertofu@xxxxxxxxx>; platform-driver-x86@xxxxxxxxxxxxxxx; >> Michal Kepien <kernel@xxxxxxxxxx> >> Subject: Re: Intel "5-button array" support on MobileStudio Pro >> >> On Mon, Nov 27, 2017 at 3:42 PM, Alex Hung <alex.hung@xxxxxxxxxxxxx> wrote: >> > On Sun, Nov 26, 2017 at 2:28 AM, Jason Gerecke <killertofu@xxxxxxxxx> wrote: >> >> On November 25, 2017 10:23:33 AM PST, Jason Gerecke >> <killertofu@xxxxxxxxx> wrote: >> >>>I've just filed a bug [1] about a device which appears to use the >> >>>"5-button array" that the patch [2] added support for earlier this >> >>>year. It seems that on this particular device (a Wacom MobileStudio >> >>>Pro), there is no "HEBC" entry in the ACPI table. Because of this, the >> >>>kernel does not load support for the "5-button array" that this device >> >>>seems to use. >> > >> > HEBC seems to be the control method that can report 5 button array; >> > however, I think we can use either one of the below to workaround >> > this: >> > >> > 1. Add a kernel parameter and a quirk (ex. DMI strings). The downside >> > is the quirk can grow if more systems do not include HEBC >> > >> > 2. Check both HEBC & BTNL (ex. the method that enables 5 button >> > array). Either one enables 5 button array and this fixes current >> > problems. >> > >> > While 2 is preferable, nothing stops future BIOS skips BTNL in the future... >> > >> > Any comments? I will create a patch for testing in a few days. >> >> Looking for a tables with HEBC and this one I only can tell that I >> would assume that tables w/o HEBC can guess HEBC returns value for >> supporting 5-button array. >> Though to be on the safe side I would rather vote for DMI based quirk After thinking through again I think staying safe with DMI quirk may be a good idea. 1. Only one system needs workaround now 2. Future systems may behave differently and we will have more test samples Jason, Please upload output of dmidecode to https://bugzilla.kernel.org/show_bug.cgi?id=197991 and I can modify intel-hid accordingly. This may take couple days before I can find a system to test. > > Andy, > > Could you provide any insight how this possibly works on Windows? > I would expect that corner cases like this should completely fall over > on Windows since it's an Intel driver and not customized per > OEM or anything. > > Maybe would it be possible for the folks at Intel > who worked on the Windows driver to share information about anything > different they're doing with regards to something like a possibly missing > HEBC? It would be good to do the same rather than bandage on a bunch > of things tied to DMI quirks. If it's not possible to obtain that information, > then I would agree DMI quirks makes sense though. If we can refer how it is done in Windows we can improve it but let's try DMI quirk first. > > Thanks, > >> >> > >> >>>Please let me know what additional information would be helpful to >> >>>provide. I am able to test kernel patches. >> >>> >> >>>[1]: https://bugzilla.kernel.org/show_bug.cgi?id=197991 >> >>> >> >>>[2]: https://patchwork.kernel.org/patch/9571353/ >> >> >>>Now instead of four in the eights place / >> >>>you’ve got three, ‘Cause you added one / >> >>>(That is to say, eight) to the two, / >> >>>But you can’t take seven from three, / >> >>>So you look at the sixty-fours.... >> >> >> -- >> With Best Regards, >> Andy Shevchenko -- Cheers, Alex Hung