On 01/31/2017 03:21 AM, Mika Westerberg wrote:
On Mon, Jan 30, 2017 at 04:52:27PM -0800, Dmitry Torokhov wrote:
On Sat, Jan 7, 2017 at 10:14 AM, Andy Shevchenko
<andriy.shevchenko@xxxxxxxxxxxxxxx> wrote:
On Fri, 2017-01-06 at 02:30 -0700, Robert R. Howell wrote:
[1.] Baytrail tablet touchscreen fails after pinctrl commit
[3ae02c14d964499da76b5ef4d8912710902e64f8]
[2.] Beginning with kernel 4.10-rc1 (and continuing with 4.10-rc2)
the
touchscreen on two Baytrail devices, an ASUS T100TA and a Toshiba
Encore
8, no longer works. They did work properly up though at least 4.9.0.
As
indicated in the DMESG output appended below, the ATML1000
touchscreen
controller fails to reset and initialize properly on boot. The reset
command issued through i2c_hid appears to time out after its 5 second
wait for completion. (There are similar problems with an i2c
orientation sensor.) Using git bisect I've found that the problem
begins with commit [3ae02c14d964499da76b5ef4d8912710902e64f8]
pinctrl: intel: set default handler to be handle_bad_irq()
I've appended the git bisect log output at the end. Reverting that
single commit (tested with 4.10-rc2) fixes the problem.
Thanks for the report. For me it looks like that commit actually
discovers other problems we have in the code somewhere.
Yeah, I confirm the issue on T100TA. I will investigate this next week.
This still happens on chromeOS Baytrail devices with -rc6 here as
well. Can we revert this and try again in 4.11 timeframe?
Are you sure you have commit 49c030962638 ("pinctrl: baytrail: Do not
add all GPIOs to IRQ domain")? That is supposed to fix the issue if I
understand correctly.
I've just tested 4.10-rc6 on an ASUS T100-TAM, a T100-TA, and a Toshiba
Encore 8 and the i2c connected touchscreen is working on all of them.
RC-6 DOES incorporate Andy Shevchenko's "[PATCH v1 1/2] pinctrl:
baytrail: Do not add all GPIOs to IRQ domain", which is what seems to
fix the problem.
In tests a couple weeks ago on 4.10-rc4, before that patch was added to
the mainline kernel, I did manually apply that patch (plus also his
"[PATCH v1 2/2] pinctrl: baytrail: Convert to use devm_*()" to rc4 and
that also produced a kernel where the touchscreen worked. It looks like
that second patch isn't incorporated in rc6, but apparently isn't
critical to fixing the touchscreen issue.
Bob Howell
--
To unsubscribe from this list: send the line "unsubscribe linux-input" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
