On Fri, Feb 16, 2018 at 7:20 PM, Kishon Vijay Abraham I kishon@ti.com wrote:

On Friday 19 January 2018 08:15 PM, Maxime Ripard wrote:

...

On Fri, Jan 19, 2018 at 05:25:41PM +0800, Chen-Yu Tsai wrote:

...

The AXP223 PMIC, like the AXP221, does not generate VBUS change interrupts when N_VBUSEN is used to drive VBUS for the OTG port on the board.

This was not noticed until recently, as most A23/A33 boards use a GPIO pin that does not support interrupts for OTG ID detection. This forces the driver to use polling. However the A33-OlinuXino uses a pin that does support interrupts, so the driver uses them. However the VBUS interrupt never fires, and the driver never gets to update the VBUS status. This results in musb timing out waiting for VBUS to rise.

This was worked around for the AXP221 by resorting to polling changes in commit 91d96f06a760 ("phy-sun4i-usb: Add workaround for missing Vbus det interrupts on A31"). This patch adds the A23 and A33 to the list of SoCs that need the workaround.

Fixes: fc1f45ed3043 ("phy-sun4i-usb: Add support for the usb-phys on the sun8i-a33 SoC") Fixes: 123dfdbcfaf5 ("phy-sun4i-usb: Add support for the usb-phys on the sun8i-a23 SoC") Cc: stable@vger.kernel.org # 4.3.x: 68dbc2ce77bb phy-sun4i-usb: Use of_match_node to get model specific config data Cc: stable@vger.kernel.org # 4.3.x: 5cf700ac9d50 phy: phy-sun4i-usb: Fix optional gpios failing probe Cc: stable@vger.kernel.org # 4.3.x: 04e59a0211ff phy-sun4i-usb: Fix irq free conditions to match request conditions Cc: stable@vger.kernel.org # 4.3.x: 91d96f06a760 phy-sun4i-usb: Add workaround for missing Vbus det interrupts on A31 Cc: stable@vger.kernel.org # 4.3.x Signed-off-by: Chen-Yu Tsai wens@csie.org

---

This list might get longer if the newer AXP8xx PMICs also have this behavior. We could switch to a boolean in the per-compatible data, or just always use polling. The yet-merged R40 support will have this issue as well, as the R40 is paired with the AXP221. But for now, I'd like to have something that is easier to backport to stable, so we can at least fix this for the A23 and A33.

Also, checkpatch.pl doesn't like the stable kernel prerequisite lines.

Acked-by: Maxime Ripard maxime.ripard@free-electrons.com

merged, thanks!

Thanks.

After some more experimentation on the A83T/AXP813, it seems this is pretty standard behavior for the AXP PMICs. If one uses N_VBUSEN to enable VBUS, the user/system should already know that VBUS is on or present. The VBUS-related logic in the PMIC, save for the VBUS status bits, is disconnected in this case.

So it looks like we either enable this workaround for most if not all of the Allwinner family, or fix this by polling in the power supply driver itself. This would require polling once every 50ms, which might be too much work for the system, but is necessary for VBUS detection to work properly for MUSB. Or we could limit polling to whenever the drivevbus regulator is enabled. This would involve creating a reference between the USB power supply and regulator in the device tree. (There doesn't seem to be a way to fetch a regulator directly by its device node.) The power supply would then register a notifier on the regulator for on/off events, and configure polling accordingly.

Any suggestions?

ChenYu