On Thu, Nov 25, 2021 at 08:16:12AM -0300, Fabio Estevam wrote:

Hi Uwe,

On Thu, Nov 25, 2021 at 5:34 AM Uwe Kleine-König
[off-list ref] wrote:

quoted

With an old style device tree using fsl,usbphy
devm_usb_get_phy_by_phandle() returning ERR_PTR(-EPROBE_DEFER) results in
ci->usb_phy = ERR_PTR(-EPROBE_DEFER) in the chipidea driver which then
chokes on that with

        Unable to handle kernel paging request at virtual address fffffe93

Handle errors other then -ENODEV as was done before v5.15-rc5 in
ci_hdrc_imx_probe().

Fixes: 8253a34bfae3 ("usb: chipidea: ci_hdrc_imx: Also search for 'phys' phandle")
Cc: stable@vger.kernel.org
Signed-off-by: Uwe Kleine-König <redacted>
---
Hello,

this patch became commit 8253a34bfae3278baca52fc1209b7c29270486ca in
v5.15-rc5. On an i.MX25 I experience the following fault:

[    2.248749] 8<--- cut here ---
[    2.259025] Unable to handle kernel paging request at virtual address fffffe93

Sorry for the breakage.

Dan has sent a fix for this:
https://www.spinics.net/lists/linux-usb/msg219148.html

Ah, I see this is already in next as
d4d2e5329ae9dfd6742c84d79f7d143d10410f1b via the usb tree. I know why I
missed that when I looked for fixes, but I spare you from the details.

Thanks and sorry for the noise,
Uwe

-- 
Pengutronix e.K.                           | Uwe Kleine-König            |
Industrial Linux Solutions                 | https://www.pengutronix.de/ |