ipq8064: fix dwc3-of-simple module unloading
authorThomas Reifferscheid <thomas@reifferscheid.org>
Mon, 27 Mar 2017 22:01:59 +0000 (00:01 +0200)
committerJohn Crispin <john@phrozen.org>
Tue, 28 Mar 2017 07:03:55 +0000 (09:03 +0200)
commit17f60b1cd260a24ef990d6622f9c5ed6951c0722
tree79700fe7db283baa5faf2761cdf312cf19cf0092
parentc75f059b0c4d09dd0da60e14c4933a9f645266d1
ipq8064: fix dwc3-of-simple module unloading

Without patch unloading the dwc3-of-simple module went stuck after
successfully removing hcd.1 during the hcd.0 removal:

root@LEDE:/# rmmod dwc3-of-simple
[   21.391846] xhci-hcd xhci-hcd.1.auto: remove, state 1
[   21.391931] usb usb4: USB disconnect, device number 1
[   21.397038] xhci-hcd xhci-hcd.1.auto: USB bus 4 deregistered
[   21.401111] xhci-hcd xhci-hcd.1.auto: remove, state 1
[   21.406685] usb usb3: USB disconnect, device number 1
[   21.412848] xhci-hcd xhci-hcd.1.auto: USB bus 3 deregistered
[   21.417248] xhci-hcd xhci-hcd.0.auto: remove, state 1
[   21.422521] usb usb2: USB disconnect, device number 1
followed by nothing.

Sometimes a stall CPU was detected, or a kernel panic,
or a reboot occurred after a couple of minutes.

At the same time unloading the dwc3 module followed by dwc3-of-simple
module was working repeatedly.

root@LEDE:/# rmmod dwc3
[   53.827328] xhci-hcd xhci-hcd.1.auto: remove, state 1
[   53.827412] usb usb4: USB disconnect, device number 1
[   53.832630] xhci-hcd xhci-hcd.1.auto: USB bus 4 deregistered
[   53.836452] xhci-hcd xhci-hcd.1.auto: remove, state 1
[   53.842314] usb usb3: USB disconnect, device number 1
[   53.848412] xhci-hcd xhci-hcd.1.auto: USB bus 3 deregistered
[   53.852542] xhci-hcd xhci-hcd.0.auto: remove, state 1
[   53.857882] usb usb2: USB disconnect, device number 1
[   53.863956] xhci-hcd xhci-hcd.0.auto: USB bus 2 deregistered
[   53.867875] xhci-hcd xhci-hcd.0.auto: remove, state 1
[   53.873696] usb usb1: USB disconnect, device number 1
[   53.879742] xhci-hcd xhci-hcd.0.auto: USB bus 1 deregistered
root@LEDE:/# rmmod dwc3-of-simple
root@LEDE:/#

For the non-working case, the code was stuck in a readl() in
http://lxr.free-electrons.com/source/drivers/usb/host/xhci.c#L91
because
http://lxr.free-electrons.com/source/drivers/usb/dwc3/dwc3-of-simple.c#L126
was disabling the wrong clocks when removing hcd.1 (it was disabling
the clock of hcd.0). That's why the readl() went stuck when removing
hcd.0

The patch however addresses the clock assignment from the .dtsi
file. Most probably it went into openwrt here:
https://dev.openwrt.org/browser/trunk/target/linux/ipq806x/patches-3.18/101-ARM-qcom-add-USB-nodes-to-ipq806x-ap148.patch?rev=45261
copied from Qualcomms attempt here: https://lkml.org/lkml/2015/11/20/116

Now unloading and repeated module loading is working just fine,
no matter if you'd remove dwc3-of-simple or dwc3.

root@LEDE:/# rmmod dwc3-of-simple
[   24.089679] xhci-hcd xhci-hcd.1.auto: remove, state 1
[   24.089765] usb usb4: USB disconnect, device number 1
[   24.094856] xhci-hcd xhci-hcd.1.auto: USB bus 4 deregistered
[   24.098963] xhci-hcd xhci-hcd.1.auto: remove, state 1
[   24.104522] usb usb3: USB disconnect, device number 1
[   24.111194] xhci-hcd xhci-hcd.1.auto: USB bus 3 deregistered
[   24.115086] xhci-hcd xhci-hcd.0.auto: remove, state 1
[   24.120396] usb usb2: USB disconnect, device number 1
[   24.126503] xhci-hcd xhci-hcd.0.auto: USB bus 2 deregistered
[   24.130347] xhci-hcd xhci-hcd.0.auto: remove, state 1
[   24.135948] usb usb1: USB disconnect, device number 1
[   24.142085] xhci-hcd xhci-hcd.0.auto: USB bus 1 deregistered
root@LEDE:/#

Fixes: dwc3-of-simple module unloading
Signed-off-by: Thomas Reifferscheid <thomas@reifferscheid.org>
target/linux/ipq806x/files-4.9/arch/arm/boot/dts/qcom-ipq8064.dtsi