log and apple dongle info:
https://pantheon.corp.google.com/storage/browser/chromiumos-test-logs/bugfiles/cr/709183

Can you please try M-58?  Guenter made a whole lot of fixes there addressing races in the way dwc3 (usb controller) was brought up and torn down.

A dogfooder was reproducing problems similar to this on M-57 with a specific cable that was really flaky (so it acted like plugs / unplugs all the time).  When we tried on M-58 it was all better.

It seemed unwise to pick all his changes to M-57 and risk introducing some other problem, so users will just have to wait till M-58 to get it fixed.

Reference on Kevin Dogfooder mailing list, see subject line "Bootloop for external monitor"

I found another Kevin in 9334.33.0, 58.0.3029.51 (Monday's build) and test it in the same scenarios. Unplug-re-plug Apple type-c dongle with hdmi cause device screen ( internal screen) lost its display ( can see mouse cursor on screen only). External monitor has its display.  

dianders@, should I install latest M58 build to double check? 

My changes were not related to HDMI. Given #4 I suspect that the problem may be on the video side. Also, I am not sure if anyone ever tested with two Apple dongles. We'll have to analyze the crash logs to see if those are USB related crashes.

groeck@, it's just one apple dongle, not two. 

Guenter: USB can still cause the crashes.  These cables often enumerate in a way that causes the USB ports to be created / deleted.  Certainly with HoHo that was the case.  

One feedback report from the dogfooder discussion (with HoHo) was 55529782918.  In that:

[    5.941772] xhci-hcd xhci-hcd.2.auto: xHCI Host Controller
[    5.942064] xhci-hcd xhci-hcd.2.auto: new USB bus registered, assigned bus number 5
[    5.942398] xhci-hcd xhci-hcd.2.auto: hcc params 0x0220fe64 hci version 0x110 quirks 0x02030010
[    5.942457] xhci-hcd xhci-hcd.2.auto: irq 232, io mem 0xfe900000
[    5.942654] usb usb5: New USB device found, idVendor=1d6b, idProduct=0002
[    5.942665] usb usb5: New USB device strings: Mfr=3, Product=2, SerialNumber=1
[    5.942673] usb usb5: Product: xHCI Host Controller
[    5.942680] usb usb5: Manufacturer: Linux 4.4.35-06796-g0d8baba xhci-hcd
[    5.942687] usb usb5: SerialNumber: xhci-hcd.2.auto
[    5.944469] hub 5-0:1.0: USB hub found
[    5.944524] hub 5-0:1.0: 1 port detected
[    5.944928] xhci-hcd xhci-hcd.2.auto: xHCI Host Controller
[    5.945903] xhci-hcd xhci-hcd.2.auto: new USB bus registered, assigned bus number 6
[    5.946013] usb usb6: We don't know the algorithms for LPM for this host, disabling LPM.
[    5.946149] usb usb6: New USB device found, idVendor=1d6b, idProduct=0003
[    5.946158] usb usb6: New USB device strings: Mfr=3, Product=2, SerialNumber=1
[    5.946166] usb usb6: Product: xHCI Host Controller
[    5.946173] usb usb6: Manufacturer: Linux 4.4.35-06796-g0d8baba xhci-hcd
[    5.947262] usb usb6: SerialNumber: xhci-hcd.2.auto
[    5.948672] hub 6-0:1.0: USB hub found
[    5.948725] hub 6-0:1.0: 1 port detected
[    5.950375] rockchip-dwc3 usb@fe900000: USB HOST connected
[    5.950569] xhci-hcd xhci-hcd.2.auto: remove, state 1
[    5.950595] usb usb6: USB disconnect, device number 1
[    5.952413] xhci-hcd xhci-hcd.2.auto: USB bus 6 deregistered
[    5.952640] xhci-hcd xhci-hcd.2.auto: remove, state 1
[    5.952669] usb usb5: USB disconnect, device number 1

--

Helen: I'm not sure I understand @4.  I think you're saying that there is no crash on R58 but that you had some other graphics issues?

Doug, you are right. M58 does not have crash issue. 

I will file another bug to address M58 display issue. 

Thanks. 

OK, thanks.  Closing this as there are no plans to pick Guenter's fixes to M-57.