Re: Dell Inspiron 5558/0VNM2T hangs at resume from suspend when USB 3 is enabled

[Diego Viola <diego.viola@gmail.com>]

On Thu, Mar 16, 2017 at 1:02 PM, Diego Viola <diego.viola@gmail.com> wrote:
> On Thu, Mar 16, 2017 at 12:51 PM, Diego Viola <diego.viola@gmail.com> wrote:
>> On Thu, Mar 16, 2017 at 12:45 PM, Diego Viola <diego.viola@gmail.com> wrote:
>>> On Thu, Mar 16, 2017 at 12:07 PM, Alan Stern <stern@rowland.harvard.edu> wrote:
>>>> On Thu, 16 Mar 2017, Ulf Hansson wrote:
>>>>
>>>>> +Alan
>>>>>
>>>>> On 15 March 2017 at 15:00, Diego Viola <diego.viola@gmail.com> wrote:
>>>>> > On Tue, Mar 14, 2017 at 4:15 PM, Diego Viola <diego.viola@gmail.com> wrote:
>>>>> >> On Tue, Mar 14, 2017 at 2:20 PM, Diego Viola <diego.viola@gmail.com> wrote:
>>>>> >>> On Thu, Mar 9, 2017 at 2:15 PM, Diego Viola <diego.viola@gmail.com> wrote:
>>>>> >>>> On Thu, Mar 9, 2017 at 11:11 AM, Diego Viola <diego.viola@gmail.com> wrote:
>>>>> >>>>> On Wed, Mar 8, 2017 at 5:40 PM, Diego Viola <diego.viola@gmail.com> wrote:
>>>>> >>>>>> Hi Greg,
>>>>> >>>>>>
>>>>> >>>>>> On Wed, Mar 8, 2017 at 5:15 PM, Greg KH <gregkh@linuxfoundation.org> wrote:
>>>>> >>>>>>> On Wed, Mar 08, 2017 at 03:49:19PM -0300, Diego Viola wrote:
>>>>> >>>>>>>> It hangs on resume from suspend if I have USB 3.0 enabled on the BIOS,
>>>>> >>>>>>>> it works fine with ehci_hcd or USB 2.0.
>>>>> >>>>>>>>
>>>>> >>>>>>>> The way I reproduce the problem is with this command:
>>>>> >>>>>>>>
>>>>> >>>>>>>> $ i3lock && systemctl suspend
>>>>> >>>>>>>>
>>>>> >>>>>>>> This is what I see on the screen when it hangs:
>>>>> >>>>>>>>
>>>>> >>>>>>>> https://dl.dropboxusercontent.com/u/6005119/dell/IMG_20170308_095000.jpg
>>>>> >>>>>>>> https://dl.dropboxusercontent.com/u/6005119/dell/IMG_20170307_133928.jpg
>>>>> >>>>>>>>
>>>>> >>>>>>>> Some logs:
>>>>> >>>>>>>>
>>>>> >>>>>>>> https://dl.dropboxusercontent.com/u/6005119/dell/dmesg1.txt
>>>>> >>>>>>>> https://dl.dropboxusercontent.com/u/6005119/dell/dmesg2.txt
>>>>> >>>>>>>>
>>>>> >>>>>>>> I'm on Arch Linux x86_64, kernel 4.9.11-1-ARCH.
>>>>> >>>>>>>>
>>>>> >>>>>>>> I also tried Linux 4.10.1 and I could reproduce this problem there as well.
>>>>> >>>>>>>>
>>>>> >>>>>>>> Please let me know if I could provide more info.
>>>>> >>>>>>>
>>>>> >>>>>>> Has any previous kernel ever worked properly before?  If so, any chance
>>>>> >>>>>>> you can use 'git bisect' to find the offending commit?
>>>>> >>>>>>
>>>>> >>>>>> I'm not sure, this is my work machine and I've only started using it
>>>>> >>>>>> recently (since about a month ago or so).
>>>>> >>>>>>
>>>>> >>>>>> I will try older kernels and see if I get any different results, I
>>>>> >>>>>> will report back in any case.
>>>>> >>>>>>
>>>>> >>>>>>>
>>>>> >>>>>>> And are you sure you have updated your bios to the latest version?
>>>>> >>>>>>
>>>>> >>>>>> Yes.
>>>>> >>>>>>
>>>>> >>>>>>>
>>>>> >>>>>>> thanks,
>>>>> >>>>>>>
>>>>> >>>>>>> greg k-h
>>>>> >>>>>>
>>>>> >>>>>> Thanks,
>>>>> >>>>>> Diego
>>>>> >>>>>
>>>>> >>>>> I found another workaround, I can suspend/resume fine with `i3lock &&
>>>>> >>>>> systemctl suspend` if I disconnect/unplug all my USB devices
>>>>> >>>>> (keyboard, mouse, etc). This with the default settings in the BIOS
>>>>> >>>>> (both USB 2.0 and 3.0 enabled).
>>>>> >>>>>
>>>>> >>>>> I'm also seeing some messages like this in dmesg:
>>>>> >>>>>
>>>>> >>>>> [   16.172190] usb 2-6: device descriptor read/64, error -110
>>>>> >>>>>
>>>>> >>>>> Would this indicate a hardware/firmware/power issue?
>>>>> >>>>>
>>>>> >>>>> Thanks,
>>>>> >>>>> Diego
>>>>> >>>>
>>>>> >>>> OK, I've built Linux 4.4.52 (I did a localmodconfig) and rebooted into
>>>>> >>>> it, I did a suspend/resume and it hanged the first time I tried to
>>>>> >>>> resume, which isn't much different than using the latest kernel.
>>>>> >>>>
>>>>> >>>> My dmesg is still being spammed with these messages:
>>>>> >>>>
>>>>> >>>> [  260.043673] usb 2-1: Device not responding to setup address.
>>>>> >>>> [  260.246918] usb 2-1: device not accepting address 15, error -71
>>>>> >>>> [  260.633662] usb 2-1: new high-speed USB device number 17 using xhci_hcd
>>>>> >>>> [  261.341340] usb 2-1: USB disconnect, device number 17
>>>>> >>>>
>>>>> >>>> I guess it's safe to assume at this point that this is a hardware problem?
>>>>> >>>>
>>>>> >>>> Thanks,
>>>>> >>>> Diego
>>>>> >>>
>>>>> >>> Hello,
>>>>> >>>
>>>>> >>> I've found something interesting and what it seems to be the cause of
>>>>> >>> my problem.
>>>>> >>>
>>>>> >>> As soon as I boot my system I can see this process being in the D-state:
>>>>> >>>
>>>>> >>> [root@myhost ~]# ps aux | grep " D"
>>>>> >>> root       269  0.0  0.0      0     0 ?        D    14:11   0:00 [rtsx_usb_ms_2]
>>>>> >>> root      1424  0.0  0.0  10788  2172 pts/2    S+   14:19   0:00 grep  D
>>>>> >>> [root@myhost ~]#
>>>>> >>>
>>>>> >>> I'm not exactly sure why that is, but if I do a 'rmmod rtsx_usb_ms'
>>>>> >>> the problem is gone. I already tried suspending/resuming ~40 times
>>>>> >>> after I disabled the module and the suspend/resume problem is gone.
>>>>>
>>>>> That's a good observation!
>>>>>
>>>>> It suspect the drivers/memstick/host/rtsx_usb_ms.c isn't behaving
>>>>> properly from PM point of view. Perhaps it tries to access its device
>>>>> while it from a runtime PM point view still is in a runtime suspended
>>>>> state. Exactly why I don't know yet.
>>>>>
>>>>> Moreover we have had issues with this driver before and its
>>>>> corresponding SD card driver in drivers/mmc/host/rtsx_usb_sdmmc.c. On
>>>>> top of that, both their corresponding devices shares the same usb mfd
>>>>> device as parent, which is managed by drivers/mfd/rtsx_usb.c.
>>>>>
>>>>> Unfortunate my knowledge about USB is still in the learning phase,
>>>>> however I know well about runtime PM ans system suspend, so perhaps I
>>>>> still might be able to help.
>>>>>
>>>>> Anyway, I have looped in Alan, let's see if he has some input to this.
>>>>
>>>> Is the rtsx_usb_ms device attached to an xHCI controller?
>>>
>>> I think so, I'm not sure.
>>>
>>> lsusb -t reveals rtsx_usb is under xhci_hcd as seen here:
>>>
>>> https://bugzilla.kernel.org/attachment.cgi?id=255301
>>>
>>> Also, I tried disabling USB 3.0 from the BIOS and I'm still able to
>>> see rtsx_usb_ms is being loaded after that and the [rtsx_usb_ms_2]
>>> also shows up as a D-state process still, but no hanging occurs when
>>> USB 3.0 (xhci_hcd) is disabled.
>>>
>>>>
>>>> How is the hang during resume related to the actions of the xhci-hcd
>>>> driver?  (You'll probably need to enable dynamic debugging for xhci-hcd
>>>> and use a network console to get the answer.)
>>>
>>> OK, I'll do this and get back with a trace.
>>>
>>>>
>>>> If this problem really is related to xhci-hcd, have you tried bringing
>>>> it to the attention of the xhci-hcd maintainer?
>>>
>>> No, not yet. I'm also not sure who the current maintainer for xhci_hcd is?
>>>
>>> modinfo says the author is Sarah Sharp but does she still maintains it?
>>>
>>>>
>>>> Are you using the most up-to-date version of the kernel?  xhci-hcd is
>>>> still getting fixes at a very high rate.
>>>
>>> Yes, I'm currently on 4.10.2-ARCH.
>>>
>>> I will keep an eye on xhci_hcd changes on the latest git and give them
>>> a try also.
>>>
>>>>
>>>> Alan Stern
>>>>
>>>>> >>>
>>>>> >>> Diego
>>>>> >>
>>>>> >> Adding Roger Tseng to the CC also.
>>>>> >>
>>>>> >> Diego
>>>>> >
>>>>> > According to this document:
>>>>> >
>>>>> > http://downloads.dell.com/manuals/all-products/esuprt_laptop/esuprt_inspiron_laptop/inspiron-15-5558-laptop_reference%20guide_en-us.pdf
>>>>> >
>>>>> > My computer only has a SD card slot and no MEMSTICK slot.
>>>>> >
>>>>> > lsusb says this though:
>>>>> >
>>>>> > Bus 001 Device 005: ID 0bda:0129 Realtek Semiconductor Corp. RTS5129
>>>>> > Card Reader Controller
>>>>> >
>>>>> > Maybe the driver gets locked up looking for the MEMSTICK slot?
>>>>>
>>>>> Yes correct!
>>>>>
>>>>> >
>>>>> > Diego
>>>>>
>>>>> Kind regards
>>>>> Uffe
>>>>>
>>>>
>>>
>>> Thanks,
>>> Diego
>>
>> lsusb -t with USB 3.0 disabled on BIOS:
>>
>> https://bugzilla.kernel.org/attachment.cgi?id=255303
>>
>> Diego
>
> Hrm, that's rtsx_usb and not rtsx_usb_ms.
>
> I'm getting confused here.
>
> Diego

CC Mathias Nyman