Re: [PATCH v2] OHCI: add a quirk for ULi M5237 blocking on reset

[Nikita Yushchenko <nyushchenko@dev.rtsoft.ru>]

23.12.2014 08:36, Arseny Solokha пишет:
>> On Sat, 6 Dec 2014, Arseny Solokha wrote:
>>
>>> From: Arseny Solokha <asolokha@kb.kras.ru>
>>>
>>> Commit 8dccddbc2368 ("OHCI: final fix for NVIDIA problems (I hope)")
>>> introduced into 3.1.9 broke boot on e.g. Freescale P2020DS development
>>> board. The code path that was previously specific to NVIDIA controllers
>>> had then become taken for all chips.
>>>
>>> However, the M5237 installed on the board wedges solid when accessing
>>> its base+OHCI_FMINTERVAL register, making it impossible to boot any
>>> kernel newer than 3.1.8 on this particular and apparently other similar
>>> machines.
>>>
>>> Don't readl() and writel() base+OHCI_FMINTERVAL on PCI ID 10b9:5237.
>>>
>>> The patch is suitable for the -next tree as well as all maintained
>>> kernels up to 3.2 inclusive.
>>>
>>> Signed-off-by: Arseny Solokha <asolokha@kb.kras.ru>
>>> ---
>>> Changes in v2:
>>> - review comments applied
>>
>> Much better this time.
>>
>> Acked-by: Alan Stern <stern@rowland.harvard.edu>
> 
> Meanwhile, I discovered a similar thread where the issue was previously
> discussed, and have tried to implement a workaround suggested in [1]. No luck so
> far: even if the code manages to check a value returned by
> readl(base + OHCI_FMINTERVAL), it then still hangs during device unregistering.
> 
> Of course, unregistering a device instead of just not accessing OHCI_FMINTERVAL
> like in my or Nikita's patch is also possible but it requires hardcoding all
> flawed device IDs instead of figuring out misbehaving devices at boot time.

In my case root cause was elsewhere - USB port that I thought was driven
by ULI, actually was not. And ULI's builtin OHCI was not
hardware-enabled (but still was available on bus). I workarounded it by
masking entire device in platform-specific quirk.

Nikita