Quoting Andiry Xu <andiry.xu@xxxxxxx>:
On 11/23/2011 04:59 PM, Tomi Orava wrote:
Quoting Tomi Orava <tomimo+linux-usb@xxxxxxxxxxxxxxxxxx>:
> On 11/13/2011 07:28 PM, Alan Stern wrote:
>> On Sun, 13 Nov 2011, Tomi Orava wrote:
>>
>>> The following log is a complete boot log with only a usb memory stick
>>> connected to usb2 port (ohci_hcd & xhci_hcd blacklisted).
>>
>> ...
>>
>>> [ 2.292054] ehci_hcd 0000:00:13.2: port 5 high speed
>>> [ 2.292059] ehci_hcd 0000:00:13.2: GetStatus port:5 status
>>> 001005 0 ACK POWER sig=se0 PE CONNECT
>>> [ 2.343037] usb 2-5: new high speed USB device number 2 using ehci_hcd
>>> [ 2.394080] ehci_hcd 0000:00:13.2: port 5 high speed
>>> [ 2.394085] ehci_hcd 0000:00:13.2: GetStatus port:5 status
>>> 001005 0 ACK POWER sig=se0 PE CONNECT
>>> [ 2.445038] usb 2-5: device descriptor read/64, error -32
>> ...
>>
>>> [ 3.309061] ehci_hcd 0000:00:13.2: port 5 high speed
>>> [ 3.309064] ehci_hcd 0000:00:13.2: GetStatus port:5 status
>>> 001005 0 ACK POWER sig=se0 PE CONNECT
>>> [ 3.360059] usb 2-5: new high speed USB device number 4 using ehci_hcd
>>> [ 3.762060] usb 2-5: device not accepting address 4, error -32
>>
>> This resembles a problem that people have reported for NVIDIA OHCI
>> controllers (they don't like to be kept in the RESET state very long).
>> But in your case it affects an EHCI controller. I have no idea why.
>>
>> Earlier you said that this showed up only with certain "problematic"
>> devices. Is the memory stick one of those devices?
>
> Not exactly, the most problematic devices have been the Cypress
> Tetrahub USB2.0 hub
> and the "Epox Bluetooth Adapter" ie:
>
> Bus 002 Device 007: ID 0a12:0001 Cambridge Silicon Radio, Ltd
> Bluetooth Dongle (HCI mode)
>
> (Although, this is a usb 1.0 device in my understanding, I don't
> have it with me right now)
>
> These two are seem to most of the time confuse the usb sub-systems so
that
> all of the OHCI/EHCI/XHCI are completely messed up. Unfortunately,
I've seen
> that also the transcend memory stick that usually works just fine,
can also
> confuse the usb controllers when inserted/removed several times in a row
> (each time the device is given enough time to stabilize & usb-storage to
> detect the inserted device) ---> So, basically it looks like the usb
support
> is somehow just unstable on this board and/or chipset. Currently, if
I don't
> mess up the controllers, the most stable device seems to be a
> Logitech Optical
> mouse connected especially to USB 3.0 port.
As a follow up to this long thread:
I just retrieved & updated the latest Asus provided UEFI bios for this
motherboard (0813) that was released a few days ago and suddenly both
the usb 2.0 & usb 3.0 systems are seeing all the connected devices
without problems. So, it might be that there is some related changes
in the latest bios release that fixes the problems with linux & M5A97
Pro motherboard but I'll still have to test if the bios setting
"IOMMU" somehow affects these previosly seen problems. The system has
only been running stable less than 24 hours for now, but it looks good.
At least this damn thing is now working. I received a reply from Asus
support
that "we don't support Linux" in relation to these usb problems --->
Great attitude from a vendor nowadays.
That's good news, also saves me the effort to find a Tetra hub. Just
a question, have you seen this issue on Windows with old BIOS? I
don't understand why BIOS affects driver when OS is running, since
the ownership of controller should be handovered to driver.
Unfortunately, this machine doesn't have windows installed at all.
I have absolutely no idea, why a bios update and/or bios iommu setting
might cause the previously mentioned usb stability issues, but the one
and only thing that was changed on that machine was bios update & bios
settings reset afterwards and suddenly everything seems to be working.
I'll update the
kernel to the latest 3.1.2 version just to see that is some of the previously
used patches required or not. The machine is just right now
occupied/in use for a day or two (as it seems to work) so I'll do the
testing after its free again.
I think that asus did something with that bios release (0813) as I did already
try out the earlier 0810 (beta) bios without any change/help. Based on
a google searches, there seem to be others having the same problems
with this particular motherboard & linux usb support. I'll check the
IOMMU bios setting though and see if that affects the usb sub-systems
somehow or not.
Tomi Orava
PS. In case somebody is trying to figure how to get the integrated
RTL8111 network card to work with linux, try enabling the "use_dac"
r8169 module parameter, that worked for me ---> ie. "modprobe -v r8169
use_dac=1". Without this parameter there is either no link or nothing
goes through the NIC.
--
To unsubscribe from this list: send the line "unsubscribe linux-usb" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
