Dell R630 unable to detect USB, can't boot

Lee Brown

unread,

Feb 4, 2018, 5:48:32 PM2/4/18

to

Booting FreeBSD11.1-RELENG from a USB stick, loader works fine keyboard
allows selection, kernel loads and starts OK, gets as far as detecting USB
devices and stops there. imgur <https://imgur.com/a/psn0H> for a
screenshot. Tried safe mode same occurrence.

CDROM is via a USB, so that's a no go and I've not had luck booting UEFI
iSCSI nor PXE yet, so I'm stuck without a keyboard (no PS/2 connectors) or
essentially any boot media.

For reference Xen 7.2 is running fine so at first blush it's not faulty
hardware.
USB3 mode on and off (ie force USB2) in the BIOS makes no difference.
All Firmwares current.

Any suggestions on how to move forward with this? I tried turning on
verbose mode, but I'm not sure how to capture that nor where to look.

Thx -- lee
_______________________________________________
freebsd-...@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/freebsd-hardware
To unsubscribe, send any mail to "freebsd-hardwa...@freebsd.org"

Hans Petter Selasky

unread,

Feb 13, 2018, 8:49:46 AM2/13/18

to

On 02/04/18 23:47, Lee Brown wrote:
> Booting FreeBSD11.1-RELENG from a USB stick, loader works fine keyboard
> allows selection, kernel loads and starts OK, gets as far as detecting USB
> devices and stops there. imgur <https://imgur.com/a/psn0H> for a
> screenshot. Tried safe mode same occurrence.
>
> CDROM is via a USB, so that's a no go and I've not had luck booting UEFI
> iSCSI nor PXE yet, so I'm stuck without a keyboard (no PS/2 connectors) or
> essentially any boot media.
>
> For reference Xen 7.2 is running fine so at first blush it's not faulty
> hardware.
> USB3 mode on and off (ie force USB2) in the BIOS makes no difference.
> All Firmwares current.
>
> Any suggestions on how to move forward with this? I tried turning on
> verbose mode, but I'm not sure how to capture that nor where to look.
>

Hi,

USB_ERR_TIMEOUT might be an indication of a PCI IRQ issue.

--HPS

Hans Petter Selasky

unread,

Feb 13, 2018, 9:05:07 AM2/13/18

to

On 02/13/18 14:45, Hans Petter Selasky wrote:
> On 02/04/18 23:47, Lee Brown wrote:
>> Booting FreeBSD11.1-RELENG from a USB stick, loader works fine keyboard
>> allows selection, kernel loads and starts OK, gets as far as detecting
>> USB
>> devices and stops there. imgur <https://imgur.com/a/psn0H> for a
>> screenshot. Tried safe mode same occurrence.
>>
>> CDROM is via a USB, so that's a no go and I've not had luck booting UEFI
>> iSCSI nor PXE yet, so I'm stuck without a keyboard (no PS/2
>> connectors) or
>> essentially any boot media.
>>
>> For reference Xen 7.2 is running fine so at first blush it's not faulty
>> hardware.
>> USB3 mode on and off (ie force USB2) in the BIOS makes no difference.
>> All Firmwares current.
>>
>> Any suggestions on how to move forward with this? I tried turning on
>> verbose mode, but I'm not sure how to capture that nor where to look.
>>
>
> Hi,
>
> USB_ERR_TIMEOUT might be an indication of a PCI IRQ issue.
>

Try to set:

hw.usb.xhci.use_polling=1

from the loader as a temporary quirk. Then try to dump information about
IRQ's like "vmstat -i" "dmesg" and so on.

Lee Brown

unread,

Feb 13, 2018, 1:54:44 PM2/13/18

to

On Tue, Feb 13, 2018 at 6:00 AM, Hans Petter Selasky <h...@selasky.org>
wrote:

> On 02/13/18 14:45, Hans Petter Selasky wrote:
>
>> On 02/04/18 23:47, Lee Brown wrote:
>>
>>> Booting FreeBSD11.1-RELENG from a USB stick, loader works fine keyboard
>>> allows selection, kernel loads and starts OK, gets as far as detecting
>>> USB
>>> devices and stops there. imgur <https://imgur.com/a/psn0H> for a
>>> screenshot. Tried safe mode same occurrence.
>>>
>>> CDROM is via a USB, so that's a no go and I've not had luck booting UEFI
>>> iSCSI nor PXE yet, so I'm stuck without a keyboard (no PS/2 connectors)
>>> or
>>> essentially any boot media.
>>>
>>> For reference Xen 7.2 is running fine so at first blush it's not faulty
>>> hardware.
>>> USB3 mode on and off (ie force USB2) in the BIOS makes no difference.
>>> All Firmwares current.
>>>
>>> Any suggestions on how to move forward with this? I tried turning on
>>> verbose mode, but I'm not sure how to capture that nor where to look.
>>>
>>>
>> Hi,
>>
>> USB_ERR_TIMEOUT might be an indication of a PCI IRQ issue.
>>
>>
> Try to set:
>
> hw.usb.xhci.use_polling=1
>
> from the loader as a temporary quirk. Then try to dump information about
> IRQ's like "vmstat -i" "dmesg" and so on.
>
> --HPS
>

> Thanks Hans, but unfortunately that made no difference. Here's my full
loader.conf

vfs.mountroot.timeout="10"
boot_multicons="YES"
boot_serial="YES"
comconsole_speed="115200"
console="comconsole,efi"
kern.cam.boot_delay="10000"
hw.usb.xhci.use_polling=1

Hans Petter Selasky

unread,

Feb 13, 2018, 4:00:32 PM2/13/18

to

Did you enable USB 3.0 in the BIOS when trying this?

--HPS

Lee Brown

unread,

Feb 13, 2018, 5:02:15 PM2/13/18

to

On Tue, Feb 13, 2018 at 12:56 PM, Hans Petter Selasky <h...@selasky.org>

> It was configured with USB3.0 OFF
I just toggled it ON, same result.

I verified the sysctl polling value from the loader before booting

Hans Petter Selasky

unread,

Feb 14, 2018, 7:37:36 AM2/14/18

to

Hi,

I suggest you bring this issue up with Dell customer support.

Further, maybe Linux XHCI / EHCI has a quirk for Dell R630.

Alternativly you can try to debug the XHCI/EHCI debug prints on your
own. If you get a timeout it means something is not communicating like
expected in the XHCI/EHCI DMA descriptors.

--HPS

Lee Brown

unread,

Feb 15, 2018, 12:26:07 AM2/15/18

to

q

On Wed, Feb 14, 2018 at 4:33 AM, Hans Petter Selasky <h...@selasky.org>

> As I only have the hardware replacement level of support and FreeBSD isn't
on their list of certified OS's, I'm probably best spending my time
debugging this issue.

I do find it curious that the loader is able to work the USB hardware but
not the kernel.

Once I figure out how to capture the kernel output, I'll crank up the
debugging and post back here.

Hans Petter Selasky

unread,

Feb 15, 2018, 3:16:01 AM2/15/18

to

On 02/15/18 06:24, Lee Brown wrote:
> I do find it curious that the loader is able to work the USB hardware but
> not the kernel.

The loader is using the BIOS's USB code. Usually there is a handover
from BIOS to OS, maybe that's where it's broken.

--HPS

Lee Brown

unread,

Feb 16, 2018, 11:53:55 AM2/16/18

to

On Thu, Feb 15, 2018 at 12:10 AM, Hans Petter Selasky <h...@selasky.org>
wrote:

> On 02/15/18 06:24, Lee Brown wrote:
>
>> I do find it curious that the loader is able to work the USB hardware but
>> not the kernel.
>>
>
> The loader is using the BIOS's USB code. Usually there is a handover from
> BIOS to OS, maybe that's where it's broken.

Got a little more information from watching slow video:

pcib1: <ACPI Host-PCI bridge> on apci0
pcib1: _OSC returned error 0x10

I think that may be AE_LIMIT, but I can't be sure.

Do you know if ACPI could be the root cause, I do see the ehci device
created later on, so it does 'see' the USB bus, but maybe that above error
prevents it from finding the hub? I don't know I'm wildly speculating here.

Maybe I should address the above first, then circle around to this.