Re: port-i386/53894 ("ehci_sync_hc: timed out" with ACPI/SMP enabled on eBox 3352DX3-AP)

[Andrius V <vezhlys%gmail.com@localhost>]

On Mon, Jan 21, 2019 at 8:40 AM Nick Hudson <nick.hudson%gmx.co.uk@localhost> wrote:
>
> On 20/01/2019 23:50, gutteridge%NetBSD.org@localhost wrote:
> > Synopsis: "ehci_sync_hc: timed out" with ACPI/SMP enabled on eBox 3352DX3-AP
>
> This sounds more like an MP/interrupt problem for your system than usb.
>
> ehcidebug=1 should shows us if we're seeing interrupts.
>

Hi,

Do you mean EHCI_DEBUG option in kernel configuration or something
else? Currently I was testing kernel which included EHCI_DEBUG option
line and I set hw.ehci.debug value as well, however I didn't see any
interrupt messages during boot. There were no vmstat output changes
either (I tried delaying it for few minutes, but it didn't give more
than few additional messages of "usb_event_thread#0@0: sc 0xc3d5d1b8
woke up", "usb_discover#1@0: called!" compared to original one I've
attached). After removing USB Ethernet device and leaving USB keyboard
only, "ehci_sync_hc: timed out" messages disappeared from boot log,
but USB keyboard still didn't power up (no LED lights lit up),
appeared in dmesg or worked nevertheless. Finally, I tried to boot
without SMP but with ACPI enabled (boot -1), the problem stayed the
same. So I would assume it can be ACPI related issue, since disabling
ACPI and SMP makes USB devices work (though with occasional
reconnection cycles). I couldn't test an opposite situation (SMP
enabled/ACPI disabled), because I am getting IDE interrupt errors in
this case and system fails to boot.

Regards,
Andrius V


On Mon, Jan 21, 2019 at 8:40 AM Nick Hudson <nick.hudson%gmx.co.uk@localhost> wrote:
>
> On 20/01/2019 23:50, gutteridge%NetBSD.org@localhost wrote:
> > Synopsis: "ehci_sync_hc: timed out" with ACPI/SMP enabled on eBox 3352DX3-AP
>
> This sounds more like an MP/interrupt problem for your system than usb.
>
> ehcidebug=1 should shows us if we're seeing interrupts.
>
> Nick
>