Someone broke USB

[Steve Kargl]

On Sun, Jul 07, 2019 at 11:53:27AM +0200, Hans Petter Selasky wrote:
> On 2019-07-07 10:05, Steve Kargl wrote:
> > On Sat, Jul 06, 2019 at 04:14:53PM -0700, Steve Kargl wrote:
> >> On Sat, Jul 06, 2019 at 03:08:13PM -0600, Ian Lepore wrote:
> >>>
> >>> It seems almost certain to be r349161 that causes the problem.
> >>>
> >>
> >> I've backed out the change, and the buildkernel is currently
> >> running.  It won't finish for an hour or so (old hardware and
> >> rebuilding another project).
> >>
> > 
> > I can confirm that r349161 is the cause of my problem.
> > 
> 
> Hi,
> 
> Can you try the latest -current with:
> debug.acpi.disabled="usb"

Reverting my tree to include r349161, updating to top-of-tree,
the new kernel boots and USB works if the above sysctl is set.

> 
> In /boot/loader.conf ?
> 
> Also can you dump the ACPI tables:
> acpidump -dt > dump.txt

This a 7720 line, 262KB file, do you want me to send it
to you in private email or put in my home directory on
freefall (i.e., kargl at freefall.freebsd.org).

> 
> In the failing and working case can you show the differences in output from:
> devinfo
> 

Here's the diff when r349161 was revert (devinfo.working) and
top-of-tree with the sysctl set.

--- devinfo.working     2019-07-07 08:48:26.906454000 -0700
+++ devinfo.acpi_disabled       2019-07-07 09:44:09.225520000 -0700
@@ -1,8 +1,8 @@
 nexus0
+  cryptosoft0
   vtvga0
   apic0
   ram0
-  cryptosoft0
   acpi0
     cpu0
       acpi_perf0
@@ -67,10 +67,10 @@
             uhub3
         uhci1
           usbus1
-            uhub6
+            uhub2
         ehci0
           usbus2
-            uhub2
+            uhub5
         hdac0
           hdacc0
             hdaa0
@@ -89,7 +89,7 @@
             uhub1
         uhci3
           usbus4
-            uhub5
+            uhub6
               ukbd0
               ums0
               uhid0

-- 
Steve