Re: Intel 100 series chipset (Skylake / Sunrise Point)

[Frank Wille <frank%phoenix.owl.de@localhost>]

Nick Hudson wrote:

>> [...] I had to disable xhci in GENERIC, otherwise my keyboard is dead.
>
> xhci has changed a lot in this time

Indeed! My fault. I was confused by a kernel panic, which disabled my USB
keyboard. Seems that nouveau has problems with my GTX970, so cnopen failed
and the kernel panicked. Then I disabled xhci together with nouveau and
could boot into multiuser with working keyboard afterwards.

But xhci has improved a lot indeed. I reenabled it and it works now nicely
with my mouse and keyboard. Thanks a lot for your hard work! :)

Even the ASMedia USB3.1 ports (1b21) are recognised and do work now.

addr 0: xHCI Root Hub, vendor 8086
addr 0: xHCI Root Hub, vendor 8086
 addr 1: Bluetooth Radio, Realtek
 addr 2: Dell USB Keyboard, Dell
 addr 3: USB Laser Mouse, Logitech
addr 0: xHCI Root Hub, vendor 1b21
addr 0: xHCI Root Hub, vendor 1b21


Sometimes there are still minor problems, though. When plugging in a
joystick:

uhidev2 at uhub1 port 1 configuration 1 interface 0
uhidev2: Saitek (0x6a3) Saitek X52 Pro Flight Control System (0x762), rev
2.00/1.23, addr 8, iclass 3/0
uhidev2: no report descriptor

Next try on next port it was ok again:

uhidev2 at uhub1 port 2 configuration 1 interface 0
uhidev2: Saitek (0x6a3) Saitek X52 Pro Flight Control System (0x762), rev
2.00/1.23, addr 9, iclass 3/0
uhid0 at uhidev2: input=15, output=0, feature=0


There are also problems with massstorage devices, like an SD card reader:

umass0 at uhub1 port 1 configuration 1 interface 0
umass0: Generic (0x7c4) Card Reader (0xa604), rev 2.00/93.21, addr 14
umass0: using SCSI over Bulk-Only
scsibus0 at umass0: 2 targets, 169 luns per target
scsibus0: detached
umass0: detached
umass0: at uhub1 port 1 (addr 14) disconnected
uhub1: port 4, set config at addr 15 failed
uhub1: device problem, disabling port 4

This happens preferably on the red colour-coded front ports (intel xhci).


> Last I tried I219 didn't work for me, but you might have more luck - 
> again if_wm.c has changed significantly.

I enabled I219 by removing the "#if 0" around the product table entries.
Don't know if there is anything else to reenable for a test.

It looks much better than a year ago (although in some cases I have also
seen timeouts here):

wm0 at pci0 dev 31 function 6: I219 V Ethernet Connection (rev. 0x31)
wm0: interrupting at msi4 vec 0
wm0: PCI-Express bus
wm0: 4096 words FLASH
wm0: Ethernet address 30:5a:3a:xx:xx:xx
ihphy0 at wm0 phy 2: i217 10/100/1000 media interface, rev. 6
ihphy0: 10baseT, 10baseT-FDX, 100baseTX, 100baseTX-FDX, 1000baseT,
1000baseT-FDX, auto

But as soon as I try to configure the device (assign an address, bring it
up) the whole system freezes. I cannot even enter ddb (maybe the same USB
kbd problems like after a panic?).


Patrick Welche wrote:

> Possibly a different chipset, but I use NetBSD on a skylake every day

But without the I219? Or does it work for you?

-- 
Frank Wille