Re: Lock of NetBSD-current with ifconfig down / up

[Masanobu SAITOH <msaitoh%execsw.org@localhost>]

Hi.

On 2022/09/18 2:55, John Klos wrote:
> Hi,
> 
> Here's a nice issue :)
> 
> Plug in ure* USB ethernet to amd64 machine running NetBSD-current (9.99.99, 22-August-2022):
> 
> [ 1791670.446266] ure0 at uhub8 port 4
> [ 1791670.446266] ure0: Realtek (0x0bda) USB 10/100/1000 LAN (0x8153), rev 2.10/30.00, addr 6
> [ 1791670.446266] ure0: RTL8153 ver 5c30
> [ 1791670.566267] rgephy0 at ure0 phy 0: RTL8251 1000BASE-T media interface, rev. 0
> [ 1791670.586267] rgephy0: 10baseT, 10baseT-FDX, 100baseTX, 100baseTX-FDX, 1000baseT-FDX, auto
> [ 1791670.586267] ure0: Ethernet address a0:ce:c8:e7:88:5f
> [ 1791673.256299] ugen1 at uhub8 port 5
> [ 1791673.256299] ugen1: VIA Labs, Inc. (0x2109) PD3.0 USB-C Device (0x8888), rev 2.01/0.01, addr 7
> 
> ifconfig ure0 up
> 
> No problem:
> 
> ure0: flags=0x8943<UP,BROADCAST,RUNNING,PROMISC,SIMPLEX,MULTICAST> mtu 1500
>     capabilities=0x3ff00<IP4CSUM_Rx,IP4CSUM_Tx,TCP4CSUM_Rx,TCP4CSUM_Tx>
>     capabilities=0x3ff00<UDP4CSUM_Rx,UDP4CSUM_Tx,TCP6CSUM_Rx,TCP6CSUM_Tx>
>     capabilities=0x3ff00<UDP6CSUM_Rx,UDP6CSUM_Tx>
>     enabled=0
>     ec_capabilities=0x1<VLAN_MTU>
>     ec_enabled=0
>     address: a0:ce:c8:e7:88:5f
>     media: Ethernet autoselect (100baseTX full-duplex)
>     status: active
>     inet6 fe80::a2ce:c8ff:fee7:885f%ure0/64 flags 0 scopeid 0x9
> 
> ifconfig ure0 down
> 
> Locks the machine. I couldn't get more information because it's 3000 miles away. There's nothing in dmesg because the machine was power cycled.

One of my environment:

# ifconfig ure0 down
[  63.2217114] panic: kernel diagnostic assertion "mii_locked(mii)" failed: file "../../../../dev/mii/mii.c", line 381
[  63.2332664] cpu0: Begin traceback...
[  63.2332664] vpanic() at netbsd:vpanic+0x183
[  63.2418447] kern_assert() at netbsd:kern_assert+0x4b
[  63.2418447] mii_down() at netbsd:mii_down+0x9a
[  63.2521575] usbnet_if_stop() at netbsd:usbnet_if_stop+0x134
[  63.2521575] ether_ioctl_reinit() at netbsd:ether_ioctl_reinit+0x78
[  63.2623342] usbnet_if_ioctl() at netbsd:usbnet_if_ioctl+0x80
[  63.2717104] doifioctl() at netbsd:doifioctl+0x322
[  63.2717104] sys_ioctl() at netbsd:sys_ioctl+0x56d
[  63.2817903] syscall() at netbsd:syscall+0x196
[  63.2817903] --- syscall (number 54) ---
[  63.2817903] netbsd:syscall+0x196:
[  63.2930753] cpu0: End traceback...








> Initially I imagined it might be due to the ure* driver, but then it happened locally.
> 
> On an amd64 system running 9.99.98 from 16-July-2022, I ran "ifconfig re1 down" and the machine locked - no ICMP, nothing for SIGINFO, no response to keyboard cnmagic.
> 
> It doesn't appear to be hardware, but here's this just because.
> 
> [     1.044097] re1 at pci2 dev 0 function 0: RealTek 8168/8111 PCIe Gigabit Ethernet (rev. 0x0c)
> [     1.044097] re1: interrupting at msix2 vec 0
> [     1.044097] re1: RTL8168G/8111G (0x4c00)
> [     1.044097] re1: Ethernet address 4c:cc:6a:01:a5:e0
> [     1.044097] re1: using 256 tx descriptors
> [     1.044097] rgephy1 at re1 phy 7: RTL8251 1000BASE-T media interface, rev. 0
> 
> I've ordered some PS/2 keyboards, because I take it that's the only way to reliably get in to the kernel debugger on amd64, unless someone knows a trick to make USB keyboards usable.
> 
> send-pr?
> 
> Thanks,
> John

-- 
-----------------------------------------------
                SAITOH Masanobu (msaitoh%execsw.org@localhost
                                 msaitoh%netbsd.org@localhost)