'devctl detach' panic after resume [Was: Re: svn commit: r333338 - in stable/11/sys: dev/bnxt kern net sys]

[Harry Schmalzbauer]

Am 08.05.2018 um 11:52 schrieb Kevin Bowling:
> On Tue, May 8, 2018 at 2:43 AM, Harry Schmalzbauer <freebsd at omnilan.de> wrote:
>> Bezüglich Kevin Bowling's Nachricht vom 08.05.2018 11:26 (localtime):
>>> iflib in stable/11 only affects bnxt at this time.
>>>
>>> You should try out HEAD and let us know for the rest of your questions.
…
> 
> At this point it would be helpful, we think e1000 is nearing pretty
> good shape and I need to become familiar with any outstanding bugs.

To my surprise, acpi-S3 seems to work in general on my "new" haswell 
desktop.
Just the NIC (i217) wasn't working after resume.
So I wanted to issue a D3-D0, which I thought could be done by 'devctl 
detach/attach'.
The following panic happened after 'devctl detach' – unfortunately swap 
partition (3G) again too small, so transcribed from photo:

Stopped at kdb_enter+0x3b: movq $0,kdb_why

Tracing pid 39575 tid 102130 td 0xfffff801a84ff580
kdb_enter()
vpanic()
panic()
_sx_assert() at _sx_assert+0xb2/frame 0xfffffe009ecb3460
e1000_acquire_swflag_ich8lan() at e1000_acqire_swflag_ich8lan+0x36/frame 
0xfffffe009ecb34a0
e1000_phy_hw_reset_generic() at e1000_phy_hw_reset_generic+0x2d/frame 
0xfffffe009ecb34d0
e1000_phy_hw_reset_ich8lan() at e1000_phy_hw_reset_ich8lan+0x6a/frame 
0xfffffe009ecb3500
em_if_detach() at em_if_detach+0x22/frame 0xfffffe009ecb3540
iflib_device_deregister() at iflib_device_deregister+0x25e/frame 
0xfffffe009ecb3590
device_detach() at device_detach+0x167/frame 0xfffffe009ecb35d0
devctl2_ioctl() at devctl2_ioctl+0x53b/frame 0xfffffe009ecb36a0
devfs_ioctl() at devfs_ioctl+0xcb/frame 0xfffffe009ecb36f0
VOP_IOCTL_APV() at VOP_IOCTL_APV+0xd9/frame 0xfffffe009ecb3720
vn_ioctl() at vn_ioctl+0x124/frame 0xfffffe009ecb3830
devfs_ioctl_f() at devfs_ioctl_f+0x1f/frame 0xfffffe009ecb3850
kern_ioctl() at kern_ioctl+0x2b9/frame 0xfffffe009ecb38b0
sys_ioctl() at sys_ioctl+0x168/frame 0xfffffe009ecb3980
amd64_syscall() at amd64_syscall+0x28c/frame 0xfffffe009ecb3ab0
fast_syscall_common() at fast_syscall_common

Shall I try to reproduce for getting a core file?

Thanks,

-harry