Re: Patched gpsd and /dev/pps0 results in "sleeping thread" kernel panic

On Tue, Aug 31, 2021 at 9:25 PM Warner Losh <imp@bsdimp.com> wrote:

>
>
> On Tue, Aug 31, 2021 at 9:13 PM Ian Lepore <ian@freebsd.org> wrote:
>
>> On Tue, 2021-08-31 at 21:07 -0600, Warner Losh wrote:
>> > On Tue, Aug 31, 2021 at 8:11 PM Craig Leres <leres@freebsd.org>
>> > wrote:
>> >
>> > >
>> > > On 8/31/21 6:54 PM, Warner Losh wrote:
>> > > > Can you post a traceback? We recently tightened up the sleeping
>> > > conditions
>> > > > and many of the panics are straight forward to fix
>> > >
>> > > Is this sufficient?
>> > >
>> >
>> > Maybe. I'm guessing that the pps driver isn't setting driver_mtx for
>> > the
>> > pps_state that it keeps, so when we're sleeping in pps_ioctl we're
>> > not
>> > dropping the the ppbus lock, which leads to this diagnostic.
>> >
>> > Warner
>> >
>> >
>>
>>
>> Exactly so, I've just been digging through the code.  It's not an easy
>> fix because the ppbus pps driver doesn't have access to the ppbus mutex
>> at the point where it's registering to be a pps driver.
>>
>
> I just came to the conclusion that either we'd need to create a call in
> ppbus
> to set this mutex, or to return its lock...
>
>
>> Craig, is there any chance you can configure your hardware so that the
>> PPS signal is connected to either the CTS or DCD pin of a uart (even
>> the same uart that's communicating with the gps receiver should work if
>> you use DCD)?  You'll get jitter and accuracy performance comparable to
>> the parallel port pin, and avoid this mutex/sleeping problem.
>>
>
> That's easier, but depending on the hardware might be harder :(
>

Either I'm missing something (likely am), or this might fix it up,
or at least get away from the warning:

https://reviews.freebsd.org/D31763

Note: I can't recall why ppbus has to be locked for this call.
This code dates from the very earliest days of locking and
so may do things simply because it seemed like a good idea
without a specific notion as to what that lock is protecting. If
so, the real fix may be to not take the lock in pps_ioctl at
all and maybe instead use a reference count (the most
often reason for 'a good idea' was to keep the device
from going away, though this is a parent lock, not a
child one so I'm less sure about that being the reason).

Warner


> Warner
>
>
>> -- Ian
>>
>>
>>
>> > >                 Craig
>> > >
>> > > toc2 4 # kgdb /boot/kernel/kernel /var/crash/vmcore.1
>> > > GNU gdb (GDB) 10.2 [GDB v10.2 for FreeBSD]
>> > > Copyright (C) 2021 Free Software Foundation, Inc.
>> > > License GPLv3+: GNU GPL version 3 or later
>> > > <http://gnu.org/licenses/gpl.html>
>> > > This is free software: you are free to change and redistribute it.
>> > > There is NO WARRANTY, to the extent permitted by law.
>> > > Type "show copying" and "show warranty" for details.
>> > > This GDB was configured as "x86_64-portbld-freebsd12.2".
>> > > Type "show configuration" for configuration details.
>> > > For bug reporting instructions, please see:
>> > > <https://www.gnu.org/software/gdb/bugs/>.
>> > > Find the GDB manual and other documentation resources online at:
>> > >      <http://www.gnu.org/software/gdb/documentation/>.
>> > >
>> > > For help, type "help".
>> > > Type "apropos word" to search for commands related to "word"...
>> > > Reading symbols from /boot/kernel/kernel...
>> > > Reading symbols from /boot/kernel.LBLNET/kernel.debug...
>> > >
>> > > Unread portion of the kernel message buffer:
>> > > Sleeping thread (tid 101408, pid 2046) owns a non-sleepable lock
>> > > KDB: stack backtrace of thread 101408:
>> > > sched_switch() at sched_switch+0x630/frame 0xfffffe0070e4f760
>> > > mi_switch() at mi_switch+0xd4/frame 0xfffffe0070e4f790
>> > > sleepq_catch_signals() at sleepq_catch_signals+0x403/frame
>> > > 0xfffffe0070e4f7e0
>> > > sleepq_timedwait_sig() at sleepq_timedwait_sig+0x14/frame
>> > > 0xfffffe0070e4f820
>> > > _sleep() at _sleep+0x1b3/frame 0xfffffe0070e4f8a0
>> > > pps_ioctl() at pps_ioctl+0x298/frame 0xfffffe0070e4f8f0
>> > > ppsioctl() at ppsioctl+0x48/frame 0xfffffe0070e4f920
>> > > devfs_ioctl() at devfs_ioctl+0xb0/frame 0xfffffe0070e4f970
>> > > VOP_IOCTL_APV() at VOP_IOCTL_APV+0x7b/frame 0xfffffe0070e4f9a0
>> > > vn_ioctl() at vn_ioctl+0x16a/frame 0xfffffe0070e4fab0
>> > > devfs_ioctl_f() at devfs_ioctl_f+0x1e/frame 0xfffffe0070e4fad0
>> > > kern_ioctl() at kern_ioctl+0x2b7/frame 0xfffffe0070e4fb30
>> > > sys_ioctl() at sys_ioctl+0xfa/frame 0xfffffe0070e4fc00
>> > > amd64_syscall() at amd64_syscall+0x387/frame 0xfffffe0070e4fd30
>> > > fast_syscall_common() at fast_syscall_common+0xf8/frame
>> > > 0xfffffe0070e4fd30
>> > > --- syscall (54, FreeBSD ELF64, sys_ioctl), rip = 0x8004c899a, rsp
>> > > =
>> > > 0x7fffdfdfc6a8, rbp = 0x7fffdfdfc730 ---
>> > > panic: sleeping thread
>> > > cpuid = 8
>> > > time = 1630456142
>> > > KDB: stack backtrace:
>> > > db_trace_self_wrapper() at db_trace_self_wrapper+0x2b/frame
>> > > 0xfffffe005ab73ab0
>> > > vpanic() at vpanic+0x17b/frame 0xfffffe005ab73b00
>> > > panic() at panic+0x43/frame 0xfffffe005ab73b60
>> > > propagate_priority() at propagate_priority+0x282/frame
>> > > 0xfffffe005ab73b90
>> > > turnstile_wait() at turnstile_wait+0x30c/frame 0xfffffe005ab73be0
>> > > __mtx_lock_sleep() at __mtx_lock_sleep+0x199/frame
>> > > 0xfffffe005ab73c70
>> > > ppcintr() at ppcintr+0x2a0/frame 0xfffffe005ab73c90
>> > > ithread_loop() at ithread_loop+0x23c/frame 0xfffffe005ab73cf0
>> > > fork_exit() at fork_exit+0x7e/frame 0xfffffe005ab73d30
>> > > fork_trampoline() at fork_trampoline+0xe/frame 0xfffffe005ab73d30
>> > > --- trap 0, rip = 0, rsp = 0, rbp = 0 ---
>> > > Uptime: 17m20s
>> > > Dumping 607 out of 12240
>> > > MB:..3%..11%..22%..32%..43%..51%..61%..72%..82%..93%
>> > >
>> > > __curthread () at /usr/src/sys/amd64/include/pcpu_aux.h:55
>> > > 55      /usr/src/sys/amd64/include/pcpu_aux.h: No such file or
>> > > directory.
>> > > (kgdb) bt
>> > > #0  __curthread () at /usr/src/sys/amd64/include/pcpu_aux.h:55
>> > > #1  doadump (textdump=1) at ../../../kern/kern_shutdown.c:371
>> > > #2  0xffffffff80b83ada in kern_reboot (howto=260)
>> > >      at ../../../kern/kern_shutdown.c:451
>> > > #3  0xffffffff80b83f33 in vpanic (fmt=<optimized out>,
>> > > ap=<optimized out>)
>> > >      at ../../../kern/kern_shutdown.c:880
>> > > #4  0xffffffff80b83d53 in panic (fmt=<unavailable>)
>> > >      at ../../../kern/kern_shutdown.c:807
>> > > #5  0xffffffff80be7152 in propagate_priority
>> > > (td=0xfffff80196959740)
>> > >      at ../../../kern/subr_turnstile.c:228
>> > > #6  0xffffffff80be7d1c in turnstile_wait (ts=0xfffff8000357a780,
>> > >      owner=<optimized out>, queue=0) at
>> > > ../../../kern/subr_turnstile.c:785
>> > > #7  0xffffffff80b62ca9 in __mtx_lock_sleep (c=0xfffff80003969bd0,
>> > >      v=<optimized out>) at ../../../kern/kern_mutex.c:654
>> > > #8  0xffffffff8086fcc0 in ppcintr (arg=0xfffff80003969b00)
>> > >      at ../../../dev/ppc/ppc.c:1546
>> > > #9  0xffffffff80b4637c in intr_event_execute_handlers (p=<optimized
>> > > out>,
>> > >      ie=0xfffff800030d9d00) at ../../../kern/kern_intr.c:1143
>> > > #10 ithread_execute_handlers (p=<optimized out>,
>> > > ie=0xfffff800030d9d00)
>> > >      at ../../../kern/kern_intr.c:1156
>> > > #11 ithread_loop (arg=0xfffff800039aa7e0) at
>> > > ../../../kern/kern_intr.c:1236
>> > > #12 0xffffffff80b42e1e in fork_exit (
>> > >      callout=0xffffffff80b46140 <ithread_loop>,
>> > > arg=0xfffff800039aa7e0,
>> > >      frame=0xfffffe005ab73d40) at ../../../kern/kern_fork.c:1080
>> > > #13 <signal handler called>
>> > > (kgdb)
>> > >
>>
>>
>>