kthread_exit and giant

From: Petri Helenius <pete_at_he.iki.fi>
Date: Mon, 5 May 2003 08:54:55 +0300
IŽve been bouncing this with Soren a few roundtrips and I thought IŽll ask a wider
audience since I could not locate the information who would be responsible for
grabbing Giant in this case, but it seems to be required to hold it when the thread
exits; (the panic occurs when ATA RAID rebuild thread completes)

FreeBSD 5.0-CURRENT #3: Sat May  3 11:09:45 GMT 2003
    root_at_kompak.vuokselantie10.fi:/usr/src/sys/i386/compile/EMTEST
kompak# atacontrol status 0
ar0: ATA RAID1 subdisks: ad4 ad6 status: REBUILDING 44% completed
kompak# panic: mutex Giant not owned at ../../../kern/kern_exit.c:122
Debugger("panic")
Stopped at      Debugger+0x54:  xchgl   %ebx,in_Debugger.0
db> trace
Debugger(c038d044,c03eedc0,c038c78b,d26f4c54,1) at Debugger+0x54
panic(c038c78b,c038c8c4,c038acdf,7a,104) at panic+0xab
_mtx_assert(c03ed140,1,c038acdf,7a,85) at _mtx_assert+0xec
exit1(c2a075f0,0,85,84,2) at exit1+0x33
kthread_exit(0,1,1,0,0) at kthread_exit+0xb6
ar_rebuild(c25e3000,d26f4d48,c038ad90,310,c2a075f0) at ar_rebuild+0x445
fork_exit(c0161c40,c25e3000,d26f4d48) at fork_exit+0xc0
fork_trampoline() at fork_trampoline+0x1a
--- trap 0x1, eip = 0, esp = 0xd26f4d7c, ebp = 0 ---
db>
Received on Sun May 04 2003 - 20:54:59 UTC

This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:37:06 UTC