I've run into what I thought was an ffs bug but really is a VM bug
after poking past the surface.  I don't know what to do past what is
here.  Maybe we need more invariants?

(kgdb) bt
#0  boot (howto=0x100) at ../../kern/kern_shutdown.c:303
#1  0xc01669a9 in panic (fmt=0xc0281a20 "vm_page_free: freeing free page") at 
../../kern/kern_shutdown.c:553
#2  0xc021b389 in vm_page_free_toq (m=0xc0535f70) at ../../vm/vm_page.c:1076
#3  0xc021b0be in vm_page_alloc (object=0xd079a900, pindex=0x19d0, page_req=0x0) at 
../../vm/vm_page.h:523
#4  0xc01930dc in allocbuf (bp=0xc37152c0, size=0x2000) at ../../kern/vfs_bio.c:2467
#5  0xc0192c86 in getblk (vp=0xd03d79c0, blkno=0xce8, size=0x2000, slpflag=0x0, 
slptimeo=0x0)
    at ../../kern/vfs_bio.c:2247
#6  0xc01fde1d in ffs_balloc (ap=0xd0428e2c) at ../../ufs/ffs/ffs_balloc.c:313
#7  0xc02097c1 in ffs_write (ap=0xd0428e6c) at vnode_if.h:1077
#8  0xc01a09e7 in vn_write (fp=0xc13bb440, uio=0xd0428edc, cred=0xc0ec5e00, flags=0x0, 
p=0xd031bbe0) at vnode_if.h:363
#9  0xc01782dd in dofilewrite (p=0xd031bbe0, fp=0xc13bb440, fd=0x5, buf=0x8061000, 
nbyte=0x7770, 
    offset=0xffffffffffffffff, flags=0x0) at ../../sys/file.h:157
#10 0xc01781c3 in write (p=0xd031bbe0, uap=0xd0428f80) at ../../kern/sys_generic.c:310
#11 0xc024b0a9 in syscall2 (frame={tf_fs = 0x2f, tf_es = 0x2f, tf_ds = 0x2f, tf_edi = 
0x8061000, tf_esi = 0x7770, 
      tf_ebp = 0xbfbff894, tf_isp = 0xd0428fd4, tf_ebx = 0x0, tf_edx = 0x0, tf_ecx = 
0x4ec, tf_eax = 0x4, 
      tf_trapno = 0x7, tf_err = 0x2, tf_eip = 0x8054d84, tf_cs = 0x1f, tf_eflags = 
0x297, tf_esp = 0xbfbff868, 
      tf_ss = 0x2f}) at ../../i386/i386/trap.c:1150
#12 0xc023f0f5 in Xint0x80_syscall ()
#13 0x8048998 in ?? ()
#14 0x80496b3 in ?? ()
#15 0x8048efd in ?? ()
#16 0x8048139 in ?? ()
(kgdb) frame 2
#2  0xc021b389 in vm_page_free_toq (m=0xc0535f70) at ../../vm/vm_page.c:1076
1076                            panic("vm_page_free: freeing free page");
(kgdb) printf "%d %d\n", m->queue, m->pc
102 101

Anyone know where to go from here?  I don't think there's anything
that could have fixed it in the meantime...

--
 Brian Fundakowski Feldman           \  FreeBSD: The Power to Serve!  /
 [EMAIL PROTECTED]                    `------------------------------'



To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-current" in the body of the message

Reply via email to