Manuel Bouyer writes:
> On Fri, Dec 03, 1999 at 01:40:58PM -0600, Greg Oster wrote:
> > Just to confirm: Frank (fvdl) fixed a bug such that the softupdates stuff
> > now plays much nicer with RAIDframe.. but I can still (consistently) tickle
>  
> > a panic under high load (w/ softdeps + RAIDframe).  So for now, unfortunate
> ly,
> > the safest thing is to not use softdeps w/ RAIDframe :( 
> 
> Just curious: what is the panic ?

I've actually seen seen two different ones:
 panic: softdep_setup_inomapdep: found inode
and
 panic: softdep_setup_blkmapdep: found block

A traceback of the former looks like:
(gdb) where
#0  kgdb_connect (verbose=0) at ../../../../arch/i386/i386/kgdb_machdep.c:244
#1  0xf02b7cf7 in kgdb_panic ()
    at ../../../../arch/i386/i386/kgdb_machdep.c:259
#2  0xf01a723d in panic (fmt=0xf0324bfc "softdep_setup_inomapdep: found inode")
    at ../../../../kern/subr_prf.c:207
#3  0xf0324c60 in softdep_setup_inomapdep (bp=0xf0ea25e0, ip=0xf302ed5c, 
    newinum=246016) at ../../../../../gnu/sys/ufs/ffs/ffs_softdep.c:994
#4  0xf028c4bd in ffs_nodealloccg (ip=0xf302ed5c, cg=62, ipref=246016, 
    mode=16877) at ../../../../ufs/ffs/ffs_alloc.c:1276
#5  0xf028a96f in ffs_hashalloc (ip=0xf302ed5c, cg=62, pref=246016, 
    size=16877, allocator=0xf028c148 <ffs_nodealloccg>)
    at ../../../../ufs/ffs/ffs_alloc.c:761
#6  0xf028a4ea in ffs_valloc (v=0xf3076d00)
    at ../../../../ufs/ffs/ffs_alloc.c:589
#7  0xf02a2861 in ufs_mkdir (v=0xf3076e54) at ../../../../sys/vnode_if.h:933
#8  0xf01c7059 in sys_mkdir (p=0xf305ab48, v=0xf3076f88, retval=0xf3076f80)
    at ../../../../sys/vnode_if.h:575
#9  0xf0320c11 in aout_sys_mkdir (p=0xf305ab48, v=0xf3076f88, 
    retval=0xf3076f80) at ../../../../compat/aout/aout_misc.c:465
#10 0xf02c35d9 in syscall (frame={tf_es = 31, tf_ds = 31, tf_edi = 172544, 
      tf_esi = 4, tf_ebp = -272639272, tf_ebx = 1, tf_edx = 172616, 
      tf_ecx = 172544, tf_eax = 136, tf_trapno = 3, tf_err = 2, 
      tf_eip = 27103, tf_cs = 23, tf_eflags = 518, tf_esp = -272639408, 
---Type <return> to continue, or q <return> to quit---
      tf_ss = 31, tf_vm86_es = 0, tf_vm86_ds = 0, tf_vm86_fs = 0, 
      tf_vm86_gs = 0}) at ../../../../arch/i386/i386/trap.c:760
(gdb)

Suggestions/fixes/etc are most welcome :)
(I can grab values of stuff from within kgdb, but I can't get a kernel dump
after it panics :( )

Later...

Greg Oster