On Fri, Jun 07, 2002 at 03:48:14PM -0000, eeh@netbsd.org wrote:
> So you get the stack dump during the trace?  Neat!
> 
> | buf-read-file            
> |
> | and here it hangs.
> 
> I'm surpised debug isn't on in all the children as well.  Could you
> try turning debug on in buf-read-file as well?

Here it is:
ok boot disk0:b -DV
Boot device: /sbus/espdma@e,8400000/esp@e,8800000/sd@0,0:b  File and args: -DV
The file just loaded does not appear to be executable.
ok debug buf-read-file
Stepper keys: <space> Down Up Continue Forth Go Help ? See $tring Quit
ok " /ofwboot" do-boot
NetBSD IEEE 1275 Bootblock
Loading file /ofwboot
from device /sbus@1f,0/espdma@e,8400000/esp@e,8800000/sd@0,0:b
ufs-open complete
.Looking for ofwboot in directory...
: buf-read-file           ( 2 fffebca9 0 fffebca2 7 fff7bff0 )
cur-offset                ( 2 fffebca9 0 fffebca2 7 fff7bff0 0 )
swap                      ( 2 fffebca9 0 fffebca2 7 0 fff7bff0 )
2dup                      ( 2 fffebca9 0 fffebca2 7 0 fff7bff0 0 fff7bff0 )
blkoff                    ( 2 fffebca9 0 fffebca2 7 0 fff7bff0 0 )
-rot                      ( 2 fffebca9 0 fffebca2 7 0 0 fff7bff0 )
2dup                      ( 2 fffebca9 0 fffebca2 7 0 0 fff7bff0 0 fff7bff0 )
lblkno                    ( 2 fffebca9 0 fffebca2 7 0 0 fff7bff0 0 )
swap                      ( 2 fffebca9 0 fffebca2 7 0 0 0 fff7bff0 )
2dup                      ( 2 fffebca9 0 fffebca2 7 0 0 0 fff7bff0 0 fff7bff0 )
cur-inode                 ( 2 fffebca9 0 fffebca2 7 0 0 0 fff7bff0 0 fff7bff0 fff73f60 )
swap                      ( 2 fffebca9 0 fffebca2 7 0 0 0 fff7bff0 0 fff73f60 fff7bff0 )
dblksize                  ( 2 fffebca9 0 fffebca2 7 0 0 0 fff7bff0 800 )
rot                       ( 2 fffebca9 0 fffebca2 7 0 0 fff7bff0 800 0 )
dup                       ( 2 fffebca9 0 fffebca2 7 0 0 fff7bff0 800 0 0 )
cur-blockno               ( 2 fffebca9 0 fffebca2 7 0 0 fff7bff0 800 0 0 60 )
l<>                       ( 2 fffebca9 0 fffebca2 7 0 0 fff7bff0 800 0 ffffffffffffffff )
l->n                      ( 2 fffebca9 0 fffebca2 7 0 0 fff7bff0 800 0 ffffffffffffffff )
if                        ( 2 fffebca9 0 fffebca2 7 0 0 fff7bff0 800 0 )
block-map                 ( 2 fffebca9 0 fffebca2 7 0 0 fff7bff0 800 51 )
dup                       ( 2 fffebca9 0 fffebca2 7 0 0 fff7bff0 800 51 51 )
l0=                       ( 2 fffebca9 0 fffebca2 7 0 0 fff7bff0 800 51 0 )
l->n                      ( 2 fffebca9 0 fffebca2 7 0 0 fff7bff0 800 51 0 )
if                        ( 2 fffebca9 0 fffebca2 7 0 0 fff7bff0 800 51 )
2dup                      ( 2 fffebca9 0 fffebca2 7 0 0 fff7bff0 800 51 800 51 )
sb-buf                    ( 2 fffebca9 0 fffebca2 7 0 0 fff7bff0 800 51 800 51 fff7bff0 )
fsbtodb                   ( 2 fffebca9 0 fffebca2 7 0 0 fff7bff0 800 51 800 144 )
cur-block                 ( 2 fffebca9 0 fffebca2 7 0 0 fff7bff0 800 51 800 144 fff73ff0 )
-rot                      ( 2 fffebca9 0 fffebca2 7 0 0 fff7bff0 800 51 fff73ff0 800 144 )
strategy                  ( 2 fffebca9 0 fffebca2 7 0 0 fff7bff0 800 51 0 )
rot                       ( 2 fffebca9 0 fffebca2 7 0 0 fff7bff0 51 0 800 )
2dup                      ( 2 fffebca9 0 fffebca2 7 0 0 fff7bff0 51 0 800 0 800 )
l<>                       ( 2 fffebca9 0 fffebca2 7 0 0 fff7bff0 51 0 800 ffffffffffffffff )
l->n                      ( 2 fffebca9 0 fffebca2 7 0 0 fff7bff0 51 0 800 ffffffffffffffff )
if                        ( 2 fffebca9 0 fffebca2 7 0 0 fff7bff0 51 0 800 )
" buf-read-file: short read."    ( 2 fffebca9 0 fffebca2 7 0 0 fff7bff0 51 0 800 f0082051 1a )
cr                       
                          ( 2 fffebca9 0 fffebca2 7 0 0 fff7bff0 51 0 800 f0082051 1a )
abort                    

And here I'm back at ok

> 
> Another option is to try an older bootblock from the ftp server 
> as a sanity check.  It sounds to me that for some reason fgen is
> no longer working correctly.

Hum, I tried the miniroot from the 1.5ZC snapshot, and I get the same behavior.
Then I tried running binstall from my 1.5.1 system, and I still get the
same behavior. So it may be a old problem.

--
Manuel Bouyer, LIP6, Universite Paris VI.           Manuel.Bouyer@lip6.fr
--