>
>When the hard disk worked well, it was partioning by HD SC setup 7.3.5(?),
>and newfsed by Mkfs 1.47(?), and installed by Installer(Version?).
>
>1.5Q install kernel failed in this disk(I forget messages in detail).
>B's Crew(commercial application) also required initialization for the
>disk(messages cannotupdate driver...etc), so I thought that boot
>records were wrong.
>
>Just at once, please you forget about boot records. NetBSD partion
>size(2400MB) or location(before is sd0g, this time is sd0a) is anxious
>for me.
>
>Takeshi Shibagaki
>ie9t-sbgk@asahi-net.or.jp

I have a similar "problem". Running NetBSD 1.5 well (a kernel I 
made), but when installing any 1.5Q kernel (20010126), the boot works 
only until the first reading on the disks. fsck says READ FAILED at 
block 16K, ls do not return any error, but shows nothing, only mount, 
df and reboot seems to work :)
When booting in multi-user, /etc/rc.d and /dev are not found, there 
are several messages like: "No such file or directory." (but the 
partitions are mounted).

Note that the 1.5Q kernel is compiling without error.
The disks are well recognized in the dmesg (no timeout or such), the 
boot output is the same on 1.5Q then 1.5 until "root file system 
type: ffs".



When trying to launch a 1.5-release kernel with the esp_diffs.gz 
patch (1.27), it's not better:
(...)
root file system type: ffs
exec /sbin/init: error 8
init: not found
panic: no init
Stopped in init at	_cpu_Debugger+0x6:	unlk	a6


Anyway, I'm running a 1.5 without problem, but, I have a Quadra 
660AV, and very interested of these esp improvments, then I'll 
continue others tests.

Maybe there is an upgraded patch for the esp (for 1.5 sources)? :-)


Hope that could help,

--
Stephane Thiell