Hi,

>>>>> On Sun, 28 Jan 2001 13:44:10 -0500, Allen Briggs <briggs@wasabisystems.com> said:

>> When I changed a partition table of same disk, sd0 didn't work well.
>> sd0(esp0:2:0): esp0: timed out [ecb 0x582068 (flags 0x1, dleft 10000, stat 0)], <state 5, nexus 0x582068, phase(l 3, c 0, p 0), resid 10000, msg(q 0,o 0) DMA active>
>> sd0(esp0:2:0): esp0: timed out [ecb 0x582068 (flags 0x41, dleft 10000, stat 0)], <state 5, nexus 0x582068, phase(l 3, c 0, p 0), resid 10000, msg(q 20,o 0) DMA active> AGAIN
>> panic: esp0 invalid status: 6

Allen> Hmmm...  What version of esp.c is that from?  1.27 or 1.28?
Allen> That looks like the errors that we saw with my patches from version 1.27.

First time is current-20010121, so maybe 1.28. Next time is maybe 1.27.
I used a 1.5J INSTALL kernel which was made by me, when installed current-20010121.
Installation might be well because of an old kernel.

Allen> This was caused by the disk disconnecting during a write operation with
Allen> data still in the FIFO.  The data in the FIFO was being counted as
Allen> "transferred" in 1.27 (I was confused about when the transfer count
Allen> registers were being updated).

Is there a relation with NetBSD patition size or boot records(from 0
to 576 sectors)? At first, 1.5Q install kernel(I made) failed because
of no reading a patition table, when I was going to install. So, I
formated by B's Crew(Formatter for Mac). At this time, boot records
was changed.

Takeshi Shibagaki
ie9t-sbgk@asahi-net.or.jp,shiba@jp.freebsd.org