Re: Hang at startup on Sony Vaio with 5.3-CURRENT

On Fri, 2004-10-22 at 13:53, Garance A Drosehn wrote:
> This is really a message from Jack Twilley <jmt_at_twilley.org>,
> but he has sent the message multiple times and neither of us
> have seen it show up in the -current mailing list.  I (Garance)

Can't help with that part, but...

> seem to remember someone else reporting a similar problem.
> I *think* it might have been Bruce Mah, but I am not sure.

A few of us (myself included) have seen this problem, although it seems
to have shown up most often on IBM ThinkPads (in my case, a T21).

> I have a Sony Vaio notebook (PCG-V505DXP) which runs 5.2-CURRENT
> but which hangs when booting 5.3-CURRENT or 6-CURRENT.
> 
> The hang occurs after the following bit:
> 
>    ata1-slave: FAILURE - ATAPI_IDENTIFY no interrupt
>    ata1-slave: FAILURE - ATAPI_IDENTIFY no interrupt
> 
> On 5.2-CURRENT, the CD drive is immediately detected after the
> second message is displayed and the boot continues normally.  On
> 5.3-CURRENT and 6-CURRENT, the system hangs for at least an hour.
> 
> The CD drive in question:
> 
>    acd0: CDRW <UJDA745 DVD/CDRW> at ata1-master PIO4
> 
> The working kernel:
> 
> FreeBSD jmt-work.twilley.org 5.2-CURRENT FreeBSD
>                   5.2-CURRENT #0: Tue Jul  6 10:35:02 PDT 2004 
> root_at_jmt-work.twilley.org:/usr/obj/usr/src/sys/JMT-WORK  i386
> 
> The failing kernels:
> 
> FreeBSD 6.0-CURRENT #14: Tue Oct  5 12:59:43 PDT 2004
> FreeBSD 5.3-STABLE  #17: Wed Oct 20 14:19:31 PDT 2004
> 
> The config file is available at
> http://www.twilley.org/~jmt/tmp/jmt-work.txt
> if that is of interest.  The dmesg output from the working kernel
> is also at http://www.twilley.org/~jmt/tmp/dmesg.txt
> 
> Please feel free to contact me if there's any more information
> which I can supply.

sos_at_ committed a fix for this to HEAD:

  1.39      +5 -8      src/sys/dev/ata/ata-queue.c

I think he's waiting for some "settling in" time before merging to
RELENG_5 and RELENG_5_3.

There's more discussion (and a patch for RELENG_5) in the thread on this
list entitled "No booting - ata1-slave: FAILURE".  I understand the root
cause is "hardware idiosyncracies", although that might be putting it
mildly.  Anyways, I'm running with the RELENG_5 patch on the
aforementioned ThinkPad and life is good.

Bruce.