special test request for 6.1-BETA1 on ATA disks

[Wilko Bulte]

On Fri, Feb 17, 2006 at 01:26:04PM -0500, John Baldwin wrote..
> On Friday 17 February 2006 02:45, Wilko Bulte wrote:
> > On Fri, Feb 17, 2006 at 08:31:37AM +0100, Wilko Bulte wrote..
> >
> > > On Thu, Feb 16, 2006 at 06:15:15PM -0500, John Baldwin wrote..
> > >
> > > > On Thursday 16 February 2006 17:52, Wilko Bulte wrote:
> > > > > On Thu, Feb 16, 2006 at 11:23:11PM +0100, Sten Spans wrote..
> > > > >
> > > > > > On Thu, 16 Feb 2006, Wilko Bulte wrote:
> > > > > > >On Thu, Feb 16, 2006 at 12:19:11AM +0100, Wilko Bulte wrote..
> > > > > > >
> > > > > > >>Folks,
> > > > > > >>
> > > > > > >>I am getting the strong impression that there is something hosed
> > > > > > >>related to ATA at least on the DS10 and DS10L.  Test installs
> > > > > > >>on ATA disks work fine, but the first boot of the installed disk
> > > > > > >>fails miserably, typically the symptom is that the first shell
> > > > > > >> won't start or similar.  I got:
> > > > > > >>
> > > > > > >>Trying to mount root from ufs:/dev/ad0a
> > > > > > >>/libexec/ld-elf.so.1: /lib/libc.so.6: Unsupported relocation type
> > > > > > >>1611047936
> > > > > > >>in non-PLT relocations
> > > > > > >>
> > > > > > >>Observed on one DS10 and one DS10L.  The problem does not
> > > > > > >>occur when you use a SCSI disk (which I typically do so I missed
> > > > > > >> this before).
> > > > > > >>
> > > > > > >>Those of you who own this machine type and have a spare IDE
> > > > > > >> drive, can you please test it?  Other reports for different
> > > > > > >> system models are interesting too of course.
> > > > > > >
> > > > > > >A check of 5.5-BETA1 would be interesting too.
> > > > > >
> > > > > > I'll give it a go this weekend on my ds10
> > > > > > (which also normally runs scsi :)
> > > > >
> > > > > Looks like it is not only DS10, my 164SX has similar probs as well.
> > > > > 5.5-BETA1 and 6.1-BETA1 have it.  In total we have seen it on 2
> > > > > DS10L, one DS10 and a 164SX now.  SCSI is just fine.
> > > >
> > > > What if you turn dma off via 'hw.ata.dma=0' in the loader?
> > >
> > > Haven't tried yet.  DS10s are currently building BETA2
> >
> > 164SX with 6.1-BETA1 boot CD with or without DMA disabling has another
> > problem :/
> >
> > ppi0: <Parallel I/O> on ppbus0
> > ppc0: interrupting at ISA irq 7
> > sio0 at port 0x3f8-0x3ff irq 4 flags 0x10 on isa0
> > sio0: type 16550A, console
> > sio0: interrupting at ISA irq 4
> > sio1 at port 0x2f8-0x2ff irq 3 flags 0x80 on isa0
> > sio1: type 16550A
> > sio1: interrupting at ISA irq 3
> > Timecounter "i8254" frequency 1193182 Hz quality 0
> > Timecounter "alpha" frequency 533176524 Hz quality 800
> > Timecounters tick every 0.976 msec
> > md0: Preloaded image </boot/mfsroot> 4423680 bytes at 0xfffffc0000a28400
> > ad0: 38166MB <Seagate ST340016A 3.19> at ata0-master WDMA2
> > acd0: CDROM <TOSHIBA CD-ROM XM-6302B/1017> at ata1-master BIOSPIO
> >
> > fatal kernel trap:
> >
> >     trap entry     = 0x4 (unaligned access fault)
> >     cpuid          = 0
> >     faulting va    = 0xfffffc00011ce5c2
> >     opcode         = 0x28
> >     register       = 0x1
> >     pc             = 0xfffffc00004cec6c
> 
> Can you build a release with a debug kernel on it?  (Just leave the 
> DEBUG_FLAGS=-g line in GENERIC)  Then reproduce this and run gdb on the 
> kernel.debug file and do 'l *<value of pc>' to see where this is happening?

OK..  I could make the panic go away by zeroing the ata HD (using NetBSD &
dd ;-).  So whatever it was exactly that caused the panic, it definitely
was dependent on the data already resident on the HD.

Unfortunately both 5.5-BETA3 and 6.1-BETA3 now give me
the problem that sysinstall does not show a disk to partition.
The disk itself is probed as ad0 alright.

Bah!

-- 
Wilko Bulte				wilko at FreeBSD.org