[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: [Xen-users] OpenSolaris 10 on x86_32 Xen-VTx



Andrew,

Thanks for the suggestions. I confirmed that the qemu-dm is not
stripped. I am going to start adding printks to the ioemu. Here is what
the qemu-log tells me.

Aravindh

domid: 16
eport: 3
qemu: the number of cpus is 1
shared page at pfn:20483, mfn: a476a
HVM Loader
Loading ROMBIOS ...
Loading Cirrus VGABIOS ...
Loading VMXAssist ...
VMX go ...
VMXAssist (Jan 26 2006)
Memory size 512 MB
E820 map:
0000000000000000 - 000000000009F800 (RAM)
000000000009F800 - 00000000000A0000 (Reserved)
00000000000A0000 - 00000000000C0000 (Type 16)
00000000000F0000 - 0000000000100000 (Reserved)
0000000000100000 - 000000001FFFE000 (RAM)
000000001FFFF000 - 0000000020000000 (Type 17)
000000001FFFE000 - 000000001FFFF000 (Type 18)
0000000020000000 - 0000000020003000 (ACPI NVS)
0000000020003000 - 000000002000D000 (ACPI Data)
00000000FEC00000 - 0000000100000000 (Type 16)

Start BIOS ...
Starting emulated 16-bit real-mode: ip=F000:FFF0
 rombios.c,v 1.138 2005/05/07 15:55:26 vruppert Exp $
HVM_PIT:guest init pit channel 0!
HVM_PIT:pass info 0xc00e90b to HV!
Remapping master: ICW2 0x8 -> 0x20
Remapping slave: ICW2 0x70 -> 0x28
VGABios $Id: vgabios.c,v 1.61 2005/05/24 16:50:50 vruppert Exp $
HVMAssist BIOS, 1 cpu, $Revision: 1.138 $ $Date: 2005/05/07 15:55:26 $

ata0-0: PCHS=4161/16/63 translation=lba LCHS=520/128/63
ata0 master: QEMU HARDDISK ATA-2 Hard-Disk (2048 MBytes)
ata0  slave: Unknown device
ata1 master: QEMU CD-ROM ATAPI-4 CD-Rom/DVD-Rom
ata1  slave: Unknown device

Booting from CD-Rom...
int13_harddisk: function 41, unmapped device for ELDL=81
int13_harddisk: function 08, unmapped device for ELDL=81
*** int 15h function AX=00C0, BX=0000 not yet supported!
*** int 15h function AX=5300, BX=0000 not yet supported!
*** int 15h function AX=5304, BX=0000 not yet supported!

> -----Original Message-----
> From: Andrew D. Ball [mailto:aball@xxxxxxxxxx]
> Sent: Monday, January 30, 2006 1:58 PM
> To: Puthiyaparambil, Aravindh
> Subject: RE: [Xen-users] OpenSolaris 10 on x86_32 Xen-VTx
> 
> image.py probably won't help too much here.  The current device model
> code is in tools/ioemu .
> 
> You may want to look for
> /var/log/qemu-dm*
> and
> /var/log/ioemu*
> 
> You may also want to check that /usr/{lib,lib64}/xen/bin/qemu-dm has
not
> been stripped ('file /usr/lib/xen/bin/qemu-dm' or
> 'file /usr/lib64/xen/bin/qemu-dm' will tell you).
> 
> I'm assuming that the device model is dying from some sort of
> segmentation fault.  Starting it with 'nohup' could also help you be
> able to perhaps attach to the qemu-dm processes -- the device model is
> launched with os.spawnve in tools/python/xen/xend/image.py .
> 
> The current device model is painful.  It does work well enough for
proof
> of concept with various flavors of Windows.  Your work to get it
working
> with OpenSolaris is interesting.
> 
> Good luck!
> Andrew
> 
> On Mon, 2006-01-30 at 12:08 -0500, Puthiyaparambil, Aravindh wrote:
> > Nope, I am not able to get a stack trace when qemu-dies. I tried the
> > CTRL-F2 and CTRL-3 in the SDL windows but that produces no results.
> > Looks like I will have to start adding prinks in image.py to begin
with.
> >
> > Thanks,
> > Aravindh
> >
> > > -----Original Message-----
> > > From: Andrew D. Ball [mailto:aball@xxxxxxxxxx]
> > > Sent: Friday, January 27, 2006 2:18 PM
> > > To: Puthiyaparambil, Aravindh
> > > Subject: RE: [Xen-users] OpenSolaris 10 on x86_32 Xen-VTx
> > >
> > > Hrm -- that's a good reason for the patch to not make it in.  Are
you
> > > able to get a stack trace if you don't set a breakpoint and
qemu-dm
> > dies
> > > as usual?
> > >
> > > I think there may be some logging capability taken directly from
QEMU
> > as
> > > well -- try pressing CTRL-F2 and CTRL-3 in the SDL window that
pops
> > up.
> > > Debugging the device model has never been simple for me. I was
hoping
> > > that I could just use gdb with it, but if that's not really
feasible
> > due
> > > to timing issues, we'll have to think of something else.
> > >
> > > I'm assuming that the qemu-dm processes die for some reason like a
> > > segmentation fault, but with them being launched by xend it's
harder
> > to
> > > find that out, since 'ulimit -c unlimited" has never make them
dump
> > core
> > > that I have been able to find.
> > >
> > > I really just want a core dump :-)
> > >
> > > There's always the old well-placed printf method or regular qemu
logs
> > > that you ask qemu-dm to produce with arguments [provided that code
is
> > > still there].  The last time I checked it wasn't incredibly easy
to
> > pass
> > > new arguments to qemu-dm -- I think I needed to modify
> > > tools/python/xen/xend/image.py in the 'createDeviceModel'
function.
> > >
> > > Good luck!
> > > Andrew
> > >
> > > On Thu, 2006-01-26 at 18:10 -0500, Puthiyaparambil, Aravindh
wrote:
> > > > I am able to do only "one" thing after ddd starts up. If I say
"run"
> > > > without any breakpoint then qemu executes and the Solaris
install
> > hangs
> > > > as always. If I set a breakpoint, then the code stops there. In
both
> > the
> > > > cases ddd/gdb is not responsive to any commands :-(
> > > >
> > > > Thanks,
> > > > Aravindh
> > > >
> > > > > -----Original Message-----
> > > > > From: Andrew D. Ball [mailto:aball@xxxxxxxxxx]
> > > > > Sent: Thursday, January 26, 2006 5:33 PM
> > > > > To: Puthiyaparambil, Aravindh
> > > > > Subject: RE: [Xen-users] OpenSolaris 10 on x86_32 Xen-VTx
> > > > >
> > > > > You're probably not doing anything wrong.  Can you continue
> > execution
> > > > > here?  What I really wanted when working on the patch was to
be
> > able
> > > > to
> > > > > get a stack trace from a dead qemu-dm process.  I kept seeing
> > defunct
> > > > > qemu-dm processes and 'ulimit -c' wasn't working to get core
dumps
> > > > from
> > > > > them.
> > > > >
> > > > > I think ddd is much like command-line gdb, or at least can be
if
> > > > that's
> > > > > what you're familiar with.  It's got some buttons for common
> > > > > functionality and displays the source code, but it still
> > communicates
> > > > > with gcc at the bottom.
> > > > >
> > > > > Good luck!
> > > > > Andrew
> > > > >
> > > > > On Thu, 2006-01-26 at 16:12 -0500, Puthiyaparambil, Aravindh
> > wrote:
> > > > > > Andrew,
> > > > > >
> > > > > > I followed your instructions in the patch and now when I
bring
> > up
> > > > the
> > > > > > DomVT, ddd fires. I am able to set breakpoints and the
program
> > > > execution
> > > > > > stops at the first one. But I am not able to do much beyond
as
> > ddd
> > > > > > reports "waiting until GDB is ready". Am I doing anything
wrong
> > > > here?
> > > > > >
> > > > > > Thanks,
> > > > > > Aravindh
> > > > > >
> > > > > > > -----Original Message-----
> > > > > > > From: Andrew D. Ball [mailto:aball@xxxxxxxxxx]
> > > > > > > Sent: Wednesday, January 25, 2006 12:14 PM
> > > > > > > To: Puthiyaparambil, Aravindh
> > > > > > > Subject: RE: [Xen-users] OpenSolaris 10 on x86_32 Xen-VTx
> > > > > > >
> > > > > > > That means that the device model has died on you.  There's
> > only
> > > > one
> > > > > > > device model at the moment -- it's mostly code taken from
the
> > qemu
> > > > > > > hypervisor to emulate whatever hardware is needed for full
> > > > > > > virtualization.
> > > > > > >
> > > > > > > You may want to try getting and building this tree of Xen
> > > > > > >
> > > > > > > http://xenbits.xensource.com/ext/xen-unstable-hvm.hg
> > > > > > >
> > > > > > > I have seen much more stability with the device model with
> > this
> > > > one.
> > > > > > > You will need to use 'hvm' instead of 'vmx' for
configurations
> > > > here,
> > > > > > but
> > > > > > > I don't know of any other differences in the interface.
> > > > > > >
> > > > > > > I put out a patch some time ago to try to help debug the
> > device
> > > > model,
> > > > > > > if you'd like to try to find out where and why it's
crashing:
> > > > > > >
> > > > > > >
> > > > > >
> > > >
> >
http://lists.xensource.com/archives/html/xen-devel/2006-01/msg00103.html
> > > > > > >
> > > > > > > Good luck!
> > > > > > > Andrew
> > > > > > > --
> > > > > > > Andrew D. Ball
> > > > > > > aball@xxxxxxxxxx
> > > > > > >
> > > > > > > "Festina Lente" $\approx$ "Make hast slowly"
> > > > > > >   -- Caesar Augustus
> > > > > > >
> > > > > > >
> > > > > > > On Wed, 2006-01-25 at 11:32 -0500, Puthiyaparambil,
Aravindh
> > > > wrote:
> > > > > > > > Andrew,
> > > > > > > >
> > > > > > > > Yes, I see defunct qemu-dm processes. But I am not sure
what
> > it
> > > > > > means.
> > > > > > > > Will OS10 require a separate device model?
> > > > > > > >
> > > > > > > > Thanks,
> > > > > > > > Aravindh
> > > > > > > >
> > > > > > > > root     15753  0.0  0.0      0     0 ?        Z
Jan24
> > 0:04
> > > > > > > > [qemu-dm] <defunct>
> > > > > > > > root     16334  0.0  0.0      0     0 ?        Z
Jan24
> > 0:06
> > > > > > > > [qemu-dm] <defunct>
> > > > > > > > root     16917  0.0  0.0      0     0 ?        Z
Jan24
> > 0:01
> > > > > > > > [qemu-dm] <defunct>
> > > > > > > > root     17022  0.0  0.0      0     0 ?        Z
Jan24
> > 0:01
> > > > > > > > [qemu-dm] <defunct>
> > > > > > > > root     17156  0.0  0.0      0     0 ?        Z
Jan24
> > 0:17
> > > > > > > > [qemu-dm] <defunct>
> > > > > > > > root     18893  0.0  0.0      0     0 ?        Z
Jan24
> > 0:00
> > > > > > > > [qemu-dm] <defunct>
> > > > > > > > root     18943  0.0  0.0      0     0 ?        Z
Jan24
> > 0:01
> > > > > > > > [qemu-dm] <defunct>
> > > > > > > > root     19400  0.0  0.0      0     0 ?        Z
Jan24
> > 0:01
> > > > > > > > [qemu-dm] <defunct>
> > > > > > > > root     19541  0.0  0.0      0     0 ?        Z
Jan24
> > 0:01
> > > > > > > > [qemu-dm] <defunct>
> > > > > > > > root     27794  1.3  1.0  31604  5576 ?        S
12:36
> > 0:01
> > > > > > > > /usr/lib/xen/bin/qemu-dm -d 18 -p 3 -m 512 -cdrom
/dev/hdd
> > -boot
> > > > d
> > > > > > > > -nic-ne2000 -vcpus 1 -domain-name OpenSolaris10 -hda
> > > > > > > > /root/xen/vm/vmxsol.img -nics 0
> > > > > > > > root     27806  0.0  1.0  31604  5576 ?        S
12:36
> > 0:00
> > > > > > > > /usr/lib/xen/bin/qemu-dm -d 18 -p 3 -m 512 -cdrom
/dev/hdd
> > -boot
> > > > d
> > > > > > > > -nic-ne2000 -vcpus 1 -domain-name OpenSolaris10 -hda
> > > > > > > > /root/xen/vm/vmxsol.img -nics 0
> > > > > > > > root     27807  0.0  1.0  31604  5576 ?        S
12:36
> > 0:00
> > > > > > > > /usr/lib/xen/bin/qemu-dm -d 18 -p 3 -m 512 -cdrom
/dev/hdd
> > -boot
> > > > d
> > > > > > > > -nic-ne2000 -vcpus 1 -domain-name OpenSolaris10 -hda
> > > > > > > > /root/xen/vm/vmxsol.img -nics 0
> > > > > > > >
> > > > > > > > > -----Original Message-----
> > > > > > > > > From: Andrew D. Ball [mailto:aball@xxxxxxxxxx]
> > > > > > > > > Sent: Wednesday, January 25, 2006 9:36 AM
> > > > > > > > > To: Puthiyaparambil, Aravindh
> > > > > > > > > Subject: Re: [Xen-users] OpenSolaris 10 on x86_32
Xen-VTx
> > > > > > > > >
> > > > > > > > > Do you see any defunct processes if you do
> > > > > > > > >
> > > > > > > > > ps aux | grep qemu-dm
> > > > > > > > >
> > > > > > > > > ?
> > > > > > > > >
> > > > > > > > > Good luck.
> > > > > > > > > Andrew
> > > > > > > > > --
> > > > > > > > > Andrew D. Ball
> > > > > > > > > aball@xxxxxxxxxx
> > > > > > > > >
> > > > > > > > > "Festina Lente" $\approx$ "Make hast slowly"
> > > > > > > > >   -- Caesar Augustus
> > > > > > > > >
> > > > > > > > > On Tue, 2006-01-24 at 17:48 -0500, Puthiyaparambil,
> > Aravindh
> > > > > > wrote:
> > > > > > > > > > I am trying to install x86_32 OpenSolaris 10 in a VT
> > domain
> > > > on
> > > > > > > > x86_32
> > > > > > > > > > Xen (changeset 8634). The domain boots off the
install
> > CD
> > > > and I
> > > > > > see
> > > > > > > > the
> > > > > > > > > > grub menu where I choose "Solaris". I then see the
> > "Booting
> > > > > > Solaris"
> > > > > > > > and
> > > > > > > > > > the loading kernel and modules lines but after that
the
> > Qemu
> > > > > > window
> > > > > > > > goes
> > > > > > > > > > blank. I have listed my config file below. Any idea
how
> > to
> > > > get
> > > > > > > > around
> > > > > > > > > > this issue?
> > > > > > > > > >
> > > > > > > > > > Thanks,
> > > > > > > > > > Aravindh
> > > > > > > > > >
> > > > > > > > > >
> > > > > > > > > > OpesnSolaris10 DomVT Config file
> > > > > > > > > > --------------------------------
> > > > > > > > > > import os, re
> > > > > > > > > > arch = os.uname()[4]
> > > > > > > > > > if re.search('64', arch):
> > > > > > > > > >     arch_libdir = 'lib64'
> > > > > > > > > > else:
> > > > > > > > > >     arch_libdir = 'lib'
> > > > > > > > > >
> > > > > > > > > > kernel = "/usr/lib/xen/boot/vmxloader"
> > > > > > > > > > builder='vmx'
> > > > > > > > > > memory = 512
> > > > > > > > > > name = "OpenSolaris10"
> > > > > > > > > > vif = [ 'type=ioemu, bridge=xenbr0' ]
> > > > > > > > > > disk = [ 'file:/root/xen/vm/vmxsol.img,ioemu:hda,w'
]
> > > > > > > > > > device_model = '/usr/' + arch_libdir +
> > '/xen/bin/qemu-dm'
> > > > > > > > > > cdrom = '/dev/hdd'
> > > > > > > > > > boot = 'd'
> > > > > > > > > > sdl=1
> > > > > > > > > > vnc=0
> > > > > > > > > > vncviewer=0
> > > > > > > > > > ne2000=0
> > > > > > > > > >
> > > > > > > > > > _______________________________________________
> > > > > > > > > > Xen-users mailing list
> > > > > > > > > > Xen-users@xxxxxxxxxxxxxxxxxxx
> > > > > > > > > > http://lists.xensource.com/xen-users
> > > > > > > > > >
> > > > > > > > >
> > > > > > > >
> > > > > > > >
> > > > > > >
> > > > > >
> > > > > >
> > > >
> > > >
> >
> >


_______________________________________________
Xen-users mailing list
Xen-users@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-users


 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.