Hello, I''m using the latest b75 ISO to try and get a solaris dom0 running on my laptop, a Sony SZ-330P. The Xen LiveCD works just fine, so I know that theoretically it should work. Build 75 also works just fine on my Sony, including sound and video, etc. But when I try and boot the Xen kernel, the system hangs. There''s no panic printed to console. If I pass a -v option at boot to .../unix, then I see that SMBIOS is the last thing to load before the hack. When I boot with -v non-Xen, the cpu coming online is the next thing to happen after SMBIOS loads. Here''s relevant dmesg lines: Oct 14 19:38:16 tksz330-b72 genunix: [ID 936769 kern.notice] cmdk0 is /pci@0,0/pci-ide@1f,2/ide@0/cmdk@0,0 Oct 14 19:38:16 tksz330-b72 unix: [ID 190185 kern.notice] SMBIOS v2.31 loaded (912 bytes) Oct 14 19:38:16 tksz330-b72 genunix: [ID 408114 kern.notice] /cpus (cpunex0) online Oct 14 19:38:16 tksz330-b72 acpica: [ID 426109 kern.notice] ACPI (exconfig-0558): Dynamic SSDT Load - OemId [ Sony] OemTableId [ N0] [20060721] Oct 14 19:38:16 tksz330-b72 cpunex: [ID 114370 kern.notice] cpu0 at cpus0 Oct 14 19:38:16 tksz330-b72 genunix: [ID 408114 kern.notice] /cpus/cpu@0 (cpudrv0) online Attached is a possibly useless picture of the hung console. It sits indefinitely at the end of the SMBIOS line. Any pointers on steps I can take to resolve this hang would be greatly appreciated. Thanks, -Tim -- /\ Timothy Kennedy \ \ IT Technologist V / \ \ / GCS Infrastructure Management (GCSIM) / /\\// /\ Tier-3 Infrastructure OS Team \/ //\\/ / Email: Timothy.Kennedy@Sun.COM / \ \ / Phone: +1-703-636-0531 \ \ AIM: tkSUNW \/ WWW: http://blogs.sun.com/tkblog
Timothy Kennedy wrote:> > Hello, > > I''m using the latest b75 ISO to try and get a solaris dom0 running on my > laptop, a Sony SZ-330P. The Xen LiveCD works just fine, so I know that > theoretically it should work. Build 75 also works just fine on my Sony, > including sound and video, etc. > > But when I try and boot the Xen kernel, the system hangs. There''s no > panic printed to console. If I pass a -v option at boot to .../unix, > then I see that SMBIOS is the last thing to load before the hack. > > When I boot with -v non-Xen, the cpu coming online is the next thing to > happen after SMBIOS loads. Here''s relevant dmesg lines: > > Oct 14 19:38:16 tksz330-b72 genunix: [ID 936769 kern.notice] cmdk0 is > /pci@0,0/pci-ide@1f,2/ide@0/cmdk@0,0 > Oct 14 19:38:16 tksz330-b72 unix: [ID 190185 kern.notice] SMBIOS v2.31 > loaded (912 bytes) > Oct 14 19:38:16 tksz330-b72 genunix: [ID 408114 kern.notice] /cpus > (cpunex0) online > Oct 14 19:38:16 tksz330-b72 acpica: [ID 426109 kern.notice] ACPI > (exconfig-0558): Dynamic SSDT Load - OemId [ Sony] OemTableId [ > N0] [20060721] > Oct 14 19:38:16 tksz330-b72 cpunex: [ID 114370 kern.notice] cpu0 at cpus0 > Oct 14 19:38:16 tksz330-b72 genunix: [ID 408114 kern.notice] /cpus/cpu@0 > (cpudrv0) online > > Attached is a possibly useless picture of the hung console. It sits > indefinitely at the end of the SMBIOS line. > > Any pointers on steps I can take to resolve this hang would be > greatly appreciated.Do you have a serial port on the machine? You can break into solaris if you can direct the Xen console to a serial port. If not, you can try a couple of things.. From the grub menu, edit the first module line and add a -kd. This will get you into the kernel debugger before you load unix. e.g. kernel$ /boot/$ISADIR/xen.gz dom0_mem=2048M module$ /platform/i86xpv/kernel/$ISADIR/unix /platform/i86xpv/kernel/$ISADIR/unix -kd From there, there''s a couple of things you can do.. You can get more detail on what modules are loading, etc... [0]> moddebug?W 0x80000000 moddebug: 0 = 0x80000000 [0]> :c The other thing you can do is to set a breakpoint on deadman... after a bunch of these, you should be the spot that hangs. You can see what was on the stack at the time and also look though the threadlist... e.g. [0]> moddebug?W 0x80000000 [0]> ::bp deadman [0]> :c [0]> ,20:c ... [0]> $c [0]> ::threadlist -v -- Mark Johnson <mark.johnson@sun.com> Sun Microsystems, Inc. (781) 442-0869
Timothy Kennedy wrote:> Mark Johnson wrote: >> >> The other thing you can do is to set a breakpoint on >> deadman... after a bunch of these, you should be the >> spot that hangs. You can see what was on the stack at the >> time and also look though the threadlist... >> >> [0]> moddebug?W 0x80000000 >> [0]> ::bp deadman >> [0]> :c >> [0]> ,20:c >> >> ... >> [0]> $c >> [0]> ::threadlist -v > > Thank you, Mark. That''s a very succinct tutorial. :) > > I don''t have a serial console, so some stuff scrolled by, but I did > finally see the panic that''s hitting me. I saw David Stewart''s email > about respinning b75, and in fact, I see b75a up on the internal server, > so I''ll download that and try again. > > In the meantime, here''s a pic of the crash. David, if this matches > your hangs/crashes, I''ll feel better about giving b75a a try.Yeah, this looks like the problem that we respun b75 for... MRJ -- Mark Johnson <mark.johnson@sun.com> Sun Microsystems, Inc. (781) 442-0869
Mark Johnson wrote:> > Yeah, this looks like the problem that we respun b75 for... >No joy with b75a for me. It still hangs, but at least with b75a kmdb itself isn''t panicing, so that''s a step in the right direction. kmdb is showing: target stopped at deadman: pushq %rbp If there''s anything under there that would be helpful, I can duplicate it when I get home and take a picture, or just copy it all down by hand. Thanks for the help, -Tim -- /\ Timothy Kennedy \ \ IT Technologist V / \ \ / GCS Infrastructure Management (GCSIM) / /\\// /\ Tier-3 Infrastructure OS Team \/ //\\/ / Email: Timothy.Kennedy@Sun.COM / \ \ / Phone: +1-703-636-0531 \ \ AIM: tkSUNW \/ WWW: http://blogs.sun.com/tkblog
Timothy Kennedy wrote:> > Hi Mark, > > I upgraded to b75a, and booting xen still results in cpu panics, and > dies. I don''t have any way other than my camera to capture screen > output of what the crash is, so all the crash info is in the attached > JPGs. > > I boot with -kd, then type: > [0]> moddebug?W 0x80000000 > [0]> ::bp deadman > [0]> :c > > result is the pictures below. picture1 is from one boot attempt, > the other two are from a separate boot attempt. > > Any ideas on what to try, or what''s causing this?I believe someone said that Casper ran into the same problem where the SUNWxvm package didn''t upgrade between b75 and b75a. you can try the following /ws/xvm-gate/public/bin/update_xvm non-debug or you can manually upgrade the package (using the path to your b75a mounted iso) # pkgrm SUNWxvmr # pkgadd -d /export/install/i386/nevada/75a/Solaris_11/Product SUNWxvmr MRJ -- Mark Johnson <mark.johnson@sun.com> Sun Microsystems, Inc. (781) 442-0869
Mark Johnson wrote:> > or you can manually upgrade the package (using the path > to your b75a mounted iso) > > # pkgrm SUNWxvmr > # pkgadd -d /export/install/i386/nevada/75a/Solaris_11/Product SUNWxvmr >Thank you, Mark. That did the trick. It''s all working swimingly now. :) -Tim -- /\ Timothy Kennedy \ \ IT Technologist V / \ \ / GCS Infrastructure Management (GCSIM) / /\\// /\ Tier-3 Infrastructure OS Team \/ //\\/ / Email: Timothy.Kennedy@Sun.COM / \ \ / Phone: +1-703-636-0531 \ \ AIM: tkSUNW \/ WWW: http://blogs.sun.com/tkblog