So I spent most of the night trying to assemble a new (to me; used) travel laptop using Ubuntu 12.10 LTS and Xen. I decided to get the user experience of installing Xen as described in one of the Ubuntu docs. Well I got it installed, but it doesn''t run. The Dom0 kernel panics (e.g., kernel panic -not syncing: Attempted to kill init), and most of the useful information is gone from the screen by the time it dies (there is part of a stackdump above it, but no sign of what actually caused the dump). This laptop has no serial port to capture the log and it is too early in the boot to try a network console. I''m at the point where I am thinking of trashing this altogether and trying Fedora. Is there anything I can do to capture the log so I can pinpoint a cause for this? Thanks, Russ
On Sat, 2013-08-03 at 11:06 -0400, Russ Pavlicek wrote:> So I spent most of the night trying to assemble a new (to me; used) > travel laptop using Ubuntu 12.10 LTS and Xen. I decided to get the > user experience of installing Xen as described in one of the Ubuntu > docs. > > Well I got it installed, but it doesn''t run. The Dom0 kernel panics > (e.g., kernel panic -not syncing: Attempted to kill init),This usually indicates that your init process has exited for some reason, often a userspace (perhaps initrd related) reason. I assume it boots with the non-Xen option just fine?> and most of > the useful information is gone from the screen by the time it dies > (there is part of a stackdump above it, but no sign of what actually > caused the dump). This laptop has no serial port to capture the log > and it is too early in the boot to try a network console. I''m at the > point where I am thinking of trashing this altogether and trying > Fedora. > > Is there anything I can do to capture the log so I can pinpoint a > cause for this?If you add "noreboot" to the hypervisor command line then it won''t automatically reboot when dom0 crashes and you may get some additional info. You could also try using the vga=ask option (also hypervisor cmdline) to select an increased resolution which includes more lines. A digital photo of whatever you manage to get might be enough to ring a bell with someone. Command line reference is at http://xenbits.xen.org/docs/unstable/misc/xen-command-line.html Ian.> > Thanks, > > Russ > > _______________________________________________ > Xen-users mailing list > Xen-users@lists.xen.org > http://lists.xen.org/xen-users
Ian, Thanks for the tips and the command line reference. After much observation and experimentation, I determined that using "acpi=noirq" in the Xen command line makes it all work. The remaining question is "Why?". The linux boot without Xen doesn''t need it, but it is essential when booting up with Xen. The doc says this is rare and the machine should be blacklisted. But I still have no idea why it is needed under Xen and not under straight Linux. The laptop is a Compaq/HP 6715b with AMD Turion 64 X2 processor. Thanks, Russ On Sat, Aug 3, 2013 at 11:29 AM, Ian Campbell <Ian.Campbell@citrix.com> wrote:> On Sat, 2013-08-03 at 11:06 -0400, Russ Pavlicek wrote: >> So I spent most of the night trying to assemble a new (to me; used) >> travel laptop using Ubuntu 12.10 LTS and Xen. I decided to get the >> user experience of installing Xen as described in one of the Ubuntu >> docs. >> >> Well I got it installed, but it doesn''t run. The Dom0 kernel panics >> (e.g., kernel panic -not syncing: Attempted to kill init), > > This usually indicates that your init process has exited for some > reason, often a userspace (perhaps initrd related) reason. > > I assume it boots with the non-Xen option just fine? > >> and most of >> the useful information is gone from the screen by the time it dies >> (there is part of a stackdump above it, but no sign of what actually >> caused the dump). This laptop has no serial port to capture the log >> and it is too early in the boot to try a network console. I''m at the >> point where I am thinking of trashing this altogether and trying >> Fedora. >> >> Is there anything I can do to capture the log so I can pinpoint a >> cause for this? > > If you add "noreboot" to the hypervisor command line then it won''t > automatically reboot when dom0 crashes and you may get some additional > info. > > You could also try using the vga=ask option (also hypervisor cmdline) to > select an increased resolution which includes more lines. > > A digital photo of whatever you manage to get might be enough to ring a > bell with someone. > > Command line reference is at > http://xenbits.xen.org/docs/unstable/misc/xen-command-line.html > > Ian. > >> >> Thanks, >> >> Russ >> >> _______________________________________________ >> Xen-users mailing list >> Xen-users@lists.xen.org >> http://lists.xen.org/xen-users > >
I''m not 100% sure what this option is needed for. The first thing I would try is a check for a newer BIOS for the system, especially if it is a bit older and doubly so since it is a laptop. The reason is that I think the error relates to a dodgy set of ACPI tables describing the interrupts which Linus is somehow able to work around but Xen isn''t. You might find that even though Linux is working it is complaining in dmesg during boot and e.g. only using the legacy IRQs. Ian. On Sat, 2013-08-03 at 22:46 -0400, Russ Pavlicek wrote:> Ian, > > Thanks for the tips and the command line reference. > > After much observation and experimentation, I determined that using > "acpi=noirq" in the Xen command line makes it all work. The remaining > question is "Why?". The linux boot without Xen doesn''t need it, but > it is essential when booting up with Xen. The doc says this is rare > and the machine should be blacklisted. But I still have no idea why > it is needed under Xen and not under straight Linux. > > The laptop is a Compaq/HP 6715b with AMD Turion 64 X2 processor. > > Thanks, > > Russ > > On Sat, Aug 3, 2013 at 11:29 AM, Ian Campbell <Ian.Campbell@citrix.com> wrote: > > On Sat, 2013-08-03 at 11:06 -0400, Russ Pavlicek wrote: > >> So I spent most of the night trying to assemble a new (to me; used) > >> travel laptop using Ubuntu 12.10 LTS and Xen. I decided to get the > >> user experience of installing Xen as described in one of the Ubuntu > >> docs. > >> > >> Well I got it installed, but it doesn''t run. The Dom0 kernel panics > >> (e.g., kernel panic -not syncing: Attempted to kill init), > > > > This usually indicates that your init process has exited for some > > reason, often a userspace (perhaps initrd related) reason. > > > > I assume it boots with the non-Xen option just fine? > > > >> and most of > >> the useful information is gone from the screen by the time it dies > >> (there is part of a stackdump above it, but no sign of what actually > >> caused the dump). This laptop has no serial port to capture the log > >> and it is too early in the boot to try a network console. I''m at the > >> point where I am thinking of trashing this altogether and trying > >> Fedora. > >> > >> Is there anything I can do to capture the log so I can pinpoint a > >> cause for this? > > > > If you add "noreboot" to the hypervisor command line then it won''t > > automatically reboot when dom0 crashes and you may get some additional > > info. > > > > You could also try using the vga=ask option (also hypervisor cmdline) to > > select an increased resolution which includes more lines. > > > > A digital photo of whatever you manage to get might be enough to ring a > > bell with someone. > > > > Command line reference is at > > http://xenbits.xen.org/docs/unstable/misc/xen-command-line.html > > > > Ian. > > > >> > >> Thanks, > >> > >> Russ > >> > >> _______________________________________________ > >> Xen-users mailing list > >> Xen-users@lists.xen.org > >> http://lists.xen.org/xen-users > > > >
Well, this is one of the few times I''ve found that upgrading a laptop''s BIOS made things seriously worse. :( The new BIOS introduces a new problem: it makes Linux think that the system is overheating, resulting in the processor moving at a snail''s pace. And, to make things even better, once you''ve upgraded the BIOS, you can''t downgrade. Just lovely. This problem has prompted folks to come up with patches like this: http://www.spinics.net/lists/linux-acpi/msg40108.html However, I don''t know that any form of this patch has made its way into the stable tree and onto an actual distribution. So I have a laptop which is sluggish under Windows and useless under Linux. Guess I need to find a patched kernel somewhere and see what I can do. So much for trying to capture the "out of box" experience, I guess. Thanks all, Russ On Mon, Aug 5, 2013 at 3:32 AM, Ian Campbell <Ian.Campbell@citrix.com> wrote:> I''m not 100% sure what this option is needed for. The first thing I > would try is a check for a newer BIOS for the system, especially if it > is a bit older and doubly so since it is a laptop. > > The reason is that I think the error relates to a dodgy set of ACPI > tables describing the interrupts which Linus is somehow able to work > around but Xen isn''t. You might find that even though Linux is working > it is complaining in dmesg during boot and e.g. only using the legacy > IRQs. > > Ian. > > On Sat, 2013-08-03 at 22:46 -0400, Russ Pavlicek wrote: >> Ian, >> >> Thanks for the tips and the command line reference. >> >> After much observation and experimentation, I determined that using >> "acpi=noirq" in the Xen command line makes it all work. The remaining >> question is "Why?". The linux boot without Xen doesn''t need it, but >> it is essential when booting up with Xen. The doc says this is rare >> and the machine should be blacklisted. But I still have no idea why >> it is needed under Xen and not under straight Linux. >> >> The laptop is a Compaq/HP 6715b with AMD Turion 64 X2 processor. >> >> Thanks, >> >> Russ >> >> On Sat, Aug 3, 2013 at 11:29 AM, Ian Campbell <Ian.Campbell@citrix.com> wrote: >> > On Sat, 2013-08-03 at 11:06 -0400, Russ Pavlicek wrote: >> >> So I spent most of the night trying to assemble a new (to me; used) >> >> travel laptop using Ubuntu 12.10 LTS and Xen. I decided to get the >> >> user experience of installing Xen as described in one of the Ubuntu >> >> docs. >> >> >> >> Well I got it installed, but it doesn''t run. The Dom0 kernel panics >> >> (e.g., kernel panic -not syncing: Attempted to kill init), >> > >> > This usually indicates that your init process has exited for some >> > reason, often a userspace (perhaps initrd related) reason. >> > >> > I assume it boots with the non-Xen option just fine? >> > >> >> and most of >> >> the useful information is gone from the screen by the time it dies >> >> (there is part of a stackdump above it, but no sign of what actually >> >> caused the dump). This laptop has no serial port to capture the log >> >> and it is too early in the boot to try a network console. I''m at the >> >> point where I am thinking of trashing this altogether and trying >> >> Fedora. >> >> >> >> Is there anything I can do to capture the log so I can pinpoint a >> >> cause for this? >> > >> > If you add "noreboot" to the hypervisor command line then it won''t >> > automatically reboot when dom0 crashes and you may get some additional >> > info. >> > >> > You could also try using the vga=ask option (also hypervisor cmdline) to >> > select an increased resolution which includes more lines. >> > >> > A digital photo of whatever you manage to get might be enough to ring a >> > bell with someone. >> > >> > Command line reference is at >> > http://xenbits.xen.org/docs/unstable/misc/xen-command-line.html >> > >> > Ian. >> > >> >> >> >> Thanks, >> >> >> >> Russ >> >> >> >> _______________________________________________ >> >> Xen-users mailing list >> >> Xen-users@lists.xen.org >> >> http://lists.xen.org/xen-users >> > >> > > >
On Wed, 2013-08-07 at 00:07 -0400, Russ Pavlicek wrote:> Well, this is one of the few times I''ve found that upgrading a > laptop''s BIOS made things seriously worse. :( > > The new BIOS introduces a new problem: it makes Linux think that the > system is overheating, resulting in the processor moving at a snail''s > pace. And, to make things even better, once you''ve upgraded the BIOS, > you can''t downgrade. Just lovely.Oh wow, I''m sorry to hear that :-(> This problem has prompted folks to come up with patches like this: > > http://www.spinics.net/lists/linux-acpi/msg40108.html > > However, I don''t know that any form of this patch has made its way > into the stable tree and onto an actual distribution.It appears to have hit mainline in February 2012 in linux.git 9bcb8118965a which seems to have been part of v3.4-rc6. So anything newer than that ought to have it. Even if you are running something earlier the distro may have picked it up, I can''t speak for Ubuntu LTS but the most recent Ubuntu ought to be new enough.> So I have a laptop which is sluggish under Windows and useless under Linux. Guess > I need to find a patched kernel somewhere and see what I can do.You can probably just install the latest Ubuntu kernel even without upgrading the whole distro. At least I''ve done that with Debian often... Ian.
>You can probably just install the latest Ubuntu kernel even without >upgrading the whole distro. At least I''ve done that with Debian often...Except that I need to be booted from the new kernel just to do the upgrade. Once the thermal condition is tripped, the machine becomes virtually useless. So I guess I will see if the latest Ubuntu release can boot up correctly and then try to upgrade or reinstall. If not, I may need to move the hard drive into another machine and try to upgrade the kernel there. Uggh. Thanks, Russ On Wed, Aug 7, 2013 at 4:26 AM, Ian Campbell <Ian.Campbell@citrix.com> wrote:> On Wed, 2013-08-07 at 00:07 -0400, Russ Pavlicek wrote: >> Well, this is one of the few times I''ve found that upgrading a >> laptop''s BIOS made things seriously worse. :( >> >> The new BIOS introduces a new problem: it makes Linux think that the >> system is overheating, resulting in the processor moving at a snail''s >> pace. And, to make things even better, once you''ve upgraded the BIOS, >> you can''t downgrade. Just lovely. > > Oh wow, I''m sorry to hear that :-( > >> This problem has prompted folks to come up with patches like this: >> >> http://www.spinics.net/lists/linux-acpi/msg40108.html >> >> However, I don''t know that any form of this patch has made its way >> into the stable tree and onto an actual distribution. > > It appears to have hit mainline in February 2012 in linux.git > 9bcb8118965a which seems to have been part of v3.4-rc6. So anything > newer than that ought to have it. Even if you are running something > earlier the distro may have picked it up, I can''t speak for Ubuntu LTS > but the most recent Ubuntu ought to be new enough. > >> So I have a laptop which is sluggish under Windows and useless under Linux. Guess >> I need to find a patched kernel somewhere and see what I can do. > > You can probably just install the latest Ubuntu kernel even without > upgrading the whole distro. At least I''ve done that with Debian often... > > Ian. > >
On Wed, 2013-08-07 at 09:51 -0400, Russ Pavlicek wrote:> >You can probably just install the latest Ubuntu kernel even without > >upgrading the whole distro. At least I''ve done that with Debian often... > > Except that I need to be booted from the new kernel just to do the > upgrade. Once the thermal condition is tripped, the machine becomes > virtually useless. > > So I guess I will see if the latest Ubuntu release can boot up > correctly and then try to upgrade or reinstall. If not, I may need to > move the hard drive into another machine and try to upgrade the kernel > there. Uggh.A newer Ubuntu live CD might be sufficient to boot and chroot into the HDD install to update your kernel. Or you might be able to pxeboot a rescue image. If Ubuntu is like Debian then the text mode netboot installation media has rescue functionality. Or you could do the upgrade in a cold storage locker ;-) Ian.
So I put aside the notions of seeing the laptop covered in blood at the meat storage locker in town (as appropriate as it seemed at the time, considering how much pain it has put me through) and installed Ubuntu 13.04 server. It works! It''s alive! Still have to determine why the onboard wireless is blowing chunks on this rev when it worked fine before, but that''s small potatoes. And now I have Xen running 4.2 instead of 4.1, which isn''t bad either. Still need the "acpi=noirq" on Xen itself, but Dom0 appears to function as expected. Thanks for the assistance! Russ On Wed, Aug 7, 2013 at 9:58 AM, Ian Campbell <Ian.Campbell@citrix.com> wrote:> On Wed, 2013-08-07 at 09:51 -0400, Russ Pavlicek wrote: >> >You can probably just install the latest Ubuntu kernel even without >> >upgrading the whole distro. At least I''ve done that with Debian often... >> >> Except that I need to be booted from the new kernel just to do the >> upgrade. Once the thermal condition is tripped, the machine becomes >> virtually useless. >> >> So I guess I will see if the latest Ubuntu release can boot up >> correctly and then try to upgrade or reinstall. If not, I may need to >> move the hard drive into another machine and try to upgrade the kernel >> there. Uggh. > > A newer Ubuntu live CD might be sufficient to boot and chroot into the > HDD install to update your kernel. Or you might be able to pxeboot a > rescue image. If Ubuntu is like Debian then the text mode netboot > installation media has rescue functionality. > > Or you could do the upgrade in a cold storage locker ;-) > > Ian. >