Xen 2.6.25 dom0 kernel ebuild

[andylyon]

Hi,

ich just tried the new patchset on a 32bit-system with i686-pc-linux-gnu-3.4.6. The following error occurs:

Code: Select all

 CC [M]  arch/x86/kernel/microcode_core.o
arch/x86/kernel/microcode_core-xen.c: In function `microcode_write':
arch/x86/kernel/microcode_core-xen.c:100: Fehler: »totalram_pages« nicht deklariert (erste Benutzung in dieser Funktion)
arch/x86/kernel/microcode_core-xen.c:100: Fehler: (Jeder nicht deklarierte Bezeichner wird nur einmal aufgeführt
arch/x86/kernel/microcode_core-xen.c:100: Fehler: für jede Funktion in der er auftritt.)
make[2]: *** [arch/x86/kernel/microcode_core.o] Fehler 1
make[1]: *** [arch/x86/kernel] Fehler 2
make: *** [arch/x86] Fehler 2

Deactivating the microcode update helped to compile successfully. It seem's there's missed something....

2bbionic

EDIT: Booting the kernel end to a kernel panic in smp-related stuff...

The microcode problem is easily fixed (Thanks to Bruce Edge), the following patch fixes the compile error and looks correct so it will be included in my next release or merged upstream.

diff -Naur arch/x86/kernel/microcode_core-xen.c.orig
arch/x86/kernel/microcode_core-xen.c

--- arch/x86/kernel/microcode_core-xen.c.orig 2009-10-14
10:45:42.000000000 -0700
+++ arch/x86/kernel/microcode_core-xen.c 2009-10-14
10:44:30.000000000 -0700
@@ -34,7 +34,7 @@
#include <linux/firmware.h>
#include <linux/uaccess.h>
#include <linux/vmalloc.h>
-
+#include <linux/swap.h>
#include <asm/microcode.h>
#include <asm/processor.h>

EDIT: microcode problem caused by me including the Xen part of a suse patch which is not required, next release will not have the patch so this problem will not happen.

I have tested 32 bit domU with two vcpu's and had no problems at all, I will try to test 32 bit dom0 but I don't actually have a 32 bit xen box here, only my laptop runs 32 bit linux and I've left it at home.

Andy

[babydoe]

I'm struggling trying to get 2.6.29-r4 dom0 compiled with gcc 3.4.6 (any flawors) on x86_64, it compiles fine with any flawors of gcc 4.3.4 (vanilla,hardenednopie,gentoo).

After going through this forum post i'm not sure if anyone got a dom0 compiled with gcc 3.4.6 on x86_64, can someone confirm (or deny) that it is possible to get it up and running with 3.4.6 ?
If yes, what do i need to do (except disabling xen ) ?

If this is not the case, can you confirm that any successful compilation here was done with 4.3.4 and that this is THE compiler to use ?

Much Kudos to Andy for the great job done bringing gentoo-xen out of the middle age ... tu

[andylyon]

I'm struggling trying to get 2.6.29-r4 dom0 compiled with gcc 3.4.6 (any flawors) on x86_64, it compiles fine with any flawors of gcc 4.3.4 (vanilla,hardenednopie,gentoo).

After going through this forum post i'm not sure if anyone got a dom0 compiled with gcc 3.4.6 on x86_64, can someone confirm (or deny) that it is possible to get it up and running with 3.4.6 ?
If yes, what do i need to do (except disabling xen ) ?

If this is not the case, can you confirm that any successful compilation here was done with 4.3.4 and that this is THE compiler to use ?

Much Kudos to Andy for the great job done bringing gentoo-xen out of the middle age ... tu

Kudos really must go to Jan Beulich as he does the hard work of forward porting the Xen patches, the changes I make to rebase the patches for vanilla are extremely minor, the total diff between the original Opensuse patches and my version is only 6k while the patches themselves total 9.7mb!

Anyway.. gcc -v returns v4.1.2 on both of my Gentoo systems, although I've just noticed my main system has 4.3.4 as well and test box also has 4.3.2, but both use 4.1.2 by default, gcc has been upgraded by portage but I'm in no rush to switch from 4.1.2, I stick to stable packages for gcc .

Andy

[andylyon]

I have uploaded xen-sources-2.6.31-r7.ebuild and 2.6.31 xen patches v7 (rebased patches.xen from opensuse kotd kernel-source-2.6.31.5-0.0.3.04f1d62.src.rpm)

32 bit dom0 is untested but all other dom0/domU archs tested and working.

Andy

[elilindner]

Hi,

Firstly thanks for the excellent patches, I have compiled using latest ebuild on AMD64 2.6.31-r7 successfully.

I'm having a little problem getting the xen power management to function. What are the specific config's I need to enable in order to get it working? As soon as I select xen Dom0 in processor setup, the cpu freq menu is empty.

Thanks for any help.

[andylyon]

Hi,

Firstly thanks for the excellent patches, I have compiled using latest ebuild on AMD64 2.6.31-r7 successfully.

I'm having a little problem getting the xen power management to function. What are the specific config's I need to enable in order to get it working? As soon as I select xen Dom0 in processor setup, the cpu freq menu is empty.

Thanks for any help.

Have you looked at http://wiki.xensource.com/xenwiki/xenpm ?

Andy

[Scorpion265]

Has anyone else had issues with the 2.6.31 kernel having a kernel panic when invoking heavy CPU usage? I seem to be having issues when I run an emerge command. Doesn't happen with the 2.6.29 kernel.

[andylyon]

Has anyone else had issues with the 2.6.31 kernel having a kernel panic when invoking heavy CPU usage? I seem to be having issues when I run an emerge command. Doesn't happen with the 2.6.29 kernel.

Yes I'm afraid I've had that issue too, 2.6.31.5 ran perfectly for over 48 hours uptime under quite heavy load then suddenly crashed with a null pointer dereference error, is that the same error that you got?

Was you kernel tainted with any binary drivers at the time that it crashed?

I am waiting for 2.6.31.6 to be released then I will test again, the errors I've had are similar to some bug reports I've read for vanilla (non xen) 2.6.31.5, possible fixes were posted but as the problem is hard to replicate and I am on holiday this week I've not been able to properly test them yet.

If .6 does not solve the problem I plan to test openSUSE 11.2 as I can then get help from Jan if the kernel still crashes.

2.6.29 seems to be the best so far.... I've not had a single crash nor has anybody else reported any with that version.

Andy

[Scorpion265]

Yes, that's exactly the same issue that I had. I agree, the 29 kernel is very stable, no issues with it so far.

Edit: I did not see any issues with modules. Have a good holiday! If you need any other ancillary testing of anything new, just let me know. I have a test server I use to validate any changes before pushing to my remote hosted servers.

[andylyon]

I have uploaded new patches and ebuild for 2.6.31 to http://code.google.com/p/gentoo-xen-ker ... loads/list

You should find that this version is much more stable.

Andy

[PinkFreud]

I have uploaded new patches and ebuild for 2.6.31 to http://code.google.com/p/gentoo-xen-ker ... loads/list

You should find that this version is much more stable.

Andy

Andy, these patches still break sysfs support:

Code: Select all

  Building modules, stage 2.
  MODPOST 2406 modules
ERROR: "xenbus_xsd_state" [drivers/xen/core/xen_sysfs.ko] undefined!
ERROR: "vmcoreinfo_size_xen" [drivers/xen/core/xen_sysfs.ko] undefined!
ERROR: "paddr_vmcoreinfo_xen" [drivers/xen/core/xen_sysfs.ko] undefined!
WARNING: modpost: Found 3 section mismatch(es).
To see full details build your kernel with:
'make CONFIG_DEBUG_SECTION_MISMATCH=y'
make[2]: *** [__modpost] Error 1
make[1]: *** [modules] Error 2

Unfortunately, running this kernel without sysfs support generates complaints like:

Code: Select all

# xm list
WARING!  Can't find hypervisor information in sysfs!
Name                                        ID   Mem VCPUs      State   Time(s)
Domain-0                                     0  3505     4     r-----    135.1

I'm also having difficulty booting a domU which starts fine in an older dom0 kernel - I'm not sure if this is related to the borked sysfs support or something else, though. The domU never boots, and xen eventually gives up trying to start it:

Code: Select all

# xm create debian-ldap1
WARING!  Can't find hypervisor information in sysfs!
Using config file "/etc/xen/debian-ldap1".

Error: Device 0 (vif) could not be connected. Hotplug scripts not working.

Any ideas on what I can do to convince sysfs to work? I'm hoping that a working sysfs will let me see working domUs as well.

In other news, I just noticed that 'Warning' is misspelled as 'Waring'. This is from the kernel or the Xen utils itself, and not my typo. :)

Thanks!

[PinkFreud]

Andy, these patches still break sysfs support:

Code: Select all

  Building modules, stage 2.
  MODPOST 2406 modules
ERROR: "xenbus_xsd_state" [drivers/xen/core/xen_sysfs.ko] undefined!
ERROR: "vmcoreinfo_size_xen" [drivers/xen/core/xen_sysfs.ko] undefined!
ERROR: "paddr_vmcoreinfo_xen" [drivers/xen/core/xen_sysfs.ko] undefined!
WARNING: modpost: Found 3 section mismatch(es).
To see full details build your kernel with:
'make CONFIG_DEBUG_SECTION_MISMATCH=y'
make[2]: *** [__modpost] Error 1
make[1]: *** [modules] Error 2

Resolved. Enabling sysfs works if you compile it in, rather than making it a module. Apparently, module support is broken.

I'm also having difficulty booting a domU which starts fine in an older dom0 kernel - I'm not sure if this is related to the borked sysfs support or something else, though. The domU never boots, and xen eventually gives up trying to start it:

Code: Select all

# xm create debian-ldap1
WARING!  Can't find hypervisor information in sysfs!
Using config file "/etc/xen/debian-ldap1".

Error: Device 0 (vif) could not be connected. Hotplug scripts not working.

Er, nevermind on this one. Silly me compiled everything else as modules, and then failed to make sure those modules were actually loaded. When I load the appropriate xen modules, it magically works. :)

[andylyon]

In other news, I just noticed that 'Warning' is misspelled as 'Waring'. This is from the kernel or the Xen utils itself, and not my typo.

Thanks!

What version of Xen are you running? I think the misspelling has already been corrected as I cannot find that string nor the substring "find hypervisor information in" in either the Xen source or the 2.6.31 kernel source tree.

Er, nevermind on this one. Silly me compiled everything else as modules, and then failed to make sure those modules were actually loaded. When I load the appropriate xen modules, it magically works.

Pleased you fixed the module issues, I had similar problems myself and I remember it was not immediately obvious what was wrong when the right modules were not loaded.

Andy

[PinkFreud]

In other news, I just noticed that 'Warning' is misspelled as 'Waring'. This is from the kernel or the Xen utils itself, and not my typo. :)

Thanks!

What version of Xen are you running? I think the misspelling has already been corrected as I cannot find that string nor the substring "find hypervisor information in" in either the Xen source or the 2.6.31 kernel source tree.

Version 3.2.0, under Debian. Looks like the offending spelling is in /usr/lib/xen-common/bin/xen-utils-version - amusingly enough, the typo appears multiple times throughout the script.

I don't know if that script is specific to the Debian Xen packages, or if it's a part of Xen itself. In either case, it's not an issue - it was just amusing to note.

[andylyon]

Just uploaded xen-patches-2.6.31-10.tar.bz2 and xen-sources-2.6.31-r10.ebuild to http://code.google.com/p/gentoo-xen-ker ... loads/list which has 2 small fixes for suspend which fixes xen save/restore.

Andy

[drescherjm]

BTW two xen ebuilds hit portage today. It looks like the versions are from this thread.

[dreamdreams]

I saw
kernel-xen-2.6.32.3-0.0.27.732e7c9.x86_64.rpm
kernel-xen-base-2.6.32.3-0.0.27.732e7c9.x86_64.rpm
from suse's ftp site
ftp://ftp.suse.com/pub/projects/kernel/ ... AD/x86_64/

But neither of these two can be unpacked. RPM reports bad magic. Wierd. Any idea?

[Mathijs]

Are the ebuilds in portage the same as these?

I have a networking problem, are there any modules outside the XEN area of the kernel that i might be missing?

[Mathijs]

My former problems are solved and these kernels compile fine.
Only problem i get, is when i use the xen console.
When i start a domain with -c i see everything come by until starting local [ok], then i get no login prompt.
i have the console=xvc0 in my domU config.
When i boot the domU with an old kernel, i get a login prompt so it should not be the domain.

Update: I worked around the problem by adding extra = "xencons=tty1 console=xvc0" together in the domU config.
However, i still think it is strange this is necessary, because with an old kernel it does work out of the box.
Does anyone have any idea what will be causing this?

[_Razorblade_]

I'm having problem with 2.6.31-xen-r1 with pcifront. When passing through a nic I get an oops when loading the nic module in the guest:

Code: Select all

Jan 22 14:55:59 vdr kernel: r8169 Gigabit Ethernet driver 2.3LK-NAPI loaded
Jan 22 14:55:59 vdr kernel: r8169 0000:07:00.0: enabling device (0000 -> 0003)
Jan 22 14:55:59 vdr kernel: ------------[ cut here ]------------
Jan 22 14:55:59 vdr kernel: kernel BUG at drivers/xen/core/evtchn.c:1714!
Jan 22 14:55:59 vdr kernel: invalid opcode: 0000 [#1] SMP
Jan 22 14:55:59 vdr kernel: last sysfs file: /sys/block/xvda1/removable
Jan 22 14:55:59 vdr kernel: Modules linked in: r8169(+) mii
Jan 22 14:55:59 vdr kernel:
Jan 22 14:55:59 vdr kernel: Pid: 377, comm: modprobe Not tainted (2.6.31-xen-r10 #3)
Jan 22 14:55:59 vdr kernel: EIP: 0061:[<c0307054>] EFLAGS: 00010212 CPU: 0
Jan 22 14:55:59 vdr kernel: EIP is at evtchn_register_pirq+0x54/0x60
Jan 22 14:55:59 vdr kernel: EAX: 00000011 EBX: 00000011 ECX: 00000000 EDX: c048ca40
Jan 22 14:55:59 vdr kernel: ESI: 00000045 EDI: ecd6dec8 EBP: ecd6dc00 ESP: ec68ddd4
Jan 22 14:55:59 vdr kernel: DS: 007b ES: 007b FS: 00d8 GS: 0033 SS: 0069
Jan 22 14:55:59 vdr kernel: Process modprobe (pid: 377, ti=ec68c000 task=ec407c00 task.ti=ec68c000)
Jan 22 14:55:59 vdr kernel: Stack:
Jan 22 14:55:59 vdr kernel: ecd6dec8 ecd6dc00 c0327e53 ec68dde7 11d6dec8 ecd6dc00 c0329364 ecd6dc00
Jan 22 14:55:59 vdr kernel: <0> c02d7ce2 00000045 00000300 c02d7d83 ecd6dc58 ec569000 ecd6dc00 ec569340
Jan 22 14:55:59 vdr kernel: <0> eda7c47e eda7d720 eda7db11 eda7db17 c01c6290 c1ba5204 ec4fbc80 ec82c19c
Jan 22 14:55:59 vdr kernel: Call Trace:
Jan 22 14:55:59 vdr kernel: [<c0327e53>] ? pcifront_enable_irq+0x33/0x50
Jan 22 14:55:59 vdr kernel: [<c0329364>] ? pcibios_enable_device+0x14/0x20
Jan 22 14:55:59 vdr kernel: [<c02d7ce2>] ? do_pci_enable_device+0x22/0x50
Jan 22 14:55:59 vdr kernel: [<c02d7d83>] ? __pci_enable_device_flags+0x73/0x90
Jan 22 14:55:59 vdr kernel: [<eda7c47e>] ? rtl8169_init_one+0xfe/0x9c9 [r8169]
Jan 22 14:55:59 vdr kernel: [<c01c6290>] ? sysfs_ilookup_test+0x0/0x10
Jan 22 14:55:59 vdr kernel: [<c01904dd>] ? iput+0x1d/0x50
Jan 22 14:55:59 vdr kernel: [<c01c6fb8>] ? sysfs_addrm_finish+0x38/0x280
Jan 22 14:55:59 vdr kernel: [<c01c680b>] ? sysfs_addrm_start+0x5b/0xb0
Jan 22 14:55:59 vdr kernel: [<c02d8bbb>] ? local_pci_probe+0xb/0x10
Jan 22 14:55:59 vdr kernel: [<c02d9449>] ? pci_device_probe+0x69/0x90
Jan 22 14:55:59 vdr kernel: [<c030155d>] ? driver_probe_device+0x7d/0x180
Jan 22 14:55:59 vdr kernel: [<c02d8c78>] ? pci_match_device+0x18/0x100
Jan 22 14:55:59 vdr kernel: [<c03016d9>] ? __driver_attach+0x79/0x80
Jan 22 14:55:59 vdr kernel: [<c02d9380>] ? pci_device_remove+0x0/0x40
Jan 22 14:55:59 vdr kernel: [<c0300e99>] ? bus_for_each_dev+0x49/0x70
Jan 22 14:55:59 vdr kernel: [<c02d9380>] ? pci_device_remove+0x0/0x40
Jan 22 14:55:59 vdr kernel: [<c03013f6>] ? driver_attach+0x16/0x20
Jan 22 14:55:59 vdr kernel: [<c0301660>] ? __driver_attach+0x0/0x80
Jan 22 14:55:59 vdr kernel: [<c030086f>] ? bus_add_driver+0x1af/0x250
Jan 22 14:55:59 vdr kernel: [<c02d9380>] ? pci_device_remove+0x0/0x40
Jan 22 14:55:59 vdr kernel: [<c0301957>] ? driver_register+0x67/0x150
Jan 22 14:55:59 vdr kernel: [<c02d983d>] ? __pci_register_driver+0x3d/0xb0
Jan 22 14:55:59 vdr kernel: [<eda82000>] ? rtl8169_init_module+0x0/0x14 [r8169]
Jan 22 14:55:59 vdr kernel: [<c010203e>] ? do_one_initcall+0x2e/0x1b0
Jan 22 14:55:59 vdr kernel: [<c014d152>] ? sys_init_module+0xb2/0x1f0
Jan 22 14:55:59 vdr kernel: [<c017dc21>] ? sys_read+0x41/0x80
Jan 22 14:55:59 vdr kernel: [<c010446d>] ? syscall_call+0x7/0xb
Jan 22 14:55:59 vdr kernel: Code: 04 24 ff 0b 44 c0 c1 e0 0c b9 50 0d 15 c0 0d 00 00 00 10 ba 00 8d 4a c0 89 04 dd e0 63 60 c0 89 d8 e8 a1 9f e4 ff 83 c4 04 5b c3 <0f> 0b eb fe 90 8d b4 26 00 00 00 00 56 89 d6 53 bb 10 00 00 00
Jan 22 14:55:59 vdr kernel: EIP: [<c0307054>] evtchn_register_pirq+0x54/0x60 SS:ESP 0069:ec68ddd4
Jan 22 14:55:59 vdr kernel: ---[ end trace 7a1dea61bdbe354f ]---

This was working fine with 2.6.29-r4 (which turns out to be the most stable xen kernel I ever ran).
domU is gentoo x86-SMP, dom0 and xen are x86_64-SMP

Anyone else seen this?


I was going to give pvops another shot after that, but that wouldn't even compile with pcifront enabled...

[2bbionic]

I have a problem, too with the actual 2.6.31-xen-r10. I tried some different kernel configs, but eacht time, i try to boot this kernel, i get the following oops:

Warning: at kernel/smp.c:369 smp_call_function_many
Hardware name:
Modules linkes in:
Pid: 0, comm:swapper Not tainted 2.6.31-xen-r10 #3

Call trace
....

Console colour VGA 80x25
console [tty0] enabled

I wrote this down from a picture i shot from monitor, because i don't know how to save this message...
If you need more info about the kernel trace, please let me know.
The system is a x86 with a E8400 C2D an 4 GB RAM
I anybody outthere could help me

Thanks,

2bbionic

[_Razorblade_]

Was this with a UP or SMP kernel?
I also noticed even more problems then the module failing when I run UP instead of SMP...

[2bbionic]

It's a SMP kernel:

Code: Select all

# CONFIG_64BIT is not set
CONFIG_X86_32=y
# CONFIG_X86_64 is not set
CONFIG_X86=y
CONFIG_OUTPUT_FORMAT="elf32-i386"
CONFIG_ARCH_DEFCONFIG="arch/x86/configs/i386_defconfig"
CONFIG_GENERIC_TIME=y
CONFIG_GENERIC_CMOS_UPDATE=y
CONFIG_LOCKDEP_SUPPORT=y
CONFIG_STACKTRACE_SUPPORT=y
CONFIG_HAVE_LATENCYTOP_SUPPORT=y
CONFIG_FAST_CMPXCHG_LOCAL=y
CONFIG_MMU=y
CONFIG_ZONE_DMA=y
CONFIG_GENERIC_ISA_DMA=y
CONFIG_GENERIC_IOMAP=y
CONFIG_GENERIC_BUG=y
CONFIG_GENERIC_HWEIGHT=y
CONFIG_ARCH_MAY_HAVE_PC_FDC=y
# CONFIG_RWSEM_GENERIC_SPINLOCK is not set
CONFIG_RWSEM_XCHGADD_ALGORITHM=y
CONFIG_ARCH_HAS_CPU_IDLE_WAIT=y
CONFIG_GENERIC_CALIBRATE_DELAY=y
# CONFIG_GENERIC_TIME_VSYSCALL is not set
CONFIG_ARCH_HAS_CPU_RELAX=y
CONFIG_ARCH_HAS_DEFAULT_IDLE=y
CONFIG_ARCH_HAS_CACHE_LINE_SIZE=y
CONFIG_HAVE_SETUP_PER_CPU_AREA=y
CONFIG_HAVE_DYNAMIC_PER_CPU_AREA=y
# CONFIG_HAVE_CPUMASK_OF_CPU_MAP is not set
CONFIG_ARCH_SUSPEND_POSSIBLE=y
# CONFIG_ZONE_DMA32 is not set
CONFIG_ARCH_POPULATES_NODE_MAP=y
# CONFIG_AUDIT_ARCH is not set
CONFIG_ARCH_SUPPORTS_OPTIMIZED_INLINING=y
CONFIG_ARCH_SUPPORTS_DEBUG_PAGEALLOC=y
CONFIG_GENERIC_HARDIRQS=y
CONFIG_GENERIC_HARDIRQS_NO__DO_IRQ=y
CONFIG_GENERIC_IRQ_PROBE=y
CONFIG_GENERIC_PENDING_IRQ=y
CONFIG_USE_GENERIC_SMP_HELPERS=y
CONFIG_X86_32_SMP=y
CONFIG_X86_NO_TSS=y
CONFIG_X86_NO_IDT=y
CONFIG_X86_32_LAZY_GS=y
CONFIG_KTIME_SCALAR=y
CONFIG_DEFCONFIG_LIST="/lib/modules/$UNAME_RELEASE/.config"
CONFIG_CONSTRUCTORS=y

Code: Select all

#
# RCU Subsystem
#
CONFIG_CLASSIC_RCU=y
# CONFIG_TREE_RCU is not set
# CONFIG_PREEMPT_RCU is not set
# CONFIG_TREE_RCU_TRACE is not set
# CONFIG_PREEMPT_RCU_TRACE is not set
CONFIG_IKCONFIG=y
CONFIG_IKCONFIG_PROC=y
CONFIG_LOG_BUF_SHIFT=16
CONFIG_HAVE_UNSTABLE_SCHED_CLOCK=y
CONFIG_GROUP_SCHED=y
CONFIG_FAIR_GROUP_SCHED=y
# CONFIG_RT_GROUP_SCHED is not set
CONFIG_USER_SCHED=y
# CONFIG_CGROUP_SCHED is not set
# CONFIG_CGROUPS is not set
# CONFIG_SYSFS_DEPRECATED_V2 is not set
# CONFIG_RELAY is not set
CONFIG_NAMESPACES=y
CONFIG_UTS_NS=y
CONFIG_IPC_NS=y
# CONFIG_USER_NS is not set
# CONFIG_PID_NS is not set
# CONFIG_NET_NS is not set
# CONFIG_BLK_DEV_INITRD is not set
# CONFIG_CC_OPTIMIZE_FOR_SIZE is not set
CONFIG_SYSCTL=y
CONFIG_ANON_INODES=y
# CONFIG_EMBEDDED is not set
CONFIG_UID16=y
CONFIG_SYSCTL_SYSCALL=y
CONFIG_KALLSYMS=y
# CONFIG_KALLSYMS_EXTRA_PASS is not set
CONFIG_HOTPLUG=y
CONFIG_PRINTK=y
CONFIG_BUG=y
CONFIG_ELF_CORE=y
CONFIG_PCSPKR_PLATFORM=y
CONFIG_BASE_FULL=y
CONFIG_FUTEX=y
CONFIG_EPOLL=y
CONFIG_SIGNALFD=y
CONFIG_TIMERFD=y
CONFIG_EVENTFD=y
CONFIG_SHMEM=y
CONFIG_AIO=y

Code: Select all

#
# Processor type and features
#
CONFIG_SMP=y
# CONFIG_SPARSE_IRQ is not set
# CONFIG_X86_MPPARSE is not set
CONFIG_X86_XEN=y
CONFIG_SCHED_OMIT_FRAME_POINTER=y
# CONFIG_M386 is not set
# CONFIG_M486 is not set
# CONFIG_M586 is not set
# CONFIG_M586TSC is not set
# CONFIG_M586MMX is not set
# CONFIG_M686 is not set
# CONFIG_MPENTIUMII is not set
# CONFIG_MPENTIUMIII is not set
# CONFIG_MPENTIUMM is not set
# CONFIG_MPENTIUM4 is not set
# CONFIG_MK6 is not set
# CONFIG_MK7 is not set
# CONFIG_MK8 is not set
# CONFIG_MCRUSOE is not set
# CONFIG_MEFFICEON is not set
# CONFIG_MWINCHIPC6 is not set
# CONFIG_MWINCHIP3D is not set
# CONFIG_MGEODEGX1 is not set
# CONFIG_MGEODE_LX is not set
# CONFIG_MCYRIXIII is not set
# CONFIG_MVIAC3_2 is not set
# CONFIG_MVIAC7 is not set
# CONFIG_MPSC is not set
CONFIG_MCORE2=y
# CONFIG_GENERIC_CPU is not set
CONFIG_X86_GENERIC=y
CONFIG_X86_CPU=y
CONFIG_X86_L1_CACHE_BYTES=64
CONFIG_X86_INTERNODE_CACHE_BYTES=64
CONFIG_X86_CMPXCHG=y
CONFIG_X86_L1_CACHE_SHIFT=6
CONFIG_X86_XADD=y
CONFIG_X86_WP_WORKS_OK=y
CONFIG_X86_INVLPG=y
CONFIG_X86_BSWAP=y
CONFIG_X86_POPAD_OK=y
CONFIG_X86_INTEL_USERCOPY=y
CONFIG_X86_USE_PPRO_CHECKSUM=y
CONFIG_X86_CMPXCHG64=y
CONFIG_X86_CMOV=y
CONFIG_X86_MINIMUM_CPU_FAMILY=4
CONFIG_X86_DEBUGCTLMSR=y
CONFIG_CPU_SUP_INTEL=y
CONFIG_CPU_SUP_CYRIX_32=y
CONFIG_CPU_SUP_AMD=y
CONFIG_CPU_SUP_CENTAUR=y
CONFIG_CPU_SUP_TRANSMETA_32=y
CONFIG_DMI=y
CONFIG_SWIOTLB=y
CONFIG_IOMMU_HELPER=y
# CONFIG_IOMMU_API is not set
CONFIG_NR_CPUS=4
CONFIG_PREEMPT_NONE=y
# CONFIG_PREEMPT_VOLUNTARY is not set
# CONFIG_PREEMPT is not set
CONFIG_X86_LOCAL_APIC=y
CONFIG_X86_IO_APIC=y
# CONFIG_X86_MCE is not set
CONFIG_VM86=y
# CONFIG_TOSHIBA is not set
# CONFIG_I8K is not set
CONFIG_MICROCODE=m
CONFIG_MICROCODE_OLD_INTERFACE=y
CONFIG_X86_MSR=y
CONFIG_X86_CPUID=y
# CONFIG_X86_CPU_DEBUG is not set
# CONFIG_NOHIGHMEM is not set
# CONFIG_HIGHMEM4G is not set
CONFIG_HIGHMEM64G=y
CONFIG_PAGE_OFFSET=0xC0000000
CONFIG_HIGHMEM=y
CONFIG_X86_PAE=y
CONFIG_ARCH_PHYS_ADDR_T_64BIT=y
CONFIG_SELECT_MEMORY_MODEL=y
CONFIG_FLATMEM_MANUAL=y
# CONFIG_DISCONTIGMEM_MANUAL is not set
# CONFIG_SPARSEMEM_MANUAL is not set
CONFIG_FLATMEM=y
CONFIG_FLAT_NODE_MEM_MAP=y
CONFIG_PAGEFLAGS_EXTENDED=y
CONFIG_SPLIT_PTLOCK_CPUS=4
CONFIG_PHYS_ADDR_T_64BIT=y
CONFIG_ZONE_DMA_FLAG=1
CONFIG_BOUNCE=y
CONFIG_VIRT_TO_BUS=y
CONFIG_HAVE_MLOCK=y
CONFIG_HAVE_MLOCKED_PAGE_BIT=y
CONFIG_DEFAULT_MMAP_MIN_ADDR=4096
# CONFIG_HIGHPTE is not set
CONFIG_MTRR=y
CONFIG_X86_PAT=y
CONFIG_SECCOMP=y
# CONFIG_CC_STACKPROTECTOR is not set
# CONFIG_HZ_100 is not set
CONFIG_HZ_250=y
# CONFIG_HZ_300 is not set

# CONFIG_HZ_1000 is not set
CONFIG_HZ=250
# CONFIG_SCHED_HRTICK is not set
# CONFIG_KEXEC is not set
CONFIG_PHYSICAL_START=0x100000
CONFIG_PHYSICAL_ALIGN=0x2000
CONFIG_HOTPLUG_CPU=y
# CONFIG_COMPAT_VDSO is not set
# CONFIG_CMDLINE_BOOL is not set

[Mathijs]

I also just found out that a domU with these kernels cannot be migrated.

I also found the reason why:

2.6.32.2 has patches for save/restore, which also affects live
migration. So 2.6.32.2 should work, while older 2.6.31./2.6.32.x kernels
will fail.

Update: Confirmed, with vanilla 2.6.32.3 live migration works, only i have no console again.
But that is something for tomorrow.

This is good to read.

Adding "h0respawn:/sbin/agetty -L 9600 hvc0 linux" to the domU's /etc/inittab made it work with no extra line needed in the domU config file.
Of course hvc0 needs to be added to /etc/securetty too.

[_Razorblade_]

You're mixing pvops with a xen-rebased kernel.
For an overview have a look here http://wiki.xensource.com/xenwiki/XenDom0Kernels