Hi, I''m getting kernel panic during boot latest xen/next. kernel BUG at drivers/xen/balloon.c:528! invalid opcode: 0000 [#1] SMP last sysfs file: CPU 1 Modules linked in: Pid: 1, comm: swapper Not tainted 2.6.32.16-oe64-00000-g9787330-dirty #18 HVM domU RIP: 0010:[<ffffffff812ffd5c>] [<ffffffff812ffd5c>] dealloc_pte_fn+0x9c/0xa0 RSP: 0018:ffff88001f4a3d90 EFLAGS: 00010296 RAX: 00000000ffffffff RBX: ffff88001e000ff8 RCX: 0000000000000000 RDX: ffff88001e000f01 RSI: ffff88001f4a3d90 RDI: 0000000000000001 RBP: ffff88001e001000 R08: 0000000000000000 R09: 0000000000000000 R10: 000000000001c888 R11: 0000000000000001 R12: ffff88001e1ff000 R13: ffff880002465000 R14: ffff88001e200000 R15: ffff880001006780 FS: 0000000000000000(0000) GS:ffff880001a20000(0000) knlGS:0000000000000000 CS: 0010 DS: 0018 ES: 0018 CR0: 000000008005003b CR2: 0000000000000000 CR3: 0000000001001000 CR4: 00000000000006a0 DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400 Process swapper (pid: 1, threadinfo ffff88001f4a2000, task ffff88001f4a17d0) Stack: ffff88001f4a3db8 0000000000000001 0000000000000000 0000000000007ff0 <0> 0000000000000002 0000000217461006 ffff88001e000ff8 ffff88001e001000 <0> ffff88001e1ff000 ffffffff810a68cf 000201d000000010 0000000000000000 Call Trace: [<ffffffff810a68cf>] ? apply_to_page_range+0x2bf/0x410 [<ffffffff812ffcc0>] ? dealloc_pte_fn+0x0/0xa0 [<ffffffff812ffbd8>] ? alloc_empty_pages_and_pagevec+0xf8/0x1e0 [<ffffffff818a3ce7>] ? netback_init+0x137/0x420 [<ffffffff818a3bb0>] ? netback_init+0x0/0x420 [<ffffffff8100a037>] ? do_one_initcall+0x27/0x180 [<ffffffff81879a50>] ? kernel_init+0x160/0x1c0 [<ffffffff81879140>] ? early_idt_handler+0x0/0x71 [<ffffffff81013d8a>] ? child_rip+0xa/0x20 [<ffffffff81879140>] ? early_idt_handler+0x0/0x71 [<ffffffff818798f0>] ? kernel_init+0x0/0x1c0 [<ffffffff81013d80>] ? child_rip+0x0/0x20 Code: c7 c6 ff ff ff ff 48 c1 ef 0c e8 e0 d8 d0 ff bf 01 00 00 00 48 89 e6 e8 33 94 d0 ff ff c8 75 0b 48 83 c4 30 31 c0 5b 5d 41 5c c3 <0f> 0b eb fe 53 48 8b 9f a0 00 00 00 48 85 f6 48 89 f1 0f 95 c2 RIP [<ffffffff812ffd5c>] dealloc_pte_fn+0x9c/0xa0 RSP <ffff88001f4a3d90> ---[ end trace a7919e7f17c0a725 ]--- Kernel panic - not syncing: Attempted to kill init! Pid: 1, comm: swapper Tainted: G D 2.6.32.16-oe64-00000-g9787330-dirty #18 Logs and config are in attachments -- Łukasz Oleś _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Jeremy Fitzhardinge
2010-Jul-15 17:34 UTC
Re: [Xen-devel] kernel BUG at drivers/xen/balloon.c:528!
On 07/15/2010 04:07 AM, Łukasz Oleś wrote:> Hi, > > I''m getting kernel panic during boot latest xen/next. >Booting PV or HVM? J> kernel BUG at drivers/xen/balloon.c:528! > invalid opcode: 0000 [#1] SMP > last sysfs file: > CPU 1 > Modules linked in: > Pid: 1, comm: swapper Not tainted 2.6.32.16-oe64-00000-g9787330-dirty > #18 HVM domU > RIP: 0010:[<ffffffff812ffd5c>] [<ffffffff812ffd5c>] dealloc_pte_fn+0x9c/0xa0 > RSP: 0018:ffff88001f4a3d90 EFLAGS: 00010296 > RAX: 00000000ffffffff RBX: ffff88001e000ff8 RCX: 0000000000000000 > RDX: ffff88001e000f01 RSI: ffff88001f4a3d90 RDI: 0000000000000001 > RBP: ffff88001e001000 R08: 0000000000000000 R09: 0000000000000000 > R10: 000000000001c888 R11: 0000000000000001 R12: ffff88001e1ff000 > R13: ffff880002465000 R14: ffff88001e200000 R15: ffff880001006780 > FS: 0000000000000000(0000) GS:ffff880001a20000(0000) knlGS:0000000000000000 > CS: 0010 DS: 0018 ES: 0018 CR0: 000000008005003b > CR2: 0000000000000000 CR3: 0000000001001000 CR4: 00000000000006a0 > DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 > DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400 > Process swapper (pid: 1, threadinfo ffff88001f4a2000, task ffff88001f4a17d0) > Stack: > ffff88001f4a3db8 0000000000000001 0000000000000000 0000000000007ff0 > <0> 0000000000000002 0000000217461006 ffff88001e000ff8 ffff88001e001000 > <0> ffff88001e1ff000 ffffffff810a68cf 000201d000000010 0000000000000000 > Call Trace: > [<ffffffff810a68cf>] ? apply_to_page_range+0x2bf/0x410 > [<ffffffff812ffcc0>] ? dealloc_pte_fn+0x0/0xa0 > [<ffffffff812ffbd8>] ? alloc_empty_pages_and_pagevec+0xf8/0x1e0 > [<ffffffff818a3ce7>] ? netback_init+0x137/0x420 > [<ffffffff818a3bb0>] ? netback_init+0x0/0x420 > [<ffffffff8100a037>] ? do_one_initcall+0x27/0x180 > [<ffffffff81879a50>] ? kernel_init+0x160/0x1c0 > [<ffffffff81879140>] ? early_idt_handler+0x0/0x71 > [<ffffffff81013d8a>] ? child_rip+0xa/0x20 > [<ffffffff81879140>] ? early_idt_handler+0x0/0x71 > [<ffffffff818798f0>] ? kernel_init+0x0/0x1c0 > [<ffffffff81013d80>] ? child_rip+0x0/0x20 > Code: c7 c6 ff ff ff ff 48 c1 ef 0c e8 e0 d8 d0 ff bf 01 00 00 00 48 > 89 e6 e8 33 94 d0 ff ff c8 75 0b 48 83 c4 30 31 c0 5b 5d 41 5c c3 <0f> > 0b eb fe 53 48 8b 9f a0 00 00 00 48 85 f6 48 89 f1 0f 95 c2 > RIP [<ffffffff812ffd5c>] dealloc_pte_fn+0x9c/0xa0 > RSP <ffff88001f4a3d90> > ---[ end trace a7919e7f17c0a725 ]--- > Kernel panic - not syncing: Attempted to kill init! > Pid: 1, comm: swapper Tainted: G D > 2.6.32.16-oe64-00000-g9787330-dirty #18 > > Logs and config are in attachments > > > > > _______________________________________________ > Xen-devel mailing list > Xen-devel@lists.xensource.com > http://lists.xensource.com/xen-devel >_______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Stefano Stabellini
2010-Jul-15 17:41 UTC
Re: [Xen-devel] kernel BUG at drivers/xen/balloon.c:528!
On Thu, 15 Jul 2010, Jeremy Fitzhardinge wrote:> On 07/15/2010 04:07 AM, Łukasz Oleś wrote: > > Hi, > > > > I''m getting kernel panic during boot latest xen/next. > > > > Booting PV or HVM? >This bug is caused by backend drivers trying to initialize in a PV on HVM guest. The last port of my series on xen/next should have the right fix. _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Łukasz Oleś
2010-Jul-16 10:53 UTC
Re: [Xen-devel] kernel BUG at drivers/xen/balloon.c:528!
W dniu 15 lipca 2010 19:41 użytkownik Stefano Stabellini <stefano.stabellini@eu.citrix.com> napisał:> On Thu, 15 Jul 2010, Jeremy Fitzhardinge wrote: >> On 07/15/2010 04:07 AM, Łukasz Oleś wrote: >> > Hi, >> > >> > I''m getting kernel panic during boot latest xen/next. >> > >> >> Booting PV or HVM? >> > > This bug is caused by backend drivers trying to initialize in a PV on > HVM guest. > The last port of my series on xen/next should have the right fix. >Now I have problem with vbd and vifs. vbd vbd-51712: 28 granting access to ring page vbd vbd-51712: 28 xenbus_dev_probe on device/vbd/51712 vbd: probe of vbd-51712 failed with error -28 vbd vbd-51728: 28 granting access to ring page vbd vbd-51728: 28 xenbus_dev_probe on device/vbd/51728 vbd: probe of vbd-51728 failed with error -28 vbd vbd-51744: 28 granting access to ring page vbd vbd-51744: 28 xenbus_dev_probe on device/vbd/51744 vbd: probe of vbd-51744 failed with error -28 Initialising Xen virtual ethernet driver. #### netfront can''t alloc tx grant refs vif vif-0: 12 creating netdev vif vif-0: 12 xenbus_dev_probe on device/vif/0 vif: probe of vif-0 failed with error -12 #### netfront can''t alloc tx grant refs vif vif-1: 12 creating netdev vif vif-1: 12 xenbus_dev_probe on device/vif/1 vif: probe of vif-1 failed with error -12 #### netfront can''t alloc tx grant refs vif vif-2: 12 creating netdev vif vif-2: 12 xenbus_dev_probe on device/vif/2 vif: probe of vif-2 failed with error -12 Full log in attachment, vm config is posted in earlier message. -- Łukasz Oleś _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Stefano Stabellini
2010-Jul-16 16:39 UTC
Re: [Xen-devel] kernel BUG at drivers/xen/balloon.c:528!
On Fri, 16 Jul 2010, Łukasz Oleś wrote:> W dniu 15 lipca 2010 19:41 użytkownik Stefano Stabellini > <stefano.stabellini@eu.citrix.com> napisał: > > On Thu, 15 Jul 2010, Jeremy Fitzhardinge wrote: > >> On 07/15/2010 04:07 AM, Łukasz Oleś wrote: > >> > Hi, > >> > > >> > I''m getting kernel panic during boot latest xen/next. > >> > > >> > >> Booting PV or HVM? > >> > > > > This bug is caused by backend drivers trying to initialize in a PV on > > HVM guest. > > The last port of my series on xen/next should have the right fix. > > > > Now I have problem with vbd and vifs. > > vbd vbd-51712: 28 granting access to ring page > vbd vbd-51712: 28 xenbus_dev_probe on device/vbd/51712 > vbd: probe of vbd-51712 failed with error -28 > vbd vbd-51728: 28 granting access to ring page > vbd vbd-51728: 28 xenbus_dev_probe on device/vbd/51728 > vbd: probe of vbd-51728 failed with error -28 > vbd vbd-51744: 28 granting access to ring page > vbd vbd-51744: 28 xenbus_dev_probe on device/vbd/51744 > vbd: probe of vbd-51744 failed with error -28 > > Initialising Xen virtual ethernet driver. > #### netfront can''t alloc tx grant refs > vif vif-0: 12 creating netdev > vif vif-0: 12 xenbus_dev_probe on device/vif/0 > vif: probe of vif-0 failed with error -12 > #### netfront can''t alloc tx grant refs > vif vif-1: 12 creating netdev > vif vif-1: 12 xenbus_dev_probe on device/vif/1 > vif: probe of vif-1 failed with error -12 > #### netfront can''t alloc tx grant refs > vif vif-2: 12 creating netdev > vif vif-2: 12 xenbus_dev_probe on device/vif/2 > vif: probe of vif-2 failed with error -12 > > Full log in attachment, vm config is posted in earlier message. >The warnings are due to the fact that CONFIG_XEN_PLATFORM_PCI is not enabled in your config but the kernel tries to initialize the pv frontends anyway, the patch I have just sent to the list should fix this issue. In any case you should enable CONFIG_XEN_PLATFORM_PCI in your kernel config, otherwise the pv frontends won''t load properly and you are not going to have any benefits running a PV on HVM kernel. I am not sure why you cannot mount your root device, maybe you didn''t specify the correct root= in your kernel command line options? _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Łukasz Oleś
2010-Jul-19 13:12 UTC
Re: [Xen-devel] kernel BUG at drivers/xen/balloon.c:528!
W dniu 16 lipca 2010 18:39 użytkownik Stefano Stabellini <stefano.stabellini@eu.citrix.com> napisał:> On Fri, 16 Jul 2010, Łukasz Oleś wrote: >> W dniu 15 lipca 2010 19:41 użytkownik Stefano Stabellini >> <stefano.stabellini@eu.citrix.com> napisał: >> > On Thu, 15 Jul 2010, Jeremy Fitzhardinge wrote: >> >> On 07/15/2010 04:07 AM, Łukasz Oleś wrote: >> >> > Hi, >> >> > >> >> > I''m getting kernel panic during boot latest xen/next. >> >> > >> >> >> >> Booting PV or HVM? >> >> >> > >> > This bug is caused by backend drivers trying to initialize in a PV on >> > HVM guest. >> > The last port of my series on xen/next should have the right fix. >> > >> >> Now I have problem with vbd and vifs. >> >> vbd vbd-51712: 28 granting access to ring page >> vbd vbd-51712: 28 xenbus_dev_probe on device/vbd/51712 >> vbd: probe of vbd-51712 failed with error -28 >> vbd vbd-51728: 28 granting access to ring page >> vbd vbd-51728: 28 xenbus_dev_probe on device/vbd/51728 >> vbd: probe of vbd-51728 failed with error -28 >> vbd vbd-51744: 28 granting access to ring page >> vbd vbd-51744: 28 xenbus_dev_probe on device/vbd/51744 >> vbd: probe of vbd-51744 failed with error -28 >> >> Initialising Xen virtual ethernet driver. >> #### netfront can''t alloc tx grant refs >> vif vif-0: 12 creating netdev >> vif vif-0: 12 xenbus_dev_probe on device/vif/0 >> vif: probe of vif-0 failed with error -12 >> #### netfront can''t alloc tx grant refs >> vif vif-1: 12 creating netdev >> vif vif-1: 12 xenbus_dev_probe on device/vif/1 >> vif: probe of vif-1 failed with error -12 >> #### netfront can''t alloc tx grant refs >> vif vif-2: 12 creating netdev >> vif vif-2: 12 xenbus_dev_probe on device/vif/2 >> vif: probe of vif-2 failed with error -12 >> >> Full log in attachment, vm config is posted in earlier message. >> > > The warnings are due to the fact that CONFIG_XEN_PLATFORM_PCI is not > enabled in your config but the kernel tries to initialize the pv > frontends anyway, the patch I have just sent to the list should fix this > issue.Yes, their gone now> > In any case you should enable CONFIG_XEN_PLATFORM_PCI in your kernel > config, otherwise the pv frontends won''t load properly and you are not > going to have any benefits running a PV on HVM kernel. > > I am not sure why you cannot mount your root device, maybe you didn''t > specify the correct root= in your kernel command line options?It works fine on bare metal and ESX. It looks like in initrd system doesn''t see any disks. When I enabled option CONFIG_XEN_PLATFORM_PCI I was finally able to boot my sytem. Is this suposed to work in this way? -- Łukasz Oleś _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Stefano Stabellini
2010-Jul-19 14:19 UTC
Re: [Xen-devel] kernel BUG at drivers/xen/balloon.c:528!
On Mon, 19 Jul 2010, Łukasz Oleś wrote:> > > > In any case you should enable CONFIG_XEN_PLATFORM_PCI in your kernel > > config, otherwise the pv frontends won''t load properly and you are not > > going to have any benefits running a PV on HVM kernel. > > > > I am not sure why you cannot mount your root device, maybe you didn''t > > specify the correct root= in your kernel command line options? > > It works fine on bare metal and ESX. It looks like in initrd system > doesn''t see any disks. > > When I enabled option CONFIG_XEN_PLATFORM_PCI I was finally able to > boot my sytem. Is this suposed to work in this way? >You are able to mount /dev/xvd* if you have CONFIG_XEN_PLATFORM_PCI (because it is required to initialize blkfront), but you should be able to mount /dev/hd* just fine without CONFIG_XEN_PLATFORM_PCI. However if you have CONFIG_XEN_PLATFORM_PCI in your config, then the kernel will unplug the emulated disks at boot so you''ll only be able to mount /dev/xvd*. _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Łukasz Oleś
2010-Jul-19 14:54 UTC
Re: [Xen-devel] kernel BUG at drivers/xen/balloon.c:528!
W dniu 19 lipca 2010 16:19 użytkownik Stefano Stabellini <stefano.stabellini@eu.citrix.com> napisał:> On Mon, 19 Jul 2010, Łukasz Oleś wrote: >> > >> > In any case you should enable CONFIG_XEN_PLATFORM_PCI in your kernel >> > config, otherwise the pv frontends won''t load properly and you are not >> > going to have any benefits running a PV on HVM kernel. >> > >> > I am not sure why you cannot mount your root device, maybe you didn''t >> > specify the correct root= in your kernel command line options? >> >> It works fine on bare metal and ESX. It looks like in initrd system >> doesn''t see any disks. >> >> When I enabled option CONFIG_XEN_PLATFORM_PCI I was finally able to >> boot my sytem. Is this suposed to work in this way? >> > but you should be able > to mount /dev/hd* just fine without CONFIG_XEN_PLATFORM_PCI.This one is not true on my system. If something in my initrd fails it launches busybox then I can run fdisk -l or cat /proc/partitions and they return nothing. Whats more intresting at the system start I can see: Probing IDE interface ide0... hda: QEMU HARDDISK, ATA DISK drive hdb: QEMU HARDDISK, ATA DISK drive hda: host max PIO4 wanted PIO255(auto-tune) selected PIO2 hda: MWDMA2 mode selected hdb: host max PIO4 wanted PIO255(auto-tune) selected PIO2 hdb: MWDMA2 mode selected Probing IDE interface ide1... hdc: QEMU HARDDISK, ATA DISK drive hdc: host max PIO4 wanted PIO255(auto-tune) selected PIO2 hdc: MWDMA2 mode selected ide0 at 0x1f0-0x1f7,0x3f6 on irq 14 ide1 at 0x170-0x177,0x376 on irq 15 ide-cd driver 5.00 but later disks disappear. Full log in earlier mail -- Łukasz Oleś _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Stefano Stabellini
2010-Jul-19 16:39 UTC
Re: [Xen-devel] kernel BUG at drivers/xen/balloon.c:528!
On Mon, 19 Jul 2010, Łukasz Oleś wrote:> W dniu 19 lipca 2010 16:19 użytkownik Stefano Stabellini > <stefano.stabellini@eu.citrix.com> napisał: > > On Mon, 19 Jul 2010, Łukasz Oleś wrote: > >> > > >> > In any case you should enable CONFIG_XEN_PLATFORM_PCI in your kernel > >> > config, otherwise the pv frontends won''t load properly and you are not > >> > going to have any benefits running a PV on HVM kernel. > >> > > >> > I am not sure why you cannot mount your root device, maybe you didn''t > >> > specify the correct root= in your kernel command line options? > >> > >> It works fine on bare metal and ESX. It looks like in initrd system > >> doesn''t see any disks. > >> > >> When I enabled option CONFIG_XEN_PLATFORM_PCI I was finally able to > >> boot my sytem. Is this suposed to work in this way? > >> > > but you should be able > > to mount /dev/hd* just fine without CONFIG_XEN_PLATFORM_PCI. > > This one is not true on my system. > > If something in my initrd fails it launches busybox then I can run > fdisk -l or cat /proc/partitions and they return nothing. Whats more > intresting at the system start I can see: > > Probing IDE interface ide0... > hda: QEMU HARDDISK, ATA DISK drive > hdb: QEMU HARDDISK, ATA DISK drive > hda: host max PIO4 wanted PIO255(auto-tune) selected PIO2 > hda: MWDMA2 mode selected > hdb: host max PIO4 wanted PIO255(auto-tune) selected PIO2 > hdb: MWDMA2 mode selected > Probing IDE interface ide1... > hdc: QEMU HARDDISK, ATA DISK drive > hdc: host max PIO4 wanted PIO255(auto-tune) selected PIO2 > hdc: MWDMA2 mode selected > ide0 at 0x1f0-0x1f7,0x3f6 on irq 14 > ide1 at 0x170-0x177,0x376 on irq 15 > ide-cd driver 5.00 > > but later disks disappear. Full log in earlier mail >Yes, I have seen that. I am not sure what is the cause of the problem because if you see those messages it means no unplug has been done so you should be able to access the emulated disks without any issues. Just to be sure, could you please add this patch to your kernel and tell me if it gets printed? diff --git a/arch/x86/xen/platform-pci-unplug.c b/arch/x86/xen/platform-pci-unplug.c index 2f7f3fb..4513202 100644 --- a/arch/x86/xen/platform-pci-unplug.c +++ b/arch/x86/xen/platform-pci-unplug.c @@ -98,8 +98,10 @@ void __init xen_unplug_emulated_devices(void) } } /* Now unplug the emulated devices */ - if (!(xen_emul_unplug & XEN_UNPLUG_IGNORE)) + if (!(xen_emul_unplug & XEN_UNPLUG_IGNORE)) { outw(xen_emul_unplug, XEN_IOPORT_UNPLUG); + printk("DEBUG unplug %x\n", xen_emul_unplug); + } xen_platform_pci_unplug = xen_emul_unplug; } _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Łukasz Oleś
2010-Jul-20 08:02 UTC
Re: [Xen-devel] kernel BUG at drivers/xen/balloon.c:528!
W dniu 19 lipca 2010 18:39 użytkownik Stefano Stabellini <stefano.stabellini@eu.citrix.com> napisał:> On Mon, 19 Jul 2010, Łukasz Oleś wrote: >> W dniu 19 lipca 2010 16:19 użytkownik Stefano Stabellini >> <stefano.stabellini@eu.citrix.com> napisał: >> > On Mon, 19 Jul 2010, Łukasz Oleś wrote: >> >> > >> >> > In any case you should enable CONFIG_XEN_PLATFORM_PCI in your kernel >> >> > config, otherwise the pv frontends won''t load properly and you are not >> >> > going to have any benefits running a PV on HVM kernel. >> >> > >> >> > I am not sure why you cannot mount your root device, maybe you didn''t >> >> > specify the correct root= in your kernel command line options? >> >> >> >> It works fine on bare metal and ESX. It looks like in initrd system >> >> doesn''t see any disks. >> >> >> >> When I enabled option CONFIG_XEN_PLATFORM_PCI I was finally able to >> >> boot my sytem. Is this suposed to work in this way? >> >> >> > but you should be able >> > to mount /dev/hd* just fine without CONFIG_XEN_PLATFORM_PCI. >> >> This one is not true on my system. >> >> If something in my initrd fails it launches busybox then I can run >> fdisk -l or cat /proc/partitions and they return nothing. Whats more >> intresting at the system start I can see: >> >> Probing IDE interface ide0... >> hda: QEMU HARDDISK, ATA DISK drive >> hdb: QEMU HARDDISK, ATA DISK drive >> hda: host max PIO4 wanted PIO255(auto-tune) selected PIO2 >> hda: MWDMA2 mode selected >> hdb: host max PIO4 wanted PIO255(auto-tune) selected PIO2 >> hdb: MWDMA2 mode selected >> Probing IDE interface ide1... >> hdc: QEMU HARDDISK, ATA DISK drive >> hdc: host max PIO4 wanted PIO255(auto-tune) selected PIO2 >> hdc: MWDMA2 mode selected >> ide0 at 0x1f0-0x1f7,0x3f6 on irq 14 >> ide1 at 0x170-0x177,0x376 on irq 15 >> ide-cd driver 5.00 >> >> but later disks disappear. Full log in earlier mail >> > > Yes, I have seen that. > I am not sure what is the cause of the problem because if you see > those messages it means no unplug has been done so you should be able to > access the emulated disks without any issues. > Just to be sure, could you please add this patch to your kernel and tell > me if it gets printed? > > > > diff --git a/arch/x86/xen/platform-pci-unplug.c b/arch/x86/xen/platform-pci-unplug.c > index 2f7f3fb..4513202 100644 > --- a/arch/x86/xen/platform-pci-unplug.c > +++ b/arch/x86/xen/platform-pci-unplug.c > @@ -98,8 +98,10 @@ void __init xen_unplug_emulated_devices(void) > } > } > /* Now unplug the emulated devices */ > - if (!(xen_emul_unplug & XEN_UNPLUG_IGNORE)) > + if (!(xen_emul_unplug & XEN_UNPLUG_IGNORE)) { > outw(xen_emul_unplug, XEN_IOPORT_UNPLUG); > + printk("DEBUG unplug %x\n", xen_emul_unplug); > + } > xen_platform_pci_unplug = xen_emul_unplug; > } >dmesg output from busybox and qemu log are in attachment -- Łukasz Oleś _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Stefano Stabellini
2010-Jul-20 14:47 UTC
Re: [Xen-devel] kernel BUG at drivers/xen/balloon.c:528!
On Tue, 20 Jul 2010, Łukasz Oleś wrote:> > diff --git a/arch/x86/xen/platform-pci-unplug.c b/arch/x86/xen/platform-pci-unplug.c > > index 2f7f3fb..4513202 100644 > > --- a/arch/x86/xen/platform-pci-unplug.c > > +++ b/arch/x86/xen/platform-pci-unplug.c > > @@ -98,8 +98,10 @@ void __init xen_unplug_emulated_devices(void) > > } > > } > > /* Now unplug the emulated devices */ > > - if (!(xen_emul_unplug & XEN_UNPLUG_IGNORE)) > > + if (!(xen_emul_unplug & XEN_UNPLUG_IGNORE)) { > > outw(xen_emul_unplug, XEN_IOPORT_UNPLUG); > > + printk("DEBUG unplug %x\n", xen_emul_unplug); > > + } > > xen_platform_pci_unplug = xen_emul_unplug; > > } > > > > dmesg output from busybox and qemu log are in attachment >As you can see from the logs xen_emul_unplug was 0 and that shouldn''t cause any unplug at all in qemu. I see that you have a complicated set of kernel command line options, what guest are you using? Is there a simple way for me to reproduce your problem? I tried to remove the platform pci driver from my kernel''s config but I am still able to mount the emulated disk without any issues. _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Łukasz Oleś
2010-Jul-21 11:51 UTC
Re: [Xen-devel] kernel BUG at drivers/xen/balloon.c:528!
W dniu 20 lipca 2010 16:47 użytkownik Stefano Stabellini> I see that you have a complicated set of kernel command line options, > what guest are you using? Is there a simple way for me to reproduce > your problem? > I tried to remove the platform pci driver from my kernel''s config but I > am still able to mount the emulated disk without any issues. >I''m using slax based system. To reproduce it firstly you need apply the aufs patches from http://aufs.sourceforge.net/. (using standalone git is faster). I also tested in on debian testing and there is the same problem. Maybe this problem is hardware specific? I will try on other machine. -- Łukasz Oleś _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Stefano Stabellini
2010-Jul-21 13:30 UTC
Re: [Xen-devel] kernel BUG at drivers/xen/balloon.c:528!
On Wed, 21 Jul 2010, Łukasz Oleś wrote:> W dniu 20 lipca 2010 16:47 użytkownik Stefano Stabellini > > I see that you have a complicated set of kernel command line options, > > what guest are you using? Is there a simple way for me to reproduce > > your problem? > > I tried to remove the platform pci driver from my kernel''s config but I > > am still able to mount the emulated disk without any issues. > > > > I''m using slax based system. To reproduce it firstly you need apply > the aufs patches from http://aufs.sourceforge.net/. (using standalone > git is faster). > I also tested in on debian testing and there is the same problem. > > Maybe this problem is hardware specific? I will try on other machine. >I tested on debian lenny and works fine for me. _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Łukasz Oleś
2010-Jul-25 20:58 UTC
Re: [Xen-devel] kernel BUG at drivers/xen/balloon.c:528!
On Wednesday 21 July 2010 15:30:45 Stefano Stabellini wrote:> On Wed, 21 Jul 2010, Łukasz Oleś wrote: > > W dniu 20 lipca 2010 16:47 użytkownik Stefano Stabellini > > > > > I see that you have a complicated set of kernel command line options, > > > what guest are you using? Is there a simple way for me to reproduce > > > your problem? > > > I tried to remove the platform pci driver from my kernel''s config but I > > > am still able to mount the emulated disk without any issues. > > > > I''m using slax based system. To reproduce it firstly you need apply > > the aufs patches from http://aufs.sourceforge.net/. (using standalone > > git is faster). > > I also tested in on debian testing and there is the same problem. > > > > Maybe this problem is hardware specific? I will try on other machine. > > I tested on debian lenny and works fine for me.I tried on four computers and it worked only on one(with XenServer). -- Łukasz Oleś _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Łukasz Oleś
2010-Jul-26 10:06 UTC
Re: [Xen-devel] kernel BUG at drivers/xen/balloon.c:528!
2010/7/25 Łukasz Oleś <lukaszoles@gmail.com>:> I tried on four computers and it worked only on one(with XenServer).With enabled option XEN_PLATFORM_PCI it doesn''t work on XenServer and works on the others. On XenServer it hangs during boot, see attachment. -- Łukasz Oleś _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Stefano Stabellini
2010-Jul-26 16:01 UTC
Re: [Xen-devel] kernel BUG at drivers/xen/balloon.c:528!
On Mon, 26 Jul 2010, Łukasz Oleś wrote:> 2010/7/25 Łukasz Oleś <lukaszoles@gmail.com>: > > I tried on four computers and it worked only on one(with XenServer). > > With enabled option XEN_PLATFORM_PCI it doesn''t work on XenServer and > works on the others. > > On XenServer it hangs during boot, see attachment. >Unfortunately XenServer doesn''t support Linux PVonHVM guests at the moment. All the others should work though. Are you still using the same vm config file you posted here? http://lists.xensource.com/archives/html/xen-devel/2010-07/msg00777.html _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Łukasz Oleś
2010-Jul-26 16:28 UTC
Re: [Xen-devel] kernel BUG at drivers/xen/balloon.c:528!
On Monday 26 July 2010 18:01:12 Stefano Stabellini wrote:> On Mon, 26 Jul 2010, Łukasz Oleś wrote: > > 2010/7/25 Łukasz Oleś <lukaszoles@gmail.com>: > > > I tried on four computers and it worked only on one(with XenServer). > > > > With enabled option XEN_PLATFORM_PCI it doesn''t work on XenServer and > > works on the others. > > > > On XenServer it hangs during boot, see attachment. > > Unfortunately XenServer doesn''t support Linux PVonHVM guests at the > moment. > All the others should work though. > > Are you still using the same vm config file you posted here? > > http://lists.xensource.com/archives/html/xen-devel/2010-07/msg00777.htmlYes, nothing changed. So, if I want to have one universal kernel which can run on top of Xen(3.4/4.x) XenServer, Vmware and Bare Metal I shouldn''t enable CONFIG_XEN_PLATFORM_PCI? -- Łukasz Oleś _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Stefano Stabellini
2010-Jul-26 16:33 UTC
Re: [Xen-devel] kernel BUG at drivers/xen/balloon.c:528!
On Mon, 26 Jul 2010, Łukasz Oleś wrote:> On Monday 26 July 2010 18:01:12 Stefano Stabellini wrote: > > On Mon, 26 Jul 2010, Łukasz Oleś wrote: > > > 2010/7/25 Łukasz Oleś <lukaszoles@gmail.com>: > > > > I tried on four computers and it worked only on one(with XenServer). > > > > > > With enabled option XEN_PLATFORM_PCI it doesn''t work on XenServer and > > > works on the others. > > > > > > On XenServer it hangs during boot, see attachment. > > > > Unfortunately XenServer doesn''t support Linux PVonHVM guests at the > > moment. > > All the others should work though. > > > > Are you still using the same vm config file you posted here? > > > > http://lists.xensource.com/archives/html/xen-devel/2010-07/msg00777.html > > Yes, nothing changed. > > So, if I want to have one universal kernel which can run on top of > Xen(3.4/4.x) XenServer, Vmware and Bare Metal I shouldn''t enable > CONFIG_XEN_PLATFORM_PCI?Yes. The next XenServer release should support PV on HVM kernels though. At that point you''ll be able to have one universal kernel with PV on HVM drivers. _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel