Hans-Christian Armingeon
2006-Jun-03 12:51 UTC
[Xen-users] irq 17: nobody cared (try booting with the "irqpoll" option)
Hi, I am running xen unstable hg from 2006-06-03 2.6.16.19 pathced with xen unstable hg from 2006-06-02 I get this error message. [ 1199.028108] irq 17: nobody cared (try booting with the "irqpoll" option) [ 1199.031054] [ 1199.031055] Call Trace: <IRQ> <ffffffff8014d754>{__report_bad_irq+48} [ 1199.034294] <ffffffff8014d9bc>{note_interrupt+539} <ffffffff8014d2ab>{__do_IRQ+190} [ 1199.040504] <ffffffff8010c934>{do_IRQ+62} <ffffffff802d0034>{evtchn_do_upcall+132} [ 1199.046403] <ffffffff8010b152>{do_hypervisor_callback+30} <EOI> [ 1199.052363] handlers: [ 1199.055141] [<ffffffff880354ab>] (ata_interrupt+0x0/0x166 [libata]) [ 1199.058267] Disabling IRQ #17 So I rebooted with irqpoll option. Is this a bug, or what''s the cause of this error message? Johnny _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Hans-Christian Armingeon
2006-Jun-04 00:48 UTC
Re: [Xen-users] irq 17: nobody cared (try booting with the "irqpoll" option)
Hi again, booting with irqpoll didn''t fix this issue. Any ideas? Johnny Am Samstag, 3. Juni 2006 14:51 schrieb Hans-Christian Armingeon:> Hi, > > I am running > xen unstable hg from 2006-06-03 > 2.6.16.19 pathced with xen unstable hg from 2006-06-02 > > I get this error message. > > [ 1199.028108] irq 17: nobody cared (try booting with the "irqpoll" option) > [ 1199.031054] > [ 1199.031055] Call Trace: <IRQ> <ffffffff8014d754>{__report_bad_irq+48} > [ 1199.034294] <ffffffff8014d9bc>{note_interrupt+539} <ffffffff8014d2ab>{__do_IRQ+190} > [ 1199.040504] <ffffffff8010c934>{do_IRQ+62} <ffffffff802d0034>{evtchn_do_upcall+132} > [ 1199.046403] <ffffffff8010b152>{do_hypervisor_callback+30} <EOI> > [ 1199.052363] handlers: > [ 1199.055141] [<ffffffff880354ab>] (ata_interrupt+0x0/0x166 [libata]) > [ 1199.058267] Disabling IRQ #17 > > So I rebooted with irqpoll option. > > Is this a bug, or what''s the cause of this error message? > > > Johnny > > _______________________________________________ > Xen-users mailing list > Xen-users@lists.xensource.com > http://lists.xensource.com/xen-users >_______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Keir Fraser
2006-Jun-04 08:03 UTC
Re: [Xen-devel] irq 17: nobody cared (try booting with the "irqpoll" option)
On 3 Jun 2006, at 13:51, Hans-Christian Armingeon wrote:> [ 1199.028108] irq 17: nobody cared (try booting with the "irqpoll" > option) > [ 1199.031054] > [ 1199.031055] Call Trace: <IRQ> > <ffffffff8014d754>{__report_bad_irq+48} > [ 1199.034294] <ffffffff8014d9bc>{note_interrupt+539} > <ffffffff8014d2ab>{__do_IRQ+190} > [ 1199.040504] <ffffffff8010c934>{do_IRQ+62} > <ffffffff802d0034>{evtchn_do_upcall+132} > [ 1199.046403] <ffffffff8010b152>{do_hypervisor_callback+30} > <EOI> > [ 1199.052363] handlers: > [ 1199.055141] [<ffffffff880354ab>] (ata_interrupt+0x0/0x166 [libata]) > [ 1199.058267] Disabling IRQ #17 > > So I rebooted with irqpoll option. > > Is this a bug, or what''s the cause of this error message?It''s caused by receiving lots of spurious interrupts on that IRQ. If you don''t see similar on native then post boot output from Xen and native Linux and we can look for differences. -- Keir _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Hans-Christian Armingeon
2006-Jun-04 22:05 UTC
[Xen-users] Re: [Xen-devel] irq 17: nobody cared (try booting with the "irqpoll" option)
Hi, thank you for your reply. Well, irqpoll didn''t fix it. I switched to nonsmp in dom0. I am doing dirty things: I export 2 USB controllers into a domU. I plugged an AVM FritzCard DSL USB into the domU. ANd I am sending faxes via capi4hylafax and hylafax. And things periodically break. cat /proc/interrupts [...] 17: 1008938 Phys-irq libata [...] The interesting thin is, that lsmod |grep ata says: sata_sil 12424 0 libata 67096 1 sata_sil scsi_mod 156696 6 sg,sd_mod,3w_xxxx,aic79xx,scsi_transport_spi,libata I have no harddisks plugged on the onboard sil S-ATA. dmesg says [...] [ 5.887645] SCSI subsystem initialized [ 5.893963] libata version 1.20 loaded. [ 5.894782] sata_sil 0000:03:05.0: version 0.9 [ 5.894806] ACPI: PCI Interrupt 0000:03:05.0[A] -> GSI 19 (level, low) -> IRQ 17 [ 5.901894] ata1: SATA max UDMA/100 cmd 0xFFFFC20000006C80 ctl 0xFFFFC20000006C8A bmdma 0xFFFFC20000006C00 irq 17 [ 5.907356] ata2: SATA max UDMA/100 cmd 0xFFFFC20000006CC0 ctl 0xFFFFC20000006CCA bmdma 0xFFFFC20000006C08 irq 17 [ 5.912833] ata3: SATA max UDMA/100 cmd 0xFFFFC20000006E80 ctl 0xFFFFC20000006E8A bmdma 0xFFFFC20000006E00 irq 17 [ 5.918288] ata4: SATA max UDMA/100 cmd 0xFFFFC20000006EC0 ctl 0xFFFFC20000006ECA bmdma 0xFFFFC20000006E08 irq 17 [ 6.124116] ata1: SATA link down (SStatus 0) [ 6.126847] scsi0 : sata_sil [ 6.328113] ata2: SATA link down (SStatus 0) [ 6.330829] scsi1 : sata_sil [ 6.532113] ata3: SATA link down (SStatus 0) [ 6.534839] scsi2 : sata_sil [ 6.736113] ata4: SATA link down (SStatus 0) [ 6.738813] scsi3 : sata_sil Maybe, unloading of sata_sil fixes this problem? But I would like to have a clean solution. Rebooting without xen is no solution - this box is in a production environment. [read on] Am Sonntag, 4. Juni 2006 10:03 schrieb Keir Fraser:> > On 3 Jun 2006, at 13:51, Hans-Christian Armingeon wrote: > > > [ 1199.028108] irq 17: nobody cared (try booting with the "irqpoll" > > option) > > [ 1199.031054] > > [ 1199.031055] Call Trace: <IRQ> > > <ffffffff8014d754>{__report_bad_irq+48} > > [ 1199.034294] <ffffffff8014d9bc>{note_interrupt+539} > > <ffffffff8014d2ab>{__do_IRQ+190} > > [ 1199.040504] <ffffffff8010c934>{do_IRQ+62} > > <ffffffff802d0034>{evtchn_do_upcall+132} > > [ 1199.046403] <ffffffff8010b152>{do_hypervisor_callback+30} > > <EOI> > > [ 1199.052363] handlers: > > [ 1199.055141] [<ffffffff880354ab>] (ata_interrupt+0x0/0x166 [libata]) > > [ 1199.058267] Disabling IRQ #17 > > > > So I rebooted with irqpoll option. > > > > Is this a bug, or what''s the cause of this error message? > > It''s caused by receiving lots of spurious interrupts on that IRQ. If > you don''t see similar on native then post boot output from Xen and > native Linux and we can look for differences.What do you mean? dmesg? cat proc/interrupts ?> > -- KeirJohnny _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Hans-Christian Armingeon
2006-Jun-04 23:10 UTC
Re: [Xen-users] Re: [Xen-devel] irq 17: nobody cared (try booting with the "irqpoll" option)
Am Montag, 5. Juni 2006 00:05 schrieb Hans-Christian Armingeon:> Hi, > > thank you for your reply. > > Well, irqpoll didn''t fix it. > > I switched to nonsmp in dom0. > > I am doing dirty things: I export 2 USB controllers into a domU. I plugged an AVM FritzCard DSL USB into the domU. ANd I am sending faxes via capi4hylafax and hylafax. And things periodically break. > > cat /proc/interrupts > [...] > 17: 1008938 Phys-irq libata > [...]Well, and on irq 17, there is the USB controller, that is "pluged" into the domU too. Are there any other things, I can do to debug this issue?> > > [...] >Johnny _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Keir Fraser
2006-Jun-05 07:51 UTC
Re: [Xen-users] Re: [Xen-devel] irq 17: nobody cared (try booting with the "irqpoll" option)
On 5 Jun 2006, at 00:10, Hans-Christian Armingeon wrote:> Well, and on irq 17, there is the USB controller, that is "pluged" > into the domU too. > > Are there any other things, I can do to debug this issue?Ah, a known issue. Jan Beulich has supplied a patch for this (email subject "replacement for noirqdebug hack"). I want to find some time this week to work on a solution I''d prefer, and float it on this list for discussion. For now, just boot with noirqdebug on your dom0 and driver domain kernel command lines. -- Keir _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Itai Tavor
2006-Jun-05 08:03 UTC
Re: [Xen-users] Re: [Xen-devel] irq 17: nobody cared (try booting with the "irqpoll" option)
On 05/06/2006, at 5:51 PM, Keir Fraser wrote:> > On 5 Jun 2006, at 00:10, Hans-Christian Armingeon wrote: > >> Well, and on irq 17, there is the USB controller, that is "pluged" >> into the domU too. >> >> Are there any other things, I can do to debug this issue? > > Ah, a known issue. Jan Beulich has supplied a patch for this (email > subject "replacement for noirqdebug hack"). I want to find some > time this week to work on a solution I''d prefer, and float it on > this list for discussion. > > For now, just boot with noirqdebug on your dom0 and driver domain > kernel command lines. > > -- KeirWill this patch deal with the problem I''m having (email thread subject "Guest domain with PCI frontend disabled IRQ, backend device dies") of both host and guest using IRQ 17? Itai _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Hans-Christian Armingeon
2006-Jun-05 09:48 UTC
Re: [Xen-users] Re: [Xen-devel] irq 17: nobody cared (try booting with the "irqpoll" option)
Hi, thank you for replying to my email. Am Montag, 5. Juni 2006 09:51 schrieb Keir Fraser:> > On 5 Jun 2006, at 00:10, Hans-Christian Armingeon wrote: > > > Well, and on irq 17, there is the USB controller, that is "pluged" > > into the domU too. > > > > Are there any other things, I can do to debug this issue? > > Ah, a known issue. Jan Beulich has supplied a patch for this (email > subject "replacement for noirqdebug hack"). I want to find some time > this week to work on a solution I''d prefer, and float it on this list > for discussion.I have extra = "swiotlb=force noirqdebug" in my config for the domU. I had to use swiotlb=force because of an oops. I removed the irqpoll option from the dom0 kernel, and replaced it with noirqdebug. Is that correct? Johnny> > For now, just boot with noirqdebug on your dom0 and driver domain > kernel command lines. > > -- KeirJohnny _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Keir Fraser
2006-Jun-05 12:57 UTC
Re: [Xen-users] Re: [Xen-devel] irq 17: nobody cared (try booting with the "irqpoll" option)
On 5 Jun 2006, at 10:48, Hans-Christian Armingeon wrote:> I have > > extra = "swiotlb=force noirqdebug" > > in my config for the domU. I had to use swiotlb=force because of an > oops. > > I removed the irqpoll option from the dom0 kernel, and replaced it > with noirqdebug. > > Is that correct?Yes. -- Keir _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Hans-Christian Armingeon
2006-Jun-05 16:58 UTC
Re: [Xen-users] Re: [Xen-devel] irq 17: nobody cared (try booting with the "irqpoll" option)
Hi Keir, my system (dom0) is very slow, after the reboot with noirqdebug. The domU is very slow, too, and consumes a lot of cpu time. What can I do? Johnny Am Montag, 5. Juni 2006 14:57 schrieb Keir Fraser:> > On 5 Jun 2006, at 10:48, Hans-Christian Armingeon wrote: > > > I have > > > > extra = "swiotlb=force noirqdebug" > > > > in my config for the domU. I had to use swiotlb=force because of an > > oops. > > > > I removed the irqpoll option from the dom0 kernel, and replaced it > > with noirqdebug. > > > > Is that correct? > > Yes. > > -- Keir > > > _______________________________________________ > Xen-devel mailing list > Xen-devel@lists.xensource.com > http://lists.xensource.com/xen-devel >_______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Keir Fraser
2006-Jun-05 17:08 UTC
Re: [Xen-users] Re: [Xen-devel] irq 17: nobody cared (try booting with the "irqpoll" option)
On 5 Jun 2006, at 17:58, Hans-Christian Armingeon wrote:> my system (dom0) is very slow, after the reboot with noirqdebug. > > The domU is very slow, too, and consumes a lot of cpu time. > > What can I do?Slower than booting without ''noirqdebug''? If you boot without noirqdebug, how long does it take before domain0 complains about one of its irqs? Does it happen only after you have started your driver domU? -- Keir _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Hans-Christian Armingeon
2006-Jun-05 17:21 UTC
Re: [Xen-users] Re: [Xen-devel] irq 17: nobody cared (try booting with the "irqpoll" option)
Am Montag, 5. Juni 2006 19:08 schrieben Sie:> > On 5 Jun 2006, at 17:58, Hans-Christian Armingeon wrote: > > > my system (dom0) is very slow, after the reboot with noirqdebug. > > > > The domU is very slow, too, and consumes a lot of cpu time. > > > > What can I do? > > Slower than booting without ''noirqdebug''?Yes. The domain with the device, that owns irq17 consumes very much cpu time, and makes a dual Opteron slow.> > If you boot without noirqdebug, how long does it take before domain0 > complains about one of its irqs? Does it happen only after you have > started your driver domU?Well, soetimes it takes a few hours, up to a few days. Johnny> > -- Keir > >_______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Keir Fraser
2006-Jun-05 17:44 UTC
Re: [Xen-users] Re: [Xen-devel] irq 17: nobody cared (try booting with the "irqpoll" option)
On 5 Jun 2006, at 18:21, Hans-Christian Armingeon wrote:>> Slower than booting without ''noirqdebug''? > > Yes. The domain with the device, that owns irq17 consumes very much > cpu time, and makes a dual Opteron slow. > >> >> If you boot without noirqdebug, how long does it take before domain0 >> complains about one of its irqs? Does it happen only after you have >> started your driver domU? > > Well, soetimes it takes a few hours, up to a few days.Until the kernel decides to disable irq17 it does less work per interrupt when booted with noirqdebug. Are you sure you weren''t still running with irqpoll? -- Keir _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Hans-Christian Armingeon
2006-Jun-05 18:07 UTC
Re: [Xen-users] Re: [Xen-devel] irq 17: nobody cared (try booting with the "irqpoll" option)
Oh, sorry Keir, I stumped in an old Bug. I have rtc support in that kernel, and didn''t remove /sbin/hwclock, so hwclock consumed all the time. That was the issue. Now the system performs normal. What can I do to debug the problem with the irq? What does the Sorry, Johnny Am Montag, 5. Juni 2006 19:21 schrieb Hans-Christian Armingeon:> Am Montag, 5. Juni 2006 19:08 schrieben Sie: > > > > On 5 Jun 2006, at 17:58, Hans-Christian Armingeon wrote: > > > > > my system (dom0) is very slow, after the reboot with noirqdebug. > > > > > > The domU is very slow, too, and consumes a lot of cpu time. > > > > > > What can I do? > > > > Slower than booting without ''noirqdebug''? > > Yes. The domain with the device, that owns irq17 consumes very much cpu time, and makes a dual Opteron slow. > > > > > If you boot without noirqdebug, how long does it take before domain0 > > complains about one of its irqs? Does it happen only after you have > > started your driver domU? > > Well, soetimes it takes a few hours, up to a few days. > > Johnny > > > > > -- Keir > > > > >_______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Keir Fraser
2006-Jun-05 18:15 UTC
Re: [Xen-users] Re: [Xen-devel] irq 17: nobody cared (try booting with the "irqpoll" option)
On 5 Jun 2006, at 19:07, Hans-Christian Armingeon wrote:> What can I do to debug the problem with the irq?Like I said, the issue is known. It can occur whenever an irq is shared by more than one driver domain (including domain 0). A patch will be forthcoming but for now noirqdebug is a perfect workaround. -- Keir _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Hans-Christian Armingeon
2006-Jun-13 08:26 UTC
Re: [Xen-users] Re: [Xen-devel] irq 17: nobody cared (try booting with the "irqpoll" option)
Hello Keir, today I got this error: [ 1146.202386] irq 17: nobody cared (try booting with the "irqpoll" option) [ 1146.205607] [ 1146.205608] Call Trace: <IRQ> <ffffffff801590a5>{__report_bad_irq+53} [ 1146.208938] <ffffffff80159333>{note_interrupt+563} <ffffffff80158a36> __do_IRQ+214} [ 1146.214881] <ffffffff8010d0d7>{do_IRQ+71} <ffffffff80311b76>{evtchn_d _upcall+166} [ 1146.220863] <ffffffff8010b6d2>{do_hypervisor_callback+30} <EOI> [ 1146.224036] <ffffffff801073aa>{hypercall_page+938} <ffffffff801073aa> hypercall_page+938} [ 1146.229934] <ffffffff8010e987>{safe_halt+119} <ffffffff801097d7>{xen_ dle+119} [ 1146.235920] <ffffffff801098a0>{cpu_idle+176} <ffffffff8058f7ca>{start kernel+506} [ 1146.241780] <ffffffff8058f102>{_sinittext+258} [ 1146.247580] handlers: [ 1146.250393] [<ffffffff88039370>] (ata_interrupt+0x0/0x1b0 [libata]) [ 1146.253445] Disabling IRQ #17 I am using xen unstable 20060610 with a 2.6.16.20 kernel. I thought, that you made a fix for interrupt sharing between dom0 and domUs. So what went wrong? Thanks in advance, Johnny Am Montag, 5. Juni 2006 20:15 schrieb Keir Fraser:> > On 5 Jun 2006, at 19:07, Hans-Christian Armingeon wrote: > > > What can I do to debug the problem with the irq? > > Like I said, the issue is known. It can occur whenever an irq is shared > by more than one driver domain (including domain 0). A patch will be > forthcoming but for now noirqdebug is a perfect workaround. > > -- Keir > > > _______________________________________________ > Xen-devel mailing list > Xen-devel@lists.xensource.com > http://lists.xensource.com/xen-devel >_______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Keir Fraser
2006-Jun-13 09:09 UTC
Re: [Xen-users] Re: [Xen-devel] irq 17: nobody cared (try booting with the "irqpoll" option)
On 13 Jun 2006, at 09:26, Hans-Christian Armingeon wrote:> I am using xen unstable 20060610 with a 2.6.16.20 kernel. > > I thought, that you made a fix for interrupt sharing between dom0 and > domUs.Yes, I did. At the point the error happened, IRQ17 was definitely in use by two separate domains on your system? -- Keir _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel
Hans-Christian Armingeon
2006-Jun-13 09:21 UTC
Re: [Xen-users] Re: [Xen-devel] irq 17: nobody cared (try booting with the "irqpoll" option)
Am Dienstag, 13. Juni 2006 11:09 schrieb Keir Fraser:> > On 13 Jun 2006, at 09:26, Hans-Christian Armingeon wrote: > > > I am using xen unstable 20060610 with a 2.6.16.20 kernel. > > > > I thought, that you made a fix for interrupt sharing between dom0 and > > domUs. > > Yes, I did. At the point the error happened, IRQ17 was definitely in > use by two separate domains on your system?Well, irq17 is in use for the onboard S-ATA, but thewre are no HDDs plugged in. And the USB Controllers are on Irq17, that are both passed to a domU. Last week, the system ran without problems with the noirqdebug option. Johnny> > -- Keir > > > _______________________________________________ > Xen-users mailing list > Xen-users@lists.xensource.com > http://lists.xensource.com/xen-users >_______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Keir Fraser
2006-Jun-13 10:14 UTC
Re: [Xen-users] Re: [Xen-devel] irq 17: nobody cared (try booting with the "irqpoll" option)
On 13 Jun 2006, at 10:21, Hans-Christian Armingeon wrote:>> Yes, I did. At the point the error happened, IRQ17 was definitely in >> use by two separate domains on your system? > > > Well, irq17 is in use for the onboard S-ATA, but thewre are no HDDs > plugged in. > And the USB Controllers are on Irq17, that are both passed to a domU. > > Last week, the system ran without problems with the noirqdebug option.If you ''cat /proc/interrupts'' in domain0 and in the domU, do both print a line for IRQ17? -- Keir _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel