Re: commit f8f559422b (KVM: MMU: fast invalidate all mmio sptes) causes hang
On Jul 18, 2013, at 3:31 AM, Gleb Natapov wrote: > On Thu, Jul 18, 2013 at 03:07:11AM +0800, Xiao Guangrong wrote: >> Ouch! >> >> The guest is hang when boot from ubuntu iso, i bisected the code and found >> it is caused by: >> >> cad55f1aed000db46dfd1b5706707f33624d8604 is the first bad commit >> commit cad55f1aed000db46dfd1b5706707f33624d8604 >> Author: Cornelia Huck >> Date: Wed Jul 3 16:30:53 2013 +0200 >> >>KVM: kvm-io: support cookies >> >>Add new functions kvm_io_bus_{read,write}_cookie() that allows users of >>the kvm io infrastructure to use a cookie value to speed up lookup of a >>device on an io bus. >> >>Signed-off-by: Cornelia Huck >>Signed-off-by: Paolo Bonzini >> >> Now, i have reset the code behind this commit and continue to install the >> guest…... >> > You are on an old queue branch (the commit was never in next). I fixed it not > long time ago. Can you > pull? Yes. The issue is gone after update the code. Thank you, Gleb! -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
Re: commit f8f559422b (KVM: MMU: fast invalidate all mmio sptes) causes hang
On 2013.07.18 at 03:27 +0800, Xiao Guangrong wrote: > > Could you please try this: > > index 460b9c6..16d1578 100644 > --- a/arch/x86/kvm/mmu.c > +++ b/arch/x86/kvm/mmu.c > @@ -2952,6 +2952,9 @@ static bool page_fault_can_be_fast(u32 error_code) > !(error_code & PFERR_WRITE_MASK)) > return false; > > + if (error_code & PFERR_RSVD_MASK) > + return false; > + > return true; > } This fixes the issue for me. Many thanks Xiao. -- Markus -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
Re: commit f8f559422b (KVM: MMU: fast invalidate all mmio sptes) causes hang
On Thu, Jul 18, 2013 at 03:07:11AM +0800, Xiao Guangrong wrote: > Ouch! > > The guest is hang when boot from ubuntu iso, i bisected the code and found it > is caused by: > > cad55f1aed000db46dfd1b5706707f33624d8604 is the first bad commit > commit cad55f1aed000db46dfd1b5706707f33624d8604 > Author: Cornelia Huck > Date: Wed Jul 3 16:30:53 2013 +0200 > > KVM: kvm-io: support cookies > > Add new functions kvm_io_bus_{read,write}_cookie() that allows users of > the kvm io infrastructure to use a cookie value to speed up lookup of a > device on an io bus. > > Signed-off-by: Cornelia Huck > Signed-off-by: Paolo Bonzini > > Now, i have reset the code behind this commit and continue to install the > guest…... > You are on an old queue branch (the commit was never in next). I fixed it not long time ago. Can you pull? -- Gleb. -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
Re: commit f8f559422b (KVM: MMU: fast invalidate all mmio sptes) causes hang
On Wed, Jul 17, 2013 at 09:00:29PM +0200, Markus Trippelsdorf wrote: > > > 10346403 qemu-system-x86-1403 [001] 1451.422260: check_mmio_spte: > > > spte fee00e3f kvm_gen 7ffc8 spte-gen 7ffc7 valid 0 > > > ... Can you trace only this event and record stack information: trace-cmd record -T -e kvmmmu:check_mmio_spte > > > > [ 236.322519] qemu-system-x86 R running task0 603367 > 0x Is this the task that now prints check_mmio_spte ad nauseam? > [ 236.322523] 0001 7fff27e6b950 > 000fee00 > [ 236.322528] 8801df8b3b90 fee000b0 8801dff0b0d1 > > [ 236.322533] 8801df8b3b90 8801f7614040 > > [ 236.322537] Call Trace: > [ 236.322544] [] ? nested_svm_get_tdp_cr3+0x20/0x20 I am assuming you are not running nested guest and this is just a glitch in the matrix, correct? > [ 236.322551] [] ? kvm_arch_vcpu_ioctl_run+0x415/0x10c0 > [ 236.322556] [] ? page_add_new_anon_rmap+0x2b/0x100 > [ 236.322561] [] ? handle_pte_fault+0x69b/0x7c0 > [ 236.322567] [] ? kvm_arch_vcpu_load+0x4a/0x1e0 > [ 236.322573] [] ? kvm_vcpu_ioctl+0x271/0x500 > [ 236.322578] [] ? do_vfs_ioctl+0x2cc/0x4a0 > [ 236.322583] [] ? SyS_ioctl+0x3c/0x80 > [ 236.322588] [] ? system_call_fastpath+0x16/0x1b -- Gleb. -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
Re: commit f8f559422b (KVM: MMU: fast invalidate all mmio sptes) causes hang
Hi, Could you please try this: index 460b9c6..16d1578 100644 --- a/arch/x86/kvm/mmu.c +++ b/arch/x86/kvm/mmu.c @@ -2952,6 +2952,9 @@ static bool page_fault_can_be_fast(u32 error_code) !(error_code & PFERR_WRITE_MASK)) return false; + if (error_code & PFERR_RSVD_MASK) + return false; + return true; } On Jul 18, 2013, at 2:42 AM, Markus Trippelsdorf wrote: > On 2013.07.18 at 01:25 +0800, Xiao Guangrong wrote: >> Hi Markus, >> >> Thanks for your report! I am downloading the ubuntu iso and will try to >> reproducible it. >> >> Could you please enable these tracepoints: >> kvmmmu/check_mmio_spte >> kvmmmu/handle_mmio_page_fault >> kvmmmu/kvm_mmu_invalidate_zap_all_pages >> kvmmmu/mark_mmio_spte >> kvm/kvm_mmio >> kvm/vcpu_match_mmio >> >> and post the output? >> > First line repeated ad nauseam. > ... > 10346367 qemu-system-x86-1403 [001] 1451.45: check_mmio_spte: > spte fee00e3f kvm_gen 7ffc8 spte-gen 7ffc7 valid 0 > 10346368 qemu-system-x86-1403 [001] 1451.46: check_mmio_spte: > spte fee00e3f kvm_gen 7ffc8 spte-gen 7ffc7 valid 0 > 10346369 qemu-system-x86-1403 [001] 1451.47: check_mmio_spte: > spte fee00e3f kvm_gen 7ffc8 spte-gen 7ffc7 valid 0 > 10346370 qemu-system-x86-1403 [001] 1451.48: check_mmio_spte: > spte fee00e3f kvm_gen 7ffc8 spte-gen 7ffc7 valid 0 > 10346371 qemu-system-x86-1403 [001] 1451.49: check_mmio_spte: > spte fee00e3f kvm_gen 7ffc8 spte-gen 7ffc7 valid 0index > 460b9c6..16d1578 100644 > --- a/arch/x86/kvm/mmu.c > +++ b/arch/x86/kvm/mmu.c > @@ -2952,6 +2952,9 @@ static bool page_fault_can_be_fast(u32 error_code) > !(error_code & PFERR_WRITE_MASK)) > return false; > > + if (error_code & PFERR_RSVD_MASK) > + return false; > + > return true; > }10346372 qemu-system-x86-1404 [002] 1451.49: kvm_mmio: > mmio write len 4 gpa 0xfee000b0 val 0x0 > 10346373 qemu-system-x86-1403 [001] 1451.422230: check_mmio_spte: > spte fee00e3f kvm_gen 7ffc8 spte-gen 7ffc7 valid 0 > 10346374 qemu-system-x86-1403 [001] 1451.422231: check_mmio_spte: > spte fee00e3f kvm_gen 7ffc8 spte-gen 7ffc7 valid 0 > 10346375 qemu-system-x86-1403 [001] 1451.422232: check_mmio_spte: > spte fee00e3f kvm_gen 7ffc8 spte-gen 7ffc7 valid 0 > 10346376 qemu-system-x86-1403 [001] 1451.422233: check_mmio_spte: > spte fee00e3f kvm_gen 7ffc8 spte-gen 7ffc7 valid 0 > 10346377 qemu-system-x86-1403 [001] 1451.422234: check_mmio_spte: > spte fee00e3f kvm_gen 7ffc8 spte-gen 7ffc7 valid 0 > 10346378 qemu-system-x86-1403 [001] 1451.422235: check_mmio_spte: > spte fee00e3f kvm_gen 7ffc8 spte-gen 7ffc7 valid 0 > 10346379 qemu-system-x86-1404 [002] 1451.422236: kvm_mmio: > mmio write len 4 gpa 0xfee00380 val 0x3c7b0 > 10346380 qemu-system-x86-1403 [001] 1451.422237: check_mmio_spte: > spte fee00e3f kvm_gen 7ffc8 spte-gen 7ffc7 valid 0 > 10346381 qemu-system-x86-1403 [001] 1451.422238: check_mmio_spte: > spte fee00e3f kvm_gen 7ffc8 spte-gen 7ffc7 valid 0 > 10346382 qemu-system-x86-1403 [001] 1451.422239: check_mmio_spte: > spte fee00e3f kvm_gen 7ffc8 spte-gen 7ffc7 valid 0 > 10346383 qemu-system-x86-1403 [001] 1451.422240: check_mmio_spte: > spte fee00e3f kvm_gen 7ffc8 spte-gen 7ffc7 valid 0 > 10346384 qemu-system-x86-1403 [001] 1451.422241: check_mmio_spte: > spte fee00e3f kvm_gen 7ffc8 spte-gen 7ffc7 valid 0 > 10346385 qemu-system-x86-1403 [001] 1451.422242: check_mmio_spte: > spte fee00e3f kvm_gen 7ffc8 spte-gen 7ffc7 valid 0 > 10346386 qemu-system-x86-1403 [001] 1451.422243: check_mmio_spte: > spte fee00e3f kvm_gen 7ffc8 spte-gen 7ffc7 valid 0 > 10346387 qemu-system-x86-1403 [001] 1451.422244: check_mmio_spte: > spte fee00e3f kvm_gen 7ffc8 spte-gen 7ffc7 valid 0 > 10346403 qemu-system-x86-1403 [001] 1451.422260: check_mmio_spte: > spte fee00e3f kvm_gen 7ffc8 spte-gen 7ffc7 valid 0 > ... > Last line repeated ad nauseam. > > Whole output is over 1.3G of size. > > -- > Markus -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
Re: commit f8f559422b (KVM: MMU: fast invalidate all mmio sptes) causes hang
Ouch! The guest is hang when boot from ubuntu iso, i bisected the code and found it is caused by: cad55f1aed000db46dfd1b5706707f33624d8604 is the first bad commit commit cad55f1aed000db46dfd1b5706707f33624d8604 Author: Cornelia Huck Date: Wed Jul 3 16:30:53 2013 +0200 KVM: kvm-io: support cookies Add new functions kvm_io_bus_{read,write}_cookie() that allows users of the kvm io infrastructure to use a cookie value to speed up lookup of a device on an io bus. Signed-off-by: Cornelia Huck Signed-off-by: Paolo Bonzini Now, i have reset the code behind this commit and continue to install the guest…... On Jul 18, 2013, at 12:57 AM, Markus Trippelsdorf wrote: > On 2013.07.17 at 19:42 +0300, Gleb Natapov wrote: >> On Wed, Jul 17, 2013 at 06:32:15PM +0200, Markus Trippelsdorf wrote: >>> The following commit: >>> commit f8f559422b6c6a05469dfde614b67789b6142cb5 >>> Author: Xiao Guangrong >>> Date: Fri Jun 7 16:51:26 2013 +0800 >>> >>>KVM: MMU: fast invalidate all mmio sptes >>> >>> causes my kvm qemu qcow2 images to hang when they switch from the boot >>> console to the desktop. For example: >>> qemu-system-x86_64 -enable-kvm -net nic,vlan=0,model=virtio -net user >>> -drive file=ubuntu,if=virtio,cache=unsafe -smp 2 -m 1024 >>> boots normally at first, but then hangs as soon as Xorg gets started.` >>> >> Which guest is is exactly and how reproducible is this? > > Ubuntu 12.04. > It's 100% reproducible or else it wouldn't have been so easily > bisectable. > >>> Perf top shows: >>> 20.99% [kernel] [k] svm_vcpu_run >>> 17.24% [kernel] [k] >>> handle_mmio_page_fault_common >>> 15.61% [kernel] [k] >>> kvm_arch_vcpu_ioctl_run >>> 12.49% [kernel] [k] >>> fast_page_fault.part.78 >>> >>> I'm running: >>> QEMU emulator version 1.4.2 >>> on an AMD PhenomII X4 CPU. >>> >> What is the output of "cat /proc/cpuinfo"? > > > processor : 0 > vendor_id : AuthenticAMD > cpu family: 16 > model : 4 > model name: AMD Phenom(tm) II X4 955 Processor > stepping : 2 > microcode : 0x1db > cpu MHz : 800.000 > cache size: 512 KB > physical id : 0 > siblings : 4 > core id : 0 > cpu cores : 4 > apicid: 0 > initial apicid: 0 > fpu : yes > fpu_exception : yes > cpuid level : 5 > wp: yes > flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov > pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb > rdtscp lm 3dnowext 3dnow constant_tsc rep_good nopl nonstop_tsc extd_apicid > pni monitor cx16 popcnt lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a > misalignsse 3dnowprefetch osvw ibs skinit wdt hw_pstate npt lbrv svm_lock > nrip_save > bogomips : 6424.73 > TLB size : 1024 4K pages > clflush size : 64 > cache_alignment : 64 > address sizes : 48 bits physical, 48 bits virtual > power management: ts ttp tm stc 100mhzsteps hwpstate > ... > *4 > > -- > Markus > -- > To unsubscribe from this list: send the line "unsubscribe kvm" in > the body of a message to majord...@vger.kernel.org > More majordomo info at http://vger.kernel.org/majordomo-info.html -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
Re: commit f8f559422b (KVM: MMU: fast invalidate all mmio sptes) causes hang
On Wed, Jul 17, 2013 at 08:42:26PM +0200, Markus Trippelsdorf wrote: > On 2013.07.18 at 01:25 +0800, Xiao Guangrong wrote: > > Hi Markus, > > > > Thanks for your report! I am downloading the ubuntu iso and will try to > > reproducible it. > > > > Could you please enable these tracepoints: > > kvmmmu/check_mmio_spte > > kvmmmu/handle_mmio_page_fault > > kvmmmu/kvm_mmu_invalidate_zap_all_pages > > kvmmmu/mark_mmio_spte > > kvm/kvm_mmio > > kvm/vcpu_match_mmio > > > > and post the output? > > > First line repeated ad nauseam. > ... > 10346367 qemu-system-x86-1403 [001] 1451.45: check_mmio_spte: > spte fee00e3f kvm_gen 7ffc8 spte-gen 7ffc7 valid 0 > 10346368 qemu-system-x86-1403 [001] 1451.46: check_mmio_spte: > spte fee00e3f kvm_gen 7ffc8 spte-gen 7ffc7 valid 0 > 10346369 qemu-system-x86-1403 [001] 1451.47: check_mmio_spte: > spte fee00e3f kvm_gen 7ffc8 spte-gen 7ffc7 valid 0 > 10346370 qemu-system-x86-1403 [001] 1451.48: check_mmio_spte: > spte fee00e3f kvm_gen 7ffc8 spte-gen 7ffc7 valid 0 > 10346371 qemu-system-x86-1403 [001] 1451.49: check_mmio_spte: > spte fee00e3f kvm_gen 7ffc8 spte-gen 7ffc7 valid 0 > 10346372 qemu-system-x86-1404 [002] 1451.49: kvm_mmio: > mmio write len 4 gpa 0xfee000b0 val 0x0 > 10346373 qemu-system-x86-1403 [001] 1451.422230: check_mmio_spte: > spte fee00e3f kvm_gen 7ffc8 spte-gen 7ffc7 valid 0 > 10346374 qemu-system-x86-1403 [001] 1451.422231: check_mmio_spte: > spte fee00e3f kvm_gen 7ffc8 spte-gen 7ffc7 valid 0 > 10346375 qemu-system-x86-1403 [001] 1451.422232: check_mmio_spte: > spte fee00e3f kvm_gen 7ffc8 spte-gen 7ffc7 valid 0 > 10346376 qemu-system-x86-1403 [001] 1451.422233: check_mmio_spte: > spte fee00e3f kvm_gen 7ffc8 spte-gen 7ffc7 valid 0 > 10346377 qemu-system-x86-1403 [001] 1451.422234: check_mmio_spte: > spte fee00e3f kvm_gen 7ffc8 spte-gen 7ffc7 valid 0 > 10346378 qemu-system-x86-1403 [001] 1451.422235: check_mmio_spte: > spte fee00e3f kvm_gen 7ffc8 spte-gen 7ffc7 valid 0 > 10346379 qemu-system-x86-1404 [002] 1451.422236: kvm_mmio: > mmio write len 4 gpa 0xfee00380 val 0x3c7b0 > 10346380 qemu-system-x86-1403 [001] 1451.422237: check_mmio_spte: > spte fee00e3f kvm_gen 7ffc8 spte-gen 7ffc7 valid 0 > 10346381 qemu-system-x86-1403 [001] 1451.422238: check_mmio_spte: > spte fee00e3f kvm_gen 7ffc8 spte-gen 7ffc7 valid 0 > 10346382 qemu-system-x86-1403 [001] 1451.422239: check_mmio_spte: > spte fee00e3f kvm_gen 7ffc8 spte-gen 7ffc7 valid 0 > 10346383 qemu-system-x86-1403 [001] 1451.422240: check_mmio_spte: > spte fee00e3f kvm_gen 7ffc8 spte-gen 7ffc7 valid 0 > 10346384 qemu-system-x86-1403 [001] 1451.422241: check_mmio_spte: > spte fee00e3f kvm_gen 7ffc8 spte-gen 7ffc7 valid 0 > 10346385 qemu-system-x86-1403 [001] 1451.422242: check_mmio_spte: > spte fee00e3f kvm_gen 7ffc8 spte-gen 7ffc7 valid 0 > 10346386 qemu-system-x86-1403 [001] 1451.422243: check_mmio_spte: > spte fee00e3f kvm_gen 7ffc8 spte-gen 7ffc7 valid 0 > 10346387 qemu-system-x86-1403 [001] 1451.422244: check_mmio_spte: > spte fee00e3f kvm_gen 7ffc8 spte-gen 7ffc7 valid 0 > 10346403 qemu-system-x86-1403 [001] 1451.422260: check_mmio_spte: > spte fee00e3f kvm_gen 7ffc8 spte-gen 7ffc7 valid 0 > ... > Last line repeated ad nauseam. > While this happens can you run "echo t > /proc/sysrq-trigger" and attach dmesg here. -- Gleb. -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
Re: commit f8f559422b (KVM: MMU: fast invalidate all mmio sptes) causes hang
On 2013.07.18 at 01:25 +0800, Xiao Guangrong wrote: > Hi Markus, > > Thanks for your report! I am downloading the ubuntu iso and will try to > reproducible it. > > Could you please enable these tracepoints: > kvmmmu/check_mmio_spte > kvmmmu/handle_mmio_page_fault > kvmmmu/kvm_mmu_invalidate_zap_all_pages > kvmmmu/mark_mmio_spte > kvm/kvm_mmio > kvm/vcpu_match_mmio > > and post the output? > First line repeated ad nauseam. ... 10346367 qemu-system-x86-1403 [001] 1451.45: check_mmio_spte: spte fee00e3f kvm_gen 7ffc8 spte-gen 7ffc7 valid 0 10346368 qemu-system-x86-1403 [001] 1451.46: check_mmio_spte: spte fee00e3f kvm_gen 7ffc8 spte-gen 7ffc7 valid 0 10346369 qemu-system-x86-1403 [001] 1451.47: check_mmio_spte: spte fee00e3f kvm_gen 7ffc8 spte-gen 7ffc7 valid 0 10346370 qemu-system-x86-1403 [001] 1451.48: check_mmio_spte: spte fee00e3f kvm_gen 7ffc8 spte-gen 7ffc7 valid 0 10346371 qemu-system-x86-1403 [001] 1451.49: check_mmio_spte: spte fee00e3f kvm_gen 7ffc8 spte-gen 7ffc7 valid 0 10346372 qemu-system-x86-1404 [002] 1451.49: kvm_mmio: mmio write len 4 gpa 0xfee000b0 val 0x0 10346373 qemu-system-x86-1403 [001] 1451.422230: check_mmio_spte: spte fee00e3f kvm_gen 7ffc8 spte-gen 7ffc7 valid 0 10346374 qemu-system-x86-1403 [001] 1451.422231: check_mmio_spte: spte fee00e3f kvm_gen 7ffc8 spte-gen 7ffc7 valid 0 10346375 qemu-system-x86-1403 [001] 1451.422232: check_mmio_spte: spte fee00e3f kvm_gen 7ffc8 spte-gen 7ffc7 valid 0 10346376 qemu-system-x86-1403 [001] 1451.422233: check_mmio_spte: spte fee00e3f kvm_gen 7ffc8 spte-gen 7ffc7 valid 0 10346377 qemu-system-x86-1403 [001] 1451.422234: check_mmio_spte: spte fee00e3f kvm_gen 7ffc8 spte-gen 7ffc7 valid 0 10346378 qemu-system-x86-1403 [001] 1451.422235: check_mmio_spte: spte fee00e3f kvm_gen 7ffc8 spte-gen 7ffc7 valid 0 10346379 qemu-system-x86-1404 [002] 1451.422236: kvm_mmio: mmio write len 4 gpa 0xfee00380 val 0x3c7b0 10346380 qemu-system-x86-1403 [001] 1451.422237: check_mmio_spte: spte fee00e3f kvm_gen 7ffc8 spte-gen 7ffc7 valid 0 10346381 qemu-system-x86-1403 [001] 1451.422238: check_mmio_spte: spte fee00e3f kvm_gen 7ffc8 spte-gen 7ffc7 valid 0 10346382 qemu-system-x86-1403 [001] 1451.422239: check_mmio_spte: spte fee00e3f kvm_gen 7ffc8 spte-gen 7ffc7 valid 0 10346383 qemu-system-x86-1403 [001] 1451.422240: check_mmio_spte: spte fee00e3f kvm_gen 7ffc8 spte-gen 7ffc7 valid 0 10346384 qemu-system-x86-1403 [001] 1451.422241: check_mmio_spte: spte fee00e3f kvm_gen 7ffc8 spte-gen 7ffc7 valid 0 10346385 qemu-system-x86-1403 [001] 1451.422242: check_mmio_spte: spte fee00e3f kvm_gen 7ffc8 spte-gen 7ffc7 valid 0 10346386 qemu-system-x86-1403 [001] 1451.422243: check_mmio_spte: spte fee00e3f kvm_gen 7ffc8 spte-gen 7ffc7 valid 0 10346387 qemu-system-x86-1403 [001] 1451.422244: check_mmio_spte: spte fee00e3f kvm_gen 7ffc8 spte-gen 7ffc7 valid 0 10346403 qemu-system-x86-1403 [001] 1451.422260: check_mmio_spte: spte fee00e3f kvm_gen 7ffc8 spte-gen 7ffc7 valid 0 ... Last line repeated ad nauseam. Whole output is over 1.3G of size. -- Markus -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
Re: commit f8f559422b (KVM: MMU: fast invalidate all mmio sptes) causes hang
On 2013.07.18 at 01:25 +0800, Xiao Guangrong wrote: > Hi Markus, > > Thanks for your report! I am downloading the ubuntu iso and will try to > reproducible it. Please note that the exact guest doesn't matter. I first noticed the hang when I tried to install Fedora 19 today. > Could you please enable these tracepoints: > kvmmmu/check_mmio_spte > kvmmmu/handle_mmio_page_fault > kvmmmu/kvm_mmu_invalidate_zap_all_pages > kvmmmu/mark_mmio_spte > kvm/kvm_mmio > kvm/vcpu_match_mmio > > and post the output? I will run this later. > Just check, your host is 64-bit or 32 bit? Both guest and host are 64-bit. -- Markus -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
Re: commit f8f559422b (KVM: MMU: fast invalidate all mmio sptes) causes hang
Hi Markus, Thanks for your report! I am downloading the ubuntu iso and will try to reproducible it. Could you please enable these tracepoints: kvmmmu/check_mmio_spte kvmmmu/handle_mmio_page_fault kvmmmu/kvm_mmu_invalidate_zap_all_pages kvmmmu/mark_mmio_spte kvm/kvm_mmio kvm/vcpu_match_mmio and post the output? Just check, your host is 64-bit or 32 bit? On Jul 18, 2013, at 12:57 AM, Markus Trippelsdorf wrote: > On 2013.07.17 at 19:42 +0300, Gleb Natapov wrote: >> On Wed, Jul 17, 2013 at 06:32:15PM +0200, Markus Trippelsdorf wrote: >>> The following commit: >>> commit f8f559422b6c6a05469dfde614b67789b6142cb5 >>> Author: Xiao Guangrong >>> Date: Fri Jun 7 16:51:26 2013 +0800 >>> >>>KVM: MMU: fast invalidate all mmio sptes >>> >>> causes my kvm qemu qcow2 images to hang when they switch from the boot >>> console to the desktop. For example: >>> qemu-system-x86_64 -enable-kvm -net nic,vlan=0,model=virtio -net user >>> -drive file=ubuntu,if=virtio,cache=unsafe -smp 2 -m 1024 >>> boots normally at first, but then hangs as soon as Xorg gets started.` >>> >> Which guest is is exactly and how reproducible is this? > > Ubuntu 12.04. > It's 100% reproducible or else it wouldn't have been so easily > bisectable. > >>> Perf top shows: >>> 20.99% [kernel] [k] svm_vcpu_run >>> 17.24% [kernel] [k] >>> handle_mmio_page_fault_common >>> 15.61% [kernel] [k] >>> kvm_arch_vcpu_ioctl_run >>> 12.49% [kernel] [k] >>> fast_page_fault.part.78 >>> >>> I'm running: >>> QEMU emulator version 1.4.2 >>> on an AMD PhenomII X4 CPU. >>> >> What is the output of "cat /proc/cpuinfo"? > > > processor : 0 > vendor_id : AuthenticAMD > cpu family: 16 > model : 4 > model name: AMD Phenom(tm) II X4 955 Processor > stepping : 2 > microcode : 0x1db > cpu MHz : 800.000 > cache size: 512 KB > physical id : 0 > siblings : 4 > core id : 0 > cpu cores : 4 > apicid: 0 > initial apicid: 0 > fpu : yes > fpu_exception : yes > cpuid level : 5 > wp: yes > flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov > pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb > rdtscp lm 3dnowext 3dnow constant_tsc rep_good nopl nonstop_tsc extd_apicid > pni monitor cx16 popcnt lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a > misalignsse 3dnowprefetch osvw ibs skinit wdt hw_pstate npt lbrv svm_lock > nrip_save > bogomips : 6424.73 > TLB size : 1024 4K pages > clflush size : 64 > cache_alignment : 64 > address sizes : 48 bits physical, 48 bits virtual > power management: ts ttp tm stc 100mhzsteps hwpstate > ... > *4 > > -- > Markus > -- > To unsubscribe from this list: send the line "unsubscribe kvm" in > the body of a message to majord...@vger.kernel.org > More majordomo info at http://vger.kernel.org/majordomo-info.html -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
Re: commit f8f559422b (KVM: MMU: fast invalidate all mmio sptes) causes hang
On 2013.07.17 at 19:42 +0300, Gleb Natapov wrote: > On Wed, Jul 17, 2013 at 06:32:15PM +0200, Markus Trippelsdorf wrote: > > The following commit: > > commit f8f559422b6c6a05469dfde614b67789b6142cb5 > > Author: Xiao Guangrong > > Date: Fri Jun 7 16:51:26 2013 +0800 > > > > KVM: MMU: fast invalidate all mmio sptes > > > > causes my kvm qemu qcow2 images to hang when they switch from the boot > > console to the desktop. For example: > > qemu-system-x86_64 -enable-kvm -net nic,vlan=0,model=virtio -net user > > -drive file=ubuntu,if=virtio,cache=unsafe -smp 2 -m 1024 > > boots normally at first, but then hangs as soon as Xorg gets started.` > > > Which guest is is exactly and how reproducible is this? Ubuntu 12.04. It's 100% reproducible or else it wouldn't have been so easily bisectable. > > Perf top shows: > > 20.99% [kernel] [k] svm_vcpu_run > > 17.24% [kernel] [k] > > handle_mmio_page_fault_common > > 15.61% [kernel] [k] > > kvm_arch_vcpu_ioctl_run > > 12.49% [kernel] [k] > > fast_page_fault.part.78 > > > > I'm running: > > QEMU emulator version 1.4.2 > > on an AMD PhenomII X4 CPU. > > > What is the output of "cat /proc/cpuinfo"? processor : 0 vendor_id : AuthenticAMD cpu family : 16 model : 4 model name : AMD Phenom(tm) II X4 955 Processor stepping: 2 microcode : 0x1db cpu MHz : 800.000 cache size : 512 KB physical id : 0 siblings: 4 core id : 0 cpu cores : 4 apicid : 0 initial apicid : 0 fpu : yes fpu_exception : yes cpuid level : 5 wp : yes flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm 3dnowext 3dnow constant_tsc rep_good nopl nonstop_tsc extd_apicid pni monitor cx16 popcnt lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt hw_pstate npt lbrv svm_lock nrip_save bogomips: 6424.73 TLB size: 1024 4K pages clflush size: 64 cache_alignment : 64 address sizes : 48 bits physical, 48 bits virtual power management: ts ttp tm stc 100mhzsteps hwpstate ... *4 -- Markus -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
Re: commit f8f559422b (KVM: MMU: fast invalidate all mmio sptes) causes hang
On Wed, Jul 17, 2013 at 06:32:15PM +0200, Markus Trippelsdorf wrote: > The following commit: > commit f8f559422b6c6a05469dfde614b67789b6142cb5 > Author: Xiao Guangrong > Date: Fri Jun 7 16:51:26 2013 +0800 > > KVM: MMU: fast invalidate all mmio sptes > > causes my kvm qemu qcow2 images to hang when they switch from the boot > console to the desktop. For example: > qemu-system-x86_64 -enable-kvm -net nic,vlan=0,model=virtio -net user -drive > file=ubuntu,if=virtio,cache=unsafe -smp 2 -m 1024 > boots normally at first, but then hangs as soon as Xorg gets started.` > Which guest is is exactly and how reproducible is this? > Perf top shows: > 20.99% [kernel] [k] svm_vcpu_run > 17.24% [kernel] [k] > handle_mmio_page_fault_common > 15.61% [kernel] [k] kvm_arch_vcpu_ioctl_run > 12.49% [kernel] [k] fast_page_fault.part.78 > > I'm running: > QEMU emulator version 1.4.2 > on an AMD PhenomII X4 CPU. > What is the output of "cat /proc/cpuinfo"? -- Gleb. -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/