diff options
Diffstat (limited to '')
| -rw-r--r-- | results/classifier/gemma3:12b/kvm/131 | 2 | ||||
| -rw-r--r-- | results/classifier/gemma3:12b/kvm/1312668 | 54 | ||||
| -rw-r--r-- | results/classifier/gemma3:12b/kvm/1318091 | 40 | ||||
| -rw-r--r-- | results/classifier/gemma3:12b/kvm/1318746 | 27 |
4 files changed, 123 insertions, 0 deletions
diff --git a/results/classifier/gemma3:12b/kvm/131 b/results/classifier/gemma3:12b/kvm/131 new file mode 100644 index 000000000..786c252fc --- /dev/null +++ b/results/classifier/gemma3:12b/kvm/131 @@ -0,0 +1,2 @@ + +QEMU's default msrs handling causes Windows 10 64 bit to crash diff --git a/results/classifier/gemma3:12b/kvm/1312668 b/results/classifier/gemma3:12b/kvm/1312668 new file mode 100644 index 000000000..858cb9efe --- /dev/null +++ b/results/classifier/gemma3:12b/kvm/1312668 @@ -0,0 +1,54 @@ + +x86 cpu nx feature: guest reboots after migrate exec + +Using instruction on +http://www.linux-kvm.org/page/Migration +I save VM state to external file and try load it, but VM starts, shows saved screen and reboots immediatly. + +Cmdline for vm state saving: + +$ sudo ./i386-softmmu/qemu-system-i386 -machine accel=kvm,kernel_irqchip=on -enable-kvm -m 512 -hda image.raw -vga std -net none -M pc -monitor stdio -cpu SandyBridge +(or -cpu "n270" , or "kvm32,+sse2,+pae,+nx") + +Monitor cmd: +(qemu) stop +(qemu) migrate_set_speed 4095m +(qemu) migrate "exec:gzip -c > STATEFILE.gz" +(qemu) q + +Cmdline for vm state loading: + +$ sudo ./i386-softmmu/qemu-system-i386 -machine accel=kvm,kernel_irqchip=on -enable-kvm -m 512 -hda image.raw -vga std -net none -M pc -monitor stdio -cpu SandyBridge -incoming "exec: gzip -c -d STATEFILE.gz" +(or -cpu "n270" , or "kvm32,+sse2,+pae,+nx") + +If I do the same without NX cpu feature (-cpu option "n270,-nx" / "SandyBridge,-nx" / "kvm32,+pae,+sse2") or on qemu-system-x86_64, VM save/load works correctly. + +Log kvm-all.c, DEBUG_KVM=y: + +QEMU 2.0.0 monitor - type 'help' for more information +(qemu) kvm_init_vcpu +...handle_io.../...handle_mmio... +kvm_cpu_exec() +shutdown +kvm_cpu_exec() +interrupt exit requested +io window exit +kvm_cpu_exec() + +Host: + + $ lsb_release -rd + Description: Ubuntu 12.04.4 LTS + Release: 12.04 + + $ uname -a + Linux <username> 3.8.0-38-generic #56~precise1 SMP Tue Apr 22 12:46:44 MSK 2014 x86_64 x86_64 x86_64 GNU/Linux + +Guest: + 1. Ubuntu 12.04 32bit + 2. WIndows 8 32bit + +Qemu: v2.0.0 +commit a9e8aeb3755bccb7b51174adcf4a3fc427e0d147 +Author: Peter Maydell <email address hidden> +Date: Thu Apr 17 13:41:45 2014 +0100 \ No newline at end of file diff --git a/results/classifier/gemma3:12b/kvm/1318091 b/results/classifier/gemma3:12b/kvm/1318091 new file mode 100644 index 000000000..0dee74c9d --- /dev/null +++ b/results/classifier/gemma3:12b/kvm/1318091 @@ -0,0 +1,40 @@ + +Perfctr MSRs not available to Guest OS on AMD Phenom II + +The AMD Phenom(tm) II X4 965 Processor (family 16, model 4, stepping 3) has the 4 architecturally supported perf counters at MSRs. The selectors are c001000-c001003, and the counters are c001004-c001007. I've verified that the MSRs are there and working by manually setting the MSRs with msr-tools to count cycles. + +The processor does not support the extended perfctr or the perfctr_nb. These are in cpuid leaf 8000_0001. Qemu also sees that these cpuid flags are not set, when I try launching with -cpu host,perfctr_core,check. However, this flag is only for the extended perfctr MSRs, which also happen to map the original four counters at c0010200. + +When I run a Guest OS, that OS is unable to use the perf counter registers from c001000-7. rdmsr and wrmsr complete, but the results are always 0. By contrast, a wrmsr to one of the c0010200 registers causes a general protection fault (as it should, since those aren't supported). + +Kernel: 3.14.0-gentoo +Qemu: 2.0.0 (gentoo) and also with 2.0.50 (commit 06b4f00d5) +Qemu command: qemu-system-x86_64 -enable-kvm -cpu host -smp 8 -m 1024 -nographic -monitor /dev/pts/4 mnt/hdd.img +cat /proc/cpuinfo: +processor : 3 +vendor_id : AuthenticAMD +cpu family : 16 +model : 4 +model name : AMD Phenom(tm) II X4 965 Processor +stepping : 3 +cpu MHz : 800.000 +cache size : 512 KB +physical id : 0 +siblings : 4 +core id : 3 +cpu cores : 4 +apicid : 3 +initial apicid : 3 +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 : 6803.79 +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 + +thanks. \ No newline at end of file diff --git a/results/classifier/gemma3:12b/kvm/1318746 b/results/classifier/gemma3:12b/kvm/1318746 new file mode 100644 index 000000000..21d0ce4a7 --- /dev/null +++ b/results/classifier/gemma3:12b/kvm/1318746 @@ -0,0 +1,27 @@ + +qemu Windows 7 BSOD when using hv-time + +When I use hv-time sub option and run CPU-Z or 3DMark (Physics Test) the Windiws 7 guest stops with BSOD (SYSTEM_SERVICE_EXCEPTION). It can be easily reproduced by running CPU-Z. It will fail every second or third time you execute CPU-Z and fail during "PCI detection". If I disable hv-time I can run CPU-Z and 3DMark (Physics Test) without any problems. QEMU was called with the following options: + +/usr/bin/taskset -c 4,5,6,7 /usr/bin/qemu-system-x86_64 \ + -machine q35,accel=kvm,kernel_irqchip=on \ + -enable-kvm \ + -serial none \ + -parallel none \ + -monitor none \ + -vga std \ + -boot order=dc \ + -cpu host,hv-time \ + -smp cores=4,threads=1,sockets=1 \ + -m 8192 \ + -k de \ + -rtc base=localtime \ + -drive file=/srv/kvm/maggie-drive0.img,id=drive0,if=none,cache=none,aio=threads \ + -mon chardev=monitor0 \ + -chardev socket,id=monitor0,path=/tmp/maggie.monitor,nowait,server \ + -netdev tap,id=net0,vhost=on,helper=/usr/lib/qemu/qemu-bridge-helper \ + -device virtio-net-pci,netdev=net0,mac=00:00:00:02:01:01 \ + -device virtio-blk-pci,drive=drive0,ioeventfd=on \ + -device ioh3420,bus=pcie.0,id=pcie0,port=1,chassis=1,multifunction=on + +I've removed the VFIO PCI passthrough line of my GPU to make reproduction easier. In any case it happens in both scenarios so VGA passthrough is not the root cause. It happens with linux-3.15-rc5 and linux-3.14.3 with patch from commit mentioned at https://bugzilla.kernel.org/show_bug.cgi?id=73721#c3 \ No newline at end of file |