NOVA Microhypervisor v7-23e7882 (x86_32): Aug 29 2016 10:39:41 [gcc 4.9.2]
[ 0] CORE:0:0:0 f:4:9:4 [0] Intel(R) Celeron(R) CPU 2.66GHz [ 0] Killed EC:0xc23ca160 SC:0xc23cb100 V:0x6 CS:0x8 EIP:0xc000002c CR2:0x0 ERR:0x0 (EXC) paul@...445...:~$
On Sun, 18 Sep 2016 05:32:25 -0400 Paul Dufresne (PD) wrote:
PD> NOVA Microhypervisor v7-23e7882 (x86_32): Aug 29 2016 10:39:41 [gcc 4.9.2] PD> PD> [ 0] CORE:0:0:0 f:4:9:4 [0] Intel(R) Celeron(R) CPU 2.66GHz PD> [ 0] Killed EC:0xc23ca160 SC:0xc23cb100 V:0x6 CS:0x8 EIP:0xc000002c PD> CR2:0x0 ERR:0x0 (EXC)
This is an invalid opcode exception and it means your CPU is rather old and does not understand the instruction at IP 0xc000002c. You can disassemble the hypervisor binary using objdump and check which instruction that is.
- Udo
Hello,
On Sun, Sep 18, 2016 at 01:20:16PM +0200, Udo Steinberg wrote:
On Sun, 18 Sep 2016 05:32:25 -0400 Paul Dufresne (PD) wrote:
PD> NOVA Microhypervisor v7-23e7882 (x86_32): Aug 29 2016 10:39:41 [gcc 4.9.2] PD> PD> [ 0] CORE:0:0:0 f:4:9:4 [0] Intel(R) Celeron(R) CPU 2.66GHz PD> [ 0] Killed EC:0xc23ca160 SC:0xc23cb100 V:0x6 CS:0x8 EIP:0xc000002c PD> CR2:0x0 ERR:0x0 (EXC)
This is an invalid opcode exception and it means your CPU is rather old and does not understand the instruction at IP 0xc000002c. You can disassemble the hypervisor binary using objdump and check which instruction that is.
With a recent build of NOVA/32 I get
objdump -ldSC build/nova_x86_32/kernel/hypervisor | grep > "c000002c:"
c000002c: 0f 35 sysexit
But I had expected those CPUs to support sysenter/sysexit.
Greets
Oh well, it seems I get the error only under Virtuabox (which is version 5.1.4). I just burn a CD with that nova.iso, and after showing fast invalid hints messages, goes to the not handled by monitor video mode. This is the same computer I tried under VirtualBox (Debian 4.6.0-1-amd64, Celeron D, 2Go).
A bit to my own surprise, VirtualBox does not give much option for the simulated CPU.
The subject sounds like Paul is running NOVA in a virtual machine. So maybe the VM does not support those instructions.
- Udo
On 18 Sep 2016, at 17:10, Christian Helmuth <christian.helmuth@...1...> wrote:
Hello,
On Sun, Sep 18, 2016 at 01:20:16PM +0200, Udo Steinberg wrote: On Sun, 18 Sep 2016 05:32:25 -0400 Paul Dufresne (PD) wrote:
PD> NOVA Microhypervisor v7-23e7882 (x86_32): Aug 29 2016 10:39:41 [gcc 4.9.2] PD> PD> [ 0] CORE:0:0:0 f:4:9:4 [0] Intel(R) Celeron(R) CPU 2.66GHz PD> [ 0] Killed EC:0xc23ca160 SC:0xc23cb100 V:0x6 CS:0x8 EIP:0xc000002c PD> CR2:0x0 ERR:0x0 (EXC)
This is an invalid opcode exception and it means your CPU is rather old and does not understand the instruction at IP 0xc000002c. You can disassemble the hypervisor binary using objdump and check which instruction that is.
With a recent build of NOVA/32 I get
objdump -ldSC build/nova_x86_32/kernel/hypervisor | grep > "c000002c:"
c000002c: 0f 35 sysexit
But I had expected those CPUs to support sysenter/sysexit.
Greets
Christian Helmuth Genode Labs
http://www.genode-labs.com/ · http://genode.org/ https://twitter.com/GenodeLabs · /ˈdʒiː.nəʊd/
Genode Labs GmbH · Amtsgericht Dresden · HRB 28424 · Sitz Dresden Geschäftsführer: Dr.-Ing. Norman Feske, Christian Helmuth
genode-main mailing list genode-main@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/genode-main
Well,I never got understanding clearly these SYSENTER/SYSEXIT instructions. Trying to google it a bit, I find this that seems to make some kind of sense (but not too much honestly). Basically, someone in similar situation was told it need to save MSR (most particularly IA32_STAR MSR), for using SYSENTER/SYSEXIT correctly: https://software.intel.com/en-us/forums/virtualization-software-development/...
Also about SysExit and link to MSR: http://wiki.osdev.org/SYSENTER
Ok, first this talk about saving MSR is for AMD instructions (SYSENTER/SYSRET), not for Intel SYSEXIT.
And indeed, the CPU simulated, a pentium 4, does not support SysEnter/SysExit.
I know because when I go to Machine menu, show journal I can see: ...
00:00:02.740255 VM: fHMEnabled=true (configured) fRecompileUser=false fRecompileSupervisor=false 00:00:02.740262 VM: fRawRing1Enabled=false CSAM=true PATM=true 00:00:02.740694 HM: HMR3Init: Falling back to raw-mode: VT-x is not available 00:00:02.740865 MM: cbHyperHeap=0x140000 (1310720) 00:00:02.746936 CPUM: fXStateHostMask=0x0; initial: 0x0; host XCR0=0x0 00:00:02.774577 CPUM: Matched host CPU INTEL 0xf/0x4/0x9 Intel_NB_Prescott2M with CPU DB entry 'Intel Pentium 4 3.00GHz' (INTEL 0xf/0x4/0x3 Intel_NB_Prescott2M) 00:00:02.774696 CPUM: MSR fudge: 0x000001db P6_LAST_BRANCH_FROM_IP ...
00:00:03.265885 Features 00:00:03.265886 Mnemonic - Description = guest (host) 00:00:03.265889 FPU - x87 FPU on Chip = 1 (1) 00:00:03.265893 VME - Virtual 8086 Mode Enhancements = 1 (1) 00:00:03.265896 DE - Debugging extensions = 1 (1) 00:00:03.265899 PSE - Page Size Extension = 1 (1) 00:00:03.265902 TSC - Time Stamp Counter = 1 (1) 00:00:03.265906 MSR - Model Specific Registers = 1 (1) 00:00:03.265909 PAE - Physical Address Extension = 1 (1) 00:00:03.265912 MCE - Machine Check Exception = 1 (1) 00:00:03.265915 CX8 - CMPXCHG8B instruction = 1 (1) 00:00:03.265918 APIC - APIC On-Chip = 1 (1) *** 00:00:03.265921 SEP - SYSENTER and SYSEXIT Present = 0 (1) *** 00:00:03.265924 MTRR - Memory Type Range Registers = 1 (1) 00:00:03.265927 PGE - PTE Global Bit = 1 (1) 00:00:03.265930 MCA - Machine Check Architecture = 1 (1) 00:00:03.265933 CMOV - Conditional Move instructions = 1 (1) 00:00:03.265936 PAT - Page Attribute Table = 1 (1) 00:00:03.265938 PSE-36 - 36-bit Page Size Extension = 1 (1) 00:00:03.265941 PSN - Processor Serial Number = 0 (0) 00:00:03.265944 CLFSH - CLFLUSH instruction = 1 (1) 00:00:03.265947 DS - Debug Store = 0 (1) 00:00:03.265951 ACPI - Thermal Mon. & Soft. Clock Ctrl. = 0 (1) 00:00:03.265954 MMX - Intel MMX Technology = 1 (1) 00:00:03.265957 FXSR - FXSAVE and FXRSTOR instructions = 1 (1) 00:00:03.265959 SSE - SSE support = 1 (1) 00:00:03.265963 SSE2 - SSE2 support = 1 (1) 00:00:03.265966 SS - Self Snoop = 0 (1) 00:00:03.265969 HTT - Hyper-Threading Technology = 0 (1) 00:00:03.265972 TM - Therm. Monitor = 0 (1) 00:00:03.265976 PBE - Pending Break Enabled = 0 (1) 00:00:03.265979 SSE3 - SSE3 support = 1 (1) 00:00:03.265982 PCLMUL - PCLMULQDQ support (for AES-GCM) = 0 (0) 00:00:03.265985 DTES64 - DS Area 64-bit Layout = 0 (1) 00:00:03.265988 MONITOR - MONITOR/MWAIT instructions = 1 (1) 00:00:03.265991 CPL-DS - CPL Qualified Debug Store = 0 (1) 00:00:03.265994 VMX - Virtual Machine Extensions = 0 (0) 00:00:03.265997 SMX - Safer Mode Extensions = 0 (0) 00:00:03.266000 EST - Enhanced SpeedStep Technology = 0 (0) 00:00:03.266002 TM2 - Terminal Monitor 2 = 0 (1) 00:00:03.266006 SSSE3 - Supplemental Streaming SIMD Extensions 3 = 0 (0) 00:00:03.266008 CNTX-ID - L1 Context ID = 0 (1) 00:00:03.266011 SDBG - Silicon Debug interface = 0 (0) 00:00:03.266014 FMA - Fused Multiply Add extensions = 0 (0) 00:00:03.266017 CX16 - CMPXCHG16B instruction = 0 (1) 00:00:03.266020 TPRUPDATE - xTPR Update Control = 0 (1) 00:00:03.266023 PDCM - Perf/Debug Capability MSR = 0 (0) 00:00:03.266026 PCID - Process Context Identifiers = 0 (0) 00:00:03.266029 DCA - Direct Cache Access = 0 (0) 00:00:03.266032 SSE4_1 - SSE4_1 support = 0 (0) 00:00:03.266035 SSE4_2 - SSE4_2 support = 0 (0) 00:00:03.266038 X2APIC - x2APIC support = 0 (0) 00:00:03.266041 MOVBE - MOVBE instruction = 0 (0) 00:00:03.266045 POPCNT - POPCNT instruction = 0 (0) 00:00:03.266048 TSCDEADL - Time Stamp Counter Deadline = 0 (0) 00:00:03.266050 AES - AES instructions = 0 (0) 00:00:03.266054 XSAVE - XSAVE instruction = 0 (0) 00:00:03.266057 OSXSAVE - OSXSAVE instruction = 0 (0) 00:00:03.266060 AVX - AVX support = 0 (0) 00:00:03.266063 F16C - 16-bit floating point conversion instructions = 0 (0) 00:00:03.266065 RDRAND - RDRAND instruction = 0 (0) 00:00:03.266068 HVP - Hypervisor Present (we're a guest) = 0 (0)
So my host CPU does support SysEnter/SysExit, but not my Guest CPU. Not sure if I can fix that... easily.
I got my hands on a Core2 Duo computer, that allows me to run this in VirtualBox without problems.
On real hardware, it had hang on boot line however... but since that computer still have a com port I might be able to figure out why when I will have enough time.
-
Christian Helmuth -
Paul Dufresne -
Udo Steinberg