On 05/01/16 12:05, Tamas K Lengyel wrote:
>
>
> On Tue, Jan 5, 2016 at 12:56 PM, Andrew Cooper
> <andrew.coop...@citrix.com <mailto:andrew.coop...@citrix.com>> wrote:
>
> On 05/01/16 11:49, Tamas K Lengyel wrote:
>> Hi all,
>> I've been stress-testing the built-in emulator using the vm_event
>> response VM_EVENT_FLAG_EMULATE feature. In the test I've turned
>> all pages non-readable by default and all trapped instructions to
>> be emulated. My test code can be found at
>> https://github.com/tklengyel/xen/compare/read_emul?expand=1.
>>
>> The following crash is reproducible and has been verified by
>> Razvan as well.
>>
>> (XEN) p2m.c:1726:d1v0 calling mem_access_emulate_one, kind 0
>> (XEN) Failed vm entry (exit reason 0x80000021) caused by invalid
>> guest state (0).
>> (XEN) ************* VMCS Area **************
>> (XEN) *** Guest State ***
>> (XEN) CR0: actual=0x000000008001003b, shadow=0x000000008001003b,
>> gh_mask=ffffffffffffffff
>> (XEN) CR4: actual=0x00000000000426f9, shadow=0x00000000000406f9,
>> gh_mask=ffffffffffffffff
>> (XEN) CR3 = 0x0000000000185000
>> (XEN) PDPTE0 = 0x0000000000186001 PDPTE1 = 0x0000000000187001
>> (XEN) PDPTE2 = 0x0000000000188001 PDPTE3 = 0x0000000000189001
>> (XEN) RSP = 0x000000008276dc28 (0x000000008276dc28) RIP =
>> 0x00000000826bce1c (0x00000000826bce1c)
>> (XEN) RFLAGS=0x00000002 (0x00000002) DR7 = 0x0000000000000400
>> (XEN) Sysenter RSP=000000008078b000 CS:RIP=0008:00000000826830c0
>> (XEN) sel attr limit base
>> (XEN) CS: 0008 0c09b ffffffff 0000000000000000
>> (XEN) DS: 0023 0c0f3 ffffffff 0000000000000000
>> (XEN) SS: 0010 0c093 ffffffff 0000000000000000
>> (XEN) ES: 0023 0c0f3 ffffffff 0000000000000000
>> (XEN) FS: 0030 04093 00003748 0000000082770c00
>> (XEN) GS: 0000 1c000 ffffffff 0000000000000000
>> (XEN) GDTR: 000003ff 0000000080b95000
>> (XEN) LDTR: 0000 1c000 ffffffff 0000000000000000
>> (XEN) IDTR: 000007ff 0000000080b95400
>> (XEN) TR: 0028 0008b 000020ab 00000000801da000
>> (XEN) EFER = 0x0000000000000000 PAT = 0x0007010600070106
>> (XEN) PreemptionTimer = 0x00000000 SM Base = 0x00000000
>> (XEN) DebugCtl = 0x0000000000000000 DebugExceptions =
>> 0x0000000000000000
>> (XEN) Interruptibility = 00000000 ActivityState = 00000000
>> (XEN) *** Host State ***
>> (XEN) RIP = 0xffff82d0802075c0 (vmx_asm_vmexit_handler) RSP =
>> 0xffff830430d97f90
>> (XEN) CS=e008 SS=0000 DS=0000 ES=0000 FS=0000 GS=0000 TR=e040
>> (XEN) FSBase=0000000000000000 GSBase=0000000000000000
>> TRBase=ffff830430d9bc00
>> (XEN) GDTBase=ffff830430d8c000 IDTBase=ffff830430d98000
>> (XEN) CR0=000000008005003b CR3=00000004136d0000 CR4=00000000000426e0
>> (XEN) Sysenter RSP=ffff830430d97fc0 CS:RIP=e008:ffff82d08024db30
>> (XEN) EFER = 0x0000000000000000 PAT = 0x0000050100070406
>> (XEN) *** Control State ***
>> (XEN) PinBased=0000003f CPUBased=b6a075fa SecondaryExec=000000eb
>> (XEN) EntryControls=000051ff ExitControls=000fefff
>> (XEN) ExceptionBitmap=00060042 PFECmask=00000000 PFECmatch=00000000
>> (XEN) VMEntry: intr_info=800000d1 errcode=00000000 ilen=00000000
>> (XEN) VMExit: intr_info=00000000 errcode=00000000 ilen=00000003
>> (XEN) reason=80000021 qualification=0000000000000000
>> (XEN) IDTVectoring: info=800000d1 errcode=00000000
>> (XEN) TSC Offset = 0x0000004ed9c86354
>> (XEN) TPR Threshold = 0x00 PostedIntrVec = 0x00
>> (XEN) EPT pointer = 0x000000041124e01e EPTP index = 0x0000
>> (XEN) Virtual processor ID = 0x0011 VMfunc controls =
>> 0000000000000000
>> (XEN) **************************************
>> (XEN) domain_crash called from vmx.c:2761
>>
>> Any tips on how to further debug this issue?
>
> Do you have a log of the instructions emulated?
>
>
> I don't. Is there an easy way to get that beside manually sprinkling
> debug messages around in the emulator?
Not trivially, sadly.
>
>
>
> Has the emulator by any chance just emulated setting CR4.PAE?
>
>
> Possibly but I don't think so as the guest has already been fully
> booted so I would not expect it to touch that.
At a guess, I think the fault is an emulated 'mov %reg, %cr3' while in
32bit PAE mode. The PDPTE{0..3} values look wonky.
I encountered a similar crash with the xen test framework in HAP mode
with a bad %cr3 update. The VMM is expected to emulate updates to
PDPTE{0..3} if writes to %cr3 are trapped. See vmx_update_guest_cr()
and vmx_load_pdptrs().
~Andrew
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
http://lists.xen.org/xen-devel
