Re: [Xen-devel] [PATCH v9 06/13] xen: Add ring 3 vmware_port support

[Don Slutz <dslutz@xxxxxxxxxxx>]

On 02/18/15 13:19, Andrew Cooper wrote:
> On 18/02/15 17:03, Don Slutz wrote:
>> On 02/17/15 09:38, Andrew Cooper wrote:
>>> On 16/02/15 23:05, Don Slutz wrote:
>>>> Summary is that VMware treats "in (%dx),%eax" (or "out %eax,(%dx)")
>>>> to port 0x5658 specially.  Note: since many operations return data
>>>> in EAX, "in (%dx),%eax" is the one to use.  The other lengths like
>>>> "in (%dx),%al" will still do things, only AL part of EAX will be
>>>> changed.  For "out %eax,(%dx)" of all lengths, EAX will remain
>>>> unchanged.
>>>>
>>>> This instruction is allowed to be used from ring 3.  To
>>>> support this the vmexit for GP needs to be enabled.  I have not
>>>> fully tested that nested HVM is doing the right thing for this.
>>>>
>>>> The support included is enough to allow VMware tools to install in a
>>>> HVM domU.
>>>>
>>>> Enable no-fault of pio in x86_emulate for VMware port
>>>>
>>>> Also adjust the emulation registers after doing a VMware
>>>> backdoor operation.
>>>>
>>>> Add new routine hvm_emulate_one_gp() to be used by the #GP fault
>>>> handler.
>>>>
>>>> Some of the best info is at:
>>>>
>>>> https://sites.google.com/site/chitchatvmback/backdoor
>>>>
>>>> Signed-off-by: Don Slutz <dslutz@xxxxxxxxxxx>
>>>> ---
>>>> v9:
>>>>    Split #GP handling (or skipping of #GP) code out of previous
>>>>    patch to help with the review process.
>>>>    Switch to x86_emulator to handle #GP
>>>>    I think the hvm_emulate_ops_gp() covers all needed ops.  Not able
>>>> to validate
>>>>    all paths though _hvm_emulate_one().
>>>>
>>>>  xen/arch/x86/hvm/emulate.c             | 62
>>>> ++++++++++++++++++++++++++++++++--
>>>>  xen/arch/x86/hvm/svm/svm.c             | 27 +++++++++++++++
>>>>  xen/arch/x86/hvm/svm/vmcb.c            |  2 ++
>>>>  xen/arch/x86/hvm/vmware/vmport.c       | 11 ++++++
>>>>  xen/arch/x86/hvm/vmx/vmcs.c            |  2 ++
>>>>  xen/arch/x86/hvm/vmx/vmx.c             | 38 +++++++++++++++++++++
>>>>  xen/arch/x86/x86_emulate/x86_emulate.c | 25 +++++++++++---
>>>>  xen/arch/x86/x86_emulate/x86_emulate.h |  8 +++++
>>>>  xen/include/asm-x86/hvm/emulate.h      |  2 ++
>>>>  xen/include/asm-x86/hvm/vmport.h       |  1 +
>>>>  10 files changed, 172 insertions(+), 6 deletions(-)
>>>>
>>>> diff --git a/xen/arch/x86/hvm/emulate.c b/xen/arch/x86/hvm/emulate.c
>>>> index 636c909..a6a6a5c 100644
>>>> --- a/xen/arch/x86/hvm/emulate.c
>>>> +++ b/xen/arch/x86/hvm/emulate.c
>>>> @@ -22,6 +22,7 @@
>>>>  #include <asm/hvm/trace.h>
>>>>  #include <asm/hvm/support.h>
>>>>  #include <asm/hvm/svm/svm.h>
>>>> +#include <asm/hvm/vmport.h>
>>>>
>>>>  static void hvmtrace_io_assist(int is_mmio, ioreq_t *p)
>>>>  {
>>>> @@ -776,6 +777,7 @@ static int hvmemul_read_io_discard(
>>>>      unsigned long *val,
>>>>      struct x86_emulate_ctxt *ctxt)
>>>>  {
>>>> +    ctxt->do_vmport = 0;
>>>
>>> This looks horribly invasive.
>>>
>>> Why are emulation changes needed?  What is wrong with the normal
>>> handling with a registered ioport handler?
>>
>> Because VMware made a bad way to provide a "hyper call".  They decided to
>> allow user access to this.  So when a #GP fault should have been
>> reported, they instead do the "hyper call".
>>
> 
> Urgh - now I remember.
> 
> Right.  In the case that vmport is active, we start intercepting #GP
> faults and emulating access.  That part of the patch looks ok.
> 
> However, the rest is very invasive to the emulation infrastructure.
> 
> Something along the lines of:
> 
> diff --git a/xen/arch/x86/x86_emulate/x86_emulate.c
> b/xen/arch/x86/x86_emulate/x86_emulate.c
> index 5e9e040..dd40d6a 100644
> --- a/xen/arch/x86/x86_emulate/x86_emulate.c
> +++ b/xen/arch/x86/x86_emulate/x86_emulate.c
> @@ -3394,7 +3394,8 @@ static int inject_swint(enum x86_swint_type type,
>                               ? insn_fetch_type(uint8_t)
>                               : (uint16_t)_regs.edx);
>          op_bytes = !(b & 1) ? 1 : (op_bytes == 8) ? 4 : op_bytes;
> -        if ( (rc = ioport_access_check(port, op_bytes, ctxt, ops)) != 0 )
> +        if ( ((rc = ioport_access_check(port, op_bytes, ctxt, ops)) !=
> 0) ||
> +             (ops->vmport_check && ((rc = ops->vmport_check(port,
> ctxt)) != 0)) )
>              goto done;
>          if ( b & 2 )
>          {
> 
> would be far less invasive and AFAICT, replace the entire rest of your
> patch.
> 
> In this case, if ioport_access_check() succeeds, or if it fails and
> vmport_check subsequently succeeds, the standard ioport dispatch will
> run, and hit vmport_ioport().
> 

Yes, but since vmport_ioport changes guest_cpu_user_regs() which
are different then _regs and get modified by:

    *ctxt->regs = _regs;
 done:
    return rc;

which will drop all changes to other registers by vmport_ioport.  This
is why I added the flag do_vmport. Which must be set in all case where
vmport_ioport is called via x86_emulate.

So I could call on it 1st.  This would make the change smaller.

   -Don Slutz



> ~Andrew
> 

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
http://lists.xen.org/xen-devel