[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Xen-devel] [PATCH v2 2/5] x86: Handle the Xen MSRs via the new guest_{rd, wr}msr() infrastructure


  • To: Jan Beulich <JBeulich@xxxxxxxx>
  • From: Andrew Cooper <andrew.cooper3@xxxxxxxxxx>
  • Date: Fri, 7 Sep 2018 17:01:21 +0100
  • Autocrypt: addr=andrew.cooper3@xxxxxxxxxx; prefer-encrypt=mutual; keydata= xsFNBFLhNn8BEADVhE+Hb8i0GV6mihnnr/uiQQdPF8kUoFzCOPXkf7jQ5sLYeJa0cQi6Penp VtiFYznTairnVsN5J+ujSTIb+OlMSJUWV4opS7WVNnxHbFTPYZVQ3erv7NKc2iVizCRZ2Kxn srM1oPXWRic8BIAdYOKOloF2300SL/bIpeD+x7h3w9B/qez7nOin5NzkxgFoaUeIal12pXSR Q354FKFoy6Vh96gc4VRqte3jw8mPuJQpfws+Pb+swvSf/i1q1+1I4jsRQQh2m6OTADHIqg2E ofTYAEh7R5HfPx0EXoEDMdRjOeKn8+vvkAwhviWXTHlG3R1QkbE5M/oywnZ83udJmi+lxjJ5 YhQ5IzomvJ16H0Bq+TLyVLO/VRksp1VR9HxCzItLNCS8PdpYYz5TC204ViycobYU65WMpzWe LFAGn8jSS25XIpqv0Y9k87dLbctKKA14Ifw2kq5OIVu2FuX+3i446JOa2vpCI9GcjCzi3oHV e00bzYiHMIl0FICrNJU0Kjho8pdo0m2uxkn6SYEpogAy9pnatUlO+erL4LqFUO7GXSdBRbw5 gNt25XTLdSFuZtMxkY3tq8MFss5QnjhehCVPEpE6y9ZjI4XB8ad1G4oBHVGK5LMsvg22PfMJ ISWFSHoF/B5+lHkCKWkFxZ0gZn33ju5n6/FOdEx4B8cMJt+cWwARAQABzSlBbmRyZXcgQ29v cGVyIDxhbmRyZXcuY29vcGVyM0BjaXRyaXguY29tPsLBegQTAQgAJAIbAwULCQgHAwUVCgkI CwUWAgMBAAIeAQIXgAUCWKD95wIZAQAKCRBlw/kGpdefoHbdD/9AIoR3k6fKl+RFiFpyAhvO 59ttDFI7nIAnlYngev2XUR3acFElJATHSDO0ju+hqWqAb8kVijXLops0gOfqt3VPZq9cuHlh IMDquatGLzAadfFx2eQYIYT+FYuMoPZy/aTUazmJIDVxP7L383grjIkn+7tAv+qeDfE+txL4 SAm1UHNvmdfgL2/lcmL3xRh7sub3nJilM93RWX1Pe5LBSDXO45uzCGEdst6uSlzYR/MEr+5Z JQQ32JV64zwvf/aKaagSQSQMYNX9JFgfZ3TKWC1KJQbX5ssoX/5hNLqxMcZV3TN7kU8I3kjK mPec9+1nECOjjJSO/h4P0sBZyIUGfguwzhEeGf4sMCuSEM4xjCnwiBwftR17sr0spYcOpqET ZGcAmyYcNjy6CYadNCnfR40vhhWuCfNCBzWnUW0lFoo12wb0YnzoOLjvfD6OL3JjIUJNOmJy RCsJ5IA/Iz33RhSVRmROu+TztwuThClw63g7+hoyewv7BemKyuU6FTVhjjW+XUWmS/FzknSi dAG+insr0746cTPpSkGl3KAXeWDGJzve7/SBBfyznWCMGaf8E2P1oOdIZRxHgWj0zNr1+ooF /PzgLPiCI4OMUttTlEKChgbUTQ+5o0P080JojqfXwbPAyumbaYcQNiH1/xYbJdOFSiBv9rpt TQTBLzDKXok86M7BTQRS4TZ/ARAAkgqudHsp+hd82UVkvgnlqZjzz2vyrYfz7bkPtXaGb9H4 Rfo7mQsEQavEBdWWjbga6eMnDqtu+FC+qeTGYebToxEyp2lKDSoAsvt8w82tIlP/EbmRbDVn 7bhjBlfRcFjVYw8uVDPptT0TV47vpoCVkTwcyb6OltJrvg/QzV9f07DJswuda1JH3/qvYu0p vjPnYvCq4NsqY2XSdAJ02HrdYPFtNyPEntu1n1KK+gJrstjtw7KsZ4ygXYrsm/oCBiVW/OgU g/XIlGErkrxe4vQvJyVwg6YH653YTX5hLLUEL1NS4TCo47RP+wi6y+TnuAL36UtK/uFyEuPy wwrDVcC4cIFhYSfsO0BumEI65yu7a8aHbGfq2lW251UcoU48Z27ZUUZd2Dr6O/n8poQHbaTd 6bJJSjzGGHZVbRP9UQ3lkmkmc0+XCHmj5WhwNNYjgbbmML7y0fsJT5RgvefAIFfHBg7fTY/i kBEimoUsTEQz+N4hbKwo1hULfVxDJStE4sbPhjbsPCrlXf6W9CxSyQ0qmZ2bXsLQYRj2xqd1 bpA+1o1j2N4/au1R/uSiUFjewJdT/LX1EklKDcQwpk06Af/N7VZtSfEJeRV04unbsKVXWZAk uAJyDDKN99ziC0Wz5kcPyVD1HNf8bgaqGDzrv3TfYjwqayRFcMf7xJaL9xXedMcAEQEAAcLB XwQYAQgACQUCUuE2fwIbDAAKCRBlw/kGpdefoG4XEACD1Qf/er8EA7g23HMxYWd3FXHThrVQ HgiGdk5Yh632vjOm9L4sd/GCEACVQKjsu98e8o3ysitFlznEns5EAAXEbITrgKWXDDUWGYxd pnjj2u+GkVdsOAGk0kxczX6s+VRBhpbBI2PWnOsRJgU2n10PZ3mZD4Xu9kU2IXYmuW+e5KCA vTArRUdCrAtIa1k01sPipPPw6dfxx2e5asy21YOytzxuWFfJTGnVxZZSCyLUO83sh6OZhJkk b9rxL9wPmpN/t2IPaEKoAc0FTQZS36wAMOXkBh24PQ9gaLJvfPKpNzGD8XWR5HHF0NLIJhgg 4ZlEXQ2fVp3XrtocHqhu4UZR4koCijgB8sB7Tb0GCpwK+C4UePdFLfhKyRdSXuvY3AHJd4CP 4JzW0Bzq/WXY3XMOzUTYApGQpnUpdOmuQSfpV9MQO+/jo7r6yPbxT7CwRS5dcQPzUiuHLK9i nvjREdh84qycnx0/6dDroYhp0DFv4udxuAvt1h4wGwTPRQZerSm4xaYegEFusyhbZrI0U9tJ B8WrhBLXDiYlyJT6zOV2yZFuW47VrLsjYnHwn27hmxTC/7tvG3euCklmkn9Sl9IAKFu29RSo d5bD8kMSCYsTqtTfT6W4A3qHGvIDta3ptLYpIAOD2sY3GYq2nf3Bbzx81wZK14JdDDHUX2Rs 6+ahAA==
  • Cc: Sergey Dyasli <sergey.dyasli@xxxxxxxxxx>, Paul Durrant <paul.durrant@xxxxxxxxxx>, Wei Liu <wei.liu2@xxxxxxxxxx>, Xen-devel <xen-devel@xxxxxxxxxxxxx>
  • Delivery-date: Fri, 07 Sep 2018 16:01:30 +0000
  • List-id: Xen developer discussion <xen-devel.lists.xenproject.org>
  • Openpgp: preference=signencrypt

On 07/09/18 16:47, Jan Beulich wrote:
>>
>>>>>> On 07.03.18 at 19:58, <andrew.cooper3@xxxxxxxxxx> wrote:
> Wow, resuming a discussion after a full half year.

What can I say?  "Guess roughly when I was told about L1TF?"

As you can see, I did quite literally drop everything and start working
on speculative side-channel fixes.

>>>> --- a/xen/arch/x86/msr.c
>>>> +++ b/xen/arch/x86/msr.c
>>>> @@ -185,6 +185,10 @@ int guest_rdmsr(const struct vcpu *v, uint32_t msr, 
>>>> uint64_t *val)
>>>>          }
>>>>  
>>>>          /* Fallthrough. */
>>>> +    case MSR_XEN_ALT_START ... MSR_XEN_ALT_START + NR_XEN_MSRS - 1:
>>> To account for what I've said on patch 1, perhaps this better would
>>> be
>>>
>>>     case MSR_HYPERVISOR_START + NR_VIRIDIAN_MSRS - 1 ... MSR_XEN_ALT_START 
>>> + NR_XEN_MSRS - 1:
>>>
>>> to produce consistent results regardless of the value of
>>> NR_VIRIDIAN_MSRS (which I suppose is somewhere specified)?
>> This demonstrates perfectly why using names here complicates things.  No
>> expression like this is going to be obvious to read.
>>
>> The raw numbers are the normal way developers think about these ranges,
>> and its trivial to spot that the ranges are adjacent.
> That's a personal thing - to me it's sometimes this way, sometimes
> the other.
>
>> Please compare this suggestion to the guest_cpuid().  The CPUID code is
>> far far clearer to read, and the more I think about this, the more I'm
>> considering switching back to raw numbers.
> That depends very much on how easily the reader can associate back
> the raw numbers. This is more likely to be the case for the rather
> common CPUID leaves or groups of leaves, and perhaps less likely for
> some of the less well known MSRs.

Right, but we are only talking about the course categorisation of the
Viridian block of MSRs, and the Xen block of MSRs.  All of them are of
the form 0x40000xxx.

The end result of this bit of code will look almost identical to the
CPUID side, with the dispatch function names making the context clear.

>
>>>> --- a/xen/include/asm-x86/msr-index.h
>>>> +++ b/xen/include/asm-x86/msr-index.h
>>>> @@ -543,5 +543,7 @@
>>>>  /* Hypervisor leaves in the 0x4xxxxxxx range. */
>>>>  #define MSR_HYPERVISOR_START            0x40000000
>>>>  #define NR_VIRIDIAN_MSRS                0x00000200
>>>> +#define MSR_XEN_ALT_START               0x40000200
>>>> +#define NR_XEN_MSRS                     0x00000100
>>> Where is this count coming from? I don't think it's part of the public
>>> interface, but if there was such an upper bound I think it should be.
>> Its not part of the public ABI, and it should not be, because we don't
>> want to impose an arbitrary limit on how many blocks of 0x100 MSRs the
>> Xen range uses.  Its an artefact of attempting to use numbers.
> "attempting to use numbers"??? How would we get away without
> using some form of number somewhere?

Sorry.  I meant "of attempting to use names".

~Andrew

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxxx
https://lists.xenproject.org/mailman/listinfo/xen-devel

 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.