[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Xen-devel] [PATCH 08/11] x86/hvm: RFC - PROBABLY BROKEN - Defer all debugging/monitor actions to {svm, vmx}_inject_event()
- To: Jan Beulich <JBeulich@xxxxxxxx>
- From: Andrew Cooper <andrew.cooper3@xxxxxxxxxx>
- Date: Fri, 8 Jun 2018 14:13:00 +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: Kevin Tian <kevin.tian@xxxxxxxxx>, tamas@xxxxxxxxxxxxx, Wei Liu <wei.liu2@xxxxxxxxxx>, Suravee Suthikulpanit <suravee.suthikulpanit@xxxxxxx>, Razvan Cojocaru <rcojocaru@xxxxxxxxxxxxxxx>, Xen-devel <xen-devel@xxxxxxxxxxxxx>, Jun Nakajima <jun.nakajima@xxxxxxxxx>, Boris Ostrovsky <boris.ostrovsky@xxxxxxxxxx>, brian.woods@xxxxxxx, Roger Pau Monne <roger.pau@xxxxxxxxxx>
- Delivery-date: Fri, 08 Jun 2018 13:13:06 +0000
- List-id: Xen developer discussion <xen-devel.lists.xenproject.org>
- Openpgp: preference=signencrypt
On 08/06/18 14:00, Jan Beulich wrote:
>>>> On 04.06.18 at 15:59, <andrew.cooper3@xxxxxxxxxx> wrote:
>> --- a/xen/arch/x86/hvm/svm/svm.c
>> +++ b/xen/arch/x86/hvm/svm/svm.c
>> @@ -1437,19 +1437,49 @@ static void svm_inject_event(const struct x86_event
>> *event)
>> switch ( _event.vector | -(_event.type == X86_EVENTTYPE_SW_INTERRUPT) )
>> {
>> case TRAP_debug:
>> - if ( regs->eflags & X86_EFLAGS_TF )
>> - {
>> - __restore_debug_registers(vmcb, curr);
>> - vmcb_set_dr6(vmcb, vmcb_get_dr6(vmcb) | DR_STEP);
>> - }
>> + /*
>> + * On AMD hardware, a #DB exception:
>> + * 1) Merges new status bits into %dr6
>> + * 2) Clears %dr7.gd and MSR_DEBUGCTL.{LBR,BTF}
>> + *
>> + * Item 1 is done by hardware before a #DB intercepted vmexit, but
>> we
>> + * may end up here from emulation so have to repeat it ourselves.
>> + * Item 2 is done by hardware when injecting a #DB exception.
>> + */
>> + __restore_debug_registers(vmcb, curr);
>> + vmcb_set_dr6(vmcb, vmcb_get_dr6(vmcb) | event->pending_dbg);
>> +
>> /* fall through */
>> case TRAP_int3:
>> if ( curr->domain->debugger_attached )
>> {
>> /* Debug/Int3: Trap to debugger. */
>> + if ( _event.vector == TRAP_int3 )
>> + {
>> + /* N.B. Can't use __update_guest_eip() for risk of
>> recusion. */
>> + regs->rip += _event.insn_len;
> Not all callers provide a non-zero insn length. Is that a potential
> problem here (and in the equivalent VMX code)?
TRAP_int3 is strictly a software exception (and/or software interrupt if
started via `int $n`), so should always have a nonzero length.
We should be able to assert this property, and I was considering some
extra checks in the {pv,hvm}_inject_event() logic, although it would
require XEN_DMOP_inject_event gaining more sanity checking than it
currently has (which is probably a good thing).
I'll fold something suitable into v2, which is already quite a bit
longer to help with monitor problem identified here.
>> @@ -2775,67 +2805,46 @@ void svm_vmexit_handler(struct cpu_user_regs *regs)
>>
>> case VMEXIT_ICEBP:
>> case VMEXIT_EXCEPTION_DB:
>> - if ( !v->domain->debugger_attached )
>> + case VMEXIT_EXCEPTION_BP:
>> + {
>> + unsigned int vec, type, len, extra;
>> +
>> + switch ( exit_reason )
>> {
>> - int rc;
>> - unsigned int trap_type;
>> + case VMEXIT_ICEBP:
> I don't understand this structuring of the code: The outer switch()
> has the exact same control expression, and there's no fall through
> - why the nesting? Move the variable declarations to the beginning
> of the outer switch() and drop the inner one? You may not even
> need all of the variables if you replicated the little bit of code
> currently shared by the three (immediately after the inner switch()).
Good question - I think this layout is a side effect of me changing my
mind several times during its development.
~Andrew
_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxxx
https://lists.xenproject.org/mailman/listinfo/xen-devel
|