xen-devel.lists.xenproject.org archive mirror
 help / color / mirror / Atom feed
From: Andrew Cooper <andrew.cooper3@citrix.com>
To: Xen-devel <xen-devel@lists.xen.org>
Cc: Andrew Cooper <andrew.cooper3@citrix.com>,
	Paul Durrant <paul.durrant@citrix.com>, Tim Deegan <tim@xen.org>,
	Jan Beulich <JBeulich@suse.com>
Subject: [PATCH v3 11/24] x86/emul: Implement singlestep as a retire flag
Date: Wed, 30 Nov 2016 13:50:28 +0000	[thread overview]
Message-ID: <1480513841-7565-12-git-send-email-andrew.cooper3@citrix.com> (raw)
In-Reply-To: <1480513841-7565-1-git-send-email-andrew.cooper3@citrix.com>

The behaviour of singlestep is to raise #DB after the instruction has been
completed, but implementing it with inject_hw_exception() causes x86_emulate()
to return X86EMUL_EXCEPTION, despite succesfully completing execution of the
instruction, including register writeback.

Instead, use a retire flag to indicate singlestep, which causes x86_emulate()
to return X86EMUL_OKAY.

Update all callers of x86_emulate() to use the new retire flag.  This fixes
the behaviour of singlestep for shadow pagetable updates and mmcfg/mmio_ro
intercepts, which previously discarded the exception.

With this change, all uses of X86EMUL_EXCEPTION from x86_emulate() are
believed to have strictly fault semantics.

Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
---
CC: Jan Beulich <JBeulich@suse.com>
CC: Tim Deegan <tim@xen.org>
CC: Paul Durrant <paul.durrant@citrix.com>

v3:
 * New
---
 xen/arch/x86/hvm/emulate.c             |  3 +++
 xen/arch/x86/mm.c                      | 11 ++++++++++-
 xen/arch/x86/mm/shadow/multi.c         | 21 ++++++++++++++++++++-
 xen/arch/x86/x86_emulate/x86_emulate.c |  9 ++++-----
 xen/arch/x86/x86_emulate/x86_emulate.h |  6 ++++++
 5 files changed, 43 insertions(+), 7 deletions(-)

diff --git a/xen/arch/x86/hvm/emulate.c b/xen/arch/x86/hvm/emulate.c
index fe62500..91c79fa 100644
--- a/xen/arch/x86/hvm/emulate.c
+++ b/xen/arch/x86/hvm/emulate.c
@@ -1788,6 +1788,9 @@ static int _hvm_emulate_one(struct hvm_emulate_ctxt *hvmemul_ctxt,
     if ( rc != X86EMUL_OKAY )
         return rc;
 
+    if ( hvmemul_ctxt->ctxt.retire.singlestep )
+        hvm_inject_hw_exception(TRAP_debug, X86_EVENT_NO_EC);
+
     new_intr_shadow = hvmemul_ctxt->intr_shadow;
 
     /* MOV-SS instruction toggles MOV-SS shadow, else we just clear it. */
diff --git a/xen/arch/x86/mm.c b/xen/arch/x86/mm.c
index b7c7122..231c7bf 100644
--- a/xen/arch/x86/mm.c
+++ b/xen/arch/x86/mm.c
@@ -5382,6 +5382,9 @@ int ptwr_do_page_fault(struct vcpu *v, unsigned long addr,
     if ( rc == X86EMUL_UNHANDLEABLE )
         goto bail;
 
+    if ( ptwr_ctxt.ctxt.retire.singlestep )
+        pv_inject_hw_exception(TRAP_debug, X86_EVENT_NO_EC);
+
     perfc_incr(ptwr_emulations);
     return EXCRET_fault_fixed;
 
@@ -5503,7 +5506,13 @@ int mmio_ro_do_page_fault(struct vcpu *v, unsigned long addr,
     else
         rc = x86_emulate(&ctxt, &mmio_ro_emulate_ops);
 
-    return rc != X86EMUL_UNHANDLEABLE ? EXCRET_fault_fixed : 0;
+    if ( rc == X86EMUL_UNHANDLEABLE )
+        return 0;
+
+    if ( ctxt.retire.singlestep )
+        pv_inject_hw_exception(TRAP_debug, X86_EVENT_NO_EC);
+
+    return EXCRET_fault_fixed;
 }
 
 void *alloc_xen_pagetable(void)
diff --git a/xen/arch/x86/mm/shadow/multi.c b/xen/arch/x86/mm/shadow/multi.c
index 9ee48a8..ddfb815 100644
--- a/xen/arch/x86/mm/shadow/multi.c
+++ b/xen/arch/x86/mm/shadow/multi.c
@@ -3422,6 +3422,16 @@ static int sh_page_fault(struct vcpu *v,
         v->arch.paging.last_write_emul_ok = 0;
 #endif
 
+    if ( emul_ctxt.ctxt.retire.singlestep )
+    {
+        if ( is_hvm_vcpu(v) )
+            hvm_inject_hw_exception(TRAP_debug, X86_EVENT_NO_EC);
+        else
+            pv_inject_hw_exception(TRAP_debug, X86_EVENT_NO_EC);
+
+        goto emulate_done;
+    }
+
 #if GUEST_PAGING_LEVELS == 3 /* PAE guest */
     if ( r == X86EMUL_OKAY ) {
         int i, emulation_count=0;
@@ -3433,7 +3443,7 @@ static int sh_page_fault(struct vcpu *v,
             shadow_continue_emulation(&emul_ctxt, regs);
             v->arch.paging.last_write_was_pt = 0;
             r = x86_emulate(&emul_ctxt.ctxt, emul_ops);
-            if ( r == X86EMUL_OKAY )
+            if ( r == X86EMUL_OKAY && !emul_ctxt.ctxt.retire.raw )
             {
                 emulation_count++;
                 if ( v->arch.paging.last_write_was_pt )
@@ -3449,6 +3459,15 @@ static int sh_page_fault(struct vcpu *v,
             {
                 perfc_incr(shadow_em_ex_fail);
                 TRACE_SHADOW_PATH_FLAG(TRCE_SFLAG_EMULATION_LAST_FAILED);
+
+                if ( emul_ctxt.ctxt.retire.singlestep )
+                {
+                    if ( is_hvm_vcpu(v) )
+                        hvm_inject_hw_exception(TRAP_debug, X86_EVENT_NO_EC);
+                    else
+                        pv_inject_hw_exception(TRAP_debug, X86_EVENT_NO_EC);
+                }
+
                 break; /* Don't emulate again if we failed! */
             }
         }
diff --git a/xen/arch/x86/x86_emulate/x86_emulate.c b/xen/arch/x86/x86_emulate/x86_emulate.c
index 8a1f1f5..0af532e 100644
--- a/xen/arch/x86/x86_emulate/x86_emulate.c
+++ b/xen/arch/x86/x86_emulate/x86_emulate.c
@@ -2417,7 +2417,6 @@ x86_emulate(
     struct x86_emulate_state state;
     int rc;
     uint8_t b, d;
-    bool tf = ctxt->regs->eflags & EFLG_TF;
     struct operand src = { .reg = PTR_POISON };
     struct operand dst = { .reg = PTR_POISON };
     enum x86_swint_type swint_type;
@@ -5415,11 +5414,11 @@ x86_emulate(
     if ( !mode_64bit() )
         _regs.eip = (uint32_t)_regs.eip;
 
-    *ctxt->regs = _regs;
+    /* Was singestepping active at the start of this instruction? */
+    if ( (rc == X86EMUL_OKAY) && (ctxt->regs->eflags & EFLG_TF) )
+        ctxt->retire.singlestep = true;
 
-    /* Inject #DB if single-step tracing was enabled at instruction start. */
-    if ( tf && (rc == X86EMUL_OKAY) && ops->inject_hw_exception )
-        rc = ops->inject_hw_exception(EXC_DB, -1, ctxt) ? : X86EMUL_EXCEPTION;
+    *ctxt->regs = _regs;
 
  done:
     _put_fpu();
diff --git a/xen/arch/x86/x86_emulate/x86_emulate.h b/xen/arch/x86/x86_emulate/x86_emulate.h
index fc28976..da8924b 100644
--- a/xen/arch/x86/x86_emulate/x86_emulate.h
+++ b/xen/arch/x86/x86_emulate/x86_emulate.h
@@ -483,6 +483,7 @@ struct x86_emulate_ctxt
             bool hlt:1;          /* Instruction HLTed. */
             bool mov_ss:1;       /* Instruction sets MOV-SS irq shadow. */
             bool sti:1;          /* Instruction sets STI irq shadow. */
+            bool singlestep:1;   /* Singlestepping was active. */
         };
     } retire;
 };
@@ -572,12 +573,17 @@ static inline int x86_emulate_wrapper(
     struct x86_emulate_ctxt *ctxt,
     const struct x86_emulate_ops *ops)
 {
+    unsigned long orig_eip = ctxt->regs->eip;
     int rc = x86_emulate(ctxt, ops);
 
     /* Retire flags should only be set for successful instruction emulation. */
     if ( rc != X86EMUL_OKAY )
         ASSERT(ctxt->retire.raw == 0);
 
+    /* All cases returning X86EMUL_EXCEPTION should have fault semantics. */
+    if ( rc == X86EMUL_EXCEPTION )
+        ASSERT(ctxt->regs->eip == orig_eip);
+
     return rc;
 }
 
-- 
2.1.4


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

  parent reply	other threads:[~2016-11-30 13:50 UTC|newest]

Thread overview: 59+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2016-11-30 13:50 [PATCH for-4.9 v3 00/24] XSA-191 followup Andrew Cooper
2016-11-30 13:50 ` [PATCH v3 01/24] x86/shadow: Fix #PFs from emulated writes crossing a page boundary Andrew Cooper
2016-11-30 13:50 ` [PATCH v3 02/24] x86/emul: Drop X86EMUL_CMPXCHG_FAILED Andrew Cooper
2016-11-30 13:50 ` [PATCH v3 03/24] x86/emul: Simplfy emulation state setup Andrew Cooper
2016-12-08  6:34   ` George Dunlap
2016-11-30 13:50 ` [PATCH v3 04/24] x86/emul: Rename hvm_trap to x86_event and move it into the emulation infrastructure Andrew Cooper
2016-11-30 13:50 ` [PATCH v3 05/24] x86/emul: Rename HVM_DELIVER_NO_ERROR_CODE to X86_EVENT_NO_EC Andrew Cooper
2016-11-30 13:50 ` [PATCH v3 06/24] x86/pv: Implement pv_inject_{event, page_fault, hw_exception}() Andrew Cooper
2016-12-01 10:06   ` Jan Beulich
2016-11-30 13:50 ` [PATCH v3 07/24] x86/emul: Clean up the naming of the retire union Andrew Cooper
2016-11-30 13:58   ` Paul Durrant
2016-11-30 14:02     ` Andrew Cooper
2016-11-30 14:05       ` Paul Durrant
2016-11-30 16:43         ` Jan Beulich
2016-12-01 10:08   ` Jan Beulich
2016-11-30 13:50 ` [PATCH v3 08/24] x86/emul: Correct the behaviour of pop %ss and interrupt shadowing Andrew Cooper
2016-12-01 10:18   ` Jan Beulich
2016-12-01 10:51     ` Andrew Cooper
2016-12-01 11:19       ` Jan Beulich
2016-11-30 13:50 ` [PATCH v3 09/24] x86/emul: Provide a wrapper to x86_emulate() to ASSERT() certain behaviour Andrew Cooper
2016-12-01 10:40   ` Jan Beulich
2016-12-01 10:58     ` Andrew Cooper
2016-12-01 11:21       ` Jan Beulich
2016-11-30 13:50 ` [PATCH v3 10/24] x86/emul: Always use fault semantics for software events Andrew Cooper
2016-11-30 17:55   ` Boris Ostrovsky
2016-12-01 10:53   ` Jan Beulich
2016-12-01 11:15     ` Andrew Cooper
2016-12-01 11:23       ` Jan Beulich
2016-11-30 13:50 ` Andrew Cooper [this message]
2016-11-30 14:28   ` [PATCH v3 11/24] x86/emul: Implement singlestep as a retire flag Paul Durrant
2016-12-01 11:16   ` Jan Beulich
2016-12-01 11:23     ` Andrew Cooper
2016-12-01 11:33       ` Tim Deegan
2016-12-01 12:05       ` Jan Beulich
2016-11-30 13:50 ` [PATCH v3 12/24] x86/emul: Remove opencoded exception generation Andrew Cooper
2016-11-30 13:50 ` [PATCH v3 13/24] x86/emul: Rework emulator event injection Andrew Cooper
2016-11-30 14:26   ` Paul Durrant
2016-12-01 11:35   ` Tim Deegan
2016-12-01 12:31   ` Jan Beulich
2016-11-30 13:50 ` [PATCH v3 14/24] x86/vmx: Use hvm_{get, set}_segment_register() rather than vmx_{get, set}_segment_register() Andrew Cooper
2016-11-30 13:50 ` [PATCH v3 15/24] x86/hvm: Reposition the modification of raw segment data from the VMCB/VMCS Andrew Cooper
2016-11-30 13:50 ` [PATCH v3 16/24] x86/emul: Avoid raising faults behind the emulators back Andrew Cooper
2016-11-30 13:50 ` [PATCH v3 17/24] x86/pv: " Andrew Cooper
2016-12-01 11:50   ` Tim Deegan
2016-12-01 12:57   ` Jan Beulich
2016-12-01 13:12     ` Andrew Cooper
2016-12-01 13:27       ` Jan Beulich
2016-11-30 13:50 ` [PATCH v3 18/24] x86/shadow: " Andrew Cooper
2016-12-01 11:39   ` Tim Deegan
2016-12-01 11:40     ` Andrew Cooper
2016-12-01 13:00   ` Jan Beulich
2016-12-01 13:15     ` Andrew Cooper
2016-11-30 13:50 ` [PATCH v3 19/24] x86/hvm: Extend the hvm_copy_*() API with a pagefault_info pointer Andrew Cooper
2016-11-30 13:50 ` [PATCH v3 20/24] x86/hvm: Reimplement hvm_copy_*_nofault() in terms of no pagefault_info Andrew Cooper
2016-11-30 13:50 ` [PATCH v3 21/24] x86/hvm: Rename hvm_copy_*_guest_virt() to hvm_copy_*_guest_linear() Andrew Cooper
2016-11-30 13:50 ` [PATCH v3 22/24] x86/hvm: Avoid __hvm_copy() raising #PF behind the emulators back Andrew Cooper
2016-11-30 14:29   ` Paul Durrant
2016-11-30 13:50 ` [PATCH v3 23/24] x86/emul: Prepare to allow use of system segments for memory references Andrew Cooper
2016-11-30 13:50 ` [PATCH v3 24/24] x86/emul: Use system-segment relative memory accesses Andrew Cooper

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=1480513841-7565-12-git-send-email-andrew.cooper3@citrix.com \
    --to=andrew.cooper3@citrix.com \
    --cc=JBeulich@suse.com \
    --cc=paul.durrant@citrix.com \
    --cc=tim@xen.org \
    --cc=xen-devel@lists.xen.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).