* [PATCH] vmx: Restore debug registers when injecting #DB traps
@ 2016-03-11 14:51 Ross Lagerwall
2016-03-11 15:41 ` Jan Beulich
0 siblings, 1 reply; 2+ messages in thread
From: Ross Lagerwall @ 2016-03-11 14:51 UTC (permalink / raw)
To: xen-devel
Cc: Ross Lagerwall, Kevin Tian, Jan Beulich, Jun Nakajima, Andrew Cooper
Commit a929bee0e652 ("x86/vmx: Fix injection of #DB traps following
XSA-156") prevents an infinite loop in certain #DB traps. However, it
changed the behavior to not call hvm_hw_inject_trap() for #DB and #AC
traps which which means that the debug registers are not restored
correctly and nullified commit b56ae5b48c38 ("VMX: fix/adjust trap
injection").
To fix this, restore the original code path through hvm_inject_trap(),
but ensure that the struct hvm_trap is populated with all the required
data.
Signed-off-by: Ross Lagerwall <ross.lagerwall@citrix.com>
---
xen/arch/x86/hvm/vmx/vmx.c | 25 ++++++++++++-------------
1 file changed, 12 insertions(+), 13 deletions(-)
diff --git a/xen/arch/x86/hvm/vmx/vmx.c b/xen/arch/x86/hvm/vmx/vmx.c
index 9c5a388..1eddba5 100644
--- a/xen/arch/x86/hvm/vmx/vmx.c
+++ b/xen/arch/x86/hvm/vmx/vmx.c
@@ -3091,24 +3091,23 @@ static int vmx_handle_eoi_write(void)
* It is the callers responsibility to ensure that this function is only used
* in the context of an appropriate vmexit.
*/
-static void vmx_propagate_intr(void)
+static void vmx_propagate_intr(unsigned long intr)
{
- unsigned long intr, tmp;
-
- __vmread(VM_EXIT_INTR_INFO, &intr);
-
- ASSERT(intr & INTR_INFO_VALID_MASK);
-
- __vmwrite(VM_ENTRY_INTR_INFO, intr);
+ struct hvm_trap trap = {
+ .vector = intr & INTR_INFO_VECTOR_MASK,
+ .type = MASK_EXTR(intr, INTR_INFO_INTR_TYPE_MASK) };
+ unsigned long tmp;
if ( intr & INTR_INFO_DELIVER_CODE_MASK )
{
__vmread(VM_EXIT_INTR_ERROR_CODE, &tmp);
- __vmwrite(VM_ENTRY_EXCEPTION_ERROR_CODE, tmp);
+ trap.error_code = tmp;
}
-
+ else
+ trap.error_code = HVM_DELIVER_NO_ERROR_CODE;
__vmread(VM_EXIT_INSTRUCTION_LEN, &tmp);
- __vmwrite(VM_ENTRY_INSTRUCTION_LEN, tmp);
+ trap.insn_len = tmp;
+ hvm_inject_trap(&trap);
}
static void vmx_idtv_reinject(unsigned long idtv_info)
@@ -3366,7 +3365,7 @@ void vmx_vmexit_handler(struct cpu_user_regs *regs)
HVMTRACE_1D(TRAP_DEBUG, exit_qualification);
write_debugreg(6, exit_qualification | DR_STATUS_RESERVED_ONE);
if ( !v->domain->debugger_attached )
- vmx_propagate_intr();
+ vmx_propagate_intr(intr_info);
else
domain_pause_for_debugger();
break;
@@ -3437,7 +3436,7 @@ void vmx_vmexit_handler(struct cpu_user_regs *regs)
break;
case TRAP_alignment_check:
HVMTRACE_1D(TRAP, vector);
- vmx_propagate_intr();
+ vmx_propagate_intr(intr_info);
break;
case TRAP_nmi:
if ( MASK_EXTR(intr_info, INTR_INFO_INTR_TYPE_MASK) !=
--
2.4.3
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
http://lists.xen.org/xen-devel
^ permalink raw reply related [flat|nested] 2+ messages in thread
* Re: [PATCH] vmx: Restore debug registers when injecting #DB traps
2016-03-11 14:51 [PATCH] vmx: Restore debug registers when injecting #DB traps Ross Lagerwall
@ 2016-03-11 15:41 ` Jan Beulich
0 siblings, 0 replies; 2+ messages in thread
From: Jan Beulich @ 2016-03-11 15:41 UTC (permalink / raw)
To: Andrew Cooper, Ross Lagerwall; +Cc: Kevin Tian, Jun Nakajima, xen-devel
>>> On 11.03.16 at 15:51, <ross.lagerwall@citrix.com> wrote:
> --- a/xen/arch/x86/hvm/vmx/vmx.c
> +++ b/xen/arch/x86/hvm/vmx/vmx.c
> @@ -3091,24 +3091,23 @@ static int vmx_handle_eoi_write(void)
> * It is the callers responsibility to ensure that this function is only used
> * in the context of an appropriate vmexit.
> */
> -static void vmx_propagate_intr(void)
> +static void vmx_propagate_intr(unsigned long intr)
> {
> - unsigned long intr, tmp;
> -
> - __vmread(VM_EXIT_INTR_INFO, &intr);
> -
> - ASSERT(intr & INTR_INFO_VALID_MASK);
> -
> - __vmwrite(VM_ENTRY_INTR_INFO, intr);
> + struct hvm_trap trap = {
> + .vector = intr & INTR_INFO_VECTOR_MASK,
> + .type = MASK_EXTR(intr, INTR_INFO_INTR_TYPE_MASK) };
Please use MASK_EXTR() for both. Also the closing brace would better
go on the next line.
> + unsigned long tmp;
>
> if ( intr & INTR_INFO_DELIVER_CODE_MASK )
> {
> __vmread(VM_EXIT_INTR_ERROR_CODE, &tmp);
> - __vmwrite(VM_ENTRY_EXCEPTION_ERROR_CODE, tmp);
> + trap.error_code = tmp;
> }
> -
> + else
> + trap.error_code = HVM_DELIVER_NO_ERROR_CODE;
> __vmread(VM_EXIT_INSTRUCTION_LEN, &tmp);
> - __vmwrite(VM_ENTRY_INSTRUCTION_LEN, tmp);
> + trap.insn_len = tmp;
For this one I was unsure already for the original change (but it had
gone in already by the time I got to see it): The VM-exit instruction
length field is undefined (i.e. not necessarily zero) for the #AC
intercept case, and also for some (most) of the #DB ones. I therefore
think this needs some qualification, with zero getting used otherwise.
Jan
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
http://lists.xen.org/xen-devel
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2016-03-11 15:41 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2016-03-11 14:51 [PATCH] vmx: Restore debug registers when injecting #DB traps Ross Lagerwall
2016-03-11 15:41 ` Jan Beulich
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).