At 13:50 +0000 on 30 Nov (1480513835), Andrew Cooper wrote:
> Use x86_emul_{hw_exception,pagefault}() rather than
> {pv,hvm}_inject_page_fault() and hvm_inject_hw_exception() to cause raised
> faults to be known to the emulator.  This requires altering the callers of
> x86_emulate() to properly re-inject the event.
> While fixing this, fix the singlestep behaviour.  Previously, an otherwise
> successful emulation would fail if singlestepping was active, as the emulator
> couldn't raise #DB.  This is unreasonable from the point of view of the guest.
> We therefore tolerate #PF/#GP/SS and #DB being raised by the emulator, but
> reject anything else as unexpected.
> Signed-off-by: Andrew Cooper <>

Please update the patch description to remove the bits about
singlestepping and #DB. With that,

Acked-by: Tim Deegan <>

Xen-devel mailing list

Reply via email to