Re: [PATCH 0/5] x86/pv: #DB vs %dr6 fixes, part 2

[Andrew Cooper <andrew.cooper3@xxxxxxxxxx>]

On 13/09/2023 12:21 am, Andrew Cooper wrote:
> Slightly RFC.  This is the next chunk of debug fixes from the bug that Jinoh
> reported.

I was trying to do a bit of due diligence before posting v2, and have
made some discoveries.

pv/emul-priv-op SingleStep vs Branch Step
  https://gitlab.com/xen-project/xen/-/work_items/170

HVM IO Breakpoints:
  https://gitlab.com/xen-project/xen/-/work_items/171


A third which I'm on the fence about is about PV guests and General
Detect.  We firmly prohibit PV guests from setting DR7.GD, but we them
play with the DR6.GD bit as if it had been asserted.

It would be easy to put GD back into the set of reserved bits for DR6,
but it also wouldn't be hard to handle GD via dr7_emul and let the PV
guest have a more-normal looking set of debug functionality.

Thoughts?

~Andrew