On Wed, Dec 13, 2023 at 11:43:49AM -0800, Deepak Gupta wrote:
On Wed, Dec 13, 2023 at 5:37 AM Mark Brown broonie@kernel.org wrote:
On Tue, Dec 12, 2023 at 04:50:38PM -0800, Deepak Gupta wrote:
How will it do that (currently _ENABLE is married to _WRITE and _PUSH) ?
That's feeling moderately firmly into "don't do that" territory to be honest, the problems of trying to modify the stack of another running thread while it's active just don't seem worth it - if you're coordinating enough to do the modifications it's probably possible to just ask the thread who's stack is being modified to do the modification itself and having an unprotected thread writing into shadow stack memory doesn't feel great.
Yeah no leanings on my side. Just wanted to articulate this scenario. Since this is new ground, we can define what's appropriate. Let's keep it this way where a thread can write to shadow stack mappings only when it itself has shadow stack enabled.
Sounds good to me - it's much easier to relax permissions later than to tighten them up.