On Thu, Sep 30, 2021 at 10:40:11AM +0200, Peter Zijlstra wrote: > On Wed, Sep 29, 2021 at 06:01:57PM -0700, Josh Poimboeuf wrote: > > > - Should we use a similar sched wrapper for /proc/$pid/stack to make its > > raciness go away? > > Alternatively, can we make /stack go away? AFAICT the semantics of that > are far worse in that it wants the actual kernel stack of a task, > blocked or not, which is a total pain in the arse (not to mention a > giant infoleak and side-channel). > > > - At the risk of triggering a much larger patch set, I suspect > > get_wchan() can be made generic ;-) It's just a glorified wrapper > > around stack_trace_save_tsk(). > > Done that for you :-) Thanks! I wasn't sure the renaming was worth the churn, but the other cleanups make it much nicer. :) Do you want me to re-collect the resulting series, or do you have a tree already for -tip for this? -Kees -- Kees Cook
