On Fri, Dec 17, 2021 at 10:51 PM Thomas Gleixner <tglx@xxxxxxxxxxxxx> wrote: > > Alexander, > > On Tue, Dec 14 2021 at 17:20, Alexander Potapenko wrote: > > When calling KMSAN-instrumented functions from non-instrumented > > functions, function parameters may not be initialized properly, leading > > to false positive reports. In particular, this happens all the time when > > calling interrupt handlers from `noinstr` IDT entries. > > > > Fortunately, x86 code has instrumentation_begin() and > > It's not only x86 code: > > kernel/entry/common.c | 3 +++ Shall this bit go into a separate patch? > > @@ -76,6 +77,7 @@ __visible noinstr void do_syscall_64(struct pt_regs *regs, int nr) > > nr = syscall_enter_from_user_mode(regs, nr); > > > > instrumentation_begin(); > > + kmsan_instrumentation_begin(regs); > > Can we please make this something like: > > instrumentation_begin_at_entry(regs); Fine, will do. Do you think it would make sense to hide it inside instrumentation_begin(), or is it ok to have both macros follow each other? > or some other sensible name which hides that kmsan gunk and avoids to > touch all of this again when KFOOSAN comes around? > > Thanks, > > tglx > > > -- Alexander Potapenko Software Engineer Google Germany GmbH Erika-Mann-Straße, 33 80636 München Geschäftsführer: Paul Manicle, Halimah DeLaine Prado Registergericht und -nummer: Hamburg, HRB 86891 Sitz der Gesellschaft: Hamburg