On Tue, 28 Sept 2021 at 22:28, Thomas Gleixner <tglx@xxxxxxxxxxxxx> wrote: > > Ard, > > On Tue, Sep 28 2021 at 15:34, Ard Biesheuvel wrote: > > On Tue, 28 Sept 2021 at 15:33, Sebastian Andrzej Siewior > > <bigeasy@xxxxxxxxxxxxx> wrote: > >> > >> On 2021-09-28 15:30:32 [+0200], Ard Biesheuvel wrote: > >> > This is generic code and the commit log only talks about arm64. How > >> > about other architectures? > >> > >> They also invoke the EFI services with disables interrupts. If they > >> provide a RTC behind spi/i2c then we end up in the same situation right? > >> Or did I misunderstand your point? > >> > > Are you sure you want to disable EFI runtime services on all x86 > > systems with PREEMPT_RT as well? > > I'm pretty sure because we've ran into inacceptable latencies with EFI > runtime services often enough. > > Since we disabled them these complaints have gone down to 0 and nobody > so far complained about their non-availability. > Excellent. > We might revisit that and make them default disabled on RT and offer a > command line option to enable them for those who really want them. > Yes this is what Sebastian's second patch provides. > Thanks, > > tglx I'll take this as an ack from you, Thomas, and queue these up for v5.16 Thanks, Ard.
