Re: [PATCH 3/4] arm64/entry-common: Make Aarch32 syscalls' availability depend on aarch32_enabled()

[Andrea della Porta]

On 13:57 Wed 18 Oct     , Mark Rutland wrote:
> On Wed, Oct 18, 2023 at 01:13:21PM +0200, Andrea della Porta wrote:
> > Another major aspect of supporting running of 32bit processes is the
> > ability to access 32bit syscalls. Such syscalls can be invoked by
> > using the svc instruction.
> > 
> > If Aarch32 emulation is disabled ensure that calling svc results
> > in the same behavior as if CONFIG_COMPAT has not been enabled (i.e.
> > a kernel panic).
> 
> It's not "emulation" it's directly supported by the hardware.

You're right. I also struggled to use this label but I just reused the same
name from Nikolai's patchset for x86, in the hope that the option would be
more recognizable (something like 'ARCH_emulation' that could be used maybe
for other platforms as well), but I agree with you that this is highly
misleading. I will change it to something more straightforward.

> 
> > 
> > Signed-off-by: Andrea della Porta <andrea.porta@xxxxxxxx>
> > ---
> >  arch/arm64/kernel/entry-common.c | 25 ++++++++++++++++++++++---
> >  1 file changed, 22 insertions(+), 3 deletions(-)
> > 
> > diff --git a/arch/arm64/kernel/entry-common.c b/arch/arm64/kernel/entry-common.c
> > index 69ff9b8c0bde..32761760d9dd 100644
> > --- a/arch/arm64/kernel/entry-common.c
> > +++ b/arch/arm64/kernel/entry-common.c
> > @@ -802,6 +802,11 @@ asmlinkage void noinstr el0t_64_error_handler(struct pt_regs *regs)
> >  }
> >  
> >  #ifdef CONFIG_COMPAT
> > +UNHANDLED(el0t, 32, sync_ni)
> > +UNHANDLED(el0t, 32, irq_ni)
> > +UNHANDLED(el0t, 32, fiq_ni)
> > +UNHANDLED(el0t, 32, error_ni)
> 
> IRQ, FIQ, and SError are not syscalls, so the commit title is bad.

Agreed. I'll call them exceptions.

> 
> > +
> >  static void noinstr el0_cp15(struct pt_regs *regs, unsigned long esr)
> >  {
> >  	enter_from_user_mode(regs);
> > @@ -821,6 +826,11 @@ static void noinstr el0_svc_compat(struct pt_regs *regs)
> >  
> >  asmlinkage void noinstr el0t_32_sync_handler(struct pt_regs *regs)
> >  {
> > +	if (!aarch32_enabled()) {
> > +		el0t_32_sync_ni_handler(regs);
> > +		return;
> > +	}
> 
> Why do we have to do this at all?
> 
> If we don't have AArch32 tasks, these paths are unreachable. Why do we need to
> check that they aren't called?
> 
> Mark.

Agreed. Please see also my previous comments here:
https://lore.kernel.org/lkml/ZTEKabxNdegsbxyv@apocalypse/
https://lore.kernel.org/lkml/ZTD0DAes-J-YQ2eu@apocalypse/

but again, that's only speculative as of now, so we can ignore that part.

Andrea