Re: [PATCH] arm64: mm: add message to die() in die_kernel_fault()

[Yue Hu]

On Thu, 30 Jul 2020 13:27:23 +0100
Catalin Marinas <catalin.marinas@xxxxxxx> wrote:

> On Thu, Jul 30, 2020 at 07:47:57PM +0800, Yue Hu wrote:
> > From: Yue Hu <huyue2@xxxxxxxxxx>
> > 
> > Just to identify the kernel fault more clearly.
> > 
> > Signed-off-by: Yue Hu <huyue2@xxxxxxxxxx>
> > ---
> >  arch/arm64/mm/fault.c | 2 +-
> >  1 file changed, 1 insertion(+), 1 deletion(-)
> > 
> > diff --git a/arch/arm64/mm/fault.c b/arch/arm64/mm/fault.c
> > index 8afb238..3a753c7 100644
> > --- a/arch/arm64/mm/fault.c
> > +++ b/arch/arm64/mm/fault.c
> > @@ -289,7 +289,7 @@ static void die_kernel_fault(const char *msg, unsigned long addr,
> >  	mem_abort_decode(esr);
> >  
> >  	show_pte(addr);
> > -	die("Oops", regs, esr);
> > +	die("Oops - Page fault", regs, esr);
> >  	bust_spinlocks(0);
> >  	do_exit(SIGKILL);
> >  }  
> 
> Don't we already print enough information prior to die()?
> 

Yes, we have. But "Oops" is a little common. Add specific message is just to avoid to
use it repeatedly by other callers just like die("Oops - BUG",,), die("Oops - KASAN",,).

Moreover, die() will call panic() if require, panic() does not know which oops it is.
We can let panic() know it for debug expansibility such as store the panic message to
reserved memory.