Re: [PATCH v2] idle/x86: remove the call to boot_init_stack_canary() from cpu_startup_entry()

[Tyrel Datwyler]

On 10/19/2018 03:31 AM, Christophe Leroy wrote:
> commit d7880812b359 ("idle: Add the stack canary init to
> cpu_startup_entry()") added the call to boot_init_stack_canary()
> in cpu_startup_entry() in an #ifdef CONFIG_X86 statement, with
> the intention to remove that #ifdef later.
> 
> While implementing stack protector for powerpc, it has been
> observed that calling boot_init_stack_canary() is also needed
> for powerpc which uses per task (TLS) stack canary like the X86.
> 
> However, calling boot_init_stack_canary() would break arches
> using global stack canary (ARM, SH, MIPS and XTENSA).
> 
> Instead of modifying the #ifdef CONFIG_X86 in an
>  #if defined(CONFIG_X86) || defined(CONFIG_PPC), powerpc
> implemented the call to boot_init_stack_canary() in the function
> calling cpu_startup_entry()
> 
> On x86, we have two functions calling cpu_startup_entry():
> - start_secondary()
> - cpu_bringup_and_idle()
> 
> start_secondary() already calls boot_init_stack_canary().
> 
> This patch adds the call to boot_init_stack_canary() in
> cpu_bringup_and_idle() and removes it from cpu_startup_entry()
> 
> Reviewed-by: Juergen Gross <jgross@xxxxxxxx>
> Signed-off-by: Christophe Leroy <christophe.leroy@xxxxxx>
> ---
>  v2: Revised commit log (#if defined  had been droped by 'git commit')
> 
>  arch/x86/xen/smp_pv.c |  1 +
>  kernel/sched/idle.c   | 15 ---------------
>  2 files changed, 1 insertion(+), 15 deletions(-)
> 
> diff --git a/arch/x86/xen/smp_pv.c b/arch/x86/xen/smp_pv.c
> index e3b18ad49889..0e05e8e23998 100644
> --- a/arch/x86/xen/smp_pv.c
> +++ b/arch/x86/xen/smp_pv.c
> @@ -88,6 +88,7 @@ static void cpu_bringup(void)
>  asmlinkage __visible void cpu_bringup_and_idle(void)
>  {
>  	cpu_bringup();
> +	boot_init_stack_canary();

I think I saw from the kbuild test robot in your first pass that this is an implicit declaration. Quick look at my local git tree and I don't see an include for stackprotector.h in this file.

-Tyrel

>  	cpu_startup_entry(CPUHP_AP_ONLINE_IDLE);
>  }
> 
> diff --git a/kernel/sched/idle.c b/kernel/sched/idle.c
> index 16f84142f2f4..f5516bae0c1b 100644
> --- a/kernel/sched/idle.c
> +++ b/kernel/sched/idle.c
> @@ -347,21 +347,6 @@ EXPORT_SYMBOL_GPL(play_idle);
> 
>  void cpu_startup_entry(enum cpuhp_state state)
>  {
> -	/*
> -	 * This #ifdef needs to die, but it's too late in the cycle to
> -	 * make this generic (ARM and SH have never invoked the canary
> -	 * init for the non boot CPUs!). Will be fixed in 3.11
> -	 */
> -#ifdef CONFIG_X86
> -	/*
> -	 * If we're the non-boot CPU, nothing set the stack canary up
> -	 * for us. The boot CPU already has it initialized but no harm
> -	 * in doing it again. This is a good place for updating it, as
> -	 * we wont ever return from this function (so the invalid
> -	 * canaries already on the stack wont ever trigger).
> -	 */
> -	boot_init_stack_canary();
> -#endif
>  	arch_cpu_idle_prepare();
>  	cpuhp_online_idle(state);
>  	while (1)
>