Re: [Patch v2 4/4] x86/speculation: Add prctl to control indirect branch speculation per process

[Thomas Gleixner]

On Fri, 5 Oct 2018, Tim Chen wrote:

> On 10/02/2018 10:58 AM, Thomas Gleixner wrote:
> > On Tue, 25 Sep 2018, Tim Chen wrote:
> >>  
> >> +void arch_set_dumpable(struct task_struct *tsk, struct mm_struct *mm, int value)
> >> +{
> >> +	if (!static_branch_unlikely(&spectre_v2_app_lite))
> >> +		return;
> >> +	if (!static_cpu_has(X86_FEATURE_STIBP))
> >> +		return;
> >> +
> >> +	if ((unsigned) value != SUID_DUMP_USER) {
> > 
> > First of all we use unsigned int and not unsigned, Aside of that why is the
> > argument not unsigned int right away?
> 
> 
> The original set_dumpable passes suid_dumpable, which was
> exposed via /proc/sys/fs/suid_dumpable and defined as int.
> It will make sense to define suid_dumpable as an unsigned int instead.
> 
> Would you like me to redefine suid_dumpable as unsigned int
> in sysctl.c in the patch revision as a separate clean up patch?

Yes, that makes sense.

Thanks,

	tglx