Re: [PATCH v4 1/2] kasan/hw-tags: introduce kasan.write_only option

[Yeoreum Yun <yeoreum.yun@arm.com>]

Hi Andery,
> On Mon, Aug 18, 2025 at 3:11 PM Yeoreum Yun <yeoreum.yun@arm.com> wrote:
> >
> > > > +           hw_enable_tag_checks_write_only()) {
> > > > +               kasan_arg_write_only == KASAN_ARG_WRITE_ONLY_OFF;
> > >
> > > Typo in == in the line above. But also I think we can just drop the
> > > line: kasan_arg_write_only is KASAN_ARG_WRITE_ONLY_ON after all, it's
> > > just not supported and thus kasan_flag_write_only is set to false to
> > > reflect that.
> >
> > Sorry :\ I've missed this fix from patch 3... this should be == to =.
> >
> > However, we couldn't remove kasan_arg_write_only check in condition.
> > If one of cpu get failed to hw_enable_tag_checks_write_only() then
> > By changing this with KASAN_ARG_WRITE_ONLY_OFF, It prevent to call
> > hw_eanble_tag_checks_write_only() in other cpu.
>
> Is it possible that the write-only mode will fail to be enabled on one
> CPU but then get enabled successfully for another?
>
> What would happen with the current code if the first CPU succeeds in
> enabling the write-only mode, and the second one fails?

The only failure case is when CPU doesn't support the MTE_STORE_ONLY feature.
Since MTE_STORE_ONLY is BOOT CPU feature, the CPUs have two state --
all CPUs have a MTE_STORE_ONLY feature or they doesn't have it.

So when boot-cpu try to enable the write-only option according to
"write_only=on" boot argument but the cpu doesn't have this feature,
hw_enable_tag_checks_write_only() will fail and prevent the calling
hw_enable_tag_checks_write_only() in other cpu by kasan_arg_write_only
as OFF to prevent other cpu call this function since it'll be failed
anyway.

So there is no case for failure -- the first CPU succeeds but second one
fails. if first one succeeds, all cpus will success ans vice versa.

This condition just to prevent to call
hw_enable_tags_checks_write_only() by other cpu if first cpu found it
doesn't support MTE_STORE_ONLY feature since other doesn't need to call
hw_eanble_tag_checks_write_only() function -- it'll be failed
so it is meaningless call.

>
> > As you said, kasan_flag_write_only reflects the state.
> > But like other option, I keep the condition to call the hw_enable_xxx()
> > by checking the "argments" and keep the "hw enable state" with
> > kasan_flag_write_only.
>
> Assuming we keep this behavior, please add a comment explaining all this.

Okay. :)

--
Sincerely,
Yeoreum Yun