Re: [PATCH] counter: interrupt-cnt: Protect enable/disable OPs with mutex

[Sverdlin, Alexander]

Thanks for the quick reply William,

On Fri, 2025-05-02 at 20:43 +0900, William Breathitt Gray wrote:
> > > From: Alexander Sverdlin <alexander.sverdlin@xxxxxxxxxxx>
> > > 
> > > Enable/disable seems to be racy on SMP, consider the following scenario:
> > > 
> > > CPU0 CPU1
> > > 
> > > interrupt_cnt_enable_write(true)
> > > {
> > >   if (priv->enabled == enable)
> > >   return 0;
> > > 
> > >   if (enable) {
> > >   priv->enabled = true;
> > >   interrupt_cnt_enable_write(false)
> > >   {
> > >   if (priv->enabled == enable)
> > >   return 0;
> > > 
> > >   if (enable) {
> > >   priv->enabled = true;
> > >   enable_irq(priv->irq);
> > >   } else {
> > >   disable_irq(priv->irq)
> > >   priv->enabled = false;
> > >   }
> > >   enable_irq(priv->irq);
> > >   } else {
> > >   disable_irq(priv->irq);
> > >   priv->enabled = false;
> > >   }
> > > 
> > > The above would result in priv->enabled == false, but IRQ left enabled.
> > > Protect both write (above race) and read (to propagate the value on SMP)
> > > callbacks with a mutex.
> > > 
> > > Signed-off-by: Alexander Sverdlin <alexander.sverdlin@xxxxxxxxxxx>
> > 
> > I've noticed that the patch has been marked as "Changes Requested" in
> > the patchwork, could it be a mistake? Because I never received any
> > change request.
> 
> Hi Alexander,
> 
> I can't comment on the patchwork status because I don't use that
> service, but I apologize nonetheless for the delay in responding to your
> patch submission. I'm hoping for an Ack from Oleksij, but this is a
> pretty straight-forward fix that I'll be happy to pick it up regardless.
> 
> Would you provide a Fixes line so the stable trees can pick this up for
> the necessary kernel versions?

shall I re-spin or would this separate tag suffice?

Fixes: a55ebd47f21f ("counter: add IRQ or GPIO based counter")

-- 
Alexander Sverdlin
Siemens AG
www.siemens.com