Re: [PATCH] tty: serial: samsung_tty: remove spinlock flags in interrupt handlers

[Andy Shevchenko]

On Fri, Mar 19, 2021 at 10:09 AM Johan Hovold <johan@xxxxxxxxxx> wrote:
>
> On Fri, Mar 19, 2021 at 06:36:39AM +0000, Song Bao Hua (Barry Song) wrote:
> >
> >
> > > -----Original Message-----
> > > From: Andy Shevchenko [mailto:andy.shevchenko@xxxxxxxxx]
> > > Sent: Tuesday, March 16, 2021 10:41 PM
> > > To: Johan Hovold <johan@xxxxxxxxxx>; Finn Thain <fthain@xxxxxxxxxxxxxxxxxxx>;
> > > Song Bao Hua (Barry Song) <song.bao.hua@xxxxxxxxxxxxx>
> > > Cc: Krzysztof Kozlowski <krzysztof.kozlowski@xxxxxxxxxxxxx>; Greg
> > > Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>; Jiri Slaby <jirislaby@xxxxxxxxxx>;
> > > linux-arm Mailing List <linux-arm-kernel@xxxxxxxxxxxxxxxxxxx>; Linux Samsung
> > > SOC <linux-samsung-soc@xxxxxxxxxxxxxxx>; open list:SERIAL DRIVERS
> > > <linux-serial@xxxxxxxxxxxxxxx>; Linux Kernel Mailing List
> > > <linux-kernel@xxxxxxxxxxxxxxx>; Hector Martin <marcan@xxxxxxxxx>; Arnd
> > > Bergmann <arnd@xxxxxxxxxx>
> > > Subject: Re: [PATCH] tty: serial: samsung_tty: remove spinlock flags in
> > > interrupt handlers
> > >
> > > On Tue, Mar 16, 2021 at 11:02 AM Johan Hovold <johan@xxxxxxxxxx> wrote:
> > > >
> > > > On Mon, Mar 15, 2021 at 07:12:12PM +0100, Krzysztof Kozlowski wrote:
> > > > > Since interrupt handler is called with disabled local interrupts, there
> > > > > is no need to use the spinlock primitives disabling interrupts as well.
> > > >
> > > > This isn't generally true due to "threadirqs" and that can lead to
> > > > deadlocks if the console code is called from hard irq context.
> > > >
> > > > Now, this is *not* the case for this particular driver since it doesn't
> > > > even bother to take the port lock in console_write(). That should
> > > > probably be fixed instead.
> > > >
> > > > See https://lore.kernel.org/r/X7kviiRwuxvPxC8O@localhost.
> > >
> > > Finn, Barry, something to check I think?
> >
> > My understanding is that spin_lock_irqsave can't protect the context
> > the console_write() is called in hardirq for threaded_irq case mainly
> > for preempt-rt scenarios as spin_lock_irqsave doesn't disable irq in
> > that case at all.
>
> Forced threaded interrupts have so far run with interrupts enabled and
> spin_lock_irqsave() would suffice on non-RT. This is about to change
> though so that drivers don't need to worry about "threadirqs":
>
>         https://lore.kernel.org/r/20210317143859.513307808@xxxxxxxxxxxxx
>
> > See:
> > https://www.kernel.org/doc/html/latest/locking/locktypes.html
> > spinlock_t and PREEMPT_RT
> > On a PREEMPT_RT kernel spinlock_t is mapped to a separate implementation
> > based on rt_mutex which changes the semantics:
> > Preemption is not disabled.
> > The hard interrupt related suffixes for spin_lock / spin_unlock operations
> > (_irq, _irqsave / _irqrestore) do not affect the CPU’s interrupt disabled
> > state.
> >
> > So if console_write() can interrupt our code in hardirq, we should
> > move to raw_spin_lock_irqsave for this driver.
>
> No, no. RT handles this by deferring console writes apparently.
>
> > I think it is almost always wrong to call spin_lock_irqsave in hardirq.
>
> Again, no. It's even been a requirement due to "threadirqs" in some
> cases (e.g. hrtimers) up until now (or rather until the above patch is
> in mainline).

By the way, a good question Imre (Cc'ed) and I have discussed is the
in-kernel documentation, i.e.
https://www.kernel.org/doc/html/latest/kernel-hacking/locking.html.
Should it be adjusted to reality?

-- 
With Best Regards,
Andy Shevchenko