linux-kernel - RE: [PATCH 1/1] tty: serial: samsung: add spin_lock for interrupt and console

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <001601d848b8$10c97490$325c5db0$@samsung.com>
Date:   Tue, 5 Apr 2022 15:40:39 +0900
From:   "Jaewon Kim" <jaewon02.kim@...sung.com>
To:     "'Greg Kroah-Hartman'" <gregkh@...uxfoundation.org>
Cc:     "'Krzysztof Kozlowski'" <krzk@...nel.org>,
        "'Alim Akhtar'" <alim.akhtar@...sung.com>,
        "'Jiri Slaby'" <jirislaby@...nel.org>,
        <linux-samsung-soc@...r.kernel.org>,
        <linux-arm-kernel@...ts.infradead.org>,
        <linux-serial@...r.kernel.org>, <linux-kernel@...r.kernel.org>,
        "'Chanho Park'" <chanho61.park@...sung.com>
Subject: RE: [PATCH 1/1] tty: serial: samsung: add spin_lock for interrupt
 and console_write

Hello

On 22. 4. 5. 14:01, Greg Kroah-Hartman wrote:
> On Tue, Apr 05, 2022 at 12:38:54PM +0900, Jaewon Kim wrote:
> > The console_write and IRQ handler can run concurrently.
> > Problems may occurs console_write is continuously executed while the
> > IRQ handler is running.
> >
> > Signed-off-by: Jaewon Kim <jaewon02.kim@...sung.com>
> > ---
> >  drivers/tty/serial/samsung_tty.c | 14 ++++++++++++++
> >  1 file changed, 14 insertions(+)
> 
> What commit does this fix?

This is not an issue caused by anohter commits.
There was potential issue from the beginning.

Other drivers were fixed, but samsung_tty was not.
PL011 patch : https://lkml.org/lkml/2012/2/1/495


> 
> >
> > diff --git a/drivers/tty/serial/samsung_tty.c
> > b/drivers/tty/serial/samsung_tty.c
> > index e1585fbae909..d362e8e114f1 100644
> > --- a/drivers/tty/serial/samsung_tty.c
> > +++ b/drivers/tty/serial/samsung_tty.c
> > @@ -2480,12 +2480,26 @@ s3c24xx_serial_console_write(struct console *co, const char *s,
> >  			     unsigned int count)
> >  {
> >  	unsigned int ucon = rd_regl(cons_uart, S3C2410_UCON);
> > +	unsigned long flags;
> > +	int locked = 1;
> 
> bool?

It is return value of spin_trylock()
I used int because mose drivers used int.
If you guide to change int to bool, I will change it.

> 
> >
> >  	/* not possible to xmit on unconfigured port */
> >  	if (!s3c24xx_port_configured(ucon))
> >  		return;
> >
> > +	local_irq_save(flags);
> > +	if (cons_uart->sysrq)
> > +		locked = 0;
> > +	else if (oops_in_progress)
> > +		locked = spin_trylock(&cons_uart->lock);
> > +	else
> > +		spin_lock(&cons_uart->lock);
> > +
> >  	uart_console_write(cons_uart, s, count,
> > s3c24xx_serial_console_putchar);
> > +
> > +	if (locked)
> > +		spin_unlock(&cons_uart->lock);
> > +	local_irq_restore(flags);
> 
> Why is irq_save required as well as a spinlock?

No special reason.
I will change spin_trylock() -? spin_trylock_irqsave().
spin_lock -> spin_lock_irqsave().
And, remove local_irq_save/restore.
It looks more clean.


> 
> thanks,
> 
> greg k-h

Thanks
Jaewon Kim