linux-kernel - Re: [PATCH] clocksource: atmel-st: Variable sr in at91rm9200_timer

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: <20190903081658.GK21922@piout.net>
Date:   Tue, 3 Sep 2019 10:16:58 +0200
From:   Alexandre Belloni <alexandre.belloni@...tlin.com>
To:     Yizhuo Zhai <yzhai003@....edu>
Cc:     Chengyu Song <csong@...ucr.edu>, Zhiyun Qian <zhiyunq@...ucr.edu>,
        Daniel Lezcano <daniel.lezcano@...aro.org>,
        Thomas Gleixner <tglx@...utronix.de>,
        Nicolas Ferre <nicolas.ferre@...rochip.com>,
        Ludovic Desroches <ludovic.desroches@...rochip.com>,
        linux-kernel@...r.kernel.org, linux-arm-kernel@...ts.infradead.org
Subject: Re: [PATCH] clocksource: atmel-st: Variable sr in
 at91rm9200_timer_interrupt() could be uninitialized

On 02/09/2019 22:56:48-0700, Yizhuo Zhai wrote:
> In function regmap_read(),  there're two places which could make the read fail.
> 
> First, if "reg" and  "map->reg_stride" are not aligned, then remap_read() will
> return -EINVAL without initialize variable "val".
> 

A quick look at of_syscon_register would show you that this is not
possible.

> Second, _regmap_read() could also fail and return error code if "val" is not
> initialized. The caller remap_read() returns the same error code, but
> at91rm9200_timer_interrupt() does not use this information.
> 

How would _regmap_read fail exactly?

> On Mon, Sep 2, 2019 at 3:37 PM Alexandre Belloni
> <alexandre.belloni@...tlin.com> wrote:
> >
> > On 02/09/2019 15:29:46-0700, Yizhuo wrote:
> > > Inside function at91rm9200_timer_interrupt(), variable sr could
> > > be uninitialized if regmap_read() fails. However, sr is used
> >
> > Could you elaborate on how this could fail?
> >
> > > to decide the control flow later in the if statement, which is
> > > potentially unsafe. We could check the return value of
> > > regmap_read() and print an error here.
> > >
> > > Signed-off-by: Yizhuo <yzhai003@....edu>
> > > ---
> > >  drivers/clocksource/timer-atmel-st.c | 8 +++++++-
> > >  1 file changed, 7 insertions(+), 1 deletion(-)
> > >
> > > diff --git a/drivers/clocksource/timer-atmel-st.c b/drivers/clocksource/timer-atmel-st.c
> > > index ab0aabfae5f0..061a3f27847e 100644
> > > --- a/drivers/clocksource/timer-atmel-st.c
> > > +++ b/drivers/clocksource/timer-atmel-st.c
> > > @@ -48,8 +48,14 @@ static inline unsigned long read_CRTR(void)
> > >  static irqreturn_t at91rm9200_timer_interrupt(int irq, void *dev_id)
> > >  {
> > >       u32 sr;
> > > +     int ret;
> > > +
> > > +     ret = regmap_read(regmap_st, AT91_ST_SR, &sr);
> > > +     if (ret) {
> > > +             pr_err("Fail to read AT91_ST_SR.\n");
> > > +             return ret;
> > > +     }
> > >
> > > -     regmap_read(regmap_st, AT91_ST_SR, &sr);
> > >       sr &= irqmask;
> > >
> > >       /*
> > > --
> > > 2.17.1
> > >
> >
> > --
> > Alexandre Belloni, Bootlin
> > Embedded Linux and Kernel engineering
> > https://bootlin.com
> 
> 
> 
> -- 
> Kind Regards,
> 
> Yizhuo Zhai
> 
> Computer Science, Graduate Student
> University of California, Riverside

-- 
Alexandre Belloni, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com