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] [thread-next>] [day] [month] [year] [list]
Message-ID: <Yxitmkw4sNx3328k@smile.fi.intel.com>
Date:   Wed, 7 Sep 2022 17:41:30 +0300
From:   Andy Shevchenko <andriy.shevchenko@...el.com>
To:     "Farber, Eliav" <farbere@...zon.com>
Cc:     jdelvare@...e.com, linux@...ck-us.net, robh+dt@...nel.org,
        p.zabel@...gutronix.de, rtanwar@...linear.com,
        linux-hwmon@...r.kernel.org, devicetree@...r.kernel.org,
        linux-kernel@...r.kernel.org, hhhawa@...zon.com, jonnyc@...zon.com
Subject: Re: [PATCH v4 06/21] hwmon: (mr75203) fix multi-channel voltage
 reading

On Wed, Sep 07, 2022 at 08:15:36AM +0300, Farber, Eliav wrote:
> On 9/6/2022 5:10 PM, Andy Shevchenko wrote:
> > On Tue, Sep 06, 2022 at 08:33:41AM +0000, Eliav Farber wrote:

...

> > > +             total_ch = ch_num * vm_num;
> > > +             in_config = devm_kcalloc(dev, total_ch + 1,
> > >                                        sizeof(*in_config), GFP_KERNEL);
> > 
> > Strictly speaking this should be `size_add(size_mul(...) ...)`
> > construction
> > from overflow.h.
> > 
> >                total_ch = size_mul(ch_num, vm_num);
> >                in_config = devm_kcalloc(dev, size_add(total_ch, 1),
> >                                         sizeof(*in_config), GFP_KERNEL);
> > 
> > Alternatively before doing all these, add a check
> > 
> >                if (array3_size(ch_num, vm_num, sizeof(*in_config)) <
> > SIZE_MAX - sizeof(*in_config))
> >                        return -EOVERFLOW;
> > 
> > But this is a bit monstrous. Seems like the above looks and feels better.
> > 
> > Also for backporting purposes perhaps it's fine to do without using
> > those macro
> > helpers.
> According to the driver code total_ch is a u32 variable while vm_num
> and ch_num are both limited to a value of 31:
> 
> #define VM_NUM_MSK GENMASK(20, 16)
> #define VM_NUM_SFT 16
> #define CH_NUM_MSK GENMASK(31, 24)
> #define CH_NUM_SFT 24
> 
> In addition the PVT Controller Series 3+ Specification mentions that
> the actual maximum values are even smaller – 8 for vm_num and 16 for
> ch_num.
> Therefore we are very far from a scenario of an overflow.
> Do you still think overflow protection in necessary?

Like I said "Strictly..." Means it's up to you, but allocations are
usually be protected against the overflows.

-- 
With Best Regards,
Andy Shevchenko


Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ