Re: [PATCH v2 2/6] Input: joystick - use usb_make_subpath()

[Alejandro Colomar <alx+linux-hardening@kernel.org>]

[-- Attachment #1: Type: text/plain, Size: 1350 bytes --]

Hi Shihuang,

> Date: 2026-10-22 15:36:39+0800
> From: Shihuang Liu <shlomojune6@gmail.com>
>
[...]
> > > -	usb_make_path(udev, pxrc->phys, sizeof(pxrc->phys));
> > > -	strlcat(pxrc->phys, "/input0", sizeof(pxrc->phys));
> > > +	usb_make_subpath(udev, pxrc->phys, sizeof(pxrc->phys), "input0");
> > 
> > To be honest, I think the composed approach is more readable.  You don't
> > need to learn a new _subpath function with all the details it has.
> > 
> > Everyone knows what a str*cat() (or str*cpy, FWIW) function does.
> > 
> > I'd rather improve this by adding strscat().  The reason for removing
> > strlcat(3) is that it is vulnerable to DoS, and that it is inefficient,
> > but it is good at producing readable code, and a replacement should not
> > degrade that.
> > 
> > 
> > Have a lovely day!
> > Alex
> > 
> Besides that, strscat() has other issues:
> 
> - The `sizeof(buf) - len` arithmetic is a classic source of
> off-by-one bugs.
> - any future change means re-deriving the offsets all over again.

The whole point of str*cat() functions is that you don't need to derive
the offsets at all.  Just look at the strlcat(3) code above: it doesn't
do any length calculations.

> I'd prefer to avoid strscat() here.
> 
> Thanks
> Shihuang


Cheers,
Alex

-- 
<https://www.alejandro-colomar.es>

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]