On Fri, Jan 17, 2025 at 03:26:13PM +0100, Greg Kroah-Hartman wrote:
> On Fri, Jan 17, 2025 at 03:42:27PM +0200, Andy Shevchenko wrote:
> > On Fri, Jan 17, 2025 at 07:11:46AM +0100, Greg Kroah-Hartman wrote:
> > > On Thu, Jan 16, 2025 at 06:05:43PM +0200, Andy Shevchenko wrote:
> > > > GCC is not happy about the buffer size:
> > > >
> > > > drivers/usb/core/hcd.c:441:48: error: ‘%s’ directive output may be truncated writing up to 64 bytes into a region of size between 35 and 99 [-Werror=format-truncation=]
> > > > 441 | snprintf (buf, sizeof buf, "%s %s %s", init_utsname()->sysname,
> > > > | ^~
> > > > 442 | init_utsname()->release, hcd->driver->description);
> > > > | ~~~~~~~~~~~~~~~~~~~~~~~
> > > >
> > > > Bump the size to get it enough for the possible strings.
...
> > > > static unsigned
> > > > rh_string(int id, struct usb_hcd const *hcd, u8 *data, unsigned len)
> > > > {
> > > > - char buf[100];
> > > > + char buf[160];
> > > > char const *s;
> > > > static char const langids[4] = {4, USB_DT_STRING, 0x09, 0x04};
> > >
> > > Worst case it's properly truncated so why do we need to worry about this
> > > "warning"?
> >
> > With CONFIG_WERROR=y it's a compilation error. My goal is to have
> > i386_defconfig and x86_64_defconfig to be compiled with `make W=1`.
>
> So you have to have W=1 enabled, right?
Yep!
> On my normal builds, with CONFIG_WERROR=y enabled, I do not see this.
>
> > > And what compiler version is giving that, I don't see that
> > > here in my build testing.
> >
> > `make W=1` (and be sure that CONFIG_WERROR=y).
>
> Ah, ok, manual work here.
>
> And I guess the error is right, ->sysname could be 64 and release can
> also be 64 bytes long, so it would be truncated.
Yeah... Should I update the commit message and issue v2?
--
With Best Regards,
Andy Shevchenko
