lib/tests/printf_kunit.c | 20 +++++++++++--------- 1 file changed, 11 insertions(+), 9 deletions(-)
This allows the compiler to check the arguments against the __printf
attribute on __test. This produces better diagnostics when incorrect
inputs are passed.
Reported-by: kernel test robot <lkp@intel.com>
Closes: https://lore.kernel.org/oe-kbuild-all/202512061600.89CKQ3ag-lkp@intel.com/
Signed-off-by: Tamir Duberstein <tamird@kernel.org>
---
Changes in v2:
- Convert test_hashed to macro rather than try to annotate it in a way
that is compatible with both clang and gcc.
- Link to v1: https://patch.msgid.link/20251206-printf-kunit-printf-attr-v1-1-1682808b51d0@gmail.com
---
lib/tests/printf_kunit.c | 20 +++++++++++---------
1 file changed, 11 insertions(+), 9 deletions(-)
diff --git a/lib/tests/printf_kunit.c b/lib/tests/printf_kunit.c
index 7617e5b8b02c..1d96cea8af65 100644
--- a/lib/tests/printf_kunit.c
+++ b/lib/tests/printf_kunit.c
@@ -266,15 +266,17 @@ hash_pointer(struct kunit *kunittest)
KUNIT_EXPECT_MEMNEQ(kunittest, buf, PTR_STR, PTR_WIDTH);
}
-static void
-test_hashed(struct kunit *kunittest, const char *fmt, const void *p)
-{
- char buf[PLAIN_BUF_SIZE];
-
- plain_hash_to_buffer(kunittest, p, buf, PLAIN_BUF_SIZE);
-
- test(buf, fmt, p);
-}
+/*
+ * This is a macro so that the compiler can compare its arguments to the
+ * __printf attribute on __test. This cannot be a function with a __printf
+ * attribute because GCC requires __printf functions to be variadic.
+ */
+#define test_hashed(kunittest, fmt, p) \
+ do { \
+ char buf[PLAIN_BUF_SIZE]; \
+ plain_hash_to_buffer(kunittest, p, buf, PLAIN_BUF_SIZE); \
+ test(buf, fmt, p); \
+ } while (0)
/*
* NULL pointers aren't hashed.
---
base-commit: 983d014aafb14ee5e4915465bf8948e8f3a723b5
change-id: 20251206-printf-kunit-printf-attr-19369fc57bf0
Best regards,
--
Tamir Duberstein <tamird@kernel.org>On Fri, Jan 16, 2026 at 11:27:03AM -0500, Tamir Duberstein wrote:
> This allows the compiler to check the arguments against the __printf
__printf() since it takes parameters, OTOH it's an attribute at the end,
so I have no strong opinion on how to spell it.
> attribute on __test. This produces better diagnostics when incorrect
__test()
*This is reference to a function.
> inputs are passed.
...
> +/*
> + * This is a macro so that the compiler can compare its arguments to the
> + * __printf attribute on __test. This cannot be a function with a __printf
> + * attribute because GCC requires __printf functions to be variadic.
As per commit message remarks.
> + */
> +#define test_hashed(kunittest, fmt, p) \
> + do { \
> + char buf[PLAIN_BUF_SIZE]; \
> + plain_hash_to_buffer(kunittest, p, buf, PLAIN_BUF_SIZE); \
> + test(buf, fmt, p); \
> + } while (0)
Make sure you used tabs to indent the \:s.
...
The downside of a macro is a killing of compile-time type checks.
--
With Best Regards,
Andy Shevchenko
On Mon, Jan 19, 2026 at 3:21 AM Andy Shevchenko
<andriy.shevchenko@linux.intel.com> wrote:
>
> On Fri, Jan 16, 2026 at 11:27:03AM -0500, Tamir Duberstein wrote:
> > This allows the compiler to check the arguments against the __printf
>
> __printf() since it takes parameters, OTOH it's an attribute at the end,
> so I have no strong opinion on how to spell it.
>
> > attribute on __test. This produces better diagnostics when incorrect
>
> __test()
>
> *This is reference to a function.
>
> > inputs are passed.
>
> ...
>
> > +/*
> > + * This is a macro so that the compiler can compare its arguments to the
> > + * __printf attribute on __test. This cannot be a function with a __printf
> > + * attribute because GCC requires __printf functions to be variadic.
>
> As per commit message remarks.
>
> > + */
> > +#define test_hashed(kunittest, fmt, p) \
> > + do { \
> > + char buf[PLAIN_BUF_SIZE]; \
> > + plain_hash_to_buffer(kunittest, p, buf, PLAIN_BUF_SIZE); \
> > + test(buf, fmt, p); \
> > + } while (0)
>
> Make sure you used tabs to indent the \:s.
>
> ...
>
> The downside of a macro is a killing of compile-time type checks.
This macro still bottoms out in __test() which is a function, so I
believe it preserves (and enhances, per the commit message)
compile-time checks.
I'll send v3 with the changes you requested above.
>
> --
> With Best Regards,
> Andy Shevchenko
>
>
On Tue 2026-01-20 10:50:09, Tamir Duberstein wrote:
> On Mon, Jan 19, 2026 at 3:21 AM Andy Shevchenko
> <andriy.shevchenko@linux.intel.com> wrote:
> >
> > On Fri, Jan 16, 2026 at 11:27:03AM -0500, Tamir Duberstein wrote:
> > > This allows the compiler to check the arguments against the __printf
> >
> > __printf() since it takes parameters, OTOH it's an attribute at the end,
> > so I have no strong opinion on how to spell it.
> >
> > > attribute on __test. This produces better diagnostics when incorrect
> >
> > __test()
> >
> > *This is reference to a function.
> >
> > > inputs are passed.
> >
> > ...
> >
> > > +/*
> > > + * This is a macro so that the compiler can compare its arguments to the
> > > + * __printf attribute on __test. This cannot be a function with a __printf
> > > + * attribute because GCC requires __printf functions to be variadic.
> >
> > As per commit message remarks.
> >
> > > + */
> > > +#define test_hashed(kunittest, fmt, p) \
> > > + do { \
> > > + char buf[PLAIN_BUF_SIZE]; \
> > > + plain_hash_to_buffer(kunittest, p, buf, PLAIN_BUF_SIZE); \
> > > + test(buf, fmt, p); \
> > > + } while (0)
> >
> > Make sure you used tabs to indent the \:s.
> >
> > ...
> >
> > The downside of a macro is a killing of compile-time type checks.
>
> This macro still bottoms out in __test() which is a function, so I
> believe it preserves (and enhances, per the commit message)
> compile-time checks.
>
> I'll send v3 with the changes you requested above.
JFYI, I am fine with this approach.
Best Regards,
Petr
© 2016 - 2026 Red Hat, Inc.