[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CAFd5g46usCBp_a_eTpgY=-mD5bLZTzyBt+aO-FTDBNkAAc9=UA@mail.gmail.com>
Date: Tue, 1 Mar 2022 14:56:35 -0500
From: Brendan Higgins <brendanhiggins@...gle.com>
To: Kees Cook <keescook@...omium.org>
Cc: "Gustavo A. R. Silva" <gustavoars@...nel.org>,
Nathan Chancellor <nathan@...nel.org>,
Nick Desaulniers <ndesaulniers@...gle.com>,
Rasmus Villemoes <linux@...musvillemoes.dk>,
Vitor Massaru Iha <vitor@...saru.org>,
Daniel Latypov <dlatypov@...gle.com>,
David Gow <davidgow@...gle.com>,
Anton Ivanov <anton.ivanov@...bridgegreys.com>,
Jeff Dike <jdike@...toit.com>,
Richard Weinberger <richard@....at>,
Masahiro Yamada <masahiroy@...nel.org>,
Arnd Bergmann <arnd@...db.de>,
Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
linux-um <linux-um@...ts.infradead.org>,
linux-kbuild <linux-kbuild@...r.kernel.org>,
KUnit Development <kunit-dev@...glegroups.com>,
llvm@...ts.linux.dev, x86@...nel.org,
linux-hardening@...r.kernel.org, shuah <shuah@...nel.org>
Subject: Re: [PATCH v3 7/7] UAPI: Introduce KUnit userspace compatibility
+Bart Van Assche - I don't know if you are still working on those
configfs tests, but I wonder if this would be interesting to you.
On Sun, Feb 27, 2022 at 1:45 PM Kees Cook <keescook@...omium.org> wrote:
>
> The original lib/test_stackinit.c, which exclusively tests toolchain
> features, was designed to also be built without the full Linux kernel
> sources so that compiler developers and distro maintainers had an easy
> way to check for toolchain behaviors. When it was ported to KUnit, this
> mode was removed to simplify the code.
>
> Add a small header that provides a minimally operational KUnit API that
> can allow unit tests that don't depend on kernel-specific behaviors
> to build and run strictly from userspace without kernel sources. Add
> userspace-build support back to the renamed lib/stackinit_kunit.c test.
Very cool. I was hoping to achieve a true UAPI KUnit via LKL, but
that's clearly a long way out. Besides you have a point with being
able to distribute just a couple of files.
My initial thought is that a UAPI for KUnit would mostly be useful for
testing parts of the kernel that have very few dependencies. However,
I remembered that Bart (cc'ed) was working on some KUnit tests that
would probably be better as userspace tests. I think there might be
some potential for a KUnit UAPI in decoupling the test library from
the test environment.
I have been thinking off and on for a little while about enabling some
kind of KUnit kselftest fusion tests, where you could manage an
environment in kselftest and use KUnit for testing in kernel
structures. I think this is kind of going from the opposite direction,
but is probably an easier place to start.
I will refrain myself from getting into too many wild ideas, but I
like where this is going. I think it is reasonable to just start off
with a duplicate implementation of KUnit as you have done here: we can
see if this has any legs, and if so, we can do something more
complicated later.
> Signed-off-by: Kees Cook <keescook@...omium.org>
I have some thoughts for some things that can be improved, but after
thinking about it; I think they are all things that make sense once we
see if there are other potential users. I think as an initial version,
this looks pretty good.
Reviewed-by: Brendan Higgins <brendanhiggins@...gle.com>
(leaving the rest of the email unclipped for Bart's benefit)
> ---
> v1: https://lore.kernel.org/lkml/20220224055145.1853657-1-keescook@chromium.org
> v2:
> - split from stackinit_kunit.c refactoring patch
> - add missing returns (Daniel)
> - report expression mismatch in assert msg (Daniel)
> - emulate kunit_test_suites() (Daniel)
> - emit valid KTAP (David)
> ---
> include/uapi/misc/kunit.h | 181 ++++++++++++++++++++++++++++++++++++++
> lib/stackinit_kunit.c | 11 +++
> 2 files changed, 192 insertions(+)
> create mode 100644 include/uapi/misc/kunit.h
>
> diff --git a/include/uapi/misc/kunit.h b/include/uapi/misc/kunit.h
> new file mode 100644
> index 000000000000..afdffda583ae
> --- /dev/null
> +++ b/include/uapi/misc/kunit.h
> @@ -0,0 +1,181 @@
> +#ifndef __UAPI_MISC_KUNIT_H__
> +#define __UAPI_MISC_KUNIT_H__
> +/*
> + * This is a light-weight userspace drop-in replacement for the in-kernel
> + * KUnit API. It seeks to implement a minimal subset of features so that
> + * a concisely written KUnit test can be made to run entirely in userspace
> + * when it doesn't actually depend on any real kernel internals.
> + *
> + * Additionally contains many refactored kernel-isms to support building
> + * and running in userspace without full kernel source.
> + */
> +
> +#define _GNU_SOURCE
> +#include <stdio.h>
> +#include <stdint.h>
> +#include <stdlib.h>
> +#include <string.h>
> +#include <stdbool.h>
> +#include <errno.h>
> +#include <sys/types.h>
> +
> +#define __user /**/
> +#define noinline __attribute__((__noinline__))
> +#define __aligned(x) __attribute__((__aligned__(x)))
> +#ifdef __clang__
> +# define __compiletime_error(message) /**/
> +#else
> +# define __compiletime_error(message) __attribute__((__error__(message)))
> +#endif
> +#define __compiletime_assert(condition, msg, prefix, suffix) \
> + do { \
> + extern void prefix ## suffix(void) __compiletime_error(msg); \
> + if (!(condition)) \
> + prefix ## suffix(); \
> + } while (0)
> +#define _compiletime_assert(condition, msg, prefix, suffix) \
> + __compiletime_assert(condition, msg, prefix, suffix)
> +#define compiletime_assert(condition, msg) \
> + _compiletime_assert(condition, msg, __compiletime_assert_, __COUNTER__)
> +#define BUILD_BUG_ON_MSG(cond, msg) compiletime_assert(!(cond), msg)
> +#define BUILD_BUG_ON(condition) \
> + BUILD_BUG_ON_MSG(condition, "BUILD_BUG_ON failed: " #condition)
> +
> +#define ARRAY_SIZE(a) (sizeof (a) / sizeof ((a)[0]))
> +
> +#define MODULE_LICENSE(str) /* str */
> +
> +typedef uint8_t u8;
> +typedef uint16_t u16;
> +typedef uint32_t u32;
> +typedef uint64_t u64;
> +
> +#define TEST_PASS 0
> +#define TEST_SKIP 1
> +#define TEST_FAIL 2
> +struct kunit {
> + int status;
> + char *msg;
> +};
> +struct kunit_case {
> + void (*run_case)(struct kunit *test);
> + const char *name;
> +};
> +struct kunit_suite {
> + const char *name;
> + const struct kunit_case *test_cases;
> +};
> +#define KUNIT_CASE(test_name) { .run_case = test_name, .name = #test_name }
> +
> +#define KUNIT_ASSERT_TRUE_MSG(test, expr, fmt, ...) \
> +do { \
> + if (!(expr)) { \
> + if (test->status != TEST_SKIP) \
> + test->status = TEST_FAIL; \
> + if (test->msg) \
> + free(test->msg); \
> + asprintf(&test->msg, fmt, ##__VA_ARGS__); \
> + return; \
> + } \
> +} while (0)
> +
> +#define KUNIT_ASSERT_EQ_MSG(test, left, right, fmt, ...) \
> + KUNIT_ASSERT_TRUE_MSG(test, (left) == (right), \
> + #left " != " #right ": " fmt, \
> + ##__VA_ARGS__)
> +
> +#define kunit_skip(test, fmt, ...) \
> +do { \
> + test->status = TEST_SKIP; \
> + if (test->msg) \
> + free(test->msg); \
> + asprintf(&test->msg, fmt, ##__VA_ARGS__); \
> + return; \
> +} while (0)
> +
> +static int do_kunit_test_suite(struct kunit_suite *suite)
> +{
> + const struct kunit_case *test_case;
> + int pass = 0, fail = 0, skip = 0;
> + int rc = 0;
> + size_t i = 0;
> +
> + printf(" TAP version 14\n");
> + for (test_case = suite->test_cases; test_case->run_case; test_case++)
> + i++;
> + printf(" 1..%zu\n", i);
> + i = 0;
> + for (test_case = suite->test_cases; test_case->run_case; test_case++) {
> + struct kunit test = { };
> +
> + i++;
> + test_case->run_case(&test);
> + switch (test.status) {
> + default:
> + case TEST_FAIL:
> + fprintf(stderr, " not ok %zu - %s%s%s",
> + i, test_case->name,
> + test.msg ? " # ERROR " : "",
> + test.msg ?: "\n");
> + rc = 1;
> + fail++;
> + break;
> + case TEST_SKIP:
> + fprintf(stdout, " ok %zu - %s # SKIP%s%s",
> + i, test_case->name,
> + test.msg ? " " : "",
> + test.msg ?: "\n");
> + skip++;
> + break;
> + case TEST_PASS:
> + fprintf(stdout, " ok %zu - %s\n",
> + i, test_case->name);
> + pass++;
> + break;
> + }
> + if (test.msg)
> + free(test.msg);
> + }
> + printf("# %s: pass:%d fail:%d skip:%d total:%zu\n",
> + suite->name, pass, fail, skip, i);
> + return rc;
> +}
> +
> +static int run_suites(char *name, struct kunit_suite *suites[], size_t count)
> +{
> + int pass = 0, fail = 0, skip = 0;
> + int one, ret = 0;
> + size_t i;
> +
> + printf("TAP version 14\n");
> + printf("1..%zu\n", count);
> + for (i = 0; i < count; ++i) {
> + one = do_kunit_test_suite(suites[i]);
> + switch (one) {
> + case TEST_SKIP:
> + skip++;
> + break;
> + case TEST_PASS:
> + pass++;
> + break;
> + default:
> + fail++;
> + break;
> + }
> + printf("%sok %zu - %s\n",
> + one == TEST_FAIL ? "not " : "",
> + i + 1, suites[i]->name);
> + ret |= one;
> + }
> + printf("# %s: pass:%d fail:%d skip:%d total:%zu\n",
> + name, pass, fail, skip, count);
> + return ret;
> +}
> +
> +#define kunit_test_suites(suite...) \
> +int main(int argc, char *argv[]) { \
> + static struct kunit_suite *suites[] = { suite }; \
> + return run_suites(argv[0], suites, ARRAY_SIZE(suites)); \
> +}
> +
> +#endif /* __UAPI_MISC_KUNIT_H__ */
> diff --git a/lib/stackinit_kunit.c b/lib/stackinit_kunit.c
> index 35c69aa425b2..6d468630c90a 100644
> --- a/lib/stackinit_kunit.c
> +++ b/lib/stackinit_kunit.c
> @@ -8,7 +8,13 @@
> * --make_option LLVM=1 \
> * --kconfig_add CONFIG_INIT_STACK_ALL_ZERO=y
> *
> + * External build example:
> + * clang -O2 -Wall -ftrivial-auto-var-init=pattern \
> + * -o stackinit_kunit stackinit_kunit.c
> + * ./stackinit_kunit
> + *
> */
> +#ifdef __KERNEL__
> #define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
>
> #include <kunit/test.h>
> @@ -17,6 +23,11 @@
> #include <linux/module.h>
> #include <linux/string.h>
>
> +#else
> +/* Userspace KUnit stub header. */
> +#include <misc/kunit.h>
> +#endif
> +
> /* Exfiltration buffer. */
> #define MAX_VAR_SIZE 128
> static u8 check_buf[MAX_VAR_SIZE];
> --
> 2.32.0
>
> --
> You received this message because you are subscribed to the Google Groups "KUnit Development" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to kunit-dev+unsubscribe@...glegroups.com.
> To view this discussion on the web visit https://groups.google.com/d/msgid/kunit-dev/20220227184517.504931-8-keescook%40chromium.org.
Powered by blists - more mailing lists