selftests/seccomp: Avoid redundant register flushes

When none of the registers have changed, don't flush them back. This can
happen if the architecture uses a non-register way to change the syscall
(e.g. arm64) , and a return value hasn't been written.

Signed-off-by: Kees Cook <keescook@chromium.org>
Link: https://lore.kernel.org/lkml/20200912110820.597135-11-keescook@chromium.org
Acked-by: Christian Brauner <christian.brauner@ubuntu.com>

diff --git a/tools/testing/selftests/seccomp/seccomp_bpf.c b/tools/testing/selftests/seccomp/seccomp_bpf.c
index 00056e0..638cea8 100644 (file)
--- a/tools/testing/selftests/seccomp/seccomp_bpf.c
+++ b/tools/testing/selftests/seccomp/seccomp_bpf.c
@@ -1859,11 +1859,12 @@ int get_syscall(struct __test_metadata *_metadata, pid_t tracee)
 void change_syscall(struct __test_metadata *_metadata,
                    pid_t tracee, int syscall, int result)
 {
-       ARCH_REGS regs;
+       ARCH_REGS orig, regs;
 
        EXPECT_EQ(0, ARCH_GETREGS(regs)) {
                return;
        }
+       orig = regs;
 
        SYSCALL_NUM_SET(regs, syscall);
 
@@ -1876,7 +1877,8 @@ void change_syscall(struct __test_metadata *_metadata,
 #endif
 
        /* Flush any register changes made. */
-       EXPECT_EQ(0, ARCH_SETREGS(regs));
+       if (memcmp(&orig, &regs, sizeof(orig)) != 0)
+               EXPECT_EQ(0, ARCH_SETREGS(regs));
 }
 
 void tracer_seccomp(struct __test_metadata *_metadata, pid_t tracee,