On 10/13/21 2:55 PM, Kirill A. Shutemov wrote:

On Tue, Oct 12, 2021 at 07:26:44PM +0100, Matthew Wilcox (Oracle) wrote:

quoted

Even unknown flags should be passed to %pGp so that we can test what it
does with them (today: nothing).

Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
---
 lib/test_printf.c | 6 ++----
 1 file changed, 2 insertions(+), 4 deletions(-)

diff --git a/lib/test_printf.c b/lib/test_printf.c
index a52c1c3a55ba..f744b0498672 100644
--- a/lib/test_printf.c
+++ b/lib/test_printf.c

@@ -608,14 +608,12 @@ page_flags_test(int section, int node, int zone, int last_cpupid,
 		int kasan_tag, int flags, const char *name, char *cmp_buf)
 {
 	unsigned long values[] = {section, node, zone, last_cpupid, kasan_tag};
-	unsigned long page_flags = 0;
+	unsigned long page_flags = flags;
 	unsigned long size = 0;
 	bool append = false;
 	int i;
 
-	flags &= PAGEFLAGS_MASK;
-	if (flags) {
-		page_flags |= flags;
+	if (flags & PAGEFLAGS_MASK) {
 		snprintf(cmp_buf + size, BUF_SIZE - size, "%s", name);
 		size = strlen(cmp_buf);
 #if SECTIONS_WIDTH || NODES_WIDTH || ZONES_WIDTH || \

Do we even need 'page_flags'? Can't we do everything on 'flags' directly?

Agreed, +1. Otherwise the change LGTM.