linux-kernel - [PATCH] selftests/mm: Add commentary about 9pfs bugs

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [day] [month] [year] [list]

Message-ID: <20250317-9pfs-comments-v1-1-9ac96043e146@google.com>
Date: Mon, 17 Mar 2025 10:20:34 +0000
From: Brendan Jackman <jackmanb@...gle.com>
To: Andrew Morton <akpm@...ux-foundation.org>, Shuah Khan <shuah@...nel.org>
Cc: linux-mm@...ck.org, linux-kselftest@...r.kernel.org, 
	linux-kernel@...r.kernel.org, Brendan Jackman <jackmanb@...gle.com>, 
	David Hildenbrand <david@...hat.com>
Subject: [PATCH] selftests/mm: Add commentary about 9pfs bugs

As discussed here:

https://lore.kernel.org/lkml/Z9RRkL1hom48z3Tt@google.com/

This code could benefit from some more commentary.

To avoid needing to comment the same thing in multiple places (I guess
more of these SKIPs will need to be added over time, for now I am only
like 20% of the way through Project Run run_vmtests.sh Successfully),
add a dummy "skip tests for this specific reason" function that
basically just serves as a hook to hang comments on.

Signed-off-by: Brendan Jackman <jackmanb@...gle.com>
---
To: David Hildenbrand <david@...hat.com>
---
 tools/testing/selftests/mm/gup_longterm.c |  6 +-----
 tools/testing/selftests/mm/map_populate.c |  8 +++-----
 tools/testing/selftests/mm/vm_util.h      | 18 ++++++++++++++++++
 3 files changed, 22 insertions(+), 10 deletions(-)

diff --git a/tools/testing/selftests/mm/gup_longterm.c b/tools/testing/selftests/mm/gup_longterm.c
index 03271442aae5aed060fd44010df552a2eedcdafc..21595b20bbc391a0e5d0ab0563ac4ce5e1e0069f 100644
--- a/tools/testing/selftests/mm/gup_longterm.c
+++ b/tools/testing/selftests/mm/gup_longterm.c
@@ -97,11 +97,7 @@ static void do_test(int fd, size_t size, enum test_type type, bool shared)
 
 	if (ftruncate(fd, size)) {
 		if (errno == ENOENT) {
-			/*
-			 * This can happen if the file has been unlinked and the
-			 * filesystem doesn't support truncating unlinked files.
-			 */
-			ksft_test_result_skip("ftruncate() failed with ENOENT\n");
+			skip_test_dodgy_fs("ftruncate()");
 		} else {
 			ksft_test_result_fail("ftruncate() failed (%s)\n", strerror(errno));
 		}
diff --git a/tools/testing/selftests/mm/map_populate.c b/tools/testing/selftests/mm/map_populate.c
index 433e54fb634f793f2eb4c53ba6b791045c9f4986..9df2636c829bf34d6d0517e126b3deda1f3ba834 100644
--- a/tools/testing/selftests/mm/map_populate.c
+++ b/tools/testing/selftests/mm/map_populate.c
@@ -18,6 +18,8 @@
 #include <unistd.h>
 #include "../kselftest.h"
 
+#include "vm_util.h"
+
 #define MMAP_SZ		4096
 
 #define BUG_ON(condition, description)						\
@@ -88,11 +90,7 @@ int main(int argc, char **argv)
 
 	ret = ftruncate(fileno(ftmp), MMAP_SZ);
 	if (ret < 0 && errno == ENOENT) {
-		/*
-		 * This probably means tmpfile() made a file on a filesystem
-		 * that doesn't handle temporary files the way we want.
-		 */
-		ksft_exit_skip("ftruncate(fileno(tmpfile())) gave ENOENT, weird filesystem?\n");
+		skip_test_dodgy_fs("ftruncate()");
 	}
 	BUG_ON(ret, "ftruncate()");
 
diff --git a/tools/testing/selftests/mm/vm_util.h b/tools/testing/selftests/mm/vm_util.h
index 0e629586556b5aae580d8e4ce7491bc93adcc4d6..6effafdc4d8a23f91f0adcb9e43d6196d651ba88 100644
--- a/tools/testing/selftests/mm/vm_util.h
+++ b/tools/testing/selftests/mm/vm_util.h
@@ -5,6 +5,7 @@
 #include <err.h>
 #include <strings.h> /* ffsl() */
 #include <unistd.h> /* _SC_PAGESIZE */
+#include "../kselftest.h"
 
 #define BIT_ULL(nr)                   (1ULL << (nr))
 #define PM_SOFT_DIRTY                 BIT_ULL(55)
@@ -32,6 +33,23 @@ static inline unsigned int pshift(void)
 	return __page_shift;
 }
 
+/*
+ * Plan 9 FS has bugs (at least on QEMU) where certain operations fail with
+ * ENOENT on unlinked files. See
+ * https://gitlab.com/qemu-project/qemu/-/issues/103 for some info about such
+ * bugs. There are rumours of NFS implementations with similar bugs.
+ *
+ * Ideally, tests should just detect filesystems known to have such issues and
+ * bail early. But 9pfs has the additional "feature" that it causes fstatfs to
+ * pass through the f_type field from the host filesystem. To avoid having to
+ * scrape /proc/mounts or some other hackery, tests can call this function when
+ * it seems such a bug might have been encountered.
+ */
+static inline void skip_test_dodgy_fs(const char *op_name)
+{
+	ksft_test_result_skip("%s failed with ENOENT. Filesystem might be buggy (9pfs?)\n", op_name);
+}
+
 uint64_t pagemap_get_entry(int fd, char *start);
 bool pagemap_is_softdirty(int fd, char *start);
 bool pagemap_is_swapped(int fd, char *start);

---
base-commit: a91aaf8dd549dcee9caab227ecaa6cbc243bbc5a
change-id: 20250317-9pfs-comments-24b6fa5417cd

Best regards,
-- 
Brendan Jackman <jackmanb@...gle.com>