[LTP] [PATCH] refluxfs: Check kernel reflink support before mount

[LTP] [PATCH] refluxfs: Check kernel reflink support before mount

[Martin Doucha]

On some systems, mkfs.xfs can format XFS partition with reflink support
but the kernel then cannot mount it. Check for kernel reflink support
in setup() before mount.

Signed-off-by: Martin Doucha <mdoucha@suse.cz>
---
 testcases/cve/refluxfs.c | 18 ++++++++++++++++--
 1 file changed, 16 insertions(+), 2 deletions(-)

diff --git a/testcases/cve/refluxfs.c b/testcases/cve/refluxfs.c
index e41f35c53..277d56bb7 100644
--- a/testcases/cve/refluxfs.c
+++ b/testcases/cve/refluxfs.c
@@ -38,6 +38,7 @@
  */
 
 #include <pwd.h>
+#include <sys/mount.h>
 
 #include "tst_test.h"
 #include "tst_safe_prw.h"
@@ -94,6 +95,15 @@ static void setup(void)
 	int probe_fd, probe_dio_fd;
 	struct stat sb;
 
+	SAFE_MKDIR(MNTPOINT, 0755);
+	TEST(mount(tst_device->dev, MNTPOINT, tst_device->fs_type, 0, NULL));
+
+	if (TST_RET == -1 && TST_ERR == EOPNOTSUPP)
+		tst_brk(TCONF, "Kernel does not support XFS reflinks");
+
+	if (TST_RET)
+		tst_brk(TBROK | TTERRNO, "Mount failed");
+
 	SAFE_STAT(MNTPOINT, &sb);
 	blksize = sb.st_blksize;
 
@@ -197,6 +207,11 @@ static void cleanup(void)
 	free(tbuf);
 	free(wbuf);
 	free(rbuf);
+
+	SAFE_SETEUID(0);
+
+	if (tst_is_mounted(MNTPOINT))
+		SAFE_UMOUNT(MNTPOINT);
 }
 
 static struct tst_test test = {
@@ -205,8 +220,7 @@ static struct tst_test test = {
 	.cleanup = cleanup,
 	.runtime = 180,
 	.needs_root = 1,
-	.mount_device = 1,
-	.mntpoint = MNTPOINT,
+	.format_device = 1,
 	.filesystems = (struct tst_fs []) {
 		{
 			.type = "xfs",
-- 
2.54.0


-- 
Mailing list info: https://lists.linux.it/listinfo/ltp

Re: [LTP] [PATCH] refluxfs: Check kernel reflink support before mount

[Cyril Hrubis]

Hi!
> +	if (TST_RET == -1 && TST_ERR == EOPNOTSUPP)
> +		tst_brk(TCONF, "Kernel does not support XFS reflinks");

Wouldn't it make more sense to add the check for EOPNOTSUPP into the
tst_test.c library?

-- 
Cyril Hrubis
chrubis@suse.cz

-- 
Mailing list info: https://lists.linux.it/listinfo/ltp

Re: [LTP] [PATCH] refluxfs: Check kernel reflink support before mount

[Andrea Cervesato via ltp]

Hi Cyril,

> > +	if (TST_RET == -1 && TST_ERR == EOPNOTSUPP)
> > +		tst_brk(TCONF, "Kernel does not support XFS reflinks");
> 
> Wouldn't it make more sense to add the check for EOPNOTSUPP into the
> tst_test.c library?

I don't know if we really need to add something in the LTP library,
since we have this check only in file_attr02

--
Andrea Cervesato
SUSE QE Automation Engineer Linux
andrea.cervesato@suse.com

-- 
Mailing list info: https://lists.linux.it/listinfo/ltp

Re: [LTP] [PATCH] refluxfs: Check kernel reflink support before mount

[Cyril Hrubis]

Hi!
> > > +	if (TST_RET == -1 && TST_ERR == EOPNOTSUPP)
> > > +		tst_brk(TCONF, "Kernel does not support XFS reflinks");
> > 
> > Wouldn't it make more sense to add the check for EOPNOTSUPP into the
> > tst_test.c library?
> 
> I don't know if we really need to add something in the LTP library,
> since we have this check only in file_attr02

I mean that we should rather than hacking around it in the tests handle
EOPNOTSUPP properly in the test libray when mounting filesystems. We do
enough magick in safe_mount() to support FUSE and handle other corner
cases, checking for EOPNOTSUPP (possibly with non-standard mount flags)
does sound like a reasonable extension.

The main point of handling the filesystems in the test library is that
cleanup (umounting) is handled properly even if the test crashes. If we
move the code back to the tests like this we lose that.

-- 
Cyril Hrubis
chrubis@suse.cz

-- 
Mailing list info: https://lists.linux.it/listinfo/ltp

Re: [LTP] [PATCH] refluxfs: Check kernel reflink support before mount

[Andrea Cervesato via ltp]

Hi Cyril,

> Hi!
> > > > +	if (TST_RET == -1 && TST_ERR == EOPNOTSUPP)
> > > > +		tst_brk(TCONF, "Kernel does not support XFS reflinks");
> > > 
> > > Wouldn't it make more sense to add the check for EOPNOTSUPP into the
> > > tst_test.c library?
> > 
> > I don't know if we really need to add something in the LTP library,
> > since we have this check only in file_attr02
> 
> I mean that we should rather than hacking around it in the tests handle
> EOPNOTSUPP properly in the test libray when mounting filesystems. We do
> enough magick in safe_mount() to support FUSE and handle other corner
> cases, checking for EOPNOTSUPP (possibly with non-standard mount flags)
> does sound like a reasonable extension.
> 
> The main point of handling the filesystems in the test library is that
> cleanup (umounting) is handled properly even if the test crashes. If we
> move the code back to the tests like this we lose that.
> 
> -- 
> Cyril Hrubis
> chrubis@suse.cz

Something like this?

diff --git a/lib/safe_macros.c b/lib/safe_macros.c
index f95c5fdc5..34430d018 100644
--- a/lib/safe_macros.c
+++ b/lib/safe_macros.c
@@ -993,9 +993,15 @@ int safe_mount(const char *file, const int lineno, void (*cleanup_fn)(void),
 			"mount.%s failed with %i", filesystemtype, rval);
 		return -1;
 	} else if (rval == -1) {
-		tst_brkm_(file, lineno, TBROK | TERRNO, cleanup_fn,
-			"mount(%s, %s, %s, %lu, %p) failed", source, target,
-			filesystemtype, mountflags, data);
+		if (errno == EOPNOTSUPP) {
+			tst_brkm_(file, lineno, TCONF | TERRNO, cleanup_fn,
+				"mount(%s, %s, %s, %lu, %p) failed with EOPNOTSUPP", source, target,
+				filesystemtype, mountflags, data);
+		} else {
+			tst_brkm_(file, lineno, TBROK | TERRNO, cleanup_fn,
+				"mount(%s, %s, %s, %lu, %p) failed", source, target,
+				filesystemtype, mountflags, data);
+		}
 	} else {
 		tst_brkm_(file, lineno, TBROK | TERRNO, cleanup_fn,
 			"Invalid mount(%s, %s, %s, %lu, %p) return value %d",
diff --git a/testcases/kernel/syscalls/file_attr/file_attr02.c b/testcases/kernel/syscalls/file_attr/file_attr02.c
index f6625985a..28c1c7ca6 100644
--- a/testcases/kernel/syscalls/file_attr/file_attr02.c
+++ b/testcases/kernel/syscalls/file_attr/file_attr02.c
@@ -45,13 +45,7 @@ static void setup(void)
 	struct stat statbuf;
 
 	SAFE_MKDIR(MNTPOINT, 0755);
-	TEST(mount(tst_device->dev, MNTPOINT, tst_device->fs_type, 0, NULL));
-
-	if (TST_RET == -1 && TST_ERR == EOPNOTSUPP)
-		tst_brk(TCONF, "Kernel does not support XFS reflinks");
-
-	if (TST_RET)
-		tst_brk(TBROK | TTERRNO, "Mount failed");
+	SAFE_MOUNT(tst_device->dev, MNTPOINT, tst_device->fs_type, 0, NULL);
 
 	SAFE_STAT(MNTPOINT, &statbuf);
 

--
Andrea Cervesato
SUSE QE Automation Engineer Linux
andrea.cervesato@suse.com

-- 
Mailing list info: https://lists.linux.it/listinfo/ltp

Re: [LTP] [PATCH] refluxfs: Check kernel reflink support before mount

[Cyril Hrubis]

Hi!
> Something like this?
>
> diff --git a/lib/safe_macros.c b/lib/safe_macros.c
> index f95c5fdc5..34430d018 100644
> --- a/lib/safe_macros.c
> +++ b/lib/safe_macros.c
> @@ -993,9 +993,15 @@ int safe_mount(const char *file, const int lineno, void (*cleanup_fn)(void),
>  			"mount.%s failed with %i", filesystemtype, rval);
>  		return -1;
>  	} else if (rval == -1) {
> -		tst_brkm_(file, lineno, TBROK | TERRNO, cleanup_fn,
> -			"mount(%s, %s, %s, %lu, %p) failed", source, target,
> -			filesystemtype, mountflags, data);
> +		if (errno == EOPNOTSUPP) {
> +			tst_brkm_(file, lineno, TCONF | TERRNO, cleanup_fn,
> +				"mount(%s, %s, %s, %lu, %p) failed with EOPNOTSUPP", source, target,
> +				filesystemtype, mountflags, data);
> +		} else {
> +			tst_brkm_(file, lineno, TBROK | TERRNO, cleanup_fn,
> +				"mount(%s, %s, %s, %lu, %p) failed", source, target,
> +				filesystemtype, mountflags, data);
> +		}
>  	} else {
>  		tst_brkm_(file, lineno, TBROK | TERRNO, cleanup_fn,
>  			"Invalid mount(%s, %s, %s, %lu, %p) return value %d",
> diff --git a/testcases/kernel/syscalls/file_attr/file_attr02.c b/testcases/kernel/syscalls/file_attr/file_attr02.c
> index f6625985a..28c1c7ca6 100644
> --- a/testcases/kernel/syscalls/file_attr/file_attr02.c
> +++ b/testcases/kernel/syscalls/file_attr/file_attr02.c
> @@ -45,13 +45,7 @@ static void setup(void)
>  	struct stat statbuf;
>  
>  	SAFE_MKDIR(MNTPOINT, 0755);
> -	TEST(mount(tst_device->dev, MNTPOINT, tst_device->fs_type, 0, NULL));
> -
> -	if (TST_RET == -1 && TST_ERR == EOPNOTSUPP)
> -		tst_brk(TCONF, "Kernel does not support XFS reflinks");
> -
> -	if (TST_RET)
> -		tst_brk(TBROK | TTERRNO, "Mount failed");
> +	SAFE_MOUNT(tst_device->dev, MNTPOINT, tst_device->fs_type, 0, NULL);
>  
>  	SAFE_STAT(MNTPOINT, &statbuf);

Yes. If we do not do this we are going to open code the check over and
over.

-- 
Cyril Hrubis
chrubis@suse.cz

-- 
Mailing list info: https://lists.linux.it/listinfo/ltp

Re: [LTP] [PATCH] refluxfs: Check kernel reflink support before mount

[Martin Doucha]

On 8/13/26 12:46, Andrea Cervesato wrote:
> Hi Cyril,
> 
> Something like this?
> 
> diff --git a/lib/safe_macros.c b/lib/safe_macros.c
> index f95c5fdc5..34430d018 100644
> --- a/lib/safe_macros.c
> +++ b/lib/safe_macros.c
> @@ -993,9 +993,15 @@ int safe_mount(const char *file, const int lineno, void (*cleanup_fn)(void),
>   			"mount.%s failed with %i", filesystemtype, rval);
>   		return -1;
>   	} else if (rval == -1) {
> -		tst_brkm_(file, lineno, TBROK | TERRNO, cleanup_fn,
> -			"mount(%s, %s, %s, %lu, %p) failed", source, target,
> -			filesystemtype, mountflags, data);
> +		if (errno == EOPNOTSUPP) {
> +			tst_brkm_(file, lineno, TCONF | TERRNO, cleanup_fn,
> +				"mount(%s, %s, %s, %lu, %p) failed with EOPNOTSUPP", source, target,
> +				filesystemtype, mountflags, data);
> +		} else {
> +			tst_brkm_(file, lineno, TBROK | TERRNO, cleanup_fn,
> +				"mount(%s, %s, %s, %lu, %p) failed", source, target,
> +				filesystemtype, mountflags, data);
> +		}
>   	} else {
>   		tst_brkm_(file, lineno, TBROK | TERRNO, cleanup_fn,
>   			"Invalid mount(%s, %s, %s, %lu, %p) return value %d",

Hi,
yes, but I think it'd be better to add a setting to struct tst_test to 
enable this TCONF check only for specific tests. By default, EOPNOTSUPP 
should still trigger TBROK.

Something like:
.mount_check_support = 1

-- 
Martin Doucha   mdoucha@suse.cz
SW Quality Engineer
SUSE LINUX, s.r.o.
CORSO IIa
Krizikova 148/34
186 00 Prague 8
Czech Republic

-- 
Mailing list info: https://lists.linux.it/listinfo/ltp

Re: [LTP] [PATCH] refluxfs: Check kernel reflink support before mount

[Cyril Hrubis]

Hi!
> > @@ -993,9 +993,15 @@ int safe_mount(const char *file, const int lineno, void (*cleanup_fn)(void),
> >   			"mount.%s failed with %i", filesystemtype, rval);
> >   		return -1;
> >   	} else if (rval == -1) {
> > -		tst_brkm_(file, lineno, TBROK | TERRNO, cleanup_fn,
> > -			"mount(%s, %s, %s, %lu, %p) failed", source, target,
> > -			filesystemtype, mountflags, data);
> > +		if (errno == EOPNOTSUPP) {
> > +			tst_brkm_(file, lineno, TCONF | TERRNO, cleanup_fn,
> > +				"mount(%s, %s, %s, %lu, %p) failed with EOPNOTSUPP", source, target,
> > +				filesystemtype, mountflags, data);
> > +		} else {
> > +			tst_brkm_(file, lineno, TBROK | TERRNO, cleanup_fn,
> > +				"mount(%s, %s, %s, %lu, %p) failed", source, target,
> > +				filesystemtype, mountflags, data);
> > +		}
> >   	} else {
> >   		tst_brkm_(file, lineno, TBROK | TERRNO, cleanup_fn,
> >   			"Invalid mount(%s, %s, %s, %lu, %p) return value %d",
> 
> Hi,
> yes, but I think it'd be better to add a setting to struct tst_test to 
> enable this TCONF check only for specific tests. By default, EOPNOTSUPP 
> should still trigger TBROK.
> 
> Something like:
> .mount_check_support = 1

Shoudn't that be per filesystem and live in tst_fs structure?

Either way we will need a flag to the safe_mount() added first...

-- 
Cyril Hrubis
chrubis@suse.cz

-- 
Mailing list info: https://lists.linux.it/listinfo/ltp

Re: [LTP] [PATCH] refluxfs: Check kernel reflink support before mount

[Martin Doucha]

On 8/13/26 16:15, Cyril Hrubis wrote:
>> Hi,
>> yes, but I think it'd be better to add a setting to struct tst_test to
>> enable this TCONF check only for specific tests. By default, EOPNOTSUPP
>> should still trigger TBROK.
>>
>> Something like:
>> .mount_check_support = 1
> 
> Shoudn't that be per filesystem and live in tst_fs structure?
> 
> Either way we will need a flag to the safe_mount() added first...

That's a good point. I'll prepare a patch.

-- 
Martin Doucha   mdoucha@suse.cz
SW Quality Engineer
SUSE LINUX, s.r.o.
CORSO IIa
Krizikova 148/34
186 00 Prague 8
Czech Republic

-- 
Mailing list info: https://lists.linux.it/listinfo/ltp

Re: [LTP] refluxfs: Check kernel reflink support before mount

[linuxtestproject.agent]

Hi Martin,

On Aug 12, 2026, Martin Doucha wrote:
> refluxfs: Check kernel reflink support before mount

Verdict - Reviewed

---
Note:

The agent can sometimes produce false positives although often its
findings are genuine. If you find issues with the review, please
comment this email or ignore the suggestions.

Regards,
LTP AI Reviewer

-- 
Mailing list info: https://lists.linux.it/listinfo/ltp

Re: [LTP] [PATCH] refluxfs: Check kernel reflink support before mount

[Andrea Cervesato via ltp]

Reviewed-by: Andrea Cervesato <andrea.cervesato@suse.com>

--
Andrea Cervesato
SUSE QE Automation Engineer Linux
andrea.cervesato@suse.com

-- 
Mailing list info: https://lists.linux.it/listinfo/ltp