[PATCH RFC 11/23] selftests/proc: check that a faulting writer doesn't hold up removal

sashiko-bot at kernel.org sashiko-bot at kernel.org
Tue Nov 3 07:18:48 PST 2026


Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Medium] selftests/proc: fail and leak scsi_debug host without CONFIG_SCSI_PROC_FS
--

commit d0cceed0cde591b286c7977ad630183fc653ded2
Author: Christian Brauner <brauner at kernel.org>

selftests/proc: check that a faulting writer doesn't hold up removal

This commit adds a procfs selftest that uses userfaultfd and scsi_debug to
verify that a writer faulting during copy-in does not block removal of a
/proc/scsi/scsi_debug/<host> entry and fails with -EIO once the fault is
resolved.

> diff --git a/tools/testing/selftests/proc/proc-remove-slow-writer.c b/tools/testing/selftests/proc/proc-remove-slow-writer.c
> new file mode 100644
> index 0000000000000..f69ca75ba87cb
> --- /dev/null
> +++ b/tools/testing/selftests/proc/proc-remove-slow-writer.c
> @@ -0,0 +1,203 @@

[ ... ]

> +/* the numbers of the hosts in /proc/scsi/scsi_debug as a bitmap */
> +static int list_hosts(unsigned long *hosts)
> +{
> +	struct dirent *d;
> +	DIR *dir;
> +
> +	memset(hosts, 0, MAX_HOSTS / 8);
> +	dir = opendir(HOST_DIR);
> +	if (!dir)
> +		/* the directory comes with the first host and goes with the last */
> +		return errno == ENOENT ? 0 : -errno;

[ ... ]

> +FIXTURE_SETUP(host)
> +{
> +	unsigned long before[MAX_HOSTS / HOST_BITS];
> +	unsigned long after[MAX_HOSTS / HOST_BITS];
> +
> +	self->host = -1;
> +	if (access(ADD_HOST, W_OK) && system("modprobe -q scsi_debug add_host=0"))
> +		SKIP(return, "no scsi_debug");
> +	if (access(ADD_HOST, W_OK))
> +		SKIP(return, ADD_HOST ": %s", strerror(errno));
> +
> +	/* host numbers are reused, so look for the one that's new */
> +	ASSERT_EQ(list_hosts(before), 0);
> +	ASSERT_EQ(add_host(1), 0);
> +	ASSERT_EQ(list_hosts(after), 0);
> +	self->host = new_host(before, after);
> +	ASSERT_GE(self->host, 0);
> +	snprintf(self->path, sizeof(self->path), HOST_DIR "/%d", self->host);
> +}
> +
> +FIXTURE_TEARDOWN(host)
> +{
> +	if (self->host >= 0 && !access(self->path, F_OK))
> +		add_host(-1);
> +}

[Severity: Medium]
What happens if the test runs on a kernel where scsi_debug is available
(CONFIG_SCSI_DEBUG is enabled), but CONFIG_SCSI_PROC_FS is disabled?

The sysfs attribute /sys/bus/pseudo/drivers/scsi_debug/add_host is still
present and writable without CONFIG_SCSI_PROC_FS, while /proc/scsi is never
created. In FIXTURE_SETUP(host), list_hosts(before) sees ENOENT on
/proc/scsi/scsi_debug and returns 0, add_host(1) creates a new scsi_debug
host in the kernel, and list_hosts(after) still sees ENOENT and returns 0
with an empty bitmap.

As a result, new_host(before, after) returns -1 and ASSERT_GE(self->host, 0)
fails the test instead of skipping for the missing prerequisite. Because
FIXTURE_SETUP(host) aborts with self->host still set to -1, the newly
created scsi_debug host is never removed by add_host(-1) and remains leaked.

Should FIXTURE_SETUP(host) check that /proc/scsi exists and skip before
calling add_host(1)?

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20261102-work-procfs-drivers-v1-0-a994bd4d3b1a@kernel.org?part=11



More information about the kexec mailing list