From: Ming Lei <ming.lei@redhat.com>
To: Jens Axboe <axboe@kernel.dk>, linux-block@vger.kernel.org
Cc: Caleb Sander Mateos <csander@purestorage.com>,
Uday Shankar <ushankar@purestorage.com>,
Yoav Cohen <yoav@nvidia.com>, Ming Lei <ming.lei@redhat.com>
Subject: [PATCH 2/3] selftests/ublk: add test for async partition scan
Date: Mon, 22 Dec 2025 00:41:42 +0800 [thread overview]
Message-ID: <20251221164145.1703448-3-ming.lei@redhat.com> (raw)
In-Reply-To: <20251221164145.1703448-1-ming.lei@redhat.com>
Add test_generic_15.sh to verify that async partition scan prevents
IO hang when reading partition tables.
The test creates ublk devices with fault_inject target and very large
delay (60s) to simulate blocked partition table reads, then kills the
daemon to verify proper state transitions without hanging:
1. Without recovery support (-r 0):
- Create device with fault_inject and 60s delay
- Kill daemon while partition scan may be blocked
- Verify device transitions to DEAD state within 10s
2. With recovery support (-r 1):
- Create device with fault_inject, 60s delay, and recovery
- Kill daemon while partition scan may be blocked
- Verify device transitions to QUIESCED state within 10s
Before the async partition scan fix, killing the daemon during
partition scan would cause deadlock as partition scan held ub->mutex
while waiting for IO. With the async fix, partition scan happens in
a work function and flush_work() ensures proper synchronization.
Add _add_ublk_dev_no_settle() helper function to skip udevadm settle,
which would otherwise hang waiting for partition scan events to
complete when partition table read is delayed.
Signed-off-by: Ming Lei <ming.lei@redhat.com>
---
tools/testing/selftests/ublk/Makefile | 1 +
tools/testing/selftests/ublk/test_common.sh | 16 +++-
.../testing/selftests/ublk/test_generic_15.sh | 80 +++++++++++++++++++
3 files changed, 93 insertions(+), 4 deletions(-)
create mode 100755 tools/testing/selftests/ublk/test_generic_15.sh
diff --git a/tools/testing/selftests/ublk/Makefile b/tools/testing/selftests/ublk/Makefile
index 837977b62417..eb0e6cfb00ad 100644
--- a/tools/testing/selftests/ublk/Makefile
+++ b/tools/testing/selftests/ublk/Makefile
@@ -22,6 +22,7 @@ TEST_PROGS += test_generic_11.sh
TEST_PROGS += test_generic_12.sh
TEST_PROGS += test_generic_13.sh
TEST_PROGS += test_generic_14.sh
+TEST_PROGS += test_generic_15.sh
TEST_PROGS += test_null_01.sh
TEST_PROGS += test_null_02.sh
diff --git a/tools/testing/selftests/ublk/test_common.sh b/tools/testing/selftests/ublk/test_common.sh
index 6f1c042de40e..ea9a5f3eb70a 100755
--- a/tools/testing/selftests/ublk/test_common.sh
+++ b/tools/testing/selftests/ublk/test_common.sh
@@ -178,8 +178,9 @@ _have_feature()
_create_ublk_dev() {
local dev_id;
local cmd=$1
+ local settle=$2
- shift 1
+ shift 2
if [ ! -c /dev/ublk-control ]; then
return ${UBLK_SKIP_CODE}
@@ -194,7 +195,10 @@ _create_ublk_dev() {
echo "fail to add ublk dev $*"
return 255
fi
- udevadm settle
+
+ if [ "$settle" = "yes" ]; then
+ udevadm settle
+ fi
if [[ "$dev_id" =~ ^[0-9]+$ ]]; then
echo "${dev_id}"
@@ -204,14 +208,18 @@ _create_ublk_dev() {
}
_add_ublk_dev() {
- _create_ublk_dev "add" "$@"
+ _create_ublk_dev "add" "yes" "$@"
+}
+
+_add_ublk_dev_no_settle() {
+ _create_ublk_dev "add" "no" "$@"
}
_recover_ublk_dev() {
local dev_id
local state
- dev_id=$(_create_ublk_dev "recover" "$@")
+ dev_id=$(_create_ublk_dev "recover" "yes" "$@")
for ((j=0;j<20;j++)); do
state=$(_get_ublk_dev_state "${dev_id}")
[ "$state" == "LIVE" ] && break
diff --git a/tools/testing/selftests/ublk/test_generic_15.sh b/tools/testing/selftests/ublk/test_generic_15.sh
new file mode 100755
index 000000000000..c241e641c340
--- /dev/null
+++ b/tools/testing/selftests/ublk/test_generic_15.sh
@@ -0,0 +1,80 @@
+#!/bin/bash
+# SPDX-License-Identifier: GPL-2.0
+
+. "$(cd "$(dirname "$0")" && pwd)"/test_common.sh
+
+TID="generic_15"
+ERR_CODE=0
+
+_test_partition_scan_no_hang()
+{
+ local recovery_flag=$1
+ local expected_state=$2
+ local dev_id
+ local state
+ local daemon_pid
+ local start_time
+ local elapsed
+
+ # Create ublk device with fault_inject target and very large delay
+ # to simulate hang during partition table read
+ # --delay_us 60000000 = 60 seconds delay
+ # Use _add_ublk_dev_no_settle to avoid udevadm settle hang waiting
+ # for partition scan events to complete
+ if [ "$recovery_flag" = "yes" ]; then
+ echo "Testing partition scan with recovery support..."
+ dev_id=$(_add_ublk_dev_no_settle -t fault_inject -q 1 -d 1 --delay_us 60000000 -r 1)
+ else
+ echo "Testing partition scan without recovery..."
+ dev_id=$(_add_ublk_dev_no_settle -t fault_inject -q 1 -d 1 --delay_us 60000000)
+ fi
+
+ _check_add_dev "$TID" $?
+
+ # The add command should return quickly because partition scan is async.
+ # Now sleep briefly to let the async partition scan work start and hit
+ # the delay in the fault_inject handler.
+ sleep 1
+
+ # Kill the ublk daemon while partition scan is potentially blocked
+ start_time=${SECONDS}
+ daemon_pid=$(_get_ublk_daemon_pid "${dev_id}")
+
+ # Kill daemon and check state transitions properly
+ state=$(__ublk_kill_daemon "${dev_id}" "${expected_state}")
+
+ elapsed=$((SECONDS - start_time))
+
+ # Verify the device transitioned to expected state
+ if [ "$state" != "${expected_state}" ]; then
+ echo "FAIL: Device state is $state, expected ${expected_state}"
+ ERR_CODE=255
+ ${UBLK_PROG} del -n "${dev_id}" > /dev/null 2>&1
+ return
+ fi
+
+ # Verify state transition happened within reasonable time (< 10s)
+ # This ensures we didn't hang waiting for partition scan
+ if [ $elapsed -ge 10 ]; then
+ echo "FAIL: State transition took $elapsed seconds (>= 10s), likely hung on partition scan"
+ ERR_CODE=255
+ ${UBLK_PROG} del -n "${dev_id}" > /dev/null 2>&1
+ return
+ fi
+
+ echo "PASS: Device transitioned to ${expected_state} in ${elapsed}s without hanging"
+
+ # Clean up the device
+ ${UBLK_PROG} del -n "${dev_id}" > /dev/null 2>&1
+}
+
+_prep_test "partition_scan" "verify async partition scan prevents IO hang"
+
+# Test 1: Without recovery support - should transition to DEAD
+_test_partition_scan_no_hang "no" "DEAD"
+
+# Test 2: With recovery support - should transition to QUIESCED
+_test_partition_scan_no_hang "yes" "QUIESCED"
+
+_cleanup_test "partition_scan"
+_show_result $TID $ERR_CODE
--
2.47.0
next prev parent reply other threads:[~2025-12-21 16:42 UTC|newest]
Thread overview: 7+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-12-21 16:41 [PATCH 0/3] ublk: scan partition in async way Ming Lei
2025-12-21 16:41 ` [PATCH 1/3] " Ming Lei
2025-12-22 16:43 ` Caleb Sander Mateos
2025-12-21 16:41 ` Ming Lei [this message]
2025-12-21 16:41 ` [PATCH 3/3] selftests/ublk: fix Makefile to rebuild on header changes Ming Lei
2025-12-22 16:48 ` Caleb Sander Mateos
2025-12-23 2:18 ` Ming Lei
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20251221164145.1703448-3-ming.lei@redhat.com \
--to=ming.lei@redhat.com \
--cc=axboe@kernel.dk \
--cc=csander@purestorage.com \
--cc=linux-block@vger.kernel.org \
--cc=ushankar@purestorage.com \
--cc=yoav@nvidia.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox