<head></head><body style="line-height: 1.5; font-size: 14px; color: rgba(38, 42, 51, 0.9); font-family: Source Han Sans;">
<div style="font-family:Source Han Sans;font-size:14px;line-height:1.5;
" data-mce-style="font-family:Source Han Sans;font-size:14px;line-height:1.5;
"><br></div><div style="font-family:Source Han Sans;font-size:14px;line-height:1.5;
" data-mce-style="font-family:Source Han Sans;font-size:14px;line-height:1.5;
">ping</div><div style="font-family:Source Han Sans;font-size:14px;line-height:1.5;
" data-mce-style="font-family:Source Han Sans;font-size:14px;line-height:1.5;
"><br></div><div class="mail-signature"> <div class="lingxi-signature-container" style="font-family: PingFang,Source Han Sans,sans-serif;"> <!-- 样式一，完整版 --> <table style="border-collapse: collapse;width: 100%;text-size-adjust:none !important;table-layout: fixed;" class="mce-item-table"> <tbody style="text-size-adjust:none !important;word-wrap:break-word; word-break:break-all;"> </tbody> </table> <!-- 样式二，无头像 --> <div id="lingxi-signature-v2-block" style="width: 100%;text-decoration: none;text-underline: none;"> <div id="lingxi-signature-v2-content" itemid="id" style="max-width: 95%;font-size: 14px;line-height: 16px;color: #7A8599;padding: 16px 16px 4px 0px;margin-left: 0px;" itemprop="5317106"> <div style="color: #232D47;border: none;margin-bottom:4px;font-size: 16px;line-height: 20px;text-underline: none;font-weight: bolder;" itemid="name">黄正兵</div> <div style="border: none;margin-bottom:8px;text-underline: none;">zhengbing.huang@easystack.cn</div> </div> </div> <!--办公品牌--> </div></div><div class="pre-mail-content"><div style="font-family:Source Han Sans;font-size:14px;line-height:1.5;
" data-mce-style="font-family:Source Han Sans;font-size:14px;line-height:1.5;
"><br></div><div style="font-family:Source Han Sans;font-size:14px;line-height:1.5;
" data-mce-style="font-family:Source Han Sans;font-size:14px;line-height:1.5;
"><br></div><div style="font-family:Source Han Sans;font-size:14px;line-height:1.5;
" data-mce-style="font-family:Source Han Sans;font-size:14px;line-height:1.5;
"><br></div><div id="isReplyContent" style="margin: 0;"><div style="color: #7d8085">Original:</div><ul style="color: #7d8085; font-size:12px; padding-left: 20px"><li>From：zhengbing . huang<<a href="mailto:zhengbing.huang@easystack.cn">zhengbing.huang@easystack.cn</a>></li><li>Date：2025-05-12 13:44:49(中国 (GMT+08:00))</li><li>To：drbd-dev<<a href="mailto:drbd-dev@lists.linbit.com">drbd-dev@lists.linbit.com</a>></li><li>Cc：zhengbing . huang<<a href="mailto:zhengbing.huang@easystack.cn">zhengbing.huang@easystack.cn</a>></li><li>Subject：[PATCH] drbd: Fix kernel hangs issue</li></ul><pre>Kernel hangs:
[433091.237135] BUG: scheduling while atomic: swapper/27/0/0x00000100
[433091.392756] CPU: 27 PID: 0 Comm: swapper/27 Kdump: loaded Tainted: G
OE --------- - - 4.18.0-372.19.1.es8_10.aarch64 #1
[433091.406962] Hardware name: WUZHOU S627K2/BC82AMDQ, BIOS 6.70
04/03/2024
[433091.414645] Call trace:
[433091.418157] dump_backtrace+0x0/0x158
[433091.422867] show_stack+0x24/0x30
[433091.427215] dump_stack+0x5c/0x74
[433091.431582] __schedule_bug+0x74/0x88
[433091.436260] __schedule+0x794/0x860
[433091.440760] schedule+0x48/0xc8
[433091.444888] schedule_timeout+0x1ac/0x2c0
[433091.449903] __dtr_disconnect_path+0x28c/0x520 [drbd_transport_rdma]
[433091.457229] dtr_disconnect_path.part.3+0x20/0x78 [drbd_transport_rdma]
[433091.464879] dtr_remove_path+0x24/0x30 [drbd_transport_rdma]
[433091.471553] drbd_destroy_path+0x34/0x60 [drbd]
[433091.477144] drbd_reclaim_path+0x44/0x50 [drbd]
[433091.482683] rcu_do_batch+0x178/0x438
[433091.487319] rcu_core+0x1d4/0x2f8
[433091.491635] rcu_core_si+0x14/0x20
[433091.496021] __do_softirq+0x118/0x320
[433091.500656] irq_exit_rcu+0x10c/0x120
[433091.505243] irq_exit+0x14/0x20
[433091.509293] __handle_domain_irq+0x70/0xc0
[433091.514266] gic_handle_irq+0xd4/0x178
[433091.518891] el1_irq+0xb8/0x140
[433091.522877] arch_cpu_idle+0x20/0x28
[433091.527295] default_idle_call+0x54/0x158
[433091.532154] do_idle+0x208/0x278
[433091.536230] cpu_startup_entry+0x28/0x30
[433091.540977] secondary_start_kernel+0x150/0x160
(gdb) l *__dtr_disconnect_path+0x28c
0xa3c is in __dtr_disconnect_path
(/usr/src/debug/kmod-drbd-kmodtool-9.2.9-10.es8_19.k372.mlnx.23.04.aarch64/obj/default-OFED/drbd/drbd_transport_rdma.c:2772).
2771 case PCS_FINISHING:
2772 t = wait_event_timeout(path->cs.wq,
2773 atomic_read(&path->cs.passive_state) == PCS_INACTIVE,
2774 HZ * 60);
So, the problem is that schedule is called in the softirq.
We should remove ops.remove_path() from call_rcu() to avoid this problem.
Signed-off-by: zhengbing.huang <zhengbing.huang@easystack.cn>
---
drbd/drbd_main.c | 5 +++--
drbd/drbd_nl.c | 2 ++
2 files changed, 5 insertions(+), 2 deletions(-)
diff --git a/drbd/drbd_main.c b/drbd/drbd_main.c
index f058736a1..d3912e019 100644
--- a/drbd/drbd_main.c
+++ b/drbd/drbd_main.c
@@ -3078,6 +3078,9 @@ static void drbd_remove_all_paths(struct drbd_connection *connection)
smp_wmb();
list_for_each_entry_safe(path, tmp, &transport->paths, list) {
+
+ transport->class->ops.remove_path(path);
+
/* Exclusive with reading state, in particular remember_state_change() */
write_lock_irq(&resource->state_rwlock);
list_del_rcu(&path->list);
@@ -3963,8 +3966,6 @@ void drbd_destroy_path(struct kref *kref)
struct drbd_connection *connection =
container_of(path->transport, struct drbd_connection, transport);
- connection->transport.class->ops.remove_path(path);
-
kref_debug_put(&connection->kref_debug, 17);
kref_put(&connection->kref, drbd_destroy_connection);
kfree(path);
diff --git a/drbd/drbd_nl.c b/drbd/drbd_nl.c
index 3a7991e74..bb5d531bf 100644
--- a/drbd/drbd_nl.c
+++ b/drbd/drbd_nl.c
@@ -4810,6 +4810,8 @@ adm_del_path(struct drbd_config_context *adm_ctx, struct genl_info *info)
/* Ensure flag visible before list manipulation. */
smp_wmb();
+ transport->class->ops.remove_path(path);
+
/* Exclusive with reading state, in particular remember_state_change() */
write_lock_irq(&resource->state_rwlock);
list_del_rcu(&path->list);
--
2.43.0
</pre></div></div>
<br></body>