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>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <ZpW61giPoe_k8mIv@slm.duckdns.org>
Date: Mon, 15 Jul 2024 14:12:06 -1000
From: Tejun Heo <tj@...nel.org>
To: Lai Jiangshan <jiangshanlai@...il.com>
Cc: linux-kernel@...r.kernel.org,
	Lai Jiangshan <jiangshan.ljs@...group.com>,
	kernel test robot <oliver.sang@...el.com>
Subject: Re: [PATCH] workqueue: Remove unneeded lockdep_assert_cpus_held()

Lai, I dropped the cpus_read_lock/unlock() from wq_affn_dfl_set() and
applied the following patch to fasttrack it as Linus already pulled
for-6.11. Let's do the removal as a separate patch.

Thanks.

>From aa8684755a283536bd8ad93141052f47a4faa5a3 Mon Sep 17 00:00:00 2001
From: Lai Jiangshan <jiangshan.ljs@...group.com>
Date: Tue, 16 Jul 2024 01:29:31 +0800
Subject: [PATCH] workqueue: Remove unneeded lockdep_assert_cpus_held()

The commit 19af45757383 ("workqueue: Remove cpus_read_lock() from
apply_wqattrs_lock()") removes the unneed cpus_read_lock() after the pwq
creations and installations have been reworked based on wq_online_cpumask
rather than cpu_online_mask making cpus_read_lock() is unneeded during
wqattrs changes.

But it desn't remove the lockdep_assert_cpus_held() checks during wqattrs
changes, which leads to complaints from lockdep reported by kernel test
robot:

[   15.726567][  T131] ------------[ cut here ]------------
[ 15.728117][ T131] WARNING: CPU: 1 PID: 131 at kernel/cpu.c:525 lockdep_assert_cpus_held (kernel/cpu.c:525)
[   15.731191][  T131] Modules linked in: floppy(+) parport_pc(+) parport qemu_fw_cfg rtc_cmos
[   15.733423][  T131] CPU: 1 PID: 131 Comm: systemd-udevd Tainted: G                T  6.10.0-rc2-00254-g19af45757383 #1 df6f039f42e8818bf9a534449362ebad1aad32e2
[   15.737011][  T131] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.16.2-debian-1.16.2-1 04/01/2014
[ 15.739760][ T131] EIP: lockdep_assert_cpus_held (kernel/cpu.c:525)
[ 15.741326][ T131] Code: 97 c2 03 72 20 83 3d f4 73 97 c2 00 74 17 55 89 e5 b8 fc bd 4d c2 ba ff ff ff ff e8 e4 57 d1 00 85 c0 74 06 5d 31 c0 31 d2 c3 <0f> 0b eb f6 90 90 90 90 90 90 90 90 90 90 90 90 90 90 55 89 e5 b8

Fix it by removing the unneeded lockdep_assert_cpus_held().
Also remove the unneed cpus_read_lock() from wq_affn_dfl_set().

tj: Dropped the removal of cpus_read_lock/unlock() in wq_affn_dfl_set() to
    keep this patch fix only.

Cc: kernel test robot <oliver.sang@...el.com>
Fixes: 19af45757383("workqueue: Remove cpus_read_lock() from apply_wqattrs_lock()")
Reported-by: kernel test robot <oliver.sang@...el.com>
Closes: https://lore.kernel.org/oe-lkp/202407141846.665c0446-lkp@intel.com
Signed-off-by: Lai Jiangshan <jiangshan.ljs@...group.com>
Signed-off-by: Tejun Heo <tj@...nel.org>
---
 kernel/workqueue.c | 12 ++----------
 1 file changed, 2 insertions(+), 10 deletions(-)

diff --git a/kernel/workqueue.c b/kernel/workqueue.c
index dfd42c28e404..1745ca788ede 100644
--- a/kernel/workqueue.c
+++ b/kernel/workqueue.c
@@ -5342,8 +5342,6 @@ static int apply_workqueue_attrs_locked(struct workqueue_struct *wq,
  *
  * Performs GFP_KERNEL allocations.
  *
- * Assumes caller has CPU hotplug read exclusion, i.e. cpus_read_lock().
- *
  * Return: 0 on success and -errno on failure.
  */
 int apply_workqueue_attrs(struct workqueue_struct *wq,
@@ -5351,8 +5349,6 @@ int apply_workqueue_attrs(struct workqueue_struct *wq,
 {
 	int ret;
 
-	lockdep_assert_cpus_held();
-
 	mutex_lock(&wq_pool_mutex);
 	ret = apply_workqueue_attrs_locked(wq, attrs);
 	mutex_unlock(&wq_pool_mutex);
@@ -5434,7 +5430,6 @@ static int alloc_and_link_pwqs(struct workqueue_struct *wq)
 	bool highpri = wq->flags & WQ_HIGHPRI;
 	int cpu, ret;
 
-	lockdep_assert_cpus_held();
 	lockdep_assert_held(&wq_pool_mutex);
 
 	wq->cpu_pwq = alloc_percpu(struct pool_workqueue *);
@@ -5695,8 +5690,7 @@ struct workqueue_struct *alloc_workqueue(const char *fmt,
 
 	/*
 	 * wq_pool_mutex protects the workqueues list, allocations of PWQs,
-	 * and the global freeze state.  alloc_and_link_pwqs() also requires
-	 * cpus_read_lock() for PWQs' affinities.
+	 * and the global freeze state.
 	 */
 	apply_wqattrs_lock();
 
@@ -6862,8 +6856,7 @@ static int workqueue_apply_unbound_cpumask(const cpumask_var_t unbound_cpumask)
  * @exclude_cpumask: the cpumask to be excluded from wq_unbound_cpumask
  *
  * This function can be called from cpuset code to provide a set of isolated
- * CPUs that should be excluded from wq_unbound_cpumask. The caller must hold
- * either cpus_read_lock or cpus_write_lock.
+ * CPUs that should be excluded from wq_unbound_cpumask.
  */
 int workqueue_unbound_exclude_cpumask(cpumask_var_t exclude_cpumask)
 {
@@ -6873,7 +6866,6 @@ int workqueue_unbound_exclude_cpumask(cpumask_var_t exclude_cpumask)
 	if (!zalloc_cpumask_var(&cpumask, GFP_KERNEL))
 		return -ENOMEM;
 
-	lockdep_assert_cpus_held();
 	mutex_lock(&wq_pool_mutex);
 
 	/*
-- 
2.45.2


Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ