linux-kernel - [PATCH 4/7] workqueue: Remember whether a work item was on a BH workqueue

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] [thread-next>] [day] [month] [year] [list]

Date: Wed, 21 Feb 2024 07:43:02 -1000
From: Tejun Heo <tj@...nel.org>
To: jiangshanlai@...il.com
Cc: torvalds@...ux-foundation.org,
	linux-kernel@...r.kernel.org,
	allen.lkml@...il.com,
	kernel-team@...a.com,
	Tejun Heo <tj@...nel.org>
Subject: [PATCH 4/7] workqueue: Remember whether a work item was on a BH workqueue

Add an off-queue flag, WORK_OFFQ_BH, that indicates whether the last
workqueue the work item was on was a BH one. This will be used to sanity
check the context cancel_sync operations can be called from for the work
item.

Signed-off-by: Tejun Heo <tj@...nel.org>
---
 include/linux/workqueue.h |  4 +++-
 kernel/workqueue.c        | 10 ++++++++--
 2 files changed, 11 insertions(+), 3 deletions(-)

diff --git a/include/linux/workqueue.h b/include/linux/workqueue.h
index 86483743ad28..7710cd52f7f0 100644
--- a/include/linux/workqueue.h
+++ b/include/linux/workqueue.h
@@ -52,9 +52,10 @@ enum work_bits {
 	 *
 	 * MSB
 	 * [ pool ID ] [ disable depth ] [ OFFQ flags ] [ STRUCT flags ]
-	 *                  16 bits          0 bits       4 or 5 bits
+	 *                  16 bits          1 bit        4 or 5 bits
 	 */
 	WORK_OFFQ_FLAG_SHIFT	= WORK_STRUCT_FLAG_BITS,
+	WORK_OFFQ_BH_BIT	= WORK_OFFQ_FLAG_SHIFT,
 	WORK_OFFQ_FLAG_END,
 	WORK_OFFQ_FLAG_BITS	= WORK_OFFQ_FLAG_END - WORK_OFFQ_FLAG_SHIFT,
 
@@ -98,6 +99,7 @@ enum wq_misc_consts {
 };
 
 /* Convenience constants - of type 'unsigned long', not 'enum'! */
+#define WORK_OFFQ_BH		(1ul << WORK_OFFQ_BH_BIT)
 #define WORK_OFFQ_FLAG_MASK	(((1ul << WORK_OFFQ_FLAG_BITS) - 1) << WORK_OFFQ_FLAG_SHIFT)
 #define WORK_OFFQ_DISABLE_MASK	(((1ul << WORK_OFFQ_DISABLE_BITS) - 1) << WORK_OFFQ_DISABLE_SHIFT)
 #define WORK_OFFQ_POOL_NONE	((1ul << WORK_OFFQ_POOL_BITS) - 1)
diff --git a/kernel/workqueue.c b/kernel/workqueue.c
index 9529ab1d9464..a03252ef3c8f 100644
--- a/kernel/workqueue.c
+++ b/kernel/workqueue.c
@@ -759,6 +759,11 @@ static int work_next_color(int color)
 	return (color + 1) % WORK_NR_COLORS;
 }
 
+static unsigned long pool_offq_flags(struct worker_pool *pool)
+{
+	return (pool->flags & POOL_BH) ? WORK_OFFQ_BH : 0;
+}
+
 /*
  * While queued, %WORK_STRUCT_PWQ is set and non flag bits of a work's data
  * contain the pointer to the queued pwq.  Once execution starts, the flag
@@ -2111,7 +2116,8 @@ static int try_to_grab_pending(struct work_struct *work, u32 cflags,
 		 * this destroys work->data needed by the next step, stash it.
 		 */
 		work_data = *work_data_bits(work);
-		set_work_pool_and_keep_pending(work, pool->id, 0);
+		set_work_pool_and_keep_pending(work, pool->id,
+					       pool_offq_flags(pool));
 
 		/* must be the last step, see the function comment */
 		pwq_dec_nr_in_flight(pwq, work_data);
@@ -3163,7 +3169,7 @@ __acquires(&pool->lock)
 	 * PENDING and queued state changes happen together while IRQ is
 	 * disabled.
 	 */
-	set_work_pool_and_clear_pending(work, pool->id, 0);
+	set_work_pool_and_clear_pending(work, pool->id, pool_offq_flags(pool));
 
 	pwq->stats[PWQ_STAT_STARTED]++;
 	raw_spin_unlock_irq(&pool->lock);
-- 
2.43.2