Received-SPF: pass (google.com: domain of linux-kernel+bounces-166726-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:4601:e00::3 as permitted sender) client-ip=2604:1380:4601:e00::3;
From: Fedor Pchelkin <pchelkin@ispras.ru>
To: Xiang Chen <chenxiang66@hisilicon.com>
Cc: Fedor Pchelkin <pchelkin@ispras.ru>,
	Christoph Hellwig <hch@lst.de>,
	Marek Szyprowski <m.szyprowski@samsung.com>,
	Robin Murphy <robin.murphy@arm.com>,
	Barry Song <song.bao.hua@hisilicon.com>,
	iommu@lists.linux.dev,
	linux-kernel@vger.kernel.org,
	Alexey Khoroshilov <khoroshilov@ispras.ru>,
	lvc-project@linuxtesting.org
Subject: [PATCH 2/2] dma-mapping: benchmark: prevent potential kthread hang
Date: Thu,  2 May 2024 19:18:26 +0300
Message-ID: <20240502161827.403338-3-pchelkin@ispras.ru>
In-Reply-To: <20240502161827.403338-1-pchelkin@ispras.ru>
References: <20240502161827.403338-1-pchelkin@ispras.ru>
Precedence: bulk
MIME-Version: 1.0
Content-Transfer-Encoding: 8bit

If some of kthreads executing map_benchmark_thread() return with an error
code (e.g. due to a memory allocation failure), then the next kthreads in
the array are not stopped and potentially loop for indefinite time.

Call kthread_stop() for each started thread as map_benchmark_thread()
expects that happening in order to exit.

Found by Linux Verification Center (linuxtesting.org).

Fixes: 65789daa8087 ("dma-mapping: add benchmark support for streaming DMA APIs")
Signed-off-by: Fedor Pchelkin <pchelkin@ispras.ru>
---
 kernel/dma/map_benchmark.c | 12 ++++++++----
 1 file changed, 8 insertions(+), 4 deletions(-)

diff --git a/kernel/dma/map_benchmark.c b/kernel/dma/map_benchmark.c
index ea938bc6c7e3..7e39a4690331 100644
--- a/kernel/dma/map_benchmark.c
+++ b/kernel/dma/map_benchmark.c
@@ -140,13 +140,17 @@ static int do_map_benchmark(struct map_benchmark_data *map)
 
 	msleep_interruptible(map->bparam.seconds * 1000);
 
-	/* wait for the completion of benchmark threads */
+	/* wait for the completion of all started benchmark threads */
 	for (i = 0; i < threads; i++) {
-		ret = kthread_stop(tsk[i]);
-		if (ret)
-			goto out;
+		int kthread_ret = kthread_stop(tsk[i]);
+
+		if (kthread_ret)
+			ret = kthread_ret;
 	}
 
+	if (ret)
+		goto out;
+
 	loops = atomic64_read(&map->loops);
 	if (likely(loops > 0)) {
 		u64 map_variance, unmap_variance;
-- 
2.45.0