]> git.hungrycats.org Git - linux/commitdiff
dmaengine: dmatest: warn user when dma test times out
authorAdam Wallis <awallis@codeaurora.org>
Thu, 2 Nov 2017 12:53:30 +0000 (08:53 -0400)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Tue, 21 Nov 2017 08:21:18 +0000 (09:21 +0100)
commit a9df21e34b422f79d9a9fa5c3eff8c2a53491be6 upstream.

Commit adfa543e7314 ("dmatest: don't use set_freezable_with_signal()")
introduced a bug (that is in fact documented by the patch commit text)
that leaves behind a dangling pointer. Since the done_wait structure is
allocated on the stack, future invocations to the DMATEST can produce
undesirable results (e.g., corrupted spinlocks). Ideally, this would be
cleaned up in the thread handler, but at the very least, the kernel
is left in a very precarious scenario that can lead to some long debug
sessions when the crash comes later.

Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=197605
Signed-off-by: Adam Wallis <awallis@codeaurora.org>
Signed-off-by: Vinod Koul <vinod.koul@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
drivers/dma/dmatest.c

index b8576fd6bd0e544730bfc07a19ce92b21ab1aba6..1c7568c0055a1da7c203eb7c7ec7e917f10b8bc1 100644 (file)
@@ -634,6 +634,7 @@ static int dmatest_func(void *data)
                         * free it this time?" dancing.  For now, just
                         * leave it dangling.
                         */
+                       WARN(1, "dmatest: Kernel stack may be corrupted!!\n");
                        dmaengine_unmap_put(um);
                        result("test timed out", total_tests, src_off, dst_off,
                               len, 0);