Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
From:   Martin Fuzzey <martin.fuzzey@flowbird.group>
To:     dmaengine@vger.kernel.org
Cc:     stable@vger.kernel.org, Shawn Guo <shawnguo@kernel.org>,
        Sascha Hauer <s.hauer@pengutronix.de>,
        Fabio Estevam <festevam@gmail.com>,
        NXP Linux Team <linux-imx@nxp.com>,
        Robin Gong <yibin.gong@nxp.com>,
        linux-arm-kernel@lists.infradead.org, linux-kernel@vger.kernel.org
Subject: [PATCH] dmaengine: imx-sdma: fix context cache
Date:   Wed, 29 Jan 2020 14:40:06 +0100
Message-Id: <1580305274-27274-1-git-send-email-martin.fuzzey@flowbird.group>
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

There is a DMA problem with the serial ports on i.MX6.

When the following sequence is performed:

1) Open a port
2) Write some data
3) Close the port
4) Open a *different* port
5) Write some data
6) Close the port

The second write sends nothing and the second close hangs.
If the first close() is omitted it works.

Adding logs to the the UART driver shows that the DMA is being setup but
the callback is never invoked for the second write.

This used to work in 4.19.

Git bisect leads to:
	ad0d92d: "dmaengine: imx-sdma: refine to load context only once"

This commit adds a "context_loaded" flag used to avoid unnecessary context
setups.
However the flag is only reset in sdma_channel_terminate_work(),
which is only invoked in a worker triggered by sdma_terminate_all() IF
there is an active descriptor.

So, if no active descriptor remains when the channel is terminated, the
flag is not reset and, when the channel is later reused the old context
is used.

Fix the problem by always resetting the flag in sdma_free_chan_resources().

Fixes: ad0d92d: "dmaengine: imx-sdma: refine to load context only once"
Cc: stable@vger.kernel.org
Signed-off-by: Martin Fuzzey <martin.fuzzey@flowbird.group>

---

The following python script may be used to reproduce the problem:

import re, serial, sys

ports=(0, 4) # Can be any ports not used (no need to connect anything) NOT console...

def get_tx_counts():
        pattern = re.compile("(\d+):.*tx:(\d+).*")
        tx_counts = {}
        with open("/proc/tty/driver/IMX-uart", "r") as f:
                for line in f:
                        match = pattern.match(line)
                        if match:
                                tx_counts[int(match.group(1))] = int(match.group(2))
        return tx_counts

before = get_tx_counts()

a = serial.Serial("/dev/ttymxc%d" % ports[0])
a.write("polop")
a.close()
b = serial.Serial("/dev/ttymxc%d" % ports[1])
b.write("test")

after = get_tx_counts()

if (after[ports[0]] - before[ports[0]]  > 0) and (after[ports[1]] - before[ports[1]] > 0):
        print "PASS"
        sys.exit(0)
else:
        print "FAIL"
        print "Before: %s" % before
        print "After: %s" % after
        sys.exit(1)
---
 drivers/dma/imx-sdma.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/drivers/dma/imx-sdma.c b/drivers/dma/imx-sdma.c
index 066b21a..332ca50 100644
--- a/drivers/dma/imx-sdma.c
+++ b/drivers/dma/imx-sdma.c
@@ -1338,6 +1338,7 @@ static void sdma_free_chan_resources(struct dma_chan *chan)
 
 	sdmac->event_id0 = 0;
 	sdmac->event_id1 = 0;
+	sdmac->context_loaded = false;
 
 	sdma_set_channel_priority(sdmac, 0);
 
-- 
1.9.1