All of lore.kernel.org
 help / color / mirror / Atom feed
* FAILED: patch "[PATCH] platform/chrome: cros_ec_rpmsg: Fix race with host command" failed to apply to 5.3-stable tree
@ 2019-10-03 12:10 gregkh
  2019-10-03 14:36 ` Sasha Levin
  0 siblings, 1 reply; 2+ messages in thread
From: gregkh @ 2019-10-03 12:10 UTC (permalink / raw)
  To: pihsun, enric.balletbo; +Cc: stable


The patch below does not apply to the 5.3-stable tree.
If someone wants it applied there, or to any other stable or longterm
tree, then please email the backport, including the original git commit
id to <stable@vger.kernel.org>.

thanks,

greg k-h

------------------ original commit in Linus's tree ------------------

From 71cddb7097e2b0feb855d7fd7d59afd12cbee4bb Mon Sep 17 00:00:00 2001
From: Pi-Hsun Shih <pihsun@chromium.org>
Date: Wed, 4 Sep 2019 14:26:13 +0800
Subject: [PATCH] platform/chrome: cros_ec_rpmsg: Fix race with host command
 when probe failed

Since the rpmsg_endpoint is created before probe is called, it's
possible that a host event is received during cros_ec_register, and
there would be some pending work in the host_event_work workqueue while
cros_ec_register is called.

If cros_ec_register fails, when the leftover work in host_event_work
run, the ec_dev from the drvdata of the rpdev could be already set to
NULL, causing kernel crash when trying to run cros_ec_get_next_event.

Fix this by creating the rpmsg_endpoint by ourself, and when
cros_ec_register fails (or on remove), destroy the endpoint first (to
make sure there's no more new calls to cros_ec_rpmsg_callback), and then
cancel all works in the host_event_work workqueue.

Cc: stable@vger.kernel.org
Fixes: 2de89fd98958 ("platform/chrome: cros_ec: Add EC host command support using rpmsg")
Signed-off-by: Pi-Hsun Shih <pihsun@chromium.org>
Signed-off-by: Enric Balletbo i Serra <enric.balletbo@collabora.com>

diff --git a/drivers/platform/chrome/cros_ec_rpmsg.c b/drivers/platform/chrome/cros_ec_rpmsg.c
index 8b6bd775cc9a..0c3738c3244d 100644
--- a/drivers/platform/chrome/cros_ec_rpmsg.c
+++ b/drivers/platform/chrome/cros_ec_rpmsg.c
@@ -41,6 +41,7 @@ struct cros_ec_rpmsg {
 	struct rpmsg_device *rpdev;
 	struct completion xfer_ack;
 	struct work_struct host_event_work;
+	struct rpmsg_endpoint *ept;
 };
 
 /**
@@ -72,7 +73,6 @@ static int cros_ec_pkt_xfer_rpmsg(struct cros_ec_device *ec_dev,
 				  struct cros_ec_command *ec_msg)
 {
 	struct cros_ec_rpmsg *ec_rpmsg = ec_dev->priv;
-	struct rpmsg_device *rpdev = ec_rpmsg->rpdev;
 	struct ec_host_response *response;
 	unsigned long timeout;
 	int len;
@@ -85,7 +85,7 @@ static int cros_ec_pkt_xfer_rpmsg(struct cros_ec_device *ec_dev,
 	dev_dbg(ec_dev->dev, "prepared, len=%d\n", len);
 
 	reinit_completion(&ec_rpmsg->xfer_ack);
-	ret = rpmsg_send(rpdev->ept, ec_dev->dout, len);
+	ret = rpmsg_send(ec_rpmsg->ept, ec_dev->dout, len);
 	if (ret) {
 		dev_err(ec_dev->dev, "rpmsg send failed\n");
 		return ret;
@@ -196,11 +196,24 @@ static int cros_ec_rpmsg_callback(struct rpmsg_device *rpdev, void *data,
 	return 0;
 }
 
+static struct rpmsg_endpoint *
+cros_ec_rpmsg_create_ept(struct rpmsg_device *rpdev)
+{
+	struct rpmsg_channel_info chinfo = {};
+
+	strscpy(chinfo.name, rpdev->id.name, RPMSG_NAME_SIZE);
+	chinfo.src = rpdev->src;
+	chinfo.dst = RPMSG_ADDR_ANY;
+
+	return rpmsg_create_ept(rpdev, cros_ec_rpmsg_callback, NULL, chinfo);
+}
+
 static int cros_ec_rpmsg_probe(struct rpmsg_device *rpdev)
 {
 	struct device *dev = &rpdev->dev;
 	struct cros_ec_rpmsg *ec_rpmsg;
 	struct cros_ec_device *ec_dev;
+	int ret;
 
 	ec_dev = devm_kzalloc(dev, sizeof(*ec_dev), GFP_KERNEL);
 	if (!ec_dev)
@@ -225,7 +238,18 @@ static int cros_ec_rpmsg_probe(struct rpmsg_device *rpdev)
 	INIT_WORK(&ec_rpmsg->host_event_work,
 		  cros_ec_rpmsg_host_event_function);
 
-	return cros_ec_register(ec_dev);
+	ec_rpmsg->ept = cros_ec_rpmsg_create_ept(rpdev);
+	if (!ec_rpmsg->ept)
+		return -ENOMEM;
+
+	ret = cros_ec_register(ec_dev);
+	if (ret < 0) {
+		rpmsg_destroy_ept(ec_rpmsg->ept);
+		cancel_work_sync(&ec_rpmsg->host_event_work);
+		return ret;
+	}
+
+	return 0;
 }
 
 static void cros_ec_rpmsg_remove(struct rpmsg_device *rpdev)
@@ -234,7 +258,7 @@ static void cros_ec_rpmsg_remove(struct rpmsg_device *rpdev)
 	struct cros_ec_rpmsg *ec_rpmsg = ec_dev->priv;
 
 	cros_ec_unregister(ec_dev);
-
+	rpmsg_destroy_ept(ec_rpmsg->ept);
 	cancel_work_sync(&ec_rpmsg->host_event_work);
 }
 
@@ -271,7 +295,6 @@ static struct rpmsg_driver cros_ec_driver_rpmsg = {
 	},
 	.probe		= cros_ec_rpmsg_probe,
 	.remove		= cros_ec_rpmsg_remove,
-	.callback	= cros_ec_rpmsg_callback,
 };
 
 module_rpmsg_driver(cros_ec_driver_rpmsg);


^ permalink raw reply related	[flat|nested] 2+ messages in thread

* Re: FAILED: patch "[PATCH] platform/chrome: cros_ec_rpmsg: Fix race with host command" failed to apply to 5.3-stable tree
  2019-10-03 12:10 FAILED: patch "[PATCH] platform/chrome: cros_ec_rpmsg: Fix race with host command" failed to apply to 5.3-stable tree gregkh
@ 2019-10-03 14:36 ` Sasha Levin
  0 siblings, 0 replies; 2+ messages in thread
From: Sasha Levin @ 2019-10-03 14:36 UTC (permalink / raw)
  To: gregkh; +Cc: pihsun, enric.balletbo, stable

On Thu, Oct 03, 2019 at 02:10:05PM +0200, gregkh@linuxfoundation.org wrote:
>
>The patch below does not apply to the 5.3-stable tree.
>If someone wants it applied there, or to any other stable or longterm
>tree, then please email the backport, including the original git commit
>id to <stable@vger.kernel.org>.
>
>thanks,
>
>greg k-h
>
>------------------ original commit in Linus's tree ------------------
>
>From 71cddb7097e2b0feb855d7fd7d59afd12cbee4bb Mon Sep 17 00:00:00 2001
>From: Pi-Hsun Shih <pihsun@chromium.org>
>Date: Wed, 4 Sep 2019 14:26:13 +0800
>Subject: [PATCH] platform/chrome: cros_ec_rpmsg: Fix race with host command
> when probe failed
>
>Since the rpmsg_endpoint is created before probe is called, it's
>possible that a host event is received during cros_ec_register, and
>there would be some pending work in the host_event_work workqueue while
>cros_ec_register is called.
>
>If cros_ec_register fails, when the leftover work in host_event_work
>run, the ec_dev from the drvdata of the rpdev could be already set to
>NULL, causing kernel crash when trying to run cros_ec_get_next_event.
>
>Fix this by creating the rpmsg_endpoint by ourself, and when
>cros_ec_register fails (or on remove), destroy the endpoint first (to
>make sure there's no more new calls to cros_ec_rpmsg_callback), and then
>cancel all works in the host_event_work workqueue.
>
>Cc: stable@vger.kernel.org
>Fixes: 2de89fd98958 ("platform/chrome: cros_ec: Add EC host command support using rpmsg")
>Signed-off-by: Pi-Hsun Shih <pihsun@chromium.org>
>Signed-off-by: Enric Balletbo i Serra <enric.balletbo@collabora.com>

I've worked around the changes introduced by 7aa703bb88243 ("mfd /
platform: cros_ec: Handle chained ECs as platform devices"). Queued for
5.3 and 5.2.

-- 
Thanks,
Sasha

^ permalink raw reply	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2019-10-03 14:36 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2019-10-03 12:10 FAILED: patch "[PATCH] platform/chrome: cros_ec_rpmsg: Fix race with host command" failed to apply to 5.3-stable tree gregkh
2019-10-03 14:36 ` Sasha Levin

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.