linux-kernel - [PATCH 4.14 038/109] typec: tcpm: fusb302: Resolve out of order messaging events

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 for Android: free password hash cracker in your pocket

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

Date:   Fri, 16 Mar 2018 16:23:07 +0100
From:   Greg Kroah-Hartman <gregkh@...uxfoundation.org>
To:     linux-kernel@...r.kernel.org
Cc:     Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
        stable@...r.kernel.org,
        Adam Thomson <Adam.Thomson.Opensource@...semi.com>,
        Guenter Roeck <linux@...ck-us.net>,
        Heikki Krogerus <heikki.krogerus@...ux.intel.com>,
        Sasha Levin <alexander.levin@...rosoft.com>
Subject: [PATCH 4.14 038/109] typec: tcpm: fusb302: Resolve out of order messaging events

4.14-stable review patch.  If anyone has any objections, please let me know.

------------------

From: Adam Thomson <Adam.Thomson.Opensource@...semi.com>


[ Upstream commit ab69f61321140ff632d560775bc226259a78dfa2 ]

The expectation in the FUSB302 driver is that a TX_SUCCESS event
should occur after a message has been sent, but before a GCRCSENT
event is raised to indicate successful receipt of a message from
the partner. However in some circumstances it is possible to see
the hardware raise a GCRCSENT event before a TX_SUCCESS event
is raised. The upshot of this is that the GCRCSENT handling portion
of code ends up reporting the GoodCRC message to TCPM because the
TX_SUCCESS event hasn't yet arrived to trigger a consumption of it.
When TX_SUCCESS is then raised by the chip it ends up consuming the
actual message that was meant for TCPM, and this incorrect sequence
results in a hard reset from TCPM.

To avoid this problem, this commit updates the message reading
code to check whether a GoodCRC message was received or not. Based
on this check it will either report that the previous transmission
has completed or it will pass the msg data to TCPM for futher
processing. This way the incorrect ordering of the events no longer
matters.

Signed-off-by: Adam Thomson <Adam.Thomson.Opensource@...semi.com>
Reviewed-by: Guenter Roeck <linux@...ck-us.net>
Acked-by: Heikki Krogerus <heikki.krogerus@...ux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@...uxfoundation.org>
Signed-off-by: Sasha Levin <alexander.levin@...rosoft.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@...uxfoundation.org>
---
 drivers/staging/typec/fusb302/fusb302.c |   21 +++++++++++++++++----
 1 file changed, 17 insertions(+), 4 deletions(-)

--- a/drivers/staging/typec/fusb302/fusb302.c
+++ b/drivers/staging/typec/fusb302/fusb302.c
@@ -1552,6 +1552,21 @@ static int fusb302_pd_read_message(struc
 	fusb302_log(chip, "PD message header: %x", msg->header);
 	fusb302_log(chip, "PD message len: %d", len);
 
+	/*
+	 * Check if we've read off a GoodCRC message. If so then indicate to
+	 * TCPM that the previous transmission has completed. Otherwise we pass
+	 * the received message over to TCPM for processing.
+	 *
+	 * We make this check here instead of basing the reporting decision on
+	 * the IRQ event type, as it's possible for the chip to report the
+	 * TX_SUCCESS and GCRCSENT events out of order on occasion, so we need
+	 * to check the message type to ensure correct reporting to TCPM.
+	 */
+	if ((!len) && (pd_header_type_le(msg->header) == PD_CTRL_GOOD_CRC))
+		tcpm_pd_transmit_complete(chip->tcpm_port, TCPC_TX_SUCCESS);
+	else
+		tcpm_pd_receive(chip->tcpm_port, msg);
+
 	return ret;
 }
 
@@ -1659,13 +1674,12 @@ static irqreturn_t fusb302_irq_intn(int
 
 	if (interrupta & FUSB_REG_INTERRUPTA_TX_SUCCESS) {
 		fusb302_log(chip, "IRQ: PD tx success");
-		/* read out the received good CRC */
 		ret = fusb302_pd_read_message(chip, &pd_msg);
 		if (ret < 0) {
-			fusb302_log(chip, "cannot read in GCRC, ret=%d", ret);
+			fusb302_log(chip,
+				    "cannot read in PD message, ret=%d", ret);
 			goto done;
 		}
-		tcpm_pd_transmit_complete(chip->tcpm_port, TCPC_TX_SUCCESS);
 	}
 
 	if (interrupta & FUSB_REG_INTERRUPTA_HARDRESET) {
@@ -1686,7 +1700,6 @@ static irqreturn_t fusb302_irq_intn(int
 				    "cannot read in PD message, ret=%d", ret);
 			goto done;
 		}
-		tcpm_pd_receive(chip->tcpm_port, &pd_msg);
 	}
 done:
 	mutex_unlock(&chip->lock);