[PATCH v3 4/8] mailbox: mpfs: check the service status in .tx_done()

Conor Dooley conor at kernel.org
Tue Mar 7 12:22:54 PST 2023


From: Conor Dooley <conor.dooley at microchip.com>

Services are supposed to generate an interrupt once completed, whether
or not they have do so successfully. What appears to be a bug in the
system controller means that interrupts are only generated for
*successful* services.

Currently, the status of a service is only checked in the
mpfs_mbox_rx_data() once an interrupt is received. As it turns out, this
is not really helpful where the potentially buggy behaviour is present,
as we'll only see the status for successes where it is moot anyway.

Jassi suggested moving the check to the .tx_done() callback instead.
This makes sense, as the busy bit that tx_done() is polling will be
lowered on completion, regardless of whether the service passed or
failed.
That allows us to check the status bits for all services, whether they
generate an interrupt or not & pass something more informative than
-EBADMSG back to the drivers implementing individual services.

Suggested-by: Jassi Brar <jassisinghbrar at gmail.com>
Signed-off-by: Conor Dooley <conor.dooley at microchip.com>
---
 drivers/mailbox/mailbox-mpfs.c | 32 ++++++++++++++++----------------
 1 file changed, 16 insertions(+), 16 deletions(-)

diff --git a/drivers/mailbox/mailbox-mpfs.c b/drivers/mailbox/mailbox-mpfs.c
index 0d176aba3462..162df49654fb 100644
--- a/drivers/mailbox/mailbox-mpfs.c
+++ b/drivers/mailbox/mailbox-mpfs.c
@@ -82,8 +82,22 @@ static bool mpfs_mbox_busy(struct mpfs_mbox *mbox)
 static bool mpfs_mbox_last_tx_done(struct mbox_chan *chan)
 {
 	struct mpfs_mbox *mbox = (struct mpfs_mbox *)chan->con_priv;
+	struct mpfs_mss_response *response = mbox->response;
+	u32 val;
+
+	if (mpfs_mbox_busy(mbox))
+		return false;
+
+	/*
+	 * The service status is stored in bits 31:16 of the SERVICES_SR
+	 * register & is only valid when the system controller is not busy.
+	 * Failed services are intended to generated interrupts, but in reality
+	 * this does not happen, so the status must be checked here.
+	 */
+	val = readl_relaxed(mbox->ctrl_base + SERVICES_SR_OFFSET);
+	response->resp_status = (val & SCB_STATUS_MASK) >> SCB_STATUS_POS;
 
-	return !mpfs_mbox_busy(mbox);
+	return true;
 }
 
 static int mpfs_mbox_send_data(struct mbox_chan *chan, void *data)
@@ -138,7 +152,7 @@ static void mpfs_mbox_rx_data(struct mbox_chan *chan)
 	struct mpfs_mbox *mbox = (struct mpfs_mbox *)chan->con_priv;
 	struct mpfs_mss_response *response = mbox->response;
 	u16 num_words = ALIGN((response->resp_size), (4)) / 4U;
-	u32 i, status;
+	u32 i;
 
 	if (!response->resp_msg) {
 		dev_err(mbox->dev, "failed to assign memory for response %d\n", -ENOMEM);
@@ -146,8 +160,6 @@ static void mpfs_mbox_rx_data(struct mbox_chan *chan)
 	}
 
 	/*
-	 * The status is stored in bits 31:16 of the SERVICES_SR register.
-	 * It is only valid when BUSY == 0.
 	 * We should *never* get an interrupt while the controller is
 	 * still in the busy state. If we do, something has gone badly
 	 * wrong & the content of the mailbox would not be valid.
@@ -158,18 +170,6 @@ static void mpfs_mbox_rx_data(struct mbox_chan *chan)
 		return;
 	}
 
-	status = readl_relaxed(mbox->ctrl_base + SERVICES_SR_OFFSET);
-
-	/*
-	 * If the status of the individual servers is non-zero, the service has
-	 * failed. The contents of the mailbox at this point are not be valid,
-	 * so don't bother reading them. Set the status so that the driver
-	 * implementing the service can handle the result.
-	 */
-	response->resp_status = (status & SCB_STATUS_MASK) >> SCB_STATUS_POS;
-	if (response->resp_status)
-		return;
-
 	for (i = 0; i < num_words; i++) {
 		response->resp_msg[i] =
 			readl_relaxed(mbox->mbox_base
-- 
2.39.2




More information about the linux-riscv mailing list