public inbox for linux-kernel@vger.kernel.org
 help / color / mirror / Atom feed
* [PATCH v2] drm/vc4: Fix sleeps during the IRQ handler for DSI transactions.
@ 2017-10-14  0:12 Eric Anholt
  2017-10-14  7:28 ` Paul Menzel
  2017-10-14  7:53 ` Boris Brezillon
  0 siblings, 2 replies; 3+ messages in thread
From: Eric Anholt @ 2017-10-14  0:12 UTC (permalink / raw)
  To: dri-devel; +Cc: linux-kernel, Boris Brezillon, Eric Anholt

VC4's DSI1 has a bug where the AXI connection is broken for 32-bit
writes from the CPU, so we use the DMA engine to DMA 32-bit values
into registers instead.  That sleeps, so we can't do it from the top
half.

As a solution, use an interrupt thread so that all our writes happen
when sleeping is is allowed.

v2: Use IRQF_ONESHOT (suggested by Boris)

Signed-off-by: Eric Anholt <eric@anholt.net>
---

Boris, that cleanup ended up working and it looks great.  Thanks!

 drivers/gpu/drm/vc4/vc4_dsi.c | 31 +++++++++++++++++++++++++++++--
 1 file changed, 29 insertions(+), 2 deletions(-)

diff --git a/drivers/gpu/drm/vc4/vc4_dsi.c b/drivers/gpu/drm/vc4/vc4_dsi.c
index 554605af344e..3b74fda5662d 100644
--- a/drivers/gpu/drm/vc4/vc4_dsi.c
+++ b/drivers/gpu/drm/vc4/vc4_dsi.c
@@ -1360,6 +1360,25 @@ static void dsi_handle_error(struct vc4_dsi *dsi,
 	*ret = IRQ_HANDLED;
 }
 
+/* Initial handler for port 1 where we need the reg_dma workaround.
+ * The register DMA writes sleep, so we can't do it in the top half.
+ * Instead we use IRQF_ONESHOT so that the IRQ gets disabled in the
+ * parent interrupt contrller until our interrupt thread is done.
+ */
+static irqreturn_t vc4_dsi_irq_defer_to_thread_handler(int irq, void *data)
+{
+	struct vc4_dsi *dsi = data;
+	u32 stat = DSI_PORT_READ(INT_STAT);
+
+	if (!stat)
+		return IRQ_NONE;
+
+	return IRQ_WAKE_THREAD;
+}
+
+/* Normal IRQ handler for port 0, or the threaded IRQ handler for port
+ * 1 where we need the reg_dma workaround.
+ */
 static irqreturn_t vc4_dsi_irq_handler(int irq, void *data)
 {
 	struct vc4_dsi *dsi = data;
@@ -1539,8 +1558,16 @@ static int vc4_dsi_bind(struct device *dev, struct device *master, void *data)
 	/* Clear any existing interrupt state. */
 	DSI_PORT_WRITE(INT_STAT, DSI_PORT_READ(INT_STAT));
 
-	ret = devm_request_irq(dev, platform_get_irq(pdev, 0),
-			       vc4_dsi_irq_handler, 0, "vc4 dsi", dsi);
+	if (dsi->reg_dma_mem) {
+		ret = devm_request_threaded_irq(dev, platform_get_irq(pdev, 0),
+						vc4_dsi_irq_defer_to_thread_handler,
+						vc4_dsi_irq_handler,
+						IRQF_ONESHOT,
+						"vc4 dsi", dsi);
+	} else {
+		ret = devm_request_irq(dev, platform_get_irq(pdev, 0),
+				       vc4_dsi_irq_handler, 0, "vc4 dsi", dsi);
+	}
 	if (ret) {
 		if (ret != -EPROBE_DEFER)
 			dev_err(dev, "Failed to get interrupt: %d\n", ret);
-- 
2.14.2

^ permalink raw reply related	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2017-10-14  8:13 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2017-10-14  0:12 [PATCH v2] drm/vc4: Fix sleeps during the IRQ handler for DSI transactions Eric Anholt
2017-10-14  7:28 ` Paul Menzel
2017-10-14  7:53 ` Boris Brezillon

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox