[PATCH v1 0/3] Remove atomic_exec

[PATCH v1 0/3] Remove atomic_exec

[Bibby Hsieh]

The atomic_exec was designed for processing continuous
packets of commands in atomic way for Mediatek DRM driver.

After we implement flush function, Mediatek DRM driver
doesn't need atomic_exec, the primary concept is to let
Mediatek DRM driver can make sure previous message done or
be aborted (if the message has not started yet) before they
send next message.

Changes since v0:
 - move the binding changes to first
 - remove unnecessary changes

Bibby Hsieh (3):
  dt-binding: gce: remove atomic_exec in mboxes property
  mailbox: mediatek: implement flush function
  mailbox: mediatek: remove implementation related to atomic_exec

 .../devicetree/bindings/mailbox/mtk-gce.txt   |  10 +-
 drivers/mailbox/mtk-cmdq-mailbox.c            | 128 ++++++++----------
 2 files changed, 64 insertions(+), 74 deletions(-)

-- 
2.18.0

[PATCH v1 0/3] Remove atomic_exec

[Bibby Hsieh]

The atomic_exec was designed for processing continuous
packets of commands in atomic way for Mediatek DRM driver.

After we implement flush function, Mediatek DRM driver
doesn't need atomic_exec, the primary concept is to let
Mediatek DRM driver can make sure previous message done or
be aborted (if the message has not started yet) before they
send next message.

Changes since v0:
 - move the binding changes to first
 - remove unnecessary changes

Bibby Hsieh (3):
  dt-binding: gce: remove atomic_exec in mboxes property
  mailbox: mediatek: implement flush function
  mailbox: mediatek: remove implementation related to atomic_exec

 .../devicetree/bindings/mailbox/mtk-gce.txt   |  10 +-
 drivers/mailbox/mtk-cmdq-mailbox.c            | 128 ++++++++----------
 2 files changed, 64 insertions(+), 74 deletions(-)

-- 
2.18.0
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

[PATCH v1 0/3] Remove atomic_exec

[Bibby Hsieh]

The atomic_exec was designed for processing continuous
packets of commands in atomic way for Mediatek DRM driver.

After we implement flush function, Mediatek DRM driver
doesn't need atomic_exec, the primary concept is to let
Mediatek DRM driver can make sure previous message done or
be aborted (if the message has not started yet) before they
send next message.

Changes since v0:
 - move the binding changes to first
 - remove unnecessary changes

Bibby Hsieh (3):
  dt-binding: gce: remove atomic_exec in mboxes property
  mailbox: mediatek: implement flush function
  mailbox: mediatek: remove implementation related to atomic_exec

 .../devicetree/bindings/mailbox/mtk-gce.txt   |  10 +-
 drivers/mailbox/mtk-cmdq-mailbox.c            | 128 ++++++++----------
 2 files changed, 64 insertions(+), 74 deletions(-)

-- 
2.18.0
_______________________________________________
Linux-mediatek mailing list
Linux-mediatek@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-mediatek

[PATCH v1 1/3] dt-binding: gce: remove atomic_exec in mboxes property

[Bibby Hsieh]

There is not any client driver using this feature now,
so remove it from binding.

Signed-off-by: Bibby Hsieh <bibby.hsieh@mediatek.com>
Reviewed-by: CK Hu <ck.hu@mediatek.com>
Reviewed-by: Matthias Brugger <matthias.bgg@gmail.com>
---
 Documentation/devicetree/bindings/mailbox/mtk-gce.txt | 10 ++++------
 1 file changed, 4 insertions(+), 6 deletions(-)

diff --git a/Documentation/devicetree/bindings/mailbox/mtk-gce.txt b/Documentation/devicetree/bindings/mailbox/mtk-gce.txt
index 7b13787ab13d..0b5b2a6bcc48 100644
--- a/Documentation/devicetree/bindings/mailbox/mtk-gce.txt
+++ b/Documentation/devicetree/bindings/mailbox/mtk-gce.txt
@@ -14,13 +14,11 @@ Required properties:
 - interrupts: The interrupt signal from the GCE block
 - clock: Clocks according to the common clock binding
 - clock-names: Must be "gce" to stand for GCE clock
-- #mbox-cells: Should be 3.
-	<&phandle channel priority atomic_exec>
+- #mbox-cells: Should be 2.
+	<&phandle channel priority>
 	phandle: Label name of a gce node.
 	channel: Channel of mailbox. Be equal to the thread id of GCE.
 	priority: Priority of GCE thread.
-	atomic_exec: GCE processing continuous packets of commands in atomic
-		way.
 
 Required properties for a client device:
 - mboxes: Client use mailbox to communicate with GCE, it should have this
@@ -54,8 +52,8 @@ Example for a client device:
 
 	mmsys: clock-controller@14000000 {
 		compatible = "mediatek,mt8173-mmsys";
-		mboxes = <&gce 0 CMDQ_THR_PRIO_LOWEST 1>,
-			 <&gce 1 CMDQ_THR_PRIO_LOWEST 1>;
+		mboxes = <&gce 0 CMDQ_THR_PRIO_LOWEST>,
+			 <&gce 1 CMDQ_THR_PRIO_LOWEST>;
 		mutex-event-eof = <CMDQ_EVENT_MUTEX0_STREAM_EOF
 				CMDQ_EVENT_MUTEX1_STREAM_EOF>;
 		mediatek,gce-client-reg = <&gce SUBSYS_1400XXXX 0x3000 0x1000>,
-- 
2.18.0

[PATCH v1 1/3] dt-binding: gce: remove atomic_exec in mboxes property

[Bibby Hsieh]

There is not any client driver using this feature now,
so remove it from binding.

Signed-off-by: Bibby Hsieh <bibby.hsieh@mediatek.com>
Reviewed-by: CK Hu <ck.hu@mediatek.com>
Reviewed-by: Matthias Brugger <matthias.bgg@gmail.com>
---
 Documentation/devicetree/bindings/mailbox/mtk-gce.txt | 10 ++++------
 1 file changed, 4 insertions(+), 6 deletions(-)

diff --git a/Documentation/devicetree/bindings/mailbox/mtk-gce.txt b/Documentation/devicetree/bindings/mailbox/mtk-gce.txt
index 7b13787ab13d..0b5b2a6bcc48 100644
--- a/Documentation/devicetree/bindings/mailbox/mtk-gce.txt
+++ b/Documentation/devicetree/bindings/mailbox/mtk-gce.txt
@@ -14,13 +14,11 @@ Required properties:
 - interrupts: The interrupt signal from the GCE block
 - clock: Clocks according to the common clock binding
 - clock-names: Must be "gce" to stand for GCE clock
-- #mbox-cells: Should be 3.
-	<&phandle channel priority atomic_exec>
+- #mbox-cells: Should be 2.
+	<&phandle channel priority>
 	phandle: Label name of a gce node.
 	channel: Channel of mailbox. Be equal to the thread id of GCE.
 	priority: Priority of GCE thread.
-	atomic_exec: GCE processing continuous packets of commands in atomic
-		way.
 
 Required properties for a client device:
 - mboxes: Client use mailbox to communicate with GCE, it should have this
@@ -54,8 +52,8 @@ Example for a client device:
 
 	mmsys: clock-controller@14000000 {
 		compatible = "mediatek,mt8173-mmsys";
-		mboxes = <&gce 0 CMDQ_THR_PRIO_LOWEST 1>,
-			 <&gce 1 CMDQ_THR_PRIO_LOWEST 1>;
+		mboxes = <&gce 0 CMDQ_THR_PRIO_LOWEST>,
+			 <&gce 1 CMDQ_THR_PRIO_LOWEST>;
 		mutex-event-eof = <CMDQ_EVENT_MUTEX0_STREAM_EOF
 				CMDQ_EVENT_MUTEX1_STREAM_EOF>;
 		mediatek,gce-client-reg = <&gce SUBSYS_1400XXXX 0x3000 0x1000>,
-- 
2.18.0
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

[PATCH v1 1/3] dt-binding: gce: remove atomic_exec in mboxes property

[Bibby Hsieh]

There is not any client driver using this feature now,
so remove it from binding.

Signed-off-by: Bibby Hsieh <bibby.hsieh@mediatek.com>
Reviewed-by: CK Hu <ck.hu@mediatek.com>
Reviewed-by: Matthias Brugger <matthias.bgg@gmail.com>
---
 Documentation/devicetree/bindings/mailbox/mtk-gce.txt | 10 ++++------
 1 file changed, 4 insertions(+), 6 deletions(-)

diff --git a/Documentation/devicetree/bindings/mailbox/mtk-gce.txt b/Documentation/devicetree/bindings/mailbox/mtk-gce.txt
index 7b13787ab13d..0b5b2a6bcc48 100644
--- a/Documentation/devicetree/bindings/mailbox/mtk-gce.txt
+++ b/Documentation/devicetree/bindings/mailbox/mtk-gce.txt
@@ -14,13 +14,11 @@ Required properties:
 - interrupts: The interrupt signal from the GCE block
 - clock: Clocks according to the common clock binding
 - clock-names: Must be "gce" to stand for GCE clock
-- #mbox-cells: Should be 3.
-	<&phandle channel priority atomic_exec>
+- #mbox-cells: Should be 2.
+	<&phandle channel priority>
 	phandle: Label name of a gce node.
 	channel: Channel of mailbox. Be equal to the thread id of GCE.
 	priority: Priority of GCE thread.
-	atomic_exec: GCE processing continuous packets of commands in atomic
-		way.
 
 Required properties for a client device:
 - mboxes: Client use mailbox to communicate with GCE, it should have this
@@ -54,8 +52,8 @@ Example for a client device:
 
 	mmsys: clock-controller@14000000 {
 		compatible = "mediatek,mt8173-mmsys";
-		mboxes = <&gce 0 CMDQ_THR_PRIO_LOWEST 1>,
-			 <&gce 1 CMDQ_THR_PRIO_LOWEST 1>;
+		mboxes = <&gce 0 CMDQ_THR_PRIO_LOWEST>,
+			 <&gce 1 CMDQ_THR_PRIO_LOWEST>;
 		mutex-event-eof = <CMDQ_EVENT_MUTEX0_STREAM_EOF
 				CMDQ_EVENT_MUTEX1_STREAM_EOF>;
 		mediatek,gce-client-reg = <&gce SUBSYS_1400XXXX 0x3000 0x1000>,
-- 
2.18.0
_______________________________________________
Linux-mediatek mailing list
Linux-mediatek@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-mediatek

Re: [PATCH v1 1/3] dt-binding: gce: remove atomic_exec in mboxes property

[Rob Herring]

On Mon, 17 Feb 2020 17:05:30 +0800, Bibby Hsieh wrote:
> There is not any client driver using this feature now,
> so remove it from binding.
> 
> Signed-off-by: Bibby Hsieh <bibby.hsieh@mediatek.com>
> Reviewed-by: CK Hu <ck.hu@mediatek.com>
> Reviewed-by: Matthias Brugger <matthias.bgg@gmail.com>
> ---
>  Documentation/devicetree/bindings/mailbox/mtk-gce.txt | 10 ++++------
>  1 file changed, 4 insertions(+), 6 deletions(-)
> 

Reviewed-by: Rob Herring <robh@kernel.org>

Re: [PATCH v1 1/3] dt-binding: gce: remove atomic_exec in mboxes property

[Rob Herring]

On Mon, 17 Feb 2020 17:05:30 +0800, Bibby Hsieh wrote:
> There is not any client driver using this feature now,
> so remove it from binding.
> 
> Signed-off-by: Bibby Hsieh <bibby.hsieh@mediatek.com>
> Reviewed-by: CK Hu <ck.hu@mediatek.com>
> Reviewed-by: Matthias Brugger <matthias.bgg@gmail.com>
> ---
>  Documentation/devicetree/bindings/mailbox/mtk-gce.txt | 10 ++++------
>  1 file changed, 4 insertions(+), 6 deletions(-)
> 

Reviewed-by: Rob Herring <robh@kernel.org>

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

Re: [PATCH v1 1/3] dt-binding: gce: remove atomic_exec in mboxes property

[Rob Herring]

On Mon, 17 Feb 2020 17:05:30 +0800, Bibby Hsieh wrote:
> There is not any client driver using this feature now,
> so remove it from binding.
> 
> Signed-off-by: Bibby Hsieh <bibby.hsieh@mediatek.com>
> Reviewed-by: CK Hu <ck.hu@mediatek.com>
> Reviewed-by: Matthias Brugger <matthias.bgg@gmail.com>
> ---
>  Documentation/devicetree/bindings/mailbox/mtk-gce.txt | 10 ++++------
>  1 file changed, 4 insertions(+), 6 deletions(-)
> 

Reviewed-by: Rob Herring <robh@kernel.org>

_______________________________________________
Linux-mediatek mailing list
Linux-mediatek@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-mediatek

[PATCH v1 2/3] mailbox: mediatek: implement flush function

[Bibby Hsieh]

For client driver which need to reorganize the command buffer, it could
use this function to flush the send command buffer.
If the channel doesn't be started (usually in waiting for event), this
function will abort it directly.

Signed-off-by: Bibby Hsieh <bibby.hsieh@mediatek.com>
---
 drivers/mailbox/mtk-cmdq-mailbox.c | 52 ++++++++++++++++++++++++++++++
 1 file changed, 52 insertions(+)

diff --git a/drivers/mailbox/mtk-cmdq-mailbox.c b/drivers/mailbox/mtk-cmdq-mailbox.c
index 9a6ce9f5a7db..0da5e2dc2c0e 100644
--- a/drivers/mailbox/mtk-cmdq-mailbox.c
+++ b/drivers/mailbox/mtk-cmdq-mailbox.c
@@ -432,10 +432,62 @@ static void cmdq_mbox_shutdown(struct mbox_chan *chan)
 {
 }
 
+static int cmdq_mbox_flush(struct mbox_chan *chan, unsigned long timeout)
+{
+	struct cmdq_thread *thread = (struct cmdq_thread *)chan->con_priv;
+	struct cmdq_task_cb *cb;
+	struct cmdq_cb_data data;
+	struct cmdq *cmdq = dev_get_drvdata(chan->mbox->dev);
+	struct cmdq_task *task, *tmp;
+	unsigned long flags;
+	u32 enable;
+
+	spin_lock_irqsave(&thread->chan->lock, flags);
+	if (list_empty(&thread->task_busy_list))
+		goto out;
+
+	WARN_ON(cmdq_thread_suspend(cmdq, thread) < 0);
+	if (!cmdq_thread_is_in_wfe(thread))
+		goto wait;
+
+	list_for_each_entry_safe(task, tmp, &thread->task_busy_list,
+				 list_entry) {
+		cb = &task->pkt->async_cb;
+		if (cb->cb) {
+			data.sta = CMDQ_CB_ERROR;
+			data.data = cb->data;
+			cb->cb(data);
+		}
+		list_del(&task->list_entry);
+		kfree(task);
+	}
+
+	cmdq_thread_resume(thread);
+	cmdq_thread_disable(cmdq, thread);
+	clk_disable(cmdq->clock);
+
+out:
+	spin_unlock_irqrestore(&thread->chan->lock, flags);
+	return 0;
+
+wait:
+	cmdq_thread_resume(thread);
+	spin_unlock_irqrestore(&thread->chan->lock, flags);
+	if (readl_poll_timeout_atomic(thread->base + CMDQ_THR_ENABLE_TASK,
+				      enable, enable == 0, 1, timeout)) {
+		dev_err(cmdq->mbox.dev, "Fail to wait GCE thread 0x%x done\n",
+			(u32)(thread->base - cmdq->base));
+
+		return -EFAULT;
+	}
+	return 0;
+}
+
 static const struct mbox_chan_ops cmdq_mbox_chan_ops = {
 	.send_data = cmdq_mbox_send_data,
 	.startup = cmdq_mbox_startup,
 	.shutdown = cmdq_mbox_shutdown,
+	.flush = cmdq_mbox_flush,
 };
 
 static struct mbox_chan *cmdq_xlate(struct mbox_controller *mbox,
-- 
2.18.0

[PATCH v1 2/3] mailbox: mediatek: implement flush function

[Bibby Hsieh]

For client driver which need to reorganize the command buffer, it could
use this function to flush the send command buffer.
If the channel doesn't be started (usually in waiting for event), this
function will abort it directly.

Signed-off-by: Bibby Hsieh <bibby.hsieh@mediatek.com>
---
 drivers/mailbox/mtk-cmdq-mailbox.c | 52 ++++++++++++++++++++++++++++++
 1 file changed, 52 insertions(+)

diff --git a/drivers/mailbox/mtk-cmdq-mailbox.c b/drivers/mailbox/mtk-cmdq-mailbox.c
index 9a6ce9f5a7db..0da5e2dc2c0e 100644
--- a/drivers/mailbox/mtk-cmdq-mailbox.c
+++ b/drivers/mailbox/mtk-cmdq-mailbox.c
@@ -432,10 +432,62 @@ static void cmdq_mbox_shutdown(struct mbox_chan *chan)
 {
 }
 
+static int cmdq_mbox_flush(struct mbox_chan *chan, unsigned long timeout)
+{
+	struct cmdq_thread *thread = (struct cmdq_thread *)chan->con_priv;
+	struct cmdq_task_cb *cb;
+	struct cmdq_cb_data data;
+	struct cmdq *cmdq = dev_get_drvdata(chan->mbox->dev);
+	struct cmdq_task *task, *tmp;
+	unsigned long flags;
+	u32 enable;
+
+	spin_lock_irqsave(&thread->chan->lock, flags);
+	if (list_empty(&thread->task_busy_list))
+		goto out;
+
+	WARN_ON(cmdq_thread_suspend(cmdq, thread) < 0);
+	if (!cmdq_thread_is_in_wfe(thread))
+		goto wait;
+
+	list_for_each_entry_safe(task, tmp, &thread->task_busy_list,
+				 list_entry) {
+		cb = &task->pkt->async_cb;
+		if (cb->cb) {
+			data.sta = CMDQ_CB_ERROR;
+			data.data = cb->data;
+			cb->cb(data);
+		}
+		list_del(&task->list_entry);
+		kfree(task);
+	}
+
+	cmdq_thread_resume(thread);
+	cmdq_thread_disable(cmdq, thread);
+	clk_disable(cmdq->clock);
+
+out:
+	spin_unlock_irqrestore(&thread->chan->lock, flags);
+	return 0;
+
+wait:
+	cmdq_thread_resume(thread);
+	spin_unlock_irqrestore(&thread->chan->lock, flags);
+	if (readl_poll_timeout_atomic(thread->base + CMDQ_THR_ENABLE_TASK,
+				      enable, enable == 0, 1, timeout)) {
+		dev_err(cmdq->mbox.dev, "Fail to wait GCE thread 0x%x done\n",
+			(u32)(thread->base - cmdq->base));
+
+		return -EFAULT;
+	}
+	return 0;
+}
+
 static const struct mbox_chan_ops cmdq_mbox_chan_ops = {
 	.send_data = cmdq_mbox_send_data,
 	.startup = cmdq_mbox_startup,
 	.shutdown = cmdq_mbox_shutdown,
+	.flush = cmdq_mbox_flush,
 };
 
 static struct mbox_chan *cmdq_xlate(struct mbox_controller *mbox,
-- 
2.18.0
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

[PATCH v1 2/3] mailbox: mediatek: implement flush function

[Bibby Hsieh]

For client driver which need to reorganize the command buffer, it could
use this function to flush the send command buffer.
If the channel doesn't be started (usually in waiting for event), this
function will abort it directly.

Signed-off-by: Bibby Hsieh <bibby.hsieh@mediatek.com>
---
 drivers/mailbox/mtk-cmdq-mailbox.c | 52 ++++++++++++++++++++++++++++++
 1 file changed, 52 insertions(+)

diff --git a/drivers/mailbox/mtk-cmdq-mailbox.c b/drivers/mailbox/mtk-cmdq-mailbox.c
index 9a6ce9f5a7db..0da5e2dc2c0e 100644
--- a/drivers/mailbox/mtk-cmdq-mailbox.c
+++ b/drivers/mailbox/mtk-cmdq-mailbox.c
@@ -432,10 +432,62 @@ static void cmdq_mbox_shutdown(struct mbox_chan *chan)
 {
 }
 
+static int cmdq_mbox_flush(struct mbox_chan *chan, unsigned long timeout)
+{
+	struct cmdq_thread *thread = (struct cmdq_thread *)chan->con_priv;
+	struct cmdq_task_cb *cb;
+	struct cmdq_cb_data data;
+	struct cmdq *cmdq = dev_get_drvdata(chan->mbox->dev);
+	struct cmdq_task *task, *tmp;
+	unsigned long flags;
+	u32 enable;
+
+	spin_lock_irqsave(&thread->chan->lock, flags);
+	if (list_empty(&thread->task_busy_list))
+		goto out;
+
+	WARN_ON(cmdq_thread_suspend(cmdq, thread) < 0);
+	if (!cmdq_thread_is_in_wfe(thread))
+		goto wait;
+
+	list_for_each_entry_safe(task, tmp, &thread->task_busy_list,
+				 list_entry) {
+		cb = &task->pkt->async_cb;
+		if (cb->cb) {
+			data.sta = CMDQ_CB_ERROR;
+			data.data = cb->data;
+			cb->cb(data);
+		}
+		list_del(&task->list_entry);
+		kfree(task);
+	}
+
+	cmdq_thread_resume(thread);
+	cmdq_thread_disable(cmdq, thread);
+	clk_disable(cmdq->clock);
+
+out:
+	spin_unlock_irqrestore(&thread->chan->lock, flags);
+	return 0;
+
+wait:
+	cmdq_thread_resume(thread);
+	spin_unlock_irqrestore(&thread->chan->lock, flags);
+	if (readl_poll_timeout_atomic(thread->base + CMDQ_THR_ENABLE_TASK,
+				      enable, enable == 0, 1, timeout)) {
+		dev_err(cmdq->mbox.dev, "Fail to wait GCE thread 0x%x done\n",
+			(u32)(thread->base - cmdq->base));
+
+		return -EFAULT;
+	}
+	return 0;
+}
+
 static const struct mbox_chan_ops cmdq_mbox_chan_ops = {
 	.send_data = cmdq_mbox_send_data,
 	.startup = cmdq_mbox_startup,
 	.shutdown = cmdq_mbox_shutdown,
+	.flush = cmdq_mbox_flush,
 };
 
 static struct mbox_chan *cmdq_xlate(struct mbox_controller *mbox,
-- 
2.18.0
_______________________________________________
Linux-mediatek mailing list
Linux-mediatek@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-mediatek

Re: [PATCH v1 2/3] mailbox: mediatek: implement flush function

[CK Hu]

Hi, Bibby:

On Mon, 2020-02-17 at 17:05 +0800, Bibby Hsieh wrote:
> For client driver which need to reorganize the command buffer, it could
> use this function to flush the send command buffer.
> If the channel doesn't be started (usually in waiting for event), this
> function will abort it directly.
> 

Reviewed-by: CK Hu <ck.hu@mediatek.com>

> Signed-off-by: Bibby Hsieh <bibby.hsieh@mediatek.com>
> ---
>  drivers/mailbox/mtk-cmdq-mailbox.c | 52 ++++++++++++++++++++++++++++++
>  1 file changed, 52 insertions(+)
> 
> diff --git a/drivers/mailbox/mtk-cmdq-mailbox.c b/drivers/mailbox/mtk-cmdq-mailbox.c
> index 9a6ce9f5a7db..0da5e2dc2c0e 100644
> --- a/drivers/mailbox/mtk-cmdq-mailbox.c
> +++ b/drivers/mailbox/mtk-cmdq-mailbox.c
> @@ -432,10 +432,62 @@ static void cmdq_mbox_shutdown(struct mbox_chan *chan)
>  {
>  }
>  
> +static int cmdq_mbox_flush(struct mbox_chan *chan, unsigned long timeout)
> +{
> +	struct cmdq_thread *thread = (struct cmdq_thread *)chan->con_priv;
> +	struct cmdq_task_cb *cb;
> +	struct cmdq_cb_data data;
> +	struct cmdq *cmdq = dev_get_drvdata(chan->mbox->dev);
> +	struct cmdq_task *task, *tmp;
> +	unsigned long flags;
> +	u32 enable;
> +
> +	spin_lock_irqsave(&thread->chan->lock, flags);
> +	if (list_empty(&thread->task_busy_list))
> +		goto out;
> +
> +	WARN_ON(cmdq_thread_suspend(cmdq, thread) < 0);
> +	if (!cmdq_thread_is_in_wfe(thread))
> +		goto wait;
> +
> +	list_for_each_entry_safe(task, tmp, &thread->task_busy_list,
> +				 list_entry) {
> +		cb = &task->pkt->async_cb;
> +		if (cb->cb) {
> +			data.sta = CMDQ_CB_ERROR;
> +			data.data = cb->data;
> +			cb->cb(data);
> +		}
> +		list_del(&task->list_entry);
> +		kfree(task);
> +	}
> +
> +	cmdq_thread_resume(thread);
> +	cmdq_thread_disable(cmdq, thread);
> +	clk_disable(cmdq->clock);
> +
> +out:
> +	spin_unlock_irqrestore(&thread->chan->lock, flags);
> +	return 0;
> +
> +wait:
> +	cmdq_thread_resume(thread);
> +	spin_unlock_irqrestore(&thread->chan->lock, flags);
> +	if (readl_poll_timeout_atomic(thread->base + CMDQ_THR_ENABLE_TASK,
> +				      enable, enable == 0, 1, timeout)) {
> +		dev_err(cmdq->mbox.dev, "Fail to wait GCE thread 0x%x done\n",
> +			(u32)(thread->base - cmdq->base));
> +
> +		return -EFAULT;
> +	}
> +	return 0;
> +}
> +
>  static const struct mbox_chan_ops cmdq_mbox_chan_ops = {
>  	.send_data = cmdq_mbox_send_data,
>  	.startup = cmdq_mbox_startup,
>  	.shutdown = cmdq_mbox_shutdown,
> +	.flush = cmdq_mbox_flush,
>  };
>  
>  static struct mbox_chan *cmdq_xlate(struct mbox_controller *mbox,

Re: [PATCH v1 2/3] mailbox: mediatek: implement flush function

[CK Hu]

Hi, Bibby:

On Mon, 2020-02-17 at 17:05 +0800, Bibby Hsieh wrote:
> For client driver which need to reorganize the command buffer, it could
> use this function to flush the send command buffer.
> If the channel doesn't be started (usually in waiting for event), this
> function will abort it directly.
> 

Reviewed-by: CK Hu <ck.hu@mediatek.com>

> Signed-off-by: Bibby Hsieh <bibby.hsieh@mediatek.com>
> ---
>  drivers/mailbox/mtk-cmdq-mailbox.c | 52 ++++++++++++++++++++++++++++++
>  1 file changed, 52 insertions(+)
> 
> diff --git a/drivers/mailbox/mtk-cmdq-mailbox.c b/drivers/mailbox/mtk-cmdq-mailbox.c
> index 9a6ce9f5a7db..0da5e2dc2c0e 100644
> --- a/drivers/mailbox/mtk-cmdq-mailbox.c
> +++ b/drivers/mailbox/mtk-cmdq-mailbox.c
> @@ -432,10 +432,62 @@ static void cmdq_mbox_shutdown(struct mbox_chan *chan)
>  {
>  }
>  
> +static int cmdq_mbox_flush(struct mbox_chan *chan, unsigned long timeout)
> +{
> +	struct cmdq_thread *thread = (struct cmdq_thread *)chan->con_priv;
> +	struct cmdq_task_cb *cb;
> +	struct cmdq_cb_data data;
> +	struct cmdq *cmdq = dev_get_drvdata(chan->mbox->dev);
> +	struct cmdq_task *task, *tmp;
> +	unsigned long flags;
> +	u32 enable;
> +
> +	spin_lock_irqsave(&thread->chan->lock, flags);
> +	if (list_empty(&thread->task_busy_list))
> +		goto out;
> +
> +	WARN_ON(cmdq_thread_suspend(cmdq, thread) < 0);
> +	if (!cmdq_thread_is_in_wfe(thread))
> +		goto wait;
> +
> +	list_for_each_entry_safe(task, tmp, &thread->task_busy_list,
> +				 list_entry) {
> +		cb = &task->pkt->async_cb;
> +		if (cb->cb) {
> +			data.sta = CMDQ_CB_ERROR;
> +			data.data = cb->data;
> +			cb->cb(data);
> +		}
> +		list_del(&task->list_entry);
> +		kfree(task);
> +	}
> +
> +	cmdq_thread_resume(thread);
> +	cmdq_thread_disable(cmdq, thread);
> +	clk_disable(cmdq->clock);
> +
> +out:
> +	spin_unlock_irqrestore(&thread->chan->lock, flags);
> +	return 0;
> +
> +wait:
> +	cmdq_thread_resume(thread);
> +	spin_unlock_irqrestore(&thread->chan->lock, flags);
> +	if (readl_poll_timeout_atomic(thread->base + CMDQ_THR_ENABLE_TASK,
> +				      enable, enable == 0, 1, timeout)) {
> +		dev_err(cmdq->mbox.dev, "Fail to wait GCE thread 0x%x done\n",
> +			(u32)(thread->base - cmdq->base));
> +
> +		return -EFAULT;
> +	}
> +	return 0;
> +}
> +
>  static const struct mbox_chan_ops cmdq_mbox_chan_ops = {
>  	.send_data = cmdq_mbox_send_data,
>  	.startup = cmdq_mbox_startup,
>  	.shutdown = cmdq_mbox_shutdown,
> +	.flush = cmdq_mbox_flush,
>  };
>  
>  static struct mbox_chan *cmdq_xlate(struct mbox_controller *mbox,

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

Re: [PATCH v1 2/3] mailbox: mediatek: implement flush function

[CK Hu]

Hi, Bibby:

On Mon, 2020-02-17 at 17:05 +0800, Bibby Hsieh wrote:
> For client driver which need to reorganize the command buffer, it could
> use this function to flush the send command buffer.
> If the channel doesn't be started (usually in waiting for event), this
> function will abort it directly.
> 

Reviewed-by: CK Hu <ck.hu@mediatek.com>

> Signed-off-by: Bibby Hsieh <bibby.hsieh@mediatek.com>
> ---
>  drivers/mailbox/mtk-cmdq-mailbox.c | 52 ++++++++++++++++++++++++++++++
>  1 file changed, 52 insertions(+)
> 
> diff --git a/drivers/mailbox/mtk-cmdq-mailbox.c b/drivers/mailbox/mtk-cmdq-mailbox.c
> index 9a6ce9f5a7db..0da5e2dc2c0e 100644
> --- a/drivers/mailbox/mtk-cmdq-mailbox.c
> +++ b/drivers/mailbox/mtk-cmdq-mailbox.c
> @@ -432,10 +432,62 @@ static void cmdq_mbox_shutdown(struct mbox_chan *chan)
>  {
>  }
>  
> +static int cmdq_mbox_flush(struct mbox_chan *chan, unsigned long timeout)
> +{
> +	struct cmdq_thread *thread = (struct cmdq_thread *)chan->con_priv;
> +	struct cmdq_task_cb *cb;
> +	struct cmdq_cb_data data;
> +	struct cmdq *cmdq = dev_get_drvdata(chan->mbox->dev);
> +	struct cmdq_task *task, *tmp;
> +	unsigned long flags;
> +	u32 enable;
> +
> +	spin_lock_irqsave(&thread->chan->lock, flags);
> +	if (list_empty(&thread->task_busy_list))
> +		goto out;
> +
> +	WARN_ON(cmdq_thread_suspend(cmdq, thread) < 0);
> +	if (!cmdq_thread_is_in_wfe(thread))
> +		goto wait;
> +
> +	list_for_each_entry_safe(task, tmp, &thread->task_busy_list,
> +				 list_entry) {
> +		cb = &task->pkt->async_cb;
> +		if (cb->cb) {
> +			data.sta = CMDQ_CB_ERROR;
> +			data.data = cb->data;
> +			cb->cb(data);
> +		}
> +		list_del(&task->list_entry);
> +		kfree(task);
> +	}
> +
> +	cmdq_thread_resume(thread);
> +	cmdq_thread_disable(cmdq, thread);
> +	clk_disable(cmdq->clock);
> +
> +out:
> +	spin_unlock_irqrestore(&thread->chan->lock, flags);
> +	return 0;
> +
> +wait:
> +	cmdq_thread_resume(thread);
> +	spin_unlock_irqrestore(&thread->chan->lock, flags);
> +	if (readl_poll_timeout_atomic(thread->base + CMDQ_THR_ENABLE_TASK,
> +				      enable, enable == 0, 1, timeout)) {
> +		dev_err(cmdq->mbox.dev, "Fail to wait GCE thread 0x%x done\n",
> +			(u32)(thread->base - cmdq->base));
> +
> +		return -EFAULT;
> +	}
> +	return 0;
> +}
> +
>  static const struct mbox_chan_ops cmdq_mbox_chan_ops = {
>  	.send_data = cmdq_mbox_send_data,
>  	.startup = cmdq_mbox_startup,
>  	.shutdown = cmdq_mbox_shutdown,
> +	.flush = cmdq_mbox_flush,
>  };
>  
>  static struct mbox_chan *cmdq_xlate(struct mbox_controller *mbox,

_______________________________________________
Linux-mediatek mailing list
Linux-mediatek@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-mediatek

Re: [PATCH v1 2/3] mailbox: mediatek: implement flush function

[Bibby Hsieh]

Hi, Jassi,

https://patchwork.kernel.org/cover/11385839/

I push a series of patches about remove atomic_exec feature in Mediatek
CMDQ driver. Could you please review them if you are free?

Thanks

On Mon, 2020-02-17 at 17:29 +0800, CK Hu wrote:
> Hi, Bibby:
> 
> On Mon, 2020-02-17 at 17:05 +0800, Bibby Hsieh wrote:
> > For client driver which need to reorganize the command buffer, it could
> > use this function to flush the send command buffer.
> > If the channel doesn't be started (usually in waiting for event), this
> > function will abort it directly.
> > 
> 
> Reviewed-by: CK Hu <ck.hu@mediatek.com>
> 
> > Signed-off-by: Bibby Hsieh <bibby.hsieh@mediatek.com>
> > ---
> >  drivers/mailbox/mtk-cmdq-mailbox.c | 52 ++++++++++++++++++++++++++++++
> >  1 file changed, 52 insertions(+)
> > 
> > diff --git a/drivers/mailbox/mtk-cmdq-mailbox.c b/drivers/mailbox/mtk-cmdq-mailbox.c
> > index 9a6ce9f5a7db..0da5e2dc2c0e 100644
> > --- a/drivers/mailbox/mtk-cmdq-mailbox.c
> > +++ b/drivers/mailbox/mtk-cmdq-mailbox.c
> > @@ -432,10 +432,62 @@ static void cmdq_mbox_shutdown(struct mbox_chan *chan)
> >  {
> >  }
> >  
> > +static int cmdq_mbox_flush(struct mbox_chan *chan, unsigned long timeout)
> > +{
> > +	struct cmdq_thread *thread = (struct cmdq_thread *)chan->con_priv;
> > +	struct cmdq_task_cb *cb;
> > +	struct cmdq_cb_data data;
> > +	struct cmdq *cmdq = dev_get_drvdata(chan->mbox->dev);
> > +	struct cmdq_task *task, *tmp;
> > +	unsigned long flags;
> > +	u32 enable;
> > +
> > +	spin_lock_irqsave(&thread->chan->lock, flags);
> > +	if (list_empty(&thread->task_busy_list))
> > +		goto out;
> > +
> > +	WARN_ON(cmdq_thread_suspend(cmdq, thread) < 0);
> > +	if (!cmdq_thread_is_in_wfe(thread))
> > +		goto wait;
> > +
> > +	list_for_each_entry_safe(task, tmp, &thread->task_busy_list,
> > +				 list_entry) {
> > +		cb = &task->pkt->async_cb;
> > +		if (cb->cb) {
> > +			data.sta = CMDQ_CB_ERROR;
> > +			data.data = cb->data;
> > +			cb->cb(data);
> > +		}
> > +		list_del(&task->list_entry);
> > +		kfree(task);
> > +	}
> > +
> > +	cmdq_thread_resume(thread);
> > +	cmdq_thread_disable(cmdq, thread);
> > +	clk_disable(cmdq->clock);
> > +
> > +out:
> > +	spin_unlock_irqrestore(&thread->chan->lock, flags);
> > +	return 0;
> > +
> > +wait:
> > +	cmdq_thread_resume(thread);
> > +	spin_unlock_irqrestore(&thread->chan->lock, flags);
> > +	if (readl_poll_timeout_atomic(thread->base + CMDQ_THR_ENABLE_TASK,
> > +				      enable, enable == 0, 1, timeout)) {
> > +		dev_err(cmdq->mbox.dev, "Fail to wait GCE thread 0x%x done\n",
> > +			(u32)(thread->base - cmdq->base));
> > +
> > +		return -EFAULT;
> > +	}
> > +	return 0;
> > +}
> > +
> >  static const struct mbox_chan_ops cmdq_mbox_chan_ops = {
> >  	.send_data = cmdq_mbox_send_data,
> >  	.startup = cmdq_mbox_startup,
> >  	.shutdown = cmdq_mbox_shutdown,
> > +	.flush = cmdq_mbox_flush,
> >  };
> >  
> >  static struct mbox_chan *cmdq_xlate(struct mbox_controller *mbox,
> 
> 

Re: [PATCH v1 2/3] mailbox: mediatek: implement flush function

[Bibby Hsieh]

Hi, Jassi,

https://patchwork.kernel.org/cover/11385839/

I push a series of patches about remove atomic_exec feature in Mediatek
CMDQ driver. Could you please review them if you are free?

Thanks

On Mon, 2020-02-17 at 17:29 +0800, CK Hu wrote:
> Hi, Bibby:
> 
> On Mon, 2020-02-17 at 17:05 +0800, Bibby Hsieh wrote:
> > For client driver which need to reorganize the command buffer, it could
> > use this function to flush the send command buffer.
> > If the channel doesn't be started (usually in waiting for event), this
> > function will abort it directly.
> > 
> 
> Reviewed-by: CK Hu <ck.hu@mediatek.com>
> 
> > Signed-off-by: Bibby Hsieh <bibby.hsieh@mediatek.com>
> > ---
> >  drivers/mailbox/mtk-cmdq-mailbox.c | 52 ++++++++++++++++++++++++++++++
> >  1 file changed, 52 insertions(+)
> > 
> > diff --git a/drivers/mailbox/mtk-cmdq-mailbox.c b/drivers/mailbox/mtk-cmdq-mailbox.c
> > index 9a6ce9f5a7db..0da5e2dc2c0e 100644
> > --- a/drivers/mailbox/mtk-cmdq-mailbox.c
> > +++ b/drivers/mailbox/mtk-cmdq-mailbox.c
> > @@ -432,10 +432,62 @@ static void cmdq_mbox_shutdown(struct mbox_chan *chan)
> >  {
> >  }
> >  
> > +static int cmdq_mbox_flush(struct mbox_chan *chan, unsigned long timeout)
> > +{
> > +	struct cmdq_thread *thread = (struct cmdq_thread *)chan->con_priv;
> > +	struct cmdq_task_cb *cb;
> > +	struct cmdq_cb_data data;
> > +	struct cmdq *cmdq = dev_get_drvdata(chan->mbox->dev);
> > +	struct cmdq_task *task, *tmp;
> > +	unsigned long flags;
> > +	u32 enable;
> > +
> > +	spin_lock_irqsave(&thread->chan->lock, flags);
> > +	if (list_empty(&thread->task_busy_list))
> > +		goto out;
> > +
> > +	WARN_ON(cmdq_thread_suspend(cmdq, thread) < 0);
> > +	if (!cmdq_thread_is_in_wfe(thread))
> > +		goto wait;
> > +
> > +	list_for_each_entry_safe(task, tmp, &thread->task_busy_list,
> > +				 list_entry) {
> > +		cb = &task->pkt->async_cb;
> > +		if (cb->cb) {
> > +			data.sta = CMDQ_CB_ERROR;
> > +			data.data = cb->data;
> > +			cb->cb(data);
> > +		}
> > +		list_del(&task->list_entry);
> > +		kfree(task);
> > +	}
> > +
> > +	cmdq_thread_resume(thread);
> > +	cmdq_thread_disable(cmdq, thread);
> > +	clk_disable(cmdq->clock);
> > +
> > +out:
> > +	spin_unlock_irqrestore(&thread->chan->lock, flags);
> > +	return 0;
> > +
> > +wait:
> > +	cmdq_thread_resume(thread);
> > +	spin_unlock_irqrestore(&thread->chan->lock, flags);
> > +	if (readl_poll_timeout_atomic(thread->base + CMDQ_THR_ENABLE_TASK,
> > +				      enable, enable == 0, 1, timeout)) {
> > +		dev_err(cmdq->mbox.dev, "Fail to wait GCE thread 0x%x done\n",
> > +			(u32)(thread->base - cmdq->base));
> > +
> > +		return -EFAULT;
> > +	}
> > +	return 0;
> > +}
> > +
> >  static const struct mbox_chan_ops cmdq_mbox_chan_ops = {
> >  	.send_data = cmdq_mbox_send_data,
> >  	.startup = cmdq_mbox_startup,
> >  	.shutdown = cmdq_mbox_shutdown,
> > +	.flush = cmdq_mbox_flush,
> >  };
> >  
> >  static struct mbox_chan *cmdq_xlate(struct mbox_controller *mbox,
> 
> 

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

Re: [PATCH v1 2/3] mailbox: mediatek: implement flush function

[Bibby Hsieh]

Hi, Jassi,

https://patchwork.kernel.org/cover/11385839/

I push a series of patches about remove atomic_exec feature in Mediatek
CMDQ driver. Could you please review them if you are free?

Thanks

On Mon, 2020-02-17 at 17:29 +0800, CK Hu wrote:
> Hi, Bibby:
> 
> On Mon, 2020-02-17 at 17:05 +0800, Bibby Hsieh wrote:
> > For client driver which need to reorganize the command buffer, it could
> > use this function to flush the send command buffer.
> > If the channel doesn't be started (usually in waiting for event), this
> > function will abort it directly.
> > 
> 
> Reviewed-by: CK Hu <ck.hu@mediatek.com>
> 
> > Signed-off-by: Bibby Hsieh <bibby.hsieh@mediatek.com>
> > ---
> >  drivers/mailbox/mtk-cmdq-mailbox.c | 52 ++++++++++++++++++++++++++++++
> >  1 file changed, 52 insertions(+)
> > 
> > diff --git a/drivers/mailbox/mtk-cmdq-mailbox.c b/drivers/mailbox/mtk-cmdq-mailbox.c
> > index 9a6ce9f5a7db..0da5e2dc2c0e 100644
> > --- a/drivers/mailbox/mtk-cmdq-mailbox.c
> > +++ b/drivers/mailbox/mtk-cmdq-mailbox.c
> > @@ -432,10 +432,62 @@ static void cmdq_mbox_shutdown(struct mbox_chan *chan)
> >  {
> >  }
> >  
> > +static int cmdq_mbox_flush(struct mbox_chan *chan, unsigned long timeout)
> > +{
> > +	struct cmdq_thread *thread = (struct cmdq_thread *)chan->con_priv;
> > +	struct cmdq_task_cb *cb;
> > +	struct cmdq_cb_data data;
> > +	struct cmdq *cmdq = dev_get_drvdata(chan->mbox->dev);
> > +	struct cmdq_task *task, *tmp;
> > +	unsigned long flags;
> > +	u32 enable;
> > +
> > +	spin_lock_irqsave(&thread->chan->lock, flags);
> > +	if (list_empty(&thread->task_busy_list))
> > +		goto out;
> > +
> > +	WARN_ON(cmdq_thread_suspend(cmdq, thread) < 0);
> > +	if (!cmdq_thread_is_in_wfe(thread))
> > +		goto wait;
> > +
> > +	list_for_each_entry_safe(task, tmp, &thread->task_busy_list,
> > +				 list_entry) {
> > +		cb = &task->pkt->async_cb;
> > +		if (cb->cb) {
> > +			data.sta = CMDQ_CB_ERROR;
> > +			data.data = cb->data;
> > +			cb->cb(data);
> > +		}
> > +		list_del(&task->list_entry);
> > +		kfree(task);
> > +	}
> > +
> > +	cmdq_thread_resume(thread);
> > +	cmdq_thread_disable(cmdq, thread);
> > +	clk_disable(cmdq->clock);
> > +
> > +out:
> > +	spin_unlock_irqrestore(&thread->chan->lock, flags);
> > +	return 0;
> > +
> > +wait:
> > +	cmdq_thread_resume(thread);
> > +	spin_unlock_irqrestore(&thread->chan->lock, flags);
> > +	if (readl_poll_timeout_atomic(thread->base + CMDQ_THR_ENABLE_TASK,
> > +				      enable, enable == 0, 1, timeout)) {
> > +		dev_err(cmdq->mbox.dev, "Fail to wait GCE thread 0x%x done\n",
> > +			(u32)(thread->base - cmdq->base));
> > +
> > +		return -EFAULT;
> > +	}
> > +	return 0;
> > +}
> > +
> >  static const struct mbox_chan_ops cmdq_mbox_chan_ops = {
> >  	.send_data = cmdq_mbox_send_data,
> >  	.startup = cmdq_mbox_startup,
> >  	.shutdown = cmdq_mbox_shutdown,
> > +	.flush = cmdq_mbox_flush,
> >  };
> >  
> >  static struct mbox_chan *cmdq_xlate(struct mbox_controller *mbox,
> 
> 

_______________________________________________
Linux-mediatek mailing list
Linux-mediatek@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-mediatek

[PATCH v1 3/3] mailbox: mediatek: remove implementation related to atomic_exec

[Bibby Hsieh]

After implement flush, client can flush the executing
command buffer or abort the still waiting for event
command buffer, so controller do not need to implement
atomic_exe feature. remove it.

Signed-off-by: Bibby Hsieh <bibby.hsieh@mediatek.com>
Reviewed-by: CK Hu <ck.hu@mediatek.com>
---
 drivers/mailbox/mtk-cmdq-mailbox.c | 76 ++++--------------------------
 1 file changed, 8 insertions(+), 68 deletions(-)

diff --git a/drivers/mailbox/mtk-cmdq-mailbox.c b/drivers/mailbox/mtk-cmdq-mailbox.c
index 0da5e2dc2c0e..b24822ad8409 100644
--- a/drivers/mailbox/mtk-cmdq-mailbox.c
+++ b/drivers/mailbox/mtk-cmdq-mailbox.c
@@ -56,7 +56,6 @@ struct cmdq_thread {
 	void __iomem		*base;
 	struct list_head	task_busy_list;
 	u32			priority;
-	bool			atomic_exec;
 };
 
 struct cmdq_task {
@@ -162,48 +161,11 @@ static void cmdq_task_insert_into_thread(struct cmdq_task *task)
 	cmdq_thread_invalidate_fetched_data(thread);
 }
 
-static bool cmdq_command_is_wfe(u64 cmd)
-{
-	u64 wfe_option = CMDQ_WFE_UPDATE | CMDQ_WFE_WAIT | CMDQ_WFE_WAIT_VALUE;
-	u64 wfe_op = (u64)(CMDQ_CODE_WFE << CMDQ_OP_CODE_SHIFT) << 32;
-	u64 wfe_mask = (u64)CMDQ_OP_CODE_MASK << 32 | 0xffffffff;
-
-	return ((cmd & wfe_mask) == (wfe_op | wfe_option));
-}
-
-/* we assume tasks in the same display GCE thread are waiting the same event. */
-static void cmdq_task_remove_wfe(struct cmdq_task *task)
-{
-	struct device *dev = task->cmdq->mbox.dev;
-	u64 *base = task->pkt->va_base;
-	int i;
-
-	dma_sync_single_for_cpu(dev, task->pa_base, task->pkt->cmd_buf_size,
-				DMA_TO_DEVICE);
-	for (i = 0; i < CMDQ_NUM_CMD(task->pkt); i++)
-		if (cmdq_command_is_wfe(base[i]))
-			base[i] = (u64)CMDQ_JUMP_BY_OFFSET << 32 |
-				  CMDQ_JUMP_PASS;
-	dma_sync_single_for_device(dev, task->pa_base, task->pkt->cmd_buf_size,
-				   DMA_TO_DEVICE);
-}
-
 static bool cmdq_thread_is_in_wfe(struct cmdq_thread *thread)
 {
 	return readl(thread->base + CMDQ_THR_WAIT_TOKEN) & CMDQ_THR_IS_WAITING;
 }
 
-static void cmdq_thread_wait_end(struct cmdq_thread *thread,
-				 unsigned long end_pa)
-{
-	struct device *dev = thread->chan->mbox->dev;
-	unsigned long curr_pa;
-
-	if (readl_poll_timeout_atomic(thread->base + CMDQ_THR_CURR_ADDR,
-			curr_pa, curr_pa == end_pa, 1, 20))
-		dev_err(dev, "GCE thread cannot run to end.\n");
-}
-
 static void cmdq_task_exec_done(struct cmdq_task *task, enum cmdq_cb_status sta)
 {
 	struct cmdq_task_cb *cb = &task->pkt->async_cb;
@@ -383,36 +345,15 @@ static int cmdq_mbox_send_data(struct mbox_chan *chan, void *data)
 		WARN_ON(cmdq_thread_suspend(cmdq, thread) < 0);
 		curr_pa = readl(thread->base + CMDQ_THR_CURR_ADDR);
 		end_pa = readl(thread->base + CMDQ_THR_END_ADDR);
-
-		/*
-		 * Atomic execution should remove the following wfe, i.e. only
-		 * wait event at first task, and prevent to pause when running.
-		 */
-		if (thread->atomic_exec) {
-			/* GCE is executing if command is not WFE */
-			if (!cmdq_thread_is_in_wfe(thread)) {
-				cmdq_thread_resume(thread);
-				cmdq_thread_wait_end(thread, end_pa);
-				WARN_ON(cmdq_thread_suspend(cmdq, thread) < 0);
-				/* set to this task directly */
-				writel(task->pa_base,
-				       thread->base + CMDQ_THR_CURR_ADDR);
-			} else {
-				cmdq_task_insert_into_thread(task);
-				cmdq_task_remove_wfe(task);
-				smp_mb(); /* modify jump before enable thread */
-			}
+		/* check boundary */
+		if (curr_pa == end_pa - CMDQ_INST_SIZE ||
+		    curr_pa == end_pa) {
+			/* set to this task directly */
+			writel(task->pa_base,
+			       thread->base + CMDQ_THR_CURR_ADDR);
 		} else {
-			/* check boundary */
-			if (curr_pa == end_pa - CMDQ_INST_SIZE ||
-			    curr_pa == end_pa) {
-				/* set to this task directly */
-				writel(task->pa_base,
-				       thread->base + CMDQ_THR_CURR_ADDR);
-			} else {
-				cmdq_task_insert_into_thread(task);
-				smp_mb(); /* modify jump before enable thread */
-			}
+			cmdq_task_insert_into_thread(task);
+			smp_mb(); /* modify jump before enable thread */
 		}
 		writel(task->pa_base + pkt->cmd_buf_size,
 		       thread->base + CMDQ_THR_END_ADDR);
@@ -501,7 +442,6 @@ static struct mbox_chan *cmdq_xlate(struct mbox_controller *mbox,
 
 	thread = (struct cmdq_thread *)mbox->chans[ind].con_priv;
 	thread->priority = sp->args[1];
-	thread->atomic_exec = (sp->args[2] != 0);
 	thread->chan = &mbox->chans[ind];
 
 	return &mbox->chans[ind];
-- 
2.18.0

[PATCH v1 3/3] mailbox: mediatek: remove implementation related to atomic_exec

[Bibby Hsieh]

After implement flush, client can flush the executing
command buffer or abort the still waiting for event
command buffer, so controller do not need to implement
atomic_exe feature. remove it.

Signed-off-by: Bibby Hsieh <bibby.hsieh@mediatek.com>
Reviewed-by: CK Hu <ck.hu@mediatek.com>
---
 drivers/mailbox/mtk-cmdq-mailbox.c | 76 ++++--------------------------
 1 file changed, 8 insertions(+), 68 deletions(-)

diff --git a/drivers/mailbox/mtk-cmdq-mailbox.c b/drivers/mailbox/mtk-cmdq-mailbox.c
index 0da5e2dc2c0e..b24822ad8409 100644
--- a/drivers/mailbox/mtk-cmdq-mailbox.c
+++ b/drivers/mailbox/mtk-cmdq-mailbox.c
@@ -56,7 +56,6 @@ struct cmdq_thread {
 	void __iomem		*base;
 	struct list_head	task_busy_list;
 	u32			priority;
-	bool			atomic_exec;
 };
 
 struct cmdq_task {
@@ -162,48 +161,11 @@ static void cmdq_task_insert_into_thread(struct cmdq_task *task)
 	cmdq_thread_invalidate_fetched_data(thread);
 }
 
-static bool cmdq_command_is_wfe(u64 cmd)
-{
-	u64 wfe_option = CMDQ_WFE_UPDATE | CMDQ_WFE_WAIT | CMDQ_WFE_WAIT_VALUE;
-	u64 wfe_op = (u64)(CMDQ_CODE_WFE << CMDQ_OP_CODE_SHIFT) << 32;
-	u64 wfe_mask = (u64)CMDQ_OP_CODE_MASK << 32 | 0xffffffff;
-
-	return ((cmd & wfe_mask) == (wfe_op | wfe_option));
-}
-
-/* we assume tasks in the same display GCE thread are waiting the same event. */
-static void cmdq_task_remove_wfe(struct cmdq_task *task)
-{
-	struct device *dev = task->cmdq->mbox.dev;
-	u64 *base = task->pkt->va_base;
-	int i;
-
-	dma_sync_single_for_cpu(dev, task->pa_base, task->pkt->cmd_buf_size,
-				DMA_TO_DEVICE);
-	for (i = 0; i < CMDQ_NUM_CMD(task->pkt); i++)
-		if (cmdq_command_is_wfe(base[i]))
-			base[i] = (u64)CMDQ_JUMP_BY_OFFSET << 32 |
-				  CMDQ_JUMP_PASS;
-	dma_sync_single_for_device(dev, task->pa_base, task->pkt->cmd_buf_size,
-				   DMA_TO_DEVICE);
-}
-
 static bool cmdq_thread_is_in_wfe(struct cmdq_thread *thread)
 {
 	return readl(thread->base + CMDQ_THR_WAIT_TOKEN) & CMDQ_THR_IS_WAITING;
 }
 
-static void cmdq_thread_wait_end(struct cmdq_thread *thread,
-				 unsigned long end_pa)
-{
-	struct device *dev = thread->chan->mbox->dev;
-	unsigned long curr_pa;
-
-	if (readl_poll_timeout_atomic(thread->base + CMDQ_THR_CURR_ADDR,
-			curr_pa, curr_pa == end_pa, 1, 20))
-		dev_err(dev, "GCE thread cannot run to end.\n");
-}
-
 static void cmdq_task_exec_done(struct cmdq_task *task, enum cmdq_cb_status sta)
 {
 	struct cmdq_task_cb *cb = &task->pkt->async_cb;
@@ -383,36 +345,15 @@ static int cmdq_mbox_send_data(struct mbox_chan *chan, void *data)
 		WARN_ON(cmdq_thread_suspend(cmdq, thread) < 0);
 		curr_pa = readl(thread->base + CMDQ_THR_CURR_ADDR);
 		end_pa = readl(thread->base + CMDQ_THR_END_ADDR);
-
-		/*
-		 * Atomic execution should remove the following wfe, i.e. only
-		 * wait event at first task, and prevent to pause when running.
-		 */
-		if (thread->atomic_exec) {
-			/* GCE is executing if command is not WFE */
-			if (!cmdq_thread_is_in_wfe(thread)) {
-				cmdq_thread_resume(thread);
-				cmdq_thread_wait_end(thread, end_pa);
-				WARN_ON(cmdq_thread_suspend(cmdq, thread) < 0);
-				/* set to this task directly */
-				writel(task->pa_base,
-				       thread->base + CMDQ_THR_CURR_ADDR);
-			} else {
-				cmdq_task_insert_into_thread(task);
-				cmdq_task_remove_wfe(task);
-				smp_mb(); /* modify jump before enable thread */
-			}
+		/* check boundary */
+		if (curr_pa == end_pa - CMDQ_INST_SIZE ||
+		    curr_pa == end_pa) {
+			/* set to this task directly */
+			writel(task->pa_base,
+			       thread->base + CMDQ_THR_CURR_ADDR);
 		} else {
-			/* check boundary */
-			if (curr_pa == end_pa - CMDQ_INST_SIZE ||
-			    curr_pa == end_pa) {
-				/* set to this task directly */
-				writel(task->pa_base,
-				       thread->base + CMDQ_THR_CURR_ADDR);
-			} else {
-				cmdq_task_insert_into_thread(task);
-				smp_mb(); /* modify jump before enable thread */
-			}
+			cmdq_task_insert_into_thread(task);
+			smp_mb(); /* modify jump before enable thread */
 		}
 		writel(task->pa_base + pkt->cmd_buf_size,
 		       thread->base + CMDQ_THR_END_ADDR);
@@ -501,7 +442,6 @@ static struct mbox_chan *cmdq_xlate(struct mbox_controller *mbox,
 
 	thread = (struct cmdq_thread *)mbox->chans[ind].con_priv;
 	thread->priority = sp->args[1];
-	thread->atomic_exec = (sp->args[2] != 0);
 	thread->chan = &mbox->chans[ind];
 
 	return &mbox->chans[ind];
-- 
2.18.0
_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

[PATCH v1 3/3] mailbox: mediatek: remove implementation related to atomic_exec

[Bibby Hsieh]

After implement flush, client can flush the executing
command buffer or abort the still waiting for event
command buffer, so controller do not need to implement
atomic_exe feature. remove it.

Signed-off-by: Bibby Hsieh <bibby.hsieh@mediatek.com>
Reviewed-by: CK Hu <ck.hu@mediatek.com>
---
 drivers/mailbox/mtk-cmdq-mailbox.c | 76 ++++--------------------------
 1 file changed, 8 insertions(+), 68 deletions(-)

diff --git a/drivers/mailbox/mtk-cmdq-mailbox.c b/drivers/mailbox/mtk-cmdq-mailbox.c
index 0da5e2dc2c0e..b24822ad8409 100644
--- a/drivers/mailbox/mtk-cmdq-mailbox.c
+++ b/drivers/mailbox/mtk-cmdq-mailbox.c
@@ -56,7 +56,6 @@ struct cmdq_thread {
 	void __iomem		*base;
 	struct list_head	task_busy_list;
 	u32			priority;
-	bool			atomic_exec;
 };
 
 struct cmdq_task {
@@ -162,48 +161,11 @@ static void cmdq_task_insert_into_thread(struct cmdq_task *task)
 	cmdq_thread_invalidate_fetched_data(thread);
 }
 
-static bool cmdq_command_is_wfe(u64 cmd)
-{
-	u64 wfe_option = CMDQ_WFE_UPDATE | CMDQ_WFE_WAIT | CMDQ_WFE_WAIT_VALUE;
-	u64 wfe_op = (u64)(CMDQ_CODE_WFE << CMDQ_OP_CODE_SHIFT) << 32;
-	u64 wfe_mask = (u64)CMDQ_OP_CODE_MASK << 32 | 0xffffffff;
-
-	return ((cmd & wfe_mask) == (wfe_op | wfe_option));
-}
-
-/* we assume tasks in the same display GCE thread are waiting the same event. */
-static void cmdq_task_remove_wfe(struct cmdq_task *task)
-{
-	struct device *dev = task->cmdq->mbox.dev;
-	u64 *base = task->pkt->va_base;
-	int i;
-
-	dma_sync_single_for_cpu(dev, task->pa_base, task->pkt->cmd_buf_size,
-				DMA_TO_DEVICE);
-	for (i = 0; i < CMDQ_NUM_CMD(task->pkt); i++)
-		if (cmdq_command_is_wfe(base[i]))
-			base[i] = (u64)CMDQ_JUMP_BY_OFFSET << 32 |
-				  CMDQ_JUMP_PASS;
-	dma_sync_single_for_device(dev, task->pa_base, task->pkt->cmd_buf_size,
-				   DMA_TO_DEVICE);
-}
-
 static bool cmdq_thread_is_in_wfe(struct cmdq_thread *thread)
 {
 	return readl(thread->base + CMDQ_THR_WAIT_TOKEN) & CMDQ_THR_IS_WAITING;
 }
 
-static void cmdq_thread_wait_end(struct cmdq_thread *thread,
-				 unsigned long end_pa)
-{
-	struct device *dev = thread->chan->mbox->dev;
-	unsigned long curr_pa;
-
-	if (readl_poll_timeout_atomic(thread->base + CMDQ_THR_CURR_ADDR,
-			curr_pa, curr_pa == end_pa, 1, 20))
-		dev_err(dev, "GCE thread cannot run to end.\n");
-}
-
 static void cmdq_task_exec_done(struct cmdq_task *task, enum cmdq_cb_status sta)
 {
 	struct cmdq_task_cb *cb = &task->pkt->async_cb;
@@ -383,36 +345,15 @@ static int cmdq_mbox_send_data(struct mbox_chan *chan, void *data)
 		WARN_ON(cmdq_thread_suspend(cmdq, thread) < 0);
 		curr_pa = readl(thread->base + CMDQ_THR_CURR_ADDR);
 		end_pa = readl(thread->base + CMDQ_THR_END_ADDR);
-
-		/*
-		 * Atomic execution should remove the following wfe, i.e. only
-		 * wait event at first task, and prevent to pause when running.
-		 */
-		if (thread->atomic_exec) {
-			/* GCE is executing if command is not WFE */
-			if (!cmdq_thread_is_in_wfe(thread)) {
-				cmdq_thread_resume(thread);
-				cmdq_thread_wait_end(thread, end_pa);
-				WARN_ON(cmdq_thread_suspend(cmdq, thread) < 0);
-				/* set to this task directly */
-				writel(task->pa_base,
-				       thread->base + CMDQ_THR_CURR_ADDR);
-			} else {
-				cmdq_task_insert_into_thread(task);
-				cmdq_task_remove_wfe(task);
-				smp_mb(); /* modify jump before enable thread */
-			}
+		/* check boundary */
+		if (curr_pa == end_pa - CMDQ_INST_SIZE ||
+		    curr_pa == end_pa) {
+			/* set to this task directly */
+			writel(task->pa_base,
+			       thread->base + CMDQ_THR_CURR_ADDR);
 		} else {
-			/* check boundary */
-			if (curr_pa == end_pa - CMDQ_INST_SIZE ||
-			    curr_pa == end_pa) {
-				/* set to this task directly */
-				writel(task->pa_base,
-				       thread->base + CMDQ_THR_CURR_ADDR);
-			} else {
-				cmdq_task_insert_into_thread(task);
-				smp_mb(); /* modify jump before enable thread */
-			}
+			cmdq_task_insert_into_thread(task);
+			smp_mb(); /* modify jump before enable thread */
 		}
 		writel(task->pa_base + pkt->cmd_buf_size,
 		       thread->base + CMDQ_THR_END_ADDR);
@@ -501,7 +442,6 @@ static struct mbox_chan *cmdq_xlate(struct mbox_controller *mbox,
 
 	thread = (struct cmdq_thread *)mbox->chans[ind].con_priv;
 	thread->priority = sp->args[1];
-	thread->atomic_exec = (sp->args[2] != 0);
 	thread->chan = &mbox->chans[ind];
 
 	return &mbox->chans[ind];
-- 
2.18.0
_______________________________________________
Linux-mediatek mailing list
Linux-mediatek@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-mediatek