mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: Benoit Parrot <bparrot@ti.com>
To: <linux-media@vger.kernel.org>, Hans Verkuil <hverkuil@xs4all.nl>
Cc: <linux-kernel@vger.kernel.org>,
	Tomi Valkeinen <tomi.valkeinen@ti.com>,
	Jyri Sarha <jsarha@ti.com>,
	Peter Ujfalusi <peter.ujfalusi@ti.com>,
	Benoit Parrot <bparrot@ti.com>
Subject: [Patch v2 12/35] media: ti-vpe: vpdma: Add abort channel desc and cleanup APIs
Date: Fri, 18 Nov 2016 17:20:22 -0600	[thread overview]
Message-ID: <20161118232045.24665-13-bparrot@ti.com> (raw)
In-Reply-To: <20161118232045.24665-1-bparrot@ti.com>

From: Nikhil Devshatwar <nikhil.nd@ti.com>

Whenever VPDMA processes a data descriptor of a list, it processes it
and sets up the channel for the DMA transaction. List manager holds the
descriptor in the list until the DMA is complete. If sync_on_channel
descriptor, or another descriptor for the same channel is present in
the FIFO, list manager keeps them until the current channel is free.

When the capture stream is closed suddenly while there are pending
descriptors in the FIFO (streamON failed, application killed), it would
keep the VPDMA in a busy state. Any further list post would fail with
EBUSY.

To avoid this, drivers need to stop the current processing list and
cleanup all the resources VPDMA has taken and also clear the internal FSM
of list manager. The state machine is cleared by issuing channel specific
abort descriptor.

Therefore, the vpdma_list_cleanup accepts an array of channels for which
abort_channel descriptors should be posted. It is driver's responsibility
to post for all the channels or the channels which were used in the last
context.

Signed-off-by: Nikhil Devshatwar <nikhil.nd@ti.com>
Signed-off-by: Benoit Parrot <bparrot@ti.com>
Acked-by: Hans Verkuil <hans.verkuil@cisco.com>
---
 drivers/media/platform/ti-vpe/vpdma.c | 75 +++++++++++++++++++++++++++++++++++
 drivers/media/platform/ti-vpe/vpdma.h |  6 +++
 2 files changed, 81 insertions(+)

diff --git a/drivers/media/platform/ti-vpe/vpdma.c b/drivers/media/platform/ti-vpe/vpdma.c
index 1a0152842a17..7808c9c1828b 100644
--- a/drivers/media/platform/ti-vpe/vpdma.c
+++ b/drivers/media/platform/ti-vpe/vpdma.c
@@ -385,6 +385,56 @@ void vpdma_unmap_desc_buf(struct vpdma_data *vpdma, struct vpdma_buf *buf)
 EXPORT_SYMBOL(vpdma_unmap_desc_buf);
 
 /*
+ * Cleanup all pending descriptors of a list
+ * First, stop the current list being processed.
+ * If the VPDMA was busy, this step makes vpdma to accept post lists.
+ * To cleanup the internal FSM, post abort list descriptor for all the
+ * channels from @channels array of size @size.
+ */
+int vpdma_list_cleanup(struct vpdma_data *vpdma, int list_num,
+		int *channels, int size)
+{
+	struct vpdma_desc_list abort_list;
+	int i, ret, timeout = 500;
+
+	write_reg(vpdma, VPDMA_LIST_ATTR,
+			(list_num << VPDMA_LIST_NUM_SHFT) |
+			(1 << VPDMA_LIST_STOP_SHFT));
+
+	if (size <= 0 || !channels)
+		return 0;
+
+	ret = vpdma_create_desc_list(&abort_list,
+		size * sizeof(struct vpdma_dtd), VPDMA_LIST_TYPE_NORMAL);
+	if (ret)
+		return ret;
+
+	for (i = 0; i < size; i++)
+		vpdma_add_abort_channel_ctd(&abort_list, channels[i]);
+
+	ret = vpdma_map_desc_buf(vpdma, &abort_list.buf);
+	if (ret)
+		return ret;
+	ret = vpdma_submit_descs(vpdma, &abort_list, list_num);
+	if (ret)
+		return ret;
+
+	while (vpdma_list_busy(vpdma, list_num) && timeout--)
+		;
+
+	if (timeout == 0) {
+		dev_err(&vpdma->pdev->dev, "Timed out cleaning up VPDMA list\n");
+		return -EBUSY;
+	}
+
+	vpdma_unmap_desc_buf(vpdma, &abort_list.buf);
+	vpdma_free_desc_buf(&abort_list.buf);
+
+	return 0;
+}
+EXPORT_SYMBOL(vpdma_list_cleanup);
+
+/*
  * create a descriptor list, the user of this list will append configuration,
  * control and data descriptors to this list, this list will be submitted to
  * VPDMA. VPDMA's list parser will go through each descriptor and perform the
@@ -629,6 +679,31 @@ void vpdma_add_sync_on_channel_ctd(struct vpdma_desc_list *list,
 }
 EXPORT_SYMBOL(vpdma_add_sync_on_channel_ctd);
 
+/*
+ * append an 'abort_channel' type control descriptor to the given descriptor
+ * list, this descriptor aborts any DMA transaction happening using the
+ * specified channel
+ */
+void vpdma_add_abort_channel_ctd(struct vpdma_desc_list *list,
+		int chan_num)
+{
+	struct vpdma_ctd *ctd;
+
+	ctd = list->next;
+	WARN_ON((void *)(ctd + 1) > (list->buf.addr + list->buf.size));
+
+	ctd->w0 = 0;
+	ctd->w1 = 0;
+	ctd->w2 = 0;
+	ctd->type_source_ctl = ctd_type_source_ctl(chan_num,
+				CTD_TYPE_ABORT_CHANNEL);
+
+	list->next = ctd + 1;
+
+	dump_ctd(ctd);
+}
+EXPORT_SYMBOL(vpdma_add_abort_channel_ctd);
+
 static void dump_dtd(struct vpdma_dtd *dtd)
 {
 	int dir, chan;
diff --git a/drivers/media/platform/ti-vpe/vpdma.h b/drivers/media/platform/ti-vpe/vpdma.h
index 32b9ed5191c5..4dafc1bcf116 100644
--- a/drivers/media/platform/ti-vpe/vpdma.h
+++ b/drivers/media/platform/ti-vpe/vpdma.h
@@ -163,6 +163,8 @@ enum vpdma_channel {
 #define VIP_CHAN_YUV_PORTB_OFFSET	2
 #define VIP_CHAN_RGB_PORTB_OFFSET	1
 
+#define VPDMA_MAX_CHANNELS		256
+
 /* flags for VPDMA data descriptors */
 #define VPDMA_DATA_ODD_LINE_SKIP	(1 << 0)
 #define VPDMA_DATA_EVEN_LINE_SKIP	(1 << 1)
@@ -219,6 +221,8 @@ void vpdma_add_cfd_adb(struct vpdma_desc_list *list, int client,
 		struct vpdma_buf *adb);
 void vpdma_add_sync_on_channel_ctd(struct vpdma_desc_list *list,
 		enum vpdma_channel chan);
+void vpdma_add_abort_channel_ctd(struct vpdma_desc_list *list,
+		int chan_num);
 void vpdma_add_out_dtd(struct vpdma_desc_list *list, int width,
 		const struct v4l2_rect *c_rect,
 		const struct vpdma_data_format *fmt, dma_addr_t dma_addr,
@@ -233,6 +237,8 @@ void vpdma_add_in_dtd(struct vpdma_desc_list *list, int width,
 		const struct vpdma_data_format *fmt, dma_addr_t dma_addr,
 		enum vpdma_channel chan, int field, u32 flags, int frame_width,
 		int frame_height, int start_h, int start_v);
+int vpdma_list_cleanup(struct vpdma_data *vpdma, int list_num,
+		int *channels, int size);
 
 /* vpdma list interrupt management */
 void vpdma_enable_list_complete_irq(struct vpdma_data *vpdma, int irq_num,
-- 
2.9.0

  parent reply	other threads:[~2016-11-18 23:21 UTC|newest]

Thread overview: 36+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2016-11-18 23:20 [Patch v2 00/35] media: ti-vpe: fixes and enhancements Benoit Parrot
2016-11-18 23:20 ` [Patch v2 01/35] media: ti-vpe: vpdma: Make vpdma library into its own module Benoit Parrot
2016-11-18 23:20 ` [Patch v2 02/35] media: ti-vpe: vpdma: Add multi-instance and multi-client support Benoit Parrot
2016-11-18 23:20 ` [Patch v2 03/35] media: ti-vpe: vpdma: Add helper to set a background color Benoit Parrot
2016-11-18 23:20 ` [Patch v2 04/35] media: ti-vpe: vpdma: Fix bus error when vpdma is writing a descriptor Benoit Parrot
2016-11-18 23:20 ` [Patch v2 05/35] media: ti-vpe: Use line average de-interlacing for first 2 frames Benoit Parrot
2016-11-18 23:20 ` [Patch v2 06/35] media: ti-vpe: vpe: Do not perform job transaction atomically Benoit Parrot
2016-11-18 23:20 ` [Patch v2 07/35] media: ti-vpe: Add support for SEQ_TB buffers Benoit Parrot
2016-11-18 23:20 ` [Patch v2 08/35] media: ti-vpe: Increasing max buffer height and width Benoit Parrot
2016-11-18 23:20 ` [Patch v2 09/35] media: ti-vpe: vpe: Return NULL for invalid buffer type Benoit Parrot
2016-11-18 23:20 ` [Patch v2 10/35] media: ti-vpe: Free vpdma buffers in vpe_release Benoit Parrot
2016-11-18 23:20 ` [Patch v2 11/35] media: ti-vpe: vpdma: Add support for setting max width height Benoit Parrot
2016-11-18 23:20 ` Benoit Parrot [this message]
2016-11-18 23:20 ` [Patch v2 13/35] media: ti-vpe: vpdma: Make list post atomic operation Benoit Parrot
2016-11-18 23:20 ` [Patch v2 14/35] media: ti-vpe: vpdma: Clear IRQs for individual lists Benoit Parrot
2016-11-18 23:20 ` [Patch v2 15/35] media: ti-vpe: vpe: configure line mode separately Benoit Parrot
2016-11-18 23:20 ` [Patch v2 16/35] media: ti-vpe: vpe: Setup srcdst parameters in start_streaming Benoit Parrot
2016-11-18 23:20 ` [Patch v2 17/35] media: ti-vpe: vpe: Post next descriptor only for list complete IRQ Benoit Parrot
2016-11-18 23:20 ` [Patch v2 18/35] media: ti-vpe: vpe: Add RGB565 and RGB5551 support Benoit Parrot
2016-11-18 23:20 ` [Patch v2 19/35] media: ti-vpe: vpdma: allocate and maintain hwlist Benoit Parrot
2016-11-18 23:20 ` [Patch v2 20/35] media: ti-vpe: vpe: Added MODULE_DEVICE_TABLE hint Benoit Parrot
2016-11-18 23:20 ` [Patch v2 21/35] media: ti-vpe: vpdma: Corrected YUV422 data type label Benoit Parrot
2016-11-18 23:20 ` [Patch v2 22/35] media: ti-vpe: vpdma: RGB data type yield inverted data Benoit Parrot
2016-11-18 23:20 ` [Patch v2 23/35] media: ti-vpe: sc: Fix incorrect optimization Benoit Parrot
2016-11-18 23:20 ` [Patch v2 24/35] media: ti-vpe: vpe: Fix vb2 buffer cleanup Benoit Parrot
2016-11-18 23:20 ` [Patch v2 25/35] media: ti-vpe: vpdma: Fix race condition for firmware loading Benoit Parrot
2016-11-18 23:20 ` [Patch v2 26/35] media: ti-vpe: vpdma: Use bidirectional cached buffers Benoit Parrot
2016-11-18 23:20 ` [Patch v2 27/35] media: ti-vpe: vpe: Fix line stride for output motion vector Benoit Parrot
2016-11-18 23:20 ` [Patch v2 28/35] media: ti-vpe: vpe: Enable DMABUF export Benoit Parrot
2016-11-18 23:20 ` [Patch v2 29/35] media: ti-vpe: Make scaler library into its own module Benoit Parrot
2016-11-18 23:20 ` [Patch v2 30/35] media: ti-vpe: scaler: Add debug support for multi-instance Benoit Parrot
2016-11-18 23:20 ` [Patch v2 31/35] media: ti-vpe: vpe: Make sure frame size dont exceed scaler capacity Benoit Parrot
2016-11-18 23:20 ` [Patch v2 32/35] media: ti-vpe: vpdma: Add RAW8 and RAW16 data types Benoit Parrot
2016-11-18 23:20 ` [Patch v2 33/35] media: ti-vpe: Make colorspace converter library into its own module Benoit Parrot
2016-11-18 23:20 ` [Patch v2 34/35] media: ti-vpe: csc: Add debug support for multi-instance Benoit Parrot
2016-11-18 23:20 ` [Patch v2 35/35] media: ti-vpe: vpe: Add proper support single and multi-plane buffer Benoit Parrot

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20161118232045.24665-13-bparrot@ti.com \
    --to=bparrot@ti.com \
    --cc=hverkuil@xs4all.nl \
    --cc=jsarha@ti.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-media@vger.kernel.org \
    --cc=peter.ujfalusi@ti.com \
    --cc=tomi.valkeinen@ti.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox

all inboxes | Powered by JetHome®