Hi Robert,
On Sun, 22 Mar 2015, Robert Jarzmik wrote:
> From: Robert Jarzmik <[email protected]>
>
> Convert pxa_camera to dmaengine. This removes all DMA registers
> manipulation in favor of the more generic dmaengine API.
>
> The functional level should be the same as before. The biggest change is
> in the videobuf_sg_splice() function, which splits a videobuf-dma into
As also commented below, I'm not sure "splice" is a good word for
splitting.
> several scatterlists for 3 planes captures (Y, U, V).
"Several" is actually exactly 3, isn't it?
>
> Signed-off-by: Robert Jarzmik <[email protected]>
> ---
> drivers/media/platform/soc_camera/pxa_camera.c | 428 ++++++++++++-------------
> 1 file changed, 211 insertions(+), 217 deletions(-)
>
> diff --git a/drivers/media/platform/soc_camera/pxa_camera.c b/drivers/media/platform/soc_camera/pxa_camera.c
> index 8b39f44..8644022 100644
> --- a/drivers/media/platform/soc_camera/pxa_camera.c
> +++ b/drivers/media/platform/soc_camera/pxa_camera.c
> @@ -28,6 +28,9 @@
> #include <linux/clk.h>
> #include <linux/sched.h>
> #include <linux/slab.h>
> +#include <linux/dmaengine.h>
> +#include <linux/dma-mapping.h>
> +#include <linux/dma/pxa-dma.h>
>
> #include <media/v4l2-common.h>
> #include <media/v4l2-dev.h>
> @@ -38,7 +41,6 @@
>
> #include <linux/videodev2.h>
>
> -#include <mach/dma.h>
> #include <linux/platform_data/camera-pxa.h>
>
> #define PXA_CAM_VERSION "0.0.6"
> @@ -175,21 +177,16 @@ enum pxa_camera_active_dma {
> DMA_V = 0x4,
> };
>
> -/* descriptor needed for the PXA DMA engine */
> -struct pxa_cam_dma {
> - dma_addr_t sg_dma;
> - struct pxa_dma_desc *sg_cpu;
> - size_t sg_size;
> - int sglen;
> -};
> -
> /* buffer for one video frame */
> struct pxa_buffer {
> /* common v4l buffer stuff -- must be first */
> struct videobuf_buffer vb;
> u32 code;
> /* our descriptor lists for Y, U and V channels */
> - struct pxa_cam_dma dmas[3];
> + struct dma_async_tx_descriptor *descs[3];
> + dma_cookie_t cookie[3];
> + struct scatterlist *sg[3];
> + int sg_len[3];
> int inwork;
> enum pxa_camera_active_dma active_dma;
> };
> @@ -207,7 +204,7 @@ struct pxa_camera_dev {
> void __iomem *base;
>
> int channels;
> - unsigned int dma_chans[3];
> + struct dma_chan *dma_chans[3];
>
> struct pxacamera_platform_data *pdata;
> struct resource *res;
> @@ -222,7 +219,6 @@ struct pxa_camera_dev {
> spinlock_t lock;
>
> struct pxa_buffer *active;
> - struct pxa_dma_desc *sg_tail[3];
> struct tasklet_struct task_eof;
>
> u32 save_cicr[5];
> @@ -259,7 +255,6 @@ static int pxa_videobuf_setup(struct videobuf_queue *vq, unsigned int *count,
> static void free_buffer(struct videobuf_queue *vq, struct pxa_buffer *buf)
> {
> struct soc_camera_device *icd = vq->priv_data;
> - struct soc_camera_host *ici = to_soc_camera_host(icd->parent);
> struct videobuf_dmabuf *dma = videobuf_to_dma(&buf->vb);
> int i;
>
> @@ -276,41 +271,82 @@ static void free_buffer(struct videobuf_queue *vq, struct pxa_buffer *buf)
> if (buf->vb.state == VIDEOBUF_NEEDS_INIT)
> return;
>
> - for (i = 0; i < ARRAY_SIZE(buf->dmas); i++) {
> - if (buf->dmas[i].sg_cpu)
> - dma_free_coherent(ici->v4l2_dev.dev,
> - buf->dmas[i].sg_size,
> - buf->dmas[i].sg_cpu,
> - buf->dmas[i].sg_dma);
> - buf->dmas[i].sg_cpu = NULL;
> + for (i = 0; i < 3 && buf->descs[i]; i++) {
> + async_tx_ack(buf->descs[i]);
> + dmaengine_tx_release(buf->descs[i]);
> + kfree(buf->sg[i]);
> + buf->descs[i] = NULL;
> + buf->sg[i] = NULL;
> + buf->sg_len[i] = 0;
> }
> videobuf_dma_unmap(vq->dev, dma);
> videobuf_dma_free(dma);
>
> buf->vb.state = VIDEOBUF_NEEDS_INIT;
> +
> + dev_dbg(icd->parent, "%s end (vb=0x%p) 0x%08lx %d\n", __func__,
> + &buf->vb, buf->vb.baddr, buf->vb.bsize);
> }
>
> -static int calculate_dma_sglen(struct scatterlist *sglist, int sglen,
> - int sg_first_ofs, int size)
> +static struct scatterlist *videobuf_sg_splice(struct scatterlist *sglist,
> + int sglen, int offset, int size,
> + int *new_sg_len)
> {
> - int i, offset, dma_len, xfer_len;
> - struct scatterlist *sg;
> + struct scatterlist *sg0, *sg, *sg_first = NULL;
> + int i, dma_len, dropped_xfer_len, dropped_remain, remain;
> + int nfirst = -1, nfirst_offset = 0, xfer_len;
>
> - offset = sg_first_ofs;
> + *new_sg_len = 0;
> + dropped_remain = offset;
> + remain = size;
> for_each_sg(sglist, sg, sglen, i) {
Ok, after something-that-felt-like-hours of looking at this code, I think,
I understand now, that first you calculate what sg elements have been used
for offset bytes, which is either 0, or the size of the Y plain, or size
of Y + U plains.
> dma_len = sg_dma_len(sg);
> -
> /* PXA27x Developer's Manual 27.4.4.1: round up to 8 bytes */
> - xfer_len = roundup(min(dma_len - offset, size), 8);
> + dropped_xfer_len = roundup(min(dma_len, dropped_remain), 8);
> + if (dropped_remain)
> + dropped_remain -= dropped_xfer_len;
> + xfer_len = dma_len - dropped_xfer_len;
> +
> + if ((nfirst < 0) && (xfer_len > 0)) {
Superfluous parentheses
> + sg_first = sg;
> + nfirst = i;
> + nfirst_offset = dropped_xfer_len;
> + }
> + if (xfer_len > 0) {
> + *new_sg_len = *new_sg_len + 1;
make it
+ (*new_sg_len)++;
> + remain -= xfer_len;
> + }
> + if (remain <= 0)
> + break;
> + }
> + WARN_ON(nfirst >= sglen);
>
> - size = max(0, size - xfer_len);
> - offset = 0;
> - if (size == 0)
> + sg0 = kmalloc_array(*new_sg_len, sizeof(struct scatterlist),
> + GFP_KERNEL);
> + if (!sg0)
> + return NULL;
> +
> + remain = size;
> + for_each_sg(sg_first, sg, *new_sg_len, i) {
> + dma_len = sg_dma_len(sg);
> + sg0[i] = *sg;
> +
> + sg0[i].offset = nfirst_offset;
> + nfirst_offset = 0;
> +
> + xfer_len = min_t(int, remain, dma_len - sg0[i].offset);
> + xfer_len = roundup(xfer_len, 8);
> + sg_dma_len(&sg0[i]) = xfer_len;
> +
> + remain -= xfer_len;
> + if (remain <= 0) {
> + sg_mark_end(&sg0[i]);
> break;
> + }
> }
>
> - BUG_ON(size != 0);
> - return i + 1;
> + return sg0;
> +}
> static void pxa_camera_dma_irq(struct pxa_camera_dev *pcdev,
> enum pxa_camera_active_dma act_dma);
>
> @@ -343,93 +379,59 @@ static void pxa_camera_dma_irq_v(void *data)
> * @channel: dma channel (0 => 'Y', 1 => 'U', 2 => 'V')
> * @cibr: camera Receive Buffer Register
> * @size: bytes to transfer
> - * @sg_first: first element of sg_list
> - * @sg_first_ofs: offset in first element of sg_list
> + * @offset: offset in videobuffer of the first byte to transfer
> *
> * Prepares the pxa dma descriptors to transfer one camera channel.
> - * Beware sg_first and sg_first_ofs are both input and output parameters.
> *
> - * Returns 0 or -ENOMEM if no coherent memory is available
> + * Returns 0 if success or -ENOMEM if no memory is available
> */
> static int pxa_init_dma_channel(struct pxa_camera_dev *pcdev,
> struct pxa_buffer *buf,
> struct videobuf_dmabuf *dma, int channel,
> - int cibr, int size,
> - struct scatterlist **sg_first, int *sg_first_ofs)
> + int cibr, int size, int offset)
> {
Hmm, ok, can you, please, explain, why you have to change this so much?
IIUC, the functionality, that you're implementing now is rather similar to
the present one - you split the global videobuf SG list into 3 SG lists
for YUV formats. Of course, details are different, you don't use
pxa_dma_desc and all the low-level values directly, you prepare a standard
SG list for your dmaengine driver. But the splitting is the same, isn't
it? And the current one seems rather good to me, because it preserves and
re-uses partially consumed pointers instead of recalculating them every
time, like you do it in your new version. What's the reason for that? Is
the current version buggy or the current approach unsuitable for your new
version?
> - struct pxa_cam_dma *pxa_dma = &buf->dmas[channel];
> - struct device *dev = pcdev->soc_host.v4l2_dev.dev;
> - struct scatterlist *sg;
> - int i, offset, sglen;
> - int dma_len = 0, xfer_len = 0;
> -
> - if (pxa_dma->sg_cpu)
> - dma_free_coherent(dev, pxa_dma->sg_size,
> - pxa_dma->sg_cpu, pxa_dma->sg_dma);
> -
> - sglen = calculate_dma_sglen(*sg_first, dma->sglen,
> - *sg_first_ofs, size);
> -
> - pxa_dma->sg_size = (sglen + 1) * sizeof(struct pxa_dma_desc);
> - pxa_dma->sg_cpu = dma_alloc_coherent(dev, pxa_dma->sg_size,
> - &pxa_dma->sg_dma, GFP_KERNEL);
> - if (!pxa_dma->sg_cpu)
> - return -ENOMEM;
> -
> - pxa_dma->sglen = sglen;
> - offset = *sg_first_ofs;
> -
> - dev_dbg(dev, "DMA: sg_first=%p, sglen=%d, ofs=%d, dma.desc=%x\n",
> - *sg_first, sglen, *sg_first_ofs, pxa_dma->sg_dma);
> -
> -
> - for_each_sg(*sg_first, sg, sglen, i) {
> - dma_len = sg_dma_len(sg);
> -
> - /* PXA27x Developer's Manual 27.4.4.1: round up to 8 bytes */
> - xfer_len = roundup(min(dma_len - offset, size), 8);
> -
> - size = max(0, size - xfer_len);
> -
> - pxa_dma->sg_cpu[i].dsadr = pcdev->res->start + cibr;
> - pxa_dma->sg_cpu[i].dtadr = sg_dma_address(sg) + offset;
> - pxa_dma->sg_cpu[i].dcmd =
> - DCMD_FLOWSRC | DCMD_BURST8 | DCMD_INCTRGADDR | xfer_len;
> -#ifdef DEBUG
> - if (!i)
> - pxa_dma->sg_cpu[i].dcmd |= DCMD_STARTIRQEN;
> -#endif
> - pxa_dma->sg_cpu[i].ddadr =
> - pxa_dma->sg_dma + (i + 1) * sizeof(struct pxa_dma_desc);
> -
> - dev_vdbg(dev, "DMA: desc.%08x->@phys=0x%08x, len=%d\n",
> - pxa_dma->sg_dma + i * sizeof(struct pxa_dma_desc),
> - sg_dma_address(sg) + offset, xfer_len);
> - offset = 0;
> -
> - if (size == 0)
> - break;
> + struct dma_chan *dma_chan = pcdev->dma_chans[channel];
> + struct scatterlist *sg = NULL;
Superfluous initialisation
> + int sglen;
> + struct dma_async_tx_descriptor *tx;
> +
> + sg = videobuf_sg_splice(dma->sglist, dma->sglen, offset, size, &sglen);
Isn't it rather a cut, than a splice function?
> + if (!sg)
> + goto fail;
> +
> + tx = dmaengine_prep_slave_sg(dma_chan, sg, sglen, DMA_DEV_TO_MEM,
> + DMA_PREP_INTERRUPT | DMA_CTRL_ACK);
> + if (!tx) {
> + dev_err(pcdev->soc_host.v4l2_dev.dev,
> + "dmaengine_prep_slave_sg failed\n");
> + goto fail;
> }
>
> - pxa_dma->sg_cpu[sglen].ddadr = DDADR_STOP;
> - pxa_dma->sg_cpu[sglen].dcmd = DCMD_FLOWSRC | DCMD_BURST8 | DCMD_ENDIRQEN;
> -
> - /*
> - * Handle 1 special case :
> - * - in 3 planes (YUV422P format), we might finish with xfer_len equal
> - * to dma_len (end on PAGE boundary). In this case, the sg element
> - * for next plane should be the next after the last used to store the
> - * last scatter gather RAM page
> - */
> - if (xfer_len >= dma_len) {
> - *sg_first_ofs = xfer_len - dma_len;
> - *sg_first = sg_next(sg);
> - } else {
> - *sg_first_ofs = xfer_len;
> - *sg_first = sg;
> + tx->callback_param = pcdev;
> + switch (channel) {
> + case 0:
> + tx->callback = pxa_camera_dma_irq_y;
> + break;
> + case 1:
> + tx->callback = pxa_camera_dma_irq_u;
> + break;
> + case 2:
> + tx->callback = pxa_camera_dma_irq_v;
> + break;
> }
>
> + buf->descs[channel] = tx;
> + buf->sg[channel] = sg;
> + buf->sg_len[channel] = sglen;
> return 0;
> +fail:
> + kfree(sg);
> +
> + dev_dbg(pcdev->soc_host.v4l2_dev.dev,
> + "%s (vb=0x%p) dma_tx=%p\n",
> + __func__, &buf->vb, tx);
> +
> + return -ENOMEM;
> }
>
> static void pxa_videobuf_set_actdma(struct pxa_camera_dev *pcdev,
> @@ -498,9 +500,7 @@ static int pxa_videobuf_prepare(struct videobuf_queue *vq,
>
> if (vb->state == VIDEOBUF_NEEDS_INIT) {
> int size = vb->size;
> - int next_ofs = 0;
> struct videobuf_dmabuf *dma = videobuf_to_dma(vb);
> - struct scatterlist *sg;
>
> ret = videobuf_iolock(vq, vb, NULL);
> if (ret)
> @@ -513,11 +513,9 @@ static int pxa_videobuf_prepare(struct videobuf_queue *vq,
> size_y = size;
> }
>
> - sg = dma->sglist;
> -
> /* init DMA for Y channel */
> - ret = pxa_init_dma_channel(pcdev, buf, dma, 0, CIBR0, size_y,
> - &sg, &next_ofs);
> + ret = pxa_init_dma_channel(pcdev, buf, dma, 0, CIBR0,
> + size_y, 0);
> if (ret) {
> dev_err(dev, "DMA initialization for Y/RGB failed\n");
> goto fail;
> @@ -526,7 +524,7 @@ static int pxa_videobuf_prepare(struct videobuf_queue *vq,
> /* init DMA for U channel */
> if (size_u)
> ret = pxa_init_dma_channel(pcdev, buf, dma, 1, CIBR1,
> - size_u, &sg, &next_ofs);
> + size_u, size_y);
> if (ret) {
> dev_err(dev, "DMA initialization for U failed\n");
> goto fail_u;
> @@ -535,7 +533,7 @@ static int pxa_videobuf_prepare(struct videobuf_queue *vq,
> /* init DMA for V channel */
> if (size_v)
> ret = pxa_init_dma_channel(pcdev, buf, dma, 2, CIBR2,
> - size_v, &sg, &next_ofs);
> + size_v, size_y + size_u);
> if (ret) {
> dev_err(dev, "DMA initialization for V failed\n");
> goto fail_v;
> @@ -550,11 +548,7 @@ static int pxa_videobuf_prepare(struct videobuf_queue *vq,
> return 0;
>
> fail_v:
> - dma_free_coherent(dev, buf->dmas[1].sg_size,
> - buf->dmas[1].sg_cpu, buf->dmas[1].sg_dma);
> fail_u:
> - dma_free_coherent(dev, buf->dmas[0].sg_size,
> - buf->dmas[0].sg_cpu, buf->dmas[0].sg_dma);
> fail:
You don't need 3 exit labels any more.
> free_buffer(vq, buf);
> out:
> @@ -578,10 +572,8 @@ static void pxa_dma_start_channels(struct pxa_camera_dev *pcdev)
>
> for (i = 0; i < pcdev->channels; i++) {
> dev_dbg(pcdev->soc_host.v4l2_dev.dev,
> - "%s (channel=%d) ddadr=%08x\n", __func__,
> - i, active->dmas[i].sg_dma);
> - DDADR(pcdev->dma_chans[i]) = active->dmas[i].sg_dma;
> - DCSR(pcdev->dma_chans[i]) = DCSR_RUN;
> + "%s (channel=%d)\n", __func__, i);
> + dma_async_issue_pending(pcdev->dma_chans[i]);
> }
> }
>
> @@ -592,7 +584,7 @@ static void pxa_dma_stop_channels(struct pxa_camera_dev *pcdev)
> for (i = 0; i < pcdev->channels; i++) {
> dev_dbg(pcdev->soc_host.v4l2_dev.dev,
> "%s (channel=%d)\n", __func__, i);
> - DCSR(pcdev->dma_chans[i]) = 0;
> + dmaengine_terminate_all(pcdev->dma_chans[i]);
> }
> }
>
> @@ -600,18 +592,12 @@ static void pxa_dma_add_tail_buf(struct pxa_camera_dev *pcdev,
> struct pxa_buffer *buf)
> {
> int i;
> - struct pxa_dma_desc *buf_last_desc;
>
> for (i = 0; i < pcdev->channels; i++) {
> - buf_last_desc = buf->dmas[i].sg_cpu + buf->dmas[i].sglen;
> - buf_last_desc->ddadr = DDADR_STOP;
> -
> - if (pcdev->sg_tail[i])
> - /* Link the new buffer to the old tail */
> - pcdev->sg_tail[i]->ddadr = buf->dmas[i].sg_dma;
> -
> - /* Update the channel tail */
> - pcdev->sg_tail[i] = buf_last_desc;
> + buf->cookie[i] = dmaengine_submit(buf->descs[i]);
> + dev_dbg(pcdev->soc_host.v4l2_dev.dev,
> + "%s (channel=%d) : submit vb=%p cookie=%d\n",
> + __func__, i, buf, buf->descs[i]->cookie);
> }
> }
>
> @@ -703,8 +689,6 @@ static void pxa_camera_wakeup(struct pxa_camera_dev *pcdev,
> struct videobuf_buffer *vb,
> struct pxa_buffer *buf)
> {
> - int i;
> -
> /* _init is used to debug races, see comment in pxa_camera_reqbufs() */
> list_del_init(&vb->queue);
> vb->state = VIDEOBUF_DONE;
> @@ -716,8 +700,6 @@ static void pxa_camera_wakeup(struct pxa_camera_dev *pcdev,
>
> if (list_empty(&pcdev->capture)) {
> pxa_camera_stop_capture(pcdev);
> - for (i = 0; i < pcdev->channels; i++)
> - pcdev->sg_tail[i] = NULL;
> return;
> }
>
> @@ -741,50 +723,41 @@ static void pxa_camera_wakeup(struct pxa_camera_dev *pcdev,
> *
> * Context: should only be called within the dma irq handler
> */
> -static void pxa_camera_check_link_miss(struct pxa_camera_dev *pcdev)
> +static void pxa_camera_check_link_miss(struct pxa_camera_dev *pcdev,
> + dma_cookie_t last_submitted,
> + dma_cookie_t last_issued)
> {
> - int i, is_dma_stopped = 1;
> + int is_dma_stopped;
>
> - for (i = 0; i < pcdev->channels; i++)
> - if (DDADR(pcdev->dma_chans[i]) != DDADR_STOP)
> - is_dma_stopped = 0;
> + is_dma_stopped = last_submitted > last_issued;
IIUC, actually last_issued should be == last_submitted, right? And your
dmaengine driver's .device_tx_status() method returns the last "really"
submitted cookie, so, if that situation occurs, that you submit a transfer
when DMA is off, your last_buf->cookie[chan] will contain it, as returned
by dmaengine_submit(), but .device_tx_status() will not return it? Then,
this comparison "last_submitted > last_issued" isn't a good test, cookies
can overrun. Maybe just check for unequal? Otherwise I think you'd have to
find a way to use dma_async_is_complete().
> dev_dbg(pcdev->soc_host.v4l2_dev.dev,
> - "%s : top queued buffer=%p, dma_stopped=%d\n",
> + "%s : top queued buffer=%p, is_dma_stopped=%d\n",
> __func__, pcdev->active, is_dma_stopped);
> if (pcdev->active && is_dma_stopped)
> pxa_camera_start_capture(pcdev);
> }
>
> -static void pxa_camera_dma_irq(int channel, struct pxa_camera_dev *pcdev,
> +static void pxa_camera_dma_irq(struct pxa_camera_dev *pcdev,
> enum pxa_camera_active_dma act_dma)
> {
> struct device *dev = pcdev->soc_host.v4l2_dev.dev;
> - struct pxa_buffer *buf;
> + struct pxa_buffer *buf, *last_buf;
> unsigned long flags;
> - u32 status, camera_status, overrun;
> + u32 camera_status, overrun;
> + int chan;
> struct videobuf_buffer *vb;
> + enum dma_status last_status;
> + dma_cookie_t last_issued;
>
> spin_lock_irqsave(&pcdev->lock, flags);
>
> - status = DCSR(channel);
> - DCSR(channel) = status;
> -
> camera_status = __raw_readl(pcdev->base + CISR);
> + dev_dbg(dev, "camera dma irq, cisr=0x%x dma=%d\n",
> + camera_status, act_dma);
> overrun = CISR_IFO_0;
> if (pcdev->channels == 3)
> overrun |= CISR_IFO_1 | CISR_IFO_2;
>
> - if (status & DCSR_BUSERR) {
> - dev_err(dev, "DMA Bus Error IRQ!\n");
> - goto out;
> - }
> -
> - if (!(status & (DCSR_ENDINTR | DCSR_STARTINTR))) {
> - dev_err(dev, "Unknown DMA IRQ source, status: 0x%08x\n",
> - status);
> - goto out;
> - }
> -
> /*
> * pcdev->active should not be NULL in DMA irq handler.
> *
> @@ -804,28 +777,39 @@ static void pxa_camera_dma_irq(int channel, struct pxa_camera_dev *pcdev,
> buf = container_of(vb, struct pxa_buffer, vb);
> WARN_ON(buf->inwork || list_empty(&vb->queue));
>
> - dev_dbg(dev, "%s channel=%d %s%s(vb=0x%p) dma.desc=%x\n",
> - __func__, channel, status & DCSR_STARTINTR ? "SOF " : "",
> - status & DCSR_ENDINTR ? "EOF " : "", vb, DDADR(channel));
> -
> - if (status & DCSR_ENDINTR) {
> - /*
> - * It's normal if the last frame creates an overrun, as there
> - * are no more DMA descriptors to fetch from QCI fifos
> - */
> - if (camera_status & overrun &&
> - !list_is_last(pcdev->capture.next, &pcdev->capture)) {
> - dev_dbg(dev, "FIFO overrun! CISR: %x\n",
> - camera_status);
> - pxa_camera_stop_capture(pcdev);
> - pxa_camera_start_capture(pcdev);
> - goto out;
> - }
> - buf->active_dma &= ~act_dma;
> - if (!buf->active_dma) {
> - pxa_camera_wakeup(pcdev, vb, buf);
> - pxa_camera_check_link_miss(pcdev);
> - }
> + /*
> + * It's normal if the last frame creates an overrun, as there
> + * are no more DMA descriptors to fetch from QCI fifos
> + */
> + switch (act_dma) {
> + case DMA_U:
> + chan = 1;
> + break;
> + case DMA_V:
> + chan = 2;
> + break;
> + default:
> + chan = 0;
> + break;
> + }
> + last_buf = list_entry(pcdev->capture.prev,
> + struct pxa_buffer, vb.queue);
> + last_status = dma_async_is_tx_complete(pcdev->dma_chans[chan],
> + last_buf->cookie[chan],
> + NULL, &last_issued);
> + if (camera_status & overrun &&
> + last_status != DMA_COMPLETE) {
Isn't the compiler suggesting parentheses here?
> + dev_dbg(dev, "FIFO overrun! CISR: %x\n",
> + camera_status);
> + pxa_camera_stop_capture(pcdev);
> + pxa_camera_start_capture(pcdev);
> + goto out;
> + }
> + buf->active_dma &= ~act_dma;
> + if (!buf->active_dma) {
> + pxa_camera_wakeup(pcdev, vb, buf);
> + pxa_camera_check_link_miss(pcdev, last_buf->cookie[chan],
> + last_issued);
> }
>
> out:
> @@ -1012,10 +996,7 @@ static void pxa_camera_clock_stop(struct soc_camera_host *ici)
> __raw_writel(0x3ff, pcdev->base + CICR0);
>
> /* Stop DMA engine */
> - DCSR(pcdev->dma_chans[0]) = 0;
> - DCSR(pcdev->dma_chans[1]) = 0;
> - DCSR(pcdev->dma_chans[2]) = 0;
> -
> + pxa_dma_stop_channels(pcdev);
Isn't calling pxa_dma_stop_channels() in pxa_camera_stop_capture() enough?
.clock_stop() should only be called after streaming has been stopped. But
it has been here since forever, so...
> pxa_camera_deactivate(pcdev);
> }
>
> @@ -1629,10 +1610,6 @@ static int pxa_camera_resume(struct device *dev)
> struct pxa_camera_dev *pcdev = ici->priv;
> int i = 0, ret = 0;
>
> - DRCMR(68) = pcdev->dma_chans[0] | DRCMR_MAPVLD;
> - DRCMR(69) = pcdev->dma_chans[1] | DRCMR_MAPVLD;
> - DRCMR(70) = pcdev->dma_chans[2] | DRCMR_MAPVLD;
> -
Will resume still work after this? Because of dmaengine resume?
> __raw_writel(pcdev->save_cicr[i++] & ~CICR0_ENB, pcdev->base + CICR0);
> __raw_writel(pcdev->save_cicr[i++], pcdev->base + CICR1);
> __raw_writel(pcdev->save_cicr[i++], pcdev->base + CICR2);
> @@ -1738,8 +1715,11 @@ static int pxa_camera_probe(struct platform_device *pdev)
> struct pxa_camera_dev *pcdev;
> struct resource *res;
> void __iomem *base;
> + struct dma_slave_config config;
I would do
+ struct dma_slave_config config = {
+ .direction = DMA_DEV_TO_MEM,
+ .src_maxburst = 8,
+ };
and have all other fields initialised to 0 automatically.
> + dma_cap_mask_t mask;
> + struct pxad_param params;
> int irq;
> - int err = 0;
> + int err = 0, i;
>
> res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
> irq = platform_get_irq(pdev, 0);
> @@ -1807,36 +1787,50 @@ static int pxa_camera_probe(struct platform_device *pdev)
> pcdev->base = base;
>
> /* request dma */
> - err = pxa_request_dma("CI_Y", DMA_PRIO_HIGH,
> - pxa_camera_dma_irq_y, pcdev);
> - if (err < 0) {
> + dma_cap_zero(mask);
> + dma_cap_set(DMA_SLAVE, mask);
Also DMA_PRIVATE?
> +
> + params.prio = 0;
> + params.drcmr = 68;
> + pcdev->dma_chans[0] =
> + dma_request_slave_channel_compat(mask, pxad_filter_fn,
> + ¶ms, &pdev->dev, "CI_Y");
> + if (!pcdev->dma_chans[0]) {
> dev_err(&pdev->dev, "Can't request DMA for Y\n");
> - return err;
> + return -ENODEV;
> }
> - pcdev->dma_chans[0] = err;
> - dev_dbg(&pdev->dev, "got DMA channel %d\n", pcdev->dma_chans[0]);
>
> - err = pxa_request_dma("CI_U", DMA_PRIO_HIGH,
> - pxa_camera_dma_irq_u, pcdev);
> - if (err < 0) {
> - dev_err(&pdev->dev, "Can't request DMA for U\n");
> + params.drcmr = 69;
> + pcdev->dma_chans[1] =
> + dma_request_slave_channel_compat(mask, pxad_filter_fn,
> + ¶ms, &pdev->dev, "CI_U");
> + if (!pcdev->dma_chans[1]) {
> + dev_err(&pdev->dev, "Can't request DMA for Y\n");
> goto exit_free_dma_y;
> }
> - pcdev->dma_chans[1] = err;
> - dev_dbg(&pdev->dev, "got DMA channel (U) %d\n", pcdev->dma_chans[1]);
>
> - err = pxa_request_dma("CI_V", DMA_PRIO_HIGH,
> - pxa_camera_dma_irq_v, pcdev);
> - if (err < 0) {
> + params.drcmr = 70;
> + pcdev->dma_chans[2] =
> + dma_request_slave_channel_compat(mask, pxad_filter_fn,
> + ¶ms, &pdev->dev, "CI_V");
> + if (!pcdev->dma_chans[2]) {
> dev_err(&pdev->dev, "Can't request DMA for V\n");
> goto exit_free_dma_u;
> }
> - pcdev->dma_chans[2] = err;
> - dev_dbg(&pdev->dev, "got DMA channel (V) %d\n", pcdev->dma_chans[2]);
>
> - DRCMR(68) = pcdev->dma_chans[0] | DRCMR_MAPVLD;
> - DRCMR(69) = pcdev->dma_chans[1] | DRCMR_MAPVLD;
> - DRCMR(70) = pcdev->dma_chans[2] | DRCMR_MAPVLD;
> + memset(&config, 0, sizeof(config));
> + config.src_addr_width = 0;
> + config.src_maxburst = 8;
> + config.direction = DMA_DEV_TO_MEM;
> + for (i = 0; i < 3; i++) {
> + config.src_addr = pcdev->res->start + CIBR0 + i * 8;
> + err = dmaengine_slave_config(pcdev->dma_chans[i], &config);
> + if (err < 0) {
> + dev_err(&pdev->dev, "dma slave config failed: %d\n",
> + err);
> + goto exit_free_dma;
> + }
> + }
>
> /* request irq */
> err = devm_request_irq(&pdev->dev, pcdev->irq, pxa_camera_irq, 0,
> @@ -1860,11 +1854,11 @@ static int pxa_camera_probe(struct platform_device *pdev)
> return 0;
>
> exit_free_dma:
> - pxa_free_dma(pcdev->dma_chans[2]);
> + dma_release_channel(pcdev->dma_chans[2]);
> exit_free_dma_u:
> - pxa_free_dma(pcdev->dma_chans[1]);
> + dma_release_channel(pcdev->dma_chans[1]);
> exit_free_dma_y:
> - pxa_free_dma(pcdev->dma_chans[0]);
> + dma_release_channel(pcdev->dma_chans[0]);
> return err;
> }
>
> @@ -1874,9 +1868,9 @@ static int pxa_camera_remove(struct platform_device *pdev)
> struct pxa_camera_dev *pcdev = container_of(soc_host,
> struct pxa_camera_dev, soc_host);
>
> - pxa_free_dma(pcdev->dma_chans[0]);
> - pxa_free_dma(pcdev->dma_chans[1]);
> - pxa_free_dma(pcdev->dma_chans[2]);
> + dma_release_channel(pcdev->dma_chans[0]);
> + dma_release_channel(pcdev->dma_chans[1]);
> + dma_release_channel(pcdev->dma_chans[2]);
>
> soc_camera_host_unregister(soc_host);
>
> --
> 2.1.4
>
Guennadi Liakhovetski <[email protected]> writes:
> Hi Robert,
>
> On Sun, 22 Mar 2015, Robert Jarzmik wrote:
>
>> From: Robert Jarzmik <[email protected]>
>>
>> Convert pxa_camera to dmaengine. This removes all DMA registers
>> manipulation in favor of the more generic dmaengine API.
>>
>> The functional level should be the same as before. The biggest change is
>> in the videobuf_sg_splice() function, which splits a videobuf-dma into
>
> As also commented below, I'm not sure "splice" is a good word for
> splitting.
Ok. I'm all ears for your best candidate :)
>> several scatterlists for 3 planes captures (Y, U, V).
>
> "Several" is actually exactly 3, isn't it?
Yup, it's 3, definitely. I can amend the commit message accordingly.
>> +static struct scatterlist *videobuf_sg_splice(struct scatterlist *sglist,
>> + int sglen, int offset, int size,
>> + int *new_sg_len)
>> {
>> - int i, offset, dma_len, xfer_len;
>> - struct scatterlist *sg;
>> + struct scatterlist *sg0, *sg, *sg_first = NULL;
>> + int i, dma_len, dropped_xfer_len, dropped_remain, remain;
>> + int nfirst = -1, nfirst_offset = 0, xfer_len;
>>
>> - offset = sg_first_ofs;
>> + *new_sg_len = 0;
>> + dropped_remain = offset;
>> + remain = size;
>> for_each_sg(sglist, sg, sglen, i) {
>
> Ok, after something-that-felt-like-hours of looking at this code, I think,
> I understand now, that first you calculate what sg elements have been used
> for offset bytes, which is either 0, or the size of the Y plain, or size
> of Y + U plains.
You can say it that way. I'd say that I calculate how to "malloc and fill" a new
scatter-gather to represent [offset, offset + size [ interval of the original
sglist.
>
>> dma_len = sg_dma_len(sg);
>> -
>> /* PXA27x Developer's Manual 27.4.4.1: round up to 8 bytes */
>> - xfer_len = roundup(min(dma_len - offset, size), 8);
>> + dropped_xfer_len = roundup(min(dma_len, dropped_remain), 8);
>> + if (dropped_remain)
>> + dropped_remain -= dropped_xfer_len;
>> + xfer_len = dma_len - dropped_xfer_len;
>> +
>> + if ((nfirst < 0) && (xfer_len > 0)) {
>
> Superfluous parentheses
Got it.
>
>> + sg_first = sg;
>> + nfirst = i;
>> + nfirst_offset = dropped_xfer_len;
>> + }
>> + if (xfer_len > 0) {
>> + *new_sg_len = *new_sg_len + 1;
>
> make it
> + (*new_sg_len)++;
Got it.
>> static void pxa_camera_dma_irq(struct pxa_camera_dev *pcdev,
>> enum pxa_camera_active_dma act_dma);
>>
>> @@ -343,93 +379,59 @@ static void pxa_camera_dma_irq_v(void *data)
>> * @channel: dma channel (0 => 'Y', 1 => 'U', 2 => 'V')
>> * @cibr: camera Receive Buffer Register
>> * @size: bytes to transfer
>> - * @sg_first: first element of sg_list
>> - * @sg_first_ofs: offset in first element of sg_list
>> + * @offset: offset in videobuffer of the first byte to transfer
>> *
>> * Prepares the pxa dma descriptors to transfer one camera channel.
>> - * Beware sg_first and sg_first_ofs are both input and output parameters.
>> *
>> - * Returns 0 or -ENOMEM if no coherent memory is available
>> + * Returns 0 if success or -ENOMEM if no memory is available
>> */
>> static int pxa_init_dma_channel(struct pxa_camera_dev *pcdev,
>> struct pxa_buffer *buf,
>> struct videobuf_dmabuf *dma, int channel,
>> - int cibr, int size,
>> - struct scatterlist **sg_first, int *sg_first_ofs)
>> + int cibr, int size, int offset)
>> {
>
> Hmm, ok, can you, please, explain, why you have to change this so much?
> IIUC, the functionality, that you're implementing now is rather similar to
> the present one - you split the global videobuf SG list into 3 SG lists
> for YUV formats. Of course, details are different, you don't use
> pxa_dma_desc and all the low-level values directly, you prepare a standard
> SG list for your dmaengine driver. But the splitting is the same, isn't
> it?
The overall splitting is the same, yes : split one global SG list into 3 SG
lists.
> And the current one seems rather good to me, because it preserves and
> re-uses partially consumed pointers instead of recalculating them every
> time, like you do it in your new version. What's the reason for that? Is
> the current version buggy or the current approach unsuitable for your new
> version?
Let's say "unsuitable", or to be more correct, let's say that I didn't found any
better idea yet. As I find that piece of code a bit complicated too, I'll tell
you what was my need for doing it, and maybe you'll/we'll come up with a better
idea.
The previous version had the good fortune to rely on a _single_ scatter-gather
list. Only the DMA descriptors list was computed to be 3 lists
(dma[channe].sg_cpu[0..n]).
The new version must have 3 separated SG lists, each one fed to a different
dmaengine channel.
So the big goal here is to compute the 3 SGs lists, which was not done
previously. That's the purpose of all that, and I couldn't find any easier
method. Would a "new_sg = sg_extract(sglist, offset, len)" have existed, this code
would have been a breeze ...
>> + struct dma_chan *dma_chan = pcdev->dma_chans[channel];
>> + struct scatterlist *sg = NULL;
>
> Superfluous initialisation
Got it.
>> + int sglen;
>> + struct dma_async_tx_descriptor *tx;
>> +
>> + sg = videobuf_sg_splice(dma->sglist, dma->sglen, offset, size, &sglen);
>
> Isn't it rather a cut, than a splice function?
Yeah, actually I meant "slice", but wrote "splice" ...
>> @@ -550,11 +548,7 @@ static int pxa_videobuf_prepare(struct videobuf_queue *vq,
>> return 0;
>>
>> fail_v:
>> - dma_free_coherent(dev, buf->dmas[1].sg_size,
>> - buf->dmas[1].sg_cpu, buf->dmas[1].sg_dma);
>> fail_u:
>> - dma_free_coherent(dev, buf->dmas[0].sg_size,
>> - buf->dmas[0].sg_cpu, buf->dmas[0].sg_dma);
>> fail:
>
> You don't need 3 exit labels any more.
Yes, I still need "fail:" I think, the other 2 will be done in next iteration.
>> @@ -741,50 +723,41 @@ static void pxa_camera_wakeup(struct pxa_camera_dev *pcdev,
>> *
>> * Context: should only be called within the dma irq handler
>> */
>> -static void pxa_camera_check_link_miss(struct pxa_camera_dev *pcdev)
>> +static void pxa_camera_check_link_miss(struct pxa_camera_dev *pcdev,
>> + dma_cookie_t last_submitted,
>> + dma_cookie_t last_issued)
>> {
>> - int i, is_dma_stopped = 1;
>> + int is_dma_stopped;
>>
>> - for (i = 0; i < pcdev->channels; i++)
>> - if (DDADR(pcdev->dma_chans[i]) != DDADR_STOP)
>> - is_dma_stopped = 0;
>> + is_dma_stopped = last_submitted > last_issued;
>
> IIUC, actually last_issued should be == last_submitted, right?
No.
last_issued == last_submitted is only true if the last descriptor "hot-chaining"
succeeded. If it didn't succeed, and the dma stopped, last_submitted >
last_issued, for as long as dma_async_issue_pending() is not called.
> And your dmaengine driver's .device_tx_status() method returns the last
> "really" submitted cookie,
If you mean the last tx cookie for which tx->submit() was called, then yes.
> so, if that situation occurs, that you submit a transfer when DMA is off, your
> last_buf->cookie[chan] will contain it, as returned by dmaengine_submit(), but
> .device_tx_status() will not return it?
Here I'm a bit lost, but I feel the first hypothesis "last_issued ==
last_submitted" blurs the discussion, so I'll let you first read my comment
about it, and then you'll tell me if you feel there is still a concern.
> Then, this comparison "last_submitted
> > last_issued" isn't a good test, cookies can overrun. Maybe just check for
> unequal? Otherwise I think you'd have to find a way to use
> dma_async_is_complete().
Yes, agreed, I'll try unequal, I just need a couple of days to think about it if
another corner case doesn't arise.
>> + if (camera_status & overrun &&
>> + last_status != DMA_COMPLETE) {
>
> Isn't the compiler suggesting parentheses here?
Not mine at least. As the "&&" has the lowest precedence, my understanding is
that the expression is correct. But I can add parenthesis for maintainability.
>> @@ -1012,10 +996,7 @@ static void pxa_camera_clock_stop(struct soc_camera_host *ici)
>> __raw_writel(0x3ff, pcdev->base + CICR0);
>>
>> /* Stop DMA engine */
>> - DCSR(pcdev->dma_chans[0]) = 0;
>> - DCSR(pcdev->dma_chans[1]) = 0;
>> - DCSR(pcdev->dma_chans[2]) = 0;
>> -
>> + pxa_dma_stop_channels(pcdev);
>
> Isn't calling pxa_dma_stop_channels() in pxa_camera_stop_capture() enough?
> .clock_stop() should only be called after streaming has been stopped. But
> it has been here since forever, so...
I don't think so, because the buffers are not released, ie. acked. The effect of
pxa_dma_stop_channels() is to call dmaengine_terminate_all(). But this last call
is only entitled to free the dma txs if they are acked, which happens when
free_buffer() is called.
So in pxa_camera_stop_capture(), as the buffers were not released, the dmaengine
txs were not released as well. In pxa_camera_clock_stop(), I think they have
been released, so dmaengine_terminate_all() will actually free the resources of
these txs, which it couldn't do sooner.
>> - DRCMR(68) = pcdev->dma_chans[0] | DRCMR_MAPVLD;
>> - DRCMR(69) = pcdev->dma_chans[1] | DRCMR_MAPVLD;
>> - DRCMR(70) = pcdev->dma_chans[2] | DRCMR_MAPVLD;
>> -
>
> Will resume still work after this? Because of dmaengine resume?
I must admit I haven't tried. And I fear I have not implemented suspend/resume
in pxa_dma dmaengine driver. I'd say it's a dmaengine problem now, and that this
chunk is correct. Yet your point is correct, pxa_dma driver has to handle
suspend/resume.
>> __raw_writel(pcdev->save_cicr[i++] & ~CICR0_ENB, pcdev->base + CICR0);
>> __raw_writel(pcdev->save_cicr[i++], pcdev->base + CICR1);
>> __raw_writel(pcdev->save_cicr[i++], pcdev->base + CICR2);
>> @@ -1738,8 +1715,11 @@ static int pxa_camera_probe(struct platform_device *pdev)
>> struct pxa_camera_dev *pcdev;
>> struct resource *res;
>> void __iomem *base;
>> + struct dma_slave_config config;
>
> I would do
>
> + struct dma_slave_config config = {
> + .direction = DMA_DEV_TO_MEM,
> + .src_maxburst = 8,
> + };
>
> and have all other fields initialised to 0 automatically.
Euh and what would do the "to 0 automatically" initialization ? It's an
automatic variable, not a static one.
>> + dma_cap_zero(mask);
>> + dma_cap_set(DMA_SLAVE, mask);
>
> Also DMA_PRIVATE?
Most certainly.
So thanks for the very detailed review, and I think we'll iterate a couple of
mails to converge. I especially have high hopes about the "slice" thing, two
minds are better than a single coffee broken one :)
Cheers.
--
Robert