Re: [PATCH v8 5/8] [media] vcodec: mediatek: Add Mediatek V4L2 Video Encoder Driver

From: Hans Verkuil
Date: Tue Apr 26 2016 - 02:21:02 EST


On 04/26/2016 04:17 AM, tiffany lin wrote:
> Hi Hans,
>
> On Mon, 2016-04-25 at 15:13 +0200, Hans Verkuil wrote:
>> On 04/25/2016 02:30 PM, Tiffany Lin wrote:
>>> Add v4l2 layer encoder driver for MT8173
>>>
>>> Signed-off-by: Tiffany Lin <tiffany.lin@xxxxxxxxxxxx>
>>>
>>> ---
>>> drivers/media/platform/Kconfig | 16 +
>>> drivers/media/platform/Makefile | 2 +
>>> drivers/media/platform/mtk-vcodec/Makefile | 14 +
>>> drivers/media/platform/mtk-vcodec/mtk_vcodec_drv.h | 339 +++++
>>> drivers/media/platform/mtk-vcodec/mtk_vcodec_enc.c | 1303 ++++++++++++++++++++
>>> drivers/media/platform/mtk-vcodec/mtk_vcodec_enc.h | 58 +
>>> .../media/platform/mtk-vcodec/mtk_vcodec_enc_drv.c | 456 +++++++
>>> .../media/platform/mtk-vcodec/mtk_vcodec_enc_pm.c | 137 ++
>>> .../media/platform/mtk-vcodec/mtk_vcodec_enc_pm.h | 26 +
>>> .../media/platform/mtk-vcodec/mtk_vcodec_intr.c | 56 +
>>> .../media/platform/mtk-vcodec/mtk_vcodec_intr.h | 27 +
>>> .../media/platform/mtk-vcodec/mtk_vcodec_util.c | 96 ++
>>> .../media/platform/mtk-vcodec/mtk_vcodec_util.h | 87 ++
>>> drivers/media/platform/mtk-vcodec/venc_drv_base.h | 62 +
>>> drivers/media/platform/mtk-vcodec/venc_drv_if.c | 107 ++
>>> drivers/media/platform/mtk-vcodec/venc_drv_if.h | 165 +++
>>> drivers/media/platform/mtk-vcodec/venc_ipi_msg.h | 210 ++++
>>> 17 files changed, 3161 insertions(+)
>>> create mode 100644 drivers/media/platform/mtk-vcodec/Makefile
>>> create mode 100644 drivers/media/platform/mtk-vcodec/mtk_vcodec_drv.h
>>> create mode 100644 drivers/media/platform/mtk-vcodec/mtk_vcodec_enc.c
>>> create mode 100644 drivers/media/platform/mtk-vcodec/mtk_vcodec_enc.h
>>> create mode 100644 drivers/media/platform/mtk-vcodec/mtk_vcodec_enc_drv.c
>>> create mode 100644 drivers/media/platform/mtk-vcodec/mtk_vcodec_enc_pm.c
>>> create mode 100644 drivers/media/platform/mtk-vcodec/mtk_vcodec_enc_pm.h
>>> create mode 100644 drivers/media/platform/mtk-vcodec/mtk_vcodec_intr.c
>>> create mode 100644 drivers/media/platform/mtk-vcodec/mtk_vcodec_intr.h
>>> create mode 100644 drivers/media/platform/mtk-vcodec/mtk_vcodec_util.c
>>> create mode 100644 drivers/media/platform/mtk-vcodec/mtk_vcodec_util.h
>>> create mode 100644 drivers/media/platform/mtk-vcodec/venc_drv_base.h
>>> create mode 100644 drivers/media/platform/mtk-vcodec/venc_drv_if.c
>>> create mode 100644 drivers/media/platform/mtk-vcodec/venc_drv_if.h
>>> create mode 100644 drivers/media/platform/mtk-vcodec/venc_ipi_msg.h
>>>
>>
>> ...
>>
>>> +static int fops_vcodec_open(struct file *file)
>>> +{
>>> + struct mtk_vcodec_dev *dev = video_drvdata(file);
>>> + struct mtk_vcodec_ctx *ctx = NULL;
>>> + int ret = 0;
>>> +
>>> + mutex_lock(&dev->dev_mutex);
>>
>> I would move this to after the kzalloc, since that doesn't need to be
>> called with the mutex held.
>>
> Will fix this in next version.
>
>>> +
>>> + ctx = kzalloc(sizeof(*ctx), GFP_KERNEL);
>>> + if (!ctx) {
>>
>> And then you can just do return -ENOMEM here.
>>
> Got it.
>>> + ret = -ENOMEM;
>>> + goto err_alloc;
>>> + }
>>> +
>>> + ctx->idx = dev->curr_max_idx;
>>
>> I'd do:
>>
>> /*
>> * Use simple counter to uniquely identify this context. Only
>> * used for logging.
>> */
>> ctx->idx = dev->curr_max_idx++;
>>
>> I would also prefer that it is call 'id' instead of idx, since that's really
>> what it is, just an ID.
>>
>> And the counter is then id_counter instead of curr_max_idx.
>>
> Got it, will fix this in next version.
>
>>> + v4l2_fh_init(&ctx->fh, video_devdata(file));
>>> + file->private_data = &ctx->fh;
>>> + v4l2_fh_add(&ctx->fh);
>>> + INIT_LIST_HEAD(&ctx->list);
>>> + ctx->dev = dev;
>>> + init_waitqueue_head(&ctx->queue);
>>> +
>>> + ctx->type = MTK_INST_ENCODER;
>>> + ret = mtk_vcodec_enc_ctrls_setup(ctx);
>>> + if (ret) {
>>> + mtk_v4l2_err("Failed to setup controls() (%d)",
>>> + ret);
>>> + goto err_ctrls_setup;
>>> + }
>>> + ctx->m2m_ctx = v4l2_m2m_ctx_init(dev->m2m_dev_enc, ctx,
>>> + &mtk_vcodec_enc_queue_init);
>>> + if (IS_ERR(ctx->m2m_ctx)) {
>>> + ret = PTR_ERR(ctx->m2m_ctx);
>>> + mtk_v4l2_err("Failed to v4l2_m2m_ctx_init() (%d)",
>>> + ret);
>>> + goto err_m2m_ctx_init;
>>> + }
>>> + mtk_vcodec_enc_set_default_params(ctx);
>>> +
>>> + if (v4l2_fh_is_singular(&ctx->fh)) {
>>> + /*
>>> + * vpu_load_firmware checks if it was loaded already and
>>> + * does nothing in that case
>>> + */
>>> + ret = vpu_load_firmware(dev->vpu_plat_dev);
>>> + if (ret < 0) {
>>> + /*
>>> + * Return 0 if downloading firmware successfully,
>>> + * otherwise it is failed
>>> + */
>>> + mtk_v4l2_err("vpu_load_firmware failed!");
>>> + goto err_load_fw;
>>> + }
>>> +
>>> + dev->enc_capability =
>>> + vpu_get_venc_hw_capa(dev->vpu_plat_dev);
>>> + mtk_v4l2_debug(0, "encoder capability %x", dev->enc_capability);
>>> + }
>>> +
>>> + mtk_v4l2_debug(2, "Create instance [%d]@%p m2m_ctx=%p ",
>>> + ctx->idx, ctx, ctx->m2m_ctx);
>>> + dev->curr_max_idx = (dev->curr_max_idx + 1) & 0xffffffffffffffff;
>>
>> and drop this line.
>>
> Got it.
>
>>> + dev->num_instances++;
>>> + list_add(&ctx->list, &dev->ctx_list);
>>> +
>>> + mutex_unlock(&dev->dev_mutex);
>>> + mtk_v4l2_debug(0, "%s encoder [%d]", dev_name(&dev->plat_dev->dev),
>>> + ctx->idx);
>>> + return ret;
>>> +
>>> + /* Deinit when failure occurred */
>>> +err_load_fw:
>>> + v4l2_m2m_ctx_release(ctx->m2m_ctx);
>>> +err_m2m_ctx_init:
>>> + v4l2_ctrl_handler_free(&ctx->ctrl_hdl);
>>> +err_ctrls_setup:
>>> + v4l2_fh_del(&ctx->fh);
>>> + v4l2_fh_exit(&ctx->fh);
>>> + kfree(ctx);
>>> +err_alloc:
>>> + mutex_unlock(&dev->dev_mutex);
>>> +
>>> + return ret;
>>> +}
>>> +
>>> +static int fops_vcodec_release(struct file *file)
>>> +{
>>> + struct mtk_vcodec_dev *dev = video_drvdata(file);
>>> + struct mtk_vcodec_ctx *ctx = fh_to_ctx(file->private_data);
>>> +
>>> + mtk_v4l2_debug(1, "[%d] encoder", ctx->idx);
>>> + mutex_lock(&dev->dev_mutex);
>>> +
>>> + mtk_vcodec_enc_release(ctx);
>>> + v4l2_fh_del(&ctx->fh);
>>> + v4l2_fh_exit(&ctx->fh);
>>> + v4l2_ctrl_handler_free(&ctx->ctrl_hdl);
>>> + v4l2_m2m_ctx_release(ctx->m2m_ctx);
>>> +
>>> + list_del_init(&ctx->list);
>>> + dev->num_instances--;
>>> + kfree(ctx);
>>> + mutex_unlock(&dev->dev_mutex);
>>> + return 0;
>>> +}
>>
>> Tiffany, besides these very small items here is there anything else you are waiting for?
>> I'm thinking that if you fix these small issues, merge Julia Lawall's fixes and repost,
>> then I can make a pull request. Mauro will still have to review it as well, but I have no
>> more comments.
>>
> For Julia Lawall's fixes, do you mean we need merge
> 1.[PATCH] VPU: mediatek: fix simple_open.cocci warnings
> 2.[PATCH] VPU: mediatek: fix platform_no_drv_owner.cocci warnings
> to "[PATCH v8 2/8] [media] VPU: mediatek: support Mediatek VPU"

Yes.

> or add these two patches to this patch series?
> We are ready for our encoder patch series, I will fix these issues and
> repost patches today.

Great! Please make sure it is rebased to the latest media_tree master branch.
I'll likely make a pull request for Mauro on Friday.

The driver will probably not make it for 4.7 as it looks likely that the
4.7 merge window will open soon. But 4.8 shouldn't be a problem at all.

Regards,

Hans