diff options
Diffstat (limited to 'drivers/staging/media/meson/vdec/vdec_helpers.c')
-rw-r--r-- | drivers/staging/media/meson/vdec/vdec_helpers.c | 449 |
1 files changed, 449 insertions, 0 deletions
diff --git a/drivers/staging/media/meson/vdec/vdec_helpers.c b/drivers/staging/media/meson/vdec/vdec_helpers.c new file mode 100644 index 000000000000..f16948bdbf2f --- /dev/null +++ b/drivers/staging/media/meson/vdec/vdec_helpers.c @@ -0,0 +1,449 @@ +// SPDX-License-Identifier: GPL-2.0+ +/* + * Copyright (C) 2018 BayLibre, SAS + * Author: Maxime Jourdan <mjourdan@baylibre.com> + */ + +#include <linux/gcd.h> +#include <media/v4l2-mem2mem.h> +#include <media/v4l2-event.h> +#include <media/videobuf2-dma-contig.h> + +#include "vdec_helpers.h" + +#define NUM_CANVAS_NV12 2 +#define NUM_CANVAS_YUV420 3 + +u32 amvdec_read_dos(struct amvdec_core *core, u32 reg) +{ + return readl_relaxed(core->dos_base + reg); +} +EXPORT_SYMBOL_GPL(amvdec_read_dos); + +void amvdec_write_dos(struct amvdec_core *core, u32 reg, u32 val) +{ + writel_relaxed(val, core->dos_base + reg); +} +EXPORT_SYMBOL_GPL(amvdec_write_dos); + +void amvdec_write_dos_bits(struct amvdec_core *core, u32 reg, u32 val) +{ + amvdec_write_dos(core, reg, amvdec_read_dos(core, reg) | val); +} +EXPORT_SYMBOL_GPL(amvdec_write_dos_bits); + +void amvdec_clear_dos_bits(struct amvdec_core *core, u32 reg, u32 val) +{ + amvdec_write_dos(core, reg, amvdec_read_dos(core, reg) & ~val); +} +EXPORT_SYMBOL_GPL(amvdec_clear_dos_bits); + +u32 amvdec_read_parser(struct amvdec_core *core, u32 reg) +{ + return readl_relaxed(core->esparser_base + reg); +} +EXPORT_SYMBOL_GPL(amvdec_read_parser); + +void amvdec_write_parser(struct amvdec_core *core, u32 reg, u32 val) +{ + writel_relaxed(val, core->esparser_base + reg); +} +EXPORT_SYMBOL_GPL(amvdec_write_parser); + +static int canvas_alloc(struct amvdec_session *sess, u8 *canvas_id) +{ + int ret; + + if (sess->canvas_num >= MAX_CANVAS) { + dev_err(sess->core->dev, "Reached max number of canvas\n"); + return -ENOMEM; + } + + ret = meson_canvas_alloc(sess->core->canvas, canvas_id); + if (ret) + return ret; + + sess->canvas_alloc[sess->canvas_num++] = *canvas_id; + return 0; +} + +static int set_canvas_yuv420m(struct amvdec_session *sess, + struct vb2_buffer *vb, u32 width, + u32 height, u32 reg) +{ + struct amvdec_core *core = sess->core; + u8 canvas_id[NUM_CANVAS_YUV420]; /* Y U V */ + dma_addr_t buf_paddr[NUM_CANVAS_YUV420]; /* Y U V */ + int ret, i; + + for (i = 0; i < NUM_CANVAS_YUV420; ++i) { + ret = canvas_alloc(sess, &canvas_id[i]); + if (ret) + return ret; + + buf_paddr[i] = + vb2_dma_contig_plane_dma_addr(vb, i); + } + + /* Y plane */ + meson_canvas_config(core->canvas, canvas_id[0], buf_paddr[0], + width, height, MESON_CANVAS_WRAP_NONE, + MESON_CANVAS_BLKMODE_LINEAR, + MESON_CANVAS_ENDIAN_SWAP64); + + /* U plane */ + meson_canvas_config(core->canvas, canvas_id[1], buf_paddr[1], + width / 2, height / 2, MESON_CANVAS_WRAP_NONE, + MESON_CANVAS_BLKMODE_LINEAR, + MESON_CANVAS_ENDIAN_SWAP64); + + /* V plane */ + meson_canvas_config(core->canvas, canvas_id[2], buf_paddr[2], + width / 2, height / 2, MESON_CANVAS_WRAP_NONE, + MESON_CANVAS_BLKMODE_LINEAR, + MESON_CANVAS_ENDIAN_SWAP64); + + amvdec_write_dos(core, reg, + ((canvas_id[2]) << 16) | + ((canvas_id[1]) << 8) | + (canvas_id[0])); + + return 0; +} + +static int set_canvas_nv12m(struct amvdec_session *sess, + struct vb2_buffer *vb, u32 width, + u32 height, u32 reg) +{ + struct amvdec_core *core = sess->core; + u8 canvas_id[NUM_CANVAS_NV12]; /* Y U/V */ + dma_addr_t buf_paddr[NUM_CANVAS_NV12]; /* Y U/V */ + int ret, i; + + for (i = 0; i < NUM_CANVAS_NV12; ++i) { + ret = canvas_alloc(sess, &canvas_id[i]); + if (ret) + return ret; + + buf_paddr[i] = + vb2_dma_contig_plane_dma_addr(vb, i); + } + + /* Y plane */ + meson_canvas_config(core->canvas, canvas_id[0], buf_paddr[0], + width, height, MESON_CANVAS_WRAP_NONE, + MESON_CANVAS_BLKMODE_LINEAR, + MESON_CANVAS_ENDIAN_SWAP64); + + /* U/V plane */ + meson_canvas_config(core->canvas, canvas_id[1], buf_paddr[1], + width, height / 2, MESON_CANVAS_WRAP_NONE, + MESON_CANVAS_BLKMODE_LINEAR, + MESON_CANVAS_ENDIAN_SWAP64); + + amvdec_write_dos(core, reg, + ((canvas_id[1]) << 16) | + ((canvas_id[1]) << 8) | + (canvas_id[0])); + + return 0; +} + +int amvdec_set_canvases(struct amvdec_session *sess, + u32 reg_base[], u32 reg_num[]) +{ + struct v4l2_m2m_buffer *buf; + u32 pixfmt = sess->pixfmt_cap; + u32 width = ALIGN(sess->width, 64); + u32 height = ALIGN(sess->height, 64); + u32 reg_cur = reg_base[0]; + u32 reg_num_cur = 0; + u32 reg_base_cur = 0; + int i = 0; + int ret; + + v4l2_m2m_for_each_dst_buf(sess->m2m_ctx, buf) { + if (!reg_base[reg_base_cur]) + return -EINVAL; + + reg_cur = reg_base[reg_base_cur] + reg_num_cur * 4; + + switch (pixfmt) { + case V4L2_PIX_FMT_NV12M: + ret = set_canvas_nv12m(sess, &buf->vb.vb2_buf, width, + height, reg_cur); + if (ret) + return ret; + break; + case V4L2_PIX_FMT_YUV420M: + ret = set_canvas_yuv420m(sess, &buf->vb.vb2_buf, width, + height, reg_cur); + if (ret) + return ret; + break; + default: + dev_err(sess->core->dev, "Unsupported pixfmt %08X\n", + pixfmt); + return -EINVAL; + } + + reg_num_cur++; + if (reg_num_cur >= reg_num[reg_base_cur]) { + reg_base_cur++; + reg_num_cur = 0; + } + + sess->fw_idx_to_vb2_idx[i++] = buf->vb.vb2_buf.index; + } + + return 0; +} +EXPORT_SYMBOL_GPL(amvdec_set_canvases); + +void amvdec_add_ts_reorder(struct amvdec_session *sess, u64 ts, u32 offset) +{ + struct amvdec_timestamp *new_ts, *tmp; + unsigned long flags; + + new_ts = kmalloc(sizeof(*new_ts), GFP_KERNEL); + new_ts->ts = ts; + new_ts->offset = offset; + + spin_lock_irqsave(&sess->ts_spinlock, flags); + + if (list_empty(&sess->timestamps)) + goto add_tail; + + list_for_each_entry(tmp, &sess->timestamps, list) { + if (ts <= tmp->ts) { + list_add_tail(&new_ts->list, &tmp->list); + goto unlock; + } + } + +add_tail: + list_add_tail(&new_ts->list, &sess->timestamps); +unlock: + spin_unlock_irqrestore(&sess->ts_spinlock, flags); +} +EXPORT_SYMBOL_GPL(amvdec_add_ts_reorder); + +void amvdec_remove_ts(struct amvdec_session *sess, u64 ts) +{ + struct amvdec_timestamp *tmp; + unsigned long flags; + + spin_lock_irqsave(&sess->ts_spinlock, flags); + list_for_each_entry(tmp, &sess->timestamps, list) { + if (tmp->ts == ts) { + list_del(&tmp->list); + kfree(tmp); + goto unlock; + } + } + dev_warn(sess->core->dev_dec, + "Couldn't remove buffer with timestamp %llu from list\n", ts); + +unlock: + spin_unlock_irqrestore(&sess->ts_spinlock, flags); +} +EXPORT_SYMBOL_GPL(amvdec_remove_ts); + +static void dst_buf_done(struct amvdec_session *sess, + struct vb2_v4l2_buffer *vbuf, + u32 field, + u64 timestamp) +{ + struct device *dev = sess->core->dev_dec; + u32 output_size = amvdec_get_output_size(sess); + + switch (sess->pixfmt_cap) { + case V4L2_PIX_FMT_NV12M: + vbuf->vb2_buf.planes[0].bytesused = output_size; + vbuf->vb2_buf.planes[1].bytesused = output_size / 2; + break; + case V4L2_PIX_FMT_YUV420M: + vbuf->vb2_buf.planes[0].bytesused = output_size; + vbuf->vb2_buf.planes[1].bytesused = output_size / 4; + vbuf->vb2_buf.planes[2].bytesused = output_size / 4; + break; + } + + vbuf->vb2_buf.timestamp = timestamp; + vbuf->sequence = sess->sequence_cap++; + + if (sess->should_stop && + atomic_read(&sess->esparser_queued_bufs) <= 2) { + const struct v4l2_event ev = { .type = V4L2_EVENT_EOS }; + + dev_dbg(dev, "Signaling EOS\n"); + v4l2_event_queue_fh(&sess->fh, &ev); + vbuf->flags |= V4L2_BUF_FLAG_LAST; + } else if (sess->should_stop) + dev_dbg(dev, "should_stop, %u bufs remain\n", + atomic_read(&sess->esparser_queued_bufs)); + + dev_dbg(dev, "Buffer %u done\n", vbuf->vb2_buf.index); + vbuf->field = field; + v4l2_m2m_buf_done(vbuf, VB2_BUF_STATE_DONE); + + /* Buffer done probably means the vififo got freed */ + schedule_work(&sess->esparser_queue_work); +} + +void amvdec_dst_buf_done(struct amvdec_session *sess, + struct vb2_v4l2_buffer *vbuf, u32 field) +{ + struct device *dev = sess->core->dev_dec; + struct amvdec_timestamp *tmp; + struct list_head *timestamps = &sess->timestamps; + u64 timestamp; + unsigned long flags; + + spin_lock_irqsave(&sess->ts_spinlock, flags); + if (list_empty(timestamps)) { + dev_err(dev, "Buffer %u done but list is empty\n", + vbuf->vb2_buf.index); + + v4l2_m2m_buf_done(vbuf, VB2_BUF_STATE_ERROR); + spin_unlock_irqrestore(&sess->ts_spinlock, flags); + return; + } + + tmp = list_first_entry(timestamps, struct amvdec_timestamp, list); + timestamp = tmp->ts; + list_del(&tmp->list); + kfree(tmp); + spin_unlock_irqrestore(&sess->ts_spinlock, flags); + + dst_buf_done(sess, vbuf, field, timestamp); + atomic_dec(&sess->esparser_queued_bufs); +} +EXPORT_SYMBOL_GPL(amvdec_dst_buf_done); + +void amvdec_dst_buf_done_offset(struct amvdec_session *sess, + struct vb2_v4l2_buffer *vbuf, + u32 offset, u32 field, bool allow_drop) +{ + struct device *dev = sess->core->dev_dec; + struct amvdec_timestamp *match = NULL; + struct amvdec_timestamp *tmp, *n; + u64 timestamp = 0; + unsigned long flags; + + spin_lock_irqsave(&sess->ts_spinlock, flags); + + /* Look for our vififo offset to get the corresponding timestamp. */ + list_for_each_entry_safe(tmp, n, &sess->timestamps, list) { + s64 delta = (s64)offset - tmp->offset; + + /* Offsets reported by codecs usually differ slightly, + * so we need some wiggle room. + * 4KiB being the minimum packet size, there is no risk here. + */ + if (delta > (-1 * (s32)SZ_4K) && delta < SZ_4K) { + match = tmp; + break; + } + + if (!allow_drop) + continue; + + /* Delete any timestamp entry that appears before our target + * (not all src packets/timestamps lead to a frame) + */ + if (delta > 0 || delta < -1 * (s32)sess->vififo_size) { + atomic_dec(&sess->esparser_queued_bufs); + list_del(&tmp->list); + kfree(tmp); + } + } + + if (!match) { + dev_dbg(dev, "Buffer %u done but can't match offset (%08X)\n", + vbuf->vb2_buf.index, offset); + } else { + timestamp = match->ts; + list_del(&match->list); + kfree(match); + } + spin_unlock_irqrestore(&sess->ts_spinlock, flags); + + dst_buf_done(sess, vbuf, field, timestamp); + if (match) + atomic_dec(&sess->esparser_queued_bufs); +} +EXPORT_SYMBOL_GPL(amvdec_dst_buf_done_offset); + +void amvdec_dst_buf_done_idx(struct amvdec_session *sess, + u32 buf_idx, u32 offset, u32 field) +{ + struct vb2_v4l2_buffer *vbuf; + struct device *dev = sess->core->dev_dec; + + vbuf = v4l2_m2m_dst_buf_remove_by_idx(sess->m2m_ctx, + sess->fw_idx_to_vb2_idx[buf_idx]); + + if (!vbuf) { + dev_err(dev, + "Buffer %u done but it doesn't exist in m2m_ctx\n", + buf_idx); + return; + } + + if (offset != -1) + amvdec_dst_buf_done_offset(sess, vbuf, offset, field, true); + else + amvdec_dst_buf_done(sess, vbuf, field); +} +EXPORT_SYMBOL_GPL(amvdec_dst_buf_done_idx); + +void amvdec_set_par_from_dar(struct amvdec_session *sess, + u32 dar_num, u32 dar_den) +{ + u32 div; + + sess->pixelaspect.numerator = sess->height * dar_num; + sess->pixelaspect.denominator = sess->width * dar_den; + div = gcd(sess->pixelaspect.numerator, sess->pixelaspect.denominator); + sess->pixelaspect.numerator /= div; + sess->pixelaspect.denominator /= div; +} +EXPORT_SYMBOL_GPL(amvdec_set_par_from_dar); + +void amvdec_src_change(struct amvdec_session *sess, u32 width, + u32 height, u32 dpb_size) +{ + static const struct v4l2_event ev = { + .type = V4L2_EVENT_SOURCE_CHANGE, + .u.src_change.changes = V4L2_EVENT_SRC_CH_RESOLUTION }; + + v4l2_ctrl_s_ctrl(sess->ctrl_min_buf_capture, dpb_size); + + /* Check if the capture queue is already configured well for our + * usecase. If so, keep decoding with it and do not send the event + */ + if (sess->width == width && + sess->height == height && + dpb_size <= sess->num_dst_bufs) { + sess->fmt_out->codec_ops->resume(sess); + return; + } + + sess->width = width; + sess->height = height; + sess->status = STATUS_NEEDS_RESUME; + + dev_dbg(sess->core->dev, "Res. changed (%ux%u), DPB size %u\n", + width, height, dpb_size); + v4l2_event_queue_fh(&sess->fh, &ev); +} +EXPORT_SYMBOL_GPL(amvdec_src_change); + +void amvdec_abort(struct amvdec_session *sess) +{ + dev_info(sess->core->dev, "Aborting decoding session!\n"); + vb2_queue_error(&sess->m2m_ctx->cap_q_ctx.q); + vb2_queue_error(&sess->m2m_ctx->out_q_ctx.q); +} +EXPORT_SYMBOL_GPL(amvdec_abort); |