From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-9.7 required=3.0 tests=DKIM_SIGNED,DKIM_VALID, HEADER_FROM_DIFFERENT_DOMAINS,INCLUDES_PATCH,MAILING_LIST_MULTI,SIGNED_OFF_BY, SPF_HELO_NONE,SPF_PASS,URIBL_BLOCKED,USER_AGENT_GIT autolearn=unavailable autolearn_force=no version=3.4.0 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by smtp.lore.kernel.org (Postfix) with ESMTP id CB68FC34050 for ; Wed, 19 Feb 2020 14:10:19 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by mail.kernel.org (Postfix) with ESMTP id 98C6320801 for ; Wed, 19 Feb 2020 14:10:19 +0000 (UTC) Authentication-Results: mail.kernel.org; dkim=pass (2048-bit key) header.d=baylibre-com.20150623.gappssmtp.com header.i=@baylibre-com.20150623.gappssmtp.com header.b="keYetL6I" Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1728209AbgBSOKS (ORCPT ); Wed, 19 Feb 2020 09:10:18 -0500 Received: from mail-wm1-f67.google.com ([209.85.128.67]:34542 "EHLO mail-wm1-f67.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1727833AbgBSOKL (ORCPT ); Wed, 19 Feb 2020 09:10:11 -0500 Received: by mail-wm1-f67.google.com with SMTP id s144so450318wme.1 for ; Wed, 19 Feb 2020 06:10:09 -0800 (PST) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=baylibre-com.20150623.gappssmtp.com; s=20150623; h=from:to:cc:subject:date:message-id:in-reply-to:references :mime-version:content-transfer-encoding; bh=bO+zCL5u73T+GZvMtGYQd6cU1mmqrjPqTwiTFGUJPt0=; b=keYetL6IKdKjosctUtfX0Up8yFA+EqwuNLnjprEIc0PeBM3/aAWaGEBV2kJ3+VPAwP h2dIyVemookn9f3QYWHGeb1QJMnmydBojCniZTLwbJysDLvJ4nNZ+iiLed9Zhwl+2+i3 LaWHnzxBpNFgM8xH6beIDIVaIF2uC3r4cRbnf9o0YUx3Pzz6G7fDLsp1Fr2jGC1If71i Z8ZEAfj7QxBueUBvjBUYQzDVPitAEAcm63H1pwkd/VdOVLYW+4YTAzrLxkDn3Os6wjwJ qKg5vL/oqEOATZsUd1F0/6aCMlq6QO13r/4IhA0knL72YTRy/Aiv2QFGR0C6t7gTbnBy 09WQ== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20161025; h=x-gm-message-state:from:to:cc:subject:date:message-id:in-reply-to :references:mime-version:content-transfer-encoding; bh=bO+zCL5u73T+GZvMtGYQd6cU1mmqrjPqTwiTFGUJPt0=; b=HlZA0eRG/64kS3kqzElUwYRKW8A0nS/h/he/xgm2RL53YlGCQqca2iJW1nWAj14R7y htT6kw+CjCZqg/GIVBNEYL02PjHcSffUSkx9aTcnHhlP3JawQts+paBsijX3kL85HoUB 4oy80AY3b+SeFxPBM18dvu1B0h4a4ueTtVAyeONI3TZUWVIKbBqNB3Cl8TfUJdP0uSiv XSC5dn5lfW4xTuiZpPwnxZ1Z+ZoyIO8t/d2q18UjT6DNSMqQFQ9VgflW60zSHCA1d5r+ kG0FclMxNa0Yp9s+sy3Ej5Es1AQIFOwmSh3lEXwfyMomtXcE/L9CDmxPrS65KBpfqFNm hBBQ== X-Gm-Message-State: APjAAAWg+fByusL3oXoSfKho9x8SQrnP3FBsbPGNsQBh1n232OAR5gu2 euQcXyU9suRp0GH7aSTQTPrXFC105CKYeA== X-Google-Smtp-Source: APXvYqzc/U2Yw16h4S7hTMYgsTEm+C8lAD5BiAA2G/gcI391coPZ4a8ehDCsSXuglNsh2CZczLC+lw== X-Received: by 2002:a1c:5f06:: with SMTP id t6mr10530418wmb.32.1582121408857; Wed, 19 Feb 2020 06:10:08 -0800 (PST) Received: from bender.baylibre.local ([2a01:e35:2ec0:82b0:4ca8:b25b:98e4:858]) by smtp.gmail.com with ESMTPSA id m68sm3182789wme.48.2020.02.19.06.10.07 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Wed, 19 Feb 2020 06:10:08 -0800 (PST) From: Neil Armstrong To: mchehab@kernel.org, hans.verkuil@cisco.com Cc: Maxime Jourdan , linux-media@vger.kernel.org, linux-amlogic@lists.infradead.org, linux-arm-kernel@lists.infradead.org, linux-kernel@vger.kernel.org, Neil Armstrong Subject: [PATCH v5 4/5] media: meson: vdec: add VP9 input support Date: Wed, 19 Feb 2020 15:09:57 +0100 Message-Id: <20200219140958.23542-5-narmstrong@baylibre.com> X-Mailer: git-send-email 2.22.0 In-Reply-To: <20200219140958.23542-1-narmstrong@baylibre.com> References: <20200219140958.23542-1-narmstrong@baylibre.com> MIME-Version: 1.0 Content-Transfer-Encoding: 8bit Sender: linux-kernel-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: linux-kernel@vger.kernel.org From: Maxime Jourdan Amlogic VP9 decoder requires an additional 16-byte payload before every frame header. The source buffer is updated in-place, then given to the Parser FIFO DMA. The FIFO DMA copies the blocks into the 16MiB parser ring buffer, then parses and copies the slice into the decoder "workspace". Signed-off-by: Maxime Jourdan Signed-off-by: Neil Armstrong --- drivers/staging/media/meson/vdec/esparser.c | 150 +++++++++++++++++++- 1 file changed, 146 insertions(+), 4 deletions(-) diff --git a/drivers/staging/media/meson/vdec/esparser.c b/drivers/staging/media/meson/vdec/esparser.c index adc5c1e81a4c..4a9aad3fafeb 100644 --- a/drivers/staging/media/meson/vdec/esparser.c +++ b/drivers/staging/media/meson/vdec/esparser.c @@ -52,6 +52,7 @@ #define PARSER_VIDEO_HOLE 0x90 #define SEARCH_PATTERN_LEN 512 +#define VP9_HEADER_SIZE 16 static DECLARE_WAIT_QUEUE_HEAD(wq); static int search_done; @@ -74,14 +75,121 @@ static irqreturn_t esparser_isr(int irq, void *dev) return IRQ_HANDLED; } +/** + * VP9 frame headers need to be appended by a 16-byte long + * Amlogic custom header + */ +static int vp9_update_header(struct amvdec_core *core, struct vb2_buffer *buf) +{ + u8 *dp; + u8 marker; + int dsize; + int num_frames, cur_frame; + int cur_mag, mag, mag_ptr; + int frame_size[8], tot_frame_size[8]; + int total_datasize = 0; + int new_frame_size; + unsigned char *old_header = NULL; + + dp = (uint8_t *)vb2_plane_vaddr(buf, 0); + dsize = vb2_get_plane_payload(buf, 0); + + if (dsize == vb2_plane_size(buf, 0)) { + dev_warn(core->dev, "%s: unable to update header\n", __func__); + return 0; + } + + marker = dp[dsize - 1]; + if ((marker & 0xe0) == 0xc0) { + num_frames = (marker & 0x7) + 1; + mag = ((marker >> 3) & 0x3) + 1; + mag_ptr = dsize - mag * num_frames - 2; + if (dp[mag_ptr] != marker) + return 0; + + mag_ptr++; + for (cur_frame = 0; cur_frame < num_frames; cur_frame++) { + frame_size[cur_frame] = 0; + for (cur_mag = 0; cur_mag < mag; cur_mag++) { + frame_size[cur_frame] |= + (dp[mag_ptr] << (cur_mag * 8)); + mag_ptr++; + } + if (cur_frame == 0) + tot_frame_size[cur_frame] = + frame_size[cur_frame]; + else + tot_frame_size[cur_frame] = + tot_frame_size[cur_frame - 1] + + frame_size[cur_frame]; + total_datasize += frame_size[cur_frame]; + } + } else { + num_frames = 1; + frame_size[0] = dsize; + tot_frame_size[0] = dsize; + total_datasize = dsize; + } + + new_frame_size = total_datasize + num_frames * VP9_HEADER_SIZE; + + if (new_frame_size >= vb2_plane_size(buf, 0)) { + dev_warn(core->dev, "%s: unable to update header\n", __func__); + return 0; + } + + for (cur_frame = num_frames - 1; cur_frame >= 0; cur_frame--) { + int framesize = frame_size[cur_frame]; + int framesize_header = framesize + 4; + int oldframeoff = tot_frame_size[cur_frame] - framesize; + int outheaderoff = oldframeoff + cur_frame * VP9_HEADER_SIZE; + u8 *fdata = dp + outheaderoff; + u8 *old_framedata = dp + oldframeoff; + + memmove(fdata + VP9_HEADER_SIZE, old_framedata, framesize); + + fdata[0] = (framesize_header >> 24) & 0xff; + fdata[1] = (framesize_header >> 16) & 0xff; + fdata[2] = (framesize_header >> 8) & 0xff; + fdata[3] = (framesize_header >> 0) & 0xff; + fdata[4] = ((framesize_header >> 24) & 0xff) ^ 0xff; + fdata[5] = ((framesize_header >> 16) & 0xff) ^ 0xff; + fdata[6] = ((framesize_header >> 8) & 0xff) ^ 0xff; + fdata[7] = ((framesize_header >> 0) & 0xff) ^ 0xff; + fdata[8] = 0; + fdata[9] = 0; + fdata[10] = 0; + fdata[11] = 1; + fdata[12] = 'A'; + fdata[13] = 'M'; + fdata[14] = 'L'; + fdata[15] = 'V'; + + if (!old_header) { + /* nothing */ + } else if (old_header > fdata + 16 + framesize) { + dev_dbg(core->dev, "%s: data has gaps, setting to 0\n", + __func__); + memset(fdata + 16 + framesize, 0, + (old_header - fdata + 16 + framesize)); + } else if (old_header < fdata + 16 + framesize) { + dev_err(core->dev, "%s: data overwritten\n", __func__); + } + old_header = fdata; + } + + return new_frame_size; +} + /* Pad the packet to at least 4KiB bytes otherwise the VDEC unit won't trigger * ISRs. * Also append a start code 000001ff at the end to trigger * the ESPARSER interrupt. */ -static u32 esparser_pad_start_code(struct amvdec_core *core, struct vb2_buffer *vb) +static u32 esparser_pad_start_code(struct amvdec_core *core, + struct vb2_buffer *vb, + u32 payload_size) { - u32 payload_size = vb2_get_plane_payload(vb, 0); u32 pad_size = 0; u8 *vaddr = vb2_plane_vaddr(vb, 0); @@ -186,13 +294,35 @@ esparser_queue(struct amvdec_session *sess, struct vb2_v4l2_buffer *vbuf) int ret; struct vb2_buffer *vb = &vbuf->vb2_buf; struct amvdec_core *core = sess->core; + struct amvdec_codec_ops *codec_ops = sess->fmt_out->codec_ops; u32 payload_size = vb2_get_plane_payload(vb, 0); dma_addr_t phy = vb2_dma_contig_plane_dma_addr(vb, 0); + u32 num_dst_bufs = 0; u32 offset; u32 pad_size; - if (esparser_vififo_get_free_space(sess) < payload_size) + /* + * When max ref frame is held by VP9, this should be -= 3 to prevent a + * shortage of CAPTURE buffers on the decoder side. + * For the future, a good enhancement of the way this is handled could + * be to notify new capture buffers to the decoding modules, so that + * they could pause when there is no capture buffer available and + * resume on this notification. + */ + if (sess->fmt_out->pixfmt == V4L2_PIX_FMT_VP9) { + if (codec_ops->num_pending_bufs) + num_dst_bufs = codec_ops->num_pending_bufs(sess); + + num_dst_bufs += v4l2_m2m_num_dst_bufs_ready(sess->m2m_ctx); + if (sess->fmt_out->pixfmt == V4L2_PIX_FMT_VP9) + num_dst_bufs -= 3; + + if (esparser_vififo_get_free_space(sess) < payload_size || + atomic_read(&sess->esparser_queued_bufs) >= num_dst_bufs) + return -EAGAIN; + } else if (esparser_vififo_get_free_space(sess) < payload_size) { return -EAGAIN; + } v4l2_m2m_src_buf_remove_by_buf(sess->m2m_ctx, vbuf); @@ -206,7 +336,19 @@ esparser_queue(struct amvdec_session *sess, struct vb2_v4l2_buffer *vbuf) vbuf->field = V4L2_FIELD_NONE; vbuf->sequence = sess->sequence_out++; - pad_size = esparser_pad_start_code(core, vb); + if (sess->fmt_out->pixfmt == V4L2_PIX_FMT_VP9) { + payload_size = vp9_update_header(core, vb); + + /* If unable to alter buffer to add headers */ + if (payload_size == 0) { + amvdec_remove_ts(sess, vb->timestamp); + v4l2_m2m_buf_done(vbuf, VB2_BUF_STATE_ERROR); + + return 0; + } + } + + pad_size = esparser_pad_start_code(core, vb, payload_size); ret = esparser_write_data(core, phy, payload_size + pad_size); if (ret <= 0) { -- 2.22.0