1
0
mirror of https://github.com/mpv-player/mpv synced 2024-12-26 00:42:57 +00:00
mpv/libvo/vo_dxr3.c

516 lines
14 KiB
C
Raw Normal View History

/*
* vo_dxr3.c - DXR3/H+ video out
*
* Copyright (C) 2002 David Holm <dholm@iname.com>
*
*/
/* ChangeLog added 2002-01-10
* 2002-02-18:
* Fixed sync problems when pausing video (while using prebuffering)
*
* 2002-02-16:
* Fixed bug which would case invalid output when using :noprebuf
* Removed equalization code, it caused problems on slow systems
*
* 2002-02-13:
* Using the swscaler instead of the old hand coded shit. (Checkout man mplayer and search for sws ;).
* Using aspect function to setup a proper mpeg1, no more hassling with odd resolutions or GOP-sizes,
* this would only create jitter on some vids!
* The swscaler sometimes exits with sig8 on mpegs, I don't know why yet (just use -vc mpegpes in this
* case, and report to me if you have any avi's etc which does this...)
*
* 2002-02-09:
* Thanks to the new control() method I have finally been able to enable the em8300 prebuffering.
* This should speed up playback on all systems, the vout cpu usage should rocket since I will be hogging
* the pci bus. Not to worry though, since frames are prebuffered it should be able to take a few blows
* if you start doing other stuff simultaneously.
*
* 2002-02-03:
* Removal of libmp1e, libavcodec has finally become faster (and it's code is helluva lot cleaner)
*
* 2002-02-02:
* Cleaned out some old code which might have slowed down writes
*
Added new syncengine thanks to a new previously undocumented feature of the em8300, this might fix playback on both slow and fast machines (more testing needed). This also requires users to get the em8300 driver from cvs until the next version is released (will probably happen this weekend) Added lots of comments, should be pretty easy to understand most of the internals now Added lots of brackets to if's for's while's etc, this is not a cosmetical thing but rather due to the fact I got some very odd bugs with else's since I didn't properly use brackets (and it's the K&R standard to have brackets everywhere) Fixed some bugs that would occur when disabling libmp1e Switched to default to the new naming scheme of device nodes, the driver will slowly switch over to this state, if it can't find devices under the new name it will try the old naming scheme I stopped opening devices in non-blocking mode, it would break the new syncengine which tries to burst data to the device (alot of times meaning it will fill the fifo pretty fast which would previously result in jerkyness on fast machines) The device now sets the initial state of the pts and speed (probably not needed, but assumption is the mother of all fuckups =) Keep the control interface open during the entire duration of the libvo device, we might need this to flush video buffers on seeking (currently not implemented, therefore seeking is broken) This is beta stuff to the driver, I will get some users to test it for me and do my best to fix seeking as soon as possible... git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@4219 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-01-17 10:33:47 +00:00
* 2002-01-17:
* Testrelease of new sync engine (using previously undocumented feature of em8300).
Added new syncengine thanks to a new previously undocumented feature of the em8300, this might fix playback on both slow and fast machines (more testing needed). This also requires users to get the em8300 driver from cvs until the next version is released (will probably happen this weekend) Added lots of comments, should be pretty easy to understand most of the internals now Added lots of brackets to if's for's while's etc, this is not a cosmetical thing but rather due to the fact I got some very odd bugs with else's since I didn't properly use brackets (and it's the K&R standard to have brackets everywhere) Fixed some bugs that would occur when disabling libmp1e Switched to default to the new naming scheme of device nodes, the driver will slowly switch over to this state, if it can't find devices under the new name it will try the old naming scheme I stopped opening devices in non-blocking mode, it would break the new syncengine which tries to burst data to the device (alot of times meaning it will fill the fifo pretty fast which would previously result in jerkyness on fast machines) The device now sets the initial state of the pts and speed (probably not needed, but assumption is the mother of all fuckups =) Keep the control interface open during the entire duration of the libvo device, we might need this to flush video buffers on seeking (currently not implemented, therefore seeking is broken) This is beta stuff to the driver, I will get some users to test it for me and do my best to fix seeking as soon as possible... git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@4219 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-01-17 10:33:47 +00:00
*
* 2002-01-15:
* Preliminary subpic support with -vc mpegpes and dvd's
* Device interfaces tries the new naming scheme by default (even though most users probably still use the old one)
*
* 2002-01-10:
* I rehauled the entire codebase. I have now changed to
* Kernighan & Ritchie codingstyle, please mail me if you
* find any inconcistencies.
*/
#include <linux/em8300.h>
#include <sys/ioctl.h>
#include <sys/stat.h>
#include <sys/types.h>
#include <sys/select.h>
#include <unistd.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <fcntl.h>
#include <stdio.h>
#include <time.h>
#include "config.h"
#include "fastmemcpy.h"
#include "video_out.h"
#include "video_out_internal.h"
#include "aspect.h"
#include "../postproc/rgb2rgb.h"
#include "../postproc/swscale.h"
#ifdef USE_LIBAVCODEC
#ifdef USE_LIBAVCODEC_SO
#include <libffmpeg/avcodec.h>
#else
#include "libavcodec/avcodec.h"
#endif
/* for video encoder */
static AVCodec *avc_codec = NULL;
static AVCodecContext *avc_context = NULL;
static AVPicture avc_picture;
int avc_outbuf_size = 100000;
#endif
char *picture_buf = NULL;
#ifdef HAVE_MMX
#include "mmx.h"
#endif
LIBVO_EXTERN (dxr3)
/* Resolutions and positions */
static int v_width, v_height;
static int s_width, s_height;
static int s_pos_x, s_pos_y;
static int d_pos_x, d_pos_y;
static int osd_w, osd_h;
static int noprebuf = 0;
static int img_format = 0;
static SwsContext * sws = NULL;
/* File descriptors */
static int fd_control = -1;
static int fd_video = -1;
static int fd_spu = -1;
static char fdv_name[80];
/* Static variable used in ioctl's */
static int ioval = 0;
static vo_info_t vo_info =
{
"DXR3/H+ video out",
"dxr3",
"David Holm <dholm@iname.com>",
""
};
uint32_t control(uint32_t request, void *data, ...)
{
uint32_t flag = 0;
switch (request) {
case VOCTRL_RESUME:
if (!noprebuf) {
ioval = EM8300_PLAYMODE_PLAY;
if (ioctl(fd_control, EM8300_IOCTL_SET_PLAYMODE, &ioval) < 0) {
printf("VO: [dxr3] Unable to set playmode!\n");
}
}
return VO_TRUE;
case VOCTRL_PAUSE:
if (!noprebuf) {
ioval = EM8300_PLAYMODE_PAUSED;
if (ioctl(fd_control, EM8300_IOCTL_SET_PLAYMODE, &ioval) < 0) {
printf("VO: [dxr3] Unable to set playmode!\n");
}
}
return VO_TRUE;
case VOCTRL_RESET:
if (!noprebuf) {
close(fd_video);
fd_video = open(fdv_name, O_WRONLY);
fsync(fd_video);
}
return VO_TRUE;
case VOCTRL_QUERY_FORMAT:
switch (*((uint32_t*)data)) {
case IMGFMT_MPEGPES:
/* Hardware accelerated | Hardware supports subpics */
flag = 0x2 | 0x8;
break;
#ifdef USE_LIBAVCODEC
case IMGFMT_YV12:
case IMGFMT_YUY2:
case IMGFMT_RGB24:
case IMGFMT_BGR24:
/* Conversion needed | OSD Supported */
flag = 0x1 | 0x4;
break;
default:
printf("VO: [dxr3] Format unsupported, mail dholm@iname.com\n");
#else
default:
printf("VO: [dxr3] You have disabled libavcodec support (Read DOCS/codecs.html)!\n");
#endif
}
if (noprebuf) {
return flag;
} else {
return (flag | 0x100);
}
}
return VO_NOTIMPL;
}
static uint32_t config(uint32_t width, uint32_t height, uint32_t d_width, uint32_t d_height, uint32_t fullscreen, char *title, uint32_t format,const vo_tune_info_t *info)
{
int tmp1, tmp2, size;
Added new syncengine thanks to a new previously undocumented feature of the em8300, this might fix playback on both slow and fast machines (more testing needed). This also requires users to get the em8300 driver from cvs until the next version is released (will probably happen this weekend) Added lots of comments, should be pretty easy to understand most of the internals now Added lots of brackets to if's for's while's etc, this is not a cosmetical thing but rather due to the fact I got some very odd bugs with else's since I didn't properly use brackets (and it's the K&R standard to have brackets everywhere) Fixed some bugs that would occur when disabling libmp1e Switched to default to the new naming scheme of device nodes, the driver will slowly switch over to this state, if it can't find devices under the new name it will try the old naming scheme I stopped opening devices in non-blocking mode, it would break the new syncengine which tries to burst data to the device (alot of times meaning it will fill the fifo pretty fast which would previously result in jerkyness on fast machines) The device now sets the initial state of the pts and speed (probably not needed, but assumption is the mother of all fuckups =) Keep the control interface open during the entire duration of the libvo device, we might need this to flush video buffers on seeking (currently not implemented, therefore seeking is broken) This is beta stuff to the driver, I will get some users to test it for me and do my best to fix seeking as soon as possible... git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@4219 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-01-17 10:33:47 +00:00
em8300_register_t reg;
/* Softzoom turned on, downscale */
Added new syncengine thanks to a new previously undocumented feature of the em8300, this might fix playback on both slow and fast machines (more testing needed). This also requires users to get the em8300 driver from cvs until the next version is released (will probably happen this weekend) Added lots of comments, should be pretty easy to understand most of the internals now Added lots of brackets to if's for's while's etc, this is not a cosmetical thing but rather due to the fact I got some very odd bugs with else's since I didn't properly use brackets (and it's the K&R standard to have brackets everywhere) Fixed some bugs that would occur when disabling libmp1e Switched to default to the new naming scheme of device nodes, the driver will slowly switch over to this state, if it can't find devices under the new name it will try the old naming scheme I stopped opening devices in non-blocking mode, it would break the new syncengine which tries to burst data to the device (alot of times meaning it will fill the fifo pretty fast which would previously result in jerkyness on fast machines) The device now sets the initial state of the pts and speed (probably not needed, but assumption is the mother of all fuckups =) Keep the control interface open during the entire duration of the libvo device, we might need this to flush video buffers on seeking (currently not implemented, therefore seeking is broken) This is beta stuff to the driver, I will get some users to test it for me and do my best to fix seeking as soon as possible... git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@4219 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-01-17 10:33:47 +00:00
/* This activates the subpicture processor, you can safely disable this and still send */
/* broken subpics to the em8300, if it's enabled and you send broken subpics you will end */
/* up in a lockup */
ioval = EM8300_SPUMODE_ON;
if (ioctl(fd_control, EM8300_IOCTL_SET_SPUMODE, &ioval) < 0) {
printf("VO: [dxr3] Unable to set subpicture mode!\n");
uninit();
return -1;
}
/* Set the playmode to play (just in case another app has set it to something else) */
ioval = EM8300_PLAYMODE_PLAY;
Added new syncengine thanks to a new previously undocumented feature of the em8300, this might fix playback on both slow and fast machines (more testing needed). This also requires users to get the em8300 driver from cvs until the next version is released (will probably happen this weekend) Added lots of comments, should be pretty easy to understand most of the internals now Added lots of brackets to if's for's while's etc, this is not a cosmetical thing but rather due to the fact I got some very odd bugs with else's since I didn't properly use brackets (and it's the K&R standard to have brackets everywhere) Fixed some bugs that would occur when disabling libmp1e Switched to default to the new naming scheme of device nodes, the driver will slowly switch over to this state, if it can't find devices under the new name it will try the old naming scheme I stopped opening devices in non-blocking mode, it would break the new syncengine which tries to burst data to the device (alot of times meaning it will fill the fifo pretty fast which would previously result in jerkyness on fast machines) The device now sets the initial state of the pts and speed (probably not needed, but assumption is the mother of all fuckups =) Keep the control interface open during the entire duration of the libvo device, we might need this to flush video buffers on seeking (currently not implemented, therefore seeking is broken) This is beta stuff to the driver, I will get some users to test it for me and do my best to fix seeking as soon as possible... git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@4219 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-01-17 10:33:47 +00:00
if (ioctl(fd_control, EM8300_IOCTL_SET_PLAYMODE, &ioval) < 0) {
printf("VO: [dxr3] Unable to set playmode!\n");
Added new syncengine thanks to a new previously undocumented feature of the em8300, this might fix playback on both slow and fast machines (more testing needed). This also requires users to get the em8300 driver from cvs until the next version is released (will probably happen this weekend) Added lots of comments, should be pretty easy to understand most of the internals now Added lots of brackets to if's for's while's etc, this is not a cosmetical thing but rather due to the fact I got some very odd bugs with else's since I didn't properly use brackets (and it's the K&R standard to have brackets everywhere) Fixed some bugs that would occur when disabling libmp1e Switched to default to the new naming scheme of device nodes, the driver will slowly switch over to this state, if it can't find devices under the new name it will try the old naming scheme I stopped opening devices in non-blocking mode, it would break the new syncengine which tries to burst data to the device (alot of times meaning it will fill the fifo pretty fast which would previously result in jerkyness on fast machines) The device now sets the initial state of the pts and speed (probably not needed, but assumption is the mother of all fuckups =) Keep the control interface open during the entire duration of the libvo device, we might need this to flush video buffers on seeking (currently not implemented, therefore seeking is broken) This is beta stuff to the driver, I will get some users to test it for me and do my best to fix seeking as soon as possible... git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@4219 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-01-17 10:33:47 +00:00
}
/* Start em8300 prebuffering and sync engine */
reg.microcode_register = 1;
reg.reg = 0;
reg.val = MVCOMMAND_SYNC;
ioctl(fd_control, EM8300_IOCTL_WRITEREG, &reg);
/* Clean buffer by syncing it */
ioval = EM8300_SUBDEVICE_VIDEO;
ioctl(fd_control, EM8300_IOCTL_FLUSH, &ioval);
ioval = EM8300_SUBDEVICE_AUDIO;
ioctl(fd_control, EM8300_IOCTL_FLUSH, &ioval);
fsync(fd_video);
if (!noprebuf) {
ioval = 0x900;
ioctl(fd_control, EM8300_IOCTL_SCR_SETSPEED, &ioval);
ioval = 0;
ioctl(fd_control, EM8300_IOCTL_SCR_SET, &ioval);
}
Added new syncengine thanks to a new previously undocumented feature of the em8300, this might fix playback on both slow and fast machines (more testing needed). This also requires users to get the em8300 driver from cvs until the next version is released (will probably happen this weekend) Added lots of comments, should be pretty easy to understand most of the internals now Added lots of brackets to if's for's while's etc, this is not a cosmetical thing but rather due to the fact I got some very odd bugs with else's since I didn't properly use brackets (and it's the K&R standard to have brackets everywhere) Fixed some bugs that would occur when disabling libmp1e Switched to default to the new naming scheme of device nodes, the driver will slowly switch over to this state, if it can't find devices under the new name it will try the old naming scheme I stopped opening devices in non-blocking mode, it would break the new syncengine which tries to burst data to the device (alot of times meaning it will fill the fifo pretty fast which would previously result in jerkyness on fast machines) The device now sets the initial state of the pts and speed (probably not needed, but assumption is the mother of all fuckups =) Keep the control interface open during the entire duration of the libvo device, we might need this to flush video buffers on seeking (currently not implemented, therefore seeking is broken) This is beta stuff to the driver, I will get some users to test it for me and do my best to fix seeking as soon as possible... git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@4219 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-01-17 10:33:47 +00:00
/* Store some variables statically that we need later in another scope */
img_format = format;
v_width = width;
v_height = height;
/* libavcodec requires a width and height that is x|16 */
aspect_save_orig(width, height);
aspect_save_prescale(d_width, d_height);
ioctl(fd_control, EM8300_IOCTL_GET_VIDEOMODE, &ioval);
if (ioval == EM8300_VIDEOMODE_NTSC) {
printf("VO: [dxr3] Setting up for NTSC.\n");
aspect_save_screenres(352, 240);
} else {
printf("VO: [dxr3] Setting up for PAL/SECAM.\n");
aspect_save_screenres(352, 288);
}
aspect(&s_width, &s_height, A_ZOOM);
/* Try to figure out whether to use widescreen output or not */
Added new syncengine thanks to a new previously undocumented feature of the em8300, this might fix playback on both slow and fast machines (more testing needed). This also requires users to get the em8300 driver from cvs until the next version is released (will probably happen this weekend) Added lots of comments, should be pretty easy to understand most of the internals now Added lots of brackets to if's for's while's etc, this is not a cosmetical thing but rather due to the fact I got some very odd bugs with else's since I didn't properly use brackets (and it's the K&R standard to have brackets everywhere) Fixed some bugs that would occur when disabling libmp1e Switched to default to the new naming scheme of device nodes, the driver will slowly switch over to this state, if it can't find devices under the new name it will try the old naming scheme I stopped opening devices in non-blocking mode, it would break the new syncengine which tries to burst data to the device (alot of times meaning it will fill the fifo pretty fast which would previously result in jerkyness on fast machines) The device now sets the initial state of the pts and speed (probably not needed, but assumption is the mother of all fuckups =) Keep the control interface open during the entire duration of the libvo device, we might need this to flush video buffers on seeking (currently not implemented, therefore seeking is broken) This is beta stuff to the driver, I will get some users to test it for me and do my best to fix seeking as soon as possible... git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@4219 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-01-17 10:33:47 +00:00
/* Anamorphic widescreen modes makes this a pain in the ass */
tmp1 = abs(d_height - ((d_width / 4) * 3));
tmp2 = abs(d_height - (int) (d_width / 2.35));
if (tmp1 < tmp2) {
ioval = EM8300_ASPECTRATIO_4_3;
printf("VO: [dxr3] Setting aspect ratio to 4:3\n");
} else {
ioval = EM8300_ASPECTRATIO_16_9;
printf("VO: [dxr3] Setting aspect ratio to 16:9\n");
}
ioctl(fd_control, EM8300_IOCTL_SET_ASPECTRATIO, &ioval);
if (format != IMGFMT_MPEGPES) {
#ifdef USE_LIBAVCODEC
avc_codec = avcodec_find_encoder(CODEC_ID_MPEG1VIDEO);
if (!avc_codec) {
printf("VO: [dxr3] Unable to find mpeg1video codec\n");
uninit();
return -1;
}
avc_context = malloc(sizeof(AVCodecContext));
memset(avc_context, 0, sizeof(avc_context));
avc_context->width = s_width;
avc_context->height = s_height;
ioctl(fd_control, EM8300_IOCTL_GET_VIDEOMODE, &ioval);
if (ioval == EM8300_VIDEOMODE_NTSC) {
avc_context->gop_size = 18;
} else {
avc_context->gop_size = 15;
}
avc_context->frame_rate = vo_fps * FRAME_RATE_BASE;
avc_context->bit_rate = 8e6;
avc_context->flags = CODEC_FLAG_HQ | CODEC_FLAG_QSCALE;
avc_context->quality = 2;
avc_context->pix_fmt = PIX_FMT_YUV420P;
if (avcodec_open(avc_context, avc_codec) < 0) {
printf("VO: [dxr3] Unable to open codec\n");
uninit();
return -1;
}
sws = getSwsContextFromCmdLine(v_width, v_height, img_format, s_width, s_height, IMGFMT_YV12);
if (!sws)
{
printf("vo_vesa: Can't initialize SwScaler\n");
return -1;
}
/* This stuff calculations the relative position of video and osd on screen */
Added new syncengine thanks to a new previously undocumented feature of the em8300, this might fix playback on both slow and fast machines (more testing needed). This also requires users to get the em8300 driver from cvs until the next version is released (will probably happen this weekend) Added lots of comments, should be pretty easy to understand most of the internals now Added lots of brackets to if's for's while's etc, this is not a cosmetical thing but rather due to the fact I got some very odd bugs with else's since I didn't properly use brackets (and it's the K&R standard to have brackets everywhere) Fixed some bugs that would occur when disabling libmp1e Switched to default to the new naming scheme of device nodes, the driver will slowly switch over to this state, if it can't find devices under the new name it will try the old naming scheme I stopped opening devices in non-blocking mode, it would break the new syncengine which tries to burst data to the device (alot of times meaning it will fill the fifo pretty fast which would previously result in jerkyness on fast machines) The device now sets the initial state of the pts and speed (probably not needed, but assumption is the mother of all fuckups =) Keep the control interface open during the entire duration of the libvo device, we might need this to flush video buffers on seeking (currently not implemented, therefore seeking is broken) This is beta stuff to the driver, I will get some users to test it for me and do my best to fix seeking as soon as possible... git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@4219 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-01-17 10:33:47 +00:00
/* Old stuff taken from the dvb driver, should be removed when introducing spuenc */
osd_w = s_width;
d_pos_x = (s_width - v_width) / 2;
if (d_pos_x < 0) {
s_pos_x = -d_pos_x;
d_pos_x = 0;
osd_w = s_width;
Added new syncengine thanks to a new previously undocumented feature of the em8300, this might fix playback on both slow and fast machines (more testing needed). This also requires users to get the em8300 driver from cvs until the next version is released (will probably happen this weekend) Added lots of comments, should be pretty easy to understand most of the internals now Added lots of brackets to if's for's while's etc, this is not a cosmetical thing but rather due to the fact I got some very odd bugs with else's since I didn't properly use brackets (and it's the K&R standard to have brackets everywhere) Fixed some bugs that would occur when disabling libmp1e Switched to default to the new naming scheme of device nodes, the driver will slowly switch over to this state, if it can't find devices under the new name it will try the old naming scheme I stopped opening devices in non-blocking mode, it would break the new syncengine which tries to burst data to the device (alot of times meaning it will fill the fifo pretty fast which would previously result in jerkyness on fast machines) The device now sets the initial state of the pts and speed (probably not needed, but assumption is the mother of all fuckups =) Keep the control interface open during the entire duration of the libvo device, we might need this to flush video buffers on seeking (currently not implemented, therefore seeking is broken) This is beta stuff to the driver, I will get some users to test it for me and do my best to fix seeking as soon as possible... git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@4219 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-01-17 10:33:47 +00:00
} else {
s_pos_x = 0;
Added new syncengine thanks to a new previously undocumented feature of the em8300, this might fix playback on both slow and fast machines (more testing needed). This also requires users to get the em8300 driver from cvs until the next version is released (will probably happen this weekend) Added lots of comments, should be pretty easy to understand most of the internals now Added lots of brackets to if's for's while's etc, this is not a cosmetical thing but rather due to the fact I got some very odd bugs with else's since I didn't properly use brackets (and it's the K&R standard to have brackets everywhere) Fixed some bugs that would occur when disabling libmp1e Switched to default to the new naming scheme of device nodes, the driver will slowly switch over to this state, if it can't find devices under the new name it will try the old naming scheme I stopped opening devices in non-blocking mode, it would break the new syncengine which tries to burst data to the device (alot of times meaning it will fill the fifo pretty fast which would previously result in jerkyness on fast machines) The device now sets the initial state of the pts and speed (probably not needed, but assumption is the mother of all fuckups =) Keep the control interface open during the entire duration of the libvo device, we might need this to flush video buffers on seeking (currently not implemented, therefore seeking is broken) This is beta stuff to the driver, I will get some users to test it for me and do my best to fix seeking as soon as possible... git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@4219 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-01-17 10:33:47 +00:00
}
osd_h = s_height;
d_pos_y = (s_height - v_width) / 2;
if (d_pos_y < 0) {
Added new syncengine thanks to a new previously undocumented feature of the em8300, this might fix playback on both slow and fast machines (more testing needed). This also requires users to get the em8300 driver from cvs until the next version is released (will probably happen this weekend) Added lots of comments, should be pretty easy to understand most of the internals now Added lots of brackets to if's for's while's etc, this is not a cosmetical thing but rather due to the fact I got some very odd bugs with else's since I didn't properly use brackets (and it's the K&R standard to have brackets everywhere) Fixed some bugs that would occur when disabling libmp1e Switched to default to the new naming scheme of device nodes, the driver will slowly switch over to this state, if it can't find devices under the new name it will try the old naming scheme I stopped opening devices in non-blocking mode, it would break the new syncengine which tries to burst data to the device (alot of times meaning it will fill the fifo pretty fast which would previously result in jerkyness on fast machines) The device now sets the initial state of the pts and speed (probably not needed, but assumption is the mother of all fuckups =) Keep the control interface open during the entire duration of the libvo device, we might need this to flush video buffers on seeking (currently not implemented, therefore seeking is broken) This is beta stuff to the driver, I will get some users to test it for me and do my best to fix seeking as soon as possible... git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@4219 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-01-17 10:33:47 +00:00
s_pos_y = -d_pos_y;
d_pos_y = 0;
osd_h = s_height;
Added new syncengine thanks to a new previously undocumented feature of the em8300, this might fix playback on both slow and fast machines (more testing needed). This also requires users to get the em8300 driver from cvs until the next version is released (will probably happen this weekend) Added lots of comments, should be pretty easy to understand most of the internals now Added lots of brackets to if's for's while's etc, this is not a cosmetical thing but rather due to the fact I got some very odd bugs with else's since I didn't properly use brackets (and it's the K&R standard to have brackets everywhere) Fixed some bugs that would occur when disabling libmp1e Switched to default to the new naming scheme of device nodes, the driver will slowly switch over to this state, if it can't find devices under the new name it will try the old naming scheme I stopped opening devices in non-blocking mode, it would break the new syncengine which tries to burst data to the device (alot of times meaning it will fill the fifo pretty fast which would previously result in jerkyness on fast machines) The device now sets the initial state of the pts and speed (probably not needed, but assumption is the mother of all fuckups =) Keep the control interface open during the entire duration of the libvo device, we might need this to flush video buffers on seeking (currently not implemented, therefore seeking is broken) This is beta stuff to the driver, I will get some users to test it for me and do my best to fix seeking as soon as possible... git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@4219 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-01-17 10:33:47 +00:00
} else {
s_pos_y = 0;
Added new syncengine thanks to a new previously undocumented feature of the em8300, this might fix playback on both slow and fast machines (more testing needed). This also requires users to get the em8300 driver from cvs until the next version is released (will probably happen this weekend) Added lots of comments, should be pretty easy to understand most of the internals now Added lots of brackets to if's for's while's etc, this is not a cosmetical thing but rather due to the fact I got some very odd bugs with else's since I didn't properly use brackets (and it's the K&R standard to have brackets everywhere) Fixed some bugs that would occur when disabling libmp1e Switched to default to the new naming scheme of device nodes, the driver will slowly switch over to this state, if it can't find devices under the new name it will try the old naming scheme I stopped opening devices in non-blocking mode, it would break the new syncengine which tries to burst data to the device (alot of times meaning it will fill the fifo pretty fast which would previously result in jerkyness on fast machines) The device now sets the initial state of the pts and speed (probably not needed, but assumption is the mother of all fuckups =) Keep the control interface open during the entire duration of the libvo device, we might need this to flush video buffers on seeking (currently not implemented, therefore seeking is broken) This is beta stuff to the driver, I will get some users to test it for me and do my best to fix seeking as soon as possible... git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@4219 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-01-17 10:33:47 +00:00
}
/* Create a pixel buffer and set up pointers for color components */
memset(&avc_picture, 0, sizeof(avc_picture));
avc_picture.linesize[0] = s_width;
avc_picture.linesize[1] = s_width / 2;
avc_picture.linesize[2] = s_width / 2;
size = s_width * s_height;
picture_buf = malloc((size * 3) / 2);
avc_picture.data[0] = picture_buf;
avc_picture.data[1] = avc_picture.data[0] + size;
avc_picture.data[2] = avc_picture.data[1] + size / 4;
Added new syncengine thanks to a new previously undocumented feature of the em8300, this might fix playback on both slow and fast machines (more testing needed). This also requires users to get the em8300 driver from cvs until the next version is released (will probably happen this weekend) Added lots of comments, should be pretty easy to understand most of the internals now Added lots of brackets to if's for's while's etc, this is not a cosmetical thing but rather due to the fact I got some very odd bugs with else's since I didn't properly use brackets (and it's the K&R standard to have brackets everywhere) Fixed some bugs that would occur when disabling libmp1e Switched to default to the new naming scheme of device nodes, the driver will slowly switch over to this state, if it can't find devices under the new name it will try the old naming scheme I stopped opening devices in non-blocking mode, it would break the new syncengine which tries to burst data to the device (alot of times meaning it will fill the fifo pretty fast which would previously result in jerkyness on fast machines) The device now sets the initial state of the pts and speed (probably not needed, but assumption is the mother of all fuckups =) Keep the control interface open during the entire duration of the libvo device, we might need this to flush video buffers on seeking (currently not implemented, therefore seeking is broken) This is beta stuff to the driver, I will get some users to test it for me and do my best to fix seeking as soon as possible... git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@4219 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-01-17 10:33:47 +00:00
if (format == IMGFMT_BGR24) {
yuv2rgb_init(24, MODE_BGR);
} else {
yuv2rgb_init(24, MODE_RGB);
Added new syncengine thanks to a new previously undocumented feature of the em8300, this might fix playback on both slow and fast machines (more testing needed). This also requires users to get the em8300 driver from cvs until the next version is released (will probably happen this weekend) Added lots of comments, should be pretty easy to understand most of the internals now Added lots of brackets to if's for's while's etc, this is not a cosmetical thing but rather due to the fact I got some very odd bugs with else's since I didn't properly use brackets (and it's the K&R standard to have brackets everywhere) Fixed some bugs that would occur when disabling libmp1e Switched to default to the new naming scheme of device nodes, the driver will slowly switch over to this state, if it can't find devices under the new name it will try the old naming scheme I stopped opening devices in non-blocking mode, it would break the new syncengine which tries to burst data to the device (alot of times meaning it will fill the fifo pretty fast which would previously result in jerkyness on fast machines) The device now sets the initial state of the pts and speed (probably not needed, but assumption is the mother of all fuckups =) Keep the control interface open during the entire duration of the libvo device, we might need this to flush video buffers on seeking (currently not implemented, therefore seeking is broken) This is beta stuff to the driver, I will get some users to test it for me and do my best to fix seeking as soon as possible... git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@4219 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-01-17 10:33:47 +00:00
}
return 0;
#endif
return -1;
} else if (format == IMGFMT_MPEGPES) {
printf("VO: [dxr3] Format: MPEG-PES (no conversion needed)\n");
return 0;
}
printf("VO: [dxr3] Format: Unsupported\n");
uninit();
return -1;
}
static const vo_info_t* get_info(void)
{
return &vo_info;
}
static void draw_alpha(int x, int y, int w, int h, unsigned char* src, unsigned char *srca, int srcstride)
{
#ifdef USE_LIBAVCODEC
vo_draw_alpha_yv12(w, h, src, srca, srcstride,
avc_picture.data[0] + (x + d_pos_x) + (y + d_pos_y) * avc_picture.linesize[0], avc_picture.linesize[0]);
Added new syncengine thanks to a new previously undocumented feature of the em8300, this might fix playback on both slow and fast machines (more testing needed). This also requires users to get the em8300 driver from cvs until the next version is released (will probably happen this weekend) Added lots of comments, should be pretty easy to understand most of the internals now Added lots of brackets to if's for's while's etc, this is not a cosmetical thing but rather due to the fact I got some very odd bugs with else's since I didn't properly use brackets (and it's the K&R standard to have brackets everywhere) Fixed some bugs that would occur when disabling libmp1e Switched to default to the new naming scheme of device nodes, the driver will slowly switch over to this state, if it can't find devices under the new name it will try the old naming scheme I stopped opening devices in non-blocking mode, it would break the new syncengine which tries to burst data to the device (alot of times meaning it will fill the fifo pretty fast which would previously result in jerkyness on fast machines) The device now sets the initial state of the pts and speed (probably not needed, but assumption is the mother of all fuckups =) Keep the control interface open during the entire duration of the libvo device, we might need this to flush video buffers on seeking (currently not implemented, therefore seeking is broken) This is beta stuff to the driver, I will get some users to test it for me and do my best to fix seeking as soon as possible... git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@4219 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-01-17 10:33:47 +00:00
#endif
}
static void draw_osd(void)
{
Added new syncengine thanks to a new previously undocumented feature of the em8300, this might fix playback on both slow and fast machines (more testing needed). This also requires users to get the em8300 driver from cvs until the next version is released (will probably happen this weekend) Added lots of comments, should be pretty easy to understand most of the internals now Added lots of brackets to if's for's while's etc, this is not a cosmetical thing but rather due to the fact I got some very odd bugs with else's since I didn't properly use brackets (and it's the K&R standard to have brackets everywhere) Fixed some bugs that would occur when disabling libmp1e Switched to default to the new naming scheme of device nodes, the driver will slowly switch over to this state, if it can't find devices under the new name it will try the old naming scheme I stopped opening devices in non-blocking mode, it would break the new syncengine which tries to burst data to the device (alot of times meaning it will fill the fifo pretty fast which would previously result in jerkyness on fast machines) The device now sets the initial state of the pts and speed (probably not needed, but assumption is the mother of all fuckups =) Keep the control interface open during the entire duration of the libvo device, we might need this to flush video buffers on seeking (currently not implemented, therefore seeking is broken) This is beta stuff to the driver, I will get some users to test it for me and do my best to fix seeking as soon as possible... git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@4219 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-01-17 10:33:47 +00:00
if (img_format != IMGFMT_MPEGPES) {
vo_draw_text(osd_w, osd_h, draw_alpha);
}
}
static uint32_t draw_frame(uint8_t * src[])
{
if (img_format == IMGFMT_MPEGPES) {
vo_mpegpes_t *p = (vo_mpegpes_t *) src[0];
if (p->id == 0x20) {
if (!noprebuf) {
ioctl(fd_spu, EM8300_IOCTL_SPU_SETPTS, &vo_pts);
}
write(fd_spu, p->data, p->size);
} else {
write(fd_video, p->data, p->size);
}
return 0;
#ifdef USE_LIBAVCODEC
} else {
int size, srcStride = (img_format == IMGFMT_YUY2) ? (v_width * 2) : (v_width * 3);
sws->swScale(sws, src, &srcStride, 0, v_height, avc_picture.data, avc_picture.linesize);
draw_osd();
size = avcodec_encode_video(avc_context, picture_buf, avc_outbuf_size, &avc_picture);
write(fd_video, picture_buf, size);
return 0;
#endif
}
return -1;
}
static void flip_page(void)
{
if (!noprebuf) {
ioctl(fd_video, EM8300_IOCTL_VIDEO_SETPTS, &vo_pts);
}
#ifdef USE_LIBAVCODEC
if (img_format == IMGFMT_YV12) {
int size = avcodec_encode_video(avc_context, picture_buf, avc_outbuf_size, &avc_picture);
if (!noprebuf) {
ioctl(fd_video, EM8300_IOCTL_VIDEO_SETPTS, &vo_pts);
}
write(fd_video, picture_buf, size);
}
#endif
}
static uint32_t draw_slice(uint8_t *srcimg[], int stride[], int w, int h, int x0, int y0)
{
#ifdef USE_LIBAVCODEC
if (img_format == IMGFMT_YV12) {
sws->swScale(sws, srcimg, stride, y0, h, avc_picture.data, avc_picture.linesize);
return 0;
}
Added new syncengine thanks to a new previously undocumented feature of the em8300, this might fix playback on both slow and fast machines (more testing needed). This also requires users to get the em8300 driver from cvs until the next version is released (will probably happen this weekend) Added lots of comments, should be pretty easy to understand most of the internals now Added lots of brackets to if's for's while's etc, this is not a cosmetical thing but rather due to the fact I got some very odd bugs with else's since I didn't properly use brackets (and it's the K&R standard to have brackets everywhere) Fixed some bugs that would occur when disabling libmp1e Switched to default to the new naming scheme of device nodes, the driver will slowly switch over to this state, if it can't find devices under the new name it will try the old naming scheme I stopped opening devices in non-blocking mode, it would break the new syncengine which tries to burst data to the device (alot of times meaning it will fill the fifo pretty fast which would previously result in jerkyness on fast machines) The device now sets the initial state of the pts and speed (probably not needed, but assumption is the mother of all fuckups =) Keep the control interface open during the entire duration of the libvo device, we might need this to flush video buffers on seeking (currently not implemented, therefore seeking is broken) This is beta stuff to the driver, I will get some users to test it for me and do my best to fix seeking as soon as possible... git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@4219 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-01-17 10:33:47 +00:00
#endif
return -1;
}
static void uninit(void)
{
printf("VO: [dxr3] Uninitializing\n");
if (sws) {
freeSwsContext(sws);
}
#ifdef USE_LIBAVCODEC
if (avc_context) {
avcodec_close(avc_context);
}
if (picture_buf) {
free(picture_buf);
Added new syncengine thanks to a new previously undocumented feature of the em8300, this might fix playback on both slow and fast machines (more testing needed). This also requires users to get the em8300 driver from cvs until the next version is released (will probably happen this weekend) Added lots of comments, should be pretty easy to understand most of the internals now Added lots of brackets to if's for's while's etc, this is not a cosmetical thing but rather due to the fact I got some very odd bugs with else's since I didn't properly use brackets (and it's the K&R standard to have brackets everywhere) Fixed some bugs that would occur when disabling libmp1e Switched to default to the new naming scheme of device nodes, the driver will slowly switch over to this state, if it can't find devices under the new name it will try the old naming scheme I stopped opening devices in non-blocking mode, it would break the new syncengine which tries to burst data to the device (alot of times meaning it will fill the fifo pretty fast which would previously result in jerkyness on fast machines) The device now sets the initial state of the pts and speed (probably not needed, but assumption is the mother of all fuckups =) Keep the control interface open during the entire duration of the libvo device, we might need this to flush video buffers on seeking (currently not implemented, therefore seeking is broken) This is beta stuff to the driver, I will get some users to test it for me and do my best to fix seeking as soon as possible... git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@4219 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-01-17 10:33:47 +00:00
}
#endif
Added new syncengine thanks to a new previously undocumented feature of the em8300, this might fix playback on both slow and fast machines (more testing needed). This also requires users to get the em8300 driver from cvs until the next version is released (will probably happen this weekend) Added lots of comments, should be pretty easy to understand most of the internals now Added lots of brackets to if's for's while's etc, this is not a cosmetical thing but rather due to the fact I got some very odd bugs with else's since I didn't properly use brackets (and it's the K&R standard to have brackets everywhere) Fixed some bugs that would occur when disabling libmp1e Switched to default to the new naming scheme of device nodes, the driver will slowly switch over to this state, if it can't find devices under the new name it will try the old naming scheme I stopped opening devices in non-blocking mode, it would break the new syncengine which tries to burst data to the device (alot of times meaning it will fill the fifo pretty fast which would previously result in jerkyness on fast machines) The device now sets the initial state of the pts and speed (probably not needed, but assumption is the mother of all fuckups =) Keep the control interface open during the entire duration of the libvo device, we might need this to flush video buffers on seeking (currently not implemented, therefore seeking is broken) This is beta stuff to the driver, I will get some users to test it for me and do my best to fix seeking as soon as possible... git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@4219 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-01-17 10:33:47 +00:00
if (fd_video) {
close(fd_video);
Added new syncengine thanks to a new previously undocumented feature of the em8300, this might fix playback on both slow and fast machines (more testing needed). This also requires users to get the em8300 driver from cvs until the next version is released (will probably happen this weekend) Added lots of comments, should be pretty easy to understand most of the internals now Added lots of brackets to if's for's while's etc, this is not a cosmetical thing but rather due to the fact I got some very odd bugs with else's since I didn't properly use brackets (and it's the K&R standard to have brackets everywhere) Fixed some bugs that would occur when disabling libmp1e Switched to default to the new naming scheme of device nodes, the driver will slowly switch over to this state, if it can't find devices under the new name it will try the old naming scheme I stopped opening devices in non-blocking mode, it would break the new syncengine which tries to burst data to the device (alot of times meaning it will fill the fifo pretty fast which would previously result in jerkyness on fast machines) The device now sets the initial state of the pts and speed (probably not needed, but assumption is the mother of all fuckups =) Keep the control interface open during the entire duration of the libvo device, we might need this to flush video buffers on seeking (currently not implemented, therefore seeking is broken) This is beta stuff to the driver, I will get some users to test it for me and do my best to fix seeking as soon as possible... git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@4219 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-01-17 10:33:47 +00:00
}
if (fd_spu) {
close(fd_spu);
Added new syncengine thanks to a new previously undocumented feature of the em8300, this might fix playback on both slow and fast machines (more testing needed). This also requires users to get the em8300 driver from cvs until the next version is released (will probably happen this weekend) Added lots of comments, should be pretty easy to understand most of the internals now Added lots of brackets to if's for's while's etc, this is not a cosmetical thing but rather due to the fact I got some very odd bugs with else's since I didn't properly use brackets (and it's the K&R standard to have brackets everywhere) Fixed some bugs that would occur when disabling libmp1e Switched to default to the new naming scheme of device nodes, the driver will slowly switch over to this state, if it can't find devices under the new name it will try the old naming scheme I stopped opening devices in non-blocking mode, it would break the new syncengine which tries to burst data to the device (alot of times meaning it will fill the fifo pretty fast which would previously result in jerkyness on fast machines) The device now sets the initial state of the pts and speed (probably not needed, but assumption is the mother of all fuckups =) Keep the control interface open during the entire duration of the libvo device, we might need this to flush video buffers on seeking (currently not implemented, therefore seeking is broken) This is beta stuff to the driver, I will get some users to test it for me and do my best to fix seeking as soon as possible... git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@4219 b3059339-0415-0410-9bf9-f77b7e298cf2
2002-01-17 10:33:47 +00:00
}
if (fd_control) {
close(fd_control);
}
}
static void check_events(void)
{
}
static uint32_t preinit(const char *arg)
{
char devname[80];
int fdflags = O_WRONLY;
/* Open the control interface */
if (arg && !strcmp("noprebuf", arg)) {
printf("VO: [dxr3] Disabling prebuffering.\n");
noprebuf = 1;
fdflags |= O_NONBLOCK;
}
if (arg && !noprebuf) {
printf("VO: [dxr3] Forcing use of device %s\n", arg);
sprintf(devname, "/dev/em8300-%s", arg);
} else {
/* Try new naming scheme by default */
sprintf(devname, "/dev/em8300-0");
}
fd_control = open(devname, fdflags);
if (fd_control < 1) {
/* Fall back to old naming scheme */
printf("VO: [dxr3] Error opening %s for writing, trying /dev/em8300 instead\n", devname);
sprintf(devname, "/dev/em8300");
fd_control = open(devname, fdflags);
if (fd_control < 1) {
printf("VO: [dxr3] Error opening /dev/em8300 for writing as well!\nBailing\n");
return -1;
}
}
/* Open the video interface */
if (arg && !noprebuf) {
sprintf(devname, "/dev/em8300_mv-%s", arg);
} else {
/* Try new naming scheme by default */
sprintf(devname, "/dev/em8300_mv-0");
}
fd_video = open(devname, fdflags);
if (fd_video < 0) {
/* Fall back to old naming scheme */
printf("VO: [dxr3] Error opening %s for writing, trying /dev/em8300_mv instead\n", devname);
sprintf(devname, "/dev/em8300_mv");
fd_video = open(devname, fdflags);
if (fd_video < 0) {
printf("VO: [dxr3] Error opening /dev/em8300_mv for writing as well!\nBailing\n");
uninit();
return -1;
}
} else {
printf("VO: [dxr3] Opened %s\n", devname);
}
strcpy(fdv_name, devname);
/* Open the subpicture interface */
if (arg && !noprebuf) {
sprintf(devname, "/dev/em8300_sp-%s", arg);
} else {
/* Try new naming scheme by default */
sprintf(devname, "/dev/em8300_sp-0");
}
fd_spu = open(devname, fdflags);
if (fd_spu < 0) {
/* Fall back to old naming scheme */
printf("VO: [dxr3] Error opening %s for writing, trying /dev/em8300_sp instead\n", devname);
sprintf(devname, "/dev/em8300_sp");
fd_spu = open(devname, fdflags);
if (fd_spu < 0) {
printf("VO: [dxr3] Error opening /dev/em8300_sp for writing as well!\nBailing\n");
uninit();
return -1;
}
}
#ifdef USE_LIBAVCODEC
avcodec_init();
avcodec_register_all();
#endif
return 0;
}