gatos-devel Mailing List for GATOS

Oops, my apologies, Serge mentioned that in his mail.

Serge, do you have the symlink asm to asm-i386 in the include dir?
That's what I see with my compiled 2.4.17.

-- 
Michael Leuchtenburg - mi...@sl...
1984 is not an instruction manual

On Sun, 31 Mar 2002, Michael Leuchtenburg wrote:

> You need to create a symlink, /usr/src/linux, to your actual linux
> kernel source directory, or you need to tell km's compile process about
> the location of the linux kernel source (is that possible?)

I thought I said in my mail, there is a symlink to it.

Observe:
lrwxrwxrwx    1 root     root           17 Mar 30 21:35 linux ->
linux-2.4.18-0.4/

Also, as I pointed out, the places the code itself is looking is wrong
(asm rather than asm-i386)

- Serge Wroclawski

You need to create a symlink, /usr/src/linux, to your actual linux
kernel source directory, or you need to tell km's compile process about
the location of the linux kernel source (is that possible?)

-- 
Michael Leuchtenburg - mi...@sl...
1984 is not an instruction manual

Hello,

I know that a development mailing list is not the most appropriate place
to ask this question, but I do not see a discussion list or anything else
that seems more appropriate. If one exists, please flame me in the right
direction. :)

I have an ATI Radeon 7500 All In Wonder and I'm using the Red Hat Linux
7.3 beta operating system (becuase the card is supported in X 4.2).

I recompiled the kernel with what I thought were the appropriate drivers.
That is, I compiled in the bttv module and everything it needs. I'm
somewhat familiar with this module since I use it on my older
computer with a Haupage WinTV card.

Unfortunately, it does not seem to work. What I get is that the
videodev module insmods fine, but the bttv module itself does not,
complaining it can't find the card:

[root@localhost serge]# /sbin/insmod bttv
Using /lib/modules/2.4.18-0.4custom/kernel/drivers/media/video/bttv.o
/lib/modules/2.4.18-0.4custom/kernel/drivers/media/video/bttv.o:
init_module: No such device
Hint: insmod errors can be caused by incorrect module parameters,
including invalid IO or IRQ parameters

This is an AGP device, and I do not see it under lspci (is this a
problem). This is a list of what I *do* see:

[root@localhost serge]# /sbin/lspci
00:00.0 Host bridge: VIA Technologies, Inc. VT8367 [KT266]
00:01.0 PCI bridge: VIA Technologies, Inc. VT8367 [KT266 AGP]
00:09.0 Multimedia audio controller: Creative Labs SB Live! EMU10k1 (rev
07)
00:09.1 Input device controller: Creative Labs SB Live! (rev 07)
00:0b.0 Ethernet controller: National Semiconductor Corporation DP83815
(MacPhyter) Ethernet Controller
00:11.0 ISA bridge: VIA Technologies, Inc.: Unknown device 3147
00:11.1 IDE interface: VIA Technologies, Inc. Bus Master IDE (rev 06)
00:11.2 USB Controller: VIA Technologies, Inc. UHCI USB (rev 23)
00:11.3 USB Controller: VIA Technologies, Inc. UHCI USB (rev 23)
00:11.5 Multimedia audio controller: VIA Technologies, Inc. AC97 Audio
Controller (rev 40)
01:00.0 VGA compatible controller: ATI Technologies Inc Radeon 7500 QW

So, upon suggestion, I decided I'd get the km module and compile it.

As above states, I'm using a 2.4.18 kernel located in
/usr/src/inux-2.4.18-0.4 with a symlink from /usr/src/linux to that.

When I try to compile the km module, it complains:

gcc -m486 -O2 -Wall -Wstrict-prototypes -pipe -fno-strict-aliasing
-fno-common -DMODULE -D__KERNEL__ -DLINUX -I/usr/src/linux/include -c km.c
In file included from km.c:3:
/usr/include/linux/autoconf.h:1:2: #error Invalid kernel header included
in userspace
In file included from /usr/src/linux/include/linux/config.h:4,
                 from /usr/src/linux/include/linux/types.h:5,
                 from km.c:9:
/usr/include/linux/autoconf.h:1:2: #error Invalid kernel header included
in userspace
In file included from /usr/src/linux/include/linux/fs.h:19,
                 from /usr/src/linux/include/linux/capability.h:17,
                 from /usr/src/linux/include/linux/binfmts.h:5,
                 from /usr/src/linux/include/linux/sched.h:9,
                 from km.c:14:
/usr/src/linux/include/linux/dcache.h: In function `dget':
/usr/src/linux/include/linux/dcache.h:247: warning: implicit declaration
of function `BUG'
In file included from /usr/src/linux/include/linux/ext3_fs_i.h:19,
                 from /usr/src/linux/include/linux/fs.h:293,
                 from /usr/src/linux/include/linux/capability.h:17,
                 from /usr/src/linux/include/linux/binfmts.h:5,
                 from /usr/src/linux/include/linux/sched.h:9,
                 from km.c:14:
/usr/src/linux/include/linux/rwsem.h: In function `down_read':
/usr/src/linux/include/linux/rwsem.h:44: warning: implicit declaration of
function `__down_read'


(I've ommitted the rest of the errors and warnings).

It appears not to find the files it needs as it says so explicitly later:

In file included from km.c:17:
/usr/src/linux/include/linux/interrupt.h:45:25: asm/hardirq.h: No such
file or directory
/usr/src/linux/include/linux/interrupt.h:46:25: asm/softirq.h: No such
file or directory
In file included from km.c:17:


This files are present under the following directory
/usr/src/linux-2.4.18-0.4/include/asm-i386

This seems to be the standard place for the asm stuff, as it appears in a
fresh kernel from kernel.org.

Is this an artifact from older kernels? A bug in what I'm doing?
Can anyone offer me some guideing light?

I did not see this problem mentioned in the archives of this list (which I
searched a bit before asking).

Thank you in advance,

- Serge Wroclawski

On Sun, 31 Mar 2002, R C wrote:

> On Sat, Mar 30, 2002 at 11:57:38PM -0800, Happy Camper wrote:
> >
> > I just bought an ATI AIW 7500.  I am wondering if there is a V4L2 driver
> > for it.  The V4L2 main page says :
> > "Octavian Purdila is part of the GATOS group and he is working on a V4L2
> > driver  for ATI graphics cards".
>
> This is _old_.  That dates back to the first kernel module attempted
> (98? 99? somewhere around there). IIRC, they got I2C working for mach64,
> then it fizzled.  No capture, no interrupt support, again, IIRC.  This
> was during the time when you had to tell X you had 800k less videoram
> than you actually did.  See gatos, the application.
>

Hmmm I do not think it was v4l2 based. The main problem this was not
developed further is because the intent was to have the _complete_ driver
in the kernel module. This raises problems with sharing the device with X.

                      Vladimir Dergachev

>could I see all of the regs that rttest dumped from windows ?
regs for windows:
fld_tmpReg1 = 0
fld_tmpReg2 = 0
fld_tmpReg3 = 0
fld_LP_CONTRAST = 106
fld_LP_BRIGHTNESS = 0
fld_CP_HUE_CNTL = 0
fld_LUMA_FILTER = 1
fld_H_SCALE_RATIO = 66901
fld_H_SHARPNESS = 0
fld_V_SCALE_RATIO = 2048
fld_V_DEINTERLACE_ON = 0
fld_V_BYPSS = 0
fld_V_DITHER_ON = 1
fld_EVENF_OFFSET = 512
fld_ODDF_OFFSET = 1536
fld_INTERLACE_DETECTED = 1
fld_VS_LINE_COUNT = 434
fld_VS_DETECTED_LINES = 525
fld_VS_ITU656_VB = 0
fld_VBI_CC_DATA = 0
fld_VBI_CC_WT = 0
fld_VBI_CC_WT_ACK = 0
fld_VBI_CC_HOLD = 0
fld_VBI_DECODE_EN = 0
fld_VBI_CC_DTO_P = 6146
fld_VBI_20BIT_DTO_P = 5468
fld_VBI_CC_LEVEL = 63
fld_VBI_20BIT_LEVEL = 89
fld_VBI_CLK_RUNIN_GAIN = 271
fld_H_VBI_WIND_START = 101
fld_H_VBI_WIND_END = 871
fld_V_VBI_WIND_START = 11
fld_V_VBI_WIND_END = 36
fld_VBI_20BIT_DATA0 = 0
fld_VBI_20BIT_DATA1 = 0
fld_VBI_20BIT_WT = 0
fld_VBI_20BIT_WT_ACK = 0
fld_VBI_20BIT_HOLD = 0
fld_VBI_CAPTURE_ENABLE = 1
fld_VBI_EDS_DATA = 0
fld_VBI_EDS_WT = 0
fld_VBI_EDS_WT_ACK = 0
fld_VBI_EDS_HOLD = 0
fld_VBI_SCALING_RATIO = 65536
fld_VBI_ALIGNER_ENABLE = 0
fld_H_ACTIVE_START = 104
fld_H_ACTIVE_END = 838
fld_V_ACTIVE_START = 37
fld_V_ACTIVE_END = 518
fld_CH_HEIGHT = 147
fld_CH_KILL_LEVEL = 142
fld_CH_AGC_ERROR_LIM = 2
fld_CH_AGC_FILTER_EN = 0
fld_CH_AGC_LOOP_SPEED = 0
fld_HUE_ADJ = 0
fld_STANDARD_SEL = 1
fld_STANDARD_YC = 0
fld_ADC_PDWN = 0
fld_INPUT_SELECT = 0
fld_ADC_PREFLO = 0
fld_H_SYNC_PULSE_WIDTH = 69
fld_HS_GENLOCKED = 1
fld_HS_SYNC_IN_WIN = 1
fld_VIN_ASYNC_RST = 0
fld_DVS_ASYNC_RST = 0
fld_VIP_VENDOR_ID = 4098
fld_VIP_DEVICE_ID = 19796
fld_VIP_REVISION_ID = 3
fld_BLACK_INT_START = 49
fld_BLACK_INT_LENGTH = 15
fld_UV_INT_START = 59
fld_U_INT_LENGTH = 15
fld_V_INT_LENGTH = 15
fld_CRDR_ACTIVE_GAIN = 109
fld_CBDB_ACTIVE_GAIN = 153
fld_DVS_DIRECTION = 1
fld_DVS_VBI_CARD8_SWAP = 0
fld_DVS_CLK_SELECT = 0
fld_CONTINUOUS_STREAM = 0
fld_DVSOUT_CLK_DRV = 1
fld_DVSOUT_DATA_DRV = 1
fld_COMB_CNTL0 = 1116831888
fld_COMB_CNTL1 = 177
fld_COMB_CNTL2 = -1
fld_COMB_LENGTH = 118883209
fld_SYNCTIP_REF0 = 57
fld_SYNCTIP_REF1 = 41
fld_CLAMP_REF = 59
fld_AGC_PEAKWHITE = 255
fld_VBI_PEAKWHITE = 210
fld_WPA_THRESHOLD = 1030
fld_WPA_TRIGGER_LO = 179
fld_WPA_TRIGGER_HIGH = 539
fld_LOCKOUT_START = 518
fld_LOCKOUT_END = 33
fld_CH_DTO_INC = 4189610
fld_PLL_SGAIN = 1
fld_PLL_FGAIN = 2
fld_CR_BURST_GAIN = 109
fld_CB_BURST_GAIN = 154
fld_VERT_LOCKOUT_START = 519
fld_VERT_LOCKOUT_END = 14
fld_H_IN_WIND_START = 100
fld_V_IN_WIND_START = 37
fld_H_OUT_WIND_WIDTH = 720
fld_V_OUT_WIND_WIDTH = 480
fld_HS_LINE_TOTAL = 909
fld_MIN_PULSE_WIDTH = 33
fld_MAX_PULSE_WIDTH = 100
fld_WIN_CLOSE_LIMIT = 160
fld_WIN_OPEN_LIMIT = 439
fld_VSYNC_INT_TRIGGER = 682
fld_VSYNC_INT_HOLD = 23
fld_VIN_M0 = 11
fld_VIN_N0 = 70
fld_MNFLIP_EN = 0
fld_VIN_P = 6
fld_REG_CLK_SEL = 0
fld_VIN_M1 = 0
fld_VIN_N1 = 0
fld_VIN_DRIVER_SEL = 0
fld_VIN_MNFLIP_REQ = 0
fld_VIN_MNFLIP_DONE = 1
fld_TV_LOCK_TO_VIN = 0
fld_TV_P_FOR_WINCLK = 4
fld_VINRST = 0
fld_VIN_CLK_SEL = 1
fld_VS_FIELD_BLANK_START = 518
fld_VS_FIELD_BLANK_END = 10
fld_VS_FIELD_IDLOCATION = 261
fld_VS_FRAME_TOTAL = 535
fld_SYNC_TIP_START = 882
fld_SYNC_TIP_LENGTH = 15
fld_GAIN_FORCE_DATA = 0
fld_GAIN_FORCE_EN = 0
fld_I_CLAMP_SEL = 3
fld_I_AGC_SEL = 3
fld_EXT_CLAMP_CAP = 1
fld_EXT_AGC_CAP = 1
fld_DECI_DITHER_EN = 1
fld_ADC_PREFHI = 0
fld_ADC_CH_GAIN_SEL = 1
fld_HS_PLL_SGAIN = 2
fld_NREn = 0
fld_NRGainCntl = 0
fld_NRBWTresh = 0
fld_NRGCTresh = 0
fld_NRCoefDespeclMode = 0
fld_GPIO_5_OE = 0
fld_GPIO_6_OE = 0
fld_GPIO_5_OUT = 0
fld_GPIO_6_OUT = 0

and to complement here is a diff (no -u) I did with the windows regs and the 
linux regs:

the values before --- are for linux (while still ntsc) and
the values after --- are for windows (ATI MMC pal_m)
4,5c4,5
< fld_LP_CONTRAST = 110
< fld_LP_BRIGHTNESS = 16368
---
>fld_LP_CONTRAST = 106
>fld_LP_BRIGHTNESS = 0
8c8
< fld_H_SCALE_RATIO = 84468
---
>fld_H_SCALE_RATIO = 66901
10,11c10,11
< fld_V_SCALE_RATIO = 1718
< fld_V_DEINTERLACE_ON = 1
---
>fld_V_SCALE_RATIO = 2048
>fld_V_DEINTERLACE_ON = 0
14,19c14,19
< fld_EVENF_OFFSET = 82
< fld_ODDF_OFFSET = 82
< fld_INTERLACE_DETECTED = 0
< fld_VS_LINE_COUNT = 6
< fld_VS_DETECTED_LINES = 7
< fld_VS_ITU656_VB = 1
---
>fld_EVENF_OFFSET = 512
>fld_ODDF_OFFSET = 1536
>fld_INTERLACE_DETECTED = 1
>fld_VS_LINE_COUNT = 434
>fld_VS_DETECTED_LINES = 525
>fld_VS_ITU656_VB = 0
30,32c30,32
< fld_H_VBI_WIND_START = 73
< fld_H_VBI_WIND_END = 870
< fld_V_VBI_WIND_START = 13
---
>fld_H_VBI_WIND_START = 101
>fld_H_VBI_WIND_END = 871
>fld_V_VBI_WIND_START = 11
39c39
< fld_VBI_CAPTURE_ENABLE = 0
---
>fld_VBI_CAPTURE_ENABLE = 1
44c44
< fld_VBI_SCALING_RATIO = 61791
---
>fld_VBI_SCALING_RATIO = 65536
46,47c46,47
< fld_H_ACTIVE_START = 112
< fld_H_ACTIVE_END = 867
---
>fld_H_ACTIVE_START = 104
>fld_H_ACTIVE_END = 838
49,51c49,51
< fld_V_ACTIVE_END = 517
< fld_CH_HEIGHT = 205
< fld_CH_KILL_LEVEL = 192
---
>fld_V_ACTIVE_END = 518
>fld_CH_HEIGHT = 147
>fld_CH_KILL_LEVEL = 142
56c56
< fld_STANDARD_SEL = 0
---					/IMPORTANT/
>fld_STANDARD_SEL = 1			must be 1 not 0
60,61c60,61
< fld_ADC_PREFLO = 3
< fld_H_SYNC_PULSE_WIDTH = 71
---
>fld_ADC_PREFLO = 0
>fld_H_SYNC_PULSE_WIDTH = 69
75c75
< fld_CBDB_ACTIVE_GAIN = 154
---
>fld_CBDB_ACTIVE_GAIN = 153
82,84c82,84
< fld_COMB_CNTL0 = 155418768		/VERY IMPORTANTE/
< fld_COMB_CNTL1 = 16
< fld_COMB_CNTL2 = 101187842
---
>fld_COMB_CNTL0 = 1116831888		/these are the right values (at least for the 
>tuner)/
>fld_COMB_CNTL1 = 177
>fld_COMB_CNTL2 = -1			0xffffffff
86c86
< fld_SYNCTIP_REF0 = 55
---
>fld_SYNCTIP_REF0 = 57
96c96
< fld_CH_DTO_INC = 4194304
---
>fld_CH_DTO_INC = 4189610		/<-correct one/
103,104c103,104
< fld_H_IN_WIND_START = 154
< fld_V_IN_WIND_START = 44
---
>fld_H_IN_WIND_START = 100
>fld_V_IN_WIND_START = 37
106c106
< fld_V_OUT_WIND_WIDTH = 240
---
>fld_V_OUT_WIND_WIDTH = 480
108,110c108,110
< fld_MIN_PULSE_WIDTH = 50
< fld_MAX_PULSE_WIDTH = 84
< fld_WIN_CLOSE_LIMIT = 77
---
>fld_MIN_PULSE_WIDTH = 33
>fld_MAX_PULSE_WIDTH = 100
>fld_WIN_CLOSE_LIMIT = 160
130c130
< fld_VS_FIELD_IDLOCATION = 1
---
>fld_VS_FIELD_IDLOCATION = 261			(it's better to have 261)
132c132
< fld_SYNC_TIP_START = 881
---
>fld_SYNC_TIP_START = 882
137c137
< fld_I_AGC_SEL = 1
---
>fld_I_AGC_SEL = 3
140c140
< fld_DECI_DITHER_EN = 0
---
>fld_DECI_DITHER_EN = 1



>completely different ? which ones ?
well, I can't remember right now, thinking about it I think there aren't 
regs completely different. sorry... you could lookup the diff and look for 
totally different values.

>usually:
>   diff -u orig/src/dir new/src/dir > name_of_diff
thanks, I did them individually (only 3 files)

_________________________________________________________________
MSN Photos is the easiest way to share and print your photos: 
http://photos.msn.com/support/worldwide.aspx

Ramiro Polla scribbled:

> Hello,
> I wrote some messages here a while ago because I wanted to get PAL_M 
> working with avview, and after some tests and a lot of help (specially 
> from deek) I got it kind of working.
> -on Windows I used rttest to get the values for all the registers, and 
> with ati.2 made it so that when I changed to pal-m it used all the 
> windows regs. With that I got the screen very well, just like with ATI 
> MMC, with no problems.
> -then I made a function for ati.2 that wrote all the regs to a file. I 
> made a diff between the windows regs and the linux regs and changing 
> them one by one found out which were the most crucial to get PAL_M.


could I see all of the regs that rttest dumped from windows ?


> -I found out PAL_M uses a screwed up standard, being that some values 
> are from PAL, some from NTSC, and some completely different, so I want 
> the help of all of you to instead of having extPAL_M under DEC_NTSC, 
> making a new DEC_PALM.
> 


completely different ? which ones ?


> I wasn't able to change the values for SetOutputVideoSize so that I 
> could get the whole video screen working, I got the same I would get 
> with NTSC:
> one fifth of the bottom of the screen is screwed up, one seventh of the 
> right of the screen is screwed up, and the actual video is shifted a bit 
> to the left, so I lose a part of the screen.
> 
> Also I had to change the COMB_CNTLx values and I did not know how to 
> change the ones for the Tuner, so I changed the Composite ones and left 
> svideo the way they were.
> 
> here are the diffs with todays' CVS (I don't know how you people 
> normally use diff, so I did mine differently):
> 


usually:
   diff -u orig/src/dir new/src/dir > name_of_diff


> The diffs for r128_video.c (I didn't change for radeon and the other one):
> <_snipped__>


-- 
  Issued as is without warranty of any kind; without even the
  implied warranty of merchantability or fitness for a
  particular purpose - GNU General Public License

  the unofficial official GATOS irc at:
  irc.openprojects.net #gatos-ati
	# Ross Coleman
-----

Hello,
I wrote some messages here a while ago because I wanted to get PAL_M working 
with avview, and after some tests and a lot of help (specially from deek) I 
got it kind of working.
-on Windows I used rttest to get the values for all the registers, and with 
ati.2 made it so that when I changed to pal-m it used all the windows regs. 
With that I got the screen very well, just like with ATI MMC, with no 
problems.
-then I made a function for ati.2 that wrote all the regs to a file. I made 
a diff between the windows regs and the linux regs and changing them one by 
one found out which were the most crucial to get PAL_M.
-I found out PAL_M uses a screwed up standard, being that some values are 
from PAL, some from NTSC, and some completely different, so I want the help 
of all of you to instead of having extPAL_M under DEC_NTSC, making a new 
DEC_PALM.

I wasn't able to change the values for SetOutputVideoSize so that I could 
get the whole video screen working, I got the same I would get with NTSC:
one fifth of the bottom of the screen is screwed up, one seventh of the 
right of the screen is screwed up, and the actual video is shifted a bit to 
the left, so I lose a part of the screen.

Also I had to change the COMB_CNTLx values and I did not know how to change 
the ones for the Tuner, so I changed the Composite ones and left svideo the 
way they were.

here are the diffs with todays' CVS (I don't know how you people normally 
use diff, so I did mine differently):

The diffs for r128_video.c (I didn't change for radeon and the other one):

--- ati.2.orig/r128_video.c	Sun Mar 31 12:20:16 2002
+++ ati.2/r128_video.c	Sun Mar 31 12:50:05 2002
@@ -197,15 +197,18 @@
     { 1, "pal-composite",		720, 240, { 1, 50 }},
     { 2, "pal-tuner",			720, 240, { 1, 50 }},
     { 3, "pal-svideo",			720, 240, { 1, 50 }},
-    { 4, "ntsc-composite",		640, 240, { 1001, 60000 }},
-    { 5, "ntsc-tuner",			640, 240, { 1001, 60000 }},
-    { 6, "ntsc-svideo",			640, 240, { 1001, 60000 }},
-    { 7, "secam-composite",		720, 240, { 1, 50 }},
-    { 8, "secam-tuner",			720, 240, { 1, 50 }},
-    { 9, "secam-svideo",		720, 240, { 1, 50 }},
-    { 10,"pal_60-composite",		768, 288, { 1, 50 }},
-    { 11,"pal_60-tuner",		768, 288, { 1, 50 }},
-    { 12,"pal_60-svideo",		768, 288, { 1, 50 }}
+    { 4, "pal_m-composite",            640, 240, { 1001, 60000 }},
+    { 5, "pal_m-tuner",                640, 240, { 1001, 60000 }},
+    { 6, "pal_m-svideo",               640, 240, { 1001, 60000 }},
+    { 7, "ntsc-composite",             640, 240, { 1001, 60000 }},
+    { 8, "ntsc-tuner",                 640, 240, { 1001, 60000 }},
+    { 9, "ntsc-svideo",                640, 240, { 1001, 60000 }},
+    { 10, "secam-composite",           720, 288, { 1, 50 }},
+    { 11, "secam-tuner",               720, 288, { 1, 50 }},
+    { 12, "secam-svideo",              720, 288, { 1, 50 }},
+    { 13,"pal_60-composite",           720, 288, { 1, 50 }},
+    { 14,"pal_60-tuner",               720, 288, { 1, 50 }},
+    { 15,"pal_60-svideo",              720, 288, { 1, 50 }}
};

#define NUM_FORMATS 12
@@ -2529,46 +2532,60 @@
		pPriv->v=24;
		break;
	case 4:
+               xf86_RT_SetStandard(pPriv->theatre,DEC_NTSC | extPAL_M);
+               xf86_RT_SetConnector(pPriv->theatre,DEC_COMPOSITE, 0);
+               pPriv->v=24;
+               break;
+       case 5:
+               xf86_RT_SetStandard(pPriv->theatre,DEC_NTSC | extPAL_M);
+               xf86_RT_SetConnector(pPriv->theatre,DEC_TUNER,0);
+               pPriv->v=24;
+               break;
+       case 6:
+               xf86_RT_SetStandard(pPriv->theatre,DEC_NTSC | extPAL_M);
+               xf86_RT_SetConnector(pPriv->theatre,DEC_SVIDEO,0);
+               pPriv->v=24;
+       case 7:
		xf86_RT_SetStandard(pPriv->theatre,DEC_NTSC | extNONE);
		xf86_RT_SetConnector(pPriv->theatre, DEC_COMPOSITE,0);
		pPriv->v=23;
		break;
-	case 5:
+	case 8:
		xf86_RT_SetStandard(pPriv->theatre,DEC_NTSC | extNONE);
		xf86_RT_SetConnector(pPriv->theatre, DEC_TUNER, 0);
		pPriv->v=23;
		break;
-	case 6:
+	case 9:
		xf86_RT_SetStandard(pPriv->theatre,DEC_NTSC | extNONE);
		xf86_RT_SetConnector(pPriv->theatre, DEC_SVIDEO, 0);
		pPriv->v=23;
		break;
-	case 7:
+	case 10:
		xf86_RT_SetStandard(pPriv->theatre,DEC_SECAM | extNONE);
		xf86_RT_SetConnector(pPriv->theatre, DEC_COMPOSITE, 0);
		pPriv->v=25;
		break;
-	case 8:
+	case 11:
		xf86_RT_SetStandard(pPriv->theatre,DEC_SECAM | extNONE);
		xf86_RT_SetConnector(pPriv->theatre, DEC_TUNER, 0);
		pPriv->v=25;
		break;
-	case 9:

+	case 12:
		xf86_RT_SetStandard(pPriv->theatre,DEC_SECAM | extNONE);
		xf86_RT_SetConnector(pPriv->theatre, DEC_SVIDEO, 0);
		pPriv->v=25;
		break;
-	case 10:
+	case 13:
		xf86_RT_SetStandard(pPriv->theatre,DEC_PAL | extPAL_60);
		xf86_RT_SetConnector(pPriv->theatre,DEC_COMPOSITE, 0);
		pPriv->v=24;
		break;
-	case 11:
+	case 14:
		xf86_RT_SetStandard(pPriv->theatre,DEC_PAL | extPAL_60);
		xf86_RT_SetConnector(pPriv->theatre,DEC_TUNER,0);
		pPriv->v=24;
		break;
-	case 12:
+	case 15:
		xf86_RT_SetStandard(pPriv->theatre,DEC_PAL | extPAL_60);
		xf86_RT_SetConnector(pPriv->theatre,DEC_SVIDEO,0);
		pPriv->v=24;
@@ -2608,15 +2625,15 @@
		pPriv->msp3430->connector = MSP3430_CONNECTOR_2;
		break;
	case 7:
-		pPriv->msp3430->standard = MSP3430_SECAM;
+		pPriv->msp3430->standard = MSP3430_NTSC;
		pPriv->msp3430->connector = MSP3430_CONNECTOR_3;
		break;
	case 8:
-		pPriv->msp3430->standard = MSP3430_SECAM;
+		pPriv->msp3430->standard = MSP3430_NTSC;
		pPriv->msp3430->connector = MSP3430_CONNECTOR_1;
		break;
	case 9:
-		pPriv->msp3430->standard = MSP3430_SECAM;
+		pPriv->msp3430->standard = MSP3430_NTSC;
		pPriv->msp3430->connector = MSP3430_CONNECTOR_2;
		break;
	case 10:
@@ -2628,6 +2645,18 @@
		pPriv->msp3430->connector = MSP3430_CONNECTOR_1;
		break;
	case 12:
+               pPriv->msp3430->standard = MSP3430_SECAM;
+               pPriv->msp3430->connector = MSP3430_CONNECTOR_2;
+               break;
+       case 13:
+               pPriv->msp3430->standard = MSP3430_SECAM;
+               pPriv->msp3430->connector = MSP3430_CONNECTOR_3;
+               break;
+       case 14:
+               pPriv->msp3430->standard = MSP3430_SECAM;
+               pPriv->msp3430->connector = MSP3430_CONNECTOR_1;
+               break;
+       case 15:
		pPriv->msp3430->standard = MSP3430_SECAM;
		pPriv->msp3430->connector = MSP3430_CONNECTOR_2;
		break;





The diffs for theatre.c:

--- ati.2.orig/theatre.c	Sun Mar 31 12:20:16 2002
+++ ati.2/theatre.c	Sun Mar 31 12:36:50 2002
@@ -1187,6 +1187,8 @@
             switch (wStandard & 0xFF00)
             {
                 case (extPAL_M):
+                    WriteRT_fld (fld_CH_DTO_INC,       RT_PALM_CH_DTO_INC);
+                    WriteRT_fld (fld_STANDARD_SEL,     RT_PAL);
                 case (extNONE):
                 case (extNTSC):
                     WriteRT_fld (fld_CR_BURST_GAIN,        
RT_NTSCM_CR_BURST_GAIN);

The diffs for theatre_reg.h:

--- ati.2.orig/theatre_reg.h	Sun Mar 31 12:24:22 2002
+++ ati.2/theatre_reg.h	Sun Mar 31 12:32:38 2002
@@ -324,7 +324,7 @@
#define     RT_PALN_COMB_CNTL0_COMPOSITE      0x09438090
#define     RT_PALN_COMB_CNTL0_SVIDEO         0x40348090

-#define     RT_PALM_COMB_CNTL0_COMPOSITE      0x09438090
+#define     RT_PALM_COMB_CNTL0_COMPOSITE      0x42918090
#define     RT_PALM_COMB_CNTL0_SVIDEO         0x40348090
/* End of filter settings. */

@@ -341,7 +341,7 @@
#define     RT_PALN_COMB_CNTL1_COMPOSITE      0x00000010
#define     RT_PALN_COMB_CNTL1_SVIDEO         0x000000A1

-#define     RT_PALM_COMB_CNTL1_COMPOSITE      0x00000010
+#define     RT_PALM_COMB_CNTL1_COMPOSITE      0x000000B1
#define     RT_PALM_COMB_CNTL1_SVIDEO         0x000000A1
/* End of filter settings. */

@@ -358,7 +358,7 @@
#define     RT_PALN_COMB_CNTL2_COMPOSITE      0x06080102
#define     RT_PALN_COMB_CNTL2_SVIDEO         0x06080102

-#define     RT_PALM_COMB_CNTL2_COMPOSITE      0x06080102
+#define     RT_PALM_COMB_CNTL2_COMPOSITE      0xFFFFFFFF
#define     RT_PALM_COMB_CNTL2_SVIDEO         0x06080102
/* End of filter settings. */

@@ -435,6 +435,8 @@
#define     RT_NTSCJ_CBDB_ACTIVE_GAIN          0x0000009F
#define     RT_NTSCJ_VERT_LOCKOUT_START        0x00000207
#define     RT_NTSCJ_VERT_LOCKOUT_END          0x0000000E
+
+#define     RT_PALM_CH_DTO_INC                 0x003FEDAA /* added by 
Ramiro Polla, I kind of had no place to put it so I put it here */

#define     RT_PAL_SYNCTIP_REF0                0x37  /* instead of 
0x00000004 - Ivo */
#define     RT_PAL_SYNCTIP_REF1                0x26  /* instead of 
0x0000000F - Ivo */


I hope someone can help



_________________________________________________________________
Chat with friends online, try MSN Messenger: http://messenger.msn.com

Are you running main or devel branch ? If devel branch then the reason
avview fails is likely because devel is too old.. I'll get around to
fixing this sometime, but not now.

                          Vladimir Dergachev

On Sat, 30 Mar 2002, Chris Jensen wrote:

> > The other funny thing that I've noticed, is that since I've updated my
> > drivers, as well as not being able to run 800x600 anymore, some colours
> > don't appear to be displayed correctly when playing videos back with xi=
ne
> > (haven't tried any other players).
> > I'm thinking there's something strange going on with my X setup.
>
> I moved my X11R6 directory and installed xfree again from source because =
I
> thought I may have installed straight over XF 4.1.0, and maybe that was
> causing something strange, but I still have the same problem using the la=
test
> drivers - 800x600 is out of sync and certain colours don't appear properl=
y
> when playing back video - they sort of appear transparent, although they'=
re
> not showing any particular image, more like static.
> If I revert to my previous xfree setup(xfree 4.2.0 with ati.2 drivers and
> modifications to r128_video.c from Chris Bare) I can get perfect DVD play=
back
> in 800x600.
> But whatever setup I use, avview still falls over in the same spot.
>
> On Thu, 28 Mar 2002 18:36, Chris Jensen wrote:
> > What else do you want to know about the card? It's an AGP All-In-Wonder=
 Pro
> > 128 with 32MB of ram on board, i'm using the r128 drivers for it (that'=
s
> > correct isn't it?) I bought it in Australia and it's outputing on s-vid=
eo
> > if that makes difference.
> > I tried the avview from today, but got the same problem.
> > Something else that's happening is that if I can only run it as root,
> > otherwise, it looks up pretty much straight away - the overlay displays=
 the
> > input, but the user interface is unresponsive, and doesn't redraw if it
> > gets covered and then unconvered, the only way to close it is to hit Ct=
rl-C
> > in the console - hitting X on the window won't do it.
> > This also happens when I'm running as root if I left the alsa mixer ope=
n
> > last time I closed the program - the only way to get it to start withou=
t
> > locking up is to delete /root/.avview_state
> >
> >
> > On Wed, 27 Mar 2002 18:22, Vladimir Dergachev wrote:
> > >=A0This looks strange.. OVERLAY_GAMMA shows that you are running recen=
t
> > >=A0drivers, but then avview should not crash that way. Can you remind =
me
> > >=A0which video card you have and also try the very latest avview (got
> > > updated again).
> >
> > Thanks for your help.
>
> --
> Chris Jensen
> ch...@dr...
>
> Public Key: http://drspirograph.com/public_key/
>
> Wait: Did you know that there's a direct correlation between the decline =
of
> Spirograph and the rise in gang activity?  Think about it.
> - Dr Spirograph (The Simpsons)
>
>
> _______________________________________________
> Gatos-devel mailing list
> Gat...@li...
> https://lists.sourceforge.net/lists/listinfo/gatos-devel
>

> The other funny thing that I've noticed, is that since I've updated=
 my
> drivers, as well as not being able to run 800x600 anymore, some col=
ours
> don't appear to be displayed correctly when playing videos back wit=
h xine
> (haven't tried any other players).
> I'm thinking there's something strange going on with my X setup.

I moved my X11R6 directory and installed xfree again from source beca=
use I=20
thought I may have installed straight over XF 4.1.0, and maybe that w=
as=20
causing something strange, but I still have the same problem using th=
e latest=20
drivers - 800x600 is out of sync and certain colours don't appear pro=
perly=20
when playing back video - they sort of appear transparent, although t=
hey're=20
not showing any particular image, more like static.
If I revert to my previous xfree setup(xfree 4.2.0 with ati.2 drivers=
 and=20
modifications to r128_video.c from Chris Bare) I can get perfect DVD =
playback=20
in 800x600.
But whatever setup I use, avview still falls over in the same spot.

On Thu, 28 Mar 2002 18:36, Chris Jensen wrote:
> What else do you want to know about the card? It's an AGP All-In-Wo=
nder Pro
> 128 with 32MB of ram on board, i'm using the r128 drivers for it (t=
hat's
> correct isn't it?) I bought it in Australia and it's outputing on s=
-video
> if that makes difference.
> I tried the avview from today, but got the same problem.
> Something else that's happening is that if I can only run it as roo=
t,
> otherwise, it looks up pretty much straight away - the overlay disp=
lays the
> input, but the user interface is unresponsive, and doesn't redraw i=
f it
> gets covered and then unconvered, the only way to close it is to hi=
t Ctrl-C
> in the console - hitting X on the window won't do it.
> This also happens when I'm running as root if I left the alsa mixer=
 open
> last time I closed the program - the only way to get it to start wi=
thout
> locking up is to delete /root/.avview_state
>
>
> On Wed, 27 Mar 2002 18:22, Vladimir Dergachev wrote:
> >=A0This looks strange.. OVERLAY_GAMMA shows that you are running r=
ecent
> >=A0drivers, but then avview should not crash that way. Can you rem=
ind me
> >=A0which video card you have and also try the very latest avview (=
got
> > updated again).
>
> Thanks for your help.

--=20
Chris Jensen
ch...@dr...

Public Key: http://drspirograph.com/public_key/

Wait: Did you know that there's a direct correlation between the decl=
ine of
Spirograph and the rise in gang activity?  Think about it.
- Dr Spirograph (The Simpsons)

I am not aware of any v4l2 drivers.. There is a v4l (not 2) driver - km.

                      Vladimir Dergachev

On Sat, 30 Mar 2002, Happy Camper wrote:

>
> I just bought an ATI AIW 7500.  I am wondering if there is a V4L2 driver
> for it.  The V4L2 main page says :
> "Octavian Purdila is part of the GATOS group and he is working on a V4L2
> driver  for ATI graphics cards".
>
> But I cant find any mention of it on the Gatos website.
>
> Any information regarding this or pointers in the right direction would
> be greatly appreciated.
>
> Thanks.
>
> Atif
>
> __________________________________________________
> Do You Yahoo!?
> Yahoo! Greetings - send holiday greetings for Easter, Passover
> http://greetings.yahoo.com/
>
> _______________________________________________
> Gatos-devel mailing list
> Gat...@li...
> https://lists.sourceforge.net/lists/listinfo/gatos-devel
>

I just bought an ATI AIW 7500.  I am wondering if there is a V4L2 driver
for it.  The V4L2 main page says :
"Octavian Purdila is part of the GATOS group and he is working on a V4L2
driver  for ATI graphics cards".

But I cant find any mention of it on the Gatos website.

Any information regarding this or pointers in the right direction would
be greatly appreciated.

Thanks.

Atif

__________________________________________________
Do You Yahoo!?
Yahoo! Greetings - send holiday greetings for Easter, Passover
http://greetings.yahoo.com/

On 31 Mar 2002, Bill Triplett wrote:

> > Not supported with regular drivers. See
> >  http://gatos.sf.net/watching_tv.php
> > Vladimir Dergachev
>
> Thanks for the tip. I was under the assumption that the latest ati.2 cvs
> was considered devel, but I see there's a separate devel one.

Yes. The main branch is "stable".. The devel branch is a little old and
contains some TV-out code that uses int10 calls. The code is flaky and
slow, though it does work in select resolutions.

If you only want TV-out you can also try putting return right after
SetupImageVideo function starts in r128_video.c - see if this helps the
picture.

                      Vladimir Dergachev

>
> I installed that one, and so far I've been able to get it to work with
> the XFree config file here:
>
>   http://dev.nethouse.com/gatostuff/XF86Config-4_tvout.txt
>
> The only mode that works so far is 800x600.
>
> Here is the log:
>
>   http://dev.nethouse.com/gatostuff/XFree86.0.log_tvout.txt
>
> Note the:
>
> (EE) R128(0): Failed to determine current TV standard
> (II) R128(0): ax=0xa016 bx=0x0002 cx=0x0000
>
> Could that be related to why the picture on the TV set looks really
> funky and psychedelic? (The monitor version of xawtv looks fine.)
>
> Thanks alot,
> Bill
>
> On 30 Mar 2002, Bill Triplett wrote:
>
> > I have a line run from my AIW card's TV out (AV OUT) to my VCR's
> > video-in RCA jack and can see the boot-up process on the TV as well as
> > the monitor. However, when I start X, screen goes black and my
> > computer crashes hard. There is gibberish in my XFree86.0.log to the
> > tune of what looks ^@'s and loaded kernel module symbol listings. The
> > file system is probably a little hung over. Even though / is mounted
> > sync with write-caching turned off... doesn't help :(
>
>
>
> _______________________________________________
> Gatos-devel mailing list
> Gat...@li...
> https://lists.sourceforge.net/lists/listinfo/gatos-devel
>

On 31 Mar 2002, Bill Triplett wrote:

> > Not supported with regular drivers. See
> >  http://gatos.sf.net/watching_tv.php
> > Vladimir Dergachev
>
> Thanks for the tip. I was under the assumption that the latest ati.2 cvs
> was considered devel, but I see there's a separate devel one.
>
> I installed that one, and so far I've been able to get it to work with
> the XFree config file here:
>
>   http://dev.nethouse.com/gatostuff/XF86Config-4_tvout.txt
>
> The only mode that works so far is 800x600.
>
> Here is the log:
>
>   http://dev.nethouse.com/gatostuff/XFree86.0.log_tvout.txt
>
> Note the:
>
> (EE) R128(0): Failed to determine current TV standard

This is always so.. For some reason the BIOS call that does this isn't
working.

> (II) R128(0): ax=0xa016 bx=0x0002 cx=0x0000
>
> Could that be related to why the picture on the TV set looks really
> funky and psychedelic? (The monitor version of xawtv looks fine.)

Funky ?

Try using 800x600 16bpp.

                       Vladimir Dergachev

>
> Thanks alot,
> Bill
>
> On 30 Mar 2002, Bill Triplett wrote:
>
> > I have a line run from my AIW card's TV out (AV OUT) to my VCR's
> > video-in RCA jack and can see the boot-up process on the TV as well as
> > the monitor. However, when I start X, screen goes black and my
> > computer crashes hard. There is gibberish in my XFree86.0.log to the
> > tune of what looks ^@'s and loaded kernel module symbol listings. The
> > file system is probably a little hung over. Even though / is mounted
> > sync with write-caching turned off... doesn't help :(
>
>
>
> _______________________________________________
> Gatos-devel mailing list
> Gat...@li...
> https://lists.sourceforge.net/lists/listinfo/gatos-devel
>

> Not supported with regular drivers. See 
>  http://gatos.sf.net/watching_tv.php
> Vladimir Dergachev 

Thanks for the tip. I was under the assumption that the latest ati.2 cvs
was considered devel, but I see there's a separate devel one. 

I installed that one, and so far I've been able to get it to work with
the XFree config file here:

  http://dev.nethouse.com/gatostuff/XF86Config-4_tvout.txt

The only mode that works so far is 800x600.

Here is the log:

  http://dev.nethouse.com/gatostuff/XFree86.0.log_tvout.txt

Note the:

(EE) R128(0): Failed to determine current TV standard
(II) R128(0): ax=0xa016 bx=0x0002 cx=0x0000

Could that be related to why the picture on the TV set looks really
funky and psychedelic? (The monitor version of xawtv looks fine.)

Thanks alot,
Bill

On 30 Mar 2002, Bill Triplett wrote: 

> I have a line run from my AIW card's TV out (AV OUT) to my VCR's 
> video-in RCA jack and can see the boot-up process on the TV as well as
> the monitor. However, when I start X, screen goes black and my 
> computer crashes hard. There is gibberish in my XFree86.0.log to the 
> tune of what looks ^@'s and loaded kernel module symbol listings. The 
> file system is probably a little hung over. Even though / is mounted 
> sync with write-caching turned off... doesn't help :( 

Not supported with regular drivers. See
http://gatos.sf.net/watching_tv.php

                        Vladimir Dergachev

On 30 Mar 2002, Bill Triplett wrote:

> Hi all,
>
> I have a line run from my AIW card's TV out (AV OUT) to my VCR's
> video-in RCA jack and can see the boot-up process on the TV as well as
> the monitor. However, when I start X, screen goes black and my computer
> crashes hard. There is gibberish in my XFree86.0.log to the tune of what
> looks ^@'s and loaded kernel module symbol listings. The file system is
> probably a little hung over. Even though / is mounted sync with
> write-caching turned off... doesn't help :(
>
> I added the composite_sync off option to the X config file, but no luck.
> This happens regardless of whether dri is enabled and I have no
> framebuffer support enabled, enabling km makes no difference.
>
> Any help would be appreciated.
>
> Kernel 2.4.19-pre5
> All In Wonder r128 Pro 32M AGP
> CVS is up to date, just checked.
>
> Thanks,
> Bill
>
>
>
> _______________________________________________
> Gatos-devel mailing list
> Gat...@li...
> https://lists.sourceforge.net/lists/listinfo/gatos-devel
>

Hi all, 

I have a line run from my AIW card's TV out (AV OUT) to my VCR's
video-in RCA jack and can see the boot-up process on the TV as well as
the monitor. However, when I start X, screen goes black and my computer
crashes hard. There is gibberish in my XFree86.0.log to the tune of what
looks ^@'s and loaded kernel module symbol listings. The file system is
probably a little hung over. Even though / is mounted sync with
write-caching turned off... doesn't help :(

I added the composite_sync off option to the X config file, but no luck.
This happens regardless of whether dri is enabled and I have no
framebuffer support enabled, enabling km makes no difference.

Any help would be appreciated.

Kernel 2.4.19-pre5
All In Wonder r128 Pro 32M AGP
CVS is up to date, just checked.

Thanks,
Bill

On Sat, 30 Mar 2002, Nick Kurshev wrote:

> Hello, Vladimir!
>
> On Sat, 30 Mar 2002 11:56:58 -0500 (EST) you wrote:
>
> >
> >
> > On Sat, 30 Mar 2002, Nick Kurshev wrote:
> >
> > > Hello, Vladimir!
> > >
> > > On Fri, 29 Mar 2002 09:15:44 -0500 (EST) you wrote:
> > >
> > > I know that well, but I told about SOURCE buffer for driver but NOT
> > > for graphics chip.
> >
> > You lost me.
> >
> > What I meant is that you program the overlay with the offset into cards
> > framebuffer. That offset should be aligned on the size of cards memory
> > bus.
> >
> This is truht for case when driver performs direct copying of source buffer
> into video memory. AFAIK gatos driver performs copying of buffer's lines instead
> of copying of whole buffer with one memcpy() call.
> So it can realign every line.

The lines need to be aligned as well.

                          Vladimir Dergachev

> >                         Vladimir Dergachev
> >
> > > >
> > > > Everything that you feed to the graphics card should be aligned on the
> > > > size of cards memory bus. For Rage128 it is (AFAIK) dwords and for Radeon
> > > > it is QWORDS. Not aligning properly will result in incorrect operation,
> > > > what kind however I do not know as documentation does not say.
> > > >
> > > >                               Vladimir Dergachev
> > > >
> > > > On Fri, 29 Mar 2002, Nick Kurshev wrote:
> > > >
> > > > > Hello!
> > > > >
> > > > > I have one question to gatos mantainers:
> > > > > what should be stride for source buffer?
> > > > >
> > > > > static int
> > > > > RADEONPutImage(
> > > > > ...
> > > > >
> > > > >    switch(id) {
> > > > > ...
> > > > >    case FOURCC_YV12:
> > > > >    case FOURCC_I420:
> > > > > 	dstPitch = ((width << 1) + 15) & ~15;
> > > > > 	new_size = ((dstPitch * height) + bpp - 1) / bpp;
> > > > > 	srcPitch = (width + 3) & ~3;
> > > > > 	s2offset = srcPitch * height;
> > > > > 	srcPitch2 = ((width >> 1) + 3) & ~3;
> > > > > 	s3offset = (srcPitch2 * (height >> 1)) + s2offset;
> > > > > 	break;
> > > > > ...
> > > > > >From this source I can see that source buffer should be 8 byte aligned, isn't?
> > > > > Is it really documented feature or is it gatos feature?
> > > > > At least mplayer(xp) currently assumed that source buffer is 1 byte aligned!
> > > > >
> > > > > I have this question because I have broken color during watching of well-known 405.avi
> > > > > with gatos drivers.
> > > > >
> > > > > Really: this movie has width 356!
> > > > > well: y-plane is 4-byte aligned (356/4)*4 = 356 - no problems
> > > > > but srcPitch2 is 4 byte aligned too!
> > > > > Therefore Y-plane should be 8 byte aligned:
> > > > > (356/8)*8 = 352!!!
> > > > >
> > > > > Any ideas!
> > > > >
> > > > > Best regards! Nick
> > > > >
> > > >
> > > >
> > > > _______________________________________________
> > > > Gatos-devel mailing list
> > > > Gat...@li...
> > > > https://lists.sourceforge.net/lists/listinfo/gatos-devel
> > > >
> > >
> > >
> > > Best regards! Nick
> > >
> >
> >
> > _______________________________________________
> > Gatos-devel mailing list
> > Gat...@li...
> > https://lists.sourceforge.net/lists/listinfo/gatos-devel
> >
>
>
> Best regards! Nick
>

Hello, Vladimir!

On Sat, 30 Mar 2002 11:56:58 -0500 (EST) you wrote:

> 
> 
> On Sat, 30 Mar 2002, Nick Kurshev wrote:
> 
> > Hello, Vladimir!
> >
> > On Fri, 29 Mar 2002 09:15:44 -0500 (EST) you wrote:
> >
> > I know that well, but I told about SOURCE buffer for driver but NOT
> > for graphics chip.
> 
> You lost me.
> 
> What I meant is that you program the overlay with the offset into cards
> framebuffer. That offset should be aligned on the size of cards memory
> bus.
> 
This is truht for case when driver performs direct copying of source buffer
into video memory. AFAIK gatos driver performs copying of buffer's lines instead
of copying of whole buffer with one memcpy() call.
So it can realign every line.
>                         Vladimir Dergachev
> 
> > >
> > > Everything that you feed to the graphics card should be aligned on the
> > > size of cards memory bus. For Rage128 it is (AFAIK) dwords and for Radeon
> > > it is QWORDS. Not aligning properly will result in incorrect operation,
> > > what kind however I do not know as documentation does not say.
> > >
> > >                               Vladimir Dergachev
> > >
> > > On Fri, 29 Mar 2002, Nick Kurshev wrote:
> > >
> > > > Hello!
> > > >
> > > > I have one question to gatos mantainers:
> > > > what should be stride for source buffer?
> > > >
> > > > static int
> > > > RADEONPutImage(
> > > > ...
> > > >
> > > >    switch(id) {
> > > > ...
> > > >    case FOURCC_YV12:
> > > >    case FOURCC_I420:
> > > > 	dstPitch = ((width << 1) + 15) & ~15;
> > > > 	new_size = ((dstPitch * height) + bpp - 1) / bpp;
> > > > 	srcPitch = (width + 3) & ~3;
> > > > 	s2offset = srcPitch * height;
> > > > 	srcPitch2 = ((width >> 1) + 3) & ~3;
> > > > 	s3offset = (srcPitch2 * (height >> 1)) + s2offset;
> > > > 	break;
> > > > ...
> > > > >From this source I can see that source buffer should be 8 byte aligned, isn't?
> > > > Is it really documented feature or is it gatos feature?
> > > > At least mplayer(xp) currently assumed that source buffer is 1 byte aligned!
> > > >
> > > > I have this question because I have broken color during watching of well-known 405.avi
> > > > with gatos drivers.
> > > >
> > > > Really: this movie has width 356!
> > > > well: y-plane is 4-byte aligned (356/4)*4 = 356 - no problems
> > > > but srcPitch2 is 4 byte aligned too!
> > > > Therefore Y-plane should be 8 byte aligned:
> > > > (356/8)*8 = 352!!!
> > > >
> > > > Any ideas!
> > > >
> > > > Best regards! Nick
> > > >
> > >
> > >
> > > _______________________________________________
> > > Gatos-devel mailing list
> > > Gat...@li...
> > > https://lists.sourceforge.net/lists/listinfo/gatos-devel
> > >
> >
> >
> > Best regards! Nick
> >
> 
> 
> _______________________________________________
> Gatos-devel mailing list
> Gat...@li...
> https://lists.sourceforge.net/lists/listinfo/gatos-devel
> 


Best regards! Nick

On Sat, 30 Mar 2002, Nick Kurshev wrote:

> Hello, Vladimir!
>
> On Fri, 29 Mar 2002 09:15:44 -0500 (EST) you wrote:
>
> I know that well, but I told about SOURCE buffer for driver but NOT
> for graphics chip.

You lost me.

What I meant is that you program the overlay with the offset into cards
framebuffer. That offset should be aligned on the size of cards memory
bus.

                        Vladimir Dergachev

> >
> > Everything that you feed to the graphics card should be aligned on the
> > size of cards memory bus. For Rage128 it is (AFAIK) dwords and for Radeon
> > it is QWORDS. Not aligning properly will result in incorrect operation,
> > what kind however I do not know as documentation does not say.
> >
> >                               Vladimir Dergachev
> >
> > On Fri, 29 Mar 2002, Nick Kurshev wrote:
> >
> > > Hello!
> > >
> > > I have one question to gatos mantainers:
> > > what should be stride for source buffer?
> > >
> > > static int
> > > RADEONPutImage(
> > > ...
> > >
> > >    switch(id) {
> > > ...
> > >    case FOURCC_YV12:
> > >    case FOURCC_I420:
> > > 	dstPitch = ((width << 1) + 15) & ~15;
> > > 	new_size = ((dstPitch * height) + bpp - 1) / bpp;
> > > 	srcPitch = (width + 3) & ~3;
> > > 	s2offset = srcPitch * height;
> > > 	srcPitch2 = ((width >> 1) + 3) & ~3;
> > > 	s3offset = (srcPitch2 * (height >> 1)) + s2offset;
> > > 	break;
> > > ...
> > > >From this source I can see that source buffer should be 8 byte aligned, isn't?
> > > Is it really documented feature or is it gatos feature?
> > > At least mplayer(xp) currently assumed that source buffer is 1 byte aligned!
> > >
> > > I have this question because I have broken color during watching of well-known 405.avi
> > > with gatos drivers.
> > >
> > > Really: this movie has width 356!
> > > well: y-plane is 4-byte aligned (356/4)*4 = 356 - no problems
> > > but srcPitch2 is 4 byte aligned too!
> > > Therefore Y-plane should be 8 byte aligned:
> > > (356/8)*8 = 352!!!
> > >
> > > Any ideas!
> > >
> > > Best regards! Nick
> > >
> >
> >
> > _______________________________________________
> > Gatos-devel mailing list
> > Gat...@li...
> > https://lists.sourceforge.net/lists/listinfo/gatos-devel
> >
>
>
> Best regards! Nick
>

Hello, Vladimir!

On Fri, 29 Mar 2002 09:15:44 -0500 (EST) you wrote:

I know that well, but I told about SOURCE buffer for driver but NOT
for graphics chip.
> 
> Everything that you feed to the graphics card should be aligned on the
> size of cards memory bus. For Rage128 it is (AFAIK) dwords and for Radeon
> it is QWORDS. Not aligning properly will result in incorrect operation,
> what kind however I do not know as documentation does not say.
> 
>                               Vladimir Dergachev
> 
> On Fri, 29 Mar 2002, Nick Kurshev wrote:
> 
> > Hello!
> >
> > I have one question to gatos mantainers:
> > what should be stride for source buffer?
> >
> > static int
> > RADEONPutImage(
> > ...
> >
> >    switch(id) {
> > ...
> >    case FOURCC_YV12:
> >    case FOURCC_I420:
> > 	dstPitch = ((width << 1) + 15) & ~15;
> > 	new_size = ((dstPitch * height) + bpp - 1) / bpp;
> > 	srcPitch = (width + 3) & ~3;
> > 	s2offset = srcPitch * height;
> > 	srcPitch2 = ((width >> 1) + 3) & ~3;
> > 	s3offset = (srcPitch2 * (height >> 1)) + s2offset;
> > 	break;
> > ...
> > >From this source I can see that source buffer should be 8 byte aligned, isn't?
> > Is it really documented feature or is it gatos feature?
> > At least mplayer(xp) currently assumed that source buffer is 1 byte aligned!
> >
> > I have this question because I have broken color during watching of well-known 405.avi
> > with gatos drivers.
> >
> > Really: this movie has width 356!
> > well: y-plane is 4-byte aligned (356/4)*4 = 356 - no problems
> > but srcPitch2 is 4 byte aligned too!
> > Therefore Y-plane should be 8 byte aligned:
> > (356/8)*8 = 352!!!
> >
> > Any ideas!
> >
> > Best regards! Nick
> >
> 
> 
> _______________________________________________
> Gatos-devel mailing list
> Gat...@li...
> https://lists.sourceforge.net/lists/listinfo/gatos-devel
> 


Best regards! Nick

On Sat, 30 Mar 2002, Parienti M. wrote:

> On Wed, 27 Mar 2002 02:27:58 -0500 (EST)
> Vladimir Dergachev <vo...@mi...> wrote:
>
> >
> > Which video mode are you running at ? Does the flickering go away if yo=
u
> > reduce resolution or refresh rate ?
>
>
> I reduce the resolution to 1024x768 and 800x600, and I have
> the same problem, even if it is less visible.
>
>
> My vertical sync is always 85 Hz (85.02 exactly for 1280x1024
> resolution, read from xvidtune), which is almost conform to
> the specification of my monitor, for these resolutions (85.0).
>
> My horizontal frequency is 91.14 (from xvidtune), and the
> specification say that it should be 91.15.
>
>
> Can my problem come from this?

It could. Try adding Vertrefresh 75 to your monitor section.

Also, is the flickering you see visible only on the thatched pattern (the
one you see when X comes up) or on solid tones as well ?

                            Vladimir Dergachev


>
>
>
> > On Mon, 25 Mar 2002, [ISO-8859-1] Micha=EBl Parienti wrote:
> >
> > > On Mon, 25 Mar 2002 08:15:46 -0500 (EST)
> > > Vladimir Dergachev <vo...@mi...> wrote:
> > >
> > > > > >
> > > > > > Are you using CRT or flat panel monitor ? Do you use monitor sw=
itch ?
> > > > >
> > > > >
> > > > > I have a Iiyama, a vision master pro 410 aka A702HT.
> > > > >
> > > > >
> > > > > I don't understand what you mean by "monitor switch". My
> > > >
> > > > A box with buttons that allows you to switch one monitor between tw=
o
> > > > computers.
> > >
> > >
> > > So the answer is "No, I haven't". I just have a simple
> > > monitor.
>
>
> --
> Micha=EBl P.
> gpg:        D4C8 F73D A000 71C7 44EF  27E6 8982 4991 7126 3CE3

On Wed, 27 Mar 2002 02:27:58 -0500 (EST)
Vladimir Dergachev <vo...@mi...> wrote:

> 
> Which video mode are you running at ? Does the flickering go away if you
> reduce resolution or refresh rate ?


I reduce the resolution to 1024x768 and 800x600, and I have
the same problem, even if it is less visible.


My vertical sync is always 85 Hz (85.02 exactly for 1280x1024
resolution, read from xvidtune), which is almost conform to
the specification of my monitor, for these resolutions (85.0).

My horizontal frequency is 91.14 (from xvidtune), and the
specification say that it should be 91.15.


Can my problem come from this?



> On Mon, 25 Mar 2002, [ISO-8859-1] Michaël Parienti wrote:
> 
> > On Mon, 25 Mar 2002 08:15:46 -0500 (EST)
> > Vladimir Dergachev <vo...@mi...> wrote:
> >
> > > > >
> > > > > Are you using CRT or flat panel monitor ? Do you use monitor switch ?
> > > >
> > > >
> > > > I have a Iiyama, a vision master pro 410 aka A702HT.
> > > >
> > > >
> > > > I don't understand what you mean by "monitor switch". My
> > >
> > > A box with buttons that allows you to switch one monitor between two
> > > computers.
> >
> >
> > So the answer is "No, I haven't". I just have a simple
> > monitor.


-- 
Michaël P.
gpg:        D4C8 F73D A000 71C7 44EF  27E6 8982 4991 7126 3CE3

Greetings all,

	Tomorrow I'm going to give your drivers a whirl on a digital alpha PC164LX 
machine.  I do have a few questions.

1) has this ever been tried before?

2)  I'm using an all in wonder radeon PCI.  In the kernel configuration, I 
only can see the DRM module available when i add agp support (or as a 
module)...I dont have agp, does this matter?  I do want to use the kernel 
module loader.

I will keep all of you informed to my success/lack thereof.

Ted Goodridge

R C scribbled:

> On Fri, Mar 29, 2002 at 12:05:04PM -0500, deek wrote:
> 
>>
>>Bahman M. Farahani scribbled:
>>
>>
>>>Is there any closed-caption support for ATI all in wonder?
>>>If not, can anyone help me to the right direction of how one can make
>>>closed-caption work for this video card?
>>>
>>>Bahman
>>>
>>>
>>Currently, there is no CC support for any of the ATI cards.
>>(AFAIK) With the exception of the AIW-classic, as this card is used
>>in conjunction with the bttv/ v4l drivers.
>>
> 
> There are old CC decoders out there, including a kernel module based one
> with interrupt support, for the AIW original, and Radeon. I wasn't happy
> with them, and got no response at the time.
> 


where ?

> 
>>I've, actually, spent considerable time playing with this very
>>thing. I've come to the conclusion that the best way would be
>>to add one more modules to ati.2... xxx_cc_drv.o
>>(or something to that effect) with an XV_ATOM toggle.
>>
>>
> 
> Umm. No. Read the Xpert archives; I already did this for the AIW
> original, and was told it would be rejected.


Perhaps. But all we are doing is drawing to an existing overlay
from within the driver by toggling the feature on/ off
I still devoutly believe that this sort of feature belongs in
ati.2 vice a subsequent piece of code.

> 
> 
>>or, perhaps, extend the functionality of xxx_video.c
>>
>>IMHO, coding these into km would be remarkably redundant.
>>Most if not all of the required/ necessary functions exist,
>>currently, in the ati.2 code base.
>>
> 
> Fairly simple way to do it, but I have not implemented it yet. To do
> this efficiently, you need interrupt support, which means putting at
> least some of it in km. usleep() latency is too unpredictable to use it
> for exact timing.
> 
> 
>>Teletext, on the other hand, belongs in KM. Or, perhaps, a hybrid 
>>adaptation of ati.2 and km. But, that [teletext] is a slightly
>>different beast all together.
>>
> 
> Not really. The solution is to provide /dev/vbi.  You setup two more
> buffers (small ones), one page'll do it, in km. You setup new, or extend
> and place at the end, in video ram.  Then you dump the raw data to
> userspace, where it gets processed. Same for CC or Teletext, but slight
> differences in data length, IIRC.
> 
> Even for the cards that can decode in hardware, it's generally faster to
> decode the CC in software (roughly 1.3k of data, every 60 hz, is nothing
> to a modern processor).  Specifically, I2C and it's busy loops take up a
> lot of CPU on those processors with a BT829 (and thus a hardware CC
> decoder).


agreed. I implied slightly different in as much as we are reading from a
character device and then drawing that information onto the overlay.
which is surpassingly easy to implement... I've done this to some extent.
just haven't figured out a reliable way to do anything with /dev/vbi
after it is created.

> 
> R C
> 


-- 
  Issued as is without warranty of any kind; without even the
  implied warranty of merchantability or fitness for a
  particular purpose - GNU General Public License

  the unofficial official GATOS irc at:
  irc.openprojects.net #gatos-ati
	# Ross Coleman
-----