Discussion:
[directfb-users] directfb-users] Can't get any hardwareaccelerationusingthesavage driver.
Tim Nicholson
2015-04-09 06:22:42 UTC
Permalink
Thanks,

I have managed to get it all to work now, there were a couple of issues.

1 The case statement for LUT8 was missing in the colour selection routine of savage4.c so my test in 8 bit mode was using software fall back. I have fixed that.

2 There is a macro that waits for the GPU hardware to become ready, it was not yielding (it just loops while incrementing a counter!) so the CPU load was much, much higher than it should be, I have sorted that out now by adding a usleep call and it is now behaving much better. I have also added a non blocking function to call to test if the GPU is busy to help with multitasking. The CPU loading is now sub 0.1% for my program which was what I had expected.

Thanks tor your help,

Tim
hi,tim
Sorry, It's too late to replay you email.
I just read the code about how the graphics initialized.
In your log, the accelerator is worked!
When I have time I will read your log again and If I found any thing usefull I will email you!
(-) [Main Thread       0.176] (  303) Core/Graphics:      dfb_graphics_core_initialize( 0x804d170, 0x8051048, 0x8051110 )
(*) DirectFB/Genefx: MMX detected and enabled
(-) [Main Thread       0.176] (  303) Direct/Modules:     direct_modules_explore_directory( 'gfxdrivers' )
(-) [Main Thread       0.176] (  303) Direct/Mem:           +    44 bytes [modules.c:232 in direct_modules_explore_directory()]
(-) [Main Thread       0.176] (  303) Direct/Mem:           +    22 bytes [modules.c:240 in direct_modules_explore_directory()] -> 0x8052838 "         libdirectfb_savage.so"
(-) [Main Thread       0.176] (  303) Direct/Modules:     Loading '/usr/local/lib/directfb-1.2-9/gfxdrivers/libdirectfb_savage.so'...
(-) [Main Thread       0.177] (  303) Direct/Modules:     Registering 'savage' ('gfxdrivers')...
(-) [Main Thread       0.177] (  303) Direct/Mem:           +     7 bytes [modules.c:134 in direct_modules_register()] -> 0x8051b28 "                        savage"
(-) [Main Thread       0.177] (  303) Direct/Modules:     ...registered.
savage4:savage4getinfo
(-) [Main Thread       0.177] (  303) Direct/Mem:           +     8 bytes [gfxcard.c:210 in dfb_graphics_core_initialize()]
savage4:savage4initdriver
savage4:savage4initdevice
(*) DirectFB/Graphics: S3 Savage4 Series 0.3 (directfb.org)‍
------------------ Original ------------------
Date:  Mon, Mar 30, 2015 10:19 PM
Subject:  Re: [directfb-users] Can't get any hardwareaccelerationusingthesavage driver.
Hi,
(!) [Main Thread       0.237] (  303) *** Assumption [core != NULL] failed *** [colorhash.c:206 in dfb_colorhash_attach()]
I also notice that calls to dfb_gfxcard_state_check() return accel = 0x00000000 which I assume means that no acceleration is available for the function it is checking. What could cause that?
Would you mind having a quick look and let me know your thoughts? It looks like the savage driver is loading correctly but I don't really know where to look from there.
When I run with --dfb:software-warn
I get messages for fill rectangle and drawline which fits with the log 
my own debug code, returned that the call to dfb_gfxcard_get_accelerator( device ) was returning FB_ACCEL_S3TWISTER_P    (0x8a)
Regards
Tim
Hi, 
First, you can set the config for acceleration use these value.
#define FB_ACCEL_SAVAGE4        0x80 /* S3 Savage4                   */
#define FB_ACCEL_SAVAGE3D       0x81 /* S3 Savage3D                  */
#define FB_ACCEL_SAVAGE3D_MV    0x82 /* S3 Savage3D-MV               */
#define FB_ACCEL_SAVAGE2000     0x83 /* S3 Savage2000                */
#define FB_ACCEL_SAVAGE_MX_MV   0x84 /* S3 Savage/MX-MV              */
#define FB_ACCEL_SAVAGE_MX      0x85 /* S3 Savage/MX                 */
#define FB_ACCEL_SAVAGE_IX_MV   0x86 /* S3 Savage/IX-MV              */
#define FB_ACCEL_SAVAGE_IX      0x87 /* S3 Savage/IX                 */
#define FB_ACCEL_PROSAVAGE_PM   0x88 /* S3 ProSavage PM133           */
#define FB_ACCEL_PROSAVAGE_KM   0x89 /* S3 ProSavage KM133           */
#define FB_ACCEL_S3TWISTER_P    0x8a /* S3 Twister                   */
#define FB_ACCEL_S3TWISTER_K    0x8b /* S3 TwisterK                  */
#define FB_ACCEL_SUPERSAVAGE    0x8c    /* S3 Supersavage               */
#define FB_ACCEL_PROSAVAGE_DDR  0x8d /* S3 ProSavage DDR             */
#define FB_ACCEL_PROSAVAGE_DDRK 0x8e /* S3 ProSavage DDR-K           */‍
Second, add some log in savage, why it does not use this video card
Good Luck!
Try do something and give me some information, so I can help you more!
------------------ Original ------------------
Date:  Fri, Mar 27, 2015 11:41 PM
Subject:  Re: [directfb-users] Can't get any hardware accelerationusingthesavage driver.
Hi,
Ok, I can give it a go, but how would I go about doing that? Could you give me some pointers, such as were to look and how to get a log? The current build is using the following, should I enable additional options to get more information?
./configure --disable-png --disable-jpeg --disable-freetype --with-gfxdrivers=savage
Thanks for your help, I am happy to do the work, I just need to a little help about how to go about it and what to look for.
Regards
Tim
Hi,tim
I just want to say you video card is so old.
But It is support by DirectFB.
Under the directory Gfxdrivers/Savage, you can find the drivers code. You can debug and find why the video driver is not work!
Or you can send me a log and I will try to find why it's no work!
------------------ Original ------------------
Date:  Fri, Mar 27, 2015 10:22 PM
Subject:  Re: [directfb-users] Can't get any hardware acceleration usingthesavage driver.
Hi,
Yes that does work, I think that is the default option if you pass no parameters isn't it?
However, The program does not seem to be running with hardware acceleration. I say that because it runs the same speed when I run it with --dfb:no-hardware! It is also the same speed as if I run using the vesa frame buffer (/dev/fb1) which does not support acceleration. I really wanted to get the GPU to do the work.
 
Regards
hi, tim
You can juse use this command
-bash-3.1# ./test --dfb:hardware
I think it will work.
"  [no-]hardware                  Enable/disable hardware acceleration\n"‍
option "no-software"  disable software fallbacks. 
This is mainly for  debugging/profiling purpose. 
It will print out a warning with the operation and flags each time a fallback would occur.
And I think it is just print some information.
 
------------------ Original ------------------
Date:  Fri, Mar 27, 2015 08:21 PM
Subject:  [directfb-users] Can't get any hardware acceleration using thesavage driver.
Hi,
I am trying to take advantage of the graphics accelerator on some embedded linux hardware. I have sucessfully built directfb and I have written some test code which runs fine but, it only appears to be using software drawing and I am at a loss as to what I need to do to utilize the hardware acceleration. Below are some details about my setup. Any help of advice would be much appreciated.
-bash-3.1$ gcc -I /usr/local/include/directfb -lm -ldirectfb directfb_test.c -o test
grub menu.lst entry
kernel /boot/linux-2.6.16.46 root=/dev/hdc4 i8042.direct=1 atkbd.set=3 video=savagefb:mtrr,ywrap,pmipal,vga=0x301 ide=nodma quiet console=tty7 loglevel=0
-bash-3.1# cat /proc/version 
-bash-3.1# fbset -i
mode "640x400-70"
    # D: 25.176 MHz, H: 31.469 kHz, V: 69.932 Hz
    geometry 640 400 640 4096 8
    timings 39721 40 24 39 9 96 2
    accel true
    rgba 8/0,8/0,8/0,0/0
endmode
    Name        : TwisterP
    Address     : 0xe0000000
    Size        : 33423360
    Type        : PACKED PIXELS
    Visual      : PSEUDOCOLOR
    XPanStep    : 4
    YPanStep    : 1
    YWrapStep   : 0
    LineLength  : 640
    MMIO Address: 0xed000000
    MMIO Size   : 524288
    Accelerator : Unknown (138)
-bash-3.1$ lspci
00:00.0 Host bridge: VIA Technologies, Inc. VT8605 [ProSavage PM133]
00:01.0 PCI bridge: VIA Technologies, Inc. VT8605 [PM133 AGP]
00:07.0 ISA bridge: VIA Technologies, Inc. VT82C686 [Apollo Super South] (rev 40)
00:07.1 IDE interface: VIA Technologies, Inc. VT82C586A/B/VT82C686/A/B/VT823x/A/C PIPC Bus Master IDE (rev 06)
00:07.2 USB Controller: VIA Technologies, Inc. VT82xxxxx UHCI USB 1.1 Controller (rev 1d)
00:07.4 Bridge: VIA Technologies, Inc. VT82C686 [Apollo Super ACPI] (rev 40)
00:07.5 Multimedia audio controller: VIA Technologies, Inc. VT82C686 AC97 Audio Controller (rev 50)
00:0e.0 Ethernet controller: National Semiconductor Corporation DP83815 (MacPhyter) Ethernet Controller
01:00.0 VGA compatible controller: S3 Inc. 86C380 [ProSavageDDR K4M266] (rev 02)
-bash-3.1# ./test
test application is starting
   ~~~~~~~~~~~~~~~~~~~~~~~~~~| DirectFB 1.2.9 |~~~~~~~~~~~~~~~~~~~~~~~~~~
        (c) 2001-2008  The world wide DirectFB Open Source Community
        (c) 2000-2004  Convergence (integrated media) GmbH
      ----------------------------------------------------------------
(*) DirectFB/Core: Single Application Core. (2015-03-23 14:57) 
(*) Direct/Thread: Started 'VT Switcher' (-1) [CRITICAL OTHER/OTHER 0/0] <8388608>...
(*) Direct/Thread: Started 'Linux Input' (-1) [INPUT OTHER/OTHER 0/0] <8388608>...
(*) DirectFB/Input: AT Raw Set 3 keyboard (1) 0.1 (directfb.org)
(*) Direct/Thread: Started 'Linux Input' (-1) [INPUT OTHER/OTHER 0/0] <8388608>...
(*) DirectFB/Input: PC Speaker (2) 0.1 (directfb.org)
(*) Direct/Thread: Started 'Linux Input' (-1) [INPUT OTHER/OTHER 0/0] <8388608>...
(*) DirectFB/Input: DELL DELL USB Keyboard (3) 0.1 (directfb.org)
(*) Direct/Thread: Started 'Keyboard Input' (-1) [INPUT OTHER/OTHER 0/0] <8388608>...
(*) DirectFB/Input: Keyboard 0.9 (directfb.org)
(*) Direct/Thread: Started 'PS/2 Input' (-1) [INPUT OTHER/OTHER 0/0] <8388608>...
(*) DirectFB/Input: IMPS/2 Mouse 1.0 (directfb.org)
(*) DirectFB/Genefx: MMX detected and enabled
(*) DirectFB/Graphics: S3 Savage4 Series 0.3 (directfb.org)
(*) DirectFB/Core/WM: Default 0.3 (directfb.org)
(*) FBDev/Surface: Allocated 640x400 8 bit LUT8 buffer (index 0) at offset 0 and pitch 640.
(*) FBDev/Mode: Setting 640x400 LUT8
(*) FBDev/Mode: Switched to 640x400 (virtual 640x800) at 8 bit (LUT8), pitch 640
(*) FBDev/Surface: Allocated 640x400 8 bit LUT8 buffer (index 1) at offset 256000 and pitch 640.
(*) FBDev/Mode: Setting 640x400 LUT8
(*) FBDev/Mode: Switched to 640x400 (virtual 640x400) at 8 bit (LUT8), pitch 640
(*) FBDev/Surface: Allocated 640x400 8 bit LUT8 buffer (index 0) at offset 0 and pitch 640.
When I run the code specifying no-software, as follows, I just get a blank screen, so I am assuming hardware acceleration is not working.
-bash-3.1# ./test --dfb:no-software
test application is starting
   ~~~~~~~~~~~~~~~~~~~~~~~~~~| DirectFB 1.2.9 |~~~~~~~~~~~~~~~~~~~~~~~~~~
        (c) 2001-2008  The world wide DirectFB Open Source Community
        (c) 2000-2004  Convergence (integrated media) GmbH
      ----------------------------------------------------------------
(*) DirectFB/Core: Single Application Core. (2015-03-23 14:57) 
(*) Direct/Thread: Started 'VT Switcher' (-1) [CRITICAL OTHER/OTHER 0/0] <8388608>...
(*) Direct/Thread: Started 'Linux Input' (-1) [INPUT OTHER/OTHER 0/0] <8388608>...
(*) DirectFB/Input: AT Raw Set 3 keyboard (1) 0.1 (directfb.org)
(*) Direct/Thread: Started 'Linux Input' (-1) [INPUT OTHER/OTHER 0/0] <8388608>...
(*) DirectFB/Input: PC Speaker (2) 0.1 (directfb.org)
(*) Direct/Thread: Started 'Linux Input' (-1) [INPUT OTHER/OTHER 0/0] <8388608>...
(*) DirectFB/Input: DELL DELL USB Keyboard (3) 0.1 (directfb.org)
(*) Direct/Thread: Started 'Keyboard Input' (-1) [INPUT OTHER/OTHER 0/0] <8388608>...
(*) DirectFB/Input: Keyboard 0.9 (directfb.org)
(*) Direct/Thread: Started 'PS/2 Input' (-1) [INPUT OTHER/OTHER 0/0] <8388608>...
(*) DirectFB/Input: IMPS/2 Mouse 1.0 (directfb.org)
(*) DirectFB/Genefx: MMX detected and enabled
(*) DirectFB/Graphics: S3 Savage4 Series 0.3 (directfb.org)
(*) DirectFB/Core/WM: Default 0.3 (directfb.org)
(*) FBDev/Surface: Allocated 640x400 8 bit LUT8 buffer (index 0) at offset 0 and pitch 640.
(*) FBDev/Mode: Setting 640x400 LUT8
(*) FBDev/Mode: Switched to 640x400 (virtual 640x800) at 8 bit (LUT8), pitch 640
(*) FBDev/Surface: Allocated 640x400 8 bit LUT8 buffer (index 1) at offset 256000 and pitch 640.
(*) FBDev/Mode: Setting 640x400 LUT8
(*) FBDev/Mode: Switched to 640x400 (virtual 640x400) at 8 bit (LUT8), pitch 640
(*) FBDev/Surface: Allocated 640x400 8 bit LUT8 buffer (index 0) at offset 0 and pitch 640
/****************************************************
* Name:                directfb_test.c
* Author:              Tim Nicholson
* Purpose:             Just for test the DirectFB
****************************************************/
#include <stdio.h>
#include <unistd.h>
#include <directfb.h>
#include <math.h>
#define HEIGHT 480
#define WIDTH 640
#define WHOLE 150
#define HALF (WHOLE/2.0)
#define MIDX (WIDTH/2.0)
#define MIDY (HEIGHT/2.0)
static struct PointF
{
double x;
double y;
} c1,c2,c3,c4,p1,p2,p3,p4,p5,p6,p7,p8,mid1,mid2;
static int x = 0;
void calc (void);
int main (int argc, char *argv[])
{
  IDirectFB             *dfb   = NULL;
  IDirectFBDisplayLayer *layer = NULL;
  IDirectFBSurface      *surface = NULL;
  printf ("test application is starting\n");
  DirectFBInit (&argc, &argv);
  DirectFBCreate (&dfb);
  dfb->GetDisplayLayer (dfb, DLID_PRIMARY, &layer);
  layer->SetCooperativeLevel (layer, DLSCL_EXCLUSIVE);
  layer->GetSurface (layer, &surface);
  mid1.x = MIDX - HALF;
  mid1.y = MIDY;
  mid2.x = MIDX + HALF;
  mid2.y = MIDY;
  while (1)
    {
      calc();
      surface->SetColor (surface, 0x00, 0x00, 0x00, 0xff);
      surface->FillRectangle (surface, 0, 0, WIDTH, HEIGHT);
      surface->SetColor (surface, 0xff, 0xff, 0xff, 0xff);
      
      surface->DrawLine (surface, p1.x, p1.y, p2.x, p2.y);
      surface->DrawLine (surface, p2.x, p2.y, p3.x, p3.y);
      surface->DrawLine (surface, p3.x, p3.y, p4.x, p4.y);
      surface->DrawLine (surface, p4.x, p4.y, p1.x, p1.y);
      surface->DrawLine (surface, p5.x, p5.y, p6.x, p6.y);
      surface->DrawLine (surface, p6.x, p6.y, p7.x, p7.y);
      surface->DrawLine (surface, p7.x, p7.y, p8.x, p8.y);
      surface->DrawLine (surface, p8.x, p8.y, p5.x, p5.y);
      surface->DrawLine (surface, p1.x, p1.y, p5.x, p5.y);
      surface->DrawLine (surface, p2.x, p2.y, p6.x, p6.y);
      surface->DrawLine (surface, p3.x, p3.y, p7.x, p7.y);
      surface->DrawLine (surface, p4.x, p4.y, p8.x, p8.y);
      dfb->WaitForSync(dfb);
    }
  surface->Release (surface);
  layer->Release (layer);
  dfb->Release (dfb);
  return 0;
}
// *********************************************************************************************
// c a l c
// *********************************************************************************************
void calc (void)
{
  int a,b,c,d;
  x++;
  a = x;
  b = x + 90;
  c = x + 180;
  d = x + 270;
  a %= 360;
  b %= 360;
  c %= 360;
  d %= 360;
  c1.x = (sin(M_PI / 180.0 * a) * WHOLE);
  c1.y = (cos(M_PI / 180.0 * a) * WHOLE);
  c2.x = (sin(M_PI / 180.0 * b) * WHOLE);
  c2.y = (cos(M_PI / 180.0 * b) * WHOLE);
  c3.x = (sin(M_PI / 180.0 * c) * WHOLE);
  c3.y = (cos(M_PI / 180.0 * c) * WHOLE);
  c4.x = (sin(M_PI / 180.0 * d) * WHOLE);
  c4.y = (cos(M_PI / 180.0 * d) * WHOLE);
  p1.x = mid1.x + c1.x;
  p2.x = mid1.x + c2.x;
  p3.x = mid1.x + c3.x;
  p4.x = mid1.x + c4.x;
  p5.x = mid2.x + c1.x;
  p6.x = mid2.x + c2.x;
  p7.x = mid2.x + c3.x;
  p8.x = mid2.x + c4.x;
  p1.y = mid1.y + c1.y;
  p2.y = mid1.y + c2.y;
  p3.y = mid1.y + c3.y;
  p4.y = mid1.y + c4.y;
  p5.y = mid2.y + c1.y;
  p6.y = mid2.y + c2.y;
  p7.y = mid2.y + c3.y;
  p8.y = mid2.y + c4.y;
}
Freeman
2015-04-09 07:10:16 UTC
Permalink
:)
Well done!








------------------ Original ------------------
From: "tim";<***@gizzy.co.uk>;
Date: Thu, Apr 9, 2015 02:22 PM
To: "Freeman"<***@foxmail.com>;
Cc: "directfb-users"<directfb-***@directfb.org>;
Subject: Re: [directfb-users] directfb-users] Can't get anyhardwareaccelerationusingthesavage driver.



Thanks,

I have managed to get it all to work now, there were a couple of issues.

1 The case statement for LUT8 was missing in the colour selection routine of savage4.c so my test in 8 bit mode was using software fall back. I have fixed that.

2 There is a macro that waits for the GPU hardware to become ready, it was not yielding (it just loops while incrementing a counter!) so the CPU load was much, much higher than it should be, I have sorted that out now by adding a usleep call and it is now behaving much better. I have also added a non blocking function to call to test if the GPU is busy to help with multitasking. The CPU loading is now sub 0.1% for my program which was what I had expected.

Thanks tor your help,

Tim



Freeman <***@foxmail.com> wrote:
hi,tim
Sorry, It's too late to replay you email.


I just read the code about how the graphics initialized.


In your log, the accelerator is worked!


When I have time I will read your log again and If I found any thing usefull I will email you!


(-) [Main Thread 0.176] ( 303) Core/Graphics: dfb_graphics_core_initialize( 0x804d170, 0x8051048, 0x8051110 )
(*) DirectFB/Genefx: MMX detected and enabled
(-) [Main Thread 0.176] ( 303) Direct/Modules: direct_modules_explore_directory( 'gfxdrivers' )
(-) [Main Thread 0.176] ( 303) Direct/Mem: + 44 bytes [modules.c:232 in direct_modules_explore_directory()]
(-) [Main Thread 0.176] ( 303) Direct/Mem: + 22 bytes [modules.c:240 in direct_modules_explore_directory()] -> 0x8052838 " libdirectfb_savage.so"
(-) [Main Thread 0.176] ( 303) Direct/Modules: Loading '/usr/local/lib/directfb-1.2-9/gfxdrivers/libdirectfb_savage.so'...
(-) [Main Thread 0.177] ( 303) Direct/Modules: Registering 'savage' ('gfxdrivers')...
(-) [Main Thread 0.177] ( 303) Direct/Mem: + 7 bytes [modules.c:134 in direct_modules_register()] -> 0x8051b28 " savage"
(-) [Main Thread 0.177] ( 303) Direct/Modules: ...registered.
savage4:savage4getinfo
(-) [Main Thread 0.177] ( 303) Direct/Mem: + 8 bytes [gfxcard.c:210 in dfb_graphics_core_initialize()]
savage4:savage4initdriver
savage4:savage4initdevice
(*) DirectFB/Graphics: S3 Savage4 Series 0.3 (directfb.org)‍







------------------ Original ------------------
From: "tim";<***@gizzy.co.uk>;
Date: Mon, Mar 30, 2015 10:19 PM
To: "Freeman"<***@foxmail.com>;
Cc: "directfb-users"<directfb-***@directfb.org>;
Subject: Re: [directfb-users] Can't get any hardwareaccelerationusingthesavage driver.




Hi,

I have attached a log file (dump.txt) for my application. I have had a look through it and the only thing that really jumps out at me is:


(!) [Main Thread 0.237] ( 303) *** Assumption [core != NULL] failed *** [colorhash.c:206 in dfb_colorhash_attach()]


I also notice that calls to dfb_gfxcard_state_check() return accel = 0x00000000 which I assume means that no acceleration is available for the function it is checking. What could cause that?


Would you mind having a quick look and let me know your thoughts? It looks like the savage driver is loading correctly but I don't really know where to look from there.


When I run with --dfb:software-warn


I get messages for fill rectangle and drawline which fits with the log


my own debug code, returned that the call to dfb_gfxcard_get_accelerator( device ) was returning FB_ACCEL_S3TWISTER_P (0x8a)


Regards




Tim




On March 27, 2015, Freeman <***@foxmail.com> wrote:
Hi,
First, you can set the config for acceleration use these value.
#define FB_ACCEL_SAVAGE4 0x80 /* S3 Savage4 */
#define FB_ACCEL_SAVAGE3D 0x81 /* S3 Savage3D */
#define FB_ACCEL_SAVAGE3D_MV 0x82 /* S3 Savage3D-MV */
#define FB_ACCEL_SAVAGE2000 0x83 /* S3 Savage2000 */
#define FB_ACCEL_SAVAGE_MX_MV 0x84 /* S3 Savage/MX-MV */
#define FB_ACCEL_SAVAGE_MX 0x85 /* S3 Savage/MX */
#define FB_ACCEL_SAVAGE_IX_MV 0x86 /* S3 Savage/IX-MV */
#define FB_ACCEL_SAVAGE_IX 0x87 /* S3 Savage/IX */
#define FB_ACCEL_PROSAVAGE_PM 0x88 /* S3 ProSavage PM133 */
#define FB_ACCEL_PROSAVAGE_KM 0x89 /* S3 ProSavage KM133 */
#define FB_ACCEL_S3TWISTER_P 0x8a /* S3 Twister */
#define FB_ACCEL_S3TWISTER_K 0x8b /* S3 TwisterK */
#define FB_ACCEL_SUPERSAVAGE 0x8c /* S3 Supersavage */
#define FB_ACCEL_PROSAVAGE_DDR 0x8d /* S3 ProSavage DDR */
#define FB_ACCEL_PROSAVAGE_DDRK 0x8e /* S3 ProSavage DDR-K */‍



Second, add some log in savage, why it does not use this video card


Good Luck!
Try do something and give me some information, so I can help you more!


------------------ Original ------------------
From: "tim";<***@gizzy.co.uk>;
Date: Fri, Mar 27, 2015 11:41 PM
To: "directfb-users"<directfb-***@directfb.org>;

Subject: Re: [directfb-users] Can't get any hardware accelerationusingthesavage driver.



Hi,

Ok, I can give it a go, but how would I go about doing that? Could you give me some pointers, such as were to look and how to get a log? The current build is using the following, should I enable additional options to get more information?


./configure --disable-png --disable-jpeg --disable-freetype --with-gfxdrivers=savage


Thanks for your help, I am happy to do the work, I just need to a little help about how to go about it and what to look for.



Regards


Tim


On March 27, 2015, Freeman <***@foxmail.com> wrote:
Hi,tim
I just want to say you video card is so old.
But It is support by DirectFB.
Under the directory Gfxdrivers/Savage, you can find the drivers code. You can debug and find why the video driver is not work!
Or you can send me a log and I will try to find why it's no work!




------------------ Original ------------------
From: "tim";<***@gizzy.co.uk>;
Date: Fri, Mar 27, 2015 10:22 PM
To: "directfb-users"<directfb-***@directfb.org>;

Subject: Re: [directfb-users] Can't get any hardware acceleration usingthesavage driver.



Hi,

Yes that does work, I think that is the default option if you pass no parameters isn't it?


However, The program does not seem to be running with hardware acceleration. I say that because it runs the same speed when I run it with --dfb:no-hardware! It is also the same speed as if I run using the vesa frame buffer (/dev/fb1) which does not support acceleration. I really wanted to get the GPU to do the work.



Regards








On March 27, 2015, Freeman <***@foxmail.com> wrote:
hi, tim
You can juse use this command
-bash-3.1# ./test --dfb:hardware
I think it will work.


" [no-]hardware Enable/disable hardware acceleration\n"‍


option "no-software" disable software fallbacks.
This is mainly for debugging/profiling purpose.
It will print out a warning with the operation and flags each time a fallback would occur.



And I think it is just print some information.







------------------ Original ------------------
From: "Tim Nicholson";<***@gizzy.co.uk>;
Date: Fri, Mar 27, 2015 08:21 PM
To: "directfb-users"<directfb-***@directfb.org>;

Subject: [directfb-users] Can't get any hardware acceleration using thesavage driver.



Hi,

I am trying to take advantage of the graphics accelerator on some embedded linux hardware. I have sucessfully built directfb and I have written some test code which runs fine but, it only appears to be using software drawing and I am at a loss as to what I need to do to utilize the hardware acceleration. Below are some details about my setup. Any help of advice would be much appreciated.


I have included the source code for my test program below, it simply rotates a wire cube using multiple DrawLine functions , I am compiling it using the following command line:


-bash-3.1$ gcc -I /usr/local/include/directfb -lm -ldirectfb directfb_test.c -o test


I have included details of various outputs below in the hope they may be of use:


grub menu.lst entry


kernel /boot/linux-2.6.16.46 root=/dev/hdc4 i8042.direct=1 atkbd.set=3 video=savagefb:mtrr,ywrap,pmipal,vga=0x301 ide=nodma quiet console=tty7 loglevel=0


-bash-3.1# cat /proc/version
Linux version 2.6.16.46 (***@sentinel) (gcc version 4.0.3) #1 PREEMPT Sat Mar 10 13:17:45 EST 2012



-bash-3.1# fbset -i


mode "640x400-70"
# D: 25.176 MHz, H: 31.469 kHz, V: 69.932 Hz
geometry 640 400 640 4096 8
timings 39721 40 24 39 9 96 2
accel true
rgba 8/0,8/0,8/0,0/0
endmode


Frame buffer device information:
Name : TwisterP
Address : 0xe0000000
Size : 33423360
Type : PACKED PIXELS
Visual : PSEUDOCOLOR
XPanStep : 4
YPanStep : 1
YWrapStep : 0
LineLength : 640
MMIO Address: 0xed000000
MMIO Size : 524288
Accelerator : Unknown (138)



-bash-3.1$ lspci
00:00.0 Host bridge: VIA Technologies, Inc. VT8605 [ProSavage PM133]
00:01.0 PCI bridge: VIA Technologies, Inc. VT8605 [PM133 AGP]
00:07.0 ISA bridge: VIA Technologies, Inc. VT82C686 [Apollo Super South] (rev 40)
00:07.1 IDE interface: VIA Technologies, Inc. VT82C586A/B/VT82C686/A/B/VT823x/A/C PIPC Bus Master IDE (rev 06)
00:07.2 USB Controller: VIA Technologies, Inc. VT82xxxxx UHCI USB 1.1 Controller (rev 1d)
00:07.4 Bridge: VIA Technologies, Inc. VT82C686 [Apollo Super ACPI] (rev 40)
00:07.5 Multimedia audio controller: VIA Technologies, Inc. VT82C686 AC97 Audio Controller (rev 50)
00:0e.0 Ethernet controller: National Semiconductor Corporation DP83815 (MacPhyter) Ethernet Controller
01:00.0 VGA compatible controller: S3 Inc. 86C380 [ProSavageDDR K4M266] (rev 02)



-bash-3.1# ./test
test application is starting


~~~~~~~~~~~~~~~~~~~~~~~~~~| DirectFB 1.2.9 |~~~~~~~~~~~~~~~~~~~~~~~~~~
(c) 2001-2008 The world wide DirectFB Open Source Community
(c) 2000-2004 Convergence (integrated media) GmbH
----------------------------------------------------------------


(*) DirectFB/Core: Single Application Core. (2015-03-23 14:57)
(*) Direct/Thread: Started 'VT Switcher' (-1) [CRITICAL OTHER/OTHER 0/0] <8388608>...
(*) Direct/Thread: Started 'Linux Input' (-1) [INPUT OTHER/OTHER 0/0] <8388608>...
(*) DirectFB/Input: AT Raw Set 3 keyboard (1) 0.1 (directfb.org)
(*) Direct/Thread: Started 'Linux Input' (-1) [INPUT OTHER/OTHER 0/0] <8388608>...
(*) DirectFB/Input: PC Speaker (2) 0.1 (directfb.org)
(*) Direct/Thread: Started 'Linux Input' (-1) [INPUT OTHER/OTHER 0/0] <8388608>...
(*) DirectFB/Input: DELL DELL USB Keyboard (3) 0.1 (directfb.org)
(*) Direct/Thread: Started 'Keyboard Input' (-1) [INPUT OTHER/OTHER 0/0] <8388608>...
(*) DirectFB/Input: Keyboard 0.9 (directfb.org)
(*) Direct/Thread: Started 'PS/2 Input' (-1) [INPUT OTHER/OTHER 0/0] <8388608>...
(*) DirectFB/Input: IMPS/2 Mouse 1.0 (directfb.org)
(*) DirectFB/Genefx: MMX detected and enabled
(*) DirectFB/Graphics: S3 Savage4 Series 0.3 (directfb.org)
(*) DirectFB/Core/WM: Default 0.3 (directfb.org)
(*) FBDev/Surface: Allocated 640x400 8 bit LUT8 buffer (index 0) at offset 0 and pitch 640.
(*) FBDev/Mode: Setting 640x400 LUT8
(*) FBDev/Mode: Switched to 640x400 (virtual 640x800) at 8 bit (LUT8), pitch 640
(*) FBDev/Surface: Allocated 640x400 8 bit LUT8 buffer (index 1) at offset 256000 and pitch 640.
(*) FBDev/Mode: Setting 640x400 LUT8
(*) FBDev/Mode: Switched to 640x400 (virtual 640x400) at 8 bit (LUT8), pitch 640
(*) FBDev/Surface: Allocated 640x400 8 bit LUT8 buffer (index 0) at offset 0 and pitch 640.



When I run the code specifying no-software, as follows, I just get a blank screen, so I am assuming hardware acceleration is not working.


-bash-3.1# ./test --dfb:no-software
test application is starting


~~~~~~~~~~~~~~~~~~~~~~~~~~| DirectFB 1.2.9 |~~~~~~~~~~~~~~~~~~~~~~~~~~
(c) 2001-2008 The world wide DirectFB Open Source Community
(c) 2000-2004 Convergence (integrated media) GmbH
----------------------------------------------------------------


(*) DirectFB/Core: Single Application Core. (2015-03-23 14:57)
(*) Direct/Thread: Started 'VT Switcher' (-1) [CRITICAL OTHER/OTHER 0/0] <8388608>...
(*) Direct/Thread: Started 'Linux Input' (-1) [INPUT OTHER/OTHER 0/0] <8388608>...
(*) DirectFB/Input: AT Raw Set 3 keyboard (1) 0.1 (directfb.org)
(*) Direct/Thread: Started 'Linux Input' (-1) [INPUT OTHER/OTHER 0/0] <8388608>...
(*) DirectFB/Input: PC Speaker (2) 0.1 (directfb.org)
(*) Direct/Thread: Started 'Linux Input' (-1) [INPUT OTHER/OTHER 0/0] <8388608>...
(*) DirectFB/Input: DELL DELL USB Keyboard (3) 0.1 (directfb.org)
(*) Direct/Thread: Started 'Keyboard Input' (-1) [INPUT OTHER/OTHER 0/0] <8388608>...
(*) DirectFB/Input: Keyboard 0.9 (directfb.org)
(*) Direct/Thread: Started 'PS/2 Input' (-1) [INPUT OTHER/OTHER 0/0] <8388608>...
(*) DirectFB/Input: IMPS/2 Mouse 1.0 (directfb.org)
(*) DirectFB/Genefx: MMX detected and enabled
(*) DirectFB/Graphics: S3 Savage4 Series 0.3 (directfb.org)
(*) DirectFB/Core/WM: Default 0.3 (directfb.org)
(*) FBDev/Surface: Allocated 640x400 8 bit LUT8 buffer (index 0) at offset 0 and pitch 640.
(*) FBDev/Mode: Setting 640x400 LUT8
(*) FBDev/Mode: Switched to 640x400 (virtual 640x800) at 8 bit (LUT8), pitch 640
(*) FBDev/Surface: Allocated 640x400 8 bit LUT8 buffer (index 1) at offset 256000 and pitch 640.
(*) FBDev/Mode: Setting 640x400 LUT8
(*) FBDev/Mode: Switched to 640x400 (virtual 640x400) at 8 bit (LUT8), pitch 640
(*) FBDev/Surface: Allocated 640x400 8 bit LUT8 buffer (index 0) at offset 0 and pitch 640









/****************************************************
* Name: directfb_test.c
* Author: Tim Nicholson
* Purpose: Just for test the DirectFB
****************************************************/
#include <stdio.h>
#include <unistd.h>
#include <directfb.h>
#include <math.h>


#define HEIGHT 480
#define WIDTH 640
#define WHOLE 150
#define HALF (WHOLE/2.0)
#define MIDX (WIDTH/2.0)
#define MIDY (HEIGHT/2.0)


static struct PointF
{
double x;
double y;
} c1,c2,c3,c4,p1,p2,p3,p4,p5,p6,p7,p8,mid1,mid2;


static int x = 0;
void calc (void);


int main (int argc, char *argv[])
{
IDirectFB *dfb = NULL;
IDirectFBDisplayLayer *layer = NULL;
IDirectFBSurface *surface = NULL;


printf ("test application is starting\n");
DirectFBInit (&argc, &argv);
DirectFBCreate (&dfb);


dfb->GetDisplayLayer (dfb, DLID_PRIMARY, &layer);
layer->SetCooperativeLevel (layer, DLSCL_EXCLUSIVE);
layer->GetSurface (layer, &surface);


mid1.x = MIDX - HALF;
mid1.y = MIDY;
mid2.x = MIDX + HALF;
mid2.y = MIDY;


while (1)
{
calc();
surface->SetColor (surface, 0x00, 0x00, 0x00, 0xff);
surface->FillRectangle (surface, 0, 0, WIDTH, HEIGHT);
surface->SetColor (surface, 0xff, 0xff, 0xff, 0xff);

surface->DrawLine (surface, p1.x, p1.y, p2.x, p2.y);
surface->DrawLine (surface, p2.x, p2.y, p3.x, p3.y);
surface->DrawLine (surface, p3.x, p3.y, p4.x, p4.y);
surface->DrawLine (surface, p4.x, p4.y, p1.x, p1.y);


surface->DrawLine (surface, p5.x, p5.y, p6.x, p6.y);
surface->DrawLine (surface, p6.x, p6.y, p7.x, p7.y);
surface->DrawLine (surface, p7.x, p7.y, p8.x, p8.y);
surface->DrawLine (surface, p8.x, p8.y, p5.x, p5.y);


surface->DrawLine (surface, p1.x, p1.y, p5.x, p5.y);
surface->DrawLine (surface, p2.x, p2.y, p6.x, p6.y);
surface->DrawLine (surface, p3.x, p3.y, p7.x, p7.y);
surface->DrawLine (surface, p4.x, p4.y, p8.x, p8.y);


dfb->WaitForSync(dfb);
}
surface->Release (surface);
layer->Release (layer);
dfb->Release (dfb);
return 0;
}


// *********************************************************************************************
// c a l c
// *********************************************************************************************
void calc (void)
{
int a,b,c,d;


x++;
a = x;
b = x + 90;
c = x + 180;
d = x + 270;


a %= 360;
b %= 360;
c %= 360;
d %= 360;


c1.x = (sin(M_PI / 180.0 * a) * WHOLE);
c1.y = (cos(M_PI / 180.0 * a) * WHOLE);
c2.x = (sin(M_PI / 180.0 * b) * WHOLE);
c2.y = (cos(M_PI / 180.0 * b) * WHOLE);
c3.x = (sin(M_PI / 180.0 * c) * WHOLE);
c3.y = (cos(M_PI / 180.0 * c) * WHOLE);
c4.x = (sin(M_PI / 180.0 * d) * WHOLE);
c4.y = (cos(M_PI / 180.0 * d) * WHOLE);


p1.x = mid1.x + c1.x;
p2.x = mid1.x + c2.x;
p3.x = mid1.x + c3.x;
p4.x = mid1.x + c4.x;


p5.x = mid2.x + c1.x;
p6.x = mid2.x + c2.x;
p7.x = mid2.x + c3.x;
p8.x = mid2.x + c4.x;


p1.y = mid1.y + c1.y;
p2.y = mid1.y + c2.y;
p3.y = mid1.y + c3.y;
p4.y = mid1.y + c4.y;


p5.y = mid2.y + c1.y;
p6.y = mid2.y + c2.y;
p7.y = mid2.y + c3.y;
p8.y = mid2.y + c4.y;
}

Loading...