mmap on /dev/fd0 on pc686 BSP not working at all.

[Pavel Pisa]

Hello Vijay, Karel and Gedare,

On Monday 21 of September 2020 19:35:26 Vijay Kumar Banerjee wrote:
> Reading the post by Pavel in the other thread, I think the pc386
> shouldn't be using the libbsd framebuffer. To be able to use the
> driver in pc386, you'll have to disable the drivers from
> rtems-littevgl. Please use `--no-drivers` option with ./waf configure
> in rtems-littlevgl and remove the "fbdev_init" from the hello app. The
> initialization of the fb is supposed to be handled by the drivers in
> the BSP.

I have fear that my analysis is not clear and simple enough.
I have tried to provide full background but I strive to
instructions there.

I see as the most viable solution with minimal code changes
to use RTEMS i386 VESA BIOS driver, I am not sure about RSB
option, I have build by configure in the past (my last i386
RTEMS build seems to be 4.12 version)

export USE_VBE_RM=1 \
../../../git/rtems/configure --target=i386-rtems4.12 --prefix=/opt/rtems4.12 \
  --build=x86_64-pc-linux-gnu \
  --enable-rtems-inlines --disable-multiprocessing --enable-cxx \
  --enable-rdbg --enable-maintainer-mode \
  --enable-networking --enable-posix --enable-itron --disable-ada \
  --disable-expada --disable-multilib --disable-docs \
  --enable-rtemsbsp=pc686 \
 

For LVGL, configure/include into build some Linux /dev/fbX driver which
is probably the most matching to RTEMS i386 BSP VESA BIOS driver.
I do not see in original LVGL driver open and mmap of Linux /dev/fb
by fast glimpse into

  git://github.com/littlevgl/lvgl.git

I have not noticed the pointer into your full source to can help there.

But the code should include call to to some IOCTLs like
FBIOGET_FSCREENINFO and FBIOGET_VSCREENINFO (fixed and variable)
and then there is somewhere mmap(), something like this

  psd->addr = mmap(NULL, psd->size, PROT_READ|PROT_WRITE, MAP_SHARED, fb, 0);

Instead of calling mmap() you assign directly address returned to the
framebuffer start

  psd->addr = fb_fix.smem_start

So unfortunately, understanding to my analysis is really 180 degrees
different direction than its results. robably my mistake, wording, ....

Best wishes,

Pavel