All of lore.kernel.org
 help / color / mirror / Atom feed
* [Buildroot] frame buffer problems
@ 2007-04-06 16:14 nina.holly at aerende.com
  2007-04-06 16:39 ` Bernhard Fischer
  2007-04-06 17:05 ` Luis R. Rodriguez
  0 siblings, 2 replies; 8+ messages in thread
From: nina.holly at aerende.com @ 2007-04-06 16:14 UTC (permalink / raw)
  To: buildroot

Hi,
    I am using Linux 2.6.10 with devfs on ARM.  I know that
devfs is deprecated/obsolete.  The framebuffer works fine
with buildroot svn version 15763.  But when I switched to
buildroot version 17224 I get a return value < 0 when I try to execute

ioctl(frame_buffer_ptr, FBIOGET_VSCREENINFO, &var_info);

I.E. I can't read the basic frame buffer information.

1.  Is this a problem in the version of buildroot that I am using?

2.  Is this because buildroot version 17224 doesn't work
    with devfs and I need to switch to using udev?

3.  Do I need to use Linux version greater than 2.6.15 (which works
    with udev) to work with buildroot version 17224?

Thanks in advance for any pointers.

Nina

^ permalink raw reply	[flat|nested] 8+ messages in thread

* [Buildroot] frame buffer problems
  2007-04-06 16:14 [Buildroot] frame buffer problems nina.holly at aerende.com
@ 2007-04-06 16:39 ` Bernhard Fischer
  2007-04-06 17:05 ` Luis R. Rodriguez
  1 sibling, 0 replies; 8+ messages in thread
From: Bernhard Fischer @ 2007-04-06 16:39 UTC (permalink / raw)
  To: buildroot

On Fri, Apr 06, 2007 at 09:14:33AM -0700, nina.holly at aerende.com wrote:
>Hi,
>    I am using Linux 2.6.10 with devfs on ARM.  I know that
>devfs is deprecated/obsolete.  The framebuffer works fine
>with buildroot svn version 15763.  But when I switched to
>buildroot version 17224 I get a return value < 0 when I try to execute
>
>ioctl(frame_buffer_ptr, FBIOGET_VSCREENINFO, &var_info);
>
>I.E. I can't read the basic frame buffer information.
>
>1.  Is this a problem in the version of buildroot that I am using?
>
>2.  Is this because buildroot version 17224 doesn't work
>    with devfs and I need to switch to using udev?
>
>3.  Do I need to use Linux version greater than 2.6.15 (which works
>    with udev) to work with buildroot version 17224?

I'm using 2.6.20.4 with binutils-2.17.50.0.14 and gcc-4.2. I have
BR2_PACKAGE_BUSYBOX_CONFIG="busybox.config"
BR2_PACKAGE_BUSYBOX_HIDE_OTHERS=y
BR2_PACKAGE_BUSYBOX_SKELETON=y
and use mdev (from buysbox trunk) to create my device nodes, dunno
about udev and don't want to know about devfs anymore ;)
I do use sysfs and configfs, though, the former is required for mdev to
work.

Works fine for me, fwiw.

^ permalink raw reply	[flat|nested] 8+ messages in thread

* [Buildroot] frame buffer problems
  2007-04-06 16:14 [Buildroot] frame buffer problems nina.holly at aerende.com
  2007-04-06 16:39 ` Bernhard Fischer
@ 2007-04-06 17:05 ` Luis R. Rodriguez
  2007-04-07 13:31   ` nina.holly at aerende.com
  1 sibling, 1 reply; 8+ messages in thread
From: Luis R. Rodriguez @ 2007-04-06 17:05 UTC (permalink / raw)
  To: buildroot

On 4/6/07, nina.holly at aerende.com <nina.holly@aerende.com> wrote:
> Hi,
>     I am using Linux 2.6.10 with devfs on ARM.  I know that
> devfs is deprecated/obsolete.  The framebuffer works fine
> with buildroot svn version 15763.  But when I switched to
> buildroot version 17224 I get a return value < 0 when I try to execute
>
> ioctl(frame_buffer_ptr, FBIOGET_VSCREENINFO, &var_info);

Was the return value -14 or -19?

  Luis

^ permalink raw reply	[flat|nested] 8+ messages in thread

* [Buildroot] frame buffer problems
  2007-04-06 17:05 ` Luis R. Rodriguez
@ 2007-04-07 13:31   ` nina.holly at aerende.com
  2007-04-10  7:00     ` Peter Korsgaard
  0 siblings, 1 reply; 8+ messages in thread
From: nina.holly at aerende.com @ 2007-04-07 13:31 UTC (permalink / raw)
  To: buildroot

The frame buffer test program I am using does the following:

A.  Open /dev/fb/0 with fb_ptr = open("/dev/fb/0", O_RDWR);
B.  ioctl(frame_buffer_ptr, FBIOGET_FSCREENINFO, &fix_info);
C.  ioctl(frame_buffer_ptr, FBIOGET_VSCREENINFO, &var_info);
D.  fbp = (char *)mmap(0, screensize, READ | WRITE, MAP_SHARED, fb_ptr, 0);


1.  If the kernel is compiled using buildroot version 17224
    and run in a buildroot version 17224 root filesystem it
    is failing with errno = -1 on

    errno = ioctl(frame_buffer_ptr, FBIOGET_FSCREENINFO, &fix_info);

2.  If the kernel is compiled using buildroot version 15763 but
    run in a buildroot 17224 root filesystem, the failure occurs
    when trying to map the framebuffer to memory using mmap with
    a return value of -1.


Nina




> On 4/6/07, nina.holly at aerende.com <nina.holly@aerende.com> wrote:
>> Hi,
>>     I am using Linux 2.6.10 with devfs on ARM.  I know that
>> devfs is deprecated/obsolete.  The framebuffer works fine
>> with buildroot svn version 15763.  But when I switched to
>> buildroot version 17224 I get a return value < 0 when I try to execute
>>
>> ioctl(frame_buffer_ptr, FBIOGET_VSCREENINFO, &var_info);
>
> Was the return value -14 or -19?
>
>   Luis
>

^ permalink raw reply	[flat|nested] 8+ messages in thread

* [Buildroot] frame buffer problems
  2007-04-07 13:31   ` nina.holly at aerende.com
@ 2007-04-10  7:00     ` Peter Korsgaard
  2007-04-12 15:00       ` nina.holly at aerende.com
  0 siblings, 1 reply; 8+ messages in thread
From: Peter Korsgaard @ 2007-04-10  7:00 UTC (permalink / raw)
  To: buildroot

>>>>> "nh" == nina holly <nina.holly@aerende.com> writes:

Hi,

nh> The frame buffer test program I am using does the following: A.
nh> Open /dev/fb/0 with fb_ptr = open("/dev/fb/0", O_RDWR);

Do you have a /dev/fb/0? What is it's major/minor numbers?
What is the return value of open?

nh> ioctl(frame_buffer_ptr, FBIOGET_FSCREENINFO, &fix_info); C.
nh> ioctl(frame_buffer_ptr, FBIOGET_VSCREENINFO, &var_info);

And of the ioctls?

-- 
Bye, Peter Korsgaard

^ permalink raw reply	[flat|nested] 8+ messages in thread

* [Buildroot] frame buffer problems
  2007-04-10  7:00     ` Peter Korsgaard
@ 2007-04-12 15:00       ` nina.holly at aerende.com
  2007-04-12 15:08         ` Peter Korsgaard
  0 siblings, 1 reply; 8+ messages in thread
From: nina.holly at aerende.com @ 2007-04-12 15:00 UTC (permalink / raw)
  To: buildroot

>>>>>> "nh" == nina holly <nina.holly@aerende.com> writes:
>
> Hi,
>
> nh> The frame buffer test program I am using does the following: A.
> nh> Open /dev/fb/0 with fb_ptr = open("/dev/fb/0", O_RDWR);
>
> Do you have a /dev/fb/0?

Yes, I do have a /dev/fb/0.

>What is it's major/minor numbers?

% ls -l /dev/fb/0
crw-rw-rw- 1 root root 29, 0 Dec 31 17:00 /dev/fb/0

> What is the return value of open?

Return value of open = -1

>
> nh> ioctl(frame_buffer_ptr, FBIOGET_FSCREENINFO, &fix_info); C.
> nh> ioctl(frame_buffer_ptr, FBIOGET_VSCREENINFO, &var_info);
>
> And of the ioctls?

Return value of ioctl for fixed info    = -1
Return value of ioctl for variable info = -1


I have another frame buffer program that draws a checkerboard
on the screen.  On the same buildroot version and board the
return values from open and the ioctls are not less than zero,
but the program crashes with a segmentation fault when pixels
are written to the screen buffer.  The return values are

Return value of open = 3
Return value of ioctl for fixed info    = 0
Return value of ioctl for variable info = 0



Thanks,

Nina

>
> --
> Bye, Peter Korsgaard
>

^ permalink raw reply	[flat|nested] 8+ messages in thread

* [Buildroot] frame buffer problems
  2007-04-12 15:00       ` nina.holly at aerende.com
@ 2007-04-12 15:08         ` Peter Korsgaard
  2007-04-13 14:45           ` nina.holly at aerende.com
  0 siblings, 1 reply; 8+ messages in thread
From: Peter Korsgaard @ 2007-04-12 15:08 UTC (permalink / raw)
  To: buildroot

>>>>> "nina" == nina holly <nina.holly@aerende.com> writes:

>> What is the return value of open?

nina> Return value of open = -1

And errno is?

nina> I have another frame buffer program that draws a checkerboard on
nina> the screen.  On the same buildroot version and board the return
nina> values from open and the ioctls are not less than zero, but the
nina> program crashes with a segmentation fault when pixels are
nina> written to the screen buffer.

Odd. Exact same binaries, or recompiled with the new uclibc?

-- 
Bye, Peter Korsgaard

^ permalink raw reply	[flat|nested] 8+ messages in thread

* [Buildroot] frame buffer problems
  2007-04-12 15:08         ` Peter Korsgaard
@ 2007-04-13 14:45           ` nina.holly at aerende.com
  0 siblings, 0 replies; 8+ messages in thread
From: nina.holly at aerende.com @ 2007-04-13 14:45 UTC (permalink / raw)
  To: buildroot

>>>>>> "nina" == nina holly <nina.holly@aerende.com> writes:
>
>>> What is the return value of open?
>
> nina> Return value of open = -1
>
> And errno is?

The strerror(errno) = "Success".

The open had returned -1 because the link from /dev/fb0 to /dev/fb/0
was gone.  Once the programs both used /dev/fb/0 their return and
errno values were the same which are listed below (except for mmap errno,
the first program returned "Success" even though the return value was
-1):

return of open = 3
errno of open = Success
return of ioctl(fixed info) = 0
errno of ioctl(fixed info) = Success
return of ioctl(var info) = 0
errno of ioctl(var info) = Success
return of mmap = -1
errno of mmap = bad address

So it looks like both programs have trouble using "mmap" on
the framebuffer /dev/fb/0 for buildroot version 17224 whereas
these programs work fine with buildroot version 15763.

These programs when run on the earlier version of buildroot 15763
that works have the following return values and errnos.

return of open = 3
errno of open = Success
return of ioctl(fixed info) = 0
errno of ioctl(fixed info) = Success
return of ioctl(var info) = 0
errno of ioctl(var info) = Success
return of mmap = 1074180096
errno of mmap = Success

TIA,

Nina


>
> nina> I have another frame buffer program that draws a checkerboard on
nina> the screen.  On the same buildroot version and board the return
nina> values from open and the ioctls are not less than zero, but the
nina> program crashes with a segmentation fault when pixels are
> nina> written to the screen buffer.
>
> Odd. Exact same binaries, or recompiled with the new uclibc?
>
> --
> Bye, Peter Korsgaard
>

^ permalink raw reply	[flat|nested] 8+ messages in thread

end of thread, other threads:[~2007-04-13 14:45 UTC | newest]

Thread overview: 8+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2007-04-06 16:14 [Buildroot] frame buffer problems nina.holly at aerende.com
2007-04-06 16:39 ` Bernhard Fischer
2007-04-06 17:05 ` Luis R. Rodriguez
2007-04-07 13:31   ` nina.holly at aerende.com
2007-04-10  7:00     ` Peter Korsgaard
2007-04-12 15:00       ` nina.holly at aerende.com
2007-04-12 15:08         ` Peter Korsgaard
2007-04-13 14:45           ` nina.holly at aerende.com

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.