[Buildroot] read-edid package - clarifications

[Buildroot] read-edid package - clarifications

[Dimitry Golubovsky]

Hi,

Thomas Petazzoni wrote:

> On Ubuntu, parse-edid is in usr/bin, get-edid is in usr/sbin. See
> http://packages.ubuntu.com/quantal/i386/read-edid/filelist

There is another package in my queue for submission which is v86d, and
this program is about to be used at about the same stage of boot as
get-edid. Ubuntu has it in /sbin per
http://packages.ubuntu.com/quantal/i386/v86d/filelist

> I don't feel strongly about this, if you prefer to keep them in bin/, it's
> fine. On embedded systems, /usr is most likely on the same partition
> as /, so it doesn't make much difference.

OK, perhaps I'll place all three (get-edid, parse-edid, v86d) in /sbin.

Baruch Siach wrote:

> read-edid seems to work perfectly on my x86_64 laptop. So keeping BR2_x86_64
> should be safe.

OK, if you tried what I submitted then I'll add x86_64 (I just do not
have a hardware to test it such way).

> As for PowerPC, I guess the low level code is selected at build time according
> to the target platform. I haven't looked at the code, though.

Well, this one uses real-mode BIOS call to VBE ROM via lrmi
(http://lrmi.sf.net - not sure if it supports powerpc). There is
another version of monitor info dump program in existence which uses
i2c which may work on something other than PC. Not the one I am
submitting.

Thanks.

-- 
Dimitry Golubovsky

Anywhere on the Web

[Buildroot] read-edid package - clarifications

[Baruch Siach]

Hi Dimitry,

On Thu, May 03, 2012 at 09:27:05AM -0400, Dimitry Golubovsky wrote:
> Baruch Siach wrote:
> 
> > read-edid seems to work perfectly on my x86_64 laptop. So keeping BR2_x86_64
> > should be safe.
> 
> OK, if you tried what I submitted then I'll add x86_64 (I just do not
> have a hardware to test it such way).

No I haven't tested your patch. I just run 'get-edid | parse-edid' from the 
read-edid package of my distro (Debian).

baruch

-- 
     http://baruch.siach.name/blog/                  ~. .~   Tk Open Systems
=}------------------------------------------------ooO--U--Ooo------------{=
   - baruch at tkos.co.il - tel: +972.2.679.5364, http://www.tkos.co.il -

[Buildroot] read-edid package - clarifications

[Dimitry Golubovsky]

Baruch,

On Thu, May 3, 2012 at 2:22 PM, Baruch Siach <baruch@tkos.co.il> wrote:
>
> No I haven't tested your patch. I just run 'get-edid | parse-edid' from the
> read-edid package of my distro (Debian).

Per http://packages.debian.org/search?keywords=read-edid&searchon=names&suite=all&section=all

I believe your read-edid is of version 2.0.x but mine is 1.4.x

I could not get through with 2.0.0 compilation for some reason -
messed up headers in liblrmi - but 1.4.x compiled and worked fine.

1.4.x is only for i386 and powerpc (which I obviously could not test).

Thanks for clarification.

I'll try to submit a refined patch later tonight.

-- 
Dimitry Golubovsky

Anywhere on the Web

[Buildroot] read-edid package - clarifications

[Baruch Siach]

Hi Dimitry,

On Thu, May 03, 2012 at 06:32:35AM -0400, Dimitry Golubovsky wrote:
> Baruch's comment:
> 
> > read-edid can only run on x86 PCs and PowerPC Macs, so it should depend on
> > (BR2_i386 || BR2_x86_64 || BR2_powerpc).
> 
> I only tested it on i386, not sure about x86-64, and definitely not on
> powerpc as it has x86 real-mode stuff in it. So perhaps I'll keep i386
> only.

read-edid seems to work perfectly on my x86_64 laptop. So keeping BR2_x86_64 
should be safe.

As for PowerPC, I guess the low level code is selected at build time according 
to the target platform. I haven't looked at the code, though.

baruch

-- 
     http://baruch.siach.name/blog/                  ~. .~   Tk Open Systems
=}------------------------------------------------ooO--U--Ooo------------{=
   - baruch at tkos.co.il - tel: +972.2.679.5364, http://www.tkos.co.il -

[Buildroot] read-edid package - clarifications

[Thomas Petazzoni]

Le Thu, 3 May 2012 06:32:35 -0400,
Dimitry Golubovsky <golubovsky@gmail.com> a ?crit :

> > non-essential binaries should probably live in usr/bin/ rather than
> > bin/.
> 
> I think such program may be needed rather early in the boot process.
> Basically it is needed when setting up the framebuffer console
> resolution. I have a real case of Via Unichrome and a Dell widescreen
> monitor where the framebuffer driver is unable to sense the monitor
> (X.org driver does fine though), so the only way for me to initialize
> the console was to run read-edid even before I am attempting to load
> the framebuffer drivers and fbcon. That's why I placed it into /bin
> (or maybe it is even better in /sbin).

On Ubuntu, parse-edid is in usr/bin, get-edid is in usr/sbin. See
http://packages.ubuntu.com/quantal/i386/read-edid/filelist. I don't
feel strongly about this, if you prefer to keep them in bin/, it's
fine. On embedded systems, /usr is most likely on the same partition
as /, so it doesn't make much difference.

Thomas
-- 
Thomas Petazzoni, Free Electrons
Kernel, drivers, real-time and embedded Linux
development, consulting, training and support.
http://free-electrons.com

[Buildroot] read-edid package - clarifications

[Dimitry Golubovsky]

Hi,

Thanks to Baruch and Thomas for the comments. I'll try to implement them.

Few thoughts though:

Thomas' comment:

> non-essential binaries should probably live in usr/bin/ rather than bin/.

I think such program may be needed rather early in the boot process.
Basically it is needed when setting up the framebuffer console
resolution. I have a real case of Via Unichrome and a Dell widescreen
monitor where the framebuffer driver is unable to sense the monitor
(X.org driver does fine though), so the only way for me to initialize
the console was to run read-edid even before I am attempting to load
the framebuffer drivers and fbcon. That's why I placed it into /bin
(or maybe it is even better in /sbin).

Baruch's comment:

> read-edid can only run on x86 PCs and PowerPC Macs, so it should depend on
> (BR2_i386 || BR2_x86_64 || BR2_powerpc).

I only tested it on i386, not sure about x86-64, and definitely not on
powerpc as it has x86 real-mode stuff in it. So perhaps I'll keep i386
only.

Thanks.

-- 
Dimitry Golubovsky

Anywhere on the Web