[PATCH 4/4] commands: version: Add framebuffer output support

Andrey Smirnov andrew.smirnov at gmail.com
Fri Feb 16 05:39:08 PST 2018


On Thu, Feb 8, 2018 at 11:29 PM, Sascha Hauer <s.hauer at pengutronix.de> wrote:
> Hi Andrey,
>
> On Mon, Feb 05, 2018 at 09:29:35AM -0800, Andrey Smirnov wrote:
>> Extend "version" command to be capable of rendeing version information
>> on framebuffer devices.
>>
>> Signed-off-by: Andrey Smirnov <andrew.smirnov at gmail.com>
>> ---
>>  commands/Kconfig   |   8 +++
>>  commands/version.c | 198 ++++++++++++++++++++++++++++++++++++++++++++++++++++-
>>  2 files changed, 204 insertions(+), 2 deletions(-)
>>
>> diff --git a/commands/Kconfig b/commands/Kconfig
>> index 095536849..1cad5d608 100644
>> --- a/commands/Kconfig
>> +++ b/commands/Kconfig
>> @@ -238,6 +238,14 @@ config CMD_VERSION
>>
>>         barebox 2014.05.0-00142-gb289373 #177 Mon May 12 20:35:55 CEST 2014
>>
>> +config CMD_VERSION_OUTPUT_TO_FRAMEBUFFER
>> +       bool
>> +       depends on CMD_VERSION
>> +       prompt "support displaying version via framebuffer"
>> +       help
>> +         Selecting this option will enable version command to output
>> +         version information onto display backed by a frambuffer
>> +
>>  config CMD_MMC_EXTCSD
>>       tristate
>>       prompt "read/write eMMC ext. CSD register"
>> diff --git a/commands/version.c b/commands/version.c
>> index 090f2dd13..ba74a993c 100644
>> --- a/commands/version.c
>> +++ b/commands/version.c
>> @@ -20,16 +20,210 @@
>>  #include <common.h>
>>  #include <command.h>
>>  #include <complete.h>
>> +#include <getopt.h>
>> +#include <fb.h>
>> +#include <gui/graphic_utils.h>
>> +#include <gui/2d-primitives.h>
>> +
>> +#define FRAME_SIZE   8
>> +
>> +struct placement {
>> +     int x, y;
>> +};
>> +
>> +typedef struct placement placement_function_t (struct screen *, int, int);
>> +
>> +struct placement
>> +placement_center(struct screen *sc, int text_width, int text_height)
>> +{
>> +     struct placement p;
>> +
>> +     p.x = (sc->info->xres - text_width) / 2;
>> +     p.y = (sc->info->yres - text_height) / 2;
>> +
>> +     return p;
>> +}
>> +
>> +struct placement
>> +placement_upper_left(struct screen *sc, int text_width, int text_height)
>> +{
>> +     struct placement p;
>> +
>> +     p.x = FRAME_SIZE;
>> +     p.y = FRAME_SIZE;
>> +
>> +     return p;
>> +}
>> +
>> +struct placement
>> +placement_upper_right(struct screen *sc, int text_width, int text_height)
>> +{
>> +     struct placement p;
>> +
>> +     p.x = sc->info->xres - 1 - text_width - FRAME_SIZE;
>> +     p.y = FRAME_SIZE;
>> +
>> +     return p;
>> +}
>> +
>> +struct placement
>> +placement_lower_left(struct screen *sc, int text_width, int text_height)
>> +{
>> +     struct placement p;
>> +
>> +     p.x = FRAME_SIZE;
>> +     p.y = sc->info->yres - 1 - text_height - FRAME_SIZE;
>> +
>> +     return p;
>> +}
>> +
>> +struct placement
>> +placement_lower_right(struct screen *sc, int text_width, int text_height)
>> +{
>> +     struct placement p;
>> +
>> +     p.x = sc->info->xres - 1 - text_width  - FRAME_SIZE;
>> +     p.y = sc->info->yres - 1 - text_height - FRAME_SIZE;
>> +
>> +     return p;
>> +}
>>
>>  static int do_version(int argc, char *argv[])
>>  {
>> -     printf ("\n%s\n", version_string);
>> +     const struct font_desc *font;
>> +     const char *fbdev = NULL;
>> +     struct screen *sc;
>> +     u8 fg[3], bg[3];
>> +     int text_width;
>> +     int text_height;
>> +
>> +     placement_function_t *place;
>> +     struct placement placement;
>> +
>> +     if (IS_ENABLED(CONFIG_CMD_VERSION_OUTPUT_TO_FRAMEBUFFER)) {
>
> This is quite much code added to the version command, given that the
> command previously only was a one-liner. I think this should rather be a
> separate command. This could depend on framebuffer and other needed
> stuff which would make the first two patches unnecessary. I think
> stubbing away graphics functions when the user is graphics-only code is
> a bit over the top.
>

OK, will change in v2.

> The placement functions look as if they should be moved to some font
> handling generic code.

OK, will do.

>
> Also it's always nice to have a C API for anything we also have as
> command. It has proven to be useful so many times.
>

Sure, will do.

>> +BAREBOX_CMD_HELP_OPT ("-f color\t", "foreground color, in hex RRGGBB format")
>> +BAREBOX_CMD_HELP_OPT ("-b color\t", "background color, in hex RRGGBB format")
>
> I'm not sure how useful these are as parameters to a command. How about
> a nv variable? This could be used by the framebuffer console aswell.
>

Sure, sounds reasonable.

Thanks,
Andrey Smirnov



More information about the barebox mailing list