Hi, for the first time I was able to read the values of a controller configuration register and the values match the default values described in the data sheet.
The trick was to use the new feature of the panel-mipi-debug driver to choose to send out “generic packets” if instructed.
And I did finally understand how the mapping of 16 bit SPI register numbers to two MIPI commands works.
This sequence:
echo gff128301 >dcs echo g0080 >dcs echo gff1283 >dcs echo g0080 >dcs echo gb0rrrrr >dcs
shows this result in dmesg:
[ 8268.330870] dsi: mipi_debug_read(g,b0, 00 01 00 30 04) -> 0
and the data sheet says that 00 01 00 30 04 are indeed the default values of the “MIPISET1 (B080h)” register.
And - maybe something I had tried wrongly - the video stream must be turned off while trying to access these registers.
So now I think I can translate the SPI command sequence we have received as C code into a shell script and we will see if the panel becomes better initialized and does no longer show the flicker effect.
Then it is quite simple to translate this into (different than the original) C code for implanting into our kernel driver.
BR, Nikolaus
Am 12.01.2015 um 14:42 schrieb Dr. H. Nikolaus Schaller hns@goldelico.com:
Hi, for the first time I was able to read the values of a controller configuration register and the values match the default values described in the data sheet.
The trick was to use the new feature of the panel-mipi-debug driver to choose to send out “generic packets” if instructed.
And I did finally understand how the mapping of 16 bit SPI register numbers to two MIPI commands works.
This sequence:
echo gff128301 >dcs echo g0080 >dcs echo gff1283 >dcs echo g0080 >dcs echo gb0rrrrr >dcs
shows this result in dmesg:
[ 8268.330870] dsi: mipi_debug_read(g,b0, 00 01 00 30 04) -> 0
and the data sheet says that 00 01 00 30 04 are indeed the default values of the “MIPISET1 (B080h)” register.
And - maybe something I had tried wrongly - the video stream must be turned off while trying to access these registers.
So now I think I can translate the SPI command sequence we have received as C code into a shell script and we will see if the panel becomes better initialized and does no longer show the flicker effect.
Indeed! Flicker is gone! Hurrah :)
http://git.goldelico.com/?p=gta04-kernel.git;a=blob;f=GTA04/root/boe-w677l;h...
It appears that the strategy to sharpen the tools (panel-mipi-debug driver) first was the right one.
Then it is quite simple to translate this into (different than the original) C code for implanting into our kernel driver.
I think I will invest this time as well so that we then can close this topic and focus on the SSD2858 driver.
BTW: I assume that I will receive the new Mainboard PCBs tomorrow and can then start populating with chips and have them soldered by end of the week.
If there are no significant bugs we will then have 3 complete sets. Each one comes with a panel adapter board with SSD2858 and one without (i.e. no chip installed but signals bridged). This will allow for comparisons.
BR, Nikolaus
Am 12.01.2015 um 15:05 schrieb Dr. H. Nikolaus Schaller hns@goldelico.com:
Am 12.01.2015 um 14:42 schrieb Dr. H. Nikolaus Schaller hns@goldelico.com:
Hi, for the first time I was able to read the values of a controller configuration register and the values match the default values described in the data sheet.
The trick was to use the new feature of the panel-mipi-debug driver to choose to send out “generic packets” if instructed.
And I did finally understand how the mapping of 16 bit SPI register numbers to two MIPI commands works.
This sequence:
echo gff128301 >dcs echo g0080 >dcs echo gff1283 >dcs echo g0080 >dcs echo gb0rrrrr >dcs
shows this result in dmesg:
[ 8268.330870] dsi: mipi_debug_read(g,b0, 00 01 00 30 04) -> 0
and the data sheet says that 00 01 00 30 04 are indeed the default values of the “MIPISET1 (B080h)” register.
And - maybe something I had tried wrongly - the video stream must be turned off while trying to access these registers.
So now I think I can translate the SPI command sequence we have received as C code into a shell script and we will see if the panel becomes better initialized and does no longer show the flicker effect.
Indeed! Flicker is gone! Hurrah :)
http://git.goldelico.com/?p=gta04-kernel.git;a=blob;f=GTA04/root/boe-w677l;h...
It appears that the strategy to sharpen the tools (panel-mipi-debug driver) first was the right one.
Then it is quite simple to translate this into (different than the original) C code for implanting into our kernel driver.
I think I will invest this time as well so that we then can close this topic and focus on the SSD2858 driver.
Done: http://git.goldelico.com/?p=gta04-kernel.git;a=commit;h=dde7e067fb88ecf4e864...
BTW: I assume that I will receive the new Mainboard PCBs tomorrow and can then start populating with chips and have them soldered by end of the week.
If there are no significant bugs we will then have 3 complete sets. Each one comes with a panel adapter board with SSD2858 and one without (i.e. no chip installed but signals bridged). This will allow for comparisons.
BR, Nikolaus _______________________________________________ Kernel mailing list Kernel@pyra-handheld.com http://pyra-handheld.com/cgi-bin/mailman/listinfo/kernel