Hi Notaz,
Am 24.06.2018 um 16:57 schrieb Grazvydas Ignotas notasas@gmail.com:
Hi,
ED: your suggestion might be good, but hard to implement.
Ok got it to work with some horrible hacks to reprogram bq24297 earlier (good luck to anyone trying to mainline that...). Binaries (for 4GB RAM, includes Tony's fixes): http://notaz.gp2x.de/misc/pyra/u-boot/ Source: https://dev.pyra-handheld.com/notaz/pyra-u-boot (not posting patches to the list unless requested to, as they're hacks, and that git already has more of such pyra-only patches anyway).
I have cherry-picked some, but there are also interesting patches which lack a Signed-off line in the commit message. This makes it impossible to work on them or at some time submit upstream (e.g. by Patrick).
Most interesting is "LC15: mask some irqs so that connecting USB doesn't power on"
How can we get them signed-off by the author (sometimes you sometimes Matthijs)?
BR, Nikolaus
Gražvydas
On Sun, Jun 24, 2018 at 5:46 PM, Michael Mrozek EvilDragon@openpandora.org wrote:
On So, 2018-06-24 at 17:01 +0300, Grazvydas Ignotas wrote:
So try to change board/pandora/pyra+lc15/pyra+lc15.c line 51 #if defined(CONFIG_SPL_BUILD) {TIMER8_PWM_EVT, (PTD | IEN | M6)}, /* GPIO8_230: keyboard backlight - pulled low */ #else {TIMER8_PWM_EVT, (PTU | IEN | M6)}, /* GPIO8_230: keyboard backlight - pulled high */ #endif
I've tried this, now it loads u-boot successfully but dies before it reaches the point where it reprograms bq24297, just after printing "DRAM: 4 GiB". It seems to be dying in initr_malloc(), no idea why. A wild guess is that code in malloc init utilizes more CPU pipelines and RAM (as it's clearing memory) which uses more power, and that goes over the current limit.
Hm. Could the CPU power usage be lowered (keep it down at 500MHz with 1 CPU core only, etc.) and upped after the limit has been raised?
Or is there anything else we could do to lower the power usage on startup - I mean, even limited, we've got 5V * 0.5A which is 2.5W... and we only need to keep it lower than that until the limit has been reached.
Hm, it it hard to enable high power for the bq24297 in MLO? I mean, if we recompile the MLO anyways, why switch off the backlight instead of enabling high power mode (if that fixes it).
I've tried this too and it doesn't work (i2c errors), probably the needed i2c bus is not properly initialized yet in MLO.
I guess adding i2c to MLO is no simple task...
-- Greetings,
Michael Mrozek
OpenPandora GmbH CEO: Michael Mrozek
Schäffbräustr. 11 85049 Ingolstadt Germany Tel.: +49 841 990 5548 http://www.openpandora.de/ HRB 4879, Amtsgericht Ingolstadt
eMail: mrozek@openpandora.org