]> git.baikalelectronics.ru Git - arm-tf.git/commit
rockchip/rk3399: fixes the typo and the WARNINGS during suspend/resume
authorCaesar Wang <wxt@rock-chips.com>
Mon, 19 Jun 2017 06:02:52 +0000 (14:02 +0800)
committerCaesar Wang <wxt@rock-chips.com>
Fri, 30 Jun 2017 00:46:20 +0000 (08:46 +0800)
commitc3710ee7a2360c946a3e06e471a6205279729145
tree610c27642ee9ac9f31350a395cd7505cdb34ce96
parentaa5b843fe8f2d8cea80fd1c06e7fc6b7c18f265c
rockchip/rk3399: fixes the typo and the WARNINGS during suspend/resume

This patch fixes the two things as follows:

1) rk3399_flash_l2_b" seems to be a typo. That's "flush", not "flash".

2) fixes the warnings log.
We always hit the warnings thing during the suspend, as below log:
..
[   51.022334] CPU5: shutdown
[   51.025069] psci: CPU5 killed.
INFO:    sdram_params->ddr_freq = 928000000
WARNING: rk3399_flash_l2_b:reg 28830380,wait

When the L2 completes the clean and invalidate sequence, it asserts the
L2FLUSHDONE signal. The SoC can now deassert L2FLUSHREQ signal and then
the L2 deasserts L2FLUSHDONE.

Then, a loop without a delay isn't really great to measure time. We should
probably add a udelay(10) or so in there and then maybe replace the WARN()
after the loop. In the actual tests, the L2 cache will take ~4ms by
default for big cluster.

In the real world that give 10ms for the enough margin, like the
ddr/cpu/cci frequency and other factors that will affect it.

Change-Id: I55788c897be232bf72e8c7b0e10cf9b06f7aa50d
Signed-off-by: Caesar Wang <wxt@rock-chips.com>
plat/rockchip/rk3399/drivers/pmu/pmu.c