Hey folks. Its been a while. Finally got back on the kernel dev train. This *mostly* seems to work, but I've ran into an issue which may require some upstream support. Basic gist, I was attempting to boot a buildroot-built intramfs, was mostly working, but it would not give me a login prompt no matter how much I badgered it, so I decided to try a flash boot. That also failed, and after much banging my head on the desk and annoying people on irc, we finally came across *why* it was failing, or at least part of it. The kernel parser could produce the mtdblockN partitions based on the devicetree, but for whatever reason, when the hisilicon,fmc-spi-nor read its superblock, instead of the magic 0x73717368 (sqsh), it read back 0x73717360. At first I thought it was a mistake of mine, as we had tried modifying the header, and 0x60 is a `, so it could be a hex editing issue, but nope. Reading the data in u-boot (2010.06, vendor fork, would like to get mainline running on it at some point) showed the correct magic, so its something with the controller driver, I guess. CC'ing the people associated with that as well, hopefully we can get to the bottom of this. Marty E. Plummer (2): clk: hisilicon: add CRG driver Hi3521a SoC arm: hisi: enable Hi3521a soc arch/arm/boot/dts/Makefile | 2 + arch/arm/mach-hisi/Kconfig | 9 ++ drivers/clk/hisilicon/Kconfig | 8 ++ drivers/clk/hisilicon/Makefile | 1 + drivers/clk/hisilicon/crg-hi3521a.c | 141 ++++++++++++++++++++++ include/dt-bindings/clock/hi3521a-clock.h | 34 ++++++ 6 files changed, 195 insertions(+) create mode 100644 drivers/clk/hisilicon/crg-hi3521a.c create mode 100644 include/dt-bindings/clock/hi3521a-clock.h -- 2.35.1
