Add Ritmix RZX-50 and RZX-27

[SiarheiVolkau]

RZX-50 - JZ4755 based handheld with 64Mb SDRAM
RZX-27 - JZ4725b based handheld with 32Mb SDRAM

These patches allows to initialize SDRAM and load & execute OpenDingux as well as original firmwares.

Also these patches unify firmware header creation and MMC boot order.

The firmware headers are separated to NAND headers and MMC headers.

In contrast to JZ4760/70 JZ4725 have slightly different "MSPL" header.

For JZ4760+ header structure is:

• MBR (512 bytes)
• MSPL header (4 bytes)
• entry point
• rest of binary

For JZ4725b it is:

• MSPL header (4b)
• entry point
• MBR (aligned to 512b boundary)
• rest of binary

JZ4755 have no specific header, but entry point is at beginning of firmware:

• entry point
• MBR (aligned to 512b boundary)
• rest of binary

So, for those 2 SoCs UBIBoot header also includes space for MBR into binary: -DHAVE_MBR option.

Regarding MMC boot order, MMC_ID and MMC_ID2 are replaced by
MMC_IDS and MMC_BLKIDS arrays

• the first defines SoCs MSC(x) in boot order.
• the second defines which /dev/mmcblk(y) corresponds the MSC(x) in MMC_IDS.

[pcercuei]

Ah, sorry! I missed your PR!

[pcercuei]

I reworked your MMC refactor, please have a look: 7728d2d

For your RZX-27 and RXZ-50 boards, I'm annoyed that a lot of code is duplicated (PLL init code, memory init code). This code is SoC-specific and not board-specific, so it should be common to these two.

If the code ends up similar to what's already in board-lepus.c (although the SoC is different), then it would be a good idea to share as much code as possible.

[SiarheiVolkau]

Hi @pcercuei ,

I reworked your MMC refactor

Looks good, but it missed one important thing for me - easy changing the boot order: its useful during development to boot from external MMC first and fall back to internal then and finally try original firmware on the internal MMC slot. Also this algo is applicable for normal usage, what do you think?

it would be a good idea to share as much code as possible

Indeed, although these two are closer to rs90 than to lepus, I'll try to unify PLL and SDRAM initialization for all three in next round.

missed your PR!

Please have a look at my other activities: 1, 2.

[pcercuei]

@SiarheiVolkau with your original MMC refactor commit ubiboot would "try all MMCs in order", I did not change that.

We could make it try to boot from external SDs first, yes, I think it makes sense.

[SiarheiVolkau]

with your original MMC refactor commit ubiboot would "try all MMCs in order", I did not change that.

But you don't use the MMC_BLKIDS which is crucial to that feature.
Imagine, we're changing boot order to { 2, 0 } to boot from external first, but the root parameter passed to the kernel will be mmcblk0p1 for MSC2 and mmcblk1p1 for MSC0, which is completely incorrect.