There are some issues with CONFIG_MMU=n in MIPS configuration
(e.g. qemu-malta_defconfig):

  * "Cannot request SDRAM region for stack" error message on start;
  * empty kseg1_ram0 region in the iomem command output.

The patch fixes the problems by forcing KSEG1 memory region usage
for CONFIG_MMU=n MIPS configurations.

Signed-off-by: Antony Pavlov <antonynpavlov@gmail.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

.bss __rel_start (OVERLAY) was used to optimize RAM size used by
barebox. Since .bss and __rel_start overlap, we should clear bss only
after __rel_start was used. Following patch changed initialization
sequence to clear .bss before __rel_start:
1e5aef61fc6a444 ("MIPS: reloc: init bss and cpu")

This make relocation buffer unusable and broke relocation with different
symptoms reported by Antony Pavlov <antonynpavlov@gmail.com>:
  * iomem output has no information on sdram regions, so memtest is unusable;
  * pc = 0xa081232c, relocation does not work, barebox is located with 8M offset
    from start of RAM. The board has 256M and relocation routine
    should move barebox code much higher;
  * pc = 0xa081232c, so barebox code works from KSEG1 not from KSEG0 as MMU=y option implies.

There is a choice of moving .bss clear sequence after __rel_start or
remove this optimization. Since the use of this optimization is minimal
and danger to trap in to similar issue is still high, i prefer to remove
this optimization.

As result of this change, we need to fix calculation of relocation are
size:
relocate_code() calculates relocation area size as a sum (barebox_image_size +
bss_len). barebox_image_size and bss_len are calculated as (__image_end -
__image_start) and (__bss_stop - __bss_start) respectively. This doesn't take
into account relocation data placed between  __image_end and __bss_start.
However relocation preserves BSS position relative to image start, as if
relocation data is still there. This causes RAM overflow during BSS
initialization in main_entry(). This problem may be hidden due to the alignment
of the `relocaddr`.

Reported-by: Antony Pavlov <antonynpavlov@gmail.com>
Fixes: 1e5aef61fc6a444 ("MIPS: reloc: init bss and cpu")
Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de>
Signed-off-by: Peter Mamonov <pmamonov@gmail.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

We need it to make flush_cache_all() work properly.

Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

It will be done in main_entry() any way.

Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

In this relocate_code() piece 'length' is greater than 'barebox_image_size':

    #define MAX_BSS_SIZE SZ_1M
    ...
    length = barebox_image_size + MAX_BSS_SIZE;
    relocaddr = ALIGN_DOWN(ram_size - barebox_image_size, SZ_64K);
    ...
    memcpy((void *)relocaddr, __image_start, length);

so 'ram_size' overflow occurs during memcpy().

Signed-off-by: Antony Pavlov <antonynpavlov@gmail.com>
Tested-by: Oleksij Rempel <o.rempel@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

this patch i a port of following patch from u-boot with some additional
integration changes and fixes of original code:
| Subject: [PATCH] MIPS: Stop building position independent code
|
| U-Boot has up until now built with -fpic for the MIPS architecture,
| producing position independent code which uses indirection through a
| global offset table, making relocation fairly straightforward as it
| simply involves patching up GOT entries.
|
| Using -fpic does however have some downsides. The biggest of these is
| that generated code is bloated in various ways. For example, function
| calls are indirected through the GOT & the t9 register:
|
|   8f998064   lw     t9,-32668(gp)
|   0320f809   jalr   t9
|
| Without -fpic the call is simply:
|
|   0f803f01   jal    be00fc04 <puts>
|
| This is more compact & faster (due to the lack of the load & the
| dependency the jump has on its result). It is also easier to read &
| debug because the disassembly shows what function is being called,
| rather than just an offset from gp which would then have to be looked up
| in the ELF to discover the target function.
|
| Another disadvantage of -fpic is that each function begins with a
| sequence to calculate the value of the gp register, for example:
|
|   3c1c0004   lui    gp,0x4
|   279c3384   addiu  gp,gp,13188
|   0399e021   addu   gp,gp,t9
|
| Without using -fpic this sequence no longer appears at the start of each
| function, reducing code size considerably.
|
| This patch switches U-Boot from building with -fpic to building with
| -fno-pic, in order to gain the benefits described above. The cost of
| this is an extra step during the build process to extract relocation
| data from the ELF & write it into a new .rel section in a compact
| format, plus the added complexity of dealing with multiple types of
| relocation rather than the single type that applied to the GOT. The
| benefit is smaller, cleaner, more debuggable code. The relocate_code()
| function is reimplemented in C to handle the new relocation scheme,
| which also makes it easier to read & debug.
|
| Taking maltael_defconfig as an example the size of u-boot.bin built
| using the Codescape MIPS 2016.05-06 toolchain (gcc 4.9.2, binutils
| 2.24.90) shrinks from 254KiB to 224KiB.
|
| Signed-off-by: Paul Burton <paul.burton@imgtec.com>

Signed-off-by: Oleksij Rempel <linux@rempel-privat.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>