this is an atempt to make the export of functions typesafe.
I replaced the jumptable void ** by a struct (jt_funcs) with function pointers.
The EXPORT_FUNC macro now has 3 fixed parameters and one
variadic parameter
The first is the name of the exported function,
the rest of the parameters are used to format a functionpointer
in the jumptable,

the EXPORT_FUNC macros are expanded three times,
1. to declare the members of the struct
2. to initialize the structmember pointers
3. to call the functions in stubs.c

Signed-off-by: Martin Dorwig <dorwig@tetronik.com>
Acked-by: Simon Glass <sjg@chromium.org>

Signed-off-by: Simon Glass <sjg@chromium.org>
(resending to the list since my tweaks are not quite trivial)

Introduce a gd->hose to save the pci hose in the early phase so that
apis in drivers/pci/pci.c can be used before relocation. Architecture
codes need assign a valid gd->hose in the early phase.

Some variables are declared as static so change them to be either
stack variable or global data member so that they can be used before
relocation, except the 'indent' used by CONFIG_PCI_SCAN_SHOW which
just affects some print format.

Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Acked-by: Simon Glass <sjg@chromium.org>

Conflicts:
	drivers/serial/serial-uclass.c

Signed-off-by: Tom Rini <trini@ti.com>

The simple malloc() implementation is used when memory is tight. It provides
a simple buffer with an incrementing pointer.

At present the implementation is inside dlmalloc. Move it into its own file
so that it is easier to find.

Rather than using relocation as a signal that the full malloc() is
available, add a special GD_FLG_FULL_MALLOC_INIT flag. This signals that the
simple malloc() should no longer be used.

In some cases, such as SPL, even the code space used by the full malloc() is
wasteful. Add a CONFIG_SYS_MALLOC_SIMPLE option to provide only the simple
malloc. In this case the full malloc is not available at all. It saves about
1KB of code space and about 0.5KB of data on Thumb 2.

Acked-by: Tom Rini <trini@ti.com>
Signed-off-by: Simon Glass <sjg@chromium.org>

In general we can't store things in the data section until we have inited
SDRAM. Some platforms allow this (e.g. those with SPL) but some don't. Move
the pointer to global_data so that it will work on all platforms.

Without this fix the serial port will not work prior to relocation with
driver model on some platforms.

Signed-off-by: Simon Glass <sjg@chromium.org>

For sandbox we have a fallback console which is used very early in
U-Boot, before serial drivers are available. Rather than try to guess
when to switch to the real console, add a flag so we can be sure. This
makes sure that sandbox can always output a panic() message, for example,
and avoids silent failure (which is very annoying in sandbox).

Signed-off-by: Simon Glass <sjg@chromium.org>

Initialise devices marked 'pre-reloc' and make them available prior to
relocation. Note that this requires pre-reloc malloc() to be available.

Signed-off-by: Simon Glass <sjg@chromium.org>

If we are to have driver model before relocation we need to support some
way of calling memory allocation routines.

The standard malloc() is pretty complicated:

1. It uses some BSS memory for its state, and BSS is not available before
relocation

2. It supports algorithms for reducing memory fragmentation and improving
performace of free(). Before relocation we could happily just not support
free().

3. It includes about 4KB of code (Thumb 2) and 1KB of data. However since
this has been loaded anyway this is not really a problem.

The simplest way to support pre-relocation malloc() is to reserve an area
of memory and allocate it in increasing blocks as needed. This
implementation does this.

To enable it, you need to define the size of the malloc() pool as described
in the README. It will be located above the pre-relocation stack on
supported architectures.

Note that this implementation is only useful on machines which have some
memory available before dram_init() is called - this includes those that
do no DRAM init (like tegra) and those that do it in SPL (quite a few
boards). Enabling driver model preior to relocation for the rest of the
boards is left for a later exercise.

Signed-off-by: Simon Glass <sjg@chromium.org>

using UBI and DM together leads in compiler error, as
both define a "struct device", so rename "struct device"
in include/dm/device.h to "struct udevice", as we use
linux code (MTD/UBI/UBIFS some USB code,...) and cannot
change the linux "struct device"

Signed-off-by: Heiko Schocher <hs@denx.de>
Cc: Simon Glass <sjg@chromium.org>
Cc: Marek Vasut <marex@denx.de>

This driver needs a data structure in SRAM before SDRAM is available.
This is not alway the case using .data section. Moving this data
structure to global_data guarantees it is writable.

Signed-off-by: York Sun <yorksun@freescale.com>
CC: Troy Kisky <troy.kisky@boundarydevices.com>