FIT: Let user specify the configuration to use ...

The images in FIT images can be opened in two different ways. They
can be either opened directly based on their names in the images/
node or as part of a configuration based on their names in the
corresponding /configuration/ node.

So far we only supported the latter. To prepare supporting the former
we return a cookie belonging to the configuration from
fit_open_configuration() which we use in fit_open_image() to refer
to the desired configuration.

While at it document fit_open_configuration().

Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

FIT: Do not pre-open images ...

Only do what fit_open_configuration() suggests: open the configuration,
but not the images in it.

Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

bootm: FIT: do not depend on FIT pre-opened images ...

When calling fit_open_configuration the FIT code already opens
the images "kernel", "ramdisk" and "dtb". This does not fit well
into the FIT code, so make the bootm code independent of these
pre-opened images so that we can drop the opening from the FIT
code in the next step.

Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

bootchooser: Drop SHELL_NONE conflict ...

Earlier versions of the bootchooser code executed scripts,
thus had to depend on shell support. We do not need this anymore, so
the dependency can be dropped.

Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

boot: Make independent of command support ...

The boot code works with command support disabled if we skip
support for boot scripts. Bail out before parsing boot scripts
and allow to build with command support disabled.

Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

startup: Load default environment earlier ...

Some files compiled into the default environment may be needed earlier in
the boot process, so move loading of the default environment to fs_initcall.
Only the default environment is loaded earlier, but not the overwriting with
the persistent environment files, so when the files are used that early, only
the compiled in version is available; it's not overwritable at that
time.

Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

state: Add 'init_from_defaults' parameter ...

The init_from_defaults parameter allows to detect if a state has been
initialized from default values, i.e. state_load failed.

Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

state: refuse to set dirty parameter via setenv ...

The dirty parameter is readonly, so refuse to set it.

Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

filetype: Add type detection at an offset ...

Add file_name_detect_type_offset to allow file system detection for loop
mounts at an offset.

Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

imd: Make all pointers into image const ...

The IMD code should be readonly and never modify any pointers. Make
all pointers const so that const pointers can be passed in to IMD.

Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

ratp: remove unused context fields ...

Signed-off-by: Aleksander Morgado <aleksander@aleksander.es>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

ratp: move barebox-specific command to ratp_bb.h ...

The ratp_run_command() is really an implementation detail of the
barebox ratp 'command' operation support. Move it to the barebox
specific header and rename it with the correct prefix.

Signed-off-by: Aleksander Morgado <aleksander@aleksander.es>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

ratp: rename global context variable ...

The 'ratp_command_ctx' seems to specify that this variable is only
used to process command operations, but it really is used in every
async operation (e.g. also in FS), so just rename it to a more generic
name for clarity.

Signed-off-by: Aleksander Morgado <aleksander@aleksander.es>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

bootargs: Fix memory leak in linux_bootargs_get() ...

globalvar_get_match() returns an allocated string which may be
empty. We have to free() the empty string aswell.

Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

uimage: Use is_tftp_fs() and cache_file() to ease TFTP workaround ...

We have helper functions now to ease file caching when a file
is on TFTP. Use them.

Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

uimage: fix memory leak in error path ...

handle->name is dynamically allocated, so free it in the error path.

Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Reported-by: Stefan Lengfeld <contact@stefanchrist.eu>

startup: create /tmp ...

Some commands create temporary files in /. Create /tmp to offer
these commands an appropriate place for storing temporary files.

Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

Merge branch 'for-next/misc'

Merge branch 'for-next/fastboot-sparse'

Merge branch 'for-next/build'

blspec: fix crash when trying to boot from loop mounts ...

When trying to boot from a loop mount path, the blspec code passes a
NULL pointer to dev_name, because cdev->dev is not set for loop mounts:

  unable to handle NULL pointer dereference at address 0x00000020
  pc : [<4fe09e28>]    lr : [<4fe02305>]
  sp : 4ffef7c0  ip : 31b75c78  fp : 31b6a388
  r10: 31b6a368  r9 : 315a4cac  r8 : 00000001
  r7 : 00000001  r6 : 311b8540  r5 : ffffffea  r4 : 31b6a3f8
  r3 : 00000000  r2 : 10000000  r1 : 00000001  r0 : 00000000
  Flags: nzcv  IRQs off  FIQs off  Mode SVC_32
  [<4fe09e28>] (dev_id+0xc/0x38) from [<4fe02305>] (blspec_scan_directory+0x3e1/0x484)
  [<4fe02305>] (blspec_scan_directory+0x3e1/0x484) from [<4fe02551>] (blspec_bootentry_provider+0x1d/0x3c)
  [<4fe02551>] (blspec_bootentry_provider+0x1d/0x3c) from [<4fe0984b>] (bootentry_create_from_name+0x23/0xdc)
  [<4fe0984b>] (bootentry_create_from_name+0x23/0xdc) from [<4fe246f9>] (do_boot+0x95/0x160)
  [<4fe246f9>] (do_boot+0x95/0x160) from [<4fe032e9>] (execute_command+0x21/0x48)
  [<4fe032e9>] (execute_command+0x21/0x48) from [<4fe088e3>] (run_list_real+0x56b/0x634)
  [<4fe088e3>] (run_list_real+0x56b/0x634) from [<4fe08249>] (parse_stream_outer+0xd9/0x164)
  [<4fe08249>] (parse_stream_outer+0xd9/0x164) from [<4fe08b6d>] (run_shell+0x31/0x60)
  [<4fe08b6d>] (run_shell+0x31/0x60) from [<4fe032e9>] (execute_command+0x21/0x48)
  [<4fe032e9>] (execute_command+0x21/0x48) from [<4fe088e3>] (run_list_real+0x56b/0x634)
  [<4fe088e3>] (run_list_real+0x56b/0x634) from [<4fe085fd>] (run_list_real+0x285/0x634)
  [<4fe444b9>] (unwind_backtrace+0x1/0x58) from [<4fe00b7d>] (panic+0x1d/0x34)
  [<4fe00b7d>] (panic+0x1d/0x34) from [<4fe42a75>] (do_exception+0xd/0x10)
  [<4fe42a75>] (do_exception+0xd/0x10) from [<4fe42ad5>] (do_data_abort+0x21/0x2c)
  [<4fe42ad5>] (do_data_abort+0x21/0x2c) from [<4fe42394>] (do_abort_6+0x48/0x54)

Fix this by checking cdev->dev as well as cdev before calling dev_name.

Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

partitions: dos: Treat all extended partition types equally ...

Creating logical DOS partitions with fdisk (Fedora 27, fdisk from
util-linux 2.30.2) results in extended partition of type 0x05 being
created on the device. Partitioned like this, device is succesfully
recognized by Linux, but, due to algorithm in dos_partition(), not
Barebox.

Looking closer at the actual contents of MBR shows that while marked
as "Extended partition with CHS addressing" that partition still have
all of the LBA adressed fields filled correctly.

Given the above and the fact that similar code in Linux
kernel (block/partitions/msdos.c) does not make a distinction between
types 0x0f and 0x05, port Linux's is_extended_partition() and convert
the code to treat types 0x05, 0x0f and 0x85 the same way.

Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

FIT: export fit_open_configuration() and fit_open_image() ...

Currently only fit_open() is exported which only opens the predefined
images "kernel", "dtb" and "ramdisk". To make the FIT code more usable
for other code which may want to open other images export
fit_open_configuration() and fit_open_image().

Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

FIT: Fix error path ...

In case of error of_unflatten_dtb() returns an ERR_PTR. Make sure
that handle->root contains NULL in this case so that we do not call
of_delete_node on the error pointer in the exit path.

Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

file_list: Add ubi flag ...

This flag is added to better support flashing UBI image with Android
sparse images. Android fastboot splits images which are bigger than the
free memory into multiple images which are then transferred in multiple
fastboot sessions. In a session which is not the first one we can no
longer detect if an image is a UBI image or not, so we need a flag to
make this explicit.

Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

filetype: Add fastboot sparse format detection ...

Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

of: base: use root_node compatible as suggestion for a hostname ...

on some SoCs we can use generic PLL and RAM initialization. In this
cases we create board file only to provide a host name.
With this patch host name will be created from device tree compatible.
For example:
 compatible = "board_vendor,board", "chip_vendor,soc"
the host name will be:
 "board"

This function will not overwrite a host name which is already set by
board or machine code.

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

ubiformat: Add ubiformat write function ...

The ubiformat C API expects an image file as argument. With upcoming
Android fastboot sparse image support we can no longer provide a
complete image anymore. With this patch we can write an image in
multiple chunks after we've formatted a MTD device with ubiformat.

ubiformat_write will skip the EC header in both the MTD device we
write to and also the image we read from, so we can flash an image
on a MTD device containing EC headers already.

Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

ubiformat: Lower message printing frequency ...

During scanning and formatting printing the messages on a serial line
has significant overhead up to the point where it takes longer to print
the messages than it takes to do the work. Lower the message printing
frequency to improve performance.

Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

device parameters: Allow dots in device names for setenv aswell ...

Since

f5d77d80f5 Allow device parameters for devices with dots in name

Devices can have dots in their name and we can still read its
device parameters. Do the same change for writing parameters so
they can be written aswell.

Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>