This option shows help and usage information. If this option is specified, other command-line options are ignored. Calling fromELF without any parameters produces the same help information.

This option produces, to standard output, a list of assembly language EQU directives that equate C++ class or C structure field names to their offsets from the base of the class or structure. The input ELF file can be a relocatable object or an image.

Use -o to redirect the output to a file. Use the INCLUDE command from armasm to load the produced file and provide access to C++ classes and C structure members by name from assembly language. See RealView Compilation Tools for BREW Assembler Guide for more information on armasm.

-fieldoffsets outputs all structure information. To output a subset of the structures, use -select select_options.

If you do not require a file that can be input to armasm, use the -text -a option to format the display addresses in a more readable form. The -a option only outputs address information for structures and static data in images because the addresses are not known in a relocatable object.

Use -select select_option together with either the -fieldoffsets or -text -a options to select only those fields that match the patterns in the option list.

Use special characters to select multiple fields:

If you are using a special character on Unix, you must enclose the options in quotes to prevent the shell expanding the selection.

Use -nolinkview to discard the section-level view (link-time view) from the ELF image and retain only the segment level view (load-time view). Discarding the link-view section level eliminates:

All that is left in the output is the program header table and the program segments. According to the ELF specification, these are all that a program loader can rely upon being present in an ELF file.

This option does not put debug information in the output files. This is the default for binary images. If -nodebug is specified, it affects all output formats. It overrides the -text -g option.

fromelf -nodebug image -o image_nodb.axf

fromelf -elf -nodebug image.axf -o image_ndb.axf

This option displays fromELF version information.

Use a text specification, for example -text -c, to display image information in text format. You can decode an ELF image or ELF object file using this option. This is the default. If no text or code output format is specified, -text is assumed.

If output_file is not specified with the -o option, the information is displayed on stdout.

If a specific text category is not specified, the default is to output header information.

If specified, the text category consists of one or more of the following:

The category selectors can be specified as one of:

If an output format is not specified, the default output format of -text is used and the individual category selectors are recognized. If another output format is specified, the selectors are ignored.

This option selects the binary or ELF output file options. code_output_format can be one of:

If you use fromELF to convert an ELF image containing multiple load regions to a binary format using any of the -bin, -ihf, -m32, -i32, or -vhx options, fromELF creates an output directory named output_file and generates one binary output file for each load region in the input image. fromELF places the output files in the output_file directory.

This option specifies the base address of the output for Motorola S-record, Intel Hex, and Extended Intellec Hex file formats. This option is available only if -m32, -i32, or -ihf is specified as the output format.

You can specify the base address as either:

All addresses encoded in the output file start at the base address n. If you do not specify a -base option, the base address is taken from the load region address.

This option outputs multiple files for multiple memory banks. This option is available only if -vhx or -bin is specified as the output format.

The format of memory_config is -widthxbanks where:

Valid configurations are:

-8x1
-8x2
-8x4
-16x1
-16x2
-32x1
-32x2
-64x1

fromELF uses the last specified configuration if more than one configuration is specified. If this option is selected for any format other than -bin or -vhx it is ignored.

If the image has one load region, fromELF generates bank files with the following naming conventions:

If the image has multiple load regions, fromELF creates a directory named output_file and generates bank files for each load region named load region0 to load region banks-1.

The memory width specified by width controls the size of the chunk of information read from the image and written to a file. The first chunk read is allocated to the first file (output_file0), the next chunk is allocated to the next file. After a chunk is allocated to the last file, allocation begins again with the first file (that is, the allocation is modulo based on the number of files). For example:

For a memory_config of -8x4

byte0 -> file0
byte1 -> file1
byte2 -> file2 
byte3 -> file3 
byte4 -> file0 
...

For a memory_config of -16x2

halfword0 -> file0 
halfword1 -> file1 
halfword3 -> file0 
...

This option specifies the name of the output file, or the name of the output directory if multiple output files are created (see the description of text_output_format and code_output_format for more information). Specifying the output file is optional with the -text output option and mandatory with all other outputs.

This option specifies the ELF file to be translated.

fromELF accepts only ARM-executable ELF files and ARM object ELF files (.o).