commit | a1b7f3521f66a19a2d4c9a6a373c5a7ab36e1473 | [log] [tgz] |
---|---|---|
author | Nico Huber <nico.h@gmx.de> | Mon Mar 25 18:32:11 2024 +0100 |
committer | Nico Huber <nico.h@gmx.de> | Mon Jul 22 10:08:47 2024 +0000 |
tree | fd996296810ab45fe99d29d8dc254f6d496f3091 | |
parent | 008a44fa1c33b8a77c90b4e9dba267ae23c01056 [diff] |
dediprog: Implement multi-i/o reads This implements i/o-mode switches and opcode handling for multi-i/o reads with protocol versions 2 and 3. The mode switching is done by a simple command that takes an enum just as our internal `io_mode` as argument. The opcode handling differs between protocol versions. For protocol v2, we keep the current behavior for single-i/o operations and only set the matching opcode. Tests with an SF600Plus-G2 have shown that the programmer automatically chooses the address length and number of dummy cycles. It is unknown, however, if it chooses these para- meters based on the opcode or the configured i/o mode. For dual-out reads, it seems to choose the wrong number of dummy cycles. Hence, we mask the respective support bit for the v2 case. For protocol v3, a new `read mode' was discovered in traces of the Dediprog Windows application. It allows to explicitly specify the opcode, the address length, and the number of dummy cycles. We call this READ_MODE_CONFIGURABLE. As this is the only way to make use of the additional command bytes of the v3 protocol, we can assume that this mode always works with v3. For partial reads, i.e. not multiples of 512B blocks, that have to go through dediprog_spi_send_command(), we temporarily disable the chosen `.spi_fast_read` function. This is necessary, because multi- io is not supported on this path. We enable dual i/o by default for protocol v3 devices. This should work out of the box with many compatible flash chips. The command- line logic is a little convoluted this way, but can be refactored once protocol v2 devices are tested. Change-Id: Ib07b1b61eccc19c7ead9f64c980b37feabfa70a8 Signed-off-by: Nico Huber <nico.h@gmx.de> Reviewed-on: https://review.sourcearcade.org/c/flashprog/+/114 Reviewed-by: Arthur Heymans <arthur@aheymans.xyz>
flashprog is a utility for detecting, reading, writing, verifying and erasing flash chips. It is often used to flash BIOS/EFI/coreboot/firmware images in-system using a supported mainboard, but it also supports flashing of network cards (NICs), SATA controller cards, and other external devices which can program flash chips.
It supports a wide range of flash chips (most commonly found in SOIC8, DIP8, SOIC16, WSON8, PLCC32, DIP32, TSOP32, and TSOP40 packages), which use various protocols such as LPC, FWH, parallel flash, or SPI.
Do not use flashprog on laptops (yet)! The embedded controller (EC) present in many laptops might interact badly with any attempts to communicate with the flash chip and may brick your laptop.
Please make a backup of your flash chip before writing to it.
Please see the flashprog(8) manpage.
To package flashprog and remove dependencies on Git, either use make export or make tarball
'make export' will export all flashprog files from the Git repository at revision HEAD into a directory named "$EXPORTDIR/flashprog-$RELEASENAME" and will additionally add a "versioninfo.inc" file in that directory to contain the Git revision of the exported tree and a date for the manual page.
'make tarball' will simply tar up the result of make export and compress it with bzip2.
The snapshot tarballs are the result of 'make tarball' and require no further processing.
To build flashprog you need to install the following software:
Linux et al:
On FreeBSD, you need the following ports:
On OpenBSD, you need the following ports:
To compile on Linux, use:
make
To compile on FreeBSD, OpenBSD or DragonFly BSD, use:
gmake
To compile on Nexenta, use:
make
To compile on Solaris, use:
gmake LDFLAGS="-L$pathtolibpci" CC="gcc -I$pathtopciheaders" CFLAGS=-O2
To compile on NetBSD (with pciutils, libftdi, libusb installed in /usr/pkg/), use:
gmake
To compile and run on Darwin/Mac OS X:
Install DirectHW from coresystems GmbH. DirectHW is available at http://www.coreboot.org/DirectHW .
To cross-compile on Linux for DOS:
Get packages of the DJGPP cross compiler and install them: djgpp-filesystem djgpp-gcc djgpp-cpp djgpp-runtime djgpp-binutils
As an alternative, the DJGPP web site offers packages for download as well: djcross-binutils-2.29.1-1ap.x86_64.rpm djcross-gcc-7.2.0-1ap.x86_64.rpm djcrx-2.05-5.x86_64.rpm
The cross toolchain packages for your distribution may have slightly different names (look for packages named djgpp).
Alternatively, you could use a script to build it from scratch: https://github.com/andrewwutw/build-djgpp
You will need the libpci and libgetopt library source trees and their compiled static libraries and header files installed in some directory say libpci-libgetopt/, which will be later specified with LIBS_BASE parameter during flashprog compilation. Easiest way to handle it is to put pciutils, libgetopt and flashprog directories in one subdirectory. There will be an extra subdirectory libpci-libgetopt created, which will contain compiled libpci and libgetopt.
Download pciutils 3.5.6 and apply http://flashprog.org/File:Pciutils-3.5.6.patch.gz Compile pciutils, using following command line:
make ZLIB=no DNS=no HOST=i386-djgpp-djgpp CROSS_COMPILE=i586-pc-msdosdjgpp- \ PREFIX=/ DESTDIR=$PWD/../libpci-libgetopt \ STRIP="--strip-program=i586-pc-msdosdjgpp-strip -s" install install-lib
Download and compile with 'make' http://flashprog.org/File:Libgetopt.tar.gz
Copy the libgetopt.a to ../libpci-libgetopt/lib and getopt.h to ../libpci-libgetopt/include
Enter the flashprog directory.
make CC=i586-pc-msdosdjgpp-gcc STRIP=i586-pc-msdosdjgpp-strip
LIBS_BASE=../libpci-libgetopt/ HAS_LIBPCI=yes CONFIG_LIBPCI_LDFLAGS=-lpci
strip
If you like, you can compress the resulting executable with UPX:
upx -9 flashprog.exe
To run flashprog.exe, download http://flashprog.org/File:Csdpmi7b.zip and unpack CWSDPMI.EXE into the current directory or one in PATH.
To cross-compile on Linux for Windows:
Get packages of the MinGW cross compiler and install them: mingw32-filesystem mingw32-cross-cpp mingw32-cross-binutils mingw32-cross-gcc mingw32-runtime mingw32-headers
The cross toolchain packages for your distribution may have slightly different names (look for packages named mingw). PCI-based programmers (internal etc.) are not supported on Windows.
Run (change CC= and STRIP= settings where appropriate) make CC=i686-w64-mingw32-gcc STRIP=i686-w64-mingw32-strip
Processor architecture dependent features:
On non-x86 architectures a few programmers don't work (yet) because they use port-based I/O which is not directly available on non-x86. Those programmers will be disabled automatically if you run "make".
Compiler quirks:
If you are using clang and if you want to enable only one driver, you may hit an overzealous compiler warning from clang. Compile with "make WARNERROR=no" to force it to continue and enjoy.
In order to install flashprog and the manpage into /usr/local, type:
make install
For installation in a different directory use DESTDIR, e.g. like this:
make DESTDIR=/usr install
If you have insufficient permissions for the destination directory, use sudo by adding sudo in front of the commands above.
main
branchp
, such as p1.0
.x
suffix, for example 1.0.x
v
, for example v1.0
or v1.0.1
-rcN
, for example v1.0-rc1
The official flashprog website is:
Available contact methods are