GitLab

This is not just for fun. We hit a real bug on BSD with the outl macros.
The macro variable tmp collided with the tmp from outer scope.

second revision, now also taking care of inb/inw/inl. While that shadowing
did not introduce bugs (yet), of course it breaks the build on BSD when
-Wshadow is enabled.

Corresponding to flashrom svn r860.
Signed-off-by: Michael Karcher <flashrom@mkarcher.dialup.fu-berlin.de>
Acked-by: Luc Verhaegen <libv@skynet.be>

.features was always regenerated in that case, triggering whole-project
recompiles even if nothing changed.

Corresponding to flashrom svn r844.
Signed-off-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>
Acked-by: Sean Nelson <audiohacked@gmail.com>

Corresponding to flashrom svn r838.
Signed-off-by: Jonathan A. Kollasch <jakllsch@kollasch.net>
Acked-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>

It also changes so -VV will spit out highly verbose messages for
debugging. This is a minimal patch to lessen impact a later patch will
convert current printf messages to the new interface.

Corresponding to flashrom svn r835.
Signed-off-by: Sean Nelson <audiohacked@gmail.com>
Acked-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>

And rename them accordingly. For now, main() just calls cli_classic(),
but alternative frontends can be switched in main().
Annotate remaining help texts with CLI dependency inside flashrom.c with
a FIXME comment.

Now people can go and create different frontends and be happy. Please
note that any other frontend will have to sort of duplicate the probing
code in cli_classic.c. Refactoring that part of the code is possible,
but not easy because we still want to print instructive help messages
for users.

Corresponding to flashrom svn r833.
Signed-off-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>
Acked-by: Sean Nelson <audiohacked@gmail.com>

Some programmers (most notably FT2232SPI) use fallback_* and noop_*,
but those functions lived inside internal.c and were unavailable if no
PCI-based programmers were compiled in.

Move those functions to the new file programmer.c. Thanks to Patrick
Georgi for finding this.

Corresponding to flashrom svn r829.
Signed-off-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>
Acked-by: Patrick Georgi <patrick.georgi@coresystems.de>

It just splits $(OBJS) in Makefile into separate lists for Programmer,
Chip, and CLI related files/objects. This should help later on figuring
out where files may go for a libflashrom library.

Corresponding to flashrom svn r824.
Signed-off-by: Sean Nelson <audiohacked@gmail.com>
Acked-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>

Libpci is no longer required to build flashrom and will not be checked
for if no PCI code is needed for the selected programmers.
libftdi is no longer checked for if FT2232 support is not selected.

With this patch, it is possible to build on pretty much every OS out
there (including Windows) without altering the Makefile.
Some gcc versions may need a CFLAGS override for a warning in
dummyflasher.c, though.

Corresponding to flashrom svn r823.
Signed-off-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>
Acked-by: Idwer Vollering <vidwer@gmail.com>

Move that function to udelay.c and compile internal.c only if really
needed. physmap.c is only needed if the programmer is internal or a PCI
card. Make its compilation conditional.

Corresponding to flashrom svn r822.
Signed-off-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>
Acked-by: Idwer Vollering <vidwer@gmail.com>

Make various pieces of code conditional on support for internal
programming. Code shared between PCI device programmers and onboard
programming is now conditional as well.

It is now possible to build only with dummy support:
make CONFIG_INTERNAL=no CONFIG_NIC3COM=no CONFIG_SATASII=no
CONFIG_DRKAISER=no CONFIG_SERPROG=no CONFIG_FT2232SPI=no

This allows building for a specific use case only, and it also
facilitates porting to a new architecture because it is possible to
focus on highlevel code only.

Note: Either internal or dummy programmer needs to be compiled in due to
the current behaviour of always picking a default programmer if -p is
not specified. Picking an arbitrary external programmer as default  
wouldn't make sense.

Build and runtime tested in all 1024 possible build combinations. The
only failures are by design as mentioned above.

Corresponding to flashrom svn r797.
Signed-off-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>
Acked-by: Sean Nelson <audiohacked@gmail.com>

Corresponding to flashrom svn r773.
Signed-off-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>
Acked-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>

The code should work on Linux/*BSD/MacOSX and relies on the serial code
implementation in serial.c. Support for additional platforms (Windows)
will have to be added to serial.c for this to work. For tests without a
Bus Pirate (or with non-functional serial code) it is possible to
#define FAKE_COMMUNICATION in buspirate_spi.c.
Thanks to Sean Nelson for the SPI mode settings code. I tweaked it a bit
to make configuration from a commandline easier should anybody want that
feature.
Tested-by: Sean Nelson <audiohacked@gmail.com>

Corresponding to flashrom svn r772.
Signed-off-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>
Acked-by: Stefan Reinauer <stepan@coresystems.de>
Acked-by: Sean Nelson <audiohacked@gmail.com>

This is the first step in enabling platform independent serprog and it
also allows other drivers to use serial port functionality without
requiring serprog.

Pure code move, no code changed.

Corresponding to flashrom svn r771.
Signed-off-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>
Acked-by: Uwe Hermann <uwe@hermann-uwe.de>

Spotted by Benjamin BELLEC <b.bellec@gmail.com>

Corresponding to flashrom svn r765.
Signed-off-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>
Acked-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>

Retrieve the proper linker flags for libftdi via pkg-config and fall
back if pkg-config isn't available or if it doesn't know libftdi.

Fix $LIBS and $FEATURE_LIBS to honor dependency order.

The original patch is from Jörg, it has been updated by Carl-Daniel to
work on the current tree and to have a fallback in case pkg-config is
not available or not working.

Corresponding to flashrom svn r762.
Signed-off-by: Jörg Mayer <jmayer@loplof.de>
Acked-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>
Acked-by: Jörg Mayer <jmayer@loplof.de>

So far, the only case where libz is needed is when a library (libpci)
pulls in libz and even then it only happens if libpci is available in a
static version only and said static version has libz requirements.

Check for libpci separately and don't require libz if it isn't needed.

Clarify the README.

Corresponding to flashrom svn r743.
Signed-off-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>
Acked-by: Stefan Reinauer <stepan@coresystems.de>

This needs more testing and investigation (partly timing related, it
seems). Reads did work in multiple cases, though.

Corresponding to flashrom svn r740.
Signed-off-by: Uwe Hermann <uwe@hermann-uwe.de>
Acked-by: Uwe Hermann <uwe@hermann-uwe.de>

The new option is '-p gfxnvidia', rest of the interface is as usual.

I tested a successful identify and read on a "RIVA TNT2 Model 64/Model 64 Pro"
card for now, erase and write did NOT work properly so far!

Please do not attempt to write/erase cards yet, unless you can recover!

In addition to the NVIDIA handling code it was required to call
programmer_shutdown() in a lot more places, otherwise the graphics card
will be disabled in the init function, but never enabled again as the
shutdown function is not called.
The shutdown handling may be changed to use atexit() later.

Corresponding to flashrom svn r737.
Signed-off-by: Uwe Hermann <uwe@hermann-uwe.de>
Acked-by: Luc Verhaegen <libv@skynet.be>

If you want support for a particular piece of hardware, just fill in
a few functions in spi_bitbang_master_table. That's it.
On top of this, the RayeR SPI flasher should be supportable in ~20 LOC.

Tested, trace looks OK.

Corresponding to flashrom svn r736.
Signed-off-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>
Acked-by: Uwe Hermann <uwe@hermann-uwe.de>

This is useful for libflashrom (you don't need wiki output in a coreboot
payload).

Wiki output is now disabled by default. If you want to enable it, run
make CONFIG_PRINT_WIKI=yes

Corresponding to flashrom svn r725.
Signed-off-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>
Acked-by: Stefan Reinauer <stepan@coresystems.de>

Example make commandline if you want only internal programmers:
make CONFIG_FT2232SPI=no CONFIG_SERPROG=no CONFIG_NIC3COM=no
CONFIG_SATASII=no CONFIG_DRKAISER=no CONFIG_DUMMY=no

Of course, all of the CONFIG_* symbols can be mixed and matched as
needed. CONFIG_FT2232SPI is special because even if it is enabled, make
will check if the headers are available and skip it otherwise.

Corresponding to flashrom svn r724.
Signed-off-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>
Acked-by: Stefan Reinauer <stepan@coresystems.de>

Fall back to svn info if svnversion fails.

Corresponding to flashrom svn r716.
Signed-off-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>
Acked-by: Uwe Hermann <uwe@hermann-uwe.de>

The vendor sold different designs under that name, the patch works with
the one that has an Actel FPGA as PCI-to-Flash bridge.

The Flash chip is a "Macronix MX29F001B" (128 KB, parallel) soldered
directly to the PCB.
Flash operations (PROBE, READ, ERASE, WRITE) work as expected.

Corresponding to flashrom svn r712.
Signed-off-by: TURBO J <turboj@gmx.de>
Acked-by: Uwe Hermann <uwe@hermann-uwe.de>

Please refer to the release notes for a high-level overview of all the
amazing changes and added features since 0.9.0.

Corresponding to flashrom svn r709.
Signed-off-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>
Acked-by: Stefan Reinauer <stepan@coresystems.de>
Acked-by: Uwe Hermann <uwe@hermann-uwe.de>

With this patch, make tarball and make export still work as expected,
but if you specify RELEASENAME=foo, then the directories and tarballs
are named flashrom-foo instead of flashrom-0.9.0-r703. This makes
release creation a lot easier. As an example, look at creating the 0.9.1
tarball: # make tarball RELEASENAME=0.9.1

Corresponding to flashrom svn r704.
Signed-off-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>
Acked-by: Stefan Reinauer <stepan@coresystems.de>

Corresponding to flashrom svn r686.
Signed-off-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>
Acked-by: Stefan Reinauer <stepan@coresystems.de>

If CONFIG_SERPROG is not set, no stubs and no data of serprog will
remain.

Side benefit: This kills a few dozen lines of code.

r678, r679 and r680 made this possible. Once "Only list available
programers in usage()" is committed, even the usage message will be
adjusted automatically.

Corresponding to flashrom svn r681.
Signed-off-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>
Acked-by: Stefan Reinauer <stepan@coresystems.de>

Serprog compilation is now controlled by a Makefile variable.
Replace munmap with physunmap where appropriate.

Corresponding to flashrom svn r671.
Signed-off-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>
Acked-by: Stefan Reinauer <stepan@coresystems.de>

Set the user and group of all files to root for tar versions which
support it. Add explanatory comments for supporting that feature with
other tar versions.

Use LC_ALL instead of LANG everywhere.

Corresponding to flashrom svn r644.
Signed-off-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>
Acked-by: Stefan Reinauer <stepan@coresystems.de>

Thanks to Uwe for the suggestion.

Corresponding to flashrom svn r637.
Signed-off-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>
Acked-by: Uwe Hermann <uwe@hermann-uwe.de>

Let %.o handle flashrom.o as well.

Corresponding to flashrom svn r626.
Signed-off-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>
Acked-by: Luc Verhaegen <libv@skynet.be>

Supports RS-232, USB serial converters (untested) and TCP streams.

All functionality is stubbed out to allow multiplatform compile testing
of the headers we use.
The real serial flasher protocol driver will be committed next.

Corresponding to flashrom svn r625.
Signed-off-by: Urja Rannikko <urjaman@gmail.com>
Signed-off-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>
Acked-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>

 - Move board_info_url struct to print.c, doesn't have to be global.

 - Simplify flashrom.c a bit by moving stuff to print.c.
   Eliminate two now-useless mini-functions in print.c.

 - Add a note that the wiki page contents are semi-automatically generated.

 - Mention date of last wiki page update as well as the flashrom revision
   that was used to generate the wiki output.

 - Also generate list of supported laptops in -z output now.

 - Add some more board URLs.

 - Add a boards_notes[] table to allow for arbitrary footnotes/comments for
   each board in the table. All notes will automatically be turned into
   wiki footnotes with correct numbers and will appear at the end of the
   respective table.

Corresponding to flashrom svn r615.
Signed-off-by: Uwe Hermann <uwe@hermann-uwe.de>
Acked-by: Uwe Hermann <uwe@hermann-uwe.de>

The new makefile structure uses indirection (.features->features) to
work around a gmake bug which resulted in immediate premature evaluation
of variable definitions with deferred evaluation.

Corresponding to flashrom svn r611.
Signed-off-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>
Acked-by: Uwe Hermann <uwe@hermann-uwe.de>

Drop no longer needed MAX macro, we have a max() function.

Corresponding to flashrom svn r601.
Signed-off-by: Uwe Hermann <uwe@hermann-uwe.de>
Acked-by: Uwe Hermann <uwe@hermann-uwe.de>

FTDI support is autodetected during compilation.

Paul writes:
There are certainly possible improvements: The code has hard-coded
values for which interface of the ftdi chip to use (interface B was
chosen because libftdi seems to have trouble with A right now), what
clock rate use for the SPI interface (I've been running at 30Mhz, but
the patch sets it to 10Mhz), and possibly others. I think this means
that per-programmer options might be a good idea at some point.

Carl-Daniel writes:
There is one additional FIXME comment in the code, but AFAICS that
problem is not solvable with current libftdi.

Corresponding to flashrom svn r598.
Signed-off-by: Paul Fox <pgf@laptop.org>
Signed-off-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>
Acked-by: Uwe Hermann <uwe@hermann-uwe.de>

Corresponding to flashrom svn r597.
Signed-off-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>
Acked-by: Joseph Smith <joe@settoplinux.org>

This strangely breaks with gmake on FreeBSD, but seems to work with make
on Linux.

Thanks Idwer Vollering <vidwer@gmail.com> for noticing and testing.

Corresponding to flashrom svn r593.
Signed-off-by: Uwe Hermann <uwe@hermann-uwe.de>
Acked-by: Uwe Hermann <uwe@hermann-uwe.de>

I sucessfully tested all operations on a Pm29F002T chip. The Pm29F002B is
untested but I assume it should also work.

Corresponding to flashrom svn r590.
Signed-off-by: Uwe Hermann <uwe@hermann-uwe.de>
Acked-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>

In particular, it should no longer be necessary to patch the makefile
for hassle-free compilation.

The targets are:
make export
make tarball
Both preserve svn revisions and the exported tree does not depend on
subversion in any way or shape.

Documentation for this feature has been added to README.

We need this for five reasons:
1. Packagers currently have to patch flashrom source to compile it on
systems without subversion. We should make it easier for them.
2. Snapshot tarballs currently have a .svn 1.5 directory included but
this will cause errors for users with older svn 1.4. Not requiring
subversion for snapshot compilation is best.
3. Since packagers seldom the svn revision in their fixup patches, some
packages out there have incorrect or no revision, only major version
numbers.
4. Releasing a new version of flashrom needs too many changes to the
makefile which have to be reverted instantly after the release. That is
unnecessary churn.
5. Making a release is easy with the change. Update the major version,
then run "make tarball".

Corresponding to flashrom svn r587.

Signed-off-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>  
Acked-by: Uwe Hermann <uwe@hermann-uwe.de>