GitLab

Constify variables where possible.
Initialize programmer-related variables explicitly in programmer_init to
allow running programmer_init from a clean state after
programmer_shutdown.
Prohibit registering programmer shutdown functions before init or after
shutdown.
Kill some dead code.
Rename global variables with namespace-polluting names.
Use a previously unused locking helper function in sst49lfxxxc.c.

This is needed for libflashrom.

Effects on the binary size of flashrom are minimal (300 bytes
shrinkage), but the data section shrinks by 4384 bytes, and that's a
good thing if flashrom is operating in constrained envionments.

Corresponding to flashrom svn r1068.
Signed-off-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>
Acked-by: Michael Karcher <flashrom@mkarcher.dialup.fu-berlin.de>

 - Fix a number of typos (found via ispell).

 - Use correct vendor names (as per their websites) consistently.

Corresponding to flashrom svn r985.
Signed-off-by: Uwe Hermann <uwe@hermann-uwe.de>
Acked-by: Uwe Hermann <uwe@hermann-uwe.de>

Corresponding to flashrom svn r982.
Signed-off-by: Sean Nelson <audiohacked@gmail.com>
Acked-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>

Some of the spi programmer drivers required chipdrivers.h, needs fixing later:
  it87spi.c
  ichspi.c
  sb600spi.c
  wbsio_spi.c
  buspirate_spi.c
  ft2232spi.c
  bitbang_spi.c
  dediprog.c

Corresponding to flashrom svn r914.
Signed-off-by: Sean Nelson <audiohacked@gmail.com>
Acked-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>

Add FEATURE_SHORT_RESET, FEATURE_LONG_RESET, and FEATURE_EITHER_RESET
rewrite jedec functions to use getaddrmask

convert write_49f002 to write_jedec_1
convert write_w39v040c to write_jedec_1
convert probe_w39v040c to probe_jedec
convert write_49lf040 to write_jedec_1
convert write_pm29f002 to write_jedec
convert write_29f040b to write_jedec_1
convert probe_29f040b to probe_jedec
convert erase_chip_29f040b to erase_chip_block_jedec
convert erase_sector_29f040b to erase_sector_jedec
convert write_m29f002b to write_jedec
convert write_m29f002t to write_jedec
convert *_29f002 to *_jedec

decouple unused files from Makefile:
am29f040b.c
en29f002a.c
m29f002.c
mx29f002.c
pm29f002.c
sst49lf040.c
w39v040c.c
w49f002u.c

Corresponding to flashrom svn r886.
Signed-off-by: Sean Nelson <audiohacked@gmail.com>
Acked-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>
Acked-by: Anders Juel Jensen <andersjjensen@gmail.com>

This allows us to have one common probe_jedec function instead of half a
dozen wrappers. The trick here is to have FEATURE_ADDR_FULL==0 and thus
default to unmasked addresses. That way, we only have to annotate chips
which need small address masks.

Corresponding to flashrom svn r840.
Signed-off-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>
Acked-by: Sean Nelson <audiohacked@gmail.com>

All functions which just call probe_jedec and then map flash registers
are replaced by probe_jedec. All functions which call probe_jedec, map
flash registers and do something else can at least eliminate mapping
flash registers.
Fix logic inversion in probe_jedec to map flash registers on success
instead of on failure.
Change a few TIMING_IGNORED to TIMING_FIXME where probe_jedec is used.

Total savings: One probe function simplified, three probe functions
eliminated.

Corresponding to flashrom svn r839.
Signed-off-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>
Acked-by: Michael Karcher <flashrom@mkarcher.dialup.fu-berlin.de>
Acked-by: Sean Nelson <audiohacked@gmail.com>

The patch converts jedec functions into mask-based generics which can
be used for many chip provided the only changes are the addresses are
converted from 0x5555/0x2AAA to 0x555/0x2AA or similar.

The patch mostly changes jedec.c, but a few other files are changed
because they use the jedec functions within their own functions.

The patch also adds a copyright line to flashchips.c because of my
recent work in converting AMD and Atmel chips to use struct erase_block.

Corresponding to flashrom svn r828.
Signed-off-by: Sean Nelson <audiohacked@gmail.com>
Acked-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>

Am29F010A/B
Am29F002(N)BB
Am29F002(N)BT
Am29F016D
Am29F040B
Am29F080B
Am29LV040B
Am29LV081B
A29040B
Pm29F002T
Pm29F002B

Change function signature of Am29 erase functions and JEDEC chip erase
to be usable with block_erasers.

Corresponding to flashrom svn r812.
Signed-off-by: Sean Nelson <audiohacked@gmail.com>
Acked-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>

If the JEDEC Toggle Bit algorithm needs more than 2^20 loops, it is a
good sign we should have used delays between toggle bit reads.

Tell the user about this. 2^20 loops need roughly a second depending on
flash bus speed. One reason for excessive loops can be a slow operation
like erase.

The Winbond W39V040C requires a 50 ms delay between toggle bit reads
during erase according to the datasheet. Turns out a 2 ms delay is
sufficient. Use a safety factor of 4 and default all erase operations
to 8 ms delay between toggle reads. This is short enough not to have
a substantial negative impact on erase times, and should improve
reliability.

This patch addresses the excessive toggle behaviour (observed on some
non-Winbond chips) and the toggle delay requirement (Winbond W39V040C).

Corresponding to flashrom svn r807.
Signed-off-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>
Acked-by: Javier Ortega Conde (aka Malkavian) <malkavian666@gmail.com>
Acked-by: Michael Karcher <flashrom@mkarcher.dialup.fu-berlin.de>

Probe_jedec() checks the delay value and issues programmer_delay based
on the value except for delays between single chip_writeb. If a chip has
zero probe_delay, delays between chip_writeb should be skipped as well.

Corresponding to flashrom svn r805.
Signed-off-by: Sean Nelson <audiohacked@gmail.com>
Acked-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>

The deleted function in en29f002a.c is reintroduced as
write_by_byte_jedec in jedec.c as it contains no chip-specific
instructions. It is not yet used in other chip drivers, as key addresses
(0x2AAA/0x5555) are often specified with less bits. After crosschecking
datasheets, most of the fixmes can probably be resolved as indicated in
them, causing significant code reduction.

The common JEDEC code for bytewise programming does not program 0xFF
at all. The chips that had a dedicated bytewise flash function which
has been changed to write_jedec_1 thus changed flashing behaviour
and the "write" test flag has been removed. This applies to: AMD
Am29F002BB/Am29F002NBB AMD Am29F002BT/Am29F002NBT (TEST_OK_PREW before)
AMIC A29002B AMIC A29002T (TEST_OK_PREW before) EON EN29F002(A)(N)B EON
EN29F002(A)(N)T (TEST_OK_PREW before) Macronix MX29F001B (TEST_OK_PREW
before) Macronix MX29F001T (TEST_OK_PREW before) Macronix MX29F002B
Macronix MX29F002T (TEST_OK_PREW before) Macronix MX29LV040

Similar analysis should be performed for the read id stuff.

Corresponding to flashrom svn r785.
Signed-off-by: Michael Karcher <flashrom@mkarcher.dialup.fu-berlin.de>
Acked-by: Sean Nelson <audiohacked@gmail.com>
Acked-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>

This patch removes the extremely dangerous unprotect_jedec function
which is not used at all within flashrom code, and renames the
misleadingly named protect_jedec function to start_program_jedec.

Calls to protect_jedec after flashing are removed, because a) on LPC
chips, the command sent by protoct_jedec is not even in the datasheet
and b) on parallel chips, the block write command issued before already
contained the software protection sequence, so software protection is
definitely enabled.

This patch also removes two clones of protect_jedec

Background: JEDEC Software Data Protection started as an optional
feature, which was disabled on the first single-voltage-flash chips.
The software data protection is the need to prefix a write with a magic
"write enable" command, while without write protection every write
access into the chip's address space modifies flash content. This magic
write enable command also tells the flash chip that the programmer
obviously support sending write-enable commands and turns off the "any
write modifies flash content" mode. There also exist a two-command (6
writes) sequence that disables Software Data Protection completey, which
should only ever be used to prepare updating with a device that can't
handle software data protection.

Corresponding to flashrom svn r783.
Signed-off-by: Michael Karcher <flashrom@mkarcher.dialup.fu-berlin.de>
Acked-by: Sean Nelson <audiohacked@gmail.com>
Acked-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>

Fix. jedec.c error handling used double negation in too many places for
no good reason. Clean up.

Corresponding to flashrom svn r779.
Signed-off-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>
Acked-by: Sean Nelson <audiohacked@gmail.com>

The automatic retry in write_page_write_jedec didn't retry flashing the
correct range, essentially rendering the functionality useless.

This patch simplifies the code and fixes the bug.

Thanks to Luke Dashjr for testing.

Mark Winbond W29C040P as supported.

Corresponding to flashrom svn r757.
Signed-off-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>
Acked-by: Luke Dashjr <luke_coreboot@dashjr.org>

Corresponding to flashrom svn r751.
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>

Annotate SST49LF004B quirks for TBL#.

Add TEST_OK_PRW which is useful when a PREW chip gets a new erase
routine.

Change a few erase function prototypes to use unsigned int instead of
int.

Corresponding to flashrom svn r731.
Signed-off-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>
Acked-by: Luc Verhaegen <libv@skynet.be>

The __func__ variant is standardized in C99 and recommended to be
used instead of __FUNCTION__ in the gcc info page.

Only _very_ old versions of gcc did not know about __func__, but we've
been using both __func__ and __FUNCTION__ for a long while now, and
nobody complained about this, so all our users seem to use recent
enough compilers.

Corresponding to flashrom svn r711.
Signed-off-by: Uwe Hermann <uwe@hermann-uwe.de>
Acked-by: Stefan Reinauer <stepan@coresystems.de>

This makes visual inspection and grepping a lot harder than necessary.
Remove line breaks where appropriate. Some error messages should end up
on stderr instead of just being displayed in verbose mode.

Thanks to Maciej Pijanka for testing.

Corresponding to flashrom svn r660.
Signed-off-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>
Acked-by: Stefan Reinauer <stepan@coresystems.de>

Tested by Urja Rannikko with external flasher.
Tested by Uwe Hermann with onboard flash.

Corresponding to flashrom svn r632.
Signed-off-by: Urja Rannikko <urjaman@gmail.com>
Signed-off-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>
Acked-by: Urja Rannikko <urjaman@gmail.com>
Acked-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>

And even when it checks if the erase worked, the result of that check is
often ignored.

Convert all erase functions and actually check return codes
almost everywhere.
Check inside all erase_* routines if erase worked, not outside.
erase_sector_jedec and erase_block_jedec have changed prototypes to
enable erase checking.

Uwe successfully tested LPC on an CK804 box and SPI on some SB600 box.

Corresponding to flashrom svn r595.
Signed-off-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>
Signed-off-by: Urja Rannikko <urjaman@gmail.com>
Acked-by: Uwe Hermann <uwe@hermann-uwe.de>

Add external programmer delay functions so external programmers can
handle the delay on their own if needed.

Corresponding to flashrom svn r578.
Signed-off-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>
Acked-by: Uwe Hermann <uwe@hermann-uwe.de>
Acked-by: Urja Rannikko <urjaman@gmail.com>

This eliminates the conflicting delay requirements for old and new chips
with the same probing sequence.

Corresponding to flashrom svn r569.
Signed-off-by: Maciej Pijanka <maciej.pijanka@gmail.com>
Acked-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>

Build-tested on 32bit x86.

Corresponding to flashrom svn r521.
Signed-off-by: Uwe Hermann <uwe@hermann-uwe.de>
Acked-by: Uwe Hermann <uwe@hermann-uwe.de>

Use chipaddr instead of volatile uint8_t * because when we access chips
in external flashers, they are not accessed via pointers at all.

Benefits: This allows us to differentiate between volatile machine
memory accesses and flash chip accesses. It also enforces usage
of chip_{read,write}[bwl] to access flash chips, so nobody will
unintentionally use pointers to access chips anymore. Some unneeded
casts are removed as well. Grepping for chip operations and machine
memory operations doesn't yield any false positives anymore.

Compile tested on 32 bit and 64 bit Linux.

Corresponding to flashrom svn r519.
Signed-off-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>
Acked-by: Uwe Hermann <uwe@hermann-uwe.de>

The semantic patch I used in r418 to make the original conversion to
accessor functions was missing one isomorphism:
a[b] <=> *(a+b)

The semantic patcher Coccinelle was used to create this patch. Semantic
patch follows:
@@
typedef uint8_t;
expression a;
volatile uint8_t *b;
@@
- b[a]
+ *(b + a)
@@
expression a;
volatile uint8_t *b;
@@
- *(b) = (a);
+ chip_writeb(a, b);
@@
volatile uint8_t *b;
@@
- *(b)
+ chip_readb(b)
@@
type T;
T b;
@@
(
 chip_readb
|
 chip_writeb
)
 (...,
- (T)
- (b)
+ b
 )

Corresponding to flashrom svn r498.
Signed-off-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>
Acked-by: Uwe Hermann <uwe@hermann-uwe.de>

When flashrom JEDEC code sends the ID command to the chip, it expects to
see IDs in the default flash location.

However, sometimes the chip does not react to the ID command, either
because it doesn't understand the command or because the command never
reached it. One way to detect this is to compare ID output with flash
chip contents for the same location. If they are identical, there is a
high chance you're not actually seeing ID output. Warn the user in that
case.

This patch helps a lot when a chip is not recognized and we want to
check if the probe responses are real IDs or just random flash chip
contents.

This should probably be added to all probe functions, but probe_jedec
is called for all sizes and thus flashrom will check this condition at
least once per size, making sure we can cross-match the warning.

Corresponding to flashrom svn r494.
Signed-off-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>
Acked-by: FENG Yu Ning <fengyuning1984@gmail.com>

Before we attempt trickery, we can simply rename the accessor functions.

Patch created with the help of Coccinelle.

Corresponding to flashrom svn r420 and coreboot v2 svn r3984.
Signed-off-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>
Acked-by: Idwer Vollering <idwer_v@hotmail.com>
Acked-by: Patrick Georgi <patrick@georgi-clan.de>

Right now we perform direct pointer manipulation without any abstraction
to read from and write to memory mapped flash chips. That makes it
impossible to drive any flasher which does not mmap the whole chip.

Using helper functions readb() and writeb() allows a driver for external
flash programmers like Paraflasher to replace readb and writeb with
calls to its own chip access routines.

This patch has the additional advantage of removing lots of unnecessary
casts to volatile uint8_t * and now-superfluous parentheses which caused
poor readability.

I used the semantic patcher Coccinelle to create this patch. The
semantic patch follows:
@@
expression a;
typedef uint8_t;
volatile uint8_t *b;
@@
- *(b) = (a);
+ writeb(a, b);
@@
volatile uint8_t *b;
@@
- *(b)
+ readb(b)
@@
type T;
T b;
@@
(
 readb
|
 writeb
)
 (...,
- (T)
- (b)
+ b
 )

In contrast to a sed script, the semantic patch performs type checking
before converting anything.

Tested-by: Joe Julian

Corresponding to flashrom svn r418 and coreboot v2 svn r3971.
Signed-off-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>
Acked-by: FENG Yu Ning <fengyuning1984@gmail.com>

Corresponding to flashrom svn r390 and coreboot v2 svn r3895.
Signed-off-by: Peter Stuge <peter@stuge.se>
Acked-by: Peter Stuge <peter@stuge.se>

We should follow data sheet timing, even if chips have been tested to answer
faster in the field.

Corresponding to flashrom svn r273 and coreboot v2 svn r3387.
Signed-off-by: Peter Stuge <peter@stuge.se>
Acked-by: Peter Stuge <peter@stuge.se>

Corresponding to flashrom svn r264 and coreboot v2 svn r3375.
Signed-off-by: Peter Stuge <peter@stuge.se>
Acked-by: Peter Stuge <peter@stuge.se>

Run time is increased a few 100ms but this is needed for reliability.
I consider this trivial.

Corresponding to flashrom svn r262 and coreboot v2 svn r3373.
Signed-off-by: Peter Stuge <peter@stuge.se>
Acked-by: Peter Stuge <peter@stuge.se>

Flash chips which can be detected by JEDEC probe routines all have
vendor IDs with correct parity. Use a parity check as additional hint
whether a vendor ID makes sense. Note: Device IDs have no parity
requirements whatsoever.

Corresponding to flashrom svn r231 and coreboot v2 svn r3308.
Signed-off-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>
Acked-by: Peter Stuge <peter@stuge.se>

The continuation ID code does not go further than checking for IDs of
the type 0x7fXX, but does this for vendor and product ID. The current
published JEDEC spec has a list where the largest vendor ID is 7 bytes
long, but all leading bytes are 0x7f. The list will grow in the future,
and using a 64bit variable will not be enough anymore.
Besides that, it seems that the location of the ID byte after the first
continuation ID byte is very vendor specific, so we may have to revisit
that code some time in the future.

(Suggestion for a new encoding:
Use a two-byte data type for the ID, the lower byte contains the only
non-0x7f byte, the upper byte contains the number of 0x7f bytes used as
prefix, which is the bank number minus 1 the vendor ID appears in.)

Add support for EON EN29F002AT.

Corresponding to flashrom svn r171 and coreboot v2 svn r3030.
Signed-off-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>
Acked-by: Corey Osgood <corey.osgood@gmail.com>

The JEDEC probe routine had a delay of 10 us after entering ID mode
and this was insufficient for the 29C020. The data sheet claims we
have to wait 10 ms, but tests have shown that 20 us suffice. Allow for
variations in chip delays with a factor of 2 safety margin.

Corresponding to flashrom svn r159 and coreboot v2 svn r2962.
Signed-off-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>
Acked-by: Peter Stuge <peter@stuge.se>

Corresponding to flashrom svn r151 and coreboot v2 svn r2873.
Signed-off-by: Uwe Hermann <uwe@hermann-uwe.de>
Acked-by: Stefan Reinauer <stepan@coresystems.de>

Also minor changes to remove tab-space combinations where possible. 

Corresponding to flashrom svn r144 and coreboot v2 svn r2850.
Signed-off-by: Ronald G. Minnich <rminnich@gmail.com>
Signed-off-by: David Hendricks <david.hendricks@gmail.com>
Acked-by: Stefan Reinauer <stepan@coresystems.de>

Corresponding to flashrom svn r136 and coreboot v2 svn r2768.
Signed-off-by: Uwe Hermann <uwe@hermann-uwe.de>
Acked-by: Uwe Hermann <uwe@hermann-uwe.de>

No changes in content of the files.

Corresponding to flashrom svn r131 and coreboot v2 svn r2751.
Signed-off-by: Uwe Hermann <uwe@hermann-uwe.de>
Acked-by: Uwe Hermann <uwe@hermann-uwe.de>