commit ec50a8e389
"cfi_flash: handle 'chip size exceeds address window' situation"
added 3rd argument to flash_get_size() but didn't fix all the
function calls from the board specific code. Many boards have
their own flash_get_size() definitions in the board code and
use them there, but some boards (e.g. tqm834x, tqm85xx, pdm360ng)
use flash_get_size() from the cfi_flash.c driver.
The bug shows up if the value of the "max_size" argument (which
is not defined when calling the function with two arguments)
happens to be less than "info->size". In this case on the
affected boards we end up with a bank of reduced size and
in the worst case might even be not able to update U-Boot or
to boot the kernel from flash:
=> fli
Bank # 1: CFI conformant FLASH (32 x 16) Size: 0 kB in 1 Sectors
AMD Standard command set, Manufacturer ID: 0x01, Device ID: 0x227E
Erase timeout: 4096 ms, write timeout: 1 ms
Buffer write timeout: 3 ms, buffer size: 64 bytes
Sector Start Addresses:
F0000000 RO
Bank # 2: CFI conformant FLASH (32 x 16) Size: 128 MB in 512 Sectors
AMD Standard command set, Manufacturer ID: 0x01, Device ID: 0x227E
Erase timeout: 4096 ms, write timeout: 1 ms
Buffer write timeout: 3 ms, buffer size: 64 bytes
Sector Start Addresses:
F8000000 F8040000 F8080000 F80C0000 F8100000
F8140000 F8180000 F81C0000 F8200000 F8240000
...
E.g., updating U-Boot is not possible now:
=> protect off ${u-boot_addr} +${u-boot_size}
Error: end address (0xf007ffff) not in flash!
Bad address format
=> era ${u-boot_addr} +${u-boot_size}
Error: end address (0xf007ffff) not in flash!
Bad address format
This patch removes the 3rd argument of flash_get_size() again
and sets "max_size" in the function itself instead of passing
it as a function argument.
Signed-off-by: Anatolij Gustschin <agust@denx.de>
This is needed for the canyonlands_nand build target. Without it
the resulting image won't fit into 4k.
Signed-off-by: Stefan Roese <sr@denx.de>
Cc: Scott Wood <scottwood@freescale.com>
Acked-by: Scott Wood <scottwood@freescale.com>
Acked-by: Stefan Roese <sr@denx.de>
Before this commit, weak symbols were not overridden by non-weak symbols
found in archive libraries when linking with recent versions of
binutils. As stated in the System V ABI, "the link editor does not
extract archive members to resolve undefined weak symbols".
This commit changes all Makefiles to use partial linking (ld -r) instead
of creating library archives, which forces all symbols to participate in
linking, allowing non-weak symbols to override weak symbols as intended.
This approach is also used by Linux, from which the gmake function
cmd_link_o_target (defined in config.mk and used in all Makefiles) is
inspired.
The name of each former library archive is preserved except for
extensions which change from ".a" to ".o". This commit updates
references accordingly where needed, in particular in some linker
scripts.
This commit reveals board configurations that exclude some features but
include source files that depend these disabled features in the build,
resulting in undefined symbols. Known such cases include:
- disabling CMD_NET but not CMD_NFS;
- enabling CONFIG_OF_LIBFDT but not CONFIG_QE.
Signed-off-by: Sebastien Carlier <sebastien.carlier@gmail.com>
Supports most types that support Read-Id and the FM25H20.
Signed-off-by: Reinhard Meyer <u-boot@emk-elektronik.de>
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
By now, the majority of architectures have working relocation
support, so the few remaining architectures have become exceptions.
To make this more obvious, we make working relocation now the default
case, and flag the remaining cases with CONFIG_NEEDS_MANUAL_RELOC.
Signed-off-by: Wolfgang Denk <wd@denx.de>
Tested-by: Heiko Schocher <hs@denx.de>
Tested-by: Reinhard Meyer <u-boot@emk-elektronik.de>
This patch adds generic support for the Samsung s3c2440 processor.
Global s3c24x0 changes to struct members converting from upper case to
lower case.
Signed-off-by: Craig Nauman <cnauman@diagraph.com>
Cc: kevin.morfitt@fearnside-systems.co.uk
Signed-off-by: Minkyu Kang <mk7.kang@samsung.com>
On some boards we have flash mapped high in the address space with
considerably small window (say 0xFE000000 and 32MB). When we install
bigger chip (say 64MB) on such a board strange things happen
(flash_write() doesn't work at all, for ex). That's because cfi_flash
driver doesn't care about window size at all.
Of course, cleanest solution would probably be to just extend address
window to be able to map the whole flash but for legacy/compatibility
reasons some people prefer just truncate the flash size and never use
the upper part.
This patch adds an option for cfi_flash driver to handle this situation
properly. To achieve this we add the new function cfi_flash_bank_size()
which can be provided by the board code and weak-aliased to default
implementation that returns value from the CONFIG_SYS_FLASH_BANKS_SIZES
array if it's defined or 0 otherwise (the last case is added for
compatibility).
If non-zero flash bank size is provided and detected chip size is bigger
than provided address window size the warning will be displayed and
flash chip will be truncated.
Signed-off-by: Ilya Yanok <yanok@emcraft.com>
Changed cfi_flash_bank_size() return type to unsigned long
to match caller function.
Signed-off-by: Wolfgang Denk <wd@denx.de>
The ECC calculations were started by writing 1 << 13 to the nand FCR register;
that value is also defined as DAVINCI_NANDFCR_4BIT_CALC_START in emif_defs.h.
This patch substitutes the macro DAVINCI_NANDFCR_4BIT_CALC_START for the
magic number '1 << 13'.
Signed-off-by: Ben Gardiner <bengardiner@nanometrics.ca>
Signed-off-by: Sandeep Paulraj <s-paulraj@ti.com>
This allows for arbitrarily long manufacturer ids following the JEDEC
standard of 0x7f continuation bytes. It also makes adding new entries
easier as it's just one element in an array. The downside is that it
increases code size a bit, but we're talking ~50 bytes.
Signed-off-by: Reinhard Meyer <u-boot@emk-elektronik.de>
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
When a CFI flash chip could not be detected an error message similar to
the following would be printed on bootup:
FLASH: ## Unknown FLASH on Bank 1 - Size = 0x01000000 = 0 MB
The printf incorrectly converted the flash size into megabytes. This
patch fixes the printing of the flash size in megabytes:
FLASH: ## Unknown FLASH on Bank 1 - Size = 0x01000000 = 16 MB
Signed-off-by: John Schmoller <jschmoller@xes-inc.com>
Signed-off-by: Peter Tyser <ptyser@xes-inc.com>
Signed-off-by: Stefan Roese <sr@denx.de>
This patch adds the Numonyx manufacturer code (0x20) to
onenand manufacturers.
Signed-off-by: Enric Balletbo i Serra <eballetbo@gmail.com>
Acked-by: Kyungmin Park <kyungmin.park@samsung.com>
Acked-by: Steve Sakoman <steve.sakoman@linaro.org>
Tested-by: Steve Sakoman <steve.sakoman@linaro.org>
Consolidate some code in mtd_get_len_incl_bad(), and fix a condition
where a valid partition could be reported as truncated if it has a
good block at the end of the device (unlikely, since the BBT is usually
there).
Fix mid-block declarations in net_part_size().
Signed-off-by: Scott Wood <scottwood@freescale.com>
Reviewed-by: Ben Gardiner <bengardiner@nanometrics.ca>
The logic to 'spread' mtd partitions needs to calculate the length in
the mtd device, including bad blocks.
This patch introduces a new function, mtd_get_len_incl_bad that can
return both the length including bad blocks and whether that length
was truncated on the device. This new function will be used by the
mtdparts spread command later in this series. The definition of the
function is #ifdef'd out in configurations that do not use the new
'mtdparts spread' command.
Signed-off-by: Ben Gardiner<bengardiner@nanometrics.ca>
CC: Scott Wood <scottwood@freescale.com>
Get rid of the several "#if 0" sections that were keeping around Linux
code that isn't relevant to U-Boot. Besides cluttering the code, these
sections make tracking upstream changes harder, rather than easier.
It's easy to discard obviously irrelevant diff hunks that patch rejects,
but it's not as easy to notice hunks that apply cleanly to the #if 0
section, but *are* relevant to U-Boot and require modification elsewhere.
Also remove suspend/resume, as this is not applicable to U-Boot. Removal
saves 232 bytes on powerpc.
Signed-off-by: Scott Wood <scottwood@freescale.com>
Tested-by: Ben Gardiner <bengardiner@nanometrics.ca>
A while back, in http://lists.denx.de/pipermail/u-boot/2009-June/054428.html,
Michele De Candia posted a patch to not count bad blocks toward the
requested size to be erased. This is desireable when you're passing in
something like $filesize, but not when you're trying to erase a partition.
Thus, a .spread subcommand (named for consistency with
http://lists.denx.de/pipermail/u-boot/2010-August/075163.html) is introduced
to make explicit the user's desire to erase for a given amount of data,
rather than to erase a specific region of the chip.
While passing $filesize to "nand erase" is useful, accidentally passing
something like $fliesize currently produces quite unpleasant results, as the
variable evaluates to nothing and U-Boot assumes that you want to erase
the entire rest of the chip/partition. To improve the safety of the
erase command, require the user to make explicit their intentions by
using a .part or .chip subcommand. This is an incompatible user interface
change, but keeping compatibility would eliminate the safety gain, and IMHO
it's worth it.
While touching nand_erase_opts(), make it accept 64-bit offsets and sizes,
fix the percentage display when erase length is rounded up, eliminate
an inconsistent warning about rounding up the erase length which only
happened when the length was less than one block (rounding up for $filesize
is normal operation), and add a diagnostic if there's an attempt to erase
beginning at a non-block boundary.
Signed-off-by: Scott Wood <scottwood@freescale.com>
Tested-by: Ben Gardiner <bengardiner@nanometrics.ca>
The underlying code in nand_base.c already supports non-page-aligned reads
and writes, but the block-skipping wrapper code did not.
With block skipping, an unaligned start address is not useful since you
really want to be starting at the beginning of a partition -- or at least
that's where you want to start checking for blocks to skip, but we don't
(yet) support that. So we still require the start address to be aligned.
An unaligned length, though, is useful for passing $filesize to the
read/write command, and handling it does not complicate block skipping.
Signed-off-by: Scott Wood <scottwood@freescale.com>
Tested-by: Ben Gardiner <bengardiner@nanometrics.ca>
This patch introduces an extra mask-field in spansion_spi_flash_params
to support flash chips with 1-byte extended ID (like the S25FL032P).
Signed-off-by: David Jander <david@protonic.nl>
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
The spansion_erase currently only works when the sector size is 64KB.
cmd[1] should contain the higher 8 bit of the 24 bit address of the
sector to be erased. Currently it is holding the sector index to be
erased which happens to be the same thing when the sector size is
64KB.
Signed-off-by: Marc-Andre Hebert <marc-andre.hebert@humanware.com>
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Adds support for Winbond's W25Q64 SPI flash. These devices are used on
(among others) Xilinx' SP601 and SP605 Spartan-6 evaluation boards.
Tested with "sf" commands.
Signed-off-by: Graeme Smecher <graeme.smecher@mail.mcgill.ca>
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Commit 2ee951ba (UBI: Enable re-initializing of the "ubi part" command)
reset mtd_devs in ubi_exit() but missed ubi_init()'s failure path.
Signed-off-by: Karl Beldan <karl.beldan@gmail.com>
Cc: Stefan Roese <sr@denx.de>
Signed-off-by: Stefan Roese <sr@denx.de>
This patch moves some ppc4xx related headers from the common include
directory (include/) to the powerpc specific one
(arch/powerpc/include/asm/). This way to common include directory is not
so cluttered with files.
Signed-off-by: Stefan Roese <sr@denx.de>
Now that the defines are moved to header files we don't need this
conditional compilation any more. Remove it.
Signed-off-by: Stefan Roese <sr@denx.de>
This patch simplifies the use of CONFIG_SYS_MAX_FLASH_BANKS_DETECT. By
moving these optional variables and defines into the common code, board
specific code is minimized. Currently only the following board use
this feature:
APC405, IDS8247, TQM834x
And IDS8247 doesn't seem to really need this feature, since its not
updating the bank number variable at all. So this patch removes the
definition of CONFIG_SYS_MAX_FLASH_BANKS_DETECT from this board port.
This new framework will be used by the upcoming lwmon5 update as well.
Signed-off-by: Stefan Roese <sr@denx.de>
Acked-by: Heiko Schocher <hs@denx.de>
Cc: Matthias Fuchs <matthias.fuchs@esd.eu>
cfi_flash_bank_addr(int bank_nr) returns the base addresses of the
requested bank. Introducing this weak default enables boards to override
this functions with a board specific version when required.
This feature will be used in the lwmon5 board update, supporting runtime
detection of 2 board revisions with different flash layouts.
Signed-off-by: Stefan Roese <sr@denx.de>
This patch removes an unecessary check in the return statement. This is
not needed, since "info" is initializes to NULL. And "info" will not be
written to again, if the flash address is not found.
Additionally "info" is not initialized to "0" but to "NULL".
Signed-off-by: Stefan Roese <sr@denx.de>
Due to a register glitch (result code <4 might show up right after the
start-calculation-bit was set), make sure the ECC has really started.
See 1c3275b656045aff9a75bb2c9f3251af1043ebb3 in the kernel.
Signed-off-by: Wolfram Sang <w.sang@pengutronix.de>
Cc: Sandeep Paulraj <s-paulraj@ti.com>
This printk was added recently and results in ugly output on systems
with no NAND:
NAND: nand_get_flash_type: unknown NAND device: Manufacturer ID: 0x00, Chip ID: 0x00 0 MiB
instead of:
NAND: 0 MiB
Signed-off-by: Steve Sakoman <steve@sakoman.com>
Acked-by: Scott Wood <scottwood@freescale.com>
Signed-off-by: Sandeep Paulraj <s-paulraj@ti.com>
I have "ported" U-boot to a in house made board with Numonyx Axcell P33/P30
256-Mbit 65nm flash chips.
After some time :( searching for bugs in our board or soft, we have
discovered that those chips have a small but annoying bug, documented in
"Numonyx Axcell P33/P30 256-Mbit Specification Update"
It states :
When customer uses [...] block unlock, the block lock status might be
altered inadvertently. Lock status might be set to either 01h or 03h
unexpectedly (00h as expected data), which leads to program/erase failure
on certain blocks.
A working workaround is given, which I have applied and tested with success :
Workaround: If the interval between 60h and its subsequent command
can be guaranteed within 20us, Option I is recommended,
otherwise Option II (involves hardware) should be selected.
Option I: The table below lists the detail command sequences:
Command
Data bus Address bus Remarks
Sequence
1 90h Block Address
Read Lock Status
2 Read Block Address + 02h
(2)(3) (1)
3 60h Block Address
(2)(3) (1) Lock/Unlock/RCR Configuration
4 D0h/01h/03h Block Address
Notes:
(1) Block Address refers to RCR configuration data only when the 60h
command sequence is used to set RCR register combined with 03h
subsequent command.
(2) For the third and fourth command sequences, the Block Address must
be the same.
(3) The interval between 60h command and its subsequent D0h/01h/2Fh/03h
commands should be less than 20us.
And here is a log comparison of a simple (destructive) flash test without
and with the workaround.
diff without-numonyx-workaround.log with-numonyx-workaround.log
-U-Boot 2010.06-00696-g22b002c-dirty (Aug 16 2010 - 15:07:47)
+U-Boot 2010.06-00696-g22b002c-dirty (Aug 16 2010 - 15:25:19)
CPU: Freescale MCF5484
CPU CLK 200 MHz BUS CLK 100 MHz
Board: Macq Electronique ME2060
I2C: ready
DRAM: 64 MiB
FLASH: 32 MiB
In: serial
Out: serial
Err: serial
Net: FEC0, FEC1
-> flinfo
Bank # 1: CFI conformant FLASH (16 x 16) Size: 32 MB in 259 Sectors
Intel Extended command set, Manufacturer ID: 0x89, Device ID: 0x8922
Erase timeout: 4096 ms, write timeout: 1 ms
Buffer write timeout: 5 ms, buffer size: 1024 bytes
Sector Start Addresses:
FE000000 RO FE008000 RO FE010000 RO FE018000 RO FE020000 RO
FE040000 RO FE060000 RO FE080000 RO FE0A0000 RO FE0C0000 RO
...
FFF80000 RO FFFA0000 RO FFFC0000 RO FFFE0000 RO
-> protect off all
Un-Protect Flash Bank # 1
................... done
-> erase all
Erase Flash Bank # 1
................... done
-> cp.b 1000000 fe000000 2000000
-Copy to Flash... Flash not Erased
+Copy to Flash... done
->
Signed-off-by: Philippe De Muyter <phdm@macqel.be>
Signed-off-by: Stefan Roese <sr@denx.de>
This patch does the following:
- Extract code to detect if sector is erased into function
sector_erased().
- Because of this, we don't have variable declarations inside the
sector loop in flash_print_info()
- Change "return" to "break" in the "if (ctrlc()) statement:
This fixes a problem with the resulting output. Before this
patch the output was:
Sector Start Addresses:
FC000000 FC020000 FC040000 =>
With this patch it is now:
Sector Start Addresses:
FC000000 FC020000 FC040000
=>
Signed-off-by: Stefan Roese <sr@denx.de>
Cc: Kim Phillips <kim.phillips@freescale.com>
Cc: Wolfgang Denk <wd@denx.de>
Fix reading and printing of CFI flashes 16-bit devices identifiers
Nowadays CFI flashes have a 16-bit device identifier. U-boot still
print them and read them as if they were only 8-bit wide. Fix that.
Before:
Intel Extended command set, Manufacturer ID: 0x89, Device ID: 0x1B
After:
Intel Extended command set, Manufacturer ID: 0x89, Device ID: 0x881B
Signed-off-by: Philippe De Muyter <phdm@macqel.be>
Signed-off-by: Stefan Roese <sr@denx.de>
While running from flash, i. e. before relocation, we have only a
limited C runtime environment without writable data segment. In this
phase, some configurations (for example with environment in EEPROM)
must not use the normal getenv(), but a special function. This
function had been called getenv_r(), with the idea that the "_r"
suffix would mean the same as in the _r_eentrant versions of some of
the C library functions (for example getdate vs. getdate_r, getgrent
vs. getgrent_r, etc.).
Unfortunately this was a misleading name, as in U-Boot the "_r"
generally means "running from RAM", i. e. _after_ relocation.
To avoid confusion, rename into getenv_f() [as "running from flash"]
Signed-off-by: Wolfgang Denk <wd@denx.de>
Acked-by: Detlev Zundel <dzu@denx.de>
Currently, 83xx, 86xx, and 85xx have a lot of duplicated code
dedicated to defining and manipulating the LBC registers. Merge
this into a single spot.
To do this, we have to decide on a common name for the data structure
that holds the lbc registers - it will now be known as fsl_lbc_t, and we
adopt a common name for the immap layouts that include the lbc - this was
previously known as either im_lbc or lbus; use the former.
In addition, create accessors for the BR/OR regs that use in/out_be32
and use those instead of the mismash of access methods currently in play.
I have done a successful ppc build all and tested a board or two from
each processor family.
Signed-off-by: Becky Bruce <beckyb@kernel.crashing.org>
Acked-by: Kim Phillips <kim.phillips@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
Rather than bang MMRs directly, use the new portmux framework to handle
the details.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Acked-by: Scott Wood <scottwood@freescale.com>
Make it easy to use GPIOs for the DEV_READY pin by using the common GPIO
framework. Also make the NAND_PLAT_INIT() define optional.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Tested-by: Thomas Chou <thomas@wytron.com.tw>
The Toshiba TC58NVG0* parts are 128Mbytes x 8 bits 3.3V parts with the 0xD1
identifier. Add these to the list of known devices IDs.
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
When the NAND part is not supported, it is useful to show the manufacturer
and device ID to help debugging and reporting.
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: Scott Wood <scottwood@freescale.com>
At the moment, the default SPI flash subsystem is quite terse. Errors and
successes both result in a generic message. So move the useful errors and
useful successes to printf output by default.
While we're here, also convert the messages to use print_size().
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Some old STMicro parts do not support JEDEC ID (0x9f). This patch
uses RES (0xab) to get Electronic ID and translates it to JEDEC ID.
Signed-off-by: Thomas Chou <thomas@wytron.com.tw>
Acked-by: Mike Frysinger <vapier@gentoo.org>
UBI: initialise update marker
The in kernel copy of a volume's update marker is not initialised from the
volume table. This means that volumes where an update was unfinnished will
not be treated as "forbidden to use". This is basically that the update
functionality was broken.
Signed-off-by: Peter Horton <zero@colonel-panic.org>
Signed-off-by: Artem Bityutskiy <Artem.Bityutskiy@nokia.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Acked-by: Stefan Roese <sr@denx.de>
The current Blackfin nand write function fills up the write buffer but
returns before it has had a chance to drain. On faster systems, this
isn't a problem as the operation finishes before the ECC registers are
read, but on slower systems the ECC may be incomplete when the core tries
to read it.
So wait for the buffer to drain once we're done writing to it.
Signed-off-by: Andrew Caldwell <Andrew.Caldwell@analog.com>
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
The current U-Boot UBI implementation is copied from Linux. In this
porting the UBI background thread was not handled correctly. Upon write
operations ubi_wl_flush() makes sure, that all queued operations, like
page-erase, are completed. But this is missing for read operations.
This patch now makes sure that such operations (like scrubbing upon
bit-flip errors) are not queued, but executed directly.
Signed-off-by: Stefan Roese <sr@denx.de>
