After several tries with ubifs, it appears empirically that constructor
provided figures for erase/write timeouts are underestimated. A timeout
of 100ms seems to work with a 5 years worn chip, and no timeouts occur
anymore.
Signed-off-by: Robert Jarzmik <robert.jarzmik@free.fr>
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
The last erase block was not accessible, as the out of bound
check was incorrectly rejecting the last block.
The read/write/erase offset checks were forbidding the usage of the
last block, because of the calculation which was considering the
byte after the last instead of the last byte.
Signed-off-by: Robert Jarzmik <robert.jarzmik@free.fr>
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
doc_probe_device() is only called from docg3_probe() which is in .init.text,
so it must be in the same section to avoid a section mismatch warning.
Signed-off-by: Ryosuke Saito <raitosyo@gmail.com>
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Flash device drivers initialize 'ecc_strength' in struct mtd_info, which is the
maximum number of bit errors that can be corrected in one writesize region.
Drivers using the nand interface intitialize 'strength' in struct nand_ecc_ctrl,
which is the maximum number of bit errors that can be corrected in one ecc step.
Nand infrastructure code translates this to 'ecc_strength'.
Also for nand drivers, the nand infrastructure code sets ecc.strength for ecc
modes NAND_ECC_SOFT, NAND_ECC_SOFT_BCH, and NAND_ECC_NONE. It is set in the
driver for all other modes.
Signed-off-by: Mike Dunn <mikedunn@newsguy.com>
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
The read function was so far requiring the reads to be aligned on page
boundaries, and be page length multiples in size. Relieve these
constraints to ease the userspace ubifs programs runs, which read ubifs
headers of 64 bytes.
Artem: squashed a later fix from Robert Jarzmik into this patch.
Signed-off-by: Robert Jarzmik <robert.jarzmik@free.fr>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Change the name of the mtd so that it is simpler, and is easier to
cope with by mtdparts.
Signed-off-by: Robert Jarzmik <robert.jarzmik@free.fr>
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
The writebufsize concept was introduce by commit
"0e4ca7e mtd: add writebufsize field to mtd_info struct" and it represents
the maximum amount of data the device writes to the media at a time. This is
an important parameter for UBIFS which is used during recovery and which
basically defines how big a corruption caused by a power cut can be.
Set it to be equivalent to mtd->writesize because this is the maximum amount
of data the driver writes at a time.
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
Acked-by: Robert Jarzmik <robert.jarzmik@free.fr>
Cc: stable@kernel.org [3.2+]
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
This patch renames all MTD functions by adding a "_" prefix:
mtd->erase -> mtd->_erase
mtd->read_oob -> mtd->_read_oob
...
The reason is that we are re-working the MTD API and from now on it is
an error to use MTD function pointers directly - we have a corresponding
API call for every pointer. By adding a leading "_" we achieve the following:
1. Make sure we convert every direct pointer users
2. A leading "_" suggests that this interface is internal and it becomes
less likely that people will use them directly
3. Make sure all the out-of-tree modules stop compiling and the owners
spot the big API change and amend them.
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
The 'struct mtd_info' object is allocated with 'kzalloc()', so it
contains only zeroes - no need to initialize various fields to 0 or
NULL.
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
As the MTD api has no use for the number of erase cycles
each block has endured, remove the function which calculated
that value.
If one day MTD api finds it usefull for wear levelling
algorithms to have this information, the function should be
put back in place.
Signed-off-by: Robert Jarzmik <robert.jarzmik@free.fr>
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
This patch takes into account checkpatch, sparse and ECC
comments.
Signed-off-by: Robert Jarzmik <robert.jarzmik@free.fr>
Signed-off-by: Artem Bityutskiy <Artem.Bityutskiy@linux.intel.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
If doc_probe_device() returned an ERR_PTR, then we accidentally saved
that to docg3_floors[floor] = mtd; which gets derefenced in the error
handling when we call doc_release_device().
I've reworked the error handling to take care of that and hopefully
make it a little simpler.
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Acked-by: Robert Jarzmik <robert.jarzmik@free.fr>
Signed-off-by: Artem Bityutskiy <Artem.Bityutskiy@linux.intel.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
As each docg3 chip has 2 protection areas (DPS0 and DPS1),
and because theses areas can prevent user access to the chip
data, add for each floor the sysfs entries which insert the
protection key into the right DPS.
Signed-off-by: Robert Jarzmik <robert.jarzmik@free.fr>
Reviewed-by: Ivan Djelic <ivan.djelic@parrot.com>
Reviewed-by: Mike Dunn <mikedunn@newsguy.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Docg3 chips can work in 3 modes : normal MLC mode, fast
mode and reliable mode. Normally, as docg3 is a MLC chip, it
should be configured to work in normal mode.
In both normal mode, each page is distinct. This
means that writing to page 12 of blocks 14,15 writes only to
that page, and reading from page 12 of blocks 14,15 reads
only from that page.
In reliable and fast modes, pages are coupled by pairs, and
are clones one of each other. This means that the available
capacity of the chip is halved. Pages are coupled in each
block, and page of index 2*n contains the same data as page
2*n+1 of the same block.
In fast mode, the reads occur a bit faster, but are a bit
less reliable that in normal mode.
When reading from page 2*n, the chip reads bytes from both
page 2*n and page 2*n+1, makes a logical and for each byte,
and returns the result. As programming a page means
"clearing bits", even if a bit was not cleared on one page
because the flash is worn out, the other page has the bit
cleared, and the result of the "AND" gives a correct result.
When writing to page 2*n, the chip writes data to both page
2*n and page 2*n+1.
Signed-off-by: Robert Jarzmik <robert.jarzmik@free.fr>
Reviewed-by: Ivan Djelic <ivan.djelic@parrot.com>
Reviewed-by: Mike Dunn <mikedunn@newsguy.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Add functions to powerdown and powerup from suspend, in
order to save power.
Signed-off-by: Robert Jarzmik <robert.jarzmik@free.fr>
Reviewed-by: Ivan Djelic <ivan.djelic@parrot.com>
Reviewed-by: Mike Dunn <mikedunn@newsguy.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Credit for discovering the BCH algorith parameters, and bit
reversing algorithm is to be give to Mike Dunn and Ivan
Djelic.
The BCH correction code relied upon the BCH library, where
all data and ECC is bit-reversed. The BCH library works
correctly when each input byte is bit-reversed, and
accordingly ECC output is also bit-reversed.
Signed-off-by: Robert Jarzmik <robert.jarzmik@free.fr>
Reviewed-by: Ivan Djelic <ivan.djelic@parrot.com>
Reviewed-by: Mike Dunn <mikedunn@newsguy.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Map the developped write and erase functions into the mtd
structure.
Signed-off-by: Robert Jarzmik <robert.jarzmik@free.fr>
Reviewed-by: Ivan Djelic <ivan.djelic@parrot.com>
Reviewed-by: Mike Dunn <mikedunn@newsguy.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Add erase capability to the docg3 driver. The erase block is
made of 2 physical blocks, as both share all 64 pages. That
makes an erase block of at least 64 kBytes.
Signed-off-by: Robert Jarzmik <robert.jarzmik@free.fr>
Reviewed-by: Ivan Djelic <ivan.djelic@parrot.com>
Reviewed-by: Mike Dunn <mikedunn@newsguy.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Add write capability to the docg3 driver. The writes are
possible on a single page (512 bytes + 16 bytes), even if
that page is split on 2 physical pages on 2 blocks (each on
one plane).
Signed-off-by: Robert Jarzmik <robert.jarzmik@free.fr>
Reviewed-by: Ivan Djelic <ivan.djelic@parrot.com>
Reviewed-by: Mike Dunn <mikedunn@newsguy.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Add OOB layout description for docg3, so that userspace can
use this information to setup the data for write_oob().
Signed-off-by: Robert Jarzmik <robert.jarzmik@free.fr>
Reviewed-by: Ivan Djelic <ivan.djelic@parrot.com>
Reviewed-by: Mike Dunn <mikedunn@newsguy.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Add support for multiple floors, ie. cascaded docg3
chips. There might be 4 docg3 chips cascaded, sharing the
same address space, and providing up to 4 times the storage
capacity of a unique chip.
Each floor will be seen as an independant mtd device.
Signed-off-by: Robert Jarzmik <robert.jarzmik@free.fr>
Reviewed-by: Ivan Djelic <ivan.djelic@parrot.com>
Reviewed-by: Mike Dunn <mikedunn@newsguy.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Fix the docg3 reads to be able to cope with all possible
data buffer / oob buffer / file mode combinations from
docg3_read_oob().
This especially ensures that raw reads do not use ECC
corrections, and AUTOOOB and PLACEOOB do use ECC
correction.
The approach is to empty docg3_read() and make it a wrapper
to docg3_read_oob(). As docg3_read_oob() handles all the
funny cases (no data buffer but oob buffer, data buffer but
no oob buffer, ...), docg3_read() is just a special use of
docg3_read_oob().
Signed-off-by: Robert Jarzmik <robert.jarzmik@free.fr>
Reviewed-by: Ivan Djelic <ivan.djelic@parrot.com>
Reviewed-by: Mike Dunn <mikedunn@newsguy.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
The protection areas boundaries were on 16bit registers, not
8bit. This is consistent with block numbers, which can
extend up to 4096 on bigger chips (and is 2048 on the
docg3).
Signed-off-by: Robert Jarzmik <robert.jarzmik@free.fr>
Reviewed-by: Ivan Djelic <ivan.djelic@parrot.com>
Reviewed-by: Mike Dunn <mikedunn@newsguy.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Writeb was incorrectly traced as a 16 bits write, instead of
a 8 bits write. Fix it by tracing the correct width.
Signed-off-by: Robert Jarzmik <robert.jarzmik@free.fr>
Reviewed-by: Ivan Djelic <ivan.djelic@parrot.com>
Reviewed-by: Mike Dunn <mikedunn@newsguy.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Change the NOP debug log verbosity to very verbose to
unburden log analysis.
Signed-off-by: Robert Jarzmik <robert.jarzmik@free.fr>
Reviewed-by: Ivan Djelic <ivan.djelic@parrot.com>
Reviewed-by: Mike Dunn <mikedunn@newsguy.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Add support for DiskOnChip G3 chips. The support is quite
limited yet :
- no flash writes/erases are implemented
- ECC fixes are not implemented
- powerdown is not implemented
- IPL handling is not yet done
On the brighter side, the chip reading does work.
Signed-off-by: Robert Jarzmik <robert.jarzmik@free.fr>
