With CONFIG_PARTITION_NEED_MTD enabled we use mtd rather than devfs
directly to create partitions on mtd devices. Since:
| commit b32cd8df87
| Author: Sascha Hauer <s.hauer@pengutronix.de>
| Date: Wed Apr 9 15:49:32 2014 +0200
|
| mtd: nand: bb: use mtd api directly
|
| The devfs layer just adds an addition indirection between mtd
| and the bb devices with no purpose. Drop it.
|
| Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
The bad block aware device creation doesn't work if this option
is disabled. With this we remove CONFIG_PARTITION_NEED_MTD and always
use mtd partitions on mtd devices.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
We need list_for_each_entry_safe in dev_remove_bb_dev since the
list entries are removed during iteration over the list.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Reported-by: Alexander Aring <alex.aring@gmail.com>
When a mtd device can have bad blocks we want to create a
bb device, so do this automatically. This allows us to
drop bb device creation from the environment.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
This refactors the code so that we get a mtd_add_bb which can
be used to create a bb dev on a mtd_info.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
The devfs layer just adds an addition indirection between mtd
and the bb devices with no purpose. Drop it.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
If the ubidetach command is given a number that passes the range check
in ubi_detach_mtd_dev, but at which index the ubi_devices array contains
a NULL pointer, ubi_get_device tries to dereference this NULL pointer
to increment the ubi device reference count.
Signed-off-by: Philipp Zabel <p.zabel@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Was tested on NAND with {writesize=4096, oobsize=224} and
{writesize=2048, oobsize=64}.
This patch will not break any NAND that was working
before. Implemented calculation way may be used for other NAND chips
with writesize == 2048 but oobsize != 64.
Signed-off-by: Dmitry Lavnikevich <d.lavnikevich@sam-solutions.com>
Signed-off-by: Grigory Milev <g.milev@sam-solutions.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Since i.MX6 has only one R/B actual pin, if there are several R/B
signals (from different NAND chips) they must be connected to this
pin.
Signed-off-by: Dmitry Lavnikevich <d.lavnikevich@sam-solutions.com>
Signed-off-by: Grigory Milev <g.milev@sam-solutions.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Since MTD is already supports >= 4GB device sizes it is better to
allow ubi valumes also be created of larger sizes.
Signed-off-by: Dmitry Lavnikevich <d.lavnikevich@sam-solutions.com>
Signed-off-by: Grigory Milev <g.milev@sam-solutions.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
MTD internal API presently uses 32-bit values to represent device
size. This patch updates them to 64-bits but leaves the external API
unchanged.
In general, changing from 32-bit to 64-bit values cause little
or no changes to the majority of the code with the following
exceptions:
- printk message formats;
- division and modulus of 64-bit values (mtd_div_by_wb,
mtd_div_by_eb may be used in some of such cases).
Was tested on phyFLEX i.MX6.
Signed-off-by: Dmitry Lavnikevich <d.lavnikevich@sam-solutions.com>
Signed-off-by: Grigory Milev <g.milev@sam-solutions.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
I am using an sam5d35 and its pmecc nand ecc controller.
In barebox, using similar atmel_nand.c the system crashes.
The problem was a call to chip->ecc.hwctl from nand_write_subpage_hwecc
(nand_base.c) when we write a sub page.
chip->ecc.hwctl function is not set when we are using PMECC controller.
As a workaround, i set NAND_NO_SUBPAGE_WRITE for PMECC controller in
order to disable sub page access in nand_write_page.
I think we can have the same problem in Linux. I do not performed
subpage nand writes from Linux but i made the following patch according
to my Barebox defect.
Signed-off-by: Herve Codina <Herve.CODINA@celad.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
By default UBI is very verbose. Decrease the verbosity, turn
printf messages into ubi_ messages and add device parameters
for the values which are not printed anymore.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
We need to setup the info->dev pointer before flash_get_size() since it
might use it for dev_dbg().
This fix a crash on boot when debug output is enabled.
Signed-off-by: Philippe Rétornaz <philippe.retornaz@epfl.ch>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
The string pointed to isn't modified, so declare it as const.
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
struct ubi_device and struct ubi_volume have devices attached
to them. Register them to make them part of the barebox device
hierarchy.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
So far mtd partitions were mtd devices, but these were not registered.
This patch changes this. mtd partitions are now registered like real
mtd devices. This makes them part of the device hierarchy.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
struct mtd_info contains members which should not be copied
to the new partition, like for example the class_dev, so
only copy selected members of struct mtd_info to the partition.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
mtd->parent is optional so we can't derefence mtd->parent->device_node
without checking mtd->parent.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
erase_info may be modified by the mtd partition code, so
set the address each time we call mtd_erase.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
This sets the id_table in the driver correctly so that the driver
can match to the various types of flashes, not only "m25p80".
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Signed-off-by: Lucas Stach <l.stach@pengutronix.de>
Add more paranioa asserts to make it easier to detect
implementation errors.
Signed-off-by: Richard Weinberger <richard@nod.at>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
On error we have to free all three temporary lists.
Reported-by: Richard Genoud <richard.genoud@gmail.com>
Signed-off-by: Richard Weinberger <richard@nod.at>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
The test:
if (!a && b)
a = b;
can be symplified in:
if (!a)
a = b;
And there's no need to test if ubi->image_seq is not null, because if it is,
it is set to image_seq.
So, we just test if image_seq is not null.
Signed-off-by: Richard Genoud <richard.genoud@gmail.com>
Signed-off-by: Richard Weinberger <richard@nod.at>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Some old UBI implementations (e.g. U-Boot) have not implemented the image
sequence feature.
So, when erase blocks are written, the image sequence in the ec header
is lost (set to zero).
UBI scan_all() takes this case into account (commits
32bc4820287a1a03982979515949e8ea56eac641 and
2eadaad67b2b6bd132eda105128d2d466298b8e3)
But fastmap scan functions (ubi_scan_fastmap() and scan_pool()) didn't.
This patch fixes the issue.
Signed-off-by: Richard Genoud <richard.genoud@gmail.com>
Signed-off-by: Richard Weinberger <richard@nod.at>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
If we find an invalid fastmap we have to scan from the very beginning.
Otherwise we leak the first 64 PEBs.
Reported-and-tested-by: Richard Genoud <richard.genoud@gmail.com>
Signed-off-by: Richard Weinberger <richard@nod.at>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
We have to set "ret", not "err" in case of an error.
Reported-and-tested-by: Richard Genoud <richard.genoud@gmail.com>
Signed-off-by: Richard Weinberger <richard@nod.at>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
If no free PEBs are available refill_wl_user_pool() must not
return with -ENOSPC immediately.
It has to block till produce_free_peb() produced a free PEB.
Reported-and-Tested-by: Richard Genoud <richard.genoud@gmail.com>
Signed-off-by: Richard Weinberger <richard@nod.at>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Onging tests uncovered that invalidate_fastmap() is broken.
It must not call ubi_wl_put_fm_peb() because all PEBs used
by the old fastmap have already been put back.
Signed-off-by: Richard Weinberger <richard@nod.at>
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
get_peb_for_wl() removes the PEB from the free list.
If the WL subsystem detects that no wear leveling is needed
it cancels the operation and drops the gained PEB.
In this case we have to put the PEB back into the free list.
This issue was introduced with commit ed4b7021c
(UBI: remove PEB from free tree in get_peb_for_wl()).
Cc: <stable@vger.kernel.org> # 3.7.x
Signed-off-by: Richard Weinberger <richard@nod.at>
Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
Sascha Hauer
Alexander Shiyan
Philipp Zabel
Dmitry Lavnikevich
Herve Codina
Philippe Rétornaz
Uwe Kleine-König
Lucas Stach
Cristiano De Alti
Richard Weinberger
Richard Genoud