The default partitions are missing in MX53 SPI NOR MTD.
It's because spi nor driver and mx53 msl include different
header files. It causes struct definitions are different.
For i.MX SPI nor driver, need to use asm/mach/flash.h by
aligning with NAND driver.
Signed-off-by: Lily Zhang <r58066@freescale.com>
Deleted the old drivers.
Broke the driver into separate files, for readability and to isolate
hardware dependencies.
Fixed bad block problems in the boot area for the i.MX23.
At this writing, UBI can't handle MTDs larger than 2GiB. If the
general use partition is larger than 2GiB, the driver will create
sub-partitions, none of which are larger than 2GiB.
Updated the default configs for the i.MX23 and i.MX28.
Other, miscellaneous changes.
Signed-off-by: Patrick Turley <patrick.turley@freescale.com>
This driver unifies the i.MX23 and i.MX28 NAND Flash drivers into a single
driver that supports both SoC's.
Signed-off-by: Patrick Turley <patrick.turley@freescale.com>
Changed the driver to detect ECC-based writes to the boot area and write a
byte to the OOB that matches the ROM behavior.
Signed-off-by: Patrick Turley <patrick.turley@freescale.com>
Changed the driver to ignore CONFIG_MTD_CONCAT, and work correctly with
CONFIG_MTD_PARTITIONS. Also turned on the GPMI driver and UBIFS in the
default config.
Signed-off-by: Patrick Turley <patrick.turley@freescale.com>
Uniform handling NFC INT bit clearance, the INT bit
will be unformed cleared in the wait_op_done function
after any NFC operation.
Signed-off-by:Jason Liu <r64343@freescale.com>
Update NFC INT wait timeout value to 1s.
1s should be enough for all platforms.
If INT bit not set for 1s timeout value, this
should be high potential of IC issues.
Signed-off-by:Jason Liu <r64343@freescale.com>
1.GPMI NAND flash scan scheme can overcome the shortcomings
of MTD NAND community code.
2.Put it to the common NAND directory can benifit I.MX NAND
driver besides GPMI NAND driver.
3.Fix the section mismatch build warnings on ALL platforms.
4.Fix the kbuild build errors of I.MX platforms.
Signed-off-by:Jason Liu <r64343@freescale.com>
1. Add wait timeout support to void dead loop in NAND driver,
2. Try best to use IRQ mode instead of POLLING mode
Signed-off-by:Jason Liu <r64343@freescale.com>
Write bootstream to kernel by kobs with BCH ECC.
Using the same combined Metadata & Block 0 BCH layout
with ROM code.
Page size 2048B; spare area 64B => ECC8, N=3, B0=512B, M=10B
Page size 4096B; spare area 128B => ECC8, N=7, B0=512B, M=10B
Page size 4096B; spare area 218B => ECC16, N=7, B0=512B, M=10B
Signed-off-by:Jason Liu <r64343@freescale.com>
Update NAND scan scheme to support new nand type. With this patch
Linux NAND driver can support new NAND flash on mx25/mx35 board while
compatible with old NAND on old boards.
Signed-off-by: Jason Liu <r64343@freescale.com>
No all regiser use _SCT. stmp3xxx_setl may write to wrong address.
please use __raw_write (value, REG_XXX_SET) to write date to SCT
register. Only REG_XXX_SCT is defined when it exist at hardware.
Signed-off-by: Frank.Li <Frank.Li@freescale.com>
Suspend then resume, kernel will print one warning message.
This is due to mtd suspend/resume framework change on 2.6.31 kernel.
Please refer to the follwing commit for detailed information
commit 15bce40cb3
[MTD] Restore suspend/resume support for mtd devices
Signed-off-by: Jason Liu <r64343@freescale.com>
Added the device identification and timing database. Adjusted the GPMI
NAND Flash driver to use this information.
Signed-off-by: Patrick Turley <patrick.turley@freescale.com>
Simplified the partitioning mechanism and, as a side effect, made the
GPMI NAND Flash driver usable (it was always crashing the kernel).
Signed-off-by: Patrick Turley <patrick.turley@freescale.com>
Nand status does not correct on MX51 TO2 and TO3.
1. Should use RBB_MODE 1 instead of 0 due to NFC bug(ENGcm09970)
For RBB_MODE 0,NFC has a bug,it doesn't sample correctly the
returned data on the I/O and therefore the INT is
asserted earlier than expected (i.e during the busy period)
2. Should use atomic status read instead of auto mode.
Auto mode status is not correct, the WP bit (IO7)
is not correct randomly, it wrongly report nand in write-protect
mode, which result in program/erase fail.
Signed-off-by:Jason Liu <r64343@freescale.com>
commit c5ce5b46af76f52dea21f467397d24c4ae6cb3ff upstream.
Do not use an unchecked variable UBI_IOCMKVOL ioctl.
Signed-off-by: Mika Westerberg <ext-mika.1.westerberg@nokia.com>
Signed-off-by: Artem Bityutskiy <Artem.Bityutskiy@nokia.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
commit 23af51ecfb04ff65bae51bd8e2270f4449abc789 upstream.
The M29W128G Numonyx flash devices are intolerant to any 0xFF command:
in the Cfi_util.c the function cfi_qry_mode_off() (that resets the device
after the autoselect mode) must have a 0xF0 command after the 0xFF command.
This fix solves also the cause of the fixup_M29W128G_write_buffer() fix,
that can be removed now.
The following patch applies to 2.6.30 kernel.
Signed-off-by: Massimo Cirillo <maxcir@gmail.com>
Acked-by: Alexey Korolev <akorolev@infradead.org>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
commit ebd5a74db74ee2db833d43ea35108a4be9cab42f upstream.
The previous implementation breaks the dts binding "mtd-physmap.txt". This
implementation fixes the issue by checking the availability of the reg
property instead of the name property.
Signed-off-by: Benjamin Krill <ben@codiert.org>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
commit 76c23c32e3b3ad48e07e07897075ab19ae1ef117 upstream.
Fix ECC Correction bug where the byte offset location were double
fliped causing correction routine to toggle the wrong byte location
in the ECC segment. The ndfc_calculate_ecc routine change the order
of getting the ECC code.
/* The NDFC uses Smart Media (SMC) bytes order */
ecc_code[0] = p[2];
ecc_code[1] = p[1];
ecc_code[2] = p[3];
But in the Correction algorithm when calculating the byte offset
location, the b1 is used as the upper part of the address. Which
again reverse the order making the final byte offset address
location incorrect.
byte_addr = (addressbits[b1] << 4) + addressbits[b0];
The order is change to read it in straight and let the correction
function to revert it to SMC order.
Signed-off-by: Feng Kan <fkan@amcc.com>
Acked-by: Victor Gallardo <vgallardo@amcc.com>
Acked-by: Prodyut Hazarika <phazarika@amcc.com>
Acked-by: Stefan Roese <sr@denx.de>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Arithmetic conversion in the mask computation makes the upper word
of the second argument passed down to mtd->read_oob(), be always 0
(assuming 'offs' being a 64-bit signed long long type, and
'mtd->writesize' being a 32-bit unsigned int type).
This patch applies over the other one adding masking in nftl_write,
"nftl: write support is broken".
Signed-off-by: Dimitri Gorokhovik <dimitri.gorokhovik@free.fr>
Cc: Tim Gardner <tim.gardner@canonical.com>
Cc: Scott James Remnant <scott@canonical.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
GCC 4.3.3 and 4.4.1 happily moves the dword load instruction out of the
loop in orion_nand_read_buf. This patch makes the instruction volatile
to avoid the issue. I've discussed this at gcc-help, refer to the thread
at
http://gcc.gnu.org/ml/gcc-help/2009-08/msg00187.html
The early clobber is added to avoid the destination registers and the
source register overlapping.
Signed-off-by: Simon Kagstrom <simon.kagstrom@netinsight.net>
Signed-off-by: Nicolas Pitre <nico@marvell.com>
The mtdblks array and its content are prone to race conditions. Introduce
the mutex mtdblks_lock in order to solve this.
[Amended by Artem Bityutskiy]
Signed-off-by: Matthias Kaehlcke <matthias@kaehlcke.net>
Signed-off-by: Artem Bityutskiy <Artem.Bityutskiy@nokia.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
Nowadays MTD devices have to be "get" before they can be
used. This has to be done with 'get_mtd_device()'. The
'blktrans_open()' function did not do this and instead
used 'try_module_get()'. Fix this.
Since 'get_mtd_device()' already gets the module, extra
'try_module_get()' is not needed.
This fixes oops when one tries to use mtdblock on top of
gluebi.
Reported-by: Holger Brunck <holger.brunck@keymile.com>
Signed-off-by: Artem Bityutskiy <Artem.Bityutskiy@nokia.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
