Add support for sequential cache reads for controllers using the generic
core helpers for their fast read/write helpers.
Sequential reads may reduce the overhead when accessing physically
continuous data by loading in cache the next page while the previous
page gets sent out on the NAND bus.
The ONFI specification provides the following additional commands to
handle sequential cached reads:
* 0x31 - READ CACHE SEQUENTIAL:
Requires the NAND chip to load the next page into cache while keeping
the current cache available for host reads.
* 0x3F - READ CACHE END:
Tells the NAND chip this is the end of the sequential cache read, the
current cache shall remain accessible for the host but no more
internal cache loading operation is required.
On the bus, a multi page read operation is currently handled like this:
00 -- ADDR1 -- 30 -- WAIT_RDY (tR+tRR) -- DATA1_IN
00 -- ADDR2 -- 30 -- WAIT_RDY (tR+tRR) -- DATA2_IN
00 -- ADDR3 -- 30 -- WAIT_RDY (tR+tRR) -- DATA3_IN
Sequential cached reads may instead be achieved with:
00 -- ADDR1 -- 30 -- WAIT_RDY (tR) -- \
31 -- WAIT_RDY (tRCBSY+tRR) -- DATA1_IN \
31 -- WAIT_RDY (tRCBSY+tRR) -- DATA2_IN \
3F -- WAIT_RDY (tRCBSY+tRR) -- DATA3_IN
Below are the read speed test results with regular reads and
sequential cached reads, on NXP i.MX6 VAR-SOM-SOLO in mapping mode with
a NAND chip characterized with the following timings:
* tR: 20 µs
* tRCBSY: 5 µs
* tRR: 20 ns
and the following geometry:
* device size: 2 MiB
* eraseblock size: 128 kiB
* page size: 2 kiB
============= Normal read @ 33MHz =================
mtd_speedtest: eraseblock read speed is 15633 KiB/s
mtd_speedtest: page read speed is 15515 KiB/s
mtd_speedtest: 2 page read speed is 15398 KiB/s
===================================================
========= Sequential cache read @ 33MHz ===========
mtd_speedtest: eraseblock read speed is 18285 KiB/s
mtd_speedtest: page read speed is 15875 KiB/s
mtd_speedtest: 2 page read speed is 16253 KiB/s
===================================================
We observe an overall speed improvement of about 5% when reading
2 pages, up to 15% when reading an entire block. This is due to the
~14us gain on each additional page read (tR - (tRCBSY + tRR)).
Co-developed-by: Miquel Raynal <miquel.raynal@bootlin.com>
Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Signed-off-by: JaimeLiao <jaimeliao.tw@gmail.com>
Tested-by: Liao Jaime <jaimeliao.tw@gmail.com>
Link: https://lore.kernel.org/linux-mtd/20230112093637.987838-4-miquel.raynal@bootlin.com
New chips may feature a lot of CS because of their extended length. As
many controllers have been designed a decade ago, they usually only
feature just a couple. This does not mean that the entire range of
these chips cannot be accessed: it is just a matter of adding more
GPIO CS in the hardware design. A DT property has been added to
describe the CS array: cs-gpios.
Here is the code parsing it this new property, allocating what needs to
be, requesting the GPIOs and returning an array with the additional
available CS. The first entries of this array are left empty and are
reserved for native CS.
Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Link: https://lore.kernel.org/linux-mtd/20210526093242.183847-3-miquel.raynal@bootlin.com
Most timings related to the bus timings are different between SDR and
NV-DDR. However, we identified 9 individual timings which are more
related to the NAND chip internals. These are common between the two
interface types. Fortunately, only these common timings are being shared
through the NAND core and its ->exec_op() interface, which allows the
writing of a simple macro checking the interface type and depending on
it, returning either the relevant SDR timing or the NV-DDR timing. This
is the purpose of the NAND_COMMON_TIMING_PS() macro.
As all this is evaluated at build time, one will immediately be notified
in case a non common timing is being accessed through this macro.
Two handy macros are also inserted at the same time, which use
PSEC_TO_NSEC or PSEC_TO_MSEC so that it is very easy to return timings
in milli-, nano- or pico-seconds, as usually requested by the internal
API.
Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Link: https://lore.kernel.org/linux-mtd/20210505213750.257417-14-miquel.raynal@bootlin.com
On a typical end product, a vendor may choose to secure some regions in
the NAND memory which are supposed to stay intact between FW upgrades.
The access to those regions will be blocked by a secure element like
Trustzone. So the normal world software like Linux kernel should not
touch these regions (including reading).
The regions are declared using a NAND chip DT property,
"secure-regions". So let's make use of this property in the raw NAND
core and skip access to the secure regions present in a system.
Signed-off-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Link: https://lore.kernel.org/linux-mtd/20210402150128.29128-4-manivannan.sadhasivam@linaro.org
These functions must be usable by the main NAND core, so their names
must be technology-agnostic as well as the parameters. Hence, we pass
a generic nand_device instead of a raw nand_chip structure.
As it seems that changing the raw NAND functions to always pass a
generic NAND device is a lost of time, we prefer to create dedicated
raw NAND wrappers that will be useful in the near future to do the
translation.
Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Link: https://lore.kernel.org/linux-mtd/20200929230124.31491-8-miquel.raynal@bootlin.com
Instead of manipulating the statically allocated structure and copy
timings around, allocate one at identification time and save it in the
nand_chip structure once it has been initialized.
All NAND chips using the same interface configuration during reset and
startup, we define a helper to retrieve a single reset interface
configuration object, shared across all NAND chips.
We use a second pointer to always have a reference on the currently
applied interface configuration, which may either point to the "best
interface configuration" or to the "default reset interface
configuration".
Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Link: https://lore.kernel.org/linux-mtd/20200529111322.7184-29-miquel.raynal@bootlin.com
The name/suffix data_interface is a bit misleading in that the field
or functions actually represent a configuration that can be applied by
the controller/chip. Let's rename all fields/functions/hooks that are
worth renaming.
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Reorder fields in this structure and pack entries by theme:
* The main descriptive structures
* The data interface details
* Bad block information
* The device layout
* Extra buffers matching the device layout
* Internal values
* External objects like the ECC controller, the ECC engine and a
private data pointer.
While at it, adapt the documentation style.
I changed on purpose the description of @oob_poi which was weird.
Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Link: https://lore.kernel.org/linux-mtd/20200529111322.7184-7-miquel.raynal@bootlin.com
This scheme has been introduced for the Davinci controller and means
that the OOB area must be read *before* the rest of the data. This has
nothing to do with the ECC in OOB placement as it could be understood
and most importantly, there is no point in having this function out of
the Davinci NAND controller driver. A DT property for this scheme has
been added but never used, even by the Davinci driver which only uses
this scheme to change the default nand_read_page().
Move the main read_page() helper into the Davinci driver and remove
the remaining boilerplate.
Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Link: https://lore.kernel.org/linux-mtd/20200526195633.11543-4-miquel.raynal@bootlin.com
NAND controller drivers can set the NAND_USE_BOUNCE_BUFFER flag to a
chip 'option' field. With this flag, the core is responsible of
providing DMA-able buffers.
The current behavior is to not force the use of a bounce buffer when
the core thinks this is not needed. So in the end the name is a bit
misleading, because in theory we will always have a DMA buffer but in
practice it will not always be a bounce buffer.
Rename this flag NAND_USES_DMA to be more accurate.
Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Link: https://lore.kernel.org/linux-mtd/20200507105241.14299-4-miquel.raynal@bootlin.com
