[PATCH v7 00/28] Allow vendor drivers to propose their own timings

From: Miquel Raynal <miquel.raynal@bootlin.com>
To: Richard Weinberger <richard@nod.at>,
	Vignesh Raghavendra <vigneshr@ti.com>,
	Tudor Ambarus <Tudor.Ambarus@microchip.com>,
	<linux-mtd@lists.infradead.org>
Cc: Michal Simek <monstr@monstr.eu>,
	Boris Brezillon <boris.brezillon@collabora.com>,
	Naga Sureshkumar Relli <nagasure@xilinx.com>,
	Thomas Petazzoni <thomas.petazzoni@bootlin.com>,
	Miquel Raynal <miquel.raynal@bootlin.com>
Subject: [PATCH v7 00/28] Allow vendor drivers to propose their own timings
Date: Fri, 29 May 2020 13:12:54 +0200	[thread overview]
Message-ID: <20200529111322.7184-1-miquel.raynal@bootlin.com> (raw)

As raised by Rickard, certain chips like Toshiba/Kioxia
TH58NVG2S3HBAI4 are not ONFI compliant and because of that, work at a
very slow pace. This chip in particular supports running at a pace
"close" to ONFI mode 4.

This series provides a way to manufacturer drivers to propose a data
interface to the core with a very simple interface. While at it, I did
some cleanup and prepared the introduction of NV-DDR timings.

Cheers,
Miquèl

Changes in v7:
* Fixed typos/reworded commit logs.
* Collected R-by tags.
* Added a patch to turn timing_mode an unsigned int (one parameter of
  onfi_fill_data_interface/interface_config).
* Thanks to Boris comments, I realized I could not introduce the
  "reset timing helper" before the change allocating the interface
  configuration dynamically. Nor it would make sense to do it
  afterwards, so I merged the two patches.
* s/nand_choose_interface/nand_choose_interface_config/.
* s/->choose_interface()/->choose_interface_config()/.
* Moved the patch hiding the ->data_interface indirection before the
  patch renaming data_interface into interface_config. This had the
  effect of dropping two patches (legacy_nand and marvell_nand
  updates).
* Dropped an always wrong condition in nand_reset() spotted by Boris.

Changes in v6:
* Collected R-by tags.
* Inserted a patch renaming massively the "data interface" name into
  "interface configuration". Followed this naming for the rest of the
  series, in particular for the new hooks.
* Re-added a comment about the specific tCCS, tR, tBERS and tPROG
  timings.
* Inverted the conditions in onfi_find_closest_sdr_mode() as advised.
* Created a helper to retrieve the reset interface configuration.
* Dropped nand_extract_sdr_timings() and kept nand_get_sdr_timings().
* Created nand_get_interface_config() instead, to retrieve an
  interface configuration provided a nand_chip object.
* Changed the scope of a variable.
* Turned onfi_fill_interface_config() into a void helper.
* Fixed typos.
* Used strncmp() to compare NAND chip model names.
* Prevented applying timing mode 0 twice.
* Created a ->current_interface_config entry in nand_chip, always
  pointing to the right interface configuration.
* Renamed the other entry ->best_interface_config.

Changes in v5:
* Collected R-by tags.
* Dropped the rename of onfi_timing_mode_default, got rid of it instead.
* Converted Hynix and Toshiba drivers to ->choose_data_interface().
* Created a static default reset data interface object.
* Allocated dynamically the timings.
* Renamed onfi_find_closest_sdr_mode() and its parameter.
* Dropped the comparison with maximum values in
  onfi_find_closest_sdr_mode().
* Use "best_timings" instead of "best_iface" as DDR is not supported
  yet.
* Dropped choose_best_vendor_sdr_timings() and unifed its content with
  choose_best_sdr_timings().
* If the specific timing set close to ONFI mode X is not supported,
  fallback to ONFI mode X first, not X - 1.

Changes in v4:
* Took Rickard patch based on my previous proposals over Github and
  tweaked a few more things:
  - Added a "generic" helper to fallback on slower ONFI modes when the
    proposed interface is not supported by the controller.
  - Fixed more kdoc.
  - Simplified the vendor driver side by providing additional helpers.
* Tweaked a little bit Rickard patch to fit the latest changes.

Miquel Raynal (27):
  mtd: rawnand: Use unsigned types for nand_chip unsigned values
  mtd: rawnand: Only use u8 instead of uint8_t in nand_chip structure
  mtd: rawnand: Create a nand_chip operations structure
  mtd: rawnand: Rename the manufacturer structure
  mtd: rawnand: Declare the nand_manufacturer structure out of nand_chip
  mtd: rawnand: Reorganize the nand_chip structure
  mtd: rawnand: Compare the actual timing values
  mtd: rawnand: Use the data interface mode entry when relevant
  mtd: rawnand: Rename nand_has_setup_data_iface()
  mtd: rawnand: Fix nand_setup_data_interface() description
  mtd: rawnand: Rename nand_init_data_interface()
  mtd: rawnand: timings: Update onfi_fill_data_interface() kernel doc
  mtd: rawnand: timings: Provide onfi_fill_data_interface() with a data
    interface
  mtd: rawnand: timings: onfi_fill_data_interface timing mode is
    unsigned
  mtd: rawnand: timings: Add a helper to find the closest ONFI mode
  mtd: rawnand: timings: Avoid redefining tR_max and tCCS_min
  mtd: rawnand: timings: Use default values for tPROG_max and tBERS_max
  mtd: rawnand: Hide the chip->data_interface indirection
  mtd: rawnand: s/data_interface/interface_config/
  mtd: rawnand: timings: Make onfi_fill_interface_config() a void helper
  mtd: rawnand: Introduce nand_choose_best_sdr_timings()
  mtd: rawnand: Add the ->choose_interface_config() hook
  mtd: rawnand: toshiba: Implement ->choose_interface_config() for
    TC58TEG5DCLTA00
  mtd: rawnand: toshiba: Implement ->choose_interface_config() for
    TC58NVG0S3E
  mtd: rawnand: hynix: Implement ->choose_interface_config() for
    H27UCG8T2ATR-BC
  mtd: rawnand: Get rid of the default ONFI timing mode
  mtd: rawnand: Allocate the interface configurations dynamically

Rickard x Andersson (1):
  mtd: rawnand: toshiba: Choose the interface configuration for
    TH58NVG2S3HBAI4

 drivers/mtd/nand/raw/ams-delta.c              |   6 +-
 drivers/mtd/nand/raw/arasan-nand-controller.c |   6 +-
 drivers/mtd/nand/raw/atmel/nand-controller.c  |  34 +--
 .../mtd/nand/raw/cadence-nand-controller.c    |   6 +-
 drivers/mtd/nand/raw/denali.c                 |   8 +-
 drivers/mtd/nand/raw/fsmc_nand.c              |   6 +-
 drivers/mtd/nand/raw/gpmi-nand/gpmi-nand.c    |   6 +-
 drivers/mtd/nand/raw/internals.h              |  22 +-
 drivers/mtd/nand/raw/marvell_nand.c           |  18 +-
 drivers/mtd/nand/raw/meson_nand.c             |  14 +-
 drivers/mtd/nand/raw/mtk_nand.c               |   6 +-
 drivers/mtd/nand/raw/mxc_nand.c               |  20 +-
 drivers/mtd/nand/raw/mxic_nand.c              |   6 +-
 drivers/mtd/nand/raw/nand_base.c              | 264 +++++++++--------
 drivers/mtd/nand/raw/nand_hynix.c             |  16 +-
 drivers/mtd/nand/raw/nand_ids.c               |  24 +-
 drivers/mtd/nand/raw/nand_legacy.c            |   7 +-
 drivers/mtd/nand/raw/nand_macronix.c          |  10 +-
 drivers/mtd/nand/raw/nand_micron.c            |   2 +-
 drivers/mtd/nand/raw/nand_timings.c           | 116 ++++++--
 drivers/mtd/nand/raw/nand_toshiba.c           |  72 ++++-
 drivers/mtd/nand/raw/s3c2410.c                |   6 +-
 drivers/mtd/nand/raw/stm32_fmc2_nand.c        |   6 +-
 drivers/mtd/nand/raw/sunxi_nand.c             |   6 +-
 drivers/mtd/nand/raw/tango_nand.c             |   4 +-
 drivers/mtd/nand/raw/tegra_nand.c             |   6 +-
 include/linux/mtd/rawnand.h                   | 270 +++++++++---------
 27 files changed, 579 insertions(+), 388 deletions(-)

-- 
2.20.1

______________________________________________________
Linux MTD discussion mailing list
http://lists.infradead.org/mailman/listinfo/linux-mtd/