platform/kernel/linux-rpi.git
4 years agospi: Only defer to thread for cleanup when needed
Mark Brown [Wed, 15 Jul 2020 16:36:10 +0000 (17:36 +0100)]
spi: Only defer to thread for cleanup when needed

Currently we always defer idling of controllers to the SPI thread, the goal
being to ensure that we're doing teardown that's not suitable for atomic
context in an appropriate context and to try to batch up more expensive
teardown operations when the system is under higher load, allowing more
work to be started before the SPI thread is scheduled. However when the
controller does not require any substantial work to idle there is no need
to do this, we can instead save the context switch and immediately mark
the controller as idle. This is particularly useful for systems where there
is frequent but not constant activity.

Signed-off-by: Mark Brown <broonie@kernel.org>
Link: https://lore.kernel.org/r/20200715163610.9475-1-broonie@kernel.org
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agospi: fix duplicated word in <linux/spi/spi.h>
Randy Dunlap [Thu, 16 Jul 2020 01:30:48 +0000 (18:30 -0700)]
spi: fix duplicated word in <linux/spi/spi.h>

Change doubled word "as" to "as a".

Change "Return: Return:" in kernel-doc notation to have only one
"Return:".

Signed-off-by: Randy Dunlap <rdunlap@infradead.org>
Cc: Mark Brown <broonie@kernel.org>
Cc: linux-spi@vger.kernel.org
Link: https://lore.kernel.org/r/40354d64-be71-3952-a980-63a76a278145@infradead.org
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agospi: omap-uwire: Use clk_prepare_enable and clk_disable_unprepare
Qing Zhang [Wed, 15 Jul 2020 05:26:46 +0000 (13:26 +0800)]
spi: omap-uwire: Use clk_prepare_enable and clk_disable_unprepare

Convert clk_enable() to clk_prepare_enable() and clk_disable() to
clk_disable_unprepare() respectively in the spi-omap-uwire.c.

Signed-off-by: Qing Zhang <zhangqing@loongson.cn>
Link: https://lore.kernel.org/r/1594790807-32319-1-git-send-email-zhangqing@loongson.cn
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agoMerge series "mtd: spi-nor: add xSPI Octal DTR support" from Pratyush Yadav <p.yadav...
Mark Brown [Tue, 14 Jul 2020 16:38:47 +0000 (17:38 +0100)]
Merge series "mtd: spi-nor: add xSPI Octal DTR support" from Pratyush Yadav <p.yadav@ti.com>:

Hi,

This series adds support for octal DTR flashes in the spi-nor framework,
and then adds hooks for the Cypress Semper and Mircom Xcella flashes to
allow running them in octal DTR mode. This series assumes that the flash
is handed to the kernel in Legacy SPI mode.

Tested on TI J721e EVM with 1-bit ECC on the Cypress flash.

Changes in v10:
- Rebase on latest linux-next/master. Drop a couple patches that made it
  in the  previous release.

- Move the code that sets 20 dummy cycles for MT35XU512ABA to its octal
  enable function. This way, if the controller doesn't support 8D mode
  20 dummy cycles won't be used.

Changes in v9:
- Do not use '& 0xff' to get the opcode LSB in spi-mxic and
  spi-zynq-qspi. The cast to u8 will do that anyway.

- Do not use if (opcode) as a check for whether the command phase exists
  in spi-zynq-qspi because the opcode 0 can be valid. Use the new
  cmd.nbytes instead.

Changes in v8:
- Move controller changes in spi-mxic to the commit which introduces
  2-byte opcodes to avoid problems when bisecting.

- Replace usage of sizeof(op->cmd.opcode) with op->cmd.nbytes.

- Extract opcode in spi-zynq-qspi instead of using &op->cmd.opcode.

Changes in v7:
- Reject ops with more than 1 command byte in
  spi_mem_default_supports_op().

- Reject ops with more than 1 command byte in atmel and mtk controllers.

- Reject ops with 0 command bytes in spi_mem_check_op().

- Set cmd.nbytes to 1 when using SPI_MEM_OP_CMD().

- Avoid endianness problems in spi-mxic.

Changes in v6:
- Instead of hard-coding 8D-8D-8D Fast Read dummy cycles to 20, find
  them out from the Profile 1.0 table.

Changes in v5:
- Do not enable stateful X-X-X modes if the reset line is broken.

- Instead of setting SNOR_READ_HWCAPS_8_8_8_DTR from Profile 1.0 table
  parsing, do it in spi_nor_info_init_params() instead based on the
  SPI_NOR_OCTAL_DTR_READ flag instead.

- Set SNOR_HWCAPS_PP_8_8_8_DTR in s28hs post_sfdp hook since this
  capability is no longer set in Profile 1.0 parsing.

- Instead of just checking for spi_nor_get_protocol_width() in
  spi_nor_octal_dtr_enable(), make sure the protocol is
  SNOR_PROTO_8_8_8_DTR since get_protocol_width() only cares about data
  width.

- Drop flag SPI_NOR_SOFT_RESET. Instead, discover soft reset capability
  via BFPT.

- Do not make an invalid Quad Enable BFPT field a fatal error. Silently
  ignore it by assuming no quad enable bit is present.

- Set dummy cycles for Cypress Semper flash to 24 instead of 20. This
  allows for 200MHz operation in 8D mode compared to the 166MHz with 20.

- Rename spi_nor_cypress_octal_enable() to
  spi_nor_cypress_octal_dtr_enable().

- Update spi-mtk-nor.c to reject DTR ops since it doesn't call
  spi_mem_default_supports_op().

Changes in v4:
- Refactor the series to use the new spi-nor framework with the
  manufacturer-specific bits separated from the core.

- Add support for Micron MT35XU512ABA.

- Use cmd.nbytes as the criteria of whether the data phase exists or not
  instead of cmd.buf.in || cmd.buf.out in spi_nor_spimem_setup_op().

- Update Read FSR to use the same dummy cycles and address width as Read
  SR.

- Fix BFPT parsing stopping too early for JESD216 rev B flashes.

- Use 2 byte reads for Read SR and FSR commands in DTR mode.

Changes in v3:
- Drop the DT properties "spi-rx-dtr" and "spi-tx-dtr". Instead, if
  later a need is felt to disable DTR in case someone has a board with
  Octal DTR capable flash but does not support DTR transactions for some
  reason, a property like "spi-no-dtr" can be added.

- Remove mode bits SPI_RX_DTR and SPI_TX_DTR.

- Remove the Cadence Quadspi controller patch to un-block this series. I
  will submit it as a separate patch.

- Rebase on latest 'master' and fix merge conflicts.

- Update read and write dirmap templates to use DTR.

- Rename 'is_dtr' to 'dtr'.

- Make 'dtr' a bitfield.

- Reject DTR ops in spi_mem_default_supports_op().

- Update atmel-quadspi to reject DTR ops. All other controller drivers
  call spi_mem_default_supports_op() so they will automatically reject
  DTR ops.

- Add support for both enabling and disabling DTR modes.

- Perform a Software Reset on flashes that support it when shutting
  down.

- Disable Octal DTR mode on suspend, and re-enable it on resume.

- Drop enum 'spi_mem_cmd_ext' and make command opcode u16 instead.
  Update spi-nor to use the 2-byte command instead of the command
  extension. Since we still need a "extension type", mode that enum to
  spi-nor and name it 'spi_nor_cmd_ext'.

- Default variable address width to 3 to fix SMPT parsing.

- Drop non-volatile change to uniform sector mode and rely on parsing
  SMPT.

Changes in v2:
- Add DT properties "spi-rx-dtr" and "spi-tx-dtr" to allow expressing
  DTR capabilities.

- Set the mode bits SPI_RX_DTR and SPI_TX_DTR when we discover the DT
  properties "spi-rx-dtr" and spi-tx-dtr".

- spi_nor_cypress_octal_enable() was updating nor->params.read[] with
  the intention of setting the correct number of dummy cycles. But this
  function is called _after_ selecting the read so setting
  nor->params.read[] will have no effect. So, update nor->read_dummy
  directly.

- Fix spi_nor_spimem_check_readop() and spi_nor_spimem_check_pp()
  passing nor->read_proto and nor->write_proto to
  spi_nor_spimem_setup_op() instead of read->proto and pp->proto
  respectively.

- Move the call to cqspi_setup_opcode_ext() inside cqspi_enable_dtr().
  This avoids repeating the 'if (f_pdata->is_dtr)
  cqspi_setup_opcode_ext()...` snippet multiple times.

- Call the default 'supports_op()' from cqspi_supports_mem_op(). This
  makes sure the buswidth requirements are also enforced along with the
  DTR requirements.

- Drop the 'is_dtr' argument from spi_check_dtr_req(). We only call it
  when a phase is DTR so it is redundant.

Pratyush Yadav (17):
  spi: spi-mem: allow specifying whether an op is DTR or not
  spi: spi-mem: allow specifying a command's extension
  spi: atmel-quadspi: reject DTR ops
  spi: spi-mtk-nor: reject DTR ops
  mtd: spi-nor: add support for DTR protocol
  mtd: spi-nor: sfdp: get command opcode extension type from BFPT
  mtd: spi-nor: sfdp: parse xSPI Profile 1.0 table
  mtd: spi-nor: core: use dummy cycle and address width info from SFDP
  mtd: spi-nor: core: do 2 byte reads for SR and FSR in DTR mode
  mtd: spi-nor: core: enable octal DTR mode when possible
  mtd: spi-nor: sfdp: do not make invalid quad enable fatal
  mtd: spi-nor: sfdp: detect Soft Reset sequence support from BFPT
  mtd: spi-nor: core: perform a Soft Reset on shutdown
  mtd: spi-nor: core: disable Octal DTR mode on suspend.
  mtd: spi-nor: core: expose spi_nor_default_setup() in core.h
  mtd: spi-nor: spansion: add support for Cypress Semper flash
  mtd: spi-nor: micron-st: allow using MT35XU512ABA in Octal DTR mode

 drivers/mtd/spi-nor/core.c      | 446 +++++++++++++++++++++++++++-----
 drivers/mtd/spi-nor/core.h      |  22 ++
 drivers/mtd/spi-nor/micron-st.c | 103 +++++++-
 drivers/mtd/spi-nor/sfdp.c      | 131 +++++++++-
 drivers/mtd/spi-nor/sfdp.h      |   8 +
 drivers/mtd/spi-nor/spansion.c  | 166 ++++++++++++
 drivers/spi/atmel-quadspi.c     |   6 +
 drivers/spi/spi-mem.c           |  16 +-
 drivers/spi/spi-mtk-nor.c       |  10 +-
 drivers/spi/spi-mxic.c          |   3 +-
 drivers/spi/spi-zynq-qspi.c     |  11 +-
 include/linux/mtd/spi-nor.h     |  53 +++-
 include/linux/spi/spi-mem.h     |  14 +-
 13 files changed, 889 insertions(+), 100 deletions(-)

--
2.27.0

base-commit: b3a9e3b9622ae10064826dccb4f7a52bd88c7407

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

4 years agospi: omap-100k: Drop include
Linus Walleij [Tue, 14 Jul 2020 07:33:57 +0000 (09:33 +0200)]
spi: omap-100k: Drop include

The OMAP-100k driver includes <linux/gpio.h> but does not
use any symbols from it, so drop the include.

Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Link: https://lore.kernel.org/r/20200714073357.34879-1-linus.walleij@linaro.org
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agospi: spi-mtk-nor: reject DTR ops
Pratyush Yadav [Tue, 23 Jun 2020 18:30:17 +0000 (00:00 +0530)]
spi: spi-mtk-nor: reject DTR ops

Double Transfer Rate (DTR) ops are added in spi-mem. But this controller
doesn't support DTR transactions. Since we don't use the default
supports_op(), which rejects all DTR ops, do that explicitly in our
supports_op().

Signed-off-by: Pratyush Yadav <p.yadav@ti.com>
Reviewed-by: Tudor Ambarus <tudor.ambarus@microchip.com>
Link: https://lore.kernel.org/r/20200623183030.26591-5-p.yadav@ti.com
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agospi: atmel-quadspi: reject DTR ops
Pratyush Yadav [Tue, 23 Jun 2020 18:30:16 +0000 (00:00 +0530)]
spi: atmel-quadspi: reject DTR ops

Double Transfer Rate (DTR) ops are added in spi-mem. But this controller
doesn't support DTR transactions. Since we don't use the default
supports_op(), which rejects all DTR ops, do that explicitly in our
supports_op().

Signed-off-by: Pratyush Yadav <p.yadav@ti.com>
Reviewed-by: Tudor Ambarus <tudor.ambarus@microchip.com>
Link: https://lore.kernel.org/r/20200623183030.26591-4-p.yadav@ti.com
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agospi: spi-mem: allow specifying a command's extension
Pratyush Yadav [Tue, 23 Jun 2020 18:30:15 +0000 (00:00 +0530)]
spi: spi-mem: allow specifying a command's extension

In xSPI mode, flashes expect 2-byte opcodes. The second byte is called
the "command extension". There can be 3 types of extensions in xSPI:
repeat, invert, and hex. When the extension type is "repeat", the same
opcode is sent twice. When it is "invert", the second byte is the
inverse of the opcode. When it is "hex" an additional opcode byte based
is sent with the command whose value can be anything.

So, make opcode a 16-bit value and add a 'nbytes', similar to how
multiple address widths are handled.

Some places use sizeof(op->cmd.opcode). Replace them with op->cmd.nbytes

The spi-mxic and spi-zynq-qspi drivers directly use op->cmd.opcode as a
buffer. Now that opcode is a 2-byte field, this can result in different
behaviour depending on if the machine is little endian or big endian.
Extract the opcode in a local 1-byte variable and use that as the buffer
instead. Both these drivers would reject multi-byte opcodes in their
supports_op() hook anyway, so we only need to worry about single-byte
opcodes for now.

The above two changes are put in this commit to keep the series
bisectable.

Signed-off-by: Pratyush Yadav <p.yadav@ti.com>
Reviewed-by: Tudor Ambarus <tudor.ambarus@microchip.com>
Link: https://lore.kernel.org/r/20200623183030.26591-3-p.yadav@ti.com
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agospi: spi-mem: allow specifying whether an op is DTR or not
Pratyush Yadav [Tue, 23 Jun 2020 18:30:14 +0000 (00:00 +0530)]
spi: spi-mem: allow specifying whether an op is DTR or not

Each phase is given a separate 'dtr' field so mixed protocols like
4S-4D-4D can be supported.

Signed-off-by: Pratyush Yadav <p.yadav@ti.com>
Reviewed-by: Tudor Ambarus <tudor.ambarus@microchip.com>
Link: https://lore.kernel.org/r/20200623183030.26591-2-p.yadav@ti.com
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agospi: imx/fsl-lpspi: Convert to GPIO descriptors
Linus Walleij [Thu, 25 Jun 2020 20:02:52 +0000 (22:02 +0200)]
spi: imx/fsl-lpspi: Convert to GPIO descriptors

This converts the two Freescale i.MX SPI drivers
Freescale i.MX (CONFIG_SPI_IMX) and Freescale i.MX LPSPI
(CONFIG_SPI_FSL_LPSPI) to use GPIO descriptors handled in
the SPI core for GPIO chip selects whether defined in
the device tree or a board file.

The reason why both are converted at the same time is
that they were both using the same platform data and
platform device population helpers when using
board files intertwining the code so this gives a cleaner
cut.

The platform device creation was passing a platform data
container from each boardfile down to the driver using
struct spi_imx_master from <linux/platform_data/spi-imx.h>,
but this was only conveying the number of chipselects and
an int * array of the chipselect GPIO numbers.

The imx27 and imx31 platforms had code passing the
now-unused platform data when creating the platform devices,
this has been repurposed to pass around GPIO descriptor
tables. The platform data struct that was just passing an
array of integers and number of chip selects for the GPIO
lines has been removed.

The number of chipselects used to be passed from the board
file, because this number also limits the number of native
chipselects that the platform can use. To deal with this we
just augment the i.MX (CONFIG_SPI_IMX) driver to support 3
chipselects if the platform does not define "num-cs" as a
device property (such as from the device tree). This covers
all the legacy boards as these use <= 3 native chip selects
(or GPIO lines, and in that case the number of chip selects
is determined by the core from the number of available
GPIO lines). Any new boards should use device tree, so
this is a reasonable simplification to cover all old
boards.

The LPSPI driver never assigned the number of chipselects
and thus always fall back to the core default of 1 chip
select if no GPIOs are defined in the device tree.

The Freescale i.MX driver was already partly utilizing
the SPI core to obtain the GPIO numbers from the device tree,
so this completes the transtion to let the core handle all
of it.

All board files and the core i.MX boardfile registration
code is augmented to account for these changes.

This has been compile-tested with the imx_v4_v5_defconfig
and the imx_v6_v7_defconfig.

Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Acked-by: Shawn Guo <shawnguo@kernel.org>
Cc: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Cc: Robin Gong <yibin.gong@nxp.com>
Cc: Trent Piepho <tpiepho@impinj.com>
Cc: Clark Wang <xiaoning.wang@nxp.com>
Cc: Shawn Guo <shawnguo@kernel.org>
Cc: Sascha Hauer <s.hauer@pengutronix.de>
Cc: Pengutronix Kernel Team <kernel@pengutronix.de>
Cc: Fabio Estevam <festevam@gmail.com>
Cc: NXP Linux Team <linux-imx@nxp.com>
Link: https://lore.kernel.org/r/20200625200252.207614-1-linus.walleij@linaro.org
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agodt-bindings: spi: Convert imx lpspi to json-schema
Anson Huang [Tue, 16 Jun 2020 04:26:15 +0000 (12:26 +0800)]
dt-bindings: spi: Convert imx lpspi to json-schema

Convert the i.MX LPSPI binding to DT schema format using json-schema

Signed-off-by: Anson Huang <Anson.Huang@nxp.com>
Reviewed-by: Rob Herring <robh@kernel.org>
Link: https://lore.kernel.org/r/1592281575-32708-4-git-send-email-Anson.Huang@nxp.com
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agodt-bindings: spi: Convert imx cspi to json-schema
Anson Huang [Tue, 16 Jun 2020 04:26:14 +0000 (12:26 +0800)]
dt-bindings: spi: Convert imx cspi to json-schema

Convert the i.MX CSPI binding to DT schema format using json-schema,
update compatible, remove obsolete properties "fsl,spi-num-chipselects"
and update the example based on latest DT file.

Signed-off-by: Anson Huang <Anson.Huang@nxp.com>
Reviewed-by: Rob Herring <robh@kernel.org>
Link: https://lore.kernel.org/r/1592281575-32708-3-git-send-email-Anson.Huang@nxp.com
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agodt-bindings: spi: Convert mxs spi to json-schema
Anson Huang [Tue, 16 Jun 2020 04:26:13 +0000 (12:26 +0800)]
dt-bindings: spi: Convert mxs spi to json-schema

Convert the MXS SPI binding to DT schema format using json-schema

Signed-off-by: Anson Huang <Anson.Huang@nxp.com>
Reviewed-by: Rob Herring <robh@kernel.org>
Link: https://lore.kernel.org/r/1592281575-32708-2-git-send-email-Anson.Huang@nxp.com
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agoSPI SUBSYSTEM: Replace HTTP links with HTTPS ones
Alexander A. Klimov [Wed, 8 Jul 2020 19:44:00 +0000 (21:44 +0200)]
SPI SUBSYSTEM: Replace HTTP links with HTTPS ones

Rationale:
Reduces attack surface on kernel devs opening the links for MITM
as HTTPS traffic is much harder to manipulate.

Deterministic algorithm:
For each file:
  If not .svg:
    For each line:
      If doesn't contain `\bxmlns\b`:
        For each link, `\bhttp://[^# \t\r\n]*(?:\w|/)`:
  If neither `\bgnu\.org/license`, nor `\bmozilla\.org/MPL\b`:
            If both the HTTP and HTTPS versions
            return 200 OK and serve the same content:
              Replace HTTP with HTTPS.

Signed-off-by: Alexander A. Klimov <grandmaster@al2klimov.de>
Link: https://lore.kernel.org/r/20200708194400.22213-1-grandmaster@al2klimov.de
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agospi: atmel: remove redundant label out_free
Colin Ian King [Thu, 9 Jul 2020 10:12:03 +0000 (11:12 +0100)]
spi: atmel: remove redundant label out_free

The error exit label out_free is no longer being used, it is redundant
and can be removed.

Cleans up warning:
drivers/spi/spi-atmel.c:1680:1: warning: label ‘out_free’ defined but not used [-Wunused-label]

Fixes: 2d9a744685bc ("spi: atmel: No need to call spi_master_put() if spi_alloc_master() failed")
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Link: https://lore.kernel.org/r/20200709101203.1374117-1-colin.king@canonical.com
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agospi: use kthread_create_worker() helper
Marek Szyprowski [Thu, 9 Jul 2020 06:50:07 +0000 (08:50 +0200)]
spi: use kthread_create_worker() helper

Use kthread_create_worker() helper to simplify the code. It uses
the kthread worker API the right way. It will eventually allow
to remove the FIXME in kthread_worker_fn() and add more consistency
checks in the future.

Signed-off-by: Marek Szyprowski <m.szyprowski@samsung.com>
Reviewed-by: Petr Mladek <pmladek@suse.com>
Link: https://lore.kernel.org/r/20200709065007.26896-1-m.szyprowski@samsung.com
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agoMerge series "spi: spi-geni-qcom: Avoid a bunch of per-transfer overhead" from Dougla...
Mark Brown [Tue, 7 Jul 2020 13:26:34 +0000 (14:26 +0100)]
Merge series "spi: spi-geni-qcom: Avoid a bunch of per-transfer overhead" from Douglas Anderson <dianders@chromium.org>:

This series tries to reduce a whole bunch of overhead in each SPI
transfer.  Much of this overhead is new with the recent interconnect
changes, but even without those changes we still had some overhead
that we could avoid.  Let's avoid all of it.

These changes are atop the Qualcomm tree to avoid merge conflicts.  If
they look good, the most expedient way to land them is probably to get
Ack's from Mark and land then via the Qualcomm tree.

Most testing was done on the Chrome OS 5.4 tree, but sanity check was
done on mainline.

Douglas Anderson (3):
  spi: spi-geni-qcom: Avoid clock setting if not needed
  spi: spi-geni-qcom: Set an autosuspend delay of 250 ms
  spi: spi-geni-qcom: Get rid of most overhead in prepare_message()

 drivers/spi/spi-geni-qcom.c | 67 ++++++++++++++++++-------------------
 1 file changed, 32 insertions(+), 35 deletions(-)

--
2.27.0.383.g050319c2ae-goog

4 years agoMerge series "spi: spi-sun6i: One fix and some improvements" from Marc Kleine-Budde...
Mark Brown [Tue, 7 Jul 2020 13:26:33 +0000 (14:26 +0100)]
Merge series "spi: spi-sun6i: One fix and some improvements" from Marc Kleine-Budde <mkl@pengutronix.de>:

Hello,

this series first fixes the calculation of the clock rate. The driver will
round up to the nearest clock rate instead of rounding down. Resulting in SPI
devices accessed with a too high SPI clock.

The remaining patches improve the performance of the driver. The changes range
from micro-optimizations like reducing MMIO writes to the controller to
reducing the number of needed interrupts in some use cases.

regards,
Marc

changes since v1:
- added Maxime Ripard's to the existing patches
- 06/10: (was 05/10 in v1)
  "spi: spi-sun6i: sun6i_spi_drain_fifo(): introduce sun6i_spi_get_rx_fifo_count() and make use of it"
  use FIELD_GET instead of open coding it
  (tnx: Maxime Ripard)
- 05/10: "spi: spi-sun6i: sun6i_spi_get_tx_fifo_count: Convert manual shift+mask to FIELD_GET()"
  new patch

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

4 years agospi: atmel: No need to call spi_master_put() if spi_alloc_master() failed
Peng Fan [Tue, 7 Jul 2020 08:50:42 +0000 (16:50 +0800)]
spi: atmel: No need to call spi_master_put() if spi_alloc_master() failed

There is no need to call spi_master_put() if spi_alloc_master()
failed, it should return -ENOMEM directly.

Signed-off-by: Peng Fan <fanpeng@loongson.cn>
Link: https://lore.kernel.org/r/1594111842-9468-1-git-send-email-fanpeng@loongson.cn
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agospi: spi-geni-qcom: Set an autosuspend delay of 250 ms
Douglas Anderson [Thu, 2 Jul 2020 00:45:08 +0000 (17:45 -0700)]
spi: spi-geni-qcom: Set an autosuspend delay of 250 ms

In commit 0e3b8a81f5df ("spi: spi-geni-qcom: Add interconnect
support") the spi_geni_runtime_suspend() and spi_geni_runtime_resume()
became a bit slower.  Measuring on my hardware I see numbers in the
hundreds of microseconds now.

Let's use autosuspend to help avoid some of the overhead.  Now if
we're doing a bunch of transfers we won't need to be constantly
chruning.

The number 250 ms for the autosuspend delay was picked a bit
arbitrarily, so if someone has measurements showing a better value we
could easily change this.

Fixes: 0e3b8a81f5df ("spi: spi-geni-qcom: Add interconnect support")
Signed-off-by: Douglas Anderson <dianders@chromium.org>
Reviewed-by: Akash Asthana<akashast@codeaurora.org>
Link: https://lore.kernel.org/r/20200701174506.2.I9b8f6bb1e7e6d8847e2ed2cf854ec55678db427f@changeid
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agospi: spi-sun6i: sun6i_spi_transfer_one(): enable RF_RDY interrupt only if needed
Marc Kleine-Budde [Mon, 6 Jul 2020 14:34:43 +0000 (16:34 +0200)]
spi: spi-sun6i: sun6i_spi_transfer_one(): enable RF_RDY interrupt only if needed

In sun6i_spi_transfer_one() the RX FIFO Ready (SUN6I_INT_CTL_RF_RDY) is
unconditionally enabled.

A RX interrupt is only needed, if more data than fits into the FIFO is going to
be received during this transfer. As the RX-FIFO is drained during transfer
complete interrupt, enable the RX FIFO Ready interrupt only if the data doesn't
fit into the FIFO.

Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
Acked-by: Maxime Ripard <mripard@kernel.org>
Link: https://lore.kernel.org/r/20200706143443.9855-11-mkl@pengutronix.de
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agospi: spi-sun6i: sun6i_spi_transfer_one(): collate write to Interrupt Control Register
Marc Kleine-Budde [Mon, 6 Jul 2020 14:34:42 +0000 (16:34 +0200)]
spi: spi-sun6i: sun6i_spi_transfer_one(): collate write to Interrupt Control Register

In sun6i_spi_transfer_one() the Interrupt Control Register is written three
times. This patch collates the three writes into one.

Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
Acked-by: Maxime Ripard <mripard@kernel.org>
Link: https://lore.kernel.org/r/20200706143443.9855-10-mkl@pengutronix.de
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agospi: spi-sun6i: sun6i_spi_fill_fifo(): remove not needed length argument
Marc Kleine-Budde [Mon, 6 Jul 2020 14:34:41 +0000 (16:34 +0200)]
spi: spi-sun6i: sun6i_spi_fill_fifo(): remove not needed length argument

The function sun6i_spi_fill_fifo() is called with a length argument of
"sspi->fifo_depth" and "SUN6I_FIFO_DEPTH".

The driver reads the number of free bytes in the FIFO from the hardware and
uses the length argument to limit this value. This is not needed as the number
of free bytes in the FIFO is always less or equal the depth of the FIFO.

This patch removes the length argument and check.

Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
Acked-by: Maxime Ripard <mripard@kernel.org>
Link: https://lore.kernel.org/r/20200706143443.9855-9-mkl@pengutronix.de
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agospi: spi-sun6i: sun6i_spi_drain_fifo(): remove not needed length argument
Marc Kleine-Budde [Mon, 6 Jul 2020 14:34:40 +0000 (16:34 +0200)]
spi: spi-sun6i: sun6i_spi_drain_fifo(): remove not needed length argument

The function sun6i_spi_drain_fifo() is called with a length argument of
"sspi->fifo_depth" and "SUN6I_FIFO_DEPTH".

The driver reads the number of available bytes to read from the FIFO from the
hardware and uses the length argument to limit this value. This is not needed
as the FIFO can contain only the fifo depth number of bytes.

This patch removes the length argument and check.

Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
Acked-by: Maxime Ripard <mripard@kernel.org>
Link: https://lore.kernel.org/r/20200706143443.9855-8-mkl@pengutronix.de
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agospi: spi-sun6i: sun6i_spi_drain_fifo(): introduce sun6i_spi_get_rx_fifo_count() and...
Marc Kleine-Budde [Mon, 6 Jul 2020 14:34:39 +0000 (16:34 +0200)]
spi: spi-sun6i: sun6i_spi_drain_fifo(): introduce sun6i_spi_get_rx_fifo_count() and make use of it

This patch introduces the function sun6i_spi_get_rx_fifo_count(), similar to
the existing sun6i_spi_get_tx_fifo_count(), to make the sun6i_spi_drain_fifo()
function a bit easier to read.

Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
Acked-by: Maxime Ripard <mripard@kernel.org>
Link: https://lore.kernel.org/r/20200706143443.9855-7-mkl@pengutronix.de
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agospi: spi-sun6i: sun6i_spi_get_tx_fifo_count: Convert manual shift+mask to FIELD_GET()
Marc Kleine-Budde [Mon, 6 Jul 2020 14:34:38 +0000 (16:34 +0200)]
spi: spi-sun6i: sun6i_spi_get_tx_fifo_count: Convert manual shift+mask to FIELD_GET()

This patch converts the manual shift+mask in sun6i_spi_get_tx_fifo_count() to
make use of FIELD_GET()

Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
Link: https://lore.kernel.org/r/20200706143443.9855-6-mkl@pengutronix.de
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agospi: spi-sun6i: sun6i_spi_transfer_one(): remove not needed masking of transfer length
Marc Kleine-Budde [Mon, 6 Jul 2020 14:34:37 +0000 (16:34 +0200)]
spi: spi-sun6i: sun6i_spi_transfer_one(): remove not needed masking of transfer length

In sun6i_spi_transfer_one() the driver ensures that the length of the transfer
is smaller or equal to SUN6I_MAX_XFER_SIZE. This means the masking of the
length to SUN6I_MAX_XFER_SIZE can be skipped when writing the transfer length
into the registers.

This patch removes the useless masking of the transfer length to
SUN6I_MAX_XFER_SIZE.

Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
Acked-by: Maxime Ripard <mripard@kernel.org>
Link: https://lore.kernel.org/r/20200706143443.9855-5-mkl@pengutronix.de
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agospi: spi-sun6i: sun6i_spi_transfer_one(): remove useless goto
Marc Kleine-Budde [Mon, 6 Jul 2020 14:34:36 +0000 (16:34 +0200)]
spi: spi-sun6i: sun6i_spi_transfer_one(): remove useless goto

This patch removes an useless goto at the end of
sun6i_spi_transfer_one().

Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
Acked-by: Maxime Ripard <mripard@kernel.org>
Link: https://lore.kernel.org/r/20200706143443.9855-4-mkl@pengutronix.de
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agospi: spi-sun6i: sun6i_spi_transfer_one(): report effectivly used speed_hz of transfer
Marc Kleine-Budde [Mon, 6 Jul 2020 14:34:35 +0000 (16:34 +0200)]
spi: spi-sun6i: sun6i_spi_transfer_one(): report effectivly used speed_hz of transfer

This patch implementes the reporting of the effectivly used speed_hz for the
transfer by setting tfr->effective_speed_hz.

See the following patch, which adds this feature to the SPI core for more
information:

    5d7e2b5ed585 spi: core: allow reporting the effectivly used speed_hz for a transfer

Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
Acked-by: Maxime Ripard <mripard@kernel.org>
Link: https://lore.kernel.org/r/20200706143443.9855-3-mkl@pengutronix.de
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agospi: spi-sun6i: sun6i_spi_transfer_one(): fix setting of clock rate
Marc Kleine-Budde [Mon, 6 Jul 2020 14:34:34 +0000 (16:34 +0200)]
spi: spi-sun6i: sun6i_spi_transfer_one(): fix setting of clock rate

A SPI transfer defines the _maximum_ speed of the SPI transfer. However the
driver doesn't take into account that the clock divider is always rounded down
(due to integer arithmetics). This results in a too high clock rate for the SPI
transfer.

E.g.: with a mclk_rate of 24 MHz and a SPI transfer speed of 10 MHz, the
original code calculates a reg of "0", which results in a effective divider of
"2" and a 12 MHz clock for the SPI transfer.

This patch fixes the issue by using DIV_ROUND_UP() instead of a plain
integer division.

While there simplify the divider calculation for the CDR1 case, use
order_base_2() instead of two ilog2() calculations.

Fixes: 3558fe900e8a ("spi: sunxi: Add Allwinner A31 SPI controller driver")
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
Acked-by: Maxime Ripard <mripard@kernel.org>
Link: https://lore.kernel.org/r/20200706143443.9855-2-mkl@pengutronix.de
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agoMerge series "Add Renesas RPC-IF support" from Sergei Shtylyov <sergei.shtylyov@cogen...
Mark Brown [Wed, 1 Jul 2020 22:21:32 +0000 (23:21 +0100)]
Merge series "Add Renesas RPC-IF support" from Sergei Shtylyov <sergei.shtylyov@cogentembedded.com>:

Hello!

Here's a set of 2 patches against Linus' repo. Renesas Reduced Pin Count
Interface (RPC-IF) allows a SPI flash or HyperFlash connected to the SoC
to be accessed via the external address space read mode or the manual mode.
The memory controller driver for RPC-IF registers either SPI or HyperFLash
subdevice, depending on the contents of the device tree subnode; it also
provides the abstract "back end" API that can be used by the "front end"
SPI/MTD drivers to talk to the real hardware...

Based on the original patch by Mason Yang <masonccyang@mxic.com.tw>.

[1/2] dt-bindings: memory: document Renesas RPC-IF bindings
[2/2] memory: add Renesas RPC-IF driver

MBR, Sergei

4 years agoMerge series "spi: bcm2835: Interrupt-handling optimisations" from Robin Murphy ...
Mark Brown [Wed, 1 Jul 2020 22:21:31 +0000 (23:21 +0100)]
Merge series "spi: bcm2835: Interrupt-handling optimisations" from Robin Murphy <robin.murphy@arm.com>:

Hi all,

Although Florian was concerned about a trivial inline check to deal with
shared IRQs adding overhead, the reality is that it would be so small as
to not be worth even thinking about unless the driver was already tuned
to squeeze out every last cycle. And a brief look over the code shows
that that clearly isn't the case.

This is an example of some of the easy low-hanging fruit that jumps out
just from code inspection. Based on disassembly and ARM1176 cycle
timings, patch #2 should save the equivalent of 2-3 shared interrupt
checks off the critical path in all cases, and patch #3 possibly up to
about 100x more. I don't have any means to test these patches, let alone
measure performance, so they're only backed by the principle that less
code - and in particular fewer memory accesses - is almost always
better.

There is almost certainly a *lot* more to be had from careful use of
relaxed I/O accessors, not doing a read-modify-write of CS at every
reset, tweaking the loops further to avoid unnecessary writebacks to
variables, and so on. However since I'm not invested in this personally
I'm not going to pursue it any further; I'm throwing these patches out
as more of a demonstration to back up my original drive-by review
comments, so if anyone want to pick them up and run with them then
please do so.

Robin.

Robin Murphy (3):
  spi: bcm3835: Tidy up bcm2835_spi_reset_hw()
  spi: bcm2835: Micro-optimise IRQ handler
  spi: bcm2835: Micro-optimise FIFO loops

 drivers/spi/spi-bcm2835.c | 45 +++++++++++++++++++--------------------
 1 file changed, 22 insertions(+), 23 deletions(-)

--
2.23.0.dirty

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

4 years agospi: lantiq-ssc: Convert to use GPIO descriptors
Linus Walleij [Thu, 25 Jun 2020 20:21:49 +0000 (22:21 +0200)]
spi: lantiq-ssc: Convert to use GPIO descriptors

This switches the Lantiq SSC driver over to use GPIO descriptor
handling in the core.

The driver was already utilizing the core to look up and request
GPIOs from the device tree so this is a pretty small change
just switching it over to use descriptors directly instead.

Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Cc: Hauke Mehrtens <hauke@hauke-m.de>
Link: https://lore.kernel.org/r/20200625202149.209276-1-linus.walleij@linaro.org
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agospi: img-spfi: Convert to use GPIO descriptors
Linus Walleij [Thu, 25 Jun 2020 20:14:22 +0000 (22:14 +0200)]
spi: img-spfi: Convert to use GPIO descriptors

This converts the IMG SPFI SPI driver to use GPIO descriptors
as obtained from the core instead of GPIO numbers.

The driver was already relying on the core code to look up
the GPIO numbers from the device tree and allocate memory for
storing state etc. By moving to use descriptors handled by
the core we can delete the setup/cleanup functions and
the device state handler that were only dealing with this.

Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Cc: Ionela Voinescu <ionela.voinescu@imgtec.com>
Cc: Sifan Naeem <sifan.naeem@imgtec.com>
Link: https://lore.kernel.org/r/20200625201422.208640-1-linus.walleij@linaro.org
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agospi: npcm-pspi: Convert to use GPIO descriptors
Linus Walleij [Thu, 25 Jun 2020 22:57:59 +0000 (00:57 +0200)]
spi: npcm-pspi: Convert to use GPIO descriptors

The Nuvoton PSPI driver already uses the core to handle GPIO
chip selects but is using the old GPIO number method and
retrieveing the GPIOs in the probe() call.

Switch it over to using GPIO descriptors saving a bunch of
code and modernizing it.

Compile tested med ARMv7 multiplatform config augmented
with the Nuvoton arch and this driver.

Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Cc: Tomer Maimon <tmaimon77@gmail.com>
Link: https://lore.kernel.org/r/20200625225759.273911-1-linus.walleij@linaro.org
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agospi: Avoid setting the chip select if we don't need to
Douglas Anderson [Mon, 29 Jun 2020 23:41:06 +0000 (16:41 -0700)]
spi: Avoid setting the chip select if we don't need to

On some SPI controllers (like spi-geni-qcom) setting the chip select
is a heavy operation.  For instance on spi-geni-qcom, with the current
code, is was measured as taking upwards of 20 us.  Even on SPI
controllers that aren't as heavy, setting the chip select is at least
something like a MMIO operation over some peripheral bus which isn't
as fast as a RAM access.

While it would be good to find ways to mitigate problems like this in
the drivers for those SPI controllers, it can also be noted that the
SPI framework could also help out.  Specifically, in some situations,
we can see the SPI framework calling the driver's set_cs() with the
same parameter several times in a row.  This is specifically observed
when looking at the way the Chrome OS EC SPI driver (cros_ec_spi)
works but other drivers likely trip it to some extent.

Let's solve this by caching the chip select state in the core and only
calling into the controller if there was a change.  We check not only
the "enable" state but also the chip select mode (active high or
active low) since controllers may care about both the mode and the
enable flag in their callback.

Signed-off-by: Douglas Anderson <dianders@chromium.org>
Link: https://lore.kernel.org/r/20200629164103.1.Ied8e8ad8bbb2df7f947e3bc5ea1c315e041785a2@changeid
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agospi: fsl: add missing __iomem annotation
Luc Van Oostenryck [Mon, 22 Jun 2020 16:26:11 +0000 (18:26 +0200)]
spi: fsl: add missing __iomem annotation

The field mspi->reg_base is annotated as an __iomem pointer. Good.

However, this field is often assigned to a temporary variable:
before being used. For example:
struct fsl_spi_reg *reg_base = mspi->reg_base;

But this variable is missing the __iomem annotation.
So, add the missing __iomem and make sparse & the bot happier.

Reported-by: kernel test robot <lkp@intel.com>
Signed-off-by: Luc Van Oostenryck <luc.vanoostenryck@gmail.com>
Link: https://lore.kernel.org/r/20200622162611.83694-1-luc.vanoostenryck@gmail.com
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agomemory: add Renesas RPC-IF driver
Sergei Shtylyov [Tue, 16 Jun 2020 20:03:48 +0000 (23:03 +0300)]
memory: add Renesas RPC-IF driver

Add the memory driver for Renesas RPC-IF which registers either SPI or
HyperFLash device depending on the contents of the device tree subnode.
It also provides the absract "back end" device APIs that can be used by
the "front end" SPI/MTD drivers to talk to the real hardware.

Based on the original patch by Mason Yang <masonccyang@mxic.com.tw>.

Signed-off-by: Sergei Shtylyov <sergei.shtylyov@cogentembedded.com>
Link: https://lore.kernel.org/r/9a3606ec-d4d0-c63a-4fb6-631ab38e621c@cogentembedded.com
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agodt-bindings: memory: document Renesas RPC-IF bindings
Sergei Shtylyov [Tue, 16 Jun 2020 20:01:42 +0000 (23:01 +0300)]
dt-bindings: memory: document Renesas RPC-IF bindings

Renesas Reduced Pin Count Interface (RPC-IF) allows a SPI flash or
HyperFlash connected to the SoC to be accessed via the external address
space read mode or the manual mode.

Document the device tree bindings for the Renesas RPC-IF found in the R-Car
gen3 SoCs.

Based on the original patch by Mason Yang <masonccyang@mxic.com.tw>.

Signed-off-by: Sergei Shtylyov <sergei.shtylyov@cogentembedded.com>
Link: https://lore.kernel.org/r/54a84c75-fa17-9976-d9a6-a69ef67c418b@cogentembedded.com
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agospi: bcm2835: Micro-optimise FIFO loops
Robin Murphy [Tue, 16 Jun 2020 00:09:29 +0000 (01:09 +0100)]
spi: bcm2835: Micro-optimise FIFO loops

The blind and counted loops are always called with nonzero count, so
convert them to do-while loops that lead to slightly more efficient
code generation. With GCC 8.3 this shaves off 1-2 instructions per
iteration in each case.

Signed-off-by: Robin Murphy <robin.murphy@arm.com>
Link: https://lore.kernel.org/r/9242863077acf9a64e4b3720e479855b88d19e82.1592261248.git.robin.murphy@arm.com
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agospi: bcm2835: Micro-optimise IRQ handler
Robin Murphy [Tue, 16 Jun 2020 00:09:28 +0000 (01:09 +0100)]
spi: bcm2835: Micro-optimise IRQ handler

The IRQ handler only needs the struct spi_controller for the sake of
the completion at the end of a transfer. Passing the struct bcm2835_spi
directly as the IRQ data allows that level of indirection to be pushed
into the completion path for the reverse lookup, and avoided entirely
in all other cases.

This saves one explicit load in the critical path, plus (for a GCC 8.3
build) two registers worth of stack frame overhead.

Signed-off-by: Robin Murphy <robin.murphy@arm.com>
Link: https://lore.kernel.org/r/6b401cb521539caffab21f05b4c8cba6c9d27c6e.1592261248.git.robin.murphy@arm.com
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agospi: bcm3835: Tidy up bcm2835_spi_reset_hw()
Robin Murphy [Tue, 16 Jun 2020 00:09:27 +0000 (01:09 +0100)]
spi: bcm3835: Tidy up bcm2835_spi_reset_hw()

It doesn't need a struct spi_controller, and every callsite has
already retrieved the appropriate struct bcm2835_spi, so just pass
that directly.

Signed-off-by: Robin Murphy <robin.murphy@arm.com>
Link: https://lore.kernel.org/r/eca458ae1a0d3934d0627f90e25d294fefd4b13d.1592261248.git.robin.murphy@arm.com
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agospi: mediatek: use correct SPI_CFG2_REG MACRO
leilk.liu [Wed, 1 Jul 2020 09:00:20 +0000 (17:00 +0800)]
spi: mediatek: use correct SPI_CFG2_REG MACRO

this patch use correct SPI_CFG2_REG offset.

Signed-off-by: leilk.liu <leilk.liu@mediatek.com>
Link: https://lore.kernel.org/r/20200701090020.7935-1-leilk.liu@mediatek.com
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agospi: omap2-mcspi: Convert to use GPIO descriptors
Linus Walleij [Thu, 25 Jun 2020 23:12:57 +0000 (01:12 +0200)]
spi: omap2-mcspi: Convert to use GPIO descriptors

The OMAP2 MCSPI has some kind of half-baked GPIO CS support:
it includes code like this:

  if (gpio_is_valid(spi->cs_gpio)) {
        ret = gpio_request(spi->cs_gpio, dev_name(&spi->dev));
(...)

But it doesn't parse the "cs-gpios" attribute in the device
tree to count the number of GPIOs or pick out the GPIO numbers
and put these in the SPI master's .cs_gpios property.

We complete the implementation of supporting CS GPIOs
from the device tree and switch it over to use the SPI core
for this.

Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Acked-by: Tony Lindgren <tony@atomide.com>
Link: https://lore.kernel.org/r/20200625231257.280615-1-linus.walleij@linaro.org
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agospi: spi-geni-qcom: Don't set the cs if it was already right
Douglas Anderson [Fri, 26 Jun 2020 22:19:50 +0000 (15:19 -0700)]
spi: spi-geni-qcom: Don't set the cs if it was already right

Setting the chip select on the Qualcomm geni SPI controller isn't
exactly cheap.  Let's cache the current setting and avoid setting the
chip select if it's already right.

Using "flashrom" to read or write the EC firmware on a Chromebook
shows roughly a 25% reduction in interrupts and a 15% speedup.

Signed-off-by: Douglas Anderson <dianders@chromium.org>
Link: https://lore.kernel.org/r/20200626151946.1.I06134fd669bf91fd387dc6ecfe21d44c202bd412@changeid
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agospi: pxa2xx: Add support for Intel Tiger Lake PCH-H
Jarkko Nikula [Thu, 25 Jun 2020 14:00:41 +0000 (17:00 +0300)]
spi: pxa2xx: Add support for Intel Tiger Lake PCH-H

Add Intel Tiger Lake PCH-H PCI IDs.

Signed-off-by: Jarkko Nikula <jarkko.nikula@linux.intel.com>
Link: https://lore.kernel.org/r/20200625140041.745804-1-jarkko.nikula@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agospi: altera: fix module autoload
Xu Yilun [Wed, 24 Jun 2020 01:31:26 +0000 (09:31 +0800)]
spi: altera: fix module autoload

Add the MODULE_DEVICE_TABLE macro for the platform_device_id table to
allow proper creation of modalias strings and fix autoloading module for
this driver.

Signed-off-by: Xu Yilun <yilun.xu@intel.com>
Signed-off-by: Russ Weight <russell.h.weight@intel.com>
Link: https://lore.kernel.org/r/1592962286-25752-3-git-send-email-yilun.xu@intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agospi: altera: fix driver matching failure of the device ID "spi_altera"
Xu Yilun [Wed, 24 Jun 2020 01:31:25 +0000 (09:31 +0800)]
spi: altera: fix driver matching failure of the device ID "spi_altera"

The driver is expected to support device ID "spi_altera" for MMIO accessed
devices, device ID "subdev_spi_altera" for indirect accessed devices. But
the platform bus will not try driver name match anymore if the platform
driver has an id_table. So the "spi_altera" should also be added to
id_table.

Signed-off-by: Xu Yilun <yilun.xu@intel.com>
Signed-off-by: Russ Weight <russell.h.weight@intel.com>
Link: https://lore.kernel.org/r/1592962286-25752-2-git-send-email-yilun.xu@intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agospi: imx: add fallback feature
Robin Gong [Tue, 16 Jun 2020 22:42:09 +0000 (06:42 +0800)]
spi: imx: add fallback feature

Add fallback pio feature in case  dma transfer failed before start.
Besides, another whole pio transfer including setup_transfer will be
issued by spi core, no need to restore jobs like commit bcd8e7761ec9 ("spi:
imx: fallback to PIO if dma setup failure").

Signed-off-by: Robin Gong <yibin.gong@nxp.com>
Link: https://lore.kernel.org/r/1592347329-28363-3-git-send-email-yibin.gong@nxp.com
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agospi: introduce fallback to pio
Robin Gong [Tue, 16 Jun 2020 22:42:08 +0000 (06:42 +0800)]
spi: introduce fallback to pio

Add fallback to pio mode in case dma transfer failed with error status
SPI_TRANS_FAIL_NO_START.
If spi client driver want to enable this feature please set xfer->error in
the proper place such as dmaengine_prep_slave_sg() failure detect(but no
any data put into spi bus yet). Besides, add master->fallback checking in
its can_dma() so that spi core could switch to pio next time. Please refer
to spi-imx.c.

Signed-off-by: Robin Gong <yibin.gong@nxp.com>
Link: https://lore.kernel.org/r/1592347329-28363-2-git-send-email-yibin.gong@nxp.com
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agoMerge series "Some small spi geni cleanups" from Stephen Boyd <swboyd@chromium.org>:
Mark Brown [Mon, 22 Jun 2020 14:58:32 +0000 (15:58 +0100)]
Merge series "Some small spi geni cleanups" from Stephen Boyd <swboyd@chromium.org>:

To follow onto Doug's latest spi geni series[1] this simplifies and
reduces the code a little more.

[1] https://lore.kernel.org/r/20200618150626.237027-1-dianders@chromium.org

Stephen Boyd (2):
  spi: spi-geni-qcom: Simplify setup_fifo_xfer()
  spi: spi-geni-qcom: Don't set {tx,rx}_rem_bytes unnecessarily

 drivers/spi/spi-geni-qcom.c | 55 +++++++++++++++++--------------------
 1 file changed, 25 insertions(+), 30 deletions(-)

base-commit: 7ba9bdcb91f694b0eaf486a825afd9c2d99532b7
--
Sent by a computer, using git, on the internet

4 years agospi: spi-geni-qcom: Don't set {tx,rx}_rem_bytes unnecessarily
Stephen Boyd [Sat, 20 Jun 2020 02:22:33 +0000 (19:22 -0700)]
spi: spi-geni-qcom: Don't set {tx,rx}_rem_bytes unnecessarily

We only need to test for these counters being non-zero when we see the
end of a transfer. If we're doing a CS change then they will already be
zero.  This implies that we don't need to set these to 0 if we're
cancelling an in flight transfer too, because we only care to test these
counters when the 'DONE' bit is set in the hardware and we've set them
to non-zero for a transfer.

This is a non-functional change, just cleanup to consolidate code.

Signed-off-by: Stephen Boyd <swboyd@chromium.org>
Reviewed-by: Douglas Anderson <dianders@chromium.org>
Link: https://lore.kernel.org/r/20200620022233.64716-3-swboyd@chromium.org
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agospi: spi-geni-qcom: Simplify setup_fifo_xfer()
Stephen Boyd [Sat, 20 Jun 2020 02:22:32 +0000 (19:22 -0700)]
spi: spi-geni-qcom: Simplify setup_fifo_xfer()

The definition of SPI_FULL_DUPLEX (3) is really SPI_TX_ONLY (1) ORed
with SPI_RX_ONLY (2). Let's drop the define and simplify the code here a
bit by collapsing the setting of 'm_cmd' into conditions that are the
same.

This is a non-functional change, just cleanup to consolidate code.

Signed-off-by: Stephen Boyd <swboyd@chromium.org>
Reviewed-by: Douglas Anderson <dianders@chromium.org>
Link: https://lore.kernel.org/r/20200620022233.64716-2-swboyd@chromium.org
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agospi: spi-fsl-dspi: Initialize completion before possible interrupt
Krzysztof Kozlowski [Mon, 22 Jun 2020 11:05:43 +0000 (13:05 +0200)]
spi: spi-fsl-dspi: Initialize completion before possible interrupt

The interrupt handler calls completion and is IRQ requested before the
completion is initialized.  Logically it should be the other way.

Signed-off-by: Krzysztof Kozlowski <krzk@kernel.org>
Link: https://lore.kernel.org/r/20200622110543.5035-4-krzk@kernel.org
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agospi: spi-fsl-dspi: Fix external abort on interrupt in resume or exit paths
Krzysztof Kozlowski [Mon, 22 Jun 2020 11:05:42 +0000 (13:05 +0200)]
spi: spi-fsl-dspi: Fix external abort on interrupt in resume or exit paths

If shared interrupt comes late, during probe error path or device remove
(could be triggered with CONFIG_DEBUG_SHIRQ), the interrupt handler
dspi_interrupt() will access registers with the clock being disabled.
This leads to external abort on non-linefetch on Toradex Colibri VF50
module (with Vybrid VF5xx):

    $ echo 4002d000.spi > /sys/devices/platform/soc/40000000.bus/4002d000.spi/driver/unbind

    Unhandled fault: external abort on non-linefetch (0x1008) at 0x8887f02c
    Internal error: : 1008 [#1] ARM
    Hardware name: Freescale Vybrid VF5xx/VF6xx (Device Tree)
    Backtrace:
      (regmap_mmio_read32le)
      (regmap_mmio_read)
      (_regmap_bus_reg_read)
      (_regmap_read)
      (regmap_read)
      (dspi_interrupt)
      (free_irq)
      (devm_irq_release)
      (release_nodes)
      (devres_release_all)
      (device_release_driver_internal)

The resource-managed framework should not be used for shared interrupt
handling, because the interrupt handler might be called after releasing
other resources and disabling clocks.

Similar bug could happen during suspend - the shared interrupt handler
could be invoked after suspending the device.  Each device sharing this
interrupt line should disable the IRQ during suspend so handler will be
invoked only in following cases:
1. None suspended,
2. All devices resumed.

Fixes: 349ad66c0ab0 ("spi:Add Freescale DSPI driver for Vybrid VF610 platform")
Signed-off-by: Krzysztof Kozlowski <krzk@kernel.org>
Tested-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Reviewed-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Cc: <stable@vger.kernel.org>
Link: https://lore.kernel.org/r/20200622110543.5035-3-krzk@kernel.org
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agospi: spi-fsl-dspi: Fix lockup if device is shutdown during SPI transfer
Krzysztof Kozlowski [Mon, 22 Jun 2020 11:05:41 +0000 (13:05 +0200)]
spi: spi-fsl-dspi: Fix lockup if device is shutdown during SPI transfer

During shutdown, the driver should unregister the SPI controller
and stop the hardware.  Otherwise the dspi_transfer_one_message() could
wait on completion infinitely.

Additionally, calling spi_unregister_controller() first in device
shutdown reverse-matches the probe function, where SPI controller is
registered at the end.

Fixes: dc234825997e ("spi: spi-fsl-dspi: Adding shutdown hook")
Reported-by: Vladimir Oltean <olteanv@gmail.com>
Signed-off-by: Krzysztof Kozlowski <krzk@kernel.org>
Tested-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Reviewed-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Cc: <stable@vger.kernel.org>
Link: https://lore.kernel.org/r/20200622110543.5035-2-krzk@kernel.org
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agospi: spi-fsl-dspi: Fix lockup if device is removed during SPI transfer
Krzysztof Kozlowski [Mon, 22 Jun 2020 11:05:40 +0000 (13:05 +0200)]
spi: spi-fsl-dspi: Fix lockup if device is removed during SPI transfer

During device removal, the driver should unregister the SPI controller
and stop the hardware.  Otherwise the dspi_transfer_one_message() could
wait on completion infinitely.

Additionally, calling spi_unregister_controller() first in device
removal reverse-matches the probe function, where SPI controller is
registered at the end.

Fixes: 05209f457069 ("spi: fsl-dspi: add missing clk_disable_unprepare() in dspi_remove()")
Reported-by: Vladimir Oltean <olteanv@gmail.com>
Signed-off-by: Krzysztof Kozlowski <krzk@kernel.org>
Cc: <stable@vger.kernel.org>
Link: https://lore.kernel.org/r/20200622110543.5035-1-krzk@kernel.org
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agospi: spi-geni-qcom: Simplify setup_fifo_xfer()
Stephen Boyd [Thu, 18 Jun 2020 23:39:58 +0000 (16:39 -0700)]
spi: spi-geni-qcom: Simplify setup_fifo_xfer()

The definition of SPI_FULL_DUPLEX (3) is really SPI_TX_ONLY (1) ORed
with SPI_RX_ONLY (2). Let's drop the define and simplify the code here a
bit by collapsing the setting of 'm_cmd' into conditions that are the
same.

This is a non-functional change, just cleanup to consolidate code.

Signed-off-by: Stephen Boyd <swboyd@chromium.org>
Reviewed-by: Douglas Anderson <dianders@chromium.org>
Cc: Douglas Anderson <dianders@chromium.org>
Link: https://lore.kernel.org/r/20200618233959.160032-1-swboyd@chromium.org
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agoMerge series "mtd: spi-nor: Move cadence-qaudspi to spi-mem framework" from Vignesh...
Mark Brown [Fri, 19 Jun 2020 13:26:55 +0000 (14:26 +0100)]
Merge series "mtd: spi-nor: Move cadence-qaudspi to spi-mem framework" from Vignesh Raghavendra <vigneshr@ti.com>:

This series is a subset of "[PATCH v12 0/4] spi: cadence-quadspi: Add
support for the Cadence QSPI controller" by Ramuthevar,Vadivel MuruganX
<vadivel.muruganx.ramuthevar@linux.intel.com> that intended to move
cadence-quadspi driver to spi-mem framework

Those patches were trying to accomplish too many things in a single set
of patches and need to split into smaller patches. This is reduced
version of above series.

Changes that are intended to make migration easy are split into separate
patches. Patches 1 to 3 drop features that cannot be supported under
spi-mem at the moment (backward compatibility is maintained).
Patch 4-5 are trivial cleanups. Patch 6 does the actual conversion to
spi-mem and patch 7 moves the driver to drivers/spi folder.

I have tested both INDAC mode (used by non TI platforms like Altera
SoCFPGA) and DAC mode (used by TI platforms) on TI EVMs.

Patches to move move bindings over to
"Documentation/devicetree/bindings/spi/" directory and also conversion
of bindig doc to YAML will be posted separately.  Support for Intel
platform would follow that.

Resend v3:
Rebased onto v5.7-c1

v3:
Split handling of probe deferral into separate patch (out of 5/6)
Split dropping of redundant WREN to separate patch (out of 5/6)
Fix a possible memleak due to lack of spi_master_put()
Parse all SPI slave nodes in cqspi_setup_flash()
Address misc comments from Tudor on v2
Rebase onto latest spi-nor/next

v2:
Rework patch 1/6 to keep "cdns,is-decoded-cs" property supported.

Ramuthevar Vadivel Murugan (2):
  mtd: spi-nor: Convert cadence-quadspi to use spi-mem framework
  spi: Move cadence-quadspi driver to drivers/spi/

Vignesh Raghavendra (6):
  mtd: spi-nor: cadence-quadspi: Make driver independent of flash
    geometry
  mtd: spi-nor: cadence-quadspi: Provide a way to disable DAC mode
  mtd: spi-nor: cadence-quadspi: Don't initialize rx_dma_complete on
    failure
  mtd: spi-nor: cadence-quadspi: Fix error path on failure to acquire
    reset lines
  mtd: spi-nor: cadence-quadspi: Handle probe deferral while requesting
    DMA channel
  mtd: spi-nor: cadence-quadspi: Drop redundant WREN in erase path

 drivers/mtd/spi-nor/controllers/Kconfig       |  11 -
 drivers/mtd/spi-nor/controllers/Makefile      |   1 -
 drivers/spi/Kconfig                           |  11 +
 drivers/spi/Makefile                          |   1 +
 .../spi-cadence-quadspi.c}                    | 541 +++++++-----------
 5 files changed, 222 insertions(+), 343 deletions(-)
 rename drivers/{mtd/spi-nor/controllers/cadence-quadspi.c => spi/spi-cadence-quadspi.c} (74%)

base-commit: b3a9e3b9622ae10064826dccb4f7a52bd88c7407

--
2.26.2

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

4 years agospi: Move cadence-quadspi driver to drivers/spi/
Ramuthevar Vadivel Murugan [Mon, 1 Jun 2020 07:04:44 +0000 (12:34 +0530)]
spi: Move cadence-quadspi driver to drivers/spi/

Now that cadence-quadspi has been converted to use spi-mem framework,
move it under drivers/spi/

Update license header to match SPI subsystem style

Signed-off-by: Ramuthevar Vadivel Murugan <vadivel.muruganx.ramuthevar@linux.intel.com>
Signed-off-by: Vignesh Raghavendra <vigneshr@ti.com>
Reviewed-by: Tudor Ambarus <tudor.ambarus@microchip.com>
Acked-by: Tudor Ambarus <tudor.ambarus@microchip.com>
Link: https://lore.kernel.org/r/20200601070444.16923-9-vigneshr@ti.com
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agomtd: spi-nor: Convert cadence-quadspi to use spi-mem framework
Ramuthevar Vadivel Murugan [Mon, 1 Jun 2020 07:04:43 +0000 (12:34 +0530)]
mtd: spi-nor: Convert cadence-quadspi to use spi-mem framework

Move cadence-quadspi driver to use spi-mem framework. This is required
to make the driver support for SPI NAND flashes in future.

Driver is feature compliant with existing SPI NOR version.

Signed-off-by: Ramuthevar Vadivel Murugan <vadivel.muruganx.ramuthevar@linux.intel.com>
Signed-off-by: Vignesh Raghavendra <vigneshr@ti.com>
Reviewed-by: Tudor Ambarus <tudor.ambarus@microchip.com>
Acked-by: Tudor Ambarus <tudor.ambarus@microchip.com>
Link: https://lore.kernel.org/r/20200601070444.16923-8-vigneshr@ti.com
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agomtd: spi-nor: cadence-quadspi: Drop redundant WREN in erase path
Vignesh Raghavendra [Mon, 1 Jun 2020 07:04:42 +0000 (12:34 +0530)]
mtd: spi-nor: cadence-quadspi: Drop redundant WREN in erase path

Drop redundant WREN command in cqspi_erase() as SPI NOR core takes care
of sending WREN command before sending erase command.

Signed-off-by: Vignesh Raghavendra <vigneshr@ti.com>
Reviewed-by: Tudor Ambarus <tudor.ambarus@microchip.com>
Acked-by: Tudor Ambarus <tudor.ambarus@microchip.com>
Link: https://lore.kernel.org/r/20200601070444.16923-7-vigneshr@ti.com
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agomtd: spi-nor: cadence-quadspi: Handle probe deferral while requesting DMA channel
Vignesh Raghavendra [Mon, 1 Jun 2020 07:04:41 +0000 (12:34 +0530)]
mtd: spi-nor: cadence-quadspi: Handle probe deferral while requesting DMA channel

dma_request_chan_by_mask() can throw EPROBE_DEFER if DMA provider
is not yet probed. Currently driver just falls back to using PIO mode
(which is less efficient) in this case. Instead return probe deferral
error as is so that driver will be re probed once DMA provider is
available.

Signed-off-by: Vignesh Raghavendra <vigneshr@ti.com>
Reviewed-by: Tudor Ambarus <tudor.ambarus@microchip.com>
Acked-by: Tudor Ambarus <tudor.ambarus@microchip.com>
Link: https://lore.kernel.org/r/20200601070444.16923-6-vigneshr@ti.com
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agomtd: spi-nor: cadence-quadspi: Fix error path on failure to acquire reset lines
Vignesh Raghavendra [Mon, 1 Jun 2020 07:04:40 +0000 (12:34 +0530)]
mtd: spi-nor: cadence-quadspi: Fix error path on failure to acquire reset lines

Make sure to undo the prior changes done by the driver when exiting due
to failure to acquire reset lines.

Signed-off-by: Vignesh Raghavendra <vigneshr@ti.com>
Reviewed-by: Tudor Ambarus <tudor.ambarus@microchip.com>
Acked-by: Tudor Ambarus <tudor.ambarus@microchip.com>
Link: https://lore.kernel.org/r/20200601070444.16923-5-vigneshr@ti.com
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agomtd: spi-nor: cadence-quadspi: Don't initialize rx_dma_complete on failure
Vignesh Raghavendra [Mon, 1 Jun 2020 07:04:39 +0000 (12:34 +0530)]
mtd: spi-nor: cadence-quadspi: Don't initialize rx_dma_complete on failure

If driver fails to acquire DMA channel then don't initialize
rx_dma_complete struct as it won't be used.

Signed-off-by: Vignesh Raghavendra <vigneshr@ti.com>
Reviewed-by: Tudor Ambarus <tudor.ambarus@microchip.com>
Acked-by: Tudor Ambarus <tudor.ambarus@microchip.com>
Link: https://lore.kernel.org/r/20200601070444.16923-4-vigneshr@ti.com
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agomtd: spi-nor: cadence-quadspi: Provide a way to disable DAC mode
Vignesh Raghavendra [Mon, 1 Jun 2020 07:04:38 +0000 (12:34 +0530)]
mtd: spi-nor: cadence-quadspi: Provide a way to disable DAC mode

Currently direct access mode is used on platforms that have AHB window
(memory mapped window) larger than flash size. This feature is limited
to TI platforms as non TI platforms have < 1MB of AHB window.
Therefore introduce a driver quirk to disable DAC mode and set it for
non TI compatibles. This is in preparation to move to spi-mem framework
where flash geometry cannot be known.

Signed-off-by: Vignesh Raghavendra <vigneshr@ti.com>
Reviewed-by: Tudor Ambarus <tudor.ambarus@microchip.com>
Acked-by: Tudor Ambarus <tudor.ambarus@microchip.com>
Link: https://lore.kernel.org/r/20200601070444.16923-3-vigneshr@ti.com
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agomtd: spi-nor: cadence-quadspi: Make driver independent of flash geometry
Vignesh Raghavendra [Mon, 1 Jun 2020 07:04:37 +0000 (12:34 +0530)]
mtd: spi-nor: cadence-quadspi: Make driver independent of flash geometry

Drop configuration of Flash size, erase size and page size
configuration. Flash size is needed only if using AHB decoder (BIT 23 of
CONFIG_REG) which is not used by the driver.
Erase size and page size are needed if IP is configured to send WREN
automatically. But since SPI NOR layer takes care of sending WREN, there
is no need to configure these fields either.

Therefore drop these in preparation to move the driver to spi-mem
framework where flash geometry is not visible to controller driver.

Signed-off-by: Vignesh Raghavendra <vigneshr@ti.com>
Reviewed-by: Tudor Ambarus <tudor.ambarus@microchip.com>
Acked-by: Tudor Ambarus <tudor.ambarus@microchip.com>
Link: https://lore.kernel.org/r/20200601070444.16923-2-vigneshr@ti.com
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agoMerge series "add regmap & indirect access support" from Xu Yilun <yilun.xu@intel...
Mark Brown [Fri, 19 Jun 2020 12:43:10 +0000 (13:43 +0100)]
Merge series "add regmap & indirect access support" from Xu Yilun <yilun.xu@intel.com>:

Updated the regmap & indirect access support for spi-altera.

Patch #1 is an 1:1 replacement of of readl/writel with regmap_read/write
Patch #2 introduced a new platform_device_id to support indirect access as
         a sub device.
Patch #3 is a minor fix.

Main changes from v1:
 - Split the regmap supporting patch to 2 patches.
 - Add a new platform_device_id to support indirect access.
 - Removed the v1 patch "move driver name string to header file". Now we
   use driver name string directly.
 - Add Yilun's Signed-off-by for Patch #3.
 - Add Tom's Reviewed-by.

Matthew Gerlach (1):
  spi: altera: fix size mismatch on 64 bit processors

Xu Yilun (2):
  spi: altera: use regmap-mmio instead of direct mmio register access
  spi: altera: support indirect access to the registers

 drivers/spi/Kconfig      |   1 +
 drivers/spi/spi-altera.c | 127 +++++++++++++++++++++++++++++++++++++++--------
 2 files changed, 107 insertions(+), 21 deletions(-)

--
2.7.4

4 years agoMerge series "spi: spi-geni-qcom: Fixes / perf improvements" from Douglas Anderson...
Mark Brown [Fri, 19 Jun 2020 12:43:09 +0000 (13:43 +0100)]
Merge series "spi: spi-geni-qcom: Fixes / perf improvements" from Douglas Anderson <dianders@chromium.org>:

This patch series is a new version of the previous patch posted:
  [PATCH v2] spi: spi-geni-qcom: Speculative fix of "nobody cared" about interrupt
  https://lore.kernel.org/r/20200317133653.v2.1.I752ebdcfd5e8bf0de06d66e767b8974932b3620e@changeid

At this point I've done enough tracing to know that there was a real
race in the old code (not just weakly ordered memory problems) and
that should be fixed with the locking patches.

While looking at this driver, I also noticed we weren't properly
noting error interrupts and also weren't actually using our FIFO
effectively, so I fixed those.

The last patch in the series addresses review feedback about dislike
for the "cur_mcmd" state variable.  It also could possibly make
"abort" work ever-so-slightly more reliably.

Changes in v4:
- Drop 'controller' in comment.
- Use Stephen's diagram to explain the race better.

Changes in v3:
- ("spi: spi-geni-qcom: No need for irqsave variant...") new for v3
- Split out some lock cleanup to previous patch.
- Don't need to read IRQ status register inside spinlock.
- Don't check for state CMD_NONE; later patch is removing state var.
- Don't hold the lock for all of setup_fifo_xfer().
- Comment about why it's safe to Ack interrupts at the end.
- Subject/desc changed since race is definitely there.
- ("spi: spi-geni-qcom: Check for error IRQs") new in v3.
- ("spi: spi-geni-qcom: Actually use our FIFO") new in v3.
- ("spi: spi-geni-qcom: Don't keep a local state variable") new in v3.

Changes in v2:
- Detect true spurious interrupt.
- Still return IRQ_NONE for state machine mismatch, but print warn.

Douglas Anderson (5):
  spi: spi-geni-qcom: No need for irqsave variant of spinlock calls
  spi: spi-geni-qcom: Mo' betta locking
  spi: spi-geni-qcom: Check for error IRQs
  spi: spi-geni-qcom: Actually use our FIFO
  spi: spi-geni-qcom: Don't keep a local state variable

 drivers/spi/spi-geni-qcom.c | 120 ++++++++++++++++++++++++------------
 1 file changed, 81 insertions(+), 39 deletions(-)

--
2.27.0.290.gba653c62da-goog

4 years agospi: npcm-fiu: Reuse BITS_PER_BYTE definition
Andy Shevchenko [Thu, 18 Jun 2020 17:01:44 +0000 (20:01 +0300)]
spi: npcm-fiu: Reuse BITS_PER_BYTE definition

No need to redefine already existing definition.
So, replace custom by generic one.

Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Link: https://lore.kernel.org/r/20200618170144.57433-1-andriy.shevchenko@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agospi: spi-geni-qcom: Don't keep a local state variable
Douglas Anderson [Thu, 18 Jun 2020 15:06:26 +0000 (08:06 -0700)]
spi: spi-geni-qcom: Don't keep a local state variable

The variable "cur_mcmd" kept track of our current state (idle, xfer,
cs, cancel).  We don't really need it, so get rid of it.  Instead:
* Use separate condition variables for "chip select done", "cancel
  done", and "abort done".  This is important so that if a "done"
  comes through (perhaps some previous interrupt finally came through)
  it can't confuse the cancel/abort function.
* Use the "done" interrupt only for when a chip select or transfer is
  done and we can tell the difference by looking at whether "cur_xfer"
  is NULL.

This is mostly a no-op change.  However, it is possible it could fix
an issue where a super delayed interrupt for a cancel command could
have confused our waiting for an abort command.

Signed-off-by: Douglas Anderson <dianders@chromium.org>
Reviewed-by: Stephen Boyd <swboyd@chromium.org>
Link: https://lore.kernel.org/r/20200618080459.v4.5.Ib1e6855405fc9c99916ab7c7dee84d73a8bf3d68@changeid
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agospi: spi-geni-qcom: Actually use our FIFO
Douglas Anderson [Thu, 18 Jun 2020 15:06:25 +0000 (08:06 -0700)]
spi: spi-geni-qcom: Actually use our FIFO

The geni hardware has a FIFO that can hold up to 64 bytes (it has 16
entries that can hold 4 bytes each), at least on the two SoCs I tested
(sdm845 and sc7180).  We configured our RX Watermark to 0, which
basically meant we got an interrupt as soon as the first 4 bytes
showed up in the FIFO.  Tracing the IRQ handler showed that we often
only read 4 or 8 bytes per IRQ handler.

I tried setting the RX Watermark to "fifo size - 2" but that just got
me a bunch of overrun errors reported.  Setting it to "fifo size - 3"
seemed to work great, though.  This made me worried that we'd start
getting overruns if we had long interrupt latency, but that doesn't
appear to be the case and delays inserted in the IRQ handler while
using "fifo size - 3" didn't cause any errors.  Presumably there is
some interaction with the poorly-documented RFR (ready for receive)
level means that "fifo size - 3" is the max.  We are the SPI master,
so it makes sense that there would be no problems with overruns, the
master should just stop clocking.

Despite "fifo size - 3" working, I chose "fifo size / 2" (8 entries =
32 bytes) which gives us a little extra time to get to the interrupt
handler and should reduce dead time on the SPI wires.  With this
setting, I often saw the IRQ handler handle 40 bytes but sometimes up
to 56 if we had bad interrupt latency.

Testing by running "flashrom -p ec -r" on a Chromebook saw interrupts
from the SPI driver cut roughly in half.  Time was roughly the same.

Fixes: 561de45f72bd ("spi: spi-geni-qcom: Add SPI driver support for GENI based QUP")
Signed-off-by: Douglas Anderson <dianders@chromium.org>
Reviewed-by: Stephen Boyd <swboyd@chromium.org>
Link: https://lore.kernel.org/r/20200618080459.v4.4.I988281f7c6ee0ed00325559bfce7539f403da69e@changeid
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agospi: spi-geni-qcom: Check for error IRQs
Douglas Anderson [Thu, 18 Jun 2020 15:06:24 +0000 (08:06 -0700)]
spi: spi-geni-qcom: Check for error IRQs

>From reading the #defines it seems like we should shout if we ever see
one of these error bits.  Let's do so.  This doesn't do anything
functional except print a yell in the log if the error bits are seen.

Signed-off-by: Douglas Anderson <dianders@chromium.org>
Reviewed-by: Stephen Boyd <swboyd@chromium.org>
Link: https://lore.kernel.org/r/20200618080459.v4.3.Id8bebdbdb4d2ed9468634343a7e6207d6cffff8a@changeid
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agospi: spi-geni-qcom: Mo' betta locking
Douglas Anderson [Thu, 18 Jun 2020 15:06:23 +0000 (08:06 -0700)]
spi: spi-geni-qcom: Mo' betta locking

If you added a bit of a delay (like a trace_printk) into the ISR for
the spi-geni-qcom driver, you would suddenly start seeing some errors
spit out.  The problem was that, though the ISR itself held a lock,
other parts of the driver didn't always grab the lock.

One example race was this:
  CPU0                                         CPU1
  ----                                         ----
  spi_geni_set_cs()
   mas->cur_mcmd = CMD_CS;
   geni_se_setup_m_cmd(...)
   wait_for_completion_timeout(&xfer_done);
                                              <INTERRUPT>
                                               geni_spi_isr()
                                                complete(&xfer_done);
   <wakeup>
   pm_runtime_put(mas->dev);
  ... // back to SPI core
  spi_geni_transfer_one()
   setup_fifo_xfer()
    mas->cur_mcmd = CMD_XFER;
                                                mas->cur_cmd = CMD_NONE; // bad!
                                                return IRQ_HANDLED;

Let's fix this.  Before we start messing with hardware, we'll grab the
lock to make sure that the IRQ handler from some previous command has
really finished.  We don't need to hold the lock unless we're in a
state where more interrupts can come in, but we at least need to make
sure the previous IRQ is done.  This lock is used exclusively to
prevent the IRQ handler and non-IRQ from stomping on each other.  The
SPI core handles all other mutual exclusion.

As part of this, we change the way that the IRQ handler detects
spurious interrupts.  Previously we checked for our state variable
being set to IRQ_NONE, but that was done outside the spinlock.  We
could move it into the spinlock, but instead let's just change it to
look for the lack of any IRQ status bits being set.  This can be done
outside the lock--the hardware certainly isn't grabbing or looking at
the spinlock when it updates its status register.

It's possible that this will fix real (but very rare) errors seen in
the field that look like:
  irq ...: nobody cared (try booting with the "irqpoll" option)

NOTE: an alternate strategy considered here was to always make the
complete() / spi_finalize_current_transfer() the very last thing in
our IRQ handler.  With such a change you could consider that we could
be "lockless".  In that case, though, we'd have to be very careful w/
memory barriers so we made sure we didn't have any bugs with weakly
ordered memory.  Using spinlocks makes the driver much easier to
understand.

Fixes: 561de45f72bd ("spi: spi-geni-qcom: Add SPI driver support for GENI based QUP")
Signed-off-by: Douglas Anderson <dianders@chromium.org>
Reviewed-by: Stephen Boyd <swboyd@chromium.org>
Link: https://lore.kernel.org/r/20200618080459.v4.2.I752ebdcfd5e8bf0de06d66e767b8974932b3620e@changeid
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agospi: altera: fix size mismatch on 64 bit processors
Matthew Gerlach [Fri, 19 Jun 2020 01:43:41 +0000 (09:43 +0800)]
spi: altera: fix size mismatch on 64 bit processors

The spi-altera driver was originally written with a 32
bit processor, where sizeof(unsigned long) is 4.  On a
64 bit processor sizeof(unsigned long) is 8.  Change the structure
member to u32 to match the actual size of the control
register.

Signed-off-by: Matthew Gerlach <matthew.gerlach@linux.intel.com>
Signed-off-by: Xu Yilun <yilun.xu@intel.com>
Reviewed-by: Tom Rix <trix@redhat.com>
Link: https://lore.kernel.org/r/1592531021-11412-4-git-send-email-yilun.xu@intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agospi: altera: support indirect access to the registers
Xu Yilun [Fri, 19 Jun 2020 01:43:40 +0000 (09:43 +0800)]
spi: altera: support indirect access to the registers

This patch adds support for indirect access to the registers via parent
regmap.

The use case is, the spi master is a sub device of a Multifunction
device, which is connected to host by some indirect bus. To support this
device type, a new platform_device_id is introduced, and the driver tries
to get parent regmap for register accessing like many MFD sub device
drivers do.

Signed-off-by: Xu Yilun <yilun.xu@intel.com>
Signed-off-by: Matthew Gerlach <matthew.gerlach@linux.intel.com>
Signed-off-by: Wu Hao <hao.wu@intel.com>
Signed-off-by: Russ Weight <russell.h.weight@intel.com>
Reviewed-by: Tom Rix <trix@redhat.com>
Link: https://lore.kernel.org/r/1592531021-11412-3-git-send-email-yilun.xu@intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agospi: altera: use regmap-mmio instead of direct mmio register access
Xu Yilun [Fri, 19 Jun 2020 01:43:39 +0000 (09:43 +0800)]
spi: altera: use regmap-mmio instead of direct mmio register access

This patch adds support for regmap. It makes preparation for supporting
different ways to access the registers.

Signed-off-by: Matthew Gerlach <matthew.gerlach@linux.intel.com>
Signed-off-by: Wu Hao <hao.wu@intel.com>
Signed-off-by: Xu Yilun <yilun.xu@intel.com>
Signed-off-by: Russ Weight <russell.h.weight@intel.com>
Reviewed-by: Tom Rix <trix@redhat.com>
Link: https://lore.kernel.org/r/1592531021-11412-2-git-send-email-yilun.xu@intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agospi: spidev: fix a potential use-after-free in spidev_release()
Zhenzhong Duan [Thu, 18 Jun 2020 03:21:25 +0000 (11:21 +0800)]
spi: spidev: fix a potential use-after-free in spidev_release()

If an spi device is unbounded from the driver before the release
process, there will be an NULL pointer reference when it's
referenced in spi_slave_abort().

Fix it by checking it's already freed before reference.

Signed-off-by: Zhenzhong Duan <zhenzhong.duan@gmail.com>
Link: https://lore.kernel.org/r/20200618032125.4650-2-zhenzhong.duan@gmail.com
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agospi: spidev: fix a race between spidev_release and spidev_remove
Zhenzhong Duan [Thu, 18 Jun 2020 03:21:24 +0000 (11:21 +0800)]
spi: spidev: fix a race between spidev_release and spidev_remove

Imagine below scene, spidev is referenced after it's freed.

spidev_release()                spidev_remove()
...
                                spin_lock_irq(&spidev->spi_lock);
                                    spidev->spi = NULL;
                                spin_unlock_irq(&spidev->spi_lock);
mutex_lock(&device_list_lock);
dofree = (spidev->spi == NULL);
if (dofree)
    kfree(spidev);
mutex_unlock(&device_list_lock);
                                mutex_lock(&device_list_lock);
                                list_del(&spidev->device_entry);
                                device_destroy(spidev_class, spidev->devt);
                                clear_bit(MINOR(spidev->devt), minors);
                                if (spidev->users == 0)
                                    kfree(spidev);
                                mutex_unlock(&device_list_lock);

Fix it by resetting spidev->spi in device_list_lock's protection.

Signed-off-by: Zhenzhong Duan <zhenzhong.duan@gmail.com>
Link: https://lore.kernel.org/r/20200618032125.4650-1-zhenzhong.duan@gmail.com
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agospi: spi-geni-qcom: No need for irqsave variant of spinlock calls
Douglas Anderson [Tue, 16 Jun 2020 10:40:46 +0000 (03:40 -0700)]
spi: spi-geni-qcom: No need for irqsave variant of spinlock calls

The driver locks its locks in two places.

In the first usage of the lock the function doing the locking already
has a sleeping call and thus we know we can't be called from interrupt
context.  That means we can use the "spin_lock_irq" variant of the
function.

In the second usage of the lock the function is the interrupt handler
and we know interrupt handlers are called with interrupts disabled.
That means we can use the "spin_lock" variant of the function.

This patch is expected to be a no-op and is just a cleanup / slight
optimization.

Signed-off-by: Douglas Anderson <dianders@chromium.org>
Reviewed-by: Stephen Boyd <swboyd@chromium.org>
Link: https://lore.kernel.org/r/20200616034044.v3.1.Ic50cccdf27d42420a63485082f8b5bf86ed1a2b6@changeid
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agospi: bcm63xx-hsspi: add reset support
Álvaro Fernández Rojas [Tue, 16 Jun 2020 17:32:35 +0000 (19:32 +0200)]
spi: bcm63xx-hsspi: add reset support

bcm63xx arch resets the HSSPI controller at early boot. However, bmips arch
needs to perform a reset when probing the driver.

Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>
Reviewed-by: Philipp Zabel <p.zabel@pengutronix.de>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Link: https://lore.kernel.org/r/20200616173235.3473149-3-noltari@gmail.com
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agospi: bcm63xx-spi: add reset support
Álvaro Fernández Rojas [Tue, 16 Jun 2020 17:32:34 +0000 (19:32 +0200)]
spi: bcm63xx-spi: add reset support

bcm63xx arch resets the SPI controller at early boot. However, bmips arch
needs to perform a reset when probing the driver.

Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>
Reviewed-by: Philipp Zabel <p.zabel@pengutronix.de>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Link: https://lore.kernel.org/r/20200616173235.3473149-2-noltari@gmail.com
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agospi: stm32-qspi: Fix error path in case of -EPROBE_DEFER
Patrice Chotard [Tue, 16 Jun 2020 11:30:35 +0000 (13:30 +0200)]
spi: stm32-qspi: Fix error path in case of -EPROBE_DEFER

In case of -EPROBE_DEFER, stm32_qspi_release() was called
in any case which unregistered driver from pm_runtime framework
even if it has not been registered yet to it. This leads to:

stm32-qspi 58003000.spi: can't setup spi0.0, status -13
spi_master spi0: spi_device register error /soc/spi@58003000/mx66l51235l@0
spi_master spi0: Failed to create SPI device for /soc/spi@58003000/mx66l51235l@0
stm32-qspi 58003000.spi: can't setup spi0.1, status -13
spi_master spi0: spi_device register error /soc/spi@58003000/mx66l51235l@1
spi_master spi0: Failed to create SPI device for /soc/spi@58003000/mx66l51235l@1

On v5.7 kernel,this issue was not "visible", qspi driver was probed
successfully.

Fixes: 9d282c17b023 ("spi: stm32-qspi: Add pm_runtime support")
Signed-off-by: Patrice Chotard <patrice.chotard@st.com>
Link: https://lore.kernel.org/r/20200616113035.4514-1-patrice.chotard@st.com
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agoMerge series "Add MSIOF support for R8A7742 SOC" from Lad Prabhakar <prabhakar.mahade...
Mark Brown [Mon, 15 Jun 2020 23:38:42 +0000 (00:38 +0100)]
Merge series "Add MSIOF support for R8A7742 SOC" from Lad Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>:

Hi All,

This patch series adds support for MSIOF on R8A7742 SOC.

Cheers,
Prabhakar

Lad Prabhakar (2):
  dt-bindings: spi: renesas,sh-msiof: Add r8a7742 support
  ARM: dts: r8a7742: Add MSIOF[0123] support

 .../devicetree/bindings/spi/renesas,sh-msiof.yaml  |  1 +
 arch/arm/boot/dts/r8a7742.dtsi                     | 64 ++++++++++++++++++++++
 2 files changed, 65 insertions(+)

--
2.7.4

4 years agoMerge series "spi: bcm63xx: add BMIPS support" from Álvaro Fernández Rojas <noltari...
Mark Brown [Mon, 15 Jun 2020 23:38:41 +0000 (00:38 +0100)]
Merge series "spi: bcm63xx: add BMIPS support" from Álvaro Fernández Rojas <noltari@gmail.com>:

BCM63xx SPI and HSSPI controller are present on several BMIPS SoCs (BCM6318,
BCM6328, BCM6358, BCM6362, BCM6368 and BCM63268).

v2: use devm_reset_control_get_exclusive

Álvaro Fernández Rojas (4):
  spi: bcm63xx-spi: add reset support
  spi: bcm63xx-spi: allow building for BMIPS
  spi: bcm63xx-hsspi: add reset support
  spi: bcm63xx-hsspi: allow building for BMIPS

 drivers/spi/Kconfig             |  4 ++--
 drivers/spi/spi-bcm63xx-hsspi.c | 17 +++++++++++++++++
 drivers/spi/spi-bcm63xx.c       | 17 +++++++++++++++++
 3 files changed, 36 insertions(+), 2 deletions(-)

--
2.27.0

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

4 years agoMerge series "Add more configuration and regmap support for spi-altera" from Xu Yilun...
Mark Brown [Mon, 15 Jun 2020 23:38:40 +0000 (00:38 +0100)]
Merge series "Add more configuration and regmap support for spi-altera" from Xu Yilun <yilun.xu@intel.com>:

This patchset adds platform_data for spi-altera, to enable more IP
configurations, and creating specific spi client devices. It also adds
regmap support, to enable the indirect access to this IP.

We have a PCIE based FPGA platform which integrates this IP to communicate
with a BMC chip (Intel MAX10) over SPI. The IP is configured as 32bit data
width. There is also an indirect access interface in FPGA for host to
access the registers of this IP. This patchset enables this use case.

Matthew Gerlach (1):
  spi: altera: fix size mismatch on 64 bit processors

Xu Yilun (5):
  spi: altera: add 32bit data width transfer support.
  spi: altera: add SPI core parameters support via platform data.
  spi: altera: add platform data for slave information.
  spi: altera: use regmap instead of direct mmio register access
  spi: altera: move driver name string to header file

 drivers/spi/Kconfig        |   1 +
 drivers/spi/spi-altera.c   | 161 +++++++++++++++++++++++++++++++++++++--------
 include/linux/spi/altera.h |  37 +++++++++++
 3 files changed, 171 insertions(+), 28 deletions(-)
 create mode 100644 include/linux/spi/altera.h

--
2.7.4

4 years agospi: spi-cavium-thunderx: flag controller as half duplex
Tim Harvey [Thu, 28 May 2020 15:46:39 +0000 (08:46 -0700)]
spi: spi-cavium-thunderx: flag controller as half duplex

The OcteonTX (TX1/ThunderX) SPI controller does not support full
duplex transactions. Set the appropriate flag such that the spi
core will return -EINVAL on such transactions requested by chip
drivers.

This is an RFC as I need someone from Marvell/Cavium to confirm
if this driver is used for other silicon that does support
full duplex transfers (in which case we will need to identify
that we are running on the ThunderX arch before setting the flag).

Signed-off-by: Tim Harvey <tharvey@gateworks.com>
Cc: Robert Richter <rrichter@marvell.com>
Link: https://lore.kernel.org/r/1590680799-5640-1-git-send-email-tharvey@gateworks.com
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agospi: add Renesas RPC-IF driver
Sergei Shtylyov [Sat, 13 Jun 2020 19:18:34 +0000 (22:18 +0300)]
spi: add Renesas RPC-IF driver

Add the SPI driver for the Renesas RPC-IF.  It's the "front end" driver
using the "back end" APIs in the main driver to talk to the real hardware.
We only implement the 'spi-mem' interface -- there's no need to implement
the usual SPI driver methods...

Based on the original patch by Mason Yang <masonccyang@mxic.com.tw>.

Signed-off-by: Sergei Shtylyov <sergei.shtylyov@cogentembedded.com>
Link: https://lore.kernel.org/r/1ece0e6c-71af-f0f1-709e-571f4b0b4853@cogentembedded.com
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agospi: bcm63xx-hsspi: allow building for BMIPS
Álvaro Fernández Rojas [Mon, 15 Jun 2020 09:09:43 +0000 (11:09 +0200)]
spi: bcm63xx-hsspi: allow building for BMIPS

bcm63xx-hsspi controller is present on several BMIPS SoCs (BCM6318, BCM6328,
BCM6362 and BCM63268).

Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>
Acked-by: Florian Fainelli <f.fainelli@gmail.com>
Link: https://lore.kernel.org/r/20200615090943.2936839-5-noltari@gmail.com
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agospi: bcm63xx-spi: allow building for BMIPS
Álvaro Fernández Rojas [Mon, 15 Jun 2020 09:09:41 +0000 (11:09 +0200)]
spi: bcm63xx-spi: allow building for BMIPS

bcm63xx-spi controller is present on several BMIPS SoCs (BCM6358, BCM6362,
BCM6368 and BCM63268).

Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>
Acked-by: Florian Fainelli <f.fainelli@gmail.com>
Link: https://lore.kernel.org/r/20200615090943.2936839-3-noltari@gmail.com
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agospi: renesas,sh-msiof: Add r8a7742 support
Lad Prabhakar [Tue, 9 Jun 2020 20:54:13 +0000 (21:54 +0100)]
spi: renesas,sh-msiof: Add r8a7742 support

Document RZ/G1H (R8A7742) SoC bindings.

Signed-off-by: Lad Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
Reviewed-by: Marian-Cristian Rotariu <marian-cristian.rotariu.rb@bp.renesas.com>
Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>
Link: https://lore.kernel.org/r/1591736054-568-2-git-send-email-prabhakar.mahadev-lad.rj@bp.renesas.com
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agospi: altera: add platform data for slave information.
Xu Yilun [Thu, 11 Jun 2020 03:25:08 +0000 (11:25 +0800)]
spi: altera: add platform data for slave information.

This patch introduces platform data for slave information, it allows
spi-altera to add new spi devices once master registration is done.

Signed-off-by: Wu Hao <hao.wu@intel.com>
Signed-off-by: Xu Yilun <yilun.xu@intel.com>
Signed-off-by: Matthew Gerlach <matthew.gerlach@linux.intel.com>
Signed-off-by: Russ Weight <russell.h.weight@intel.com>
Reviewed-by: Tom Rix <trix@redhat.com>
Link: https://lore.kernel.org/r/1591845911-10197-4-git-send-email-yilun.xu@intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agospi: altera: add SPI core parameters support via platform data.
Xu Yilun [Thu, 11 Jun 2020 03:25:07 +0000 (11:25 +0800)]
spi: altera: add SPI core parameters support via platform data.

This patch introduced SPI core parameters in platform data, it
allows passing these SPI core parameters via platform data.

Signed-off-by: Wu Hao <hao.wu@intel.com>
Signed-off-by: Xu Yilun <yilun.xu@intel.com>
Signed-off-by: Matthew Gerlach <matthew.gerlach@linux.intel.com>
Signed-off-by: Russ Weight <russell.h.weight@intel.com>
Reviewed-by: Tom Rix <trix@redhat.com>
Link: https://lore.kernel.org/r/1591845911-10197-3-git-send-email-yilun.xu@intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agospi: altera: add 32bit data width transfer support.
Xu Yilun [Thu, 11 Jun 2020 03:25:06 +0000 (11:25 +0800)]
spi: altera: add 32bit data width transfer support.

Add support for 32bit width data register, then it supports 32bit
data width spi slave device and spi transfers.

Signed-off-by: Xu Yilun <yilun.xu@intel.com>
Signed-off-by: Wu Hao <hao.wu@intel.com>
Signed-off-by: Matthew Gerlach <matthew.gerlach@linux.intel.com>
Signed-off-by: Russ Weight <russell.h.weight@intel.com>
Reviewed-by: Tom Rix <trix@redhat.com>
Link: https://lore.kernel.org/r/1591845911-10197-2-git-send-email-yilun.xu@intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agoMerge existing fixes from spi/for-5.8
Mark Brown [Mon, 15 Jun 2020 15:16:05 +0000 (16:16 +0100)]
Merge existing fixes from spi/for-5.8

4 years agospi: uapi: spidev: Use TABs for alignment
Geert Uytterhoeven [Sat, 13 Jun 2020 07:37:54 +0000 (09:37 +0200)]
spi: uapi: spidev: Use TABs for alignment

The UAPI <linux/spi/spidev.h> uses TABs for alignment.
Convert the recently introduced spaces to TABs to restore consistency.

Fixes: 7bb64402a092136 ("spi: tools: Add macro definitions to fix build errors")
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Link: https://lore.kernel.org/r/20200613073755.15906-1-geert+renesas@glider.be
Signed-off-by: Mark Brown <broonie@kernel.org>
4 years agoLinux 5.8-rc1
Linus Torvalds [Sun, 14 Jun 2020 19:45:04 +0000 (12:45 -0700)]
Linux 5.8-rc1

4 years agoMerge tag 'LSM-add-setgid-hook-5.8-author-fix' of git://github.com/micah-morton/linux
Linus Torvalds [Sun, 14 Jun 2020 18:39:31 +0000 (11:39 -0700)]
Merge tag 'LSM-add-setgid-hook-5.8-author-fix' of git://github.com/micah-morton/linux

Pull SafeSetID update from Micah Morton:
 "Add additional LSM hooks for SafeSetID

  SafeSetID is capable of making allow/deny decisions for set*uid calls
  on a system, and we want to add similar functionality for set*gid
  calls.

  The work to do that is not yet complete, so probably won't make it in
  for v5.8, but we are looking to get this simple patch in for v5.8
  since we have it ready.

  We are planning on the rest of the work for extending the SafeSetID
  LSM being merged during the v5.9 merge window"

* tag 'LSM-add-setgid-hook-5.8-author-fix' of git://github.com/micah-morton/linux:
  security: Add LSM hooks to set*gid syscalls

4 years agosecurity: Add LSM hooks to set*gid syscalls
Thomas Cedeno [Tue, 9 Jun 2020 17:22:13 +0000 (10:22 -0700)]
security: Add LSM hooks to set*gid syscalls

The SafeSetID LSM uses the security_task_fix_setuid hook to filter
set*uid() syscalls according to its configured security policy. In
preparation for adding analagous support in the LSM for set*gid()
syscalls, we add the requisite hook here. Tested by putting print
statements in the security_task_fix_setgid hook and seeing them get hit
during kernel boot.

Signed-off-by: Thomas Cedeno <thomascedeno@google.com>
Signed-off-by: Micah Morton <mortonm@chromium.org>
4 years agoMerge tag 'for-5.8-part2-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/kdave...
Linus Torvalds [Sun, 14 Jun 2020 16:47:25 +0000 (09:47 -0700)]
Merge tag 'for-5.8-part2-tag' of git://git./linux/kernel/git/kdave/linux

Pull btrfs updates from David Sterba:
 "This reverts the direct io port to iomap infrastructure of btrfs
  merged in the first pull request. We found problems in invalidate page
  that don't seem to be fixable as regressions or without changing iomap
  code that would not affect other filesystems.

  There are four reverts in total, but three of them are followup
  cleanups needed to revert a43a67a2d715 cleanly. The result is the
  buffer head based implementation of direct io.

  Reverts are not great, but under current circumstances I don't see
  better options"

* tag 'for-5.8-part2-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/kdave/linux:
  Revert "btrfs: switch to iomap_dio_rw() for dio"
  Revert "fs: remove dio_end_io()"
  Revert "btrfs: remove BTRFS_INODE_READDIO_NEED_LOCK"
  Revert "btrfs: split btrfs_direct_IO to read and write part"