In case the data transfer failure happens, instead of returning
immediatelly, make sure the DMA is disabled, status register is
cleared and the bounce buffer is stopped.
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Dinh Nguyen <dinguyen@opensource.altera.com>
Cc: Pantelis Antoniou <panto@antoniou-consulting.com>
Cc: Tom Rini <trini@konsulko.com>
Endless timeouts are bad, since if we get stuck in one, we have no
way out. Zap this one by implementing proper timeout.
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Dinh Nguyen <dinguyen@opensource.altera.com>
Cc: Pantelis Antoniou <panto@antoniou-consulting.com>
Cc: Tom Rini <trini@konsulko.com>
The driver didn't stop the bounce buffer in case a data transfer
failed. This would lead to memory leakage if the communication
between the CPU and the card is unreliable. Add the missing call
to stop the bounce buffer.
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Dinh Nguyen <dinguyen@opensource.altera.com>
Cc: Pantelis Antoniou <panto@antoniou-consulting.com>
Cc: Tom Rini <trini@konsulko.com>
Derived from Tegra124, modified as appropriate during T210
board bringup. Cleaned up debug statements to conserve
string space, too. This also adds misc 64-bit changes
from Thierry Reding/Stephen Warren.
Signed-off-by: Tom Warren <twarren@nvidia.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
Signed-off-by: Stephen Warren <swarren@nvidia.com>
DCIMVAC is upgraded to DCCIMVAC for the individual processor
(Cortex-A7) that the DCIMVAC is executed on.
We should follow the linux dma follow. Before DMA read, first
invalidate dcache then after DMA read, invalidate dcache again.
With the DMA direction DMA_FROM_DEVICE, the dcache need be
invalidated again after the DMA completion. The reason is
that we need explicity make sure the dcache been invalidated
thus to get the DMA'ed memory correctly from the physical memory.
Any cache-line fill during the DMA operations such as the
pre-fetching can cause the DMA coherency issue, thus CPU get the stale data.
Signed-off-by: Peng Fan <Peng.Fan@freescale.com>
Signed-off-by: Ye.Li <B37916@freescale.com>
Signed-off-by: Nitin Garg <nitin.garg@freescale.com>
Signed-off-by: Jason Liu <r64343@freescale.com>
Reviewed-by: Stefano Babic <sbabic@denx.de>
This patch extracts checking for valid SD card "eGON.BT0" signature from
`board_mmc_init` into function `sunxi_mmc_has_egon_boot_signature`.
Buffer for mmc sector is allocated and freed at runtime. `panic` is
triggered on malloc failure.
Signed-off-by: Daniel Kochmański <dkochmanski@turtle-solutions.eu>
CC: Roy Spliet <r.spliet@ultimaker.com>
Cc: Ian Campbell <ijc@hellion.org.uk>
[hdegoede@redhat.com: Small bugfix to make it work for devs other then mmc0]
Acked-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
`mmc_initialize` might be called multiple times leading to the mmc-controllers
being initialised twice, and initialising the `mmc_devices` list head twice
which may lead to memory leaks.
Signed-off-by: Daniel Kochmański <dkochmanski@turtle-solutions.eu>
CC: Roy Spliet <r.spliet@ultimaker.com>
Cc: Ian Campbell <ijc@hellion.org.uk>
CC: Pantelis Antoniou <panto@antoniou-consulting.com>
Acked-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Add a test to confirm that we can probe this device. Since there is no
MMC stack support in sandbox at present, this is as far as the test goes.
Signed-off-by: Simon Glass <sjg@chromium.org>
Add basic support for MMC, providing a uclass which can set up an MMC
device. This allows MMC drivers to move to using driver model.
Signed-off-by: Simon Glass <sjg@chromium.org>
While implementing SDIO/MMC SPL booting for the Marvell Armada 38x, the
following problem occured. The SPL runs in internal SRAM which is
the L2 cache locked to memory. When the MMC buffers now are located
on the stack (or bss), the SDIO controller (SDHCI) can't write into
this L2 cache memory.
This patch introduces a method to use a fixed buffer that will be
used for all transfers by defining CONFIG_FIXED_SDHCI_ALIGNED_BUFFER.
This way, the board can use this buffer address located in SDRAM
for all transfers. This solves this SPL problem on the A38x and
should only be used in the SPL U-Boot version.
Tested for SPL booting on Marvell Armada 38x DB-88F6820-GP board.
Signed-off-by: Stefan Roese <sr@denx.de>
Cc: Pantelis Antoniou <panto@antoniou-consulting.com>
Cc: Luka Perkov <luka.perkov@sartura.hr>
The loop counter based timeout detection does not work on the Armada
38x based board (DB-88F6820-GP). At least with dcache enabled a
timeout is detected. Without dcache enabled, the timeout does not
occur. Increasing the loop counter solves this issue. But a better
solution is to use a timer based timeout detection instead. This
patch now implements this timer based detection.
Signed-off-by: Stefan Roese <sr@denx.de>
Cc: Pantelis Antoniou <panto@antoniou-consulting.com>
Cc: Luka Perkov <luka.perkov@sartura.hr>
Patch 33fe2fb8df titled
"mmc: bcm283x: Remove get_timer_us() from mmc driver"
incorrectly replaced ad-hoc get_timer_us() function
with a plain get_timer(). The get_timer() operates in
mSec units instead of uSec though, which caused very
slow operation of the driver.
Restore the original behavior of the driver, but avoid
get_timer_us() and use timer_get_us() instead. The later
is part of the standard API.
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Jakub Kiciński <moorray3@wp.pl>
Cc: Stephen Warren <swarren@wwwdotorg.org>
Tested-by: Jakub Kicinski <kubakici@wp.pl>
We want to see if the requested start or total block count are
unaligned. We discard the whole numbers and only care about the
remainder. Update the code to use div_u64_rem here and add a comment.
Cc: Hans de Goede <hdegoede@redhat.com>
Cc: Pantelis Antoniou <pantelis.antoniou@konsulko.com>
Cc: Bernhard Nortmann <bernhard.nortmann@web.de>
Reported-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Tom Rini <trini@konsulko.com>
Tested-by: Bernhard Nortmann <bernhard.nortmann@web.de>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
The get_timer_us() function is something which is no longer
existing in case we use generic timer framework, so replace
it with get_timer().
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Stephen Warren <swarren@wwwdotorg.org>
Cc: Tyler Baker <tyler.baker@linaro.org>
If the mmc device is non-removable (as indicated by the device tree), set
the flag so that users of the device know.
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Joe Hershberger <joe.hershberger@ni.com>
We need to clear the allocated memory explicitly as the included
struct sdhci_host has function pointers. Those are compared to NULL to
test if this (optional) feature is supported. Leaving them undefined let
u-boot jump to arbitrary memory.
Signed-off-by: Alexander Stein <alexanders83@web.de>
Freescale eMMC44 adapter card uses Micron N2M400FDB311A3CF eMMC
memory. According to the silicon datasheet, secure erase timeout
is 600ms. So increase erase timeout value from 250ms to 600ms.
Signed-off-by: Yangbo Lu <yangbo.lu@freescale.com>
Cc: York Sun <yorksun@freescale.com>
Timeout interrupt also work for response busy command(R1b) like
cmd38/cmd6. So need to set it accordingly. Current code only
set timeout for data command.
Signed-off-by: Kevin Liu <kliu5@marvell.com>
Signed-off-by: Rob Herring <robh@kernel.org>
Cc: Pantelis Antoniou <panto@antoniou-consulting.com>
High capacity support is not a host capability, but a device capability
that is queried via the OCR. The flag in the operating conditions
request argument can just be set unconditionally. This matches the Linux
implementation.
[panto] Hand merged and renumbering MMC_MODE_DDR_52MHz.
Signed-off-by: Rob Herring <robh@kernel.org>
Signed-off-by: Pantelis Antoniou <pantelis.antoniou@konsulko.com>
Cc: Pantelis Antoniou <pantelis.antoniou@konsulko.com>
Starting part of device initialization sets the init_in_progress flag
only if the MMC card did not yet come to ready state and needs to continue
polling. If the card is SD or if the MMC card became ready quickly,
the flag is not set and (if using pre-initialization) the starting
phase will be re-executed from mmc_init function.
Set the init_in_progress flag in all non-error cases. Also, move flags
setting statements around so that the flags are not set in error paths.
Also, IN_PROGRESS return status becomes unnecessary, so get rid of it.
Signed-off-by: Andrew Gabbasov <andrew_gabbasov@mentor.com>
The polling loops in sd_send_op_cond and mmc_complete_op_cond functions
check the ready flag state at the end of the loop, that is after executing
a delay inside the loop, which, in case of exiting with no error,
is not needed. Also, one of these loops, as well as the loop
in mmc_send_status, have the delay just before exiting on timeout
conditions.
Restructure all these loops to check the respective conditions before making
a delay for the next loop pass, and to appropriately exit without the delay.
Signed-off-by: Andrew Gabbasov <andrew_gabbasov@mentor.com>
Some MMC cards come to ready state quite quickly, so that the respective
flag appears to be set in mmc_send_op_cond already. In this case trying
to continue polling the card with CMD1 in mmc_complete_op_cond is incorrect
and may lead to unpredictable results. So check the flag before polling
and skip it appropriately.
Signed-off-by: Andrew Gabbasov <andrew_gabbasov@mentor.com>
The previous change to use 'ocr' structure field for storing send_op_cond
command response also stopped using command response directly
outside of mmc_send_op_cond_iter(). Now it becomes possible to use
command structure in mmc_send_op_cond_iter() locally, removing a necessity
to pass it as an argument from the caller.
Signed-off-by: Andrew Gabbasov <andrew_gabbasov@mentor.com>
The 'op_cond_response' field in mmc structure contains the response
from the last SEND_OP_COND MMC command while making iterational
polling of the card. Later it is copied to 'ocr' field, designed
to contain the OCR register value, which is actually the same
response from the same command. So, these fields have actually
the same data, just in different time periods. It's easier to use
the same 'ocr' field in both cases at once, without temporary using
of the 'op_cond_response' field.
Signed-off-by: Andrew Gabbasov <andrew_gabbasov@mentor.com>
The SD clock could be generated by platform clock or peripheral
clock for some platforms. This patch adds peripheral clock
support for T1024/T1040/T2080. To enable it, define
CONFIG_FSL_ESDHC_USE_PERIPHERAL_CLK.
Signed-off-by: Yangbo Lu <yangbo.lu@freescale.com>
Cc: York Sun <yorksun@freescale.com>
Cc: Pantelis Antoniou <panto@antoniou-consulting.com>
Reviewed-by: York Sun <yorksun@freescale.com>
Add adapter card type identification support by reading
FPGA STAT_PRES1 register SDHC Card ID[0:2] bits. To use this function,
define CONFIG_FSL_ESDHC_ADAPTER_IDENT.
Signed-off-by: Yangbo Lu <yangbo.lu@freescale.com>
Cc: York Sun <yorksun@freescale.com>
Cc: Pantelis Antoniou <panto@antoniou-consulting.com>
[York Sun: resolve conflicts in README.fsl-esdhc]
Reviewed-by: York Sun <yorksun@freescale.com>
The driver-model gpio functions may return another value then -1 as error,
make the sunxi mmc code properly handle this.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Acked-by: Ian Campbell <ijc@hellion.org.uk>
sun6i and newer (derived) SoCs such as the sun8i-a23, sun8i-a33 and sun9i
have a various things in common, like having separate ahb reset control
registers, the SID living inside the pmic, custom pmic busses, new style
watchdog, etc.
This commit introduces a new hidden SUNXI_GEN_SUN6I Kconfig bool which can be
used to check for these features avoiding the need for an ever growing list
of "#if defined CONFIG_MACH_SUN?I" conditionals as we add support for more
"new style" sunxi SoCs.
Note that this commit changes the behavior of the gmac and hdmi code for
sun8i and the upcoming sun9i devices. This does not matter as sun8i does
not have gmac nor hdmi, and sun9i has new hardware-blocks for these so
the old code will not work there.
Also this is intentional as if a sun8i / sun9i variant which does use the
old hwblocks shows up then the GEN_SUN6I code paths will be the right ones
to use.
For completeness this also adds a SUNXI_GEN_SUN4I bool for A10/A13/A20.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Acked-by: Ian Campbell <ijc@hellion.org.uk>
Remove the quirk SDHCI_QUIRK_NO_CD as it is not
required.
Signed-off-by: Siva Durga Prasad Paladugu <sivadur@xilinx.com>
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Change addresses to unsigned long to be compatible with 64-bit builds.
Regardless of fixing warnings, the device is still only 32-bit capable.
Signed-off-by: Rob Herring <robh@kernel.org>
Cc: Pantelis Antoniou <panto@antoniou-consulting.com>
Change addresses to unsigned long to be compatible with 64-bit builds.
Regardless of fixing warnings, the device is still only 32-bit capable.
Signed-off-by: Rob Herring <robh@kernel.org>
Cc: Pantelis Antoniou <panto@antoniou-consulting.com>
SDHCI_HOST_CONTROL is a byte-sized register, so don't write to it
as if it were a long, as that would result in clobbering the three
registers following.
Signed-off-by: Matt Reimer <mreimer@sdgsystems.com>
Properly mask SELBASECLK by using an actual mask rather than the
number of bits to shift in order to create the mask.
Signed-off-by: Matt Reimer <mreimer@sdgsystems.com>
Acked-by: Jaehoon Chung <jh80.chung@samsung.com>
Add support of the DDR mode for eSDHC driver.
Enable it for i.MX6 SoC family only.
Signed-off-by: Volodymyr Riazantsev <volodymyr.riazantsev@globallogic.com>
Reviewed-by: York Sun <yorksun@freescale.com>
Fix bus width switching from 8-bit mode down to 4-bit or 1-bit modes on
Samsung SoCs using SDHCI_QUIRK_USE_WIDE8. These SoCs report controller
version 2.0 yet they support 8-bit bus widths. If 8-bit mode was
previously enabled and then an operation like "mmc dev" caused a switch
back down to 4-bit or 1-bit mode, WIDE8 was left set, causing failures.
This problem was manifested by "mmc dev" timing out.
Signed-off-by: Matt Reimer <mreimer@sdgsystems.com>
Depending on the boot priority, the eMMC/SD cards,
can be initialized with the same numbers for each boot.
To be sure which mmc device is SD and which is eMMC,
this info is printed by 'mmc list' command, when
the init is done.
Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
Cc: Pantelis Antoniou <panto@antoniou-consulting.com>
Before this commit, the mmc devices were always registered
in the same order. So dwmmc channel 0 was registered as mmc 0,
channel 1 as mmc 1, etc.
In case of possibility to boot from more then one device,
the CONFIG_SYS_MMC_ENV_DEV should always point to right mmc device.
This can be achieved by init boot device as first, so it will be
always registered as mmc 0. Thanks to this, the 'saveenv' command
will work fine for all mmc boot devices.
Exynos based boards usually uses mmc host channels configuration:
- 0, or 0+1 for 8 bit - as a default boot device (usually eMMC)
- 2 for 4bit - as an optional boot device (usually SD card slot)
And usually the boot order is defined by OM pin configuration,
which can be changed in a few ways, eg.
- Odroid U3 - eMMC card insertion -> first boot from eMMC
- Odroid X2/XU3 - boot priority jumper
By this commit, Exynos dwmmc driver will check the OM pin configuration,
and then try to init the boot device and register it as mmc 0.
Signed-off-by: Przemyslaw Marczak <p.marczak@samsung.com>
Cc: Minkyu Kang <mk7.kang@samsung.com>
Cc: Jaehoon Chung <jh80.chung@samsung.com>
Cc: Pantelis Antoniou <panto@antoniou-consulting.com>
Cc: Simon Glass <sjg@chromium.org>
Cc: Akshay Saraswat <akshay.s@samsung.com>
High Capacity (e)MMC cards work fine on sun4i / sun5i, and not having this
capability set causes u-boot to not recognize the eMMC on an Utoo P66 A13
tablet, so always set it thereby fixing this.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Acked-by: Ian Campbell <ijc@hellion.org.uk>
