Some MMC peripherals require specific power cycle sequence, where some
registers need to be written between the regulator is turned off and then
back on. This is the case for the MMC IP embedded in STM32MP1 SoC.
In STM32MP157 reference manual [1], the power cycle sequence is:
1. Reset the SDMMC with the RCC.SDMMCxRST register bit. This will reset
the SDMMC to the reset state and the CPSM and DPSM to the Idle state.
2. Disable the Vcc power to the card.
3. Set the SDMMC in power-cycle state. This will make that the
SDMMC_D[7:0], SDMMC_CMD and SDMMC_CK are driven low, to prevent the card
from being supplied through the signal lines.
4. After minimum 1ms enable the Vcc power to the card.
5. After the power ramp period set the SDMMC to the power-off state for
minimum 1ms. The SDMMC_D[7:0], SDMMC_CMD and SDMMC_CK are set to
drive “1”.
6. After the 1ms delay set the SDMMC to power-on state in which the
SDMMC_CK clock will be enabled.
7. After 74 SDMMC_CK cycles the first command can be sent to the card.
The step 3. cannot be handled by the current framework implementation.
A new callback (host_power_cycle) is created, and called in
mmc_power_cycle(), after mmc_power_off().
The incorrect power cycle sequence has shown some boot failures on
STM32MP1 with some SD-cards, especially on cold boots when the input
frequency is low (<= 25MHz).
Those failures are no more seen with this correct power cycle sequence.
[1] https://www.st.com/resource/en/reference_manual/DM00327659.pdf
Signed-off-by: Yann Gautier <yann.gautier@st.com>
In SPL, all the available mmc devices gets initialized during boot.
This might not work in cases where clocks are not available for
certain mmc devices(other than boot device) and the support for
enabling device might not be ready.
Texas Instruments' K3 J721E device having a central system controller
(dmsc) is one such example falling in this category. Below is the
sequence for the failing scenario:
- ROM comes up in SD mode and loads SPL by just initialing SD card.
- SPL loads dmsc firmware from SD Card.
Since ROM has enabled SD, SPL need not enable the SD, just need
to re initialize the card. But SPL is trying to initialize other MMC
instances which are in disabled state. Since dmsc firmware is not yet
available, devices cannot be enabled. So in SPL, initialize only the
mmc device that is needed.
Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
Reviewed-by: Peng Fan <peng.fan@nxp.com>
At present there is only one control for this and it is used for both SPL
and TPL. But SPL might have a lot more space than TPL so the extra cost of
a full printf() might be acceptable.
Split the option into two, providing separate SPL and TPL controls. The
TPL setting defaults to the same as SPL.
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
It's quite hard to figure out time units for various function that have
timeout parameters. This leads to possible errors when one forgets to
convert ms to us, for example. Let's rename those parameters
correspondingly to 'timeout_us' and 'timeout_ms' to prevent such issues
further.
While at it, add time units info as comments to struct mmc fields.
This commit doesn't change the behavior, only renames parameters names.
Buildman should report no changes at all.
Signed-off-by: Sam Protsenko <semen.protsenko@linaro.org>
Reviewed-by: Peng Fan <peng.fan@nxp.com>
Reviewed-by: Igor Opaniuk <igor.opaniuk@gmail.com>
mmc_wait_dat0() expects timeout argument to be in usec units. But some
overlying functions operate on timeout in msec units. Convert timeout
from msec to usec when passing it to mmc_wait_dat0().
This fixes 'avb' commands on BeagleBoard X15, because next chain was
failing:
get_partition() -> mmc_switch_part() -> __mmc_switch() ->
mmc_wait_dat0()
when passing incorrect timeout from __mmc_switch() to mmc_wait_dat0().
Fixes: bb98b8c5c0 ("mmc: During a switch, poll on dat0 if available and check the final status")
Signed-off-by: Sam Protsenko <semen.protsenko@linaro.org>
Reviewed-by: Eugeniu Rosca <rosca.eugeniu@gmail.com>
Tested-by: Eugeniu Rosca <rosca.eugeniu@gmail.com>
Reviewed-by: Peng Fan <peng.fan@nxp.com>
Tested-by: Igor Opaniuk <igor.opaniuk@gmail.com>
Reviewed-by: Igor Opaniuk <igor.opaniuk@gmail.com>
The comment about init op being NULL used to be next to the NULL check
code. Commit 8ca51e51c1 ("dm: mmc: Add a way to use driver model for
MMC operations") separated the comment from the code. Put them back
together.
Fixes: 8ca51e51c1 ("dm: mmc: Add a way to use driver model for MMC operations")
Signed-off-by: Baruch Siach <baruch@tkos.co.il>
Reviewed-by: Peng Fan <peng.fan@nxp.com>
eMMC 5.1+ supports HS400 Enhances Strobe mode without the need for
tuning procedure.
The flow is as following:
- set HS_TIMIMG (Highspeed)
- Host change freq to <= 52Mhz
- set the bus width to Enhanced strobe and DDR8Bit(CMD6),
EXT_CSD[183] = 0x86 instead of 0x80
- set HS_TIMING to 0x3 (HS400)
- Host change freq to <= 200Mhz
- Host select HS400 enhanced strobe complete
Signed-off-by: Peng Fan <peng.fan@nxp.com>
The MMC mode and width are fixed for MMC SPI host hence we skip
sd_select_mode_and_width() and mmc_select_mode_and_width() for
MMC SPI host.
Signed-off-by: Anup Patel <anup.patel@wdc.com>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
Tested-by: Bin Meng <bmeng.cn@gmail.com>
Accessing the boot partition had been error prone with HS200 and HS400 and
was disabled. The driver first switched to a lesser mode and then switched
the partition access. It was mostly due to a bad handling of the switch and
has been fixed, so let's remove this 'feature'
Signed-off-by: Jean-Jacques Hiblot <jjhiblot@ti.com>
The switch operation can sometimes make the bus unreliable, in that case
the send_status parameter should be false to indicate not to poll using
CMD13. If polling on dat0 is possible, we should use it to detect the end
of the operation.
At the end of the operation it is safe to use CMD13 to get the status of
the card. It is important to do so because the operation may have failed.
Signed-off-by: Jean-Jacques Hiblot <jjhiblot@ti.com>
The e-MMC spec allows the e-MMC to specify a timeout for the partition
switch command. It can take up to 2550 ms. There is no lower limit to this
value in the spec, but do as the the linux driver does and force it to be
at least 300ms.
Signed-off-by: Jean-Jacques Hiblot <jjhiblot@ti.com>
Starting with rev 4.5, the eMMC can define a generic timeout for the
SWITCH command.
Following Linux Kernel code, the timeout also changed from 1000 -> 500
Signed-off-by: Jean-Jacques Hiblot <jjhiblot@ti.com>
Using the DAT0 line as a rdy/busy line is an alternative to reading the
status register of the card. It especially useful in situation where the
bus is not in a good shape, like when modes are switched.
This is also how the linux driver behaves.
Note of warning: As per the specification, while polling on DAT0 the CLK
must not turned off: "[...] Without a clock edge the Device (unless
previously disconnected by a deselect command (CMD7)) will force the DAT0
line down, forever. [...]"
Signed-off-by: Jean-Jacques Hiblot <jjhiblot@ti.com>
mmc_send_status() is currently used to poll the card until it is ready, not
actually returning the status of the card.
Make it return the status and add another function to poll the card.
Also remove the 'extern' declaration in the mmc-private.h header to comply
with the coding standard.
Signed-off-by: Jean-Jacques Hiblot <jjhiblot@ti.com>
This reverts commit 318a7a576b.
The last and only user of this callback had been the omap_hsmmc driver.
It is not used anymore. Removing the callback.
Signed-off-by: Jean-Jacques Hiblot <jjhiblot@ti.com>
U-Boot code currently only applies this restriction to HS200 mode,
extend this to HS400 mode as well.
Currently U-Boot code not support accessing boot partition in HS200/400
mode. This needs more check.
Signed-off-by: Marek Vasut <marek.vasut+renesas@gmail.com>
Cc: Jean-Jacques Hiblot <jjhiblot@ti.com>
Cc: Nobuhiro Iwamatsu <iwamatsu@nigauri.org>
Cc: Peng Fan <peng.fan@nxp.com>
Reviewed-by: Peng Fan <peng.fan@nxp.com>
After the commit b9a2a0e2e9 ("mmc: Add support for downgrading
HS200/HS400 to HS mode"), it add a parameter in mmc_set_card_speed()
which indicates that the HS200/HS400 to HS downgrade is happening.
During the HS400 initialization, first select to HS200, and config
the related clock rate, then downgrade to HS mode. So here also need
to config the downgrade value to be true for two reasons. First,
make sure in the function mmc_set_card_speed(), after switch to HS
mode, first config the clock rate, then read the EXT_CSD, avoid
receiving data of EXT_CSD in HS mode at 200MHz. Second, after issue
the MMC_CMD_SWITCH command, it need to wait a bit then switch bus
properties.
Test on i.MX8QM MEK board, some Micron eMMC will stuck in transfer
mode in this case, and USDHC will never get data transfer complete
status, cause the uboot hang.
Signed-off-by: Haibo Chen <haibo.chen@nxp.com>
Acked-by: Marek Vasut <marek.vasut@gmail.com>
The tiny printf implementation only supports %x format specifier,
it does not support %X . Since it makes little difference whether
the debug output prints hex numbers in capitals or not, change it
to %x and make the MMC_TRACE output work with tiny printf too.
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Jaehoon Chung <jh80.chung@samsung.com>
Cc: Simon Glass <sjg@chromium.org>
Older kernel versions or systems which do not connect eMMC reset line
properly may not be able to handle situations where either the eMMC
is left in HS200/HS400 mode or SD card in UHS modes by the bootloader
and may misbehave. Downgrade the eMMC to HS/HS52 mode and/or SD card
to non-UHS mode before booting the kernel to allow such older kernels
to work with modern U-Boot.
Signed-off-by: Marek Vasut <marek.vasut+renesas@gmail.com>
Cc: Tom Rini <trini@konsulko.com>
When using CMD6 to switch eMMC card timing from HS200/HS400 to HS/legacy,
do not poll for the completion status using CMD13, but rather wait 50mS.
Once the card receives the CMD6 and starts executing it, the bus is in
undefined state until both the card finishes executing the command and
until the controller switches the bus to matching timing configuration.
During this time, it is not possible to transport any commands or data
across the bus, which includes the CMD13.
Signed-off-by: Marek Vasut <marek.vasut+renesas@gmail.com>
Cc: Jaehoon Chung <jh80.chung@samsung.com>
The mmc_select_mode_and_width() function can be called while the card
is in HS200/HS400 mode and can be used to downgrade the card to lower
mode, e.g. HS. This is used for example by mmc_boot_part_access_chk()
which cannot access the card in HS200/HS400 mode and which is in turn
called by saveenv if env is in the MMC.
In such case, forcing the card clock to legacy frequency cannot work.
Instead, the card must be switched to HS mode first, from which it can
then be reprogrammed as needed.
However, this procedure needs additional code changes, since the current
implementation checks whether the card correctly switched to HS mode in
mmc_set_card_speed(). The check only expects that the card will be going
to HS mode from lower modes, not from higher modes, hence add a parameter
which indicates that the HS200/HS400 to HS downgrade is happening. This
makes the code send the switch command first, reconfigure the controller
next and finally perform the EXT_CSD readback check. The last two steps
cannot be done in reverse order as the card is already in HS mode when
the clock are being switched on the controller side.
Signed-off-by: Marek Vasut <marek.vasut+renesas@gmail.com>
Cc: Jaehoon Chung <jh80.chung@samsung.com>
Commit d0851c8937 ("blk: Call part_init() in the post_probe() method")
removed the call to part_init() in mmc.c, as this is done by the DM_MMC
framework.
However Allwinner is (still) relying on a non-DM MMC driver, so we are
now missing the implicit partition init, leading to failing MMC accesses
due to the missing partition information.
Bring the call back just for non-DM MMC driver to fix this regression.
Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Tested-by: Soeren Moch <smoch@web.de>
According to SD Specifications Part 1 Physical Layer Simplified Specification
Version 6.00 August 29, 2018, section 4.3.10 (Switch Function Command) and
section 5.6 (SCR register), SD cards version 1.00 and 1.01 do not support the
SD CMD 6.
Currently, U-Boot will issue CMD 6 unconditionally in sd_set_card_speed()
while configuring the bus for selected frequency. This will make SD cards
version 1.00 and 1.01 time out and thus fail detection altogether.
Fix this by not sending CMD 6 on such cards. Tested on Matsushita Electric
Industrial Co., Ltd. Japan RP-SD008B / Victor 8MB SD card, CU-SD008, which
is correctly detected with this patch as:
Device: sd@ee160000
Manufacturer ID: 1
OEM: 5041
Name: S008B
Bus Speed: 25000000
Mode : SD Legacy
Rd Block Len: 512
SD version 1.0
High Capacity: No
Capacity: 6.5 MiB
Bus Width: 4-bit
Erase Group Size: 512 Bytes
Signed-off-by: Marek Vasut <marek.vasut+renesas@gmail.com>
Cc: Masahiro Yamada <yamada.masahiro@socionext.com>
part_init() is currently called in every DM BLK driver, either
in its bind() or probe() method. However we can use the BLK
uclass driver's post_probe() method to do it automatically.
Update all DM BLK drivers to adopt this change.
Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Add HS400 support.
Selecting HS400 needs first select HS200 according to spec, so use
a dedicated function for HS400.
Add HS400 related macros.
Remove the restriction of only using the low 6 bits of
EXT_CSD_CARD_TYPE, using all the 8 bits.
Signed-off-by: Peng Fan <peng.fan@nxp.com>
Cc: Jaehoon Chung <jh80.chung@samsung.com>
Cc: Jean-Jacques Hiblot <jjhiblot@ti.com>
Cc: Stefano Babic <sbabic@denx.de>
Cc: Kishon Vijay Abraham I <kishon@ti.com>
Cc: Faiz Abbas <faiz_abbas@ti.com>
Cc: Marek Vasut <marex@denx.de>
Signed-off-by: Peng Fan <peng.fan@nxp.com>
This code is useful for testing the existance of devices that
do not have card detect capabilities. This breaks out the core
functionality and leaves the actual init logic and error reporting
in mmc_start_init().
Signed-off-by: Jon Nettleton <jon@solid-run.com>
Signed-off-by: Baruch Siach <baruch@tkos.co.il>
Reviewed-by: Stefano Babic <sbabic@denx.de>
Drop all the extra content from the MMC core, so that tiny MMC support
is really tiny, no fancy anything. That means the tiny MMC support does
only 1-bit transfers at default speed settings. Moreover, this patch
drops duplicate instance of struct mmc mmc_static, which wasted about
360 bytes. Furthermore, since MMC tiny supports only one controller
at all times, get rid of mallocating the ext csd backup and replace
it with static array. All in all, this patch saves ~4 kiB of bloat
from the MMC core, which on platforms with severe limitations can be
beneficial.
Signed-off-by: Marek Vasut <marek.vasut@gmail.com>
Cc: Jaehoon Chung <jh80.chung@samsung.com>
[trini: Fixup checkpatch.pl style warnings]
Signed-off-by: Tom Rini <trini@konsulko.com>
This patch changed the datatype of variable "start" from uint to ulong
to work properly on 64-bit machines as well. Also the return type of
get_timer() function is ulong.
Signed-off-by: Vipul Kumar <vipul.kumar@xilinx.com>
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Add the debug message for checking the mmc clock status.
It's helpful to debug the controlling clock.
Signed-off-by: Jaehoon Chung <jh80.chung@samsung.com>
mmc_set_clock() function has the disable argument as bool type.
When mmc_set_clock is called, it might be passed to "true" or "false".
But it's too confusion whether clock is enabled or disabled with only
"true" and "false".
To prevent the confusion, replace to MMC_CLK_ENABLE/DISABLE macro from
true/false.
Signed-off-by: Jaehoon Chung <jh80.chung@samsung.com>
When U-Boot started using SPDX tags we were among the early adopters and
there weren't a lot of other examples to borrow from. So we picked the
area of the file that usually had a full license text and replaced it
with an appropriate SPDX-License-Identifier: entry. Since then, the
Linux Kernel has adopted SPDX tags and they place it as the very first
line in a file (except where shebangs are used, then it's second line)
and with slightly different comment styles than us.
In part due to community overlap, in part due to better tag visibility
and in part for other minor reasons, switch over to that style.
This commit changes all instances where we have a single declared
license in the tag as both the before and after are identical in tag
contents. There's also a few places where I found we did not have a tag
and have introduced one.
Signed-off-by: Tom Rini <trini@konsulko.com>
sd_read_ssr returns 0, means no error.
Fixes: 5b2e72f32721484("mmc: read ssr only if MMC write support is enabled")
Signed-off-by: Peng Fan <peng.fan@nxp.com>
Cc: Jaehoon Chung <jh80.chung@samsung.com>
Cc: Jean-Jacques Hiblot <jjhiblot@ti.com>
Staticize the function since it's only used in mmc.c .
Signed-off-by: Marek Vasut <marek.vasut@gmail.com>
Cc: Jaehoon Chung <jh80.chung@samsung.com>
Cc: Tom Rini <trini@konsulko.com>
MMC card with EXT_CSD_REV value 9 will trigger off-by-one
bug while accessing mmc_versions array. The patch fix that.
Signed-off-by: Alexander Kochetkov <al.kochet@gmail.com>
If some configs are disabled, number of freqs array will not assigned to
correct value with bus_mode.
Synchornize the ordering with enum bus_mode in mmc.h.
Signed-off-by: Jaehoon Chung <jh80.chung@samsung.com>
The correspondence between mmc versions as used in u-boot and the version
numbers reported in register EXT_CSD_REV is wrong for versions above and
including MMC_VERSION_4_41. All those versions were shifted by one:
real 4.5 hardware appeared to be MMC_VERSION_5_0.
Fix this by adding the missing version in the correspondence table.
Reported-by: eil Eilmsteiner Heribert <eil@keba.com>
Signed-off-by: Jean-Jacques Hiblot <jjhiblot@ti.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Sam Protsenko <semen.protsenko@linaro.org>
After settings the speed of the sd with the switch command, a check is
done to make sure that the new speed has been set. The current check has a
masking error: speed are encoded on 4 bits only.
Fix it by masking the upper bits.
This fixes a problem seen with QEmu emulating a vexpress-a15.
Reported-by: Jonathan Gray <jsg@jsg.id.au>
Signed-off-by: Jean-Jacques Hiblot <jjhiblot@ti.com>
Tested-by: Jonathan Gray <jsg@jsg.id.au>
Use pr_* log functions from Linux. They can be enabled/disabled
via CONFIG_LOGLEVEL.
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Reviewed-by: Lukasz Majewski <lukma@denx.de>
Reviewed-by: Jaehoon Chung <jh80.chung@samsung.com>
When clock is enabling, it's assigned to 0 as mmc->clock.
Then it can't initialize any card.
Fix to assign to correct clock value as mmc->cfg->f_min or f_max.
Fixes: 9546eb92cb ("mmc: fix the wrong disabling clock")
Signed-off-by: Jaehoon Chung <jh80.chung@samsung.com>
Tested-by: Guillaume GARDET <guillaume.gardet@free.fr>
Tested-by: Anand Moon <linux.amoon@gmail.com>
Tested-by: Stephen Warren <swarren@nvidia.com>
When power is off, clock is not disabling.
Because it's passed to 1, mmc->clock should be set to f_min value.
Some drivers can't initialize the eMMC/SD card with current status.
This patch is to fix the disabling clock value to 0.
Fixes: 2e7410d76a ("mmc: disable the mmc clock during power off")
Signed-off-by: Jaehoon Chung <jh80.chung@samsung.com>
Reviewed-by: Jean-Jacques Hiblot <jjhiblot@ti.com>
Tested-by: Guillaume GARDET <guillaume.gardet@free.fr>
Tested-by: Anand Moon <linux.amoon@gmail.com>
hc_wp_grp_size is needed only if hardware partitionning is used.
On ARM removing it saves about 30 bytes of code space.
Signed-off-by: Jean-Jacques Hiblot <jjhiblot@ti.com>
This information is only used by the "mmc info" command.
On ARM removing this information from SPL saves about 140 of code space.
Signed-off-by: Jean-Jacques Hiblot <jjhiblot@ti.com>
Also remove erase_grp_size and write_bl_len from struct mmc as they are
not used anymore. On ARM, removing them saves about 100 bytes of code
space in SPL.
Signed-off-by: Jean-Jacques Hiblot <jjhiblot@ti.com>
The content of ssr is useful only for erase operations.
on ARM, removing sd_read_ssr() saves around 300 bytes.
Signed-off-by: Jean-Jacques Hiblot <jjhiblot@ti.com>
Not all boards have an eMMC and not all users have a need for this.
Allow to compile it out. By default it is still included.
Signed-off-by: Jean-Jacques Hiblot <jjhiblot@ti.com>
