u-boot.git/drivers/mmc/fsl_esdhc_imx.c, branch master

global: Correct duplicate U_BOOT_DRIVER entry names

2026-04-07T17:32:56+00:00

The U_BOOT_DRIVER macro creates a list of drivers used at link time, and
all entries here must be unique. This in turn means that all entries in
the code should also be unique in order to not lead to build failures
later with unexpected build combinations. Typically, the problem we have
here is when a driver is obviously based on another driver and didn't
update this particular field and so while the name field reflects
something unique the linker entry itself is not. In a few places this
provides a more suitable string name as well, however.

Reviewed-by: Marek Vasut 
Reviewed-by: Svyatoslav Ryhel  # Tegra
Reviewed-by: Peter Robinson 
Reviewed-by: Heiko Schocher 
Reviewed-by: Simon Glass 
Signed-off-by: Tom Rini

mmc: fsl_esdhc_imx: Skip voltage switching for fixed 1.8V regulator

2026-02-28T18:31:49+00:00

When using a fixed 1.8V regulator for vqmmc (indicated by vs18_enable),
attempting to change the voltage produces spurious errors since the
regulator cannot be adjusted. The driver currently attempts the voltage
change, receives -ENOSYS from the regulator subsystem, and reports:

  Setting to 1.8V error: -38
  esdhc_set_voltage error -5

Fix this by checking vs18_enable early in esdhc_set_voltage() and
returning -ENOTSUPP for all voltage switch requests, not just 3.3V.
This prevents unnecessary regulator operations and eliminates the
error messages when the hardware is correctly configured with a fixed
1.8V supply.

Signed-off-by: Kory Maincent

mmc: fsl_esdhc_imx: Cannot test unsigned to be < 0

2025-12-04T15:38:16+00:00

Testing an unisgned member of a struct to be <= 0 will only detect the
case when it is 0. So correct this error test to a working version that
will behave as expected.

Signed-off-by: Andrew Goodbody

mmc: fsl_esdhc: Migrate ESDHCI_QUIRK_BROKEN_TIMEOUT_VALUE to Kconfig

2025-03-13T18:15:50+00:00

The flag for enabling the ESDHCI_QUIRK_BROKEN_TIMEOUT_VALUE quirk can be
handled easily enough in Kconfig. This lets us remove a function but not
obviously correct usage of the IS_ENABLED() macro.

Signed-off-by: Tom Rini 
Reviewed-by: Peng Fan

mmc: fsl_esdhc: Convert to use livetree API for fdt access

2024-10-25T12:07:43+00:00

Convert to using livetree API functions.

Without this if livetree is enabled (OF_LIVE) the fsl_esdhc_imx
driver will fail to read vendor-specific properties from the dt.

Signed-off-by: Tim Harvey 
Reviewed-by: Jaehoon Chung

mmc: fsl_esdhc_imx: Reset tuning logic

2024-10-04T12:15:15+00:00

When supporting partition reset for SoC such as i.MX95 , the Linux
Kernel may have configured the tuning, while after force reset by
wdog or else, uboot CMD0 will never pass unless config RSTT to reset
tuning logic.

Since RSTA and RSTT are independent, so need both to be reseted in the
controller.

Acked-by: Haibo Chen 
Signed-off-by: Peng Fan

mmc: fsl_esdhc_imx: Avoid resource leak

2024-10-04T12:15:15+00:00

The memory of priv and plat are leaked if max_bus_width is wrong.

Signed-off-by: Ye Li 
Reviewed-by: Peng Fan 
Signed-off-by: Peng Fan

mmc: fsl_esdhc_imx: Fix host_caps issue for non-DM driver

2024-10-04T12:15:15+00:00

The plat->cfg is wrongly memset to 0, so the host_caps value configured
in fsl_esdhc_initialize is reset. Remove the unnecessary memset since
plat is allocated via calloc.

Reviewed-by: Peng Fan 
Signed-off-by: Ye Li 
Signed-off-by: Peng Fan

mmc: fsl_esdhc_imx: Send 80 clocks before IDLE command

2024-10-04T12:15:15+00:00

According to SD and MMC spec, 74 clocks must be sent to device after
power stable. This is need in reinit ops for DM MMC or init ops for
non-DM MMC after power cycle.

So set the INTIA to send 80 clocks in esdhc_init_common and move
its calling from probe to reinit.

However, on 8MQ EVK and 8QXP MEK with some brands of SD cards, sending
80 clocks may not work well.

The root cause is related with power up time.  According to spec, after
power stable, host shall supply at least 74 SD clocks to the SD card with
the maximum of 1ms. However, the power ram up time is related with the
characteristic of SD card. At the moment of sending 74 SD clocks, the
power probably not ram up to the operating level on the problematic
cards. Then cause the cards not ready.

This patch changes to send SD clock with 1ms duration to replace 80
SD clocks (0.2ms at 400Khz clock).
This way meets the spec requirement as well, and adds the margin for
power ram up time to be compatible with the problematic SD cards.
This is also aligned with implementation which has FORCE clock
always on.

Reviewed-and-tested-by: Haibo Chen 
Signed-off-by: Ye Li 
Signed-off-by: Peng Fan

mmc: fsl_esdhc_imx: Enable AHB/IPG clk with clk bulk API

2024-10-04T12:15:15+00:00

With partition reset supported for i.MX8QM/QXP/95 and etc, when linux
mmc runtime suspended, the mmc clks are gated off. While at same time
system controller reset Cortex-A cores because of various reasons(
WDOG timeout and etc), with SPL run again, only enable PER clk is not
enough, also need to enable AHB/IPG clk, here use clk bulk API to enable
all the clocks.

Signed-off-by: Peng Fan