They're still in the operating range and appear stable on my devices. YMMV.

Description: uSDHC AHB_BUS and IPG_CLK clocks must be synchronized.
Due to current physical design implementation, AHB_BUS and IPG_CLK must come from
same clock source to maintain clock sync.
Workaround: Set AHB_BUS and IPG_CLK to clock source from PLL1.

After sys1_pll_266m gate off/on, seems need to sync the USDHC AHB bus and USDHC IPG_clk again. (Here usdhc AHB BUS source from nand_usdhc_bus.)
This sync is handle by hardware, and maybe need some time, during this sync period, usdhc operation may has issue.

I just double check our local v5.10 branch, already revert the commit b04383b6 (clk: imx8mq: Define gates for pll1/2 fixed dividers).
So to fix this issue, one method is revert this patch, another method is keep the 'nand_usdhc_bus' always on. Add change like this:
Signed-off-by: Angus Ainslie <angus@akkea.ca>

see
https://lore.kernel.org/linux-arm-kernel/1571192067-19600-2-git-send-email-Anson.Huang@nxp.com/

The original version layout seems to have changed with the driver upgrade
so print out the old layout as well as the new one.
Signed-off-by: Angus Ainslie <angus@akkea.ca>

As suggested by Marek Vasut <marex@denx.de> the COEX test should and
not an or.
Signed-off-by: Angus Ainslie <angus@akkea.ca>

In file included from /var/lib/dkms/rs9116/1.2.20/build/rsi_91x_hal.c:37:
/var/lib/dkms/rs9116/1.2.20/build/rsi_sdio.h:44: error: "SDIO_VENDOR_ID_RSI" redefined [-Werror]
 #define SDIO_VENDOR_ID_RSI              0x041B

In file included from /var/lib/dkms/rs9116/1.2.20/build/rsi_sdio.h:40,
                 from /var/lib/dkms/rs9116/1.2.20/build/rsi_91x_hal.c:37:
./include/linux/mmc/sdio_ids.h:124: note: this is the location of the previous definition
 #define SDIO_VENDOR_ID_RSI   0x041b

The inner do { ... } while loop is completely useless, all it does
is iterate over a switch-case statement, one bit at a time. This
can easily be replaced by simple if (status & bit) { ... } tests
for each bit. No functional change.
Signed-off-by: Marek Vasut <marex@denx.de>

The interrupt handling of the RS911x is particularly heavy. For each RX
packet, the card does three SDIO transactions, one to read interrupt
status register, one to RX buffer length, one to read the RX packet(s).
This translates to ~330 uS per one cycle of interrupt handler. In case
there is more incoming traffic, this will be more.

The drivers/mmc/core/sdio_irq.c has the following comment, quote "Just
like traditional hard IRQ handlers, we expect SDIO IRQ handlers to be
quick and to the point, so that the holding of the host lock does not
cover too much work that doesn't require that lock to be held."

The RS911x interrupt handler does not fit that. This patch therefore
changes it such that the entire IRQ handler is moved to the RX thread
instead, and the interrupt handler only wakes the RX thread.

This is OK, because the interrupt handler only does things which can
also be done in the RX thread, that is, it checks for firmware loading
error(s), it checks buffer status, it checks whether a packet arrived
and if so, reads out the packet and passes it to network stack.

Moreover, this change permits removal of a code which allocated an
skbuff only to get 4-byte-aligned buffer, read up to 8kiB of data
into the skbuff, queue this skbuff into local private queue, then in
RX thread, this buffer is dequeued, the data in the skbuff as passed
to the RSI driver core, and the skbuff is deallocated. All this is
replaced by directly calling the RSI driver core with local buffer.
Signed-off-by: Marek Vasut <marex@denx.de>

Signed-off-by: Sebastian Krzyszkowiak <sebastian.krzyszkowiak@puri.sm>
Signed-off-by: Angus Ainslie <angus@akkea.ca>

Also drop the HW_SCAN_OFFLOAD
Signed-off-by: Sebastian Krzyszkowiak <sebastian.krzyszkowiak@puri.sm>
Signed-off-by: Angus Ainslie <angus@akkea.ca>

Import the updated driver from here

https://github.com/SiliconLabs/RS911X-nLink-OSDSigned-off-by: Angus Ainslie <angus@akkea.ca>

Clean up in preperation for a new code drop
Signed-off-by: Angus Ainslie <angus@akkea.ca>

This violates USB 2.0 spec, but any port should be able to handle
it these days (and USB 3.0 spec even requires it).

The HP detect gpio bounces at 150 ms so increase the debounce time to
1000 ms.
Signed-off-by: Angus Ainslie (Purism) <angus@akkea.ca>

The screen doesn't turn on at all with BACKLIGHT_LED built-in and
simple_ondemand devfreq patchset at the moment.

We don't need the status update unless it's plugged in.

With broken CC lines the cable sometimes shows as plugged in even when it
isn't. Look further into the status register to see if it is connected and
disable the STATUS interrupt to stop the irq storm.
Signed-off-by: Angus Ainslie <angus@akkea.ca>

The imx8m-ddrc driver can dynamically measure memory usage using ddr
pmu. Enable this by adding a "fsl,ddr-pmu" phandle reference.
Signed-off-by: Leonard Crestez <leonard.crestez@nxp.com>

The imx8m ddrc has a performance monitoring block attached which can
be used to measure bandwidth usage and automatically adjust frequency.

There is already a perf driver for the ddr pmu so instead of
implementing a devfreq-events driver use the in-kernel perf API to
implement get_dev_status directly.
Signed-off-by: Leonard Crestez <leonard.crestez@nxp.com>

Like other devfreq drivers imx8m-ddrc can measure memory bandwith usage.
This is only enabled if a link is provided to the imx8m ddr pmu.

This is referenced with a custom fsl,ddr-pmu property because
devfreq-events was deprecated.
Signed-off-by: Leonard Crestez <leonard.crestez@nxp.com>
Acked-by: Chanwoo Choi <cw00.choi@samsung.com>

Add a new public API to fetch a pointer to struct pmu from a devicetree
node devicetree node. This is meant to be used by other drivers to
create in-kernel counter for custom hardware PMUs with automatically
allocated pmu->type.

This is implementated by adding a new "parent_dev" field which pmu
drivers can optionally fill. This parent device is set as the parent of
the pmu->dev created on the pmu_bus. This also has the nice side-effect
of creating additional symlinks inside sysfs.

The actual per_get_pmu_by_node function is implemented through
bus_find_device and callers are asked to "put_device(pmu->dev)" when
done. This might cause problems if pmu_bus is initialize late but
consumers should handle errors by returning -EPROBE_DEFER anyway.
Signed-off-by: Leonard Crestez <leonard.crestez@nxp.com>