[Letux-kernel] [PATCH] arm: boot: dts: lc15: hardcode txs02612 mux to eMMC

Nick Elsmore nicholaselsmore at gmail.com
Mon Aug 10 21:34:33 CEST 2020


There are some pretty serious issues regarding the current hardware
strategy of the txs02612 mmc mux and compatibility with the Linux mmc
stack.  I would first like to propose disabling its functionality with the
following patch and supporting just the eMMC until a viable method of
handling it can be found.  The reasoning is as follows.

The mmc subsystem does not directly support multiplexing like other
subsystems do (ie i2c).  The current strategy involves intercepting mmc
request operations to the cards on the downstream side of the mux and
switching the mux before calling the parent host's request op.

Ideally the tx02612 device node would be a subnode of the root mmc host
node.  However, the mmc host driver does not directly probe child nodes
without first performing a handshake on the bus.  Therefore the subdriver
would only be probed if a valid device was connected in the default state
of the mux at the time the root host is probed.  Therefore, in order to
guarantee the txs02612 driver is probed, the bootloader and kernel would
need to ensure that a valid device is connected on the SDIO lines to ensure
the mux driver came up.  This should work as the tx02612 driver would be
registered as a card on the root host and the tx02612 driver would allocate
and register two new hosts for the eMMC and uSD which would properly go
through the SDIO/SD/MMC initialization procedures when they are
registered.  There also may be an issue if the default card used for
detection goes through the full probe procedure and is registered on the
root host AND the new host.  However, since the driver even being probed in
the first place relies on the default state of the GPIO set by the
bootloader, this implementation seems shaky to me.

Another option is to register the tx02612 driver as a platform driver and
link in the device_node in the dts in a method similar to the one below.
This way we can be sure the driver is always probed, regardless of the mmc
handshake.

txs02612 {
     compatible = "ti,txs02612";
     select-gpio = <&gpio3 18 0>;    /* gpio3_82 soft-select */
     host = <&mmc2>;
     status = "disabled";
 };

With this sort of link, the mmc device_node can be grabbed using
of_parse_phandle.  However, getting the device and then mmc_host from this
device_node is not easy since the mmc subsystem doesn't currently provide a
method of accessing host nodes from their of_node by iterating through the
class devices like some other subsystems provide (ie gpio, fpga-mgr).  So
far, I have not found a viable method to access the linked node as we don't
have access to the correct resources within the mmc subsystem.  This method
also poses the same issue exhibted by the first method, where the card may
be registered on both the root host and the newly registered host.

The current implementation allows for manual switching between the eMMC and
uSD from userspace using a sysfs hook.  This works, but I worry about its
behavior when interfacing with the mmc subsystem.  The subsystem is not
aware of the switching and so this method will likely pose issues with
regard to filesystem cache sync and similar.  The host is using polling to
handle card detect, and so if the card disappears due to unsynchronized
userland switching, it may view this as a card disconnect, even though the
card is still powered.  Because of this, a number of risks are possible
including improper handling of the card itself (potentially causing
damage), improper or missed filesystem cache synchronization, and data
corruption.

Until I or someone else can find a safe and (hopefully) graceful solution
to this issue, I am proposing to simply disable the tx02612 driver and
hardcode the mux to the eMMC for now with the following patch.

Signed-off-by: Nick Elsmore <nicholaselsmore at gmail.com>
---

index d59d9fd4c8af..60822100bcdf 100644
--- a/arch/arm/boot/dts/omap5-letux-cortex15-common.dtsi
+++ b/arch/arm/boot/dts/omap5-letux-cortex15-common.dtsi
@@ -248,6 +248,7 @@ mmcmux: txs02612 at 0 {

 		#adress-cells = <1>;
 		#size-cells = <0>;
+		status = "disabled";

 		eMMC: port at 0 {
 			reg = <0>;
@@ -268,9 +269,25 @@ / { /* eMMC / uSD-card */
 	txs02612 {
 		compatible = "ti,txs02612";
 		select-gpio = <&gpio3 18 0>;	/* gpio3_82 soft-select */
+		status = "disabled";
 	};
 };

+/* hardcode txs02612 mmc mux to eMMC input for now */
+&gpio3 {
+	txs02612_soft_control {
+		gpio-hog;
+		gpios = <18 GPIO_ACTIVE_HIGH>;
+		output-low;
+		label = "txs02612-soft-control";
+	};
+};
+
+&mmc3 { /* WiFi-SDIO */
+ /* has been predefined for a wl1271 in ompa5-common.dtsi */
+};
+
+
 &mmc3 { /* WiFi-SDIO */
 	/* has been predefined for a wl1271 in ompa5-common.dtsi */
 };
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.goldelico.com/pipermail/letux-kernel/attachments/20200810/01fdab64/attachment.htm>


More information about the Letux-kernel mailing list