Skip to contentSkip to conversation

spi: aspeed: SPI1 mode control driver enabling sysfs control

RFC In review Kun Yi 3 participants 9 messages Oct 22, 2016 – Oct 25, 2016

Subsystems (3)
lore
Share & tools
lore

Kun Yi

author
#0 lore raw

cover spi: aspeed: SPI1 mode control driver enabling sysfs control

Hello,

I'm looking for feedback on a simple Aspeed SPI1 mode control driver that does two things mainly:
1. Expose a sysfs interface for userland control of SPI1 interface.
2. Optionally override the SPI1 mode at probe if "start_mode" is configured in device tree.

Aspeed 24XX/25XX support three interface modes for their SPI1 interface, which can be configured via HW strap register. In the typical case of connecting a host flash chip SPI1 interface:
- SPI master mode makes BMC the master to the host SPI flash
- SPI pass-through mode make whichever external SPI interface connected to SYSSPI be the master
- There are also "disable" and "debug" modes. Aspeed 25XX specified the debug mode is reserved though.

In particular, SPI pass-through mode would enable an external SPI master to directly program host SPI chip and possibly serve as an alternative way to LPC for host CPU accessing SPI flash. I think by providing a sysfs handle, userspace applications can have a flexible way to mux host SPI flash ownership between BMC/external on the fly.

Overriding the SPI mode at probe time can be useful to make sure 2500-smc flash controller always identifies the flash chip and create /dev/mtd files correctly.

I wrote the driver as a standalone misc driver for the ease of initial implementation. There are other options such as merging the functionality into the SMC driver itself, possibly enabled by optional device tree arguments/nodes, but I just want to send the patches out first to gather comments on perceived usefulness of the feature and how you think the driver should be constructed.

Feedback greatly appreciated!

Regards,
Kun

author
#1 lore raw

RFC 1/2 spi: aspeed: ASPEED 24XX/25XX SPI1 disable/passthrough/master mode control

Simple driver that exposes userspace sysfs to control SPI1 mode. It also optionally override initial SPI mode if given "start_mode" parameter in the device tree.

---
 drivers/misc/Kconfig           |   6 ++
 drivers/misc/Makefile          |   1 +
 drivers/misc/aspeed-spi-mode.c | 191 +++++++++++++++++++++++++++++++++++++++++
 3 files changed, 198 insertions(+)
 create mode 100644 drivers/misc/aspeed-spi-mode.c
drivers/misc/Kconfig +6 -0
index 4617ddc..ea9b713 100644
@@ -809,6 +809,12 @@ config ASPEED_BT_IPMI_HOST
 	help
 	  Support for the Aspeed BT ipmi host.
 
+config SPI_MODECTRL_ASPEED
+        tristate "Support for SPI1 host flash mode change on Aspeed 24xx/25xx"
+        ---help---
+          Support for the Aspeed 24xx/25xx sysfs control toggle between
+	  disabled/master/pass-through modes.
+
 source "drivers/misc/c2port/Kconfig"
 source "drivers/misc/eeprom/Kconfig"
 source "drivers/misc/cb710/Kconfig"
drivers/misc/Makefile +1 -0
index 724861b..0cce3d6 100644
@@ -58,3 +58,4 @@ obj-$(CONFIG_VEXPRESS_SYSCFG)	+= vexpress-syscfg.o
 obj-$(CONFIG_CXL_BASE)		+= cxl/
 obj-$(CONFIG_PANEL)             += panel.o
 obj-$(CONFIG_ASPEED_BT_IPMI_HOST)	+= bt-host.o
+obj-$(CONFIG_SPI_MODECTRL_ASPEED)	+= aspeed-spi-mode.o
drivers/misc/aspeed-spi-mode.c new file +191 -0

191 lines not on this page · raw message

Signature · 1 line
2.8.0.rc3.226.g39d4020
author
#2 lore raw

RFC 2/2 spi: aspeed: documentation and device tree bindings for ASPEED SPI1 mode controller

In reply to Kun Yi

There is no accurate HW representation for the driver at the moment, thus it is represented as a "virtual" platform device node. The driver is placed in misc for the same reason.

---
 .../devicetree/bindings/misc/aspeed-spi-mode.txt   | 29 ++++++++++++++++++++++
 arch/arm/boot/dts/aspeed-ast2500-evb.dts           | 10 ++++++++
 2 files changed, 39 insertions(+)
 create mode 100644 Documentation/devicetree/bindings/misc/aspeed-spi-mode.txt
Documentation/devicetree/bindings/misc/aspeed-spi-mode.txt new file +29 -0
new file mode 100644
index 0000000..540b101
@@ -0,0 +1,29 @@
+Device tree entry for SPI1 mode controller on AST24XX/AST25XX.
+
+This binding enables the SPI1 mode controller virtual driver that exposes
+the SPI1 interface toggle capability to userspace through sysfs interface. If
+optional argument "start_mode" is defined, it also configures the SPI mode.
+
+Required Properties:
+- #address-cells	: should be 1
+- #size-cells		: should be 1
+- compatible		: should be "aspeed, aspeed_spi_mode"
+- reg			: addresses and sizes of HW strap register and HW strap
+			  reset register. Should be < 0x1e6e2070 0x4 > and
+			  < 0x1e6e207c 0x4 > respectively.
+
+Optional Properties:
+- start_mode		: spi mode to override HW strap setting during probe.
+			  Can be 0(disabled) 1(master) or 3(pass-through).
+			  If not defined, HW strap or pinctrl setting is kept.
+
+Example:
+	spi_mode_ctrl: spi_mode {
+		#address-cells = <1>;
+		#size-cells = <1>;
+		compatible = "aspeed,aspeed_spi_mode";
+		start_mode = <1>;
+		status = "okay";
+		reg = < 0x1e6e2070 0x4
+			0x1e6e207c 0x4 >;
+	};
arch/arm/boot/dts/aspeed-ast2500-evb.dts +10 -0
index 33b3907..f50812a 100644
@@ -18,6 +18,16 @@
 	memory {
 		reg = <0x80000000 0x20000000>;
 	};
+
+	spi_mode_ctrl: spi_mode {
+		#address-cells = <1>;
+		#size-cells = <1>;
+		compatible = "aspeed,aspeed_spi_mode";
+		start_mode = <1>;
+		status = "okay";
+		reg = < 0x1e6e2070 0x4
+			0x1e6e207c 0x4 >;
+	};
 };
 
 &fmc {
Signature · 1 line
2.8.0.rc3.226.g39d4020
maintainer
#3 lore raw
In reply to Kun Yi

Hell Kun,

Quoted 6 lines · Kun Yi

Hello,

I'm looking for feedback on a simple Aspeed SPI1 mode control driver that does two things mainly:
1. Expose a sysfs interface for userland control of SPI1 interface.
2. Optionally override the SPI1 mode at probe if "start_mode" is configured in device tree.

Why do you plan to override the mode at boot time? Is this for firmware updates?

We plan to move away from having the NOR flash accessible from both the host and the BMC for all platforms. There's no good way to indicate ownership, and the Linux mtd subsystem does not have a good way of relinquishing nor reestablishing control. For instance, there's no way to "unmount" the MTD device, and there's no way to force a re-scan of partitions if you modify the layout.

I think there are some issues with your approach that break pinctrl/pinmux that Andrew will highlight.

Cheers,

Joel

Quoted 34 lines

Aspeed 24XX/25XX support three interface modes for their SPI1 interface, which can be configured via HW strap register. In the typical case of connecting a host flash chip SPI1 interface:
- SPI master mode makes BMC the master to the host SPI flash
- SPI pass-through mode make whichever external SPI interface connected to SYSSPI be the master
- There are also "disable" and "debug" modes. Aspeed 25XX specified the debug mode is reserved though.

In particular, SPI pass-through mode would enable an external SPI master to directly program host SPI chip and possibly serve as an alternative way to LPC for host CPU accessing SPI flash. I think by providing a sysfs handle, userspace applications can have a flexible way to mux host SPI flash ownership between BMC/external on the fly.

Overriding the SPI mode at probe time can be useful to make sure 2500-smc flash controller always identifies the flash chip and create /dev/mtd files correctly.

I wrote the driver as a standalone misc driver for the ease of initial implementation. There are other options such as merging the functionality into the SMC driver itself, possibly enabled by optional device tree arguments/nodes, but I just want to send the patches out first to gather comments on perceived usefulness of the feature and how you think the driver should be constructed.

Feedback greatly appreciated!

Regards,
Kun

_______________________________________________
openbmc mailing list
[email protected]
https://lists.ozlabs.org/listinfo/openbmc

author
#4 lore raw

Hi Joel, thanks for commenting!

Quoted 6 lines · Joel Stanley

Hell Kun,

Quoted 3 lines · Kun Yi

Hello,

I'm looking for feedback on a simple Aspeed SPI1 mode control driver

Quoted 9 lines

that does two things mainly:

Quoted 2 lines

1. Expose a sysfs interface for userland control of SPI1 interface.
2. Optionally override the SPI1 mode at probe if "start_mode" is

configured in

Quoted 1 line

device tree.

Why do you plan to override the mode at boot time? Is this for firmware updates?

More important here is the ability to change the SPI mode at *run time*. One use case is for platforms without eSPI/LPC support. When host CPU is booting and BIOS needs to read SPI chip, the most straightforward way for BMC to allow host access to the SPI chip is to toggle to SPI pass-through mode. Thus the core purpose of the driver is for enabling SPI pass-through mode on demand.

Overriding the SPI mode at boot time is only to make sure the host SPI chip can be recognized correctly by the SMC driver. Because currently pinctrl treats HW strap as read-only, mis-configured HW strap pins will cause SMC driver to fail reading SPI flash ID and not register it in MTD subsystem, and there is no easy way to probe again later. Though enabling pinctrl to set HW strap register might be a better approach as Andrew mentioned.

Quoted 7 lines

We plan to move away from having the NOR flash accessible from both the host and the BMC for all platforms. There's no good way to indicate ownership, and the Linux mtd subsystem does not have a good way of relinquishing nor reestablishing control. For instance, there's no way to "unmount" the MTD device, and there's no way to force a re-scan of partitions if you modify the layout.

I saw some discussions related on IRC but wasn't clear about the details. Could you elaborate a little? Do you plan to only allow BMC daemons to access both host/BMC NOR flashes?

I got the impression that mtd will stick around for now. Do we have a time frame for deprecating that?

Quoted 3 lines

I think there are some issues with your approach that break pinctrl/pinmux that Andrew will highlight.

Sure I will reply to those in Andrew's email.

Quoted 52 lines

Cheers,

Joel

Quoted 3 lines

Aspeed 24XX/25XX support three interface modes for their SPI1 interface, which can be configured via HW strap register. In the typical case of

connecting

Quoted 3 lines

a host flash chip SPI1 interface:
- SPI master mode makes BMC the master to the host SPI flash
- SPI pass-through mode make whichever external SPI interface connected

to SYSSPI

Quoted 2 lines

be the master
- There are also "disable" and "debug" modes. Aspeed 25XX specified the

debug

Quoted 3 lines

mode is reserved though.

In particular, SPI pass-through mode would enable an external SPI master

to

Quoted 1 line

directly program host SPI chip and possibly serve as an alternative way

to LPC

Quoted 1 line

for host CPU accessing SPI flash. I think by providing a sysfs handle,

userspace

Quoted 1 line

applications can have a flexible way to mux host SPI flash ownership

between

Quoted 3 lines

BMC/external on the fly.

Overriding the SPI mode at probe time can be useful to make sure

2500-smc flash

Quoted 1 line

controller always identifies the flash chip and create /dev/mtd files

correctly.

Quoted 2 lines

I wrote the driver as a standalone misc driver for the ease of initial

implementation.

Quoted 1 line

There are other options such as merging the functionality into the SMC

driver

Quoted 1 line

itself, possibly enabled by optional device tree arguments/nodes, but I

just want to

Quoted 12 lines

send the patches out first to gather comments on perceived usefulness of the feature and how you think the driver should be constructed.

Feedback greatly appreciated!

Regards,
Kun

_______________________________________________
openbmc mailing list
[email protected]
https://lists.ozlabs.org/listinfo/openbmc

Signature · 2 lines
Regards,
Kun
#5 lore raw
In reply to Kun Yi

Hi Kun,

Quoted 22 lines · Kun Yi

Hello,

I'm looking for feedback on a simple Aspeed SPI1 mode control driver that does two things mainly:
1. Expose a sysfs interface for userland control of SPI1 interface.
2. Optionally override the SPI1 mode at probe if "start_mode" is configured in device tree.

Aspeed 24XX/25XX support three interface modes for their SPI1 interface, which can be configured via HW strap register. In the typical case of connecting a host flash chip SPI1 interface:
- SPI master mode makes BMC the master to the host SPI flash
- SPI pass-through mode make whichever external SPI interface connected to SYSSPI be the master
- There are also "disable" and "debug" modes. Aspeed 25XX specified the debug mode is reserved though.

In particular, SPI pass-through mode would enable an external SPI master to directly program host SPI chip and possibly serve as an alternative way to LPC for host CPU accessing SPI flash. I think by providing a sysfs handle, userspace applications can have a flexible way to mux host SPI flash ownership between BMC/external on the fly.

Quoted 8 lines

Overriding the SPI mode at probe time can be useful to make sure 2500-smc flash controller always identifies the flash chip and create /dev/mtd files correctly.

I wrote the driver as a standalone misc driver for the ease of initial implementation. There are other options such as merging the functionality into the SMC driver itself, possibly enabled by optional device tree arguments/nodes, but I just want to send the patches out first to gather comments on perceived usefulness of the feature and how you think the driver should be constructed.

At least one issue I have with the implementation is that the strap register is currently "owned" by the pinctrl driver, but the patch as it stands has no interaction with this subsystem. Arguably the pinctrl driver is where this should be implemented, as it encapsulates pinmux whose aim is to manage the functionality made available on the SoC's pins.

Currently the pinctrl driver treats the strapping register as read- only. That should probably change in light of what you're trying to do here, though only to the extent of allowing the SPI mode switch.

Cheers,

Andrew

author
#6 lore raw

Hi Andrew,

One question I have though: I think pinmux/pinctrl driver is supposed to be a boot time static pin configure module and parses DT and configures accordingly. Is it possible to use pinmux to dynamically modify the pin states? This is essentially the problem I'm trying to deal with.

Thanks,
Kun

Quoted 62 lines · Andrew Jeffery

Hi Kun,

Quoted 3 lines · Kun Yi

Hello,

I'm looking for feedback on a simple Aspeed SPI1 mode control driver

that does two things mainly:

Quoted 2 lines

1. Expose a sysfs interface for userland control of SPI1 interface.
2. Optionally override the SPI1 mode at probe if "start_mode" is

configured in

Quoted 4 lines

device tree.

Aspeed 24XX/25XX support three interface modes for their SPI1 interface, which can be configured via HW strap register. In the typical case of

connecting

Quoted 3 lines

a host flash chip SPI1 interface:
- SPI master mode makes BMC the master to the host SPI flash
- SPI pass-through mode make whichever external SPI interface connected

to SYSSPI

Quoted 2 lines

be the master
- There are also "disable" and "debug" modes. Aspeed 25XX specified the

debug

Quoted 3 lines

mode is reserved though.

In particular, SPI pass-through mode would enable an external SPI master

to

Quoted 1 line

directly program host SPI chip and possibly serve as an alternative way

to LPC

Quoted 1 line

for host CPU accessing SPI flash. I think by providing a sysfs handle,

userspace

Quoted 1 line

applications can have a flexible way to mux host SPI flash ownership

between

Quoted 3 lines

BMC/external on the fly.

Overriding the SPI mode at probe time can be useful to make sure

2500-smc flash

Quoted 1 line

controller always identifies the flash chip and create /dev/mtd files

correctly.

Quoted 2 lines

I wrote the driver as a standalone misc driver for the ease of initial

implementation.

Quoted 1 line

There are other options such as merging the functionality into the SMC

driver

Quoted 1 line

itself, possibly enabled by optional device tree arguments/nodes, but I

just want to

Quoted 2 lines

send the patches out first to gather comments on perceived usefulness of the feature and how you think the driver should be constructed.

At least one issue I have with the implementation is that the strap register is currently "owned" by the pinctrl driver, but the patch as it stands has no interaction with this subsystem. Arguably the pinctrl driver is where this should be implemented, as it encapsulates pinmux whose aim is to manage the functionality made available on the SoC's pins.

Currently the pinctrl driver treats the strapping register as read- only. That should probably change in light of what you're trying to do here, though only to the extent of allowing the SPI mode switch.

Cheers,

Andrew

Signature · 2 lines
Regards,
Kun
#7 lore raw
Quoted 6 lines · Kun Yi

Hi Andrew,

One question I have though: I think pinmux/pinctrl driver is supposed to be a boot time static pin configure module and parses DT and configures accordingly. Is it possible to use pinmux to dynamically modify the pin states?

Yes, see "Runtime pinmuxing" here (at the bottom of the document):

https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/tree/Documentation/pinctrl.txt

So you'd need to patch the FMC driver that Milton and Cedric have been working on to (de)activate pin states, likely at probe() and remove() (this implies you'll need to somehow orchestrate a remove/switch, switch/probe dance). Alternatively, provide a way to suspend operations on the MTD and switch the pins between master and passthru in the FMC driver. Some complications are that this functionality is only available for SPI1, and you might need to do a bit of work to guarantee sensible outcomes for operations that might span the mode switch.

Cheers,

Andrew

author
#8 lore raw

Thanks Andrew that was helpful! I will look into pinctrl and FMC driver then.

Thanks,
Kun

Quoted 25 lines · Andrew Jeffery
Quoted 6 lines · Kun Yi

Hi Andrew,

One question I have though: I think pinmux/pinctrl driver is supposed to be a boot time static pin configure module and parses DT and configures accordingly. Is it possible to use pinmux to dynamically modify the pin states?

Yes, see "Runtime pinmuxing" here (at the bottom of the document):

https://git.kernel.org/cgit/linux/kernel/git/torvalds/
linux.git/tree/Documentation/pinctrl.txt

So you'd need to patch the FMC driver that Milton and Cedric have been working on to (de)activate pin states, likely at probe() and remove() (this implies you'll need to somehow orchestrate a remove/switch, switch/probe dance). Alternatively, provide a way to suspend operations on the MTD and switch the pins between master and passthru in the FMC driver. Some complications are that this functionality is only available for SPI1, and you might need to do a bit of work to guarantee sensible outcomes for operations that might span the mode switch.

Cheers,

Andrew

Signature · 2 lines
Regards,
Kun