Add support for Colibri Vybrid Modules

This adds initial support for Colibri VF50/VF61 based on Freescale Vybrid SoC.

- CPU clocked at 396/500 MHz
- DDR3 at 396MHz
- for VF50, use PLL2 as memory clock (synchronous mode)
- for VF61, use PLL1 as memory clock (asynchronous mode)
- Console on UART0 (Colibri UART_A)
- Ethernet on FEC1
- PLL5 based RMII clocking (E.g. No external crystal)
- UART_A and UART_C I/O muxing
- Boot from NAND by default

Tested on Colibri VF50/VF61 booting using serial loader over UART.

Signed-off-by Sanchayan Maity <[email protected]>

Enable USB support on Toradex Colibri Vybrid Modules.

Signed-off-by Sanchayan Maity <[email protected]>

---
 board/toradex/colibri_vf/colibri_vf.c | 25 +++++++++++++++++++++--
 include/configs/colibri_vf.h          | 37

+++++++++++++++++++++++++++++++++++ 2 files changed, 60 insertions(+), 2 deletions(-)

(quoted diff) +0 -0

b/board/toradex/colibri_vf/colibri_vf.c index e7bc6c1..054e6f9 100644

Quoted 9 lines · Sanchayan Maity

Enable USB support on Toradex Colibri Vybrid Modules.

Signed-off-by Sanchayan Maity <[email protected]>

---
 board/toradex/colibri_vf/colibri_vf.c | 25 +++++++++++++++++++++--
 include/configs/colibri_vf.h          | 37

+++++++++++++++++++++++++++++++++++ 2 files changed, 60 insertions(+), 2 deletions(-)

board/toradex/colibri_vf/colibri_vf.c static void clock_init(void) · line 201

Quoted 11 lines

(quoted diff) +0 -0

b/board/toradex/colibri_vf/colibri_vf.c index e7bc6c1..054e6f9 100644

board/toradex/colibri_vf/colibri_vf.c +2 -1

@@ -197,7 +197,8 @@ static void clock_init(void)
 	clrsetbits_le32(&ccm->ccgr0, CCM_REG_CTRL_MASK,
 			CCM_CCGR0_UART0_CTRL_MASK);
 	clrsetbits_le32(&ccm->ccgr1, CCM_REG_CTRL_MASK,
-			CCM_CCGR1_PIT_CTRL_MASK | CCM_CCGR1_WDOGA5_CTRL_MASK);
+			CCM_CCGR1_PIT_CTRL_MASK | CCM_CCGR1_WDOGA5_CTRL_MASK |
+			CCM_CCGR1_USBC0_CTRL_MASK);

You probably want to flip these CCM regs only if CONFIG_CI_UDC is enabled or something along those lines, right ?

Best regards,
Marek Vasut

Hello,

Quoted 22 lines · Marek Vasut

Quoted 21 lines · Sanchayan Maity

Enable USB support on Toradex Colibri Vybrid Modules.

Signed-off-by Sanchayan Maity <[email protected]>

---

 board/toradex/colibri_vf/colibri_vf.c | 25 +++++++++++++++++++++--
 include/configs/colibri_vf.h          | 37

+++++++++++++++++++++++++++++++++++ 2 files changed, 60 insertions(+), 2 deletions(-)

(quoted diff) +0 -0

b/board/toradex/colibri_vf/colibri_vf.c index e7bc6c1..054e6f9 100644

board/toradex/colibri_vf/colibri_vf.c +0 -0

@@ -197,7 +197,8 @@ static void clock_init(void)
 
 	clrsetbits_le32(&ccm->ccgr0, CCM_REG_CTRL_MASK,
 
 			CCM_CCGR0_UART0_CTRL_MASK);
 

Quoted 8 lines

board/toradex/colibri_vf/colibri_vf.c static void clock_init(void) · line 201

Quoted 5 lines

board/toradex/colibri_vf/colibri_vf.c +2 -1

 	clrsetbits_le32(&ccm->ccgr1, CCM_REG_CTRL_MASK,
 
-			CCM_CCGR1_PIT_CTRL_MASK | CCM_CCGR1_WDOGA5_CTRL_MASK);
+			CCM_CCGR1_PIT_CTRL_MASK | CCM_CCGR1_WDOGA5_CTRL_MASK |
+			CCM_CCGR1_USBC0_CTRL_MASK);

You probably want to flip these CCM regs only if CONFIG_CI_UDC is enabled or something along those lines, right ?

Yes, that would be be better. Will send a new corrected patch for this. I assume since there are no issues with the rest of the patchset so sending a respin of this last patch would do?!

Thanks for the feedback and catching this.

Enable USB support on Toradex Colibri Vybrid Modules.

Signed-off-by Sanchayan Maity <[email protected]>

Hello,

This is the third version of the patchset which adds support for the Toradex Colibri Vybrid VF50 and VF61 modules. Boot up has been tested using the serial loader over UART. Compile tested for vf610twr_defconfig and vf610twr_nand_defconfig as well.

First patch in the series refactors the DDR related code for use by both the tower board and colibri modules. It also introduces a DDR3 based JEDEC timing structure.

Second third and fourth patch in this series are improvement patches related to RTC, SoC/CPU detection and caches.

Fifth patch introduces USB support for Vybrid modules. Much of the code is similar to the ehci-mx6 driver. Both host and client modes are working and DFU has also been tested with client. Currently, we restrict the ports to be in one of host and client mode.

Sixth patch adds the actual support for the Colibri modules.

Comments and feedback are most welcome. Thanks for the feedback till now.

The patchset is based and tested on the latest master branch as of this writing.

Discussion on the version 2 of the patchset can be found at the below link:
https://www.mail-archive.com/u-boot at lists.denx.de/msg168727.html

Discussion on the version 1 of the patchset can be found at the below link:
https://www.mail-archive.com/u-boot at lists.denx.de/msg168136.html

Changes since v2:
- Rework the USB driver to use register + offset method in light of discussion which Fabio Estevam pointed out instead of the regular struct{} method which v2 used. The discussion is at the below link:
https://www.marc.info/?l=u-boot&m=142609602127309&w=2

- Reorder the patchset, putting the USB support in the end and add an additional patch for adding USB support to Colibri modules. By chance if more discussions happen on the USB support, this allows picking up of atleast the first patches on which no issues have been reported so far.

- The register definitions have been moved under arch/arm/include/asm/ imx-common in the regs-usbphy.h file. This was agreed on after discussion with Marek and some input from Peter Chen. Since it is not clear if SoC's other than Freescale's use the Sigmatel Phy's which seem to be use in iMX/VF/MXS, put the USH PHY register definitions in imx-common rather than include/usb in a chipidea specific file.

- Remove setting of a PLL divisor select which was added for USB but is actually not required considering default value. It also seems to break USB after my latest rebase. The file in question concerning the change is colibri_vf.c. PLL divisor selects the PLL Multiplication factor which by default is 0, setting Fout = Fref * 20 giving 480MHz. The earlier patch set this to 1 giving Fout = Fref * 22 where Fref = 24MHz.

- Rebased on the latest master branch.

Changes since v1:
- Rework the USB driver to use register offsets using the regular struct {} method

- Some cleanups and fixes in the sixth patch for the colibri_vf.h file which takes care of environment variables in uboot

- Purge some useless defines in the fifth and sixth patch which were related to USB.

- Sanchayan.

Sanchayan Maity (5):
ARM: vf610: Move DDR3 initialization to imx-common
ARM: vf610: Add SoC and CPU type detection
ARM: vf610: Initial integration for Colibri VF50/VF61
usb: host: Add ehci-vf USB driver for ARM Vybrid SoC's
colibri_vf: Enable USB support for Colibri Vybrid

Stefan Agner (2):
ARM: vf610: Enable external 32KHz oscillator
ARM: vf610: Enable caches

 arch/arm/Kconfig                              |   5 +
 arch/arm/cpu/armv7/vf610/generic.c            |  45 ++-
 arch/arm/imx-common/Makefile                  |   3 +
 arch/arm/imx-common/ddrmc-vf610.c             | 278 +++++++++++++++++++
 arch/arm/include/asm/arch-vf610/crm_regs.h    |  11 +
 arch/arm/include/asm/arch-vf610/ddrmc-vf610.h |  72 +++++
 arch/arm/include/asm/arch-vf610/imx-regs.h    |  36 ++-
 arch/arm/include/asm/imx-common/regs-usbphy.h |  26 ++
 board/freescale/vf610twr/vf610twr.c           | 295 +++++---------------
 board/toradex/colibri_vf/Kconfig              |  18 ++
 board/toradex/colibri_vf/MAINTAINERS          |   6 +
 board/toradex/colibri_vf/Makefile             |   7 +
 board/toradex/colibri_vf/colibri_vf.c         | 382 ++++++++++++++++++++++++++
 board/toradex/colibri_vf/imximage.cfg         |  17 ++
 configs/colibri_vf_defconfig                  |   3 +
 drivers/usb/host/Makefile                     |   1 +
 drivers/usb/host/ehci-vf.c                    | 164 +++++++++++
 include/configs/colibri_vf.h                  | 271 ++++++++++++++++++
 18 files changed, 1405 insertions(+), 235 deletions(-)
 create mode 100644 arch/arm/imx-common/ddrmc-vf610.c
 create mode 100644 arch/arm/include/asm/arch-vf610/ddrmc-vf610.h
 create mode 100644 arch/arm/include/asm/imx-common/regs-usbphy.h
 create mode 100644 board/toradex/colibri_vf/Kconfig
 create mode 100644 board/toradex/colibri_vf/MAINTAINERS
 create mode 100644 board/toradex/colibri_vf/Makefile
 create mode 100644 board/toradex/colibri_vf/colibri_vf.c
 create mode 100644 board/toradex/colibri_vf/imximage.cfg
 create mode 100644 configs/colibri_vf_defconfig
 create mode 100644 drivers/usb/host/ehci-vf.c
 create mode 100644 include/configs/colibri_vf.h

Signature · 1 line

2.3.5

Hello,

Quoted 70 lines · Sanchayan Maity

Hello,

This is the third version of the patchset which adds support for the Toradex Colibri Vybrid VF50 and VF61 modules. Boot up has been tested using the serial loader over UART. Compile tested for vf610twr_defconfig and vf610twr_nand_defconfig as well.

First patch in the series refactors the DDR related code for use by both the tower board and colibri modules. It also introduces a DDR3 based JEDEC timing structure.

Second third and fourth patch in this series are improvement patches related to RTC, SoC/CPU detection and caches.

Fifth patch introduces USB support for Vybrid modules. Much of the code is similar to the ehci-mx6 driver. Both host and client modes are working and DFU has also been tested with client. Currently, we restrict the ports to be in one of host and client mode.

Sixth patch adds the actual support for the Colibri modules.

Comments and feedback are most welcome. Thanks for the feedback till now.

The patchset is based and tested on the latest master branch as of this writing.

Discussion on the version 2 of the patchset can be found at the below link:
https://www.mail-archive.com/u-boot at lists.denx.de/msg168727.html

Discussion on the version 1 of the patchset can be found at the below link:
https://www.mail-archive.com/u-boot at lists.denx.de/msg168136.html

Changes since v2:
- Rework the USB driver to use register + offset method in light of discussion which Fabio Estevam pointed out instead of the regular struct{} method which v2 used. The discussion is at the below link:
https://www.marc.info/?l=u-boot&m=142609602127309&w=2

- Reorder the patchset, putting the USB support in the end and add an additional patch for adding USB support to Colibri modules. By chance if more discussions happen on the USB support, this allows picking up of atleast the first patches on which no issues have been reported so far.

- The register definitions have been moved under arch/arm/include/asm/ imx-common in the regs-usbphy.h file. This was agreed on after discussion with Marek and some input from Peter Chen. Since it is not clear if SoC's other than Freescale's use the Sigmatel Phy's which seem to be use in iMX/VF/MXS, put the USH PHY register definitions in imx-common rather than include/usb in a chipidea specific file.

- Remove setting of a PLL divisor select which was added for USB but is actually not required considering default value. It also seems to break USB after my latest rebase. The file in question concerning the change is colibri_vf.c. PLL divisor selects the PLL Multiplication factor which by default is 0, setting Fout = Fref * 20 giving 480MHz. The earlier patch set this to 1 giving Fout = Fref * 22 where Fref = 24MHz.

- Rebased on the latest master branch.

Changes since v1:
- Rework the USB driver to use register offsets using the regular struct {} method

- Some cleanups and fixes in the sixth patch for the colibri_vf.h file which takes care of environment variables in uboot

Quoted 2 lines

- Purge some useless defines in the fifth and sixth patch which were related to USB.

Ping!?

Anything preventing this patch from getting applied?

Quoted 74 lines · maitysanchayan at gmail.com

Hello,

Quoted 71 lines · Sanchayan Maity

Hello,

This is the third version of the patchset which adds support for the Toradex Colibri Vybrid VF50 and VF61 modules. Boot up has been tested using the serial loader over UART. Compile tested for vf610twr_defconfig and vf610twr_nand_defconfig as well.

First patch in the series refactors the DDR related code for use by both the tower board and colibri modules. It also introduces a DDR3 based JEDEC timing structure.

Second third and fourth patch in this series are improvement patches related to RTC, SoC/CPU detection and caches.

Fifth patch introduces USB support for Vybrid modules. Much of the code is similar to the ehci-mx6 driver. Both host and client modes are working and DFU has also been tested with client. Currently, we restrict the ports to be in one of host and client mode.

Sixth patch adds the actual support for the Colibri modules.

Comments and feedback are most welcome. Thanks for the feedback till now.

The patchset is based and tested on the latest master branch as of this writing.

Discussion on the version 2 of the patchset can be found at the below link:
https://www.mail-archive.com/u-boot at lists.denx.de/msg168727.html

Discussion on the version 1 of the patchset can be found at the below link:
https://www.mail-archive.com/u-boot at lists.denx.de/msg168136.html

Changes since v2:
- Rework the USB driver to use register + offset method in light of discussion which Fabio Estevam pointed out instead of the regular struct{} method which v2 used. The discussion is at the below link:
https://www.marc.info/?l=u-boot&m=142609602127309&w=2

- Reorder the patchset, putting the USB support in the end and add an additional patch for adding USB support to Colibri modules. By chance if more discussions happen on the USB support, this allows picking up of atleast the first patches on which no issues have been reported so far.

- The register definitions have been moved under arch/arm/include/asm/ imx-common in the regs-usbphy.h file. This was agreed on after discussion with Marek and some input from Peter Chen. Since it is not clear if SoC's other than Freescale's use the Sigmatel Phy's which seem to be use in iMX/VF/MXS, put the USH PHY register definitions in imx-common rather than include/usb in a chipidea specific file.

- Remove setting of a PLL divisor select which was added for USB but is actually not required considering default value. It also seems to break USB after my latest rebase. The file in question concerning the change is colibri_vf.c. PLL divisor selects the PLL Multiplication factor which by default is 0, setting Fout = Fref * 20 giving 480MHz. The earlier patch set this to 1 giving Fout = Fref * 22 where Fref = 24MHz.

- Rebased on the latest master branch.

Changes since v1:
- Rework the USB driver to use register offsets using the regular struct {} method

- Some cleanups and fixes in the sixth patch for the colibri_vf.h file which takes care of environment variables in uboot

Quoted 6 lines

Quoted 2 lines

- Purge some useless defines in the fifth and sixth patch which were related to USB.

Ping!?

Anything preventing this patch from getting applied?

I'll pick this up soon, thanks!

In order to avoid code duplication, move the DDR3 initialization to the common place under imx-common. Currently ROW_DIFF and COL_DIFF can be chosen from the board file. The JEDEC timings are specified using a common ddr3_jedec_timings structure.

Signed-off-by Stefan Agner <[email protected]>

Signed-off-by Sanchayan Maity <[email protected]>

From: Stefan Agner <[email protected]>

Enable the SCSC (Slow Clock Source Controller) and select the external 32KHz oscillator. This improves the accuracy of the RTC.

Signed-off-by Sanchayan Maity <[email protected]>

Vybrid product family consists of several rather similar SoC which can be determined by softare during boot time. This allows use of variable ${soc} for Linux device tree files. Detect VF5xx CPU's by reading the CPU count register. We can determine the second number of the CPU type (VF6x0) which indicates the presence of a L2 cache.

Signed-off-by Stefan Agner <[email protected]>

Signed-off-by Sanchayan Maity <[email protected]>

From: Stefan Agner <[email protected]>

Enables caches which provides a rather huge speedup of the boot loader. Also mark the on-chip RAM as cachable since this is the area U-Boot runs from.

Signed-off-by Sanchayan Maity <[email protected]>

This adds initial support for Colibri VF50/VF61 based on Freescale Vybrid SoC.

- CPU clocked at 396/500 MHz
- DDR3 at 396MHz
- for VF50, use PLL2 as memory clock (synchronous mode)
- for VF61, use PLL1 as memory clock (asynchronous mode)
- Console on UART0 (Colibri UART_A)
- Ethernet on FEC1
- PLL5 based RMII clocking (E.g. No external crystal)
- UART_A and UART_C I/O muxing
- Boot from NAND by default

Tested on Colibri VF50/VF61 booting using serial loader over UART.

Signed-off-by Sanchayan Maity <[email protected]>

Acked-by Stefan Agner <[email protected]>

This driver adds support for the USB peripheral on Freescale Vybrid SoC's.

Signed-off-by Sanchayan Maity <[email protected]>

Enable USB support on Toradex Colibri Vybrid Modules.

Signed-off-by Sanchayan Maity <[email protected]>

Acked-by Marek Vasut <[email protected]>

Quoted 78 lines · Tom Rini

Quoted 77 lines · maitysanchayan at gmail.com

Hello,

Quoted 73 lines · Sanchayan Maity

Hello,

This is the third version of the patchset which adds support for the Toradex Colibri Vybrid VF50 and VF61 modules. Boot up has been tested using the serial loader over UART. Compile tested for vf610twr_defconfig and vf610twr_nand_defconfig as well.

First patch in the series refactors the DDR related code for use by both the tower board and colibri modules. It also introduces a DDR3 based JEDEC timing structure.

Second third and fourth patch in this series are improvement patches related to RTC, SoC/CPU detection and caches.

Fifth patch introduces USB support for Vybrid modules. Much of the code is similar to the ehci-mx6 driver. Both host and client modes are working and DFU has also been tested with client. Currently, we restrict the ports to be in one of host and client mode.

Sixth patch adds the actual support for the Colibri modules.

Comments and feedback are most welcome. Thanks for the feedback till now.

The patchset is based and tested on the latest master branch as of this writing.

Discussion on the version 2 of the patchset can be found at the below link:
https://www.mail-archive.com/u-boot at lists.denx.de/msg168727.html

Discussion on the version 1 of the patchset can be found at the below link:
https://www.mail-archive.com/u-boot at lists.denx.de/msg168136.html

Changes since v2:
- Rework the USB driver to use register + offset method in light of discussion which Fabio Estevam pointed out instead of the regular struct{} method which v2 used. The discussion is at the below link:
https://www.marc.info/?l=u-boot&m=142609602127309&w=2

- Reorder the patchset, putting the USB support in the end and add an additional patch for adding USB support to Colibri modules. By chance if more discussions happen on the USB support, this allows picking up of atleast the first patches on which no issues have been reported so far.

- The register definitions have been moved under arch/arm/include/asm/ imx-common in the regs-usbphy.h file. This was agreed on after discussion with Marek and some input from Peter Chen. Since it is not clear if SoC's other than Freescale's use the Sigmatel Phy's which seem to be use in iMX/VF/MXS, put the USH PHY register definitions in imx-common rather than include/usb in a chipidea specific file.

- Remove setting of a PLL divisor select which was added for USB but is actually not required considering default value. It also seems to break USB after my latest rebase. The file in question concerning the change is colibri_vf.c. PLL divisor selects the PLL Multiplication factor which by default is 0, setting Fout = Fref * 20 giving 480MHz. The earlier patch set this to 1 giving Fout = Fref * 22 where Fref = 24MHz.

- Rebased on the latest master branch.

Changes since v1:
- Rework the USB driver to use register offsets using the regular struct {} method

- Some cleanups and fixes in the sixth patch for the colibri_vf.h file which takes care of environment variables in uboot

- Purge some useless defines in the fifth and sixth patch which were related to USB.

Quoted 5 lines

Quoted 3 lines

Ping!?

Anything preventing this patch from getting applied?

I'll pick this up soon, thanks!

This should go through u-boot-imx though ;-)

Quoted 7 lines

Quoted 5 lines

Quoted 3 lines

Ping!?

Anything preventing this patch from getting applied?

I'll pick this up soon, thanks!

This should go through u-boot-imx though ;-)

For the record, since they aren't quite imx platforms I didn't want to throw another SoC on someone elses plate. Stefano, do you want to handle all the Vybrid stuf in the future? Thanks!

On Monday, April 27, 2015 at 07:54:15 PM, Tom Rini wrote:

[...]

Quoted 6 lines

Quoted 6 lines

Quoted 5 lines

Quoted 3 lines

Ping!?

Anything preventing this patch from getting applied?

I'll pick this up soon, thanks!

Quoted 5 lines

Quoted 1 line

This should go through u-boot-imx though ;-)

For the record, since they aren't quite imx platforms I didn't want to throw another SoC on someone elses plate. Stefano, do you want to handle all the Vybrid stuf in the future? Thanks!

How are those not imx platform ? They look imx inside out ... it's just the naming which is weird.

Quoted 5 lines

Quoted 3 lines

Quoted 1 line

Anything preventing this patch from getting applied?

I'll pick this up soon, thanks!

This should go through u-boot-imx though ;-)

For the record, since they aren't quite imx platforms I didn't want to throw another SoC on someone elses plate. Stefano, do you want to handle all the Vybrid stuf in the future? Thanks!

Quoted 12 lines · Marek Vasut

On Monday, April 27, 2015 at 07:54:15 PM, Tom Rini wrote:

[...]

Quoted 8 lines

Quoted 7 lines

Quoted 5 lines

Quoted 3 lines

Ping!?

Anything preventing this patch from getting applied?

I'll pick this up soon, thanks!

This should go through u-boot-imx though ;-)

Quoted 6 lines

Quoted 3 lines

For the record, since they aren't quite imx platforms I didn't want to throw another SoC on someone elses plate. Stefano, do you want to handle all the Vybrid stuf in the future? Thanks!

How are those not imx platform ? They look imx inside out ... it's just the naming which is weird.

I agree; this is very close to i.MX and can go through the imx tree.