net: ti: icssg: Add prueth support
Subsystems (6)
cover net: ti: icssg: Add prueth support
The series adds support for icssg_prueth functionality on u-boot. This series is based on top of master branch. rproc init needs to be done from uboot command prompt. The pru/rtu firmware loading is done by prueth driver soon after config paramters are setup.
Currently only slice0/1 of icssg2 instance on am6-evm is supported. i.e Both slices of icssg2 instance are supported.
On u-boot prompt following commands to test icssg2_port 0 on am654-evm:
setenv ethact pruss2_eth; setenv serverip 172.24.191.45; fatload mmc 1 ${pru0loadaddr} am65x-pru0-prueth-fw.elf; fatload mmc 1 ${rtu0loadaddr} am65x-rtu0-prueth-fw.elf; rproc init; setenv autoload no; dhcp; tftp 0x82000000 Image; tftp 0x83000000 k3-am654-base-board.dtb; booti 0x82000000 - 0x83000000
This tests tftp on prueth.
Note: Uboot ethernet driver architecture supports once instance per probe. So only one of the ports are supported per instance. So DT of prueth node should have either ethernet-mii0 or ethernet-mii1.
Keerthy (14):
net: eth-uclass: eth_get_dev based on SEQ_ALIAS instead of probe order
net: eth-uclass: call stop only for active devices
misc: uclass: Introduce misc_init_by_ofnode
soc: ti: pruss: add a misc driver for PRUSS in TI SoCs
remoteproc: pruss: add PRU remoteproc driver
net: ti: icssg-prueth: Add ICSSG ethernet driver
net: ti: icssg-prueth: Workaround to shutdown the prueth firmware
arm: dts: k3-am65-main: Add msmc_ram node
arm: dts: k3-am654-base-board-u-boot: Add icssg specific msmc_ram
carveout nodes
arm: dts: k3-am65-main: Add scm_conf node
arm: dts: k3-am65-main: Add pruss nodes for ICSSG2
arm64: dts: ti: am654-base-board: add ICSSG2 Ethernet support
configs: am65x_evm_a53_defconfig: Enable CONFIG_REMOTEPROC_TI_PRU
configs: am65x_evm_a53_defconfig: Enable CONFIG_CMD_REMOTEPROC
arch/arm/dts/k3-am65-main.dtsi | 212 ++++++++ arch/arm/dts/k3-am65.dtsi | 4 +- arch/arm/dts/k3-am654-base-board-u-boot.dtsi | 130 +++++ configs/am65x_evm_a53_defconfig | 5 + drivers/misc/misc-uclass.c | 25 + drivers/net/ti/Kconfig | 8 + drivers/net/ti/Makefile | 1 + drivers/net/ti/icssg-prueth.c | 525 +++++++++++++++++++ drivers/net/ti/icssg.h | 31 ++ drivers/net/ti/icssg_classifier.c | 397 ++++++++++++++ drivers/remoteproc/Kconfig | 11 + drivers/remoteproc/Makefile | 1 + drivers/remoteproc/pru_rproc.c | 384 ++++++++++++++ drivers/soc/ti/Kconfig | 13 + drivers/soc/ti/Makefile | 1 + drivers/soc/ti/pruss.c | 143 +++++ include/misc.h | 9 + include/ti-pruss.h | 13 + net/eth-uclass.c | 7 +- 19 files changed, 1916 insertions(+), 4 deletions(-) create mode 100644 drivers/net/ti/icssg-prueth.c create mode 100644 drivers/net/ti/icssg.h create mode 100644 drivers/net/ti/icssg_classifier.c create mode 100644 drivers/remoteproc/pru_rproc.c create mode 100644 drivers/soc/ti/pruss.c create mode 100644 include/ti-pruss.h
Signature · 1 line
2.17.1
PATCH 1/14 net: eth-uclass: eth_get_dev based on SEQ_ALIAS instead of probe order
In case of multiple eth interfaces currently eth_get_dev fetches the device based on the probe order which can be random hence try with the alias.
--- net/eth-uclass.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)
net/eth-uclass.c +2 -2
@@ -68,8 +68,8 @@ struct udevice *eth_get_dev(void) uc_priv =eth_get_uclass_priv ();if (!uc_priv->current) - eth_errno =uclass_first_device (UCLASS_ETH , - &uc_priv->current); + eth_errno =uclass_get_device_by_seq (UCLASS_ETH , +0 , &uc_priv->current);return uc_priv->current; }
Signature · 1 line
2.17.1
PATCH 1/14 net: eth-uclass: eth_get_dev based on SEQ_ALIAS instead of probe order
Quoted 14 lines · Keerthy
In case of multiple eth interfaces currently
eth_get_devfetches the device based on the probe order which can be random hence try with the alias.Signed-off-by Keerthy <[email protected]>--- net/eth-uclass.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)net/eth-uclass.c +0
-0
net/eth-uclass.c struct udevice *eth_get_dev(void) · line 74
Quoted 9 lines
net/eth-uclass.c +2
-2@@ -68,8 +68,8 @@ struct udevice *eth_get_dev(void) uc_priv =eth_get_uclass_priv ();if (!uc_priv->current) - eth_errno =uclass_first_device (UCLASS_ETH , - &uc_priv->current); + eth_errno =uclass_get_device_by_seq (UCLASS_ETH , +0 , &uc_priv->current);return uc_priv->current; }
This breaks networking on qemu_arm / qemu_arm64 and no networking device is found.
Signature · 8 lines
Tom -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 833 bytes Desc: not available URL: <http://lists.denx.de/pipermail/u-boot/attachments/20191025/0c72e6cc/attachment.sig>
PATCH 1/14 net: eth-uclass: eth_get_dev based on SEQ_ALIAS instead of probe order
Quoted 17 lines · Tom Rini
Quoted 15 lines · Keerthy
In case of multiple eth interfaces currently
eth_get_devfetches the device based on the probe order which can be random hence try with the alias.Signed-off-by Keerthy <[email protected]>--- net/eth-uclass.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)net/eth-uclass.c +0
-0@@ -68,8 +68,8 @@ struct udevice *eth_get_dev(void)
Quoted 11 lines
net/eth-uclass.c struct udevice *eth_get_dev(void) · line 74
Quoted 8 lines
net/eth-uclass.c +2
-2uc_priv =eth_get_uclass_priv ();if (!uc_priv->current) - eth_errno =uclass_first_device (UCLASS_ETH , - &uc_priv->current); + eth_errno =uclass_get_device_by_seq (UCLASS_ETH , +0 , &uc_priv->current);return uc_priv->current; }This breaks networking on
qemu_arm/qemu_arm64and no networking device is found.
Okay. That is pretty strange. So seems like uclass_get_device_by_seq is failing for some reason.
Quoted 1 line
PATCH 1/14 net: eth-uclass: eth_get_dev based on SEQ_ALIAS instead of probe order
Quoted 31 lines · keerthy
Quoted 11 lines · Tom Rini
Quoted 9 lines · Keerthy
In case of multiple eth interfaces currently
eth_get_devfetches the device based on the probe order which can be random hence try with the alias.Signed-off-by Keerthy <[email protected]>---
net/eth-uclass.c| 4 ++--1 file changed, 2 insertions(+), 2 deletions(-)Quoted 16 lines
net/eth-uclass.c struct udevice *eth_get_dev(void) · line 73
Quoted 13 lines
net/eth-uclass.c +2
-2@@ -68,8 +68,8 @@ struct udevice *eth_get_dev(void) uc_priv = eth_get_uclass_priv(); if (!uc_priv->current) - eth_errno = uclass_first_device(UCLASS_ETH, - &uc_priv->current); + eth_errno = uclass_get_device_by_seq(UCLASS_ETH, + 0, &uc_priv->current); return uc_priv->current; }This breaks networking on
qemu_arm/qemu_arm64and no networking device is found.
Quoted 2 lines
Okay. That is pretty strange. So seems like
uclass_get_device_by_seqis failing for some reason.
because not everyone populates aliases. You can do something like below:
eth_errno = uclass_get_device_by_seq(UCLASS_ETH, 0, &uc_priv->current); if (eth_errno || !uc_priv->current) eth_errno = uclass_first_device(UCLASS_ETH, &uc_priv->current);
Thanks and regards,
Lokesh
Quoted 2 lines
Quoted 2 lines
PATCH 1/14 net: eth-uclass: eth_get_dev based on SEQ_ALIAS instead of probe order
Quoted 34 lines · Lokesh Vutla
Quoted 31 lines · keerthy
Quoted 27 lines · Tom Rini
Quoted 14 lines · Keerthy
In case of multiple eth interfaces currently
eth_get_devfetches the device based on the probe order which can be random hence try with the alias.Signed-off-by Keerthy <[email protected]>--- net/eth-uclass.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)net/eth-uclass.c +0
-0net/eth-uclass.c struct udevice *eth_get_dev(void) · line 73
Quoted 8 lines
net/eth-uclass.c +2
-2@@ -68,8 +68,8 @@ struct udevice *eth_get_dev(void) uc_priv = eth_get_uclass_priv(); if (!uc_priv->current) - eth_errno = uclass_first_device(UCLASS_ETH, - &uc_priv->current); + eth_errno = uclass_get_device_by_seq(UCLASS_ETH, + 0, &uc_priv->current); return uc_priv->current; }This breaks networking on
qemu_arm/qemu_arm64and no networking device is found.
Quoted 8 lines
Quoted 2 lines
Okay. That is pretty strange. So seems like
uclass_get_device_by_seqis failing for some reason.because not everyone populates aliases. You can do something like below:
eth_errno = uclass_get_device_by_seq(UCLASS_ETH, 0, &uc_priv->current); if (eth_errno || !uc_priv->current) eth_errno = uclass_first_device(UCLASS_ETH, &uc_priv->current);
Thanks Lokesh. I will add that check.
Quoted 7 lines
Thanks and regards,
LokeshQuoted 2 lines
Quoted 2 lines
PATCH 2/14 net: eth-uclass: call stop only for active devices
In reply to KeerthyCurrently stop is being called unconditionally without even checking if start is called which will result in crash where multiple instances are present and stop gets called even without calling start.
--- net/eth-uclass.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)
net/eth-uclass.c +2 -1
@@ -557,7 +557,8 @@ static int eth_pre_remove(struct udevice *dev) {struct eth_pdata *pdata = dev->platdata; -eth_get_ops (dev)->stop (dev); +if (eth_is_active (dev)) +eth_get_ops (dev)->stop (dev);/* clear the MAC address */ memset (pdata->enetaddr,0 ,ARP_HLEN );
Signature · 1 line
2.17.1
PATCH 3/14 misc: uclass: Introduce misc_init_by_ofnode
In reply to KeerthyIntroduce misc_init_by_ofnode to probe a misc device using its ofnode.
--- drivers/misc/misc-uclass.c | 25 +++++++++++++++++++++++++ include/misc.h | 9 +++++++++ 2 files changed, 34 insertions(+)
drivers/misc/misc-uclass.c +25 -0
@@ -5,6 +5,8 @@#include <common.h> #include <dm.h> +#include <dm/device-internal.h> +#include <dm/uclass-internal.h> #include <errno.h> #include <misc.h>
@@ -65,6 +67,29 @@ int misc_set_enabled(struct udevice *dev, bool val)return ops->set_enabled (dev, val); } +int misc_init_by_ofnode (ofnode node) +{ +struct udevice *dev =NULL ; +int ret; +long temp1, temp2; + + temp1 =ofnode_to_offset (node); + +for (ret =uclass_find_first_device (UCLASS_MISC , &dev); dev; + ret =uclass_find_next_device (&dev)) { + temp2 =ofnode_to_offset (dev_ofnode (dev)); +if (temp1 == temp2) { + ret =device_probe (dev); +if (ret) +debug ("%s: Failed to initialize - %d\n" , + dev->name, ret); +return ret; + } + } + +return -ENODEV ; +} +UCLASS_DRIVER (misc) = { .id =UCLASS_MISC , .name ="misc" ,
include/misc.h +9 -0
@@ -76,6 +76,15 @@ int misc_call(struct udevice *dev, int msgid, void *tx_msg, int tx_size,*/ int misc_set_enabled (struct udevice *dev,bool val); +/** +* misc_init_by_ofnode() - Probe a misc device by using ofnode. +* @node: ofnode of the misc device. +* +* A misc device is probed using ofnode. +* +* Return: -ve on error, 0 on success +*/ +int misc_init_by_ofnode (ofnode node);/* * struct misc_ops - Driver model Misc operations *
Signature · 1 line
2.17.1
PATCH 3/14 misc: uclass: Introduce misc_init_by_ofnode
Quoted 31 lines · Keerthy
Introduce
misc_init_by_ofnodeto probe a misc device using its ofnode.Signed-off-by Keerthy <[email protected]>--- drivers/misc/misc-uclass.c | 25 +++++++++++++++++++++++++ include/misc.h | 9 +++++++++ 2 files changed, 34 insertions(+)drivers/misc/misc-uclass.c +7
-0@@ -5,6 +5,8 @@#include <common.h> #include <dm.h> +#include <dm/device-internal.h> +#include <dm/uclass-internal.h> #include <errno.h> #include <misc.h> @@ -65,6 +67,29 @@ int misc_set_enabled(struct udevice *dev, bool val)return ops->set_enabled (dev, val); } +int misc_init_by_ofnode (ofnode node) +{ +struct udevice *dev =NULL ; +int ret; +long temp1, temp2;
drivers/misc/misc-uclass.c int misc_set_enabled(struct udevice *dev, bool val) · line 84
Quoted 10 lines
drivers/misc/misc-uclass.c +10
-0+ + temp1 =ofnode_to_offset (node); + +for (ret =uclass_find_first_device (UCLASS_MISC , &dev); dev; + ret =uclass_find_next_device (&dev)) { + temp2 =ofnode_to_offset (dev_ofnode (dev)); +if (temp1 == temp2) { + ret =device_probe (dev); +if (ret) + debug("%s: Failed to initialize - %d\n",
Doesn't that put a 'minus' in front of an already negative number? Might look more confusing than it needs to be. I've often seen the formatting specifier for errors be like "...(%d)...".
Quoted 18 lines
drivers/misc/misc-uclass.c +8
-0+ dev->name, ret); +return ret; + } + } + +return -ENODEV ; +} +UCLASS_DRIVER (misc) = { .id =UCLASS_MISC , .name ="misc" ,include/misc.h +0
-0@@ -76,6 +76,15 @@ int misc_call(struct udevice *dev, int msgid, void *tx_msg, int tx_size,*/ int misc_set_enabled (struct udevice *dev,bool val);
include/misc.h int misc_call(struct udevice *dev, int msgid, void *tx_msg, int tx_size, · line 87
Quoted 10 lines
include/misc.h +9
-0+/** +* misc_init_by_ofnode() - Probe a misc device by using ofnode. +* @node: ofnode of the misc device. +* +* A misc device is probed using ofnode. +* +* Return: -ve on error, 0 on success +*/ +int misc_init_by_ofnode (ofnode node);
Missing blank line here.
--
Andreas Dannenberg
Texas Instruments Inc
Quoted 10 lines
include/misc.h +0
-0/* * struct misc_ops - Driver model Misc operations * Signature · 6 lines
2.17.1 _______________________________________________ U-Boot mailing list U-Boot at lists.denx.de https://lists.denx.de/listinfo/u-boot
PATCH 3/14 misc: uclass: Introduce misc_init_by_ofnode
In reply to KeerthyHi Keerthy,
Quoted 3 lines · Keerthy
Introduce
misc_init_by_ofnodeto probe a misc device using its ofnode.
What is the purpose of this? All patches should have a motivation.
Quoted 21 lines
Signed-off-by Keerthy <[email protected]>--- drivers/misc/misc-uclass.c | 25 +++++++++++++++++++++++++ include/misc.h | 9 +++++++++ 2 files changed, 34 insertions(+)drivers/misc/misc-uclass.c +2
-0@@ -5,6 +5,8 @@#include <common.h> #include <dm.h> +#include <dm/device-internal.h> +#include <dm/uclass-internal.h> #include <errno.h> #include <misc.h>
drivers/misc/misc-uclass.c int misc_set_enabled(struct udevice *dev, bool val) · line 76
Quoted 10 lines
drivers/misc/misc-uclass.c +7
-0@@ -65,6 +67,29 @@ int misc_set_enabled(struct udevice *dev, bool val)return ops->set_enabled (dev, val); } +int misc_init_by_ofnode (ofnode node) +{ +struct udevice *dev =NULL ; +int ret; +long temp1, temp2; + + temp1 =ofnode_to_offset (node);
Why convert to offset? Can you just compare the node?
drivers/misc/misc-uclass.c int misc_set_enabled(struct udevice *dev, bool val) · line 79
Quoted 3 lines
drivers/misc/misc-uclass.c +3
-0+ +for (ret =uclass_find_first_device (UCLASS_MISC , &dev); dev; + ret =uclass_find_next_device (&dev)) {
uclass_foreach_dev(dev, UCLASS_MISC) {
drivers/misc/misc-uclass.c int misc_set_enabled(struct udevice *dev, bool val) · line 80
Quoted 1 line
drivers/misc/misc-uclass.c +1
-0+ temp2 =ofnode_to_offset (dev_ofnode (dev));
Again hopefully you can avoid this conversion. It won't work with livetree.
drivers/misc/misc-uclass.c int misc_set_enabled(struct udevice *dev, bool val) · line 82
Quoted 2 lines
drivers/misc/misc-uclass.c +2
-0+if (temp1 == temp2) { + ret =device_probe (dev);
Hmm can you use uclass_foreach_dev_probe() above instead?
Quoted 36 lines
drivers/misc/misc-uclass.c +10
-0+if (ret) +debug ("%s: Failed to initialize - %d\n" , + dev->name, ret); +return ret; + } + } + +return -ENODEV ; +} +UCLASS_DRIVER (misc) = { .id =UCLASS_MISC , .name ="misc" ,include/misc.h +9
-1@@ -76,6 +76,15 @@ int misc_call(struct udevice *dev, int msgid, void *tx_msg, int tx_size,*/ int misc_set_enabled (struct udevice *dev,bool val); +/** +* misc_init_by_ofnode() - Probe a misc device by using ofnode. +* @node: ofnode of the misc device. +* +* A misc device is probed using ofnode. +* +* Return: -ve on error, 0 on success +*/ +int misc_init_by_ofnode (ofnode node);/* * struct misc_ops - Driver model Misc operations * -- 2.17.1
PATCH 3/14 misc: uclass: Introduce misc_init_by_ofnode
Quoted 8 lines · Simon Glass
Hi Keerthy,
Quoted 3 lines · Keerthy
Introduce
misc_init_by_ofnodeto probe a misc device using its ofnode.What is the purpose of this? All patches should have a motivation.
Okay. I will add more details in the commit log. In my case pruss which is the pru subsystem which is a misc device needs to be probed before pru device which is an rproc class device. I will try to figure out a cleaner way if it exists.
Quoted 25 lines
Quoted 25 lines
Signed-off-by Keerthy <[email protected]>--- drivers/misc/misc-uclass.c | 25 +++++++++++++++++++++++++ include/misc.h | 9 +++++++++ 2 files changed, 34 insertions(+)drivers/misc/misc-uclass.c +2
-0@@ -5,6 +5,8 @@#include <common.h> #include <dm.h> +#include <dm/device-internal.h> +#include <dm/uclass-internal.h> #include <errno.h> #include <misc.h> @@ -65,6 +67,29 @@ int misc_set_enabled(struct udevice *dev, bool val)return ops->set_enabled (dev, val); }
Quoted 9 lines
drivers/misc/misc-uclass.c int misc_set_enabled(struct udevice *dev, bool val) · line 76
Quoted 7 lines
drivers/misc/misc-uclass.c +7
-0+int misc_init_by_ofnode (ofnode node) +{ +struct udevice *dev =NULL ; +int ret; +long temp1, temp2; + + temp1 =ofnode_to_offset (node);Why convert to offset? Can you just compare the node?
Yes okay.
Quoted 6 lines
drivers/misc/misc-uclass.c int misc_set_enabled(struct udevice *dev, bool val) · line 79
Quoted 4 lines
drivers/misc/misc-uclass.c +3
-0+ +for (ret =uclass_find_first_device (UCLASS_MISC , &dev); dev; + ret =uclass_find_next_device (&dev)) {
uclass_foreach_dev(dev, UCLASS_MISC){
okay
Quoted 4 lines
drivers/misc/misc-uclass.c int misc_set_enabled(struct udevice *dev, bool val) · line 80
Quoted 2 lines
drivers/misc/misc-uclass.c +1
-0+ temp2 =ofnode_to_offset (dev_ofnode (dev));Again hopefully you can avoid this conversion. It won't work with livetree.
okay
Quoted 5 lines
drivers/misc/misc-uclass.c int misc_set_enabled(struct udevice *dev, bool val) · line 82
Quoted 3 lines
drivers/misc/misc-uclass.c +2
-0+if (temp1 == temp2) { + ret =device_probe (dev);Hmm can you use
uclass_foreach_dev_probe()above instead?
okay. I will try to figure out a cleaner way for my case.
Thanks for the review.
Quoted 38 lines
Quoted 38 lines
drivers/misc/misc-uclass.c +10
-0+if (ret) +debug ("%s: Failed to initialize - %d\n" , + dev->name, ret); +return ret; + } + } + +return -ENODEV ; +} +UCLASS_DRIVER (misc) = { .id =UCLASS_MISC , .name ="misc" ,include/misc.h +9
-1@@ -76,6 +76,15 @@ int misc_call(struct udevice *dev, int msgid, void *tx_msg, int tx_size,*/ int misc_set_enabled (struct udevice *dev,bool val); +/** +* misc_init_by_ofnode() - Probe a misc device by using ofnode. +* @node: ofnode of the misc device. +* +* A misc device is probed using ofnode. +* +* Return: -ve on error, 0 on success +*/ +int misc_init_by_ofnode (ofnode node);/* * struct misc_ops - Driver model Misc operations * -- 2.17.1
PATCH 4/14 soc: ti: pruss: add a misc driver for PRUSS in TI SoCs
In reply to KeerthyThe Programmable Real-Time Unit - Industrial Communication Subsystem (PRU-ICSS) is present of various TI SoCs such as AM335x or AM437x or the AM654x family. Each SoC can have one or more PRUSS instances that may or may not be identical.
The PRUSS consists of dual 32-bit RISC cores called the Programmable Real-Time Units (PRUs), some shared, data and instruction memories, some internal peripheral modules, and an interrupt controller. The programmable nature of the PRUs provide flexibility to implement custom peripheral interfaces, fast real-time responses, or specialized data handling.
Add support for pruss driver. Currently am654x family is supported.
--- drivers/soc/ti/Kconfig | 13 ++++ drivers/soc/ti/Makefile | 1 + drivers/soc/ti/pruss.c | 128 ++++++++++++++++++++++++++++++++++++++++ include/ti-pruss.h | 12 ++++ 4 files changed, 154 insertions(+) create mode 100644 drivers/soc/ti/pruss.c create mode 100644 include/ti-pruss.h
drivers/soc/ti/Kconfig +13 -0
@@ -23,4 +23,17 @@ config TI_KEYSTONE_SERDES SerDes driver for Keystone SoC used for ethernet support on TI K2 platforms. +config TI_PRUSS + bool "Support for TI's K3 based Pruss driver" + depends on DM + depends on ARCH_K3 + depends on OF_CONTROL + depends on SYSCON + help + + Support for TI PRU-ICSSG subsystem. + + Currently supported on AM65xx SoCs Say Y here to support the + Programmable Realtime Unit (PRU). + endif # SOC_TI
drivers/soc/ti/Makefile +1 -0
@@ -2,3 +2,4 @@ obj-$(CONFIG_TI_K3_NAVSS_RINGACC) += k3-navss-ringacc.o obj-$(CONFIG_TI_KEYSTONE_SERDES) += keystone_serdes.o +obj-$(CONFIG_TI_PRUSS) += pruss.o
drivers/soc/ti/pruss.c new file +128 -0
128 lines not on this page · raw message
include/ti-pruss.h new file +12 -0
@@ -0,0 +1,12 @@ +// SPDX-License-Identifier: GPL-2.0 + +struct pruss { +phys_addr_t pruss_dram0; +phys_addr_t pruss_dram1; +phys_addr_t pruss_shrdram2; +phys_size_t pruss_dram0sz; +phys_size_t pruss_dram1sz; +phys_size_t pruss_shrdram2sz; +}; + +int pruss_request_shrmem_region (struct udevice *dev,phys_addr_t *loc);
Signature · 1 line
2.17.1
PATCH 4/14 soc: ti: pruss: add a misc driver for PRUSS in TI SoCs
Quoted 31 lines · Keerthy
The Programmable Real-Time Unit - Industrial Communication Subsystem (PRU-ICSS) is present of various TI SoCs such as AM335x or AM437x or the AM654x family. Each SoC can have one or more PRUSS instances that may or may not be identical.
The PRUSS consists of dual 32-bit RISC cores called the Programmable Real-Time Units (PRUs), some shared, data and instruction memories, some internal peripheral modules, and an interrupt controller. The programmable nature of the PRUs provide flexibility to implement custom peripheral interfaces, fast real-time responses, or specialized data handling.
Add support for pruss driver. Currently am654x family is supported.
Signed-off-by Keerthy <[email protected]>--- drivers/soc/ti/Kconfig | 13 ++++ drivers/soc/ti/Makefile | 1 + drivers/soc/ti/pruss.c | 128 ++++++++++++++++++++++++++++++++++++++++ include/ti-pruss.h | 12 ++++ 4 files changed, 154 insertions(+) create mode 100644 drivers/soc/ti/pruss.c create mode 100644 include/ti-pruss.hdrivers/soc/ti/Kconfig +0
-0@@ -23,4 +23,17 @@ config TI_KEYSTONE_SERDES SerDes driver for Keystone SoC used for ethernet support on TI
drivers/soc/ti/Kconfig config TI_KEYSTONE_SERDES · line 33
Quoted 10 lines
drivers/soc/ti/Kconfig +8
-0K2 platforms. +config TI_PRUSS + bool "Support for TI's K3 based Pruss driver" + depends on DM + depends on ARCH_K3 + depends on OF_CONTROL + depends on SYSCON + help +
Usually there is no blank line below 'help', the next line should follow immediately to better indicate those things belong together.
drivers/soc/ti/Kconfig config TI_KEYSTONE_SERDES · line 36
Quoted 3 lines
drivers/soc/ti/Kconfig +3
-0+ Support for TI PRU-ICSSG subsystem. + + Currently supported on AM65xx SoCs Say Y here to support the
First sentence is missing the period ('.')
Quoted 125 lines
drivers/soc/ti/Kconfig +2
-0+ Programmable Realtime Unit (PRU). + endif # SOC_TIdrivers/soc/ti/Makefile +1
-0@@ -2,3 +2,4 @@ obj-$(CONFIG_TI_K3_NAVSS_RINGACC) += k3-navss-ringacc.o obj-$(CONFIG_TI_KEYSTONE_SERDES) += keystone_serdes.o +obj-$(CONFIG_TI_PRUSS) += pruss.odrivers/soc/ti/pruss.c new file +107
-0@@ -0,0 +1,128 @@ +// SPDX-License-Identifier: GPL-2.0 +/* +* PRU-ICSS platform driver for various TI SoCs +* +* Copyright (C) 2019 Texas Instruments Incorporated - http://www.ti.com/ +* Keerthy <[email protected]> +*/ + +#include <common.h> +#include <dm.h> +#include <dm/of_access.h> +#include <errno.h> +#include <clk.h> +#include <reset.h> +#include <regmap.h> +#include <syscon.h> +#include <asm/io.h> +#include <power-domain.h> +#include <ti-pruss.h> + +#define PRUSS_CFG_IEPCLK 0x30 +#define ICSSG_CFG_CORE_SYNC 0x3c + +/* PRUSS_IEPCLK register bits */ +#define PRUSS_IEPCLK_IEP_OCP_CLK_EN BIT(0) + +/* ICSSG CORE_SYNC register bits */ +#define ICSSG_CORE_VBUSP_SYNC_EN BIT(0) + +/** +* enum pruss_mem - PRUSS memory range identifiers +*/ +enum pruss_mem { +PRUSS_MEM_DRAM0 =0 , +PRUSS_MEM_DRAM1 , +PRUSS_MEM_SHRD_RAM2 , +PRUSS_MEM_MAX , +}; + +int pruss_request_shrmem_region (struct udevice *dev,phys_addr_t *loc) +{ +struct pruss *priv; + + priv =dev_get_priv (dev); +if (!priv || !priv->pruss_shrdram2) +return -EINVAL ; + + *loc = priv->pruss_shrdram2; + +return 0 ; +} + +static int pruss_bind (struct udevice *dev) +{ +return dm_scan_fdt_dev (dev); +} + +/** +* pruss_probe() - Basic probe +* @dev: corresponding k3 device +* +* Return: 0 if all goes good, else appropriate error message. +*/ +static int pruss_probe (struct udevice *dev) +{ +struct pruss *priv; +int ret, idx; +ofnode sub_node, node, memories; +struct regmap *regmap_cfg; +struct udevice *syscon; + + priv =dev_get_priv (dev); + node =dev_ofnode (dev); + sub_node =ofnode_find_subnode (node,"cfg" ); + memories =ofnode_find_subnode (node,"memories" ); + + idx =ofnode_stringlist_search (memories,"reg-names" ,"dram0" ); + priv->pruss_dram0 =ofnode_get_addr_size_index (memories, idx, + (u64 *)&priv->pruss_dram0sz); + idx =ofnode_stringlist_search (memories,"reg-names" ,"dram1" ); + priv->pruss_dram1 =ofnode_get_addr_size_index (memories, idx, + (u64 *)&priv->pruss_dram1sz); + idx =ofnode_stringlist_search (memories,"reg-names" ,"shrdram2" ); + priv->pruss_shrdram2 =ofnode_get_addr_size_index (memories, idx, + (u64 *)&priv->pruss_shrdram2sz); + + ret =uclass_get_device_by_ofnode (UCLASS_SYSCON , sub_node, + &syscon); + + regmap_cfg =syscon_get_regmap (syscon); + +/* +* The CORE block uses two multiplexers to allow software to +* select one of three source clocks (ICSSGn_CORE_CLK, ICSSGn_ICLK or +* ICSSGn_IEP_CLK) for the final clock source of the CORE block. +* The user needs to configure ICSSG_CORE_SYNC_REG[0] CORE_VBUSP_SYNC_EN +* bit & ICSSG_IEPCLK_REG[0] IEP_OCP_CLK_EN bit in order to select the +* clock source to the CORE block. +*/ + ret =regmap_update_bits (regmap_cfg,ICSSG_CFG_CORE_SYNC , +ICSSG_CORE_VBUSP_SYNC_EN , +ICSSG_CORE_VBUSP_SYNC_EN ); +if (ret) +return ret; + ret =regmap_update_bits (regmap_cfg,PRUSS_CFG_IEPCLK , +PRUSS_IEPCLK_IEP_OCP_CLK_EN , +PRUSS_IEPCLK_IEP_OCP_CLK_EN );
drivers/soc/ti/pruss.c line 117
Quoted 10 lines
drivers/soc/ti/pruss.c new file +10
-0+if (ret) +return ret; + +dev_dbg (dev,"pruss successfully probed %s\n" , dev->name); + +return 0 ; +} + +static const struct udevice_id pruss_ids[] = { + { .compatible ="ti,am654-icssg" },
There should be a space prior to the closing curly brace.
Quoted 17 lines
drivers/soc/ti/pruss.c new file +11
-0+ {} +}; + +U_BOOT_DRIVER (pruss) = { + .name ="pruss" , + .of_match = pruss_ids, + .id =UCLASS_MISC , + .bind = pruss_bind, + .probe = pruss_probe, + .priv_auto_alloc_size =sizeof (struct pruss ), +};include/ti-pruss.h new file +0
-0
include/ti-pruss.h line 2
Quoted 2 lines
include/ti-pruss.h new file +2
-0@@ -0,0 +1,12 @@ +// SPDX-License-Identifier: GPL-2.0 +
How about adding some include guards (#ifndef...) to the header file?
--
Andreas Dannenberg
Texas Instruments Inc
Quoted 17 lines
include/ti-pruss.h new file +10
-0+struct pruss { +phys_addr_t pruss_dram0; +phys_addr_t pruss_dram1; +phys_addr_t pruss_shrdram2; +phys_size_t pruss_dram0sz; +phys_size_t pruss_dram1sz; +phys_size_t pruss_shrdram2sz; +}; + +int pruss_request_shrmem_region (struct udevice *dev,phys_addr_t *loc);Signature · 6 lines
2.17.1 _______________________________________________ U-Boot mailing list U-Boot at lists.denx.de https://lists.denx.de/listinfo/u-boot
PATCH 5/14 remoteproc: pruss: add PRU remoteproc driver
In reply to KeerthyThe Programmable Real-Time Unit Subsystem (PRUSS) consists of dual 32-bit RISC cores (Programmable Real-Time Units, or PRUs) for program execution. This patch adds a remoteproc platform driver for managing the individual PRU RISC cores life cycle.
The driver currently supports the AM65xx SoC
--- drivers/remoteproc/Kconfig | 11 + drivers/remoteproc/Makefile | 1 + drivers/remoteproc/pru_rproc.c | 384 +++++++++++++++++++++++++++++++++ 3 files changed, 396 insertions(+) create mode 100644 drivers/remoteproc/pru_rproc.c
drivers/remoteproc/Kconfig +11 -0
@@ -61,4 +61,15 @@ config REMOTEPROC_TI_POWER help Say 'y' here to add support for TI power processors such as those found on certain TI keystone and OMAP generation SoCs. + +config REMOTEPROC_TI_PRU + bool "Support for TI's K3 based PRU remoteproc driver" + select REMOTEPROC + depends on DM + depends on TI_PRUSS + depends on ARCH_K3 + depends on OF_CONTROL + help + Say 'y' here to add support for TI' K3 remoteproc driver. + endmenu
drivers/remoteproc/Makefile +1 -0
@@ -12,3 +12,4 @@ obj-$(CONFIG_REMOTEPROC_SANDBOX) += sandbox_testproc.o obj-$(CONFIG_REMOTEPROC_STM32_COPRO) += stm32_copro.o obj-$(CONFIG_REMOTEPROC_TI_K3_ARM64) += ti_k3_arm64_rproc.o obj-$(CONFIG_REMOTEPROC_TI_POWER) += ti_power_proc.o +obj-$(CONFIG_REMOTEPROC_TI_PRU) += pru_rproc.o
drivers/remoteproc/pru_rproc.c new file +384 -0
384 lines not on this page · raw message
Signature · 1 line
2.17.1
PATCH 6/14 net: ti: icssg-prueth: Add ICSSG ethernet driver
In reply to KeerthyThis is the Ethernet driver for TI SoCs with the ICSSG PRU Sub-system running EMAC firmware. This driver caters to either of the slices(pru/rtu pair) of the icssg subsystem.
Following are the firmwares needed to run cores:
am65x-pru0-prueth-fw.elf for pru0 of slice0
am65x-rtu0-prueth-fw.elf for rtu0 of slice0
am65x-pru1-prueth-fw.elf for pru1 of slice1
am65x-rtu1-prueth-fw.elf for rtu1 of slice1
One and exactly one of the slices is supported as the u-boot ethernet supports probing one interface at a time.
--- drivers/net/ti/Kconfig | 8 + drivers/net/ti/Makefile | 1 + drivers/net/ti/icssg-prueth.c | 517 ++++++++++++++++++++++++++++++ drivers/net/ti/icssg.h | 31 ++ drivers/net/ti/icssg_classifier.c | 397 +++++++++++++++++++++++ 5 files changed, 954 insertions(+) create mode 100644 drivers/net/ti/icssg-prueth.c create mode 100644 drivers/net/ti/icssg.h create mode 100644 drivers/net/ti/icssg_classifier.c
drivers/net/ti/Kconfig +8 -0
@@ -26,3 +26,11 @@ config TI_AM65_CPSW_NUSS help This driver supports TI K3 MCU CPSW Nuss Ethernet controller in Texas Instruments K3 AM65x SoCs. + +config TI_AM64_ICSSG_PRUETH + bool "TI Gigabit PRU Ethernet driver" + depends on ARCH_K3 + select PHYLIB + help + Support Gigabit Ethernet ports over the ICSSG PRU Subsystem + This subsystem is available starting with the AM65 platform.
drivers/net/ti/Makefile +1 -0
@@ -6,3 +6,4 @@ obj-$(CONFIG_DRIVER_TI_CPSW) += cpsw.o cpsw-common.o cpsw_mdio.o obj-$(CONFIG_DRIVER_TI_EMAC) += davinci_emac.o obj-$(CONFIG_DRIVER_TI_KEYSTONE_NET) += keystone_net.o cpsw_mdio.o obj-$(CONFIG_TI_AM65_CPSW_NUSS) += am65-cpsw-nuss.o cpsw_mdio.o +obj-$(CONFIG_TI_AM64_ICSSG_PRUETH) += icssg-prueth.o cpsw_mdio.o icssg_classifier.o
drivers/net/ti/icssg-prueth.c new file +517 -0
517 lines not on this page · raw message
drivers/net/ti/icssg.h new file +31 -0
@@ -0,0 +1,31 @@ +/* SPDX-License-Identifier: GPL-2.0+ */ +/* +* Texas Instruments K3 AM65 Ethernet Switch SubSystem Driver +* +* Copyright (C) 2019, Texas Instruments, Incorporated +* +*/ +void icssg_class_set_mac_addr (struct regmap *miig_rt,int slice,u8 *mac); +void icssg_class_disable (struct regmap *miig_rt,int slice); +void icssg_class_default (struct regmap *miig_rt,int slice); +void icssg_class_promiscuous (struct regmap *miig_rt,int slice); + +/* Config area lies in shared RAM */ +#define ICSSG_CONFIG_OFFSET_SLICE0 0 +#define ICSSG_CONFIG_OFFSET_SLICE1 0x8000 + +/* Load time Fiwmware Configuration */ +struct icssg_config { +__le32 status;/* Firmware status */ +__le32 addr_lo;/* MSMC Buffer pool base address low. */ +__le32 addr_hi;/* MSMC Buffer pool base address high. Must be 0 */ +__le32 tx_buf_sz[16 ];/* Array of buffer pool sizes */ +__le32 num_tx_threads;/* Number of active egress threads, 1 to 4 */ +__le32 tx_rate_lim_en;/* Bitmask: Egress rate limit enable per thread */ +__le32 rx_flow_id;/* RX flow id for first rx ring */ +__le32 rx_mgr_flow_id;/* RX flow id for the first management ring */ +__le32 flags;/* TBD */ +__le32 n_burst;/* for debug */ +__le32 rtu_status;/* RTU status */ +__le32 info;/* reserved */ +} __packed;
drivers/net/ti/icssg_classifier.c new file +397 -0
397 lines not on this page · raw message
Signature · 1 line
2.17.1
PATCH 6/14 net: ti: icssg-prueth: Add ICSSG ethernet driver
Quoted 59 lines · Keerthy
This is the Ethernet driver for TI SoCs with the ICSSG PRU Sub-system running EMAC firmware. This driver caters to either of the slices(pru/rtu pair) of the icssg subsystem.
Following are the firmwares needed to run cores:
am65x-pru0-prueth-fw.elf for pru0 of slice0
am65x-rtu0-prueth-fw.elf for rtu0 of slice0
am65x-pru1-prueth-fw.elf for pru1 of slice1
am65x-rtu1-prueth-fw.elf for rtu1 of slice1One and exactly one of the slices is supported as the u-boot ethernet supports probing one interface at a time.
Signed-off-by Keerthy <[email protected]>--- drivers/net/ti/Kconfig | 8 + drivers/net/ti/Makefile | 1 + drivers/net/ti/icssg-prueth.c | 517 ++++++++++++++++++++++++++++++ drivers/net/ti/icssg.h | 31 ++ drivers/net/ti/icssg_classifier.c | 397 +++++++++++++++++++++++ 5 files changed, 954 insertions(+) create mode 100644 drivers/net/ti/icssg-prueth.c create mode 100644 drivers/net/ti/icssg.h create mode 100644 drivers/net/ti/icssg_classifier.cdrivers/net/ti/Kconfig +8
-0@@ -26,3 +26,11 @@ config TI_AM65_CPSW_NUSS help This driver supports TI K3 MCU CPSW Nuss Ethernet controller in Texas Instruments K3 AM65x SoCs. + +config TI_AM64_ICSSG_PRUETH + bool "TI Gigabit PRU Ethernet driver" + depends on ARCH_K3 + select PHYLIB + help + Support Gigabit Ethernet ports over the ICSSG PRU Subsystem + This subsystem is available starting with the AM65 platform.drivers/net/ti/Makefile +1
-0@@ -6,3 +6,4 @@ obj-$(CONFIG_DRIVER_TI_CPSW) += cpsw.o cpsw-common.o cpsw_mdio.o obj-$(CONFIG_DRIVER_TI_EMAC) += davinci_emac.o obj-$(CONFIG_DRIVER_TI_KEYSTONE_NET) += keystone_net.o cpsw_mdio.o obj-$(CONFIG_TI_AM65_CPSW_NUSS) += am65-cpsw-nuss.o cpsw_mdio.o +obj-$(CONFIG_TI_AM64_ICSSG_PRUETH) += icssg-prueth.o cpsw_mdio.o icssg_classifier.odrivers/net/ti/icssg-prueth.c new file +0
-0
drivers/net/ti/icssg-prueth.c line 3
Quoted 3 lines
drivers/net/ti/icssg-prueth.c new file +3
-0@@ -0,0 +1,517 @@ +// SPDX-License-Identifier: GPL-2.0+ +/* +* Texas Instruments K3 AM65 PRU Ethernet Driver
If the driver is AM65x specific that should be reflected in the driver name, like already done for CPSW_NUSS, for consistency and scalability sake.
Quoted 167 lines
drivers/net/ti/icssg-prueth.c new file +167
-0+* +* Copyright (C) 2019, Texas Instruments, Incorporated +* +*/ + +#include <common.h> +#include <asm/io.h> +#include <asm/processor.h> +#include <clk.h> +#include <dm.h> +#include <dm/lists.h> +#include <dm/device.h> +#include <dma-uclass.h> +#include <dm/of_access.h> +#include <fs_loader.h> +#include <miiphy.h> +#include <misc.h> +#include <net.h> +#include <phy.h> +#include <power-domain.h> +#include <linux/soc/ti/ti-udma.h> +#include <regmap.h> +#include <remoteproc.h> +#include <syscon.h> +#include <ti-pruss.h> + +#include "cpsw_mdio.h" +#include "icssg.h" + +#define ICSS_SLICE0 0 +#define ICSS_SLICE1 1 + +#ifdef PKTSIZE_ALIGN +#define UDMA_RX_BUF_SIZE PKTSIZE_ALIGN +#else +#define UDMA_RX_BUF_SIZE ALIGN(1522, ARCH_DMA_MINALIGN) +#endif + +#ifdef PKTBUFSRX +#define UDMA_RX_DESC_NUM PKTBUFSRX +#else +#define UDMA_RX_DESC_NUM 4 +#endif + +enum prueth_mac { +PRUETH_MAC0 =0 , +PRUETH_MAC1 , +PRUETH_NUM_MACS , +}; + +enum prueth_port { +PRUETH_PORT_HOST =0 ,/* host side port */ +PRUETH_PORT_MII0 ,/* physical port MII 0 */ +PRUETH_PORT_MII1 ,/* physical port MII 1 */ +}; + +/* Config region lies in shared RAM */ +#define ICSS_CONFIG_OFFSET_SLICE0 0 +#define ICSS_CONFIG_OFFSET_SLICE1 0x8000 + +/* Firmware flags */ +#define ICSS_SET_RUN_FLAG_VLAN_ENABLE BIT(0)/* switch only */ +#define ICSS_SET_RUN_FLAG_FLOOD_UNICAST BIT(1)/* switch only */ +#define ICSS_SET_RUN_FLAG_PROMISC BIT(2)/* MAC only */ +#define ICSS_SET_RUN_FLAG_MULTICAST_PROMISC BIT(3)/* MAC only */ + +/* CTRLMMR_ICSSG_RGMII_CTRL register bits */ +#define ICSSG_CTRL_RGMII_ID_MODE BIT(24) + +/** +* enum pruss_pru_id - PRU core identifiers +*/ +enum pruss_pru_id { +PRUSS_PRU0 =0 , +PRUSS_PRU1 , +PRUSS_NUM_PRUS , +}; + +struct prueth { +struct udevice *dev; +struct regmap *miig_rt; +fdt_addr_t mdio_base; +phys_addr_t pruss_shrdram2; +struct mii_dev *bus; +u32 port_id; +u32 sram_pa; +struct phy_device *phydev; +bool has_phy; +ofnode phy_node; +u32 phy_addr; +ofnode eth_node[PRUETH_NUM_MACS ]; +struct icssg_config config[PRUSS_NUM_PRUS ]; +u32 mdio_freq; +int phy_interface; +struct clk mdiofck; +struct dma dma_tx; +struct dma dma_rx; +u32 rx_next; +u32 rx_pend; +int slice; +}; + +static int icssg_phy_init (struct udevice *dev) +{ +struct prueth *priv =dev_get_priv (dev); +struct phy_device *phydev; +u32 supported =PHY_GBIT_FEATURES ; +int ret; + + phydev =phy_connect (priv->bus, + priv->phy_addr, + priv->dev, + priv->phy_interface); + +if (!phydev) { +dev_err (dev,"phy_connect() failed\n" ); +return -ENODEV ; + } + + phydev->supported &= supported; + phydev->advertising = phydev->supported; + +#ifdef CONFIG_DM_ETH +if (ofnode_valid (priv->phy_node)) + phydev->node = priv->phy_node; +#endif + + priv->phydev = phydev; + ret =phy_config (phydev); +if (ret <0 ) +pr_err ("phy_config() failed: %d" , ret); + +return ret; +} + +static int icssg_mdio_init (struct udevice *dev) +{ +struct prueth *prueth =dev_get_priv (dev); + + prueth->bus =cpsw_mdio_init (dev->name, prueth->mdio_base, + prueth->mdio_freq, +clk_get_rate (&prueth->mdiofck)); +if (!prueth->bus) +return -EFAULT ; + +return 0 ; +} + +static void icssg_config_set (struct prueth *prueth) +{ +void __iomem *va; + + va = (void __iomem *)prueth->pruss_shrdram2 + prueth->slice * +ICSSG_CONFIG_OFFSET_SLICE1 ; + +memcpy_toio (va, &prueth->config[0 ],sizeof (prueth->config[0 ])); +} + +static int prueth_start (struct udevice *dev) +{ +struct prueth *priv =dev_get_priv (dev); +struct eth_pdata *pdata = dev->platdata; +int ret, i; +char tx_chn_name[16 ]; +char rx_chn_name[16 ]; + + icssg_class_set_mac_addr(priv->miig_rt, priv->slice,
drivers/net/ti/icssg-prueth.c line 180
Quoted 10 lines
drivers/net/ti/icssg-prueth.c new file +10
-0+ (u8 *)pdata->enetaddr); +icssg_class_default (priv->miig_rt, priv->slice); + +/* To differentiate channels for SLICE0 vs SLICE1 */ +snprintf (tx_chn_name,sizeof (tx_chn_name),"tx%d-0" , priv->slice); +snprintf (rx_chn_name,sizeof (rx_chn_name),"rx%d" , priv->slice); + + ret =dma_get_by_name (dev, tx_chn_name, &priv->dma_tx); +if (ret) +dev_err (dev,"TX dma get failed %d\n" , ret);
Should we return on errors? Same comment applies to the statements immediately following this...
Quoted 35 lines
drivers/net/ti/icssg-prueth.c new file +35
-0+ + ret =dma_get_by_name (dev, rx_chn_name, &priv->dma_rx); +if (ret) +dev_err (dev,"RX dma get failed %d\n" , ret); + +for (i =0 ; i <UDMA_RX_DESC_NUM ; i++) { + ret =dma_prepare_rcv_buf (&priv->dma_rx, + net_rx_packets[i], +UDMA_RX_BUF_SIZE ); +if (ret) +dev_err (dev,"RX dma add buf failed %d\n" , ret); + } + + ret =dma_enable (&priv->dma_tx); +if (ret) { +dev_err (dev,"TX dma_enable failed %d\n" , ret); +return ret; + } + + ret =dma_enable (&priv->dma_rx); +if (ret) { +dev_err (dev,"RX dma_enable failed %d\n" , ret); + goto rx_fail; + } + + ret =phy_startup (priv->phydev); +if (ret) { +dev_err (dev,"phy_startup failed\n" ); + goto phy_fail; + } + +return 0 ; +phy_fail: +dma_disable (&priv->dma_rx); +rx_fail:
drivers/net/ti/icssg-prueth.c line 225
Quoted 10 lines
drivers/net/ti/icssg-prueth.c new file +10
-0+dma_disable (&priv->dma_tx); + +return ret; +} + +void prueth_print_buf (ulong addr,const void *data,uint width, +uint count,uint linelen) +{ +print_buffer (addr, data, width, count, linelen); +}
This function does not appear to be used? If so please remove...
Quoted 153 lines
drivers/net/ti/icssg-prueth.c new file +153
-0+ +static int prueth_send (struct udevice *dev,void *packet,int length) +{ +struct prueth *priv =dev_get_priv (dev); +int ret; + + ret =dma_send (&priv->dma_tx, packet, length,NULL ); + +return ret; +} + +static int prueth_recv (struct udevice *dev,int flags,uchar **packetp) +{ +struct prueth *priv =dev_get_priv (dev); +int ret; + +/* try to receive a new packet */ + ret =dma_receive (&priv->dma_rx, (void **)packetp,NULL ); + +return ret; +} + +static int prueth_free_pkt (struct udevice *dev,uchar *packet,int length) +{ +struct prueth *priv =dev_get_priv (dev); +int ret =0 ; + +if (length >0 ) { +u32 pkt = priv->rx_next %UDMA_RX_DESC_NUM ; + +dev_dbg (dev,"%s length:%d pkt:%u\n" , __func__, length, pkt); + + ret =dma_prepare_rcv_buf (&priv->dma_rx, + net_rx_packets[pkt], +UDMA_RX_BUF_SIZE ); + priv->rx_next++; + } + +return ret; +} + +static void prueth_stop (struct udevice *dev) +{ +struct prueth *priv =dev_get_priv (dev); + +icssg_class_disable (priv->miig_rt, priv->slice); + +phy_shutdown (priv->phydev); + +dma_disable (&priv->dma_tx); +dma_free (&priv->dma_tx); + +dma_disable (&priv->dma_rx); +dma_free (&priv->dma_rx); +} + +static const struct eth_ops prueth_ops = { + .start = prueth_start, + .send = prueth_send, + .recv = prueth_recv, + .free_pkt = prueth_free_pkt, + .stop = prueth_stop, +}; + +static int icssg_ofdata_parse_phy (struct udevice *dev,ofnode port_np) +{ +struct prueth *priv =dev_get_priv (dev); +struct ofnode_phandle_args out_args; +const char *phy_mode; +int ret =0 ; + + phy_mode =ofnode_read_string (port_np,"phy-mode" ); +if (phy_mode) { + priv->phy_interface = +phy_get_interface_by_name (phy_mode); +if (priv->phy_interface ==-1 ) { +dev_err (dev,"Invalid PHY mode '%s'\n" , + phy_mode); + ret = -EINVAL ; + goto out; + } + } + + ret =ofnode_parse_phandle_with_args (port_np,"phy-handle" , +NULL ,0 ,0 , &out_args); +if (ret) { +dev_err (dev,"can't parse phy-handle port (%d)\n" , ret); + ret =0 ; + } + + priv->phy_node = out_args.node; + ret =ofnode_read_u32 (priv->phy_node,"reg" , &priv->phy_addr); +if (ret) +dev_err (dev,"failed to get phy_addr port (%d)\n" , ret); + +out: +return ret; +} + +static int prueth_config_rgmiidelay (struct prueth *prueth, +ofnode eth_np) +{ +struct regmap *ctrl_mmr; +u32 val; +int ret =0 ; +ofnode node; +u32 tmp[2 ]; + + ret =ofnode_read_u32_array (eth_np,"syscon-rgmii-delay" , tmp,2 ); +if (ret) { +dev_err (dev,"no syscon-rgmii-delay\n" ); +return ret; + } + + node =ofnode_get_by_phandle (tmp[0 ]); +if (!ofnode_valid (node)) { +dev_err (dev,"can't get syscon-rgmii-delay node\n" ); +return -EINVAL ; + } + + ctrl_mmr =syscon_node_to_regmap (node); +if (!ctrl_mmr) { +dev_err (dev,"can't get ctrl_mmr regmap\n" ); +return -EINVAL ; + } + +if (ofnode_read_bool (eth_np,"enable-rgmii-delay" )) + val =0 ; +else + val =ICSSG_CTRL_RGMII_ID_MODE ; + +regmap_update_bits (ctrl_mmr, tmp[1 ],ICSSG_CTRL_RGMII_ID_MODE , val); + +return 0 ; +} + +static int prueth_probe (struct udevice *dev) +{ +struct prueth *prueth; +int ret =0 , i; +ofnode eth0_node, eth1_node, node, pruss_node, mdio_node, sram_node; +u32 phandle, err, sp; +struct udevice **prussdev =NULL ; +struct icssg_config *config; + + prueth =dev_get_priv (dev); + prueth->dev = dev; + err =ofnode_read_u32 (dev_ofnode (dev),"prus" , &phandle); +if (err) +return err; + + node =ofnode_get_by_phandle (phandle); +if (!ofnode_valid (node))
drivers/net/ti/icssg-prueth.c line 388
Quoted 10 lines
drivers/net/ti/icssg-prueth.c new file +10
-0+return -EINVAL ; + + pruss_node =ofnode_get_parent (node); + err =misc_init_by_ofnode (pruss_node); +if (err) +return err; + + ret =device_find_global_by_ofnode (pruss_node, prussdev); +if (ret) +dev_err (dev,"error getting the pruss dev\n" );
Should we not exit here?
Quoted 37 lines
drivers/net/ti/icssg-prueth.c new file +37
-0+ + ret =pruss_request_shrmem_region (*prussdev, &prueth->pruss_shrdram2); +if (ret) +return ret; + + node =dev_ofnode (dev); + eth0_node =ofnode_find_subnode (node,"ethernet-mii0" ); + eth1_node =ofnode_find_subnode (node,"ethernet-mii1" ); +/* one node must be present and available else we fail */ +if (!ofnode_valid (eth0_node) && !ofnode_valid (eth1_node)) { +dev_err (dev,"neither ethernet-mii0 nor ethernet-mii1 node available\n" ); +return -ENODEV ; + } + +/* +* Exactly one node must be present as uboot ethernet framework does +* not support two interfaces in a single probe. So Device Tree should +* have exactly one of mii0 or mii1 interface. +*/ +if (ofnode_valid (eth0_node) &&ofnode_valid (eth1_node)) { +dev_err (dev,"Both slices cannot be supported\n" ); +return -EINVAL ; + } + +if (ofnode_valid (eth0_node)) { + prueth->slice =0 ; +icssg_ofdata_parse_phy (dev, eth0_node); + prueth->eth_node[PRUETH_MAC0 ] = eth0_node; + } + +if (ofnode_valid (eth1_node)) { + prueth->slice =1 ; +icssg_ofdata_parse_phy (dev, eth1_node); + prueth->eth_node[PRUETH_MAC0 ] = eth1_node; + } + + prueth->miig_rt =syscon_regmap_lookup_by_phandle (dev,"mii-g-rt" );
drivers/net/ti/icssg-prueth.c line 435
Quoted 10 lines
drivers/net/ti/icssg-prueth.c new file +10
-0+if (!prueth->miig_rt) { +dev_err (dev,"couldn't get mii-g-rt syscon regmap\n" ); +return -ENODEV ; + } + + ret =clk_get_by_name (dev,"mdio_fck" , &prueth->mdiofck); +if (ret) { +dev_err (dev,"failed to get clock %d\n" , ret); +return ret; + }
Blank line here for consistency with other statements.
Quoted 45 lines
drivers/net/ti/icssg-prueth.c new file +45
-0+ ret =clk_enable (&prueth->mdiofck); +if (ret) { +dev_err (dev,"clk_enable failed %d\n" , ret); +return ret; + } + + ret =ofnode_read_u32 (dev_ofnode (dev),"sram" , &sp); +if (ret) { +dev_err (dev,"sram node fetch failed %d\n" , ret); +return ret; + } + + sram_node =ofnode_get_by_phandle (sp); +if (!ofnode_valid (node)) +return -EINVAL ; + + prueth->sram_pa =ofnode_get_addr (sram_node); + +if (!prueth->slice) { + ret =prueth_config_rgmiidelay (prueth, eth0_node); +if (ret) { +dev_err (dev,"prueth_config_rgmiidelay failed\n" ); +return ret; + } + }else { + ret =prueth_config_rgmiidelay (prueth, eth1_node); +if (ret) { +dev_err (dev,"prueth_config_rgmiidelay failed\n" ); +return ret; + } + } + + mdio_node =ofnode_find_subnode (pruss_node,"mdio" ); + prueth->mdio_base =ofnode_get_addr (mdio_node); +ofnode_read_u32 (mdio_node,"bus_freq" , &prueth->mdio_freq); + + ret =icssg_mdio_init (dev); +if (ret) +return ret; + + ret =icssg_phy_init (dev); +if (ret) { +dev_err (dev,"phy_init failed\n" ); + goto out; + }
drivers/net/ti/icssg-prueth.c line 490
Quoted 10 lines
drivers/net/ti/icssg-prueth.c new file +10
-0+ +/* Set Load time configuration */ + config = &prueth->config[0 ]; +memset (config,0 ,sizeof (*config)); + config->addr_lo =cpu_to_le32 (lower_32_bits (prueth->sram_pa)); + config->addr_hi =cpu_to_le32 (upper_32_bits (prueth->sram_pa)); + config->num_tx_threads =0 ; + config->rx_flow_id =0 ;/* flow id for host port */ + +for (i =8 ; i <16 ; i++)
tx_buf_sz is 16 elements in size. What is different about elements 0...7 so that they don't need to get intialized?
Quoted 33 lines
drivers/net/ti/icssg-prueth.c new file +27
-0+ config->tx_buf_sz[i] =cpu_to_le32 (0x1800 ); + +icssg_config_set (prueth); + +return 0 ; +out: +cpsw_mdio_free (prueth->bus); +clk_disable (&prueth->mdiofck); + +return ret; +} + +static const struct udevice_id prueth_ids[] = { + { .compatible ="ti,am654-icssg-prueth" }, + { } +}; + +U_BOOT_DRIVER (prueth) = { + .name ="prueth" , + .id =UCLASS_ETH , + .of_match = prueth_ids, + .probe = prueth_probe, + .ops = &prueth_ops, + .priv_auto_alloc_size =sizeof (struct prueth ), + .platdata_auto_alloc_size =sizeof (struct eth_pdata ), + .flags =DM_FLAG_ALLOC_PRIV_DMA , +};drivers/net/ti/icssg.h new file +0
-0
drivers/net/ti/icssg.h line 7
Quoted 7 lines
drivers/net/ti/icssg.h new file +7
-0@@ -0,0 +1,31 @@ +/* SPDX-License-Identifier: GPL-2.0+ */ +/* +* Texas Instruments K3 AM65 Ethernet Switch SubSystem Driver +* +* Copyright (C) 2019, Texas Instruments, Incorporated +* +*/
Header file needs include guards.
Needs blank line here.
drivers/net/ti/icssg.h line 12
Quoted 5 lines
drivers/net/ti/icssg.h new file +5
-0+void icssg_class_set_mac_addr (struct regmap *miig_rt,int slice,u8 *mac); +void icssg_class_disable (struct regmap *miig_rt,int slice); +void icssg_class_default (struct regmap *miig_rt,int slice); +void icssg_class_promiscuous (struct regmap *miig_rt,int slice); +
Typically the order of declarations in header files is...
1) defines
2) types
3) function prototypes
Quoted 25 lines
drivers/net/ti/icssg.h new file +19
-0+/* Config area lies in shared RAM */ +#define ICSSG_CONFIG_OFFSET_SLICE0 0 +#define ICSSG_CONFIG_OFFSET_SLICE1 0x8000 + +/* Load time Fiwmware Configuration */ +struct icssg_config { +__le32 status;/* Firmware status */ +__le32 addr_lo;/* MSMC Buffer pool base address low. */ +__le32 addr_hi;/* MSMC Buffer pool base address high. Must be 0 */ +__le32 tx_buf_sz[16 ];/* Array of buffer pool sizes */ +__le32 num_tx_threads;/* Number of active egress threads, 1 to 4 */ +__le32 tx_rate_lim_en;/* Bitmask: Egress rate limit enable per thread */ +__le32 rx_flow_id;/* RX flow id for first rx ring */ +__le32 rx_mgr_flow_id;/* RX flow id for the first management ring */ +__le32 flags;/* TBD */ +__le32 n_burst;/* for debug */ +__le32 rtu_status;/* RTU status */ +__le32 info;/* reserved */ +} __packed;drivers/net/ti/icssg_classifier.c new file +0
-0
drivers/net/ti/icssg_classifier.c line 4
Quoted 4 lines
drivers/net/ti/icssg_classifier.c new file +4
-0@@ -0,0 +1,397 @@ +// SPDX-License-Identifier: GPL-2.0 +/* Texas Instruments ICSSG Ethernet Driver +* +* Copyright (C) 2018 Texas Instruments Incorporated - http://www.ti.com/
2019
Quoted 378 lines
drivers/net/ti/icssg_classifier.c new file +378
-0+* +*/ + +#include <common.h> +#include <asm/io.h> +#include <asm/processor.h> +#include <clk.h> +#include <dm.h> +#include <dm/lists.h> +#include <dm/device.h> +#include <dma-uclass.h> +#include <dm/of_access.h> +#include <miiphy.h> +#include <net.h> +#include <phy.h> +#include <power-domain.h> +#include <linux/soc/ti/ti-udma.h> +#include <syscon.h> +#include <ti-pruss.h> +#include <regmap.h> + +#include "cpsw_mdio.h" +#include "icssg.h" + +#define ICSSG_NUM_CLASSIFIERS 16 +#define ICSSG_NUM_FT1_SLOTS 8 +#define ICSSG_NUM_FT3_SLOTS 16 + +/* Filter 1 - FT1 */ +#define FT1_NUM_SLOTS 8 +#define FT1_SLOT_SIZE 0x10/* bytes */ + +/* offsets from FT1 slot base i.e. slot 1 start */ +#define FT1_DA0 0x0 +#define FT1_DA1 0x4 +#define FT1_DA0_MASK 0x8 +#define FT1_DA1_MASK 0xc + +#define FT1_N_REG (slize, n, reg) (offs[slice].ft1_slot_base + FT1_SLOT_SIZE * (n) + (reg)) + +#define FT1_LEN_MASK GENMASK(19, 16) +#define FT1_LEN_SHIFT 16 +#define FT1_LEN (len) (((len) << FT1_LEN_SHIFT) & FT1_LEN_MASK) + +#define FT1_MATCH_SLOT (n) (GENMASK(23, 16) & (BIT(n) << 16)) + +enum ft1_cfg_type { +FT1_CFG_TYPE_DISABLED =0 , +FT1_CFG_TYPE_EQ , +FT1_CFG_TYPE_GT , +FT1_CFG_TYPE_LT , +}; + +#define FT1_CFG_SHIFT (n) (2 * (n)) +#define FT1_CFG_MASK (n) (0x3 << FT1_CFG_SHIFT((n))) + +/* Filter 3 - FT3 */ +#define FT3_NUM_SLOTS 16 +#define FT3_SLOT_SIZE 0x20/* bytes */ + +/* offsets from FT3 slot n's base */ +#define FT3_START 0 +#define FT3_START_AUTO 0x4 +#define FT3_START_OFFSET 0x8 +#define FT3_JUMP_OFFSET 0xc +#define FT3_LEN 0x10 +#define FT3_CFG 0x14 +#define FT3_T 0x18 +#define FT3_T_MASK 0x1c + +#define FT3_N_REG (slize, n, reg) (offs[slice].ft3_slot_base + FT3_SLOT_SIZE * (n) + (reg)) + +/* offsets from rx_class n's base */ +#define RX_CLASS_AND_EN 0 +#define RX_CLASS_OR_EN 0x4 + +#define RX_CLASS_NUM_SLOTS 16 +#define RX_CLASS_EN_SIZE 0x8/* bytes */ + +#define RX_CLASS_N_REG (slice, n, reg) (offs[slice].rx_class_base + RX_CLASS_EN_SIZE * (n) + (reg)) + +/* RX Class Gates */ +#define RX_CLASS_GATES_SIZE 0x4/* bytes */ + +#define RX_CLASS_GATES_N_REG (slice, n) (offs[slice].rx_class_gates_base + RX_CLASS_GATES_SIZE * (n)) + +#define RX_CLASS_GATES_ALLOW_MASK BIT(6) +#define RX_CLASS_GATES_RAW_MASK BIT(5) +#define RX_CLASS_GATES_PHASE_MASK BIT(4) + +/* RX Class traffic data matching bits */ +#define RX_CLASS_FT_UC BIT(31) +#define RX_CLASS_FT_MC BIT(30) +#define RX_CLASS_FT_BC BIT(29) +#define RX_CLASS_FT_FW BIT(28) +#define RX_CLASS_FT_RCV BIT(27) +#define RX_CLASS_FT_VLAN BIT(26) +#define RX_CLASS_FT_DA_P BIT(25) +#define RX_CLASS_FT_DA_I BIT(24) +#define RX_CLASS_FT_FT1_MATCH_MASK GENMASK(23, 16) +#define RX_CLASS_FT_FT1_MATCH_SHIFT 16 +#define RX_CLASS_FT_FT3_MATCH_MASK GENMASK(15, 0) +#define RX_CLASS_FT_FT3_MATCH_SHIFT 0 + +enum rx_class_sel_type { +RX_CLASS_SEL_TYPE_OR =0 , +RX_CLASS_SEL_TYPE_AND =1 , +RX_CLASS_SEL_TYPE_OR_AND_AND =2 , +RX_CLASS_SEL_TYPE_OR_OR_AND =3 , +}; + +#define FT1_CFG_SHIFT (n) (2 * (n)) +#define FT1_CFG_MASK (n) (0x3 << FT1_CFG_SHIFT((n))) + +#define RX_CLASS_SEL_SHIFT (n) (2 * (n)) +#define RX_CLASS_SEL_MASK (n) (0x3 << RX_CLASS_SEL_SHIFT((n))) + +#define ICSSG_CFG_OFFSET 0 +#define RGMII_CFG_OFFSET 4 + +#define ICSSG_CFG_RX_L2_G_EN BIT(2) + +/* these are register offsets per PRU */ +struct miig_rt_offsets { +u32 mac0; +u32 mac1; +u32 ft1_start_len; +u32 ft1_cfg; +u32 ft1_slot_base; +u32 ft3_slot_base; +u32 ft3_p_base; +u32 ft_rx_ptr; +u32 rx_class_base; +u32 rx_class_cfg1; +u32 rx_class_cfg2; +u32 rx_class_gates_base; +u32 rx_green; +u32 rx_rate_cfg_base; +u32 rx_rate_src_sel0; +u32 rx_rate_src_sel1; +u32 tx_rate_cfg_base; +u32 stat_base; +u32 tx_hsr_tag; +u32 tx_hsr_seq; +u32 tx_vlan_type; +u32 tx_vlan_ins; +}; + +static struct miig_rt_offsets offs[] = { +/* PRU0 */ + { +0x8 , +0xc , +0x80 , +0x84 , +0x88 , +0x108 , +0x308 , +0x408 , +0x40c , +0x48c , +0x490 , +0x494 , +0x4d4 , +0x4e4 , +0x504 , +0x508 , +0x50c , +0x54c , +0x63c , +0x640 , +0x644 , +0x648 , + }, +/* PRU1 */ + { +0x10 , +0x14 , +0x64c , +0x650 , +0x654 , +0x6d4 , +0x8d4 , +0x9d4 , +0x9d8 , +0xa58 , +0xa5c , +0xa60 , +0xaa0 , +0xab0 , +0xad0 , +0xad4 , +0xad8 , +0xb18 , +0xc08 , +0xc0c , +0xc10 , +0xc14 , + }, +}; + +static void rx_class_ft1_cfg_set_type (struct regmap *miig_rt,int slice,int n, +enum ft1_cfg_type type) +{ +u32 offset; + + offset = offs[slice].ft1_cfg; +regmap_update_bits (miig_rt, offset,FT1_CFG_MASK (n), + type <<FT1_CFG_SHIFT (n)); +} + +static void rx_class_sel_set_type (struct regmap *miig_rt,int slice,int n, +enum rx_class_sel_type type) +{ +u32 offset; + + offset = offs[slice].rx_class_cfg1; +regmap_update_bits (miig_rt, offset,RX_CLASS_SEL_MASK (n), + type <<RX_CLASS_SEL_SHIFT (n)); +} + +static void rx_class_set_and (struct regmap *miig_rt,int slice,int n, +u32 data) +{ +u32 offset; + + offset =RX_CLASS_N_REG (slice, n,RX_CLASS_AND_EN ); +regmap_write (miig_rt, offset, data); +} + +static void rx_class_set_or (struct regmap *miig_rt,int slice,int n, +u32 data) +{ +u32 offset; + + offset =RX_CLASS_N_REG (slice, n,RX_CLASS_OR_EN ); +regmap_write (miig_rt, offset, data); +} + +/* disable all RX traffic */ +void icssg_class_disable (struct regmap *miig_rt,int slice) +{ +u32 data, offset; +int n; + +/* Enable RX_L2_G */ +regmap_update_bits (miig_rt,ICSSG_CFG_OFFSET ,ICSSG_CFG_RX_L2_G_EN , +ICSSG_CFG_RX_L2_G_EN ); + +for (n =0 ; n <ICSSG_NUM_CLASSIFIERS ; n++) { +/* AND_EN = 0 */ +rx_class_set_and (miig_rt, slice, n,0 ); +/* OR_EN = 0 */ +rx_class_set_or (miig_rt, slice, n,0 ); + +/* set CFG1 to OR */ +rx_class_sel_set_type (miig_rt, slice, n,RX_CLASS_SEL_TYPE_OR ); + +/* configure gate */ + offset =RX_CLASS_GATES_N_REG (slice, n); +regmap_read (miig_rt, offset, &data); +/* clear class_raw */ + data &= ~RX_CLASS_GATES_RAW_MASK ; +/* set allow and phase mask */ + data |=RX_CLASS_GATES_ALLOW_MASK |RX_CLASS_GATES_PHASE_MASK ; +regmap_write (miig_rt, offset, data); + } + +/* FT1 uses 6 bytes of DA address */ + offset = offs[slice].ft1_start_len; +regmap_write (miig_rt, offset,FT1_LEN (6 )); + +/* FT1 type EQ */ +for (n =0 ; n <ICSSG_NUM_FT1_SLOTS ; n++) +rx_class_ft1_cfg_set_type (miig_rt, slice, n,FT1_CFG_TYPE_EQ ); + +/* FT1[0] DA compare address 00-00-00-00-00-00 */ + offset =FT1_N_REG (slice,0 ,FT1_DA0 ); +regmap_write (miig_rt, offset,0 ); + offset =FT1_N_REG (slice,0 ,FT1_DA1 ); +regmap_write (miig_rt, offset,0 ); + +/* FT1[0] mask FE-FF-FF-FF-FF-FF */ + offset =FT1_N_REG (slice,0 ,FT1_DA0_MASK ); +regmap_write (miig_rt, offset,0 ); + offset =FT1_N_REG (slice,0 ,FT1_DA1_MASK ); +regmap_write (miig_rt, offset,0 ); + +/* clear CFG2 */ +regmap_write (miig_rt, offs[slice].rx_class_cfg2,0 ); +} + +void icssg_class_default (struct regmap *miig_rt,int slice) +{ +u32 offset, data; +int n; + +/* defaults */ +icssg_class_disable (miig_rt, slice); + +/* FT1 uses 6 bytes of DA address */ + offset = offs[slice].ft1_start_len; +regmap_write (miig_rt, offset,FT1_LEN (6 )); + +/* FT1 slots to EQ */ +for (n =0 ; n <ICSSG_NUM_FT1_SLOTS ; n++) +rx_class_ft1_cfg_set_type (miig_rt, slice, n,FT1_CFG_TYPE_EQ ); + +/* FT1[0] DA compare address 00-00-00-00-00-00 */ + offset =FT1_N_REG (slice,0 ,FT1_DA0 ); +regmap_write (miig_rt, offset,0 ); + offset =FT1_N_REG (slice,0 ,FT1_DA1 ); +regmap_write (miig_rt, offset,0 ); + +/* FT1[0] mask 00-00-00-00-00-00 */ + offset =FT1_N_REG (slice,0 ,FT1_DA0_MASK ); +regmap_write (miig_rt, offset,0 ); + offset =FT1_N_REG (slice,0 ,FT1_DA1_MASK ); +regmap_write (miig_rt, offset,0 ); + +/* Setup Classifier 4 */ +/* match on Broadcast or MAC_PRU address */ + data =RX_CLASS_FT_BC |RX_CLASS_FT_DA_P ; +rx_class_set_or (miig_rt, slice,4 , data); + +/* set CFG1 for OR_OR_AND for classifier 4 */ +rx_class_sel_set_type (miig_rt, slice,4 ,RX_CLASS_SEL_TYPE_OR_OR_AND ); + +/* ungate classifier 4 */ + offset =RX_CLASS_GATES_N_REG (slice,4 ); +regmap_read (miig_rt, offset, &data); + data |=RX_CLASS_GATES_RAW_MASK ; +regmap_write (miig_rt, offset, data); + +/* clear CFG2 */ +regmap_write (miig_rt, offs[slice].rx_class_cfg2,0 ); +} + +void icssg_class_promiscuous (struct regmap *miig_rt,int slice) +{ +u32 data; +u32 offset; +int n; + +/* defaults */ +icssg_class_disable (miig_rt, slice); + +/* FT1 uses 6 bytes of DA address */ + offset = offs[slice].ft1_start_len; +regmap_write (miig_rt, offset,FT1_LEN (6 )); + +/* FT1 type EQ */ +for (n =0 ; n <ICSSG_NUM_FT1_SLOTS ; n++) +rx_class_ft1_cfg_set_type (miig_rt, slice, n,FT1_CFG_TYPE_EQ ); + +/* FT1[0] DA compare address 00-00-00-00-00-00 */ + offset =FT1_N_REG (slice,0 ,FT1_DA0 ); +regmap_write (miig_rt, offset,0 ); + offset =FT1_N_REG (slice,0 ,FT1_DA1 ); +regmap_write (miig_rt, offset,0 ); + +/* FT1[0] mask FE-FF-FF-FF-FF-FF */ + offset =FT1_N_REG (slice,0 ,FT1_DA0_MASK ); +regmap_write (miig_rt, offset,0xfffffffe ); + offset =FT1_N_REG (slice,0 ,FT1_DA1_MASK ); +regmap_write (miig_rt, offset,0xffff ); + +/* Setup Classifier 4 */ +/* match on multicast, broadcast or unicast (ft1-0 address) */ + data =RX_CLASS_FT_MC |RX_CLASS_FT_BC |FT1_MATCH_SLOT (0 ); +rx_class_set_or (miig_rt, slice,4 , data); + +/* set CFG1 for OR_OR_AND for classifier 4 */ +rx_class_sel_set_type (miig_rt, slice,4 ,RX_CLASS_SEL_TYPE_OR_OR_AND ); + +/* ungate classifier 4 */ + offset =RX_CLASS_GATES_N_REG (slice,4 ); +regmap_read (miig_rt, offset, &data);
drivers/net/ti/icssg_classifier.c line 392
Quoted 10 lines
drivers/net/ti/icssg_classifier.c new file +10
-0+ data |=RX_CLASS_GATES_RAW_MASK ; +regmap_write (miig_rt, offset, data); +} + +void icssg_class_set_mac_addr (struct regmap *miig_rt,int slice,u8 *mac) +{ +u32 mac0, mac1; + + mac0 = mac[0 ] | mac[1 ] <<8 | + mac[2 ] <<16 | mac[3 ] <<24 ;
This could be in one line ^^
--
Andreas Dannenberg
Texas Instruments Inc
Quoted 12 lines
drivers/net/ti/icssg_classifier.c new file +5
-0+ mac1 = mac[4 ] | mac[5 ] <<8 ; + +regmap_write (miig_rt, offs[slice].mac0, mac0); +regmap_write (miig_rt, offs[slice].mac1, mac1); +}Signature · 6 lines
2.17.1 _______________________________________________ U-Boot mailing list U-Boot at lists.denx.de https://lists.denx.de/listinfo/u-boot
PATCH 7/14 net: ti: icssg-prueth: Workaround to shutdown the prueth firmware
In reply to KeerthyCurrently shutdown command is yet to be implemented. So work around clearing the task manager of ICSSG.
--- drivers/net/ti/icssg-prueth.c | 8 ++++++++ drivers/soc/ti/pruss.c | 15 +++++++++++++++ include/ti-pruss.h | 1 + 3 files changed, 24 insertions(+)
drivers/net/ti/icssg-prueth.c +8 -0
@@ -84,6 +84,7 @@ struct prueth {struct regmap *miig_rt;fdt_addr_t mdio_base;phys_addr_t pruss_shrdram2; +phys_addr_t tmaddr;struct mii_dev *bus;u32 port_id;u32 sram_pa;
@@ -277,6 +278,9 @@ static void prueth_stop(struct udevice *dev)dma_disable (&priv->dma_rx);dma_free (&priv->dma_rx); + +/* Workaround for shutdown command */ +writel (0x0 , priv->tmaddr + priv->slice *0x200 ); }static const struct eth_ops prueth_ops = {
@@ -391,6 +395,10 @@ static int prueth_probe(struct udevice *dev)if (ret)return ret; + ret =pruss_request_tm_region (*prussdev, &prueth->tmaddr); +if (ret) +return ret; + node =dev_ofnode (dev); eth0_node =ofnode_find_subnode (node,"ethernet-mii0" ); eth1_node =ofnode_find_subnode (node,"ethernet-mii1" );
drivers/soc/ti/pruss.c +15 -0
@@ -21,6 +21,8 @@#define PRUSS_CFG_IEPCLK 0x30#define ICSSG_CFG_CORE_SYNC 0x3c +#define ICSSG_TASK_MGR_OFFSET 0x2a000 +/* PRUSS_IEPCLK register bits */ #define PRUSS_IEPCLK_IEP_OCP_CLK_EN BIT(0)
@@ -37,6 +39,19 @@ enum pruss_mem { PRUSS_MEM_MAX, }; +int pruss_request_tm_region (struct udevice *dev,phys_addr_t *loc) +{ +struct pruss *priv; + + priv =dev_get_priv (dev); +if (!priv || !priv->pruss_dram0) +return -EINVAL ; + + *loc = priv->pruss_dram0 +ICSSG_TASK_MGR_OFFSET ; + +return 0 ; +} +int pruss_request_shrmem_region (struct udevice *dev,phys_addr_t *loc) {struct pruss *priv;
include/ti-pruss.h +1 -0
@@ -10,3 +10,4 @@ struct pruss { };int pruss_request_shrmem_region (struct udevice *dev,phys_addr_t *loc); +int pruss_request_tm_region (struct udevice *dev,phys_addr_t *loc);
Signature · 1 line
2.17.1
PATCH 8/14 arm: dts: k3-am65-main: Add msmc_ram node
In reply to KeerthyAdd msmc_ram node needed for prueth
--- arch/arm/dts/k3-am65-main.dtsi | 21 +++++++++++++++++++++ arch/arm/dts/k3-am65.dtsi | 4 +++- 2 files changed, 24 insertions(+), 1 deletion(-)
arch/arm/dts/k3-am65-main.dtsi +21 -0
@@ -6,6 +6,27 @@ */ &cbass_main { + msmc_ram: sram at 70000000 { + compatible = "mmio-sram"; + reg = <0x0 0x70000000 0x0 0x200000>; + #address-cells = <1>; + #size-cells = <1>; + ranges = <0x0 0x0 0x70000000 0x200000>; + u-boot,dm-spl; + + atf-sram at 0 { + reg = <0x0 0x20000>; + }; + + sysfw-sram at f0000 { + reg = <0xf0000 0x10000>; + }; + + l3cache-sram at 100000 { + reg = <0x100000 0x100000>; + }; + }; + gic500: interrupt-controller at 1800000 { compatible = "arm,gic-v3"; #address-cells = <2>;
arch/arm/dts/k3-am65.dtsi +3 -1
@@ -76,7 +76,9 @@ <0x00 0x42040000 0x00 0x42040000 0x00 0x03ac2400>, <0x00 0x45100000 0x00 0x45100000 0x00 0x00c24000>, <0x00 0x46000000 0x00 0x46000000 0x00 0x00200000>, - <0x00 0x47000000 0x00 0x47000000 0x00 0x00068400>; + <0x00 0x47000000 0x00 0x47000000 0x00 0x00068400>, + <0x07 0x00000000 0x07 0x00000000 0x01 0x0000000>, + <0x00 0x70000000 0x00 0x70000000 0x00 0x00200000>; cbass_mcu: interconnect at 28380000 { compatible = "simple-bus";
Signature · 1 line
2.17.1
PATCH 8/14 arm: dts: k3-am65-main: Add msmc_ram node
Quoted 45 lines · Keerthy
Add
msmc_ramnode needed for pruethSigned-off-by Keerthy <[email protected]>--- arch/arm/dts/k3-am65-main.dtsi | 21 +++++++++++++++++++++ arch/arm/dts/k3-am65.dtsi | 4 +++- 2 files changed, 24 insertions(+), 1 deletion(-)arch/arm/dts/k3-am65-main.dtsi +21
-0@@ -6,6 +6,27 @@ */ &cbass_main { + msmc_ram: sram at 70000000 { + compatible = "mmio-sram"; + reg = <0x0 0x70000000 0x0 0x200000>; + #address-cells = <1>; + #size-cells = <1>; + ranges = <0x0 0x0 0x70000000 0x200000>; + u-boot,dm-spl; + + atf-sram at 0 { + reg = <0x0 0x20000>; + }; + + sysfw-sram at f0000 { + reg = <0xf0000 0x10000>; + }; + + l3cache-sram at 100000 { + reg = <0x100000 0x100000>; + }; + }; + gic500: interrupt-controller at 1800000 { compatible = "arm,gic-v3"; #address-cells = <2>;arch/arm/dts/k3-am65.dtsi +0
-0
arch/arm/dts/k3-am65.dtsi line 80
Quoted 6 lines
arch/arm/dts/k3-am65.dtsi +2
-1@@ -76,7 +76,9 @@ <0x00 0x42040000 0x00 0x42040000 0x00 0x03ac2400>, <0x00 0x45100000 0x00 0x45100000 0x00 0x00c24000>, <0x00 0x46000000 0x00 0x46000000 0x00 0x00200000>, - <0x00 0x47000000 0x00 0x47000000 0x00 0x00068400>; + <0x00 0x47000000 0x00 0x47000000 0x00 0x00068400>, + <0x07 0x00000000 0x07 0x00000000 0x01 0x0000000>,
Should align the width of the last block of zeroes with the other numbers...
--
Andreas Dannenberg
Texas Instruments Inc
Quoted 11 lines
arch/arm/dts/k3-am65.dtsi +1
-0+ <0x00 0x70000000 0x00 0x70000000 0x00 0x00200000>; cbass_mcu: interconnect at 28380000 { compatible = "simple-bus";Signature · 6 lines
2.17.1 _______________________________________________ U-Boot mailing list U-Boot at lists.denx.de https://lists.denx.de/listinfo/u-boot
PATCH 9/14 arm: dts: k3-am654-base-board-u-boot: Add icssg specific msmc_ram carveout nodes
In reply to KeerthyAdd icssg specific msmc_ram carveout nodes
---
Carving out as there is no msmc-ram driver.
Suggestions welcome.
arch/arm/dts/k3-am654-base-board-u-boot.dtsi | 12 ++++++++++++ 1 file changed, 12 insertions(+)
arch/arm/dts/k3-am654-base-board-u-boot.dtsi +12 -0
@@ -361,3 +361,15 @@ &wkup_i2c0 { u-boot,dm-spl; }; + +&msmc_ram { + icssg0_sram: icssg0-sram at 40000 { + reg = <0x40000 0x10000>; + }; + icssg1_sram: icssg1-sram at 50000 { + reg = <0x50000 0x10000>; + }; + icssg2_sram: icssg2-sram at 60000 { + reg = <0x60000 0x10000>; + }; +};
Signature · 1 line
2.17.1
PATCH 10/14 arm: dts: k3-am65-main: Add scm_conf node
In reply to KeerthyAdd scm_conf node needed for prueth.
--- arch/arm/dts/k3-am65-main.dtsi | 8 ++++++++ 1 file changed, 8 insertions(+)
arch/arm/dts/k3-am65-main.dtsi +8 -0
@@ -50,6 +50,14 @@ }; }; + scm_conf: scm_conf at 100000 { + compatible = "syscon", "simple-mfd"; + reg = <0 0x00100000 0 0x1c000>; + #address-cells = <1>; + #size-cells = <1>; + ranges = <0x0 0x0 0x00100000 0x1c000>; + }; + secure_proxy_main: mailbox at 32c00000 { compatible = "ti,am654-secure-proxy"; #mbox-cells = <1>;
Signature · 1 line
2.17.1
PATCH 11/14 arm: dts: k3-am65-main: Add pruss nodes for ICSSG2
In reply to KeerthyAdd pruss nodes. Add nodes as in the 4.19 integration kernel.
--- arch/arm/dts/k3-am65-main.dtsi | 183 +++++++++++++++++++++++++++++++++ 1 file changed, 183 insertions(+)
arch/arm/dts/k3-am65-main.dtsi +183 -0
@@ -5,6 +5,8 @@ * Copyright (C) 2016-2018 Texas Instruments Incorporated - http://www.ti.com/ */ +#include <dt-bindings/dma/k3-udma.h> + &cbass_main { msmc_ram: sram at 70000000 { compatible = "mmio-sram";
@@ -172,4 +174,185 @@ clocks = <&k3_clks 113 1>; power-domains = <&k3_pds 113 TI_SCI_PD_EXCLUSIVE>; }; + + icssg_soc_bus2: pruss-soc-bus at b226004 { + compatible = "ti,am654-icssg-soc-bus", "simple-bus"; + reg = <0x00 0x0b226004 0x00 0x4>; + power-domains = <&k3_pds 64>; + #address-cells = <1>; + #size-cells = <1>; + ranges = <0x0b200000 0x00 0x0b200000 0x100000>; + dma-ranges; + + icssg2: icssg at b200000 { + compatible = "ti,am654-icssg"; + reg = <0xb200000 0x80000>; + interrupts = <GIC_SPI 270 IRQ_TYPE_LEVEL_HIGH>, + <GIC_SPI 271 IRQ_TYPE_LEVEL_HIGH>, + <GIC_SPI 272 IRQ_TYPE_LEVEL_HIGH>, + <GIC_SPI 273 IRQ_TYPE_LEVEL_HIGH>, + <GIC_SPI 274 IRQ_TYPE_LEVEL_HIGH>, + <GIC_SPI 275 IRQ_TYPE_LEVEL_HIGH>, + <GIC_SPI 276 IRQ_TYPE_LEVEL_HIGH>, + <GIC_SPI 277 IRQ_TYPE_LEVEL_HIGH>; + interrupt-names = "host2", "host3", "host4", + "host5", "host6", "host7", + "host8", "host9"; + #address-cells = <1>; + #size-cells = <1>; + ranges; + dma-ranges; + ti,psil-base = <0x4300>; /* ICSSG2 PSIL thread start */ + + ti,psil-config0 { + linux,udma-mode = <UDMA_PKT_MODE>; + statictr-type = <PSIL_STATIC_TR_NONE>; + ti,needs-epib; + ti,psd-size = <16>; + }; + + ti,psil-config1 { + linux,udma-mode = <UDMA_PKT_MODE>; + statictr-type = <PSIL_STATIC_TR_NONE>; + ti,needs-epib; + ti,psd-size = <16>; + }; + + ti,psil-config2 { + linux,udma-mode = <UDMA_PKT_MODE>; + statictr-type = <PSIL_STATIC_TR_NONE>; + ti,needs-epib; + ti,psd-size = <16>; + }; + + ti,psil-config3 { + linux,udma-mode = <UDMA_PKT_MODE>; + statictr-type = <PSIL_STATIC_TR_NONE>; + ti,needs-epib; + ti,psd-size = <16>; + }; + + ti,psil-config4 { + linux,udma-mode = <UDMA_PKT_MODE>; + statictr-type = <PSIL_STATIC_TR_NONE>; + ti,needs-epib; + ti,psd-size = <16>; + }; + + ti,psil-config5 { + linux,udma-mode = <UDMA_PKT_MODE>; + statictr-type = <PSIL_STATIC_TR_NONE>; + ti,needs-epib; + ti,psd-size = <16>; + }; + + ti,psil-config6 { + linux,udma-mode = <UDMA_PKT_MODE>; + statictr-type = <PSIL_STATIC_TR_NONE>; + ti,needs-epib; + ti,psd-size = <16>; + }; + + ti,psil-config7 { + linux,udma-mode = <UDMA_PKT_MODE>; + statictr-type = <PSIL_STATIC_TR_NONE>; + ti,needs-epib; + ti,psd-size = <16>; + }; + + icssg2_mem: memories at b200000 { + reg = <0xb200000 0x2000>, + <0xb202000 0x2000>, + <0xb210000 0x10000>; + reg-names = "dram0", "dram1", + "shrdram2"; + }; + + icssg2_cfg: cfg at b226000 { + compatible = "syscon"; + reg = <0xb226000 0x200>; + }; + + icssg2_iep: iep at b22e000 { + compatible = "syscon"; + reg = <0xb22e000 0x1000>; + }; + + icssg2_mii_rt: mii-rt at b232000 { + compatible = "syscon"; + reg = <0xb232000 0x100>; + }; + + icssg2_mii_g_rt: mii-g-rt at b233000 { + compatible = "syscon"; + reg = <0xb233000 0x1000>; + }; + + icssg2_intc: interrupt-controller at b220000 { + compatible = "ti,am654-icssg-intc"; + reg = <0xb220000 0x2000>; + interrupt-controller; + #interrupt-cells = <1>; + }; + + pru2_0: pru at b234000 { + compatible = "ti,am654-pru"; + reg = <0xb234000 0x4000>, + <0xb222000 0x100>, + <0xb222400 0x100>; + reg-names = "iram", "control", "debug"; + firmware-name = "am65x-pru2_0-fw"; + interrupt-parent = <&icssg2_intc>; + interrupts = <16>, <17>; + interrupt-names = "vring", "kick"; + }; + + rtu2_0: rtu at b204000 { + compatible = "ti,am654-rtu"; + reg = <0xb204000 0x2000>, + <0xb223000 0x100>, + <0xb223400 0x100>; + reg-names = "iram", "control", "debug"; + firmware-name = "am65x-rtu2_0-fw"; + interrupt-parent = <&icssg2_intc>; + interrupts = <20>, <21>; + interrupt-names = "vring", "kick"; + }; + + pru2_1: pru at b238000 { + compatible = "ti,am654-pru"; + reg = <0xb238000 0x4000>, + <0xb224000 0x100>, + <0xb224400 0x100>; + reg-names = "iram", "control", "debug"; + firmware-name = "am65x-pru2_1-fw"; + interrupt-parent = <&icssg2_intc>; + interrupts = <18>, <19>; + interrupt-names = "vring", "kick"; + }; + + rtu2_1: rtu at b206000 { + compatible = "ti,am654-rtu"; + reg = <0xb206000 0x2000>, + <0xb223800 0x100>, + <0xb223c00 0x100>; + reg-names = "iram", "control", "debug"; + firmware-name = "am65x-rtu2_1-fw"; + interrupt-parent = <&icssg2_intc>; + interrupts = <22>, <23>; + interrupt-names = "vring", "kick"; + }; + + icssg2_mdio: mdio at b232400 { + compatible = "ti,davinci_mdio"; + reg = <0xb232400 0x100>; + clocks = <&k3_clks 64 3>; + clock-names = "fck"; + #address-cells = <1>; + #size-cells = <0>; + bus_freq = <1000000>; + status = "disabled"; + }; + }; + }; };
Signature · 1 line
2.17.1
PATCH 11/14 arm: dts: k3-am65-main: Add pruss nodes for ICSSG2
Hi Keerthy,
Quoted 23 lines · Keerthy
Add pruss nodes. Add nodes as in the 4.19 integration kernel.
Signed-off-by Keerthy <[email protected]>--- arch/arm/dts/k3-am65-main.dtsi | 183 +++++++++++++++++++++++++++++++++ 1 file changed, 183 insertions(+)arch/arm/dts/k3-am65-main.dtsi +2
-0@@ -5,6 +5,8 @@ * Copyright (C) 2016-2018 Texas Instruments Incorporated - http://www.ti.com/ */ +#include <dt-bindings/dma/k3-udma.h> + &cbass_main { msmc_ram: sram at 70000000 { compatible = "mmio-sram";@@ -172,4 +174,185 @@ clocks = <&k3_clks 113 1>; power-domains = <&k3_pds 113 TI_SCI_PD_EXCLUSIVE>;
arch/arm/dts/k3-am65-main.dtsi line 185
Quoted 10 lines
arch/arm/dts/k3-am65-main.dtsi +9
-0}; + + icssg_soc_bus2: pruss-soc-bus at b226004 { + compatible = "ti,am654-icssg-soc-bus", "simple-bus"; + reg = <0x00 0x0b226004 0x00 0x4>; + power-domains = <&k3_pds 64>; + #address-cells = <1>; + #size-cells = <1>; + ranges = <0x0b200000 0x00 0x0b200000 0x100000>; + dma-ranges;
So, the SoC bus nodes are rejected on the kernel-side, and the overall PRUSS bindings are still a work in progress.
Quoted 5 lines
arch/arm/dts/k3-am65-main.dtsi +5
-0+ + icssg2: icssg at b200000 { + compatible = "ti,am654-icssg"; + reg = <0xb200000 0x80000>; + interrupts = <GIC_SPI 270 IRQ_TYPE_LEVEL_HIGH>,
arch/arm/dts/k3-am65-main.dtsi line 200
Quoted 10 lines
arch/arm/dts/k3-am65-main.dtsi +10
-0+ <GIC_SPI 271 IRQ_TYPE_LEVEL_HIGH>, + <GIC_SPI 272 IRQ_TYPE_LEVEL_HIGH>, + <GIC_SPI 273 IRQ_TYPE_LEVEL_HIGH>, + <GIC_SPI 274 IRQ_TYPE_LEVEL_HIGH>, + <GIC_SPI 275 IRQ_TYPE_LEVEL_HIGH>, + <GIC_SPI 276 IRQ_TYPE_LEVEL_HIGH>, + <GIC_SPI 277 IRQ_TYPE_LEVEL_HIGH>; + interrupt-names = "host2", "host3", "host4", + "host5", "host6", "host7", + "host8", "host9";
These are going away as well, and being moved into the INTC node.
Quoted 51 lines
arch/arm/dts/k3-am65-main.dtsi +51
-0+ #address-cells = <1>; + #size-cells = <1>; + ranges; + dma-ranges; + ti,psil-base = <0x4300>; /* ICSSG2 PSIL thread start */ + + ti,psil-config0 { + linux,udma-mode = <UDMA_PKT_MODE>; + statictr-type = <PSIL_STATIC_TR_NONE>; + ti,needs-epib; + ti,psd-size = <16>; + }; + + ti,psil-config1 { + linux,udma-mode = <UDMA_PKT_MODE>; + statictr-type = <PSIL_STATIC_TR_NONE>; + ti,needs-epib; + ti,psd-size = <16>; + }; + + ti,psil-config2 { + linux,udma-mode = <UDMA_PKT_MODE>; + statictr-type = <PSIL_STATIC_TR_NONE>; + ti,needs-epib; + ti,psd-size = <16>; + }; + + ti,psil-config3 { + linux,udma-mode = <UDMA_PKT_MODE>; + statictr-type = <PSIL_STATIC_TR_NONE>; + ti,needs-epib; + ti,psd-size = <16>; + }; + + ti,psil-config4 { + linux,udma-mode = <UDMA_PKT_MODE>; + statictr-type = <PSIL_STATIC_TR_NONE>; + ti,needs-epib; + ti,psd-size = <16>; + }; + + ti,psil-config5 { + linux,udma-mode = <UDMA_PKT_MODE>; + statictr-type = <PSIL_STATIC_TR_NONE>; + ti,needs-epib; + ti,psd-size = <16>; + }; + + ti,psil-config6 { + linux,udma-mode = <UDMA_PKT_MODE>; + statictr-type = <PSIL_STATIC_TR_NONE>;
arch/arm/dts/k3-am65-main.dtsi line 261
Quoted 10 lines
arch/arm/dts/k3-am65-main.dtsi +10
-0+ ti,needs-epib; + ti,psd-size = <16>; + }; + + ti,psil-config7 { + linux,udma-mode = <UDMA_PKT_MODE>; + statictr-type = <PSIL_STATIC_TR_NONE>; + ti,needs-epib; + ti,psd-size = <16>; + };
These are still up for debate. So, I think this whole series is a bit premature.
regards
Suman
Quoted 98 lines
arch/arm/dts/k3-am65-main.dtsi +96
-0+ + icssg2_mem: memories at b200000 { + reg = <0xb200000 0x2000>, + <0xb202000 0x2000>, + <0xb210000 0x10000>; + reg-names = "dram0", "dram1", + "shrdram2"; + }; + + icssg2_cfg: cfg at b226000 { + compatible = "syscon"; + reg = <0xb226000 0x200>; + }; + + icssg2_iep: iep at b22e000 { + compatible = "syscon"; + reg = <0xb22e000 0x1000>; + }; + + icssg2_mii_rt: mii-rt at b232000 { + compatible = "syscon"; + reg = <0xb232000 0x100>; + }; + + icssg2_mii_g_rt: mii-g-rt at b233000 { + compatible = "syscon"; + reg = <0xb233000 0x1000>; + }; + + icssg2_intc: interrupt-controller at b220000 { + compatible = "ti,am654-icssg-intc"; + reg = <0xb220000 0x2000>; + interrupt-controller; + #interrupt-cells = <1>; + }; + + pru2_0: pru at b234000 { + compatible = "ti,am654-pru"; + reg = <0xb234000 0x4000>, + <0xb222000 0x100>, + <0xb222400 0x100>; + reg-names = "iram", "control", "debug"; + firmware-name = "am65x-pru2_0-fw"; + interrupt-parent = <&icssg2_intc>; + interrupts = <16>, <17>; + interrupt-names = "vring", "kick"; + }; + + rtu2_0: rtu at b204000 { + compatible = "ti,am654-rtu"; + reg = <0xb204000 0x2000>, + <0xb223000 0x100>, + <0xb223400 0x100>; + reg-names = "iram", "control", "debug"; + firmware-name = "am65x-rtu2_0-fw"; + interrupt-parent = <&icssg2_intc>; + interrupts = <20>, <21>; + interrupt-names = "vring", "kick"; + }; + + pru2_1: pru at b238000 { + compatible = "ti,am654-pru"; + reg = <0xb238000 0x4000>, + <0xb224000 0x100>, + <0xb224400 0x100>; + reg-names = "iram", "control", "debug"; + firmware-name = "am65x-pru2_1-fw"; + interrupt-parent = <&icssg2_intc>; + interrupts = <18>, <19>; + interrupt-names = "vring", "kick"; + }; + + rtu2_1: rtu at b206000 { + compatible = "ti,am654-rtu"; + reg = <0xb206000 0x2000>, + <0xb223800 0x100>, + <0xb223c00 0x100>; + reg-names = "iram", "control", "debug"; + firmware-name = "am65x-rtu2_1-fw"; + interrupt-parent = <&icssg2_intc>; + interrupts = <22>, <23>; + interrupt-names = "vring", "kick"; + }; + + icssg2_mdio: mdio at b232400 { + compatible = "ti,davinci_mdio"; + reg = <0xb232400 0x100>; + clocks = <&k3_clks 64 3>; + clock-names = "fck"; + #address-cells = <1>; + #size-cells = <0>; + bus_freq = <1000000>; + status = "disabled"; + }; + }; + }; };
PATCH 12/14 arm64: dts: ti: am654-base-board: add ICSSG2 Ethernet support
In reply to KeerthyICSSG2 provide dual Gigabit Ethernet support. Currently mdio clock is part of this node and also the icssg2_rgmii_pins_default pinmux node has the mdio pins as there is no davinci mdio driver.
Currently icssg2 instances are supported. Either mii0 or mii1 can be enabled at a time.
--- arch/arm/dts/k3-am654-base-board-u-boot.dtsi | 118 +++++++++++++++++++ 1 file changed, 118 insertions(+)
arch/arm/dts/k3-am654-base-board-u-boot.dtsi +118 -0
@@ -16,6 +16,63 @@ serial2 = &main_uart0; ethernet0 = &cpsw_port1; }; + + /* Dual Ethernet application node on PRU-ICSSG2 */ + pruss2_eth: pruss2_eth { + compatible = "ti,am654-icssg-prueth"; + pinctrl-names = "default"; + pinctrl-0 = <&icssg2_rgmii_pins_default>; + sram = <&icssg2_sram>; + clocks = <&k3_clks 64 3>; + clock-names = "mdio_fck"; + u-boot,dm-spl; + + prus = <&pru2_0>, <&rtu2_0>, <&pru2_1>, <&rtu2_1>; + firmware-name = "ti-pruss/am65x-pru0-prueth-fw.elf", + "ti-pruss/am65x-rtu0-prueth-fw.elf", + "ti-pruss/am65x-pru1-prueth-fw.elf", + "ti-pruss/am65x-rtu1-prueth-fw.elf"; + mii-g-rt = <&icssg2_mii_g_rt>; + dma-coherent; + dmas = <&mcu_udmap &icssg2 0 UDMA_DIR_TX>, /* egress slice 0 */ + <&mcu_udmap &icssg2 1 UDMA_DIR_TX>, /* egress slice 0 */ + <&mcu_udmap &icssg2 2 UDMA_DIR_TX>, /* egress slice 0 */ + <&mcu_udmap &icssg2 3 UDMA_DIR_TX>, /* mgmnt cmd slice 0 */ + <&mcu_udmap &icssg2 4 UDMA_DIR_TX>, /* egress slice 1 */ + <&mcu_udmap &icssg2 5 UDMA_DIR_TX>, /* egress slice 1 */ + <&mcu_udmap &icssg2 6 UDMA_DIR_TX>, /* egress slice 1 */ + <&mcu_udmap &icssg2 7 UDMA_DIR_TX>, /* mgmnt cmd slice 1 */ + + <&mcu_udmap &icssg2 0 UDMA_DIR_RX>, /* ingress slice 0 */ + <&mcu_udmap &icssg2 1 UDMA_DIR_RX>, /* ingress slice 1 */ + <&mcu_udmap &icssg2 2 UDMA_DIR_RX>, /* mgmnt rsp slice 0 */ + <&mcu_udmap &icssg2 3 UDMA_DIR_RX>; /* mgmnt rsp slice 1 */ + dma-names = "tx0-0", "tx0-1", "tx0-2", "tx0-3", + "tx1-0", "tx1-1", "tx1-2", "tx1-3", + "rx0", "rx1", + "rxmgm0", "rxmgm1"; + + pruss2_emac0: ethernet-mii0 { + phy-handle = <&pruss2_eth0_phy>; + phy-mode = "rgmii-id"; + syscon-rgmii-delay = <&scm_conf 0x4120>; + /* Filled in by bootloader */ + local-mac-address = [00 00 00 00 00 00]; + }; + +/* + * Commenting out the second mii interface as the framework + * supports one interface in a single probe + * So either mii1 or mii2 can be used. In case mii1 is needed + * uncomment mii1 and comment out mii0 + pruss2_emac1: ethernet-mii1 { + phy-handle = <&pruss2_eth1_phy>; + phy-mode = "rgmii-id"; + syscon-rgmii-delay = <&scm_conf 0x4124>; + local-mac-address = [00 00 00 00 00 00]; + }; +*/ + }; }; &cbass_main{
@@ -274,6 +331,47 @@ u-boot,dm-spl; }; + icssg2_rgmii_pins_default: icssg2_rgmii_pins_default { + pinctrl-single,pins = < + AM65X_IOPAD(0x00ac, PIN_INPUT , 2) /* (AH15) PRG2_PRU1_GPO0.PRG2_RGMII2_RD0 */ + AM65X_IOPAD(0x00b0, PIN_INPUT , 2) /* (AC16) PRG2_PRU1_GPO1.PRG2_RGMII2_RD1 */ + AM65X_IOPAD(0x00b4, PIN_INPUT , 2) /* (AD17) PRG2_PRU1_GPO2.PRG2_RGMII2_RD2 */ + AM65X_IOPAD(0x00b8, PIN_INPUT , 2) /* (AH14) PRG2_PRU1_GPO3.PRG2_RGMII2_RD3 */ + AM65X_IOPAD(0x00cc, PIN_OUTPUT , 2) /* (AD15) PRG2_PRU1_GPO8.PRG2_RGMII2_TD0 */ + AM65X_IOPAD(0x00d0, PIN_OUTPUT , 2) /* (AF14) PRG2_PRU1_GPO9.PRG2_RGMII2_TD1 */ + AM65X_IOPAD(0x00d4, PIN_OUTPUT , 2) /* (AC15) PRG2_PRU1_GPO10.PRG2_RGMII2_TD2 */ + AM65X_IOPAD(0x00d8, PIN_OUTPUT , 2) /* (AD14) PRG2_PRU1_GPO11.PRG2_RGMII2_TD3 */ + AM65X_IOPAD(0x00dc, PIN_INPUT , 2) /* (AE14) PRG2_PRU1_GPO16.PRG2_RGMII2_TXC */ + AM65X_IOPAD(0x00c4, PIN_OUTPUT , 2) /* (AC17) PRG2_PRU1_GPO6.PRG2_RGMII2_TX_CTL */ + AM65X_IOPAD(0x00c0, PIN_INPUT , 2) /* (AG15) PRG2_PRU1_GPO5.PRG2_RGMII2_RXC */ + AM65X_IOPAD(0x00bc, PIN_INPUT , 2) /* (AG14) PRG2_PRU1_GPO4.PRG2_RGMII2_RX_CTL */ + + AM65X_IOPAD(0x0078, PIN_INPUT , 2) /* (AF18) PRG2_PRU0_GPO0.PRG2_RGMII1_RD0 */ + AM65X_IOPAD(0x007c, PIN_INPUT , 2) /* (AE18) PRG2_PRU0_GPO1.PRG2_RGMII1_RD1 */ + AM65X_IOPAD(0x0080, PIN_INPUT , 2) /* (AH17) PRG2_PRU0_GPO2.PRG2_RGMII1_RD2 */ + AM65X_IOPAD(0x0084, PIN_INPUT , 2) /* (AG18) PRG2_PRU0_GPO3.PRG2_RGMII1_RD3 */ + AM65X_IOPAD(0x0098, PIN_OUTPUT , 2) /* (AH16) PRG2_PRU0_GPO8.PRG2_RGMII1_TD0 */ + AM65X_IOPAD(0x009c, PIN_OUTPUT , 2) /* (AG16) PRG2_PRU0_GPO9.PRG2_RGMII1_TD1 */ + AM65X_IOPAD(0x00a0, PIN_OUTPUT , 2) /* (AF16) PRG2_PRU0_GPO10.PRG2_RGMII1_TD2 */ + AM65X_IOPAD(0x00a4, PIN_OUTPUT , 2) /* (AE16) PRG2_PRU0_GPO11.PRG2_RGMII1_TD3 */ + AM65X_IOPAD(0x00a8, PIN_INPUT , 2) /* (AD16) PRG2_PRU0_GPO16.PRG2_RGMII1_TXC */ + AM65X_IOPAD(0x0090, PIN_OUTPUT , 2) /* (AE17) PRG2_PRU0_GPO6.PRG2_RGMII1_TX_CTL */ + AM65X_IOPAD(0x008c, PIN_INPUT , 2) /* (AF17) PRG2_PRU0_GPO5.PRG2_RGMII1_RXC */ + AM65X_IOPAD(0x0088, PIN_INPUT , 2) /* (AG17) PRG2_PRU0_GPO4.PRG2_RGMII1_RX_CTL */ + /* HACK to get MDIO Pins in the right state */ + AM65X_IOPAD(0x0094, PIN_INPUT , 2) /* (AC19) PRG2_PRU0_GPO7.PRG2_MDIO0_MDIO */ + AM65X_IOPAD(0x00c8, PIN_OUTPUT , 2) /* (AE15) PRG2_PRU1_GPO7.PRG2_MDIO0_MDC */ + >; + u-boot,dm-spl; + }; + + icssg2_mdio_pins_default: icssg2_mdio_pins_default { + pinctrl-single,pins = < + AM65X_IOPAD(0x0094, PIN_INPUT , 2) /* (AC19) PRG2_PRU0_GPO7.PRG2_MDIO0_MDIO */ + AM65X_IOPAD(0x00c8, PIN_OUTPUT , 2) /* (AE15) PRG2_PRU1_GPO7.PRG2_MDIO0_MDC */ + >; + u-boot,dm-spl; + }; }; &main_pmx1 {
@@ -373,3 +471,23 @@ reg = <0x60000 0x10000>; }; }; + +&icssg2_mdio{ + status = "okay"; + pinctrl-names = "default"; + pinctrl-0 = <&icssg2_mdio_pins_default>; + + pruss2_eth0_phy: ethernet-phy at 0 { + reg = <0>; + ti,rx-internal-delay = <DP83867_RGMIIDCTL_2_00_NS>; + ti,tx-internal-delay = <DP83867_RGMIIDCTL_2_00_NS>; + ti,fifo-depth = <DP83867_PHYCR_FIFO_DEPTH_4_B_NIB>; + }; + + pruss2_eth1_phy: ethernet-phy at 3 { + reg = <3>; + ti,rx-internal-delay = <DP83867_RGMIIDCTL_2_00_NS>; + ti,tx-internal-delay = <DP83867_RGMIIDCTL_2_00_NS>; + ti,fifo-depth = <DP83867_PHYCR_FIFO_DEPTH_4_B_NIB>; + }; +};
Signature · 1 line
2.17.1
PATCH 12/14 arm64: dts: ti: am654-base-board: add ICSSG2 Ethernet support
Quoted 82 lines · Keerthy
ICSSG2 provide dual Gigabit Ethernet support. Currently mdio clock is part of this node and also the
icssg2_rgmii_pins_defaultpinmux node has the mdio pins as there is no davinci mdio driver.Currently icssg2 instances are supported. Either mii0 or mii1 can be enabled at a time.
Signed-off-by Keerthy <[email protected]>--- arch/arm/dts/k3-am654-base-board-u-boot.dtsi | 118 +++++++++++++++++++ 1 file changed, 118 insertions(+)arch/arm/dts/k3-am654-base-board-u-boot.dtsi +57
-0@@ -16,6 +16,63 @@ serial2 = &main_uart0; ethernet0 = &cpsw_port1; }; + + /* Dual Ethernet application node on PRU-ICSSG2 */ + pruss2_eth: pruss2_eth { + compatible = "ti,am654-icssg-prueth"; + pinctrl-names = "default"; + pinctrl-0 = <&icssg2_rgmii_pins_default>; + sram = <&icssg2_sram>; + clocks = <&k3_clks 64 3>; + clock-names = "mdio_fck"; + u-boot,dm-spl; + + prus = <&pru2_0>, <&rtu2_0>, <&pru2_1>, <&rtu2_1>; + firmware-name = "ti-pruss/am65x-pru0-prueth-fw.elf", + "ti-pruss/am65x-rtu0-prueth-fw.elf", + "ti-pruss/am65x-pru1-prueth-fw.elf", + "ti-pruss/am65x-rtu1-prueth-fw.elf"; + mii-g-rt = <&icssg2_mii_g_rt>; + dma-coherent; + dmas = <&mcu_udmap &icssg2 0 UDMA_DIR_TX>, /* egress slice 0 */ + <&mcu_udmap &icssg2 1 UDMA_DIR_TX>, /* egress slice 0 */ + <&mcu_udmap &icssg2 2 UDMA_DIR_TX>, /* egress slice 0 */ + <&mcu_udmap &icssg2 3 UDMA_DIR_TX>, /* mgmnt cmd slice 0 */ + <&mcu_udmap &icssg2 4 UDMA_DIR_TX>, /* egress slice 1 */ + <&mcu_udmap &icssg2 5 UDMA_DIR_TX>, /* egress slice 1 */ + <&mcu_udmap &icssg2 6 UDMA_DIR_TX>, /* egress slice 1 */ + <&mcu_udmap &icssg2 7 UDMA_DIR_TX>, /* mgmnt cmd slice 1 */ + + <&mcu_udmap &icssg2 0 UDMA_DIR_RX>, /* ingress slice 0 */ + <&mcu_udmap &icssg2 1 UDMA_DIR_RX>, /* ingress slice 1 */ + <&mcu_udmap &icssg2 2 UDMA_DIR_RX>, /* mgmnt rsp slice 0 */ + <&mcu_udmap &icssg2 3 UDMA_DIR_RX>; /* mgmnt rsp slice 1 */ + dma-names = "tx0-0", "tx0-1", "tx0-2", "tx0-3", + "tx1-0", "tx1-1", "tx1-2", "tx1-3", + "rx0", "rx1", + "rxmgm0", "rxmgm1"; + + pruss2_emac0: ethernet-mii0 { + phy-handle = <&pruss2_eth0_phy>; + phy-mode = "rgmii-id"; + syscon-rgmii-delay = <&scm_conf 0x4120>; + /* Filled in by bootloader */ + local-mac-address = [00 00 00 00 00 00]; + }; + +/* + * Commenting out the second mii interface as the framework + * supports one interface in a single probe + * So either mii1 or mii2 can be used. In case mii1 is needed + * uncomment mii1 and comment out mii0 + pruss2_emac1: ethernet-mii1 { + phy-handle = <&pruss2_eth1_phy>; + phy-mode = "rgmii-id"; + syscon-rgmii-delay = <&scm_conf 0x4124>; + local-mac-address = [00 00 00 00 00 00]; + }; +*/ + }; }; &cbass_main{
arch/arm/dts/k3-am654-base-board-u-boot.dtsi line 336
Quoted 6 lines
arch/arm/dts/k3-am654-base-board-u-boot.dtsi +3
-0@@ -274,6 +331,47 @@ u-boot,dm-spl; }; + icssg2_rgmii_pins_default: icssg2_rgmii_pins_default { + pinctrl-single,pins = < + AM65X_IOPAD(0x00ac, PIN_INPUT , 2) /* (AH15) PRG2_PRU1_GPO0.PRG2_RGMII2_RD0 */
Why are there spaces behind the pin direction definitions (PIN_INPUT[space], ...)? I've never seen that before. The pinmix definitions are supposed to be a 1:1 copy from our AM64x pinmux tool for consistency sake. Is the pinmux tool broken?
--
Andreas Dannenberg
Texas Instruments Inc
Quoted 72 lines
arch/arm/dts/k3-am654-base-board-u-boot.dtsi +58
-0+ AM65X_IOPAD(0x00b0, PIN_INPUT , 2) /* (AC16) PRG2_PRU1_GPO1.PRG2_RGMII2_RD1 */ + AM65X_IOPAD(0x00b4, PIN_INPUT , 2) /* (AD17) PRG2_PRU1_GPO2.PRG2_RGMII2_RD2 */ + AM65X_IOPAD(0x00b8, PIN_INPUT , 2) /* (AH14) PRG2_PRU1_GPO3.PRG2_RGMII2_RD3 */ + AM65X_IOPAD(0x00cc, PIN_OUTPUT , 2) /* (AD15) PRG2_PRU1_GPO8.PRG2_RGMII2_TD0 */ + AM65X_IOPAD(0x00d0, PIN_OUTPUT , 2) /* (AF14) PRG2_PRU1_GPO9.PRG2_RGMII2_TD1 */ + AM65X_IOPAD(0x00d4, PIN_OUTPUT , 2) /* (AC15) PRG2_PRU1_GPO10.PRG2_RGMII2_TD2 */ + AM65X_IOPAD(0x00d8, PIN_OUTPUT , 2) /* (AD14) PRG2_PRU1_GPO11.PRG2_RGMII2_TD3 */ + AM65X_IOPAD(0x00dc, PIN_INPUT , 2) /* (AE14) PRG2_PRU1_GPO16.PRG2_RGMII2_TXC */ + AM65X_IOPAD(0x00c4, PIN_OUTPUT , 2) /* (AC17) PRG2_PRU1_GPO6.PRG2_RGMII2_TX_CTL */ + AM65X_IOPAD(0x00c0, PIN_INPUT , 2) /* (AG15) PRG2_PRU1_GPO5.PRG2_RGMII2_RXC */ + AM65X_IOPAD(0x00bc, PIN_INPUT , 2) /* (AG14) PRG2_PRU1_GPO4.PRG2_RGMII2_RX_CTL */ + + AM65X_IOPAD(0x0078, PIN_INPUT , 2) /* (AF18) PRG2_PRU0_GPO0.PRG2_RGMII1_RD0 */ + AM65X_IOPAD(0x007c, PIN_INPUT , 2) /* (AE18) PRG2_PRU0_GPO1.PRG2_RGMII1_RD1 */ + AM65X_IOPAD(0x0080, PIN_INPUT , 2) /* (AH17) PRG2_PRU0_GPO2.PRG2_RGMII1_RD2 */ + AM65X_IOPAD(0x0084, PIN_INPUT , 2) /* (AG18) PRG2_PRU0_GPO3.PRG2_RGMII1_RD3 */ + AM65X_IOPAD(0x0098, PIN_OUTPUT , 2) /* (AH16) PRG2_PRU0_GPO8.PRG2_RGMII1_TD0 */ + AM65X_IOPAD(0x009c, PIN_OUTPUT , 2) /* (AG16) PRG2_PRU0_GPO9.PRG2_RGMII1_TD1 */ + AM65X_IOPAD(0x00a0, PIN_OUTPUT , 2) /* (AF16) PRG2_PRU0_GPO10.PRG2_RGMII1_TD2 */ + AM65X_IOPAD(0x00a4, PIN_OUTPUT , 2) /* (AE16) PRG2_PRU0_GPO11.PRG2_RGMII1_TD3 */ + AM65X_IOPAD(0x00a8, PIN_INPUT , 2) /* (AD16) PRG2_PRU0_GPO16.PRG2_RGMII1_TXC */ + AM65X_IOPAD(0x0090, PIN_OUTPUT , 2) /* (AE17) PRG2_PRU0_GPO6.PRG2_RGMII1_TX_CTL */ + AM65X_IOPAD(0x008c, PIN_INPUT , 2) /* (AF17) PRG2_PRU0_GPO5.PRG2_RGMII1_RXC */ + AM65X_IOPAD(0x0088, PIN_INPUT , 2) /* (AG17) PRG2_PRU0_GPO4.PRG2_RGMII1_RX_CTL */ + /* HACK to get MDIO Pins in the right state */ + AM65X_IOPAD(0x0094, PIN_INPUT , 2) /* (AC19) PRG2_PRU0_GPO7.PRG2_MDIO0_MDIO */ + AM65X_IOPAD(0x00c8, PIN_OUTPUT , 2) /* (AE15) PRG2_PRU1_GPO7.PRG2_MDIO0_MDC */ + >; + u-boot,dm-spl; + }; + + icssg2_mdio_pins_default: icssg2_mdio_pins_default { + pinctrl-single,pins = < + AM65X_IOPAD(0x0094, PIN_INPUT , 2) /* (AC19) PRG2_PRU0_GPO7.PRG2_MDIO0_MDIO */ + AM65X_IOPAD(0x00c8, PIN_OUTPUT , 2) /* (AE15) PRG2_PRU1_GPO7.PRG2_MDIO0_MDC */ + >; + u-boot,dm-spl; + }; }; &main_pmx1 {@@ -373,3 +471,23 @@ reg = <0x60000 0x10000>; }; }; + +&icssg2_mdio{ + status = "okay"; + pinctrl-names = "default"; + pinctrl-0 = <&icssg2_mdio_pins_default>; + + pruss2_eth0_phy: ethernet-phy at 0 { + reg = <0>; + ti,rx-internal-delay = <DP83867_RGMIIDCTL_2_00_NS>; + ti,tx-internal-delay = <DP83867_RGMIIDCTL_2_00_NS>; + ti,fifo-depth = <DP83867_PHYCR_FIFO_DEPTH_4_B_NIB>; + }; + + pruss2_eth1_phy: ethernet-phy at 3 { + reg = <3>; + ti,rx-internal-delay = <DP83867_RGMIIDCTL_2_00_NS>; + ti,tx-internal-delay = <DP83867_RGMIIDCTL_2_00_NS>; + ti,fifo-depth = <DP83867_PHYCR_FIFO_DEPTH_4_B_NIB>; + }; +};Signature · 6 lines
2.17.1 _______________________________________________ U-Boot mailing list U-Boot at lists.denx.de https://lists.denx.de/listinfo/u-boot
PATCH 13/14 configs: am65x_evm_a53_defconfig: Enable CONFIG_REMOTEPROC_TI_PRU
In reply to KeerthyEnable CONFIG_REMOTEPROC_TI_PRU and related configs. With that enable ICSSG_PRUETH config.
--- configs/am65x_evm_a53_defconfig | 4 ++++ 1 file changed, 4 insertions(+)
configs/am65x_evm_a53_defconfig +4 -0
@@ -48,6 +48,7 @@ CONFIG_DM=y CONFIG_SPL_DM=y CONFIG_SPL_DM_SEQ_ALIAS=y CONFIG_REGMAP=y +CONFIG_SYSCON=y CONFIG_SPL_REGMAP=y CONFIG_CLK=y CONFIG_SPL_CLK=y
@@ -69,6 +70,7 @@ CONFIG_PHY_TI=y CONFIG_PHY_FIXED=y CONFIG_DM_ETH=y CONFIG_TI_AM65_CPSW_NUSS=y +CONFIG_TI_AM64_ICSSG_PRUETH=y CONFIG_PHY=y CONFIG_PINCTRL=y # CONFIG_PINCTRL_GENERIC is not set
@@ -77,10 +79,12 @@ CONFIG_SPL_PINCTRL=y CONFIG_PINCTRL_SINGLE=y CONFIG_POWER_DOMAIN=y CONFIG_TI_SCI_POWER_DOMAIN=y +CONFIG_REMOTEPROC_TI_PRU=y CONFIG_DM_RESET=y CONFIG_RESET_TI_SCI=y CONFIG_DM_SERIAL=y CONFIG_SOC_TI=y +CONFIG_TI_PRUSS=y CONFIG_SYSRESET=y CONFIG_SPL_SYSRESET=y CONFIG_SYSRESET_TI_SCI=y
Signature · 1 line
2.17.1
PATCH 14/14 configs: am65x_evm_a53_defconfig: Enable CONFIG_CMD_REMOTEPROC
In reply to KeerthyEnable CONFIG_CMD_REMOTEPROC needed to load prueth firmwares.
--- configs/am65x_evm_a53_defconfig | 1 + 1 file changed, 1 insertion(+)
configs/am65x_evm_a53_defconfig +1 -0
@@ -34,6 +34,7 @@ CONFIG_CMD_ASKENV=y CONFIG_CMD_GPT=y CONFIG_CMD_I2C=y CONFIG_CMD_MMC=y +CONFIG_CMD_REMOTEPROC=y # CONFIG_CMD_SETEXPR is not set CONFIG_CMD_TIME=y # CONFIG_ISO_PARTITION is not set
Signature · 1 line
2.17.1
cover net: ti: icssg: Add prueth support
In reply to KeerthyKeerthy,
Quoted 5 lines · Keerthy
The series adds support for
icssg_pruethfunctionality on u-boot. This series is based on top of master branch. rproc init needs to be done from uboot command prompt. The pru/rtu firmware loading is done by prueth driver soon after config paramters are setup.
for everybody's benefit, where does this firmware come from so one can experiment with this patch series?
--
Andreas Dannenberg
Texas Instruments Inc
Quoted 66 lines
Currently only slice0/1 of icssg2 instance on am6-evm is supported. i.e Both slices of icssg2 instance are supported.
On u-boot prompt following commands to test
icssg2_port0 on am654-evm:setenv ethact
pruss2_eth; setenv serverip 172.24.191.45; fatload mmc 1 ${pru0loadaddr} am65x-pru0-prueth-fw.elf; fatload mmc 1 ${rtu0loadaddr} am65x-rtu0-prueth-fw.elf; rproc init; setenv autoload no; dhcp; tftp0x82000000Image; tftp0x83000000k3-am654-base-board.dtb; booti0x82000000-0x83000000This tests tftp on prueth.
Note: Uboot ethernet driver architecture supports once instance per probe. So only one of the ports are supported per instance. So DT of prueth node should have either ethernet-mii0 or ethernet-mii1.
Keerthy (14):
net: eth-uclass:eth_get_devbased onSEQ_ALIASinstead of probe order
net: eth-uclass: call stop only for active devices
misc: uclass: Introducemisc_init_by_ofnode
soc: ti: pruss: add a misc driver for PRUSS in TI SoCs
remoteproc: pruss: add PRU remoteproc driver
net: ti: icssg-prueth: Add ICSSG ethernet driver
net: ti: icssg-prueth: Workaround to shutdown the prueth firmware
arm: dts: k3-am65-main: Addmsmc_ramnode
arm: dts: k3-am654-base-board-u-boot: Add icssg specificmsmc_ram
carveout nodes
arm: dts: k3-am65-main: Addscm_confnode
arm: dts: k3-am65-main: Add pruss nodes for ICSSG2
arm64: dts: ti: am654-base-board: add ICSSG2 Ethernet support
configs:am65x_evm_a53_defconfig: EnableCONFIG_REMOTEPROC_TI_PRU
configs:am65x_evm_a53_defconfig: EnableCONFIG_CMD_REMOTEPROCarch/arm/dts/k3-am65-main.dtsi | 212 ++++++++ arch/arm/dts/k3-am65.dtsi | 4 +- arch/arm/dts/k3-am654-base-board-u-boot.dtsi | 130 +++++ configs/am65x_evm_a53_defconfig | 5 + drivers/misc/misc-uclass.c | 25 + drivers/net/ti/Kconfig | 8 + drivers/net/ti/Makefile | 1 + drivers/net/ti/icssg-prueth.c | 525 +++++++++++++++++++ drivers/net/ti/icssg.h | 31 ++ drivers/net/ti/icssg_classifier.c | 397 ++++++++++++++ drivers/remoteproc/Kconfig | 11 + drivers/remoteproc/Makefile | 1 + drivers/remoteproc/pru_rproc.c | 384 ++++++++++++++ drivers/soc/ti/Kconfig | 13 + drivers/soc/ti/Makefile | 1 + drivers/soc/ti/pruss.c | 143 +++++ include/misc.h | 9 + include/ti-pruss.h | 13 + net/eth-uclass.c | 7 +- 19 files changed, 1916 insertions(+), 4 deletions(-) create mode 100644 drivers/net/ti/icssg-prueth.c create mode 100644 drivers/net/ti/icssg.h create mode 100644 drivers/net/ti/icssg_classifier.c create mode 100644 drivers/remoteproc/pru_rproc.c create mode 100644 drivers/soc/ti/pruss.c create mode 100644 include/ti-pruss.hSignature · 6 lines
2.17.1 _______________________________________________ U-Boot mailing list U-Boot at lists.denx.de https://lists.denx.de/listinfo/u-boot
cover net: ti: icssg: Add prueth support
Andreas,
Quoted 12 lines · Andreas Dannenberg
Keerthy,
Quoted 5 lines · Keerthy
The series adds support for
icssg_pruethfunctionality on u-boot. This series is based on top of master branch. rproc init needs to be done from uboot command prompt. The pru/rtu firmware loading is done by prueth driver soon after config paramters are setup.for everybody's benefit, where does this firmware come from so one can experiment with this patch series?
From here
http://git.yoctoproject.org/cgit/cgit.cgi/meta-ti/tree/recipes-bsp/prueth-fw/prueth-fw-am65x_git.bb
http://git.yoctoproject.org/cgit/cgit.cgi/meta-ti/tree/recipes-bsp/emac-lld/emac-lld.inc
http://git.ti.com/keystone-rtos/emac-lld/trees/master/firmware/icss_dualmac/src
cheers,
-roger
Quoted 70 lines
--
Andreas Dannenberg
Texas Instruments IncQuoted 66 lines
Currently only slice0/1 of icssg2 instance on am6-evm is supported. i.e Both slices of icssg2 instance are supported.
On u-boot prompt following commands to test
icssg2_port0 on am654-evm:setenv ethact
pruss2_eth; setenv serverip 172.24.191.45; fatload mmc 1 ${pru0loadaddr} am65x-pru0-prueth-fw.elf; fatload mmc 1 ${rtu0loadaddr} am65x-rtu0-prueth-fw.elf; rproc init; setenv autoload no; dhcp; tftp0x82000000Image; tftp0x83000000k3-am654-base-board.dtb; booti0x82000000-0x83000000This tests tftp on prueth.
Note: Uboot ethernet driver architecture supports once instance per probe. So only one of the ports are supported per instance. So DT of prueth node should have either ethernet-mii0 or ethernet-mii1.
Keerthy (14):
net: eth-uclass:eth_get_devbased onSEQ_ALIASinstead of probe order
net: eth-uclass: call stop only for active devices
misc: uclass: Introducemisc_init_by_ofnode
soc: ti: pruss: add a misc driver for PRUSS in TI SoCs
remoteproc: pruss: add PRU remoteproc driver
net: ti: icssg-prueth: Add ICSSG ethernet driver
net: ti: icssg-prueth: Workaround to shutdown the prueth firmware
arm: dts: k3-am65-main: Addmsmc_ramnode
arm: dts: k3-am654-base-board-u-boot: Add icssg specificmsmc_ram
carveout nodes
arm: dts: k3-am65-main: Addscm_confnode
arm: dts: k3-am65-main: Add pruss nodes for ICSSG2
arm64: dts: ti: am654-base-board: add ICSSG2 Ethernet support
configs:am65x_evm_a53_defconfig: EnableCONFIG_REMOTEPROC_TI_PRU
configs:am65x_evm_a53_defconfig: EnableCONFIG_CMD_REMOTEPROCarch/arm/dts/k3-am65-main.dtsi | 212 ++++++++ arch/arm/dts/k3-am65.dtsi | 4 +- arch/arm/dts/k3-am654-base-board-u-boot.dtsi | 130 +++++ configs/am65x_evm_a53_defconfig | 5 + drivers/misc/misc-uclass.c | 25 + drivers/net/ti/Kconfig | 8 + drivers/net/ti/Makefile | 1 + drivers/net/ti/icssg-prueth.c | 525 +++++++++++++++++++ drivers/net/ti/icssg.h | 31 ++ drivers/net/ti/icssg_classifier.c | 397 ++++++++++++++ drivers/remoteproc/Kconfig | 11 + drivers/remoteproc/Makefile | 1 + drivers/remoteproc/pru_rproc.c | 384 ++++++++++++++ drivers/soc/ti/Kconfig | 13 + drivers/soc/ti/Makefile | 1 + drivers/soc/ti/pruss.c | 143 +++++ include/misc.h | 9 + include/ti-pruss.h | 13 + net/eth-uclass.c | 7 +- 19 files changed, 1916 insertions(+), 4 deletions(-) create mode 100644 drivers/net/ti/icssg-prueth.c create mode 100644 drivers/net/ti/icssg.h create mode 100644 drivers/net/ti/icssg_classifier.c create mode 100644 drivers/remoteproc/pru_rproc.c create mode 100644 drivers/soc/ti/pruss.c create mode 100644 include/ti-pruss.hSignature · 6 lines
2.17.1 _______________________________________________ U-Boot mailing list U-Boot at lists.denx.de https://lists.denx.de/listinfo/u-boot
Signature · 2 lines
Texas Instruments Finland Oy, Porkkalankatu 22, 00180 Helsinki. Y-tunnus/Business ID: 0615521-4. Kotipaikka/Domicile: Helsinki
cover net: ti: icssg: Add prueth support
Quoted 12 lines · Roger Quadros
Andreas,
Quoted 9 lines · Andreas Dannenberg
Keerthy,
Quoted 5 lines · Keerthy
The series adds support for
icssg_pruethfunctionality on u-boot. This series is based on top of master branch. rproc init needs to be done from uboot command prompt. The pru/rtu firmware loading is done by prueth driver soon after config paramters are setup.
Quoted 8 lines
Quoted 3 lines
for everybody's benefit, where does this firmware come from so one can experiment with this patch series?
From here
http://git.yoctoproject.org/cgit/cgit.cgi/meta-ti/tree/recipes-bsp/prueth-fw/prueth-fw-am65x_git.bb
http://git.yoctoproject.org/cgit/cgit.cgi/meta-ti/tree/recipes-bsp/emac-lld/emac-lld.inchttp://git.ti.com/keystone-rtos/emac-lld/trees/master/firmware/icss_dualmac/src
Prebuilt firmware can be obtained from AM65 procSDK [1] rootfs at /lib/firmware/ti-pruss/am65x*.elf
[1] http://software-dl.ti.com/processor-sdk-linux/esd/AM65X/latest/index_FDS.html
cheers,
-roger
Quoted 72 lines
Quoted 71 lines
--
Andreas Dannenberg
Texas Instruments IncQuoted 66 lines
Currently only slice0/1 of icssg2 instance on am6-evm is supported. i.e Both slices of icssg2 instance are supported.
On u-boot prompt following commands to test
icssg2_port0 on am654-evm:setenv ethact
pruss2_eth; setenv serverip 172.24.191.45; fatload mmc 1 ${pru0loadaddr} am65x-pru0-prueth-fw.elf; fatload mmc 1 ${rtu0loadaddr} am65x-rtu0-prueth-fw.elf; rproc init; setenv autoload no; dhcp; tftp0x82000000Image; tftp0x83000000k3-am654-base-board.dtb; booti0x82000000-0x83000000This tests tftp on prueth.
Note: Uboot ethernet driver architecture supports once instance per probe. So only one of the ports are supported per instance. So DT of prueth node should have either ethernet-mii0 or ethernet-mii1.
Keerthy (14):
net: eth-uclass:eth_get_devbased onSEQ_ALIASinstead of probe order
net: eth-uclass: call stop only for active devices
misc: uclass: Introducemisc_init_by_ofnode
soc: ti: pruss: add a misc driver for PRUSS in TI SoCs
remoteproc: pruss: add PRU remoteproc driver
net: ti: icssg-prueth: Add ICSSG ethernet driver
net: ti: icssg-prueth: Workaround to shutdown the prueth firmware
arm: dts: k3-am65-main: Addmsmc_ramnode
arm: dts: k3-am654-base-board-u-boot: Add icssg specificmsmc_ram
carveout nodes
arm: dts: k3-am65-main: Addscm_confnode
arm: dts: k3-am65-main: Add pruss nodes for ICSSG2
arm64: dts: ti: am654-base-board: add ICSSG2 Ethernet support
configs:am65x_evm_a53_defconfig: EnableCONFIG_REMOTEPROC_TI_PRU
configs:am65x_evm_a53_defconfig: EnableCONFIG_CMD_REMOTEPROCarch/arm/dts/k3-am65-main.dtsi | 212 ++++++++ arch/arm/dts/k3-am65.dtsi | 4 +- arch/arm/dts/k3-am654-base-board-u-boot.dtsi | 130 +++++ configs/am65x_evm_a53_defconfig | 5 + drivers/misc/misc-uclass.c | 25 + drivers/net/ti/Kconfig | 8 + drivers/net/ti/Makefile | 1 + drivers/net/ti/icssg-prueth.c | 525 +++++++++++++++++++ drivers/net/ti/icssg.h | 31 ++ drivers/net/ti/icssg_classifier.c | 397 ++++++++++++++ drivers/remoteproc/Kconfig | 11 + drivers/remoteproc/Makefile | 1 + drivers/remoteproc/pru_rproc.c | 384 ++++++++++++++ drivers/soc/ti/Kconfig | 13 + drivers/soc/ti/Makefile | 1 + drivers/soc/ti/pruss.c | 143 +++++ include/misc.h | 9 + include/ti-pruss.h | 13 + net/eth-uclass.c | 7 +- 19 files changed, 1916 insertions(+), 4 deletions(-) create mode 100644 drivers/net/ti/icssg-prueth.c create mode 100644 drivers/net/ti/icssg.h create mode 100644 drivers/net/ti/icssg_classifier.c create mode 100644 drivers/remoteproc/pru_rproc.c create mode 100644 drivers/soc/ti/pruss.c create mode 100644 include/ti-pruss.hSignature · 6 lines
2.17.1 _______________________________________________ U-Boot mailing list U-Boot at lists.denx.de https://lists.denx.de/listinfo/u-boot
Signature · 2 lines
Texas Instruments Finland Oy, Porkkalankatu 22, 00180 Helsinki. Y-tunnus/Business ID: 0615521-4. Kotipaikka/Domicile: Helsinki