Lattice MachXO2 Slave SPI FPGA Manager support
Subsystems (2)
v8 cover Lattice MachXO2 Slave SPI FPGA Manager support
Hi all,
this series adds support for the Lattice MachXO2 FPGA chip, programmed over Slave SPI.
Tested on a raspberry pi3, beaglebone black (little endian mode) and imx6 hummingboard (big endian mode) + bugblat's pif2 fpga hat (machxo2 7000HC) or tinyfpga A1/A2 (machxo2 256HC / 1200HC), in SPI slave mode with varying bus speed.
Changes since v7:
* misc style fixes, among which:
* static const u8 cmd[]
* #ifdef DEBUG ... #endif reorder
* introduction of get_err_string()
* direct returns to reduce line count
* indentation
Paolo Pisati (2):
dt: bindings: fpga: add lattice machxo2 slave spi binding description
fpga: lattice machxo2: Add Lattice MachXO2 support
.../bindings/fpga/lattice-machxo2-spi.txt | 29 ++ drivers/fpga/Kconfig | 8 + drivers/fpga/Makefile | 1 + drivers/fpga/machxo2-spi.c | 388 +++++++++++++++++++++ 4 files changed, 426 insertions(+) create mode 100644 Documentation/devicetree/bindings/fpga/lattice-machxo2-spi.txt create mode 100644 drivers/fpga/machxo2-spi.c
Signature · 1 line
2.7.4
PATCH v8 1/2 dt: bindings: fpga: add lattice machxo2 slave spi binding description
Add dt binding documentation details for Lattice MachXO2 FPGA configuration over Slave SPI interface.
--- .../bindings/fpga/lattice-machxo2-spi.txt | 29 ++++++++++++++++++++++ 1 file changed, 29 insertions(+) create mode 100644 Documentation/devicetree/bindings/fpga/lattice-machxo2-spi.txt
Documentation/devicetree/bindings/fpga/lattice-machxo2-spi.txt new file +29 -0
@@ -0,0 +1,29 @@ +Lattice MachXO2 Slave SPI FPGA Manager + +Lattice MachXO2 FPGAs support a method of loading the bitstream over +a 'slave SPI' interface. + +See 'MachXO2ProgrammingandConfigurationUsageGuide.pdf' on www.latticesemi.com + +Required properties: +- compatible: should contain "lattice,machxo2-slave-spi" +- reg: spi chip select of the FPGA + +Example for full FPGA configuration: + + fpga-region0 { + compatible = "fpga-region"; + fpga-mgr = <&fpga_mgr_spi>; + #address-cells = <0x1>; + #size-cells = <0x1>; + }; + + spi1: spi@2000 { + ... + + fpga_mgr_spi: fpga-mgr@0 { + compatible = "lattice,machxo2-slave-spi"; + spi-max-frequency = <8000000>; + reg = <0>; + }; + };
Signature · 1 line
2.7.4
PATCH v8 2/2 fpga: lattice machxo2: Add Lattice MachXO2 support
In reply to Paolo PisatiThis patch adds support to the FPGA manager for programming MachXO2 device’s internal flash memory, via slave SPI.
--- drivers/fpga/Kconfig | 8 + drivers/fpga/Makefile | 1 + drivers/fpga/machxo2-spi.c | 388 +++++++++++++++++++++++++++++++++++++++++++++ 3 files changed, 397 insertions(+) create mode 100644 drivers/fpga/machxo2-spi.c
drivers/fpga/Kconfig +8 -0
@@ -38,6 +38,14 @@ config FPGA_MGR_ALTERA_PS_SPI FPGA manager driver support for Altera Arria/Cyclone/Stratix using the passive serial interface over SPI. +config FPGA_MGR_MACHXO2_SPI + tristate "Lattice MachXO2 SPI" + depends on SPI + help + FPGA manager driver support for Lattice MachXO2 configuration + over slave SPI interface. + + config FPGA_MGR_SOCFPGA tristate "Altera SOCFPGA FPGA Manager" depends on ARCH_SOCFPGA || COMPILE_TEST
drivers/fpga/Makefile +1 -0
@@ -10,6 +10,7 @@ obj-$(CONFIG_FPGA) += fpga-mgr.o obj-$(CONFIG_FPGA_MGR_ALTERA_CVP) += altera-cvp.o obj-$(CONFIG_FPGA_MGR_ALTERA_PS_SPI) += altera-ps-spi.o obj-$(CONFIG_FPGA_MGR_ICE40_SPI) += ice40-spi.o +obj-$(CONFIG_FPGA_MGR_MACHXO2_SPI) += machxo2-spi.o obj-$(CONFIG_FPGA_MGR_SOCFPGA) += socfpga.o obj-$(CONFIG_FPGA_MGR_SOCFPGA_A10) += socfpga-a10.o obj-$(CONFIG_FPGA_MGR_TS73XX) += ts73xx-fpga.o
drivers/fpga/machxo2-spi.c new file +388 -0
388 lines not on this page · raw message
Signature · 1 line
2.7.4
Re: [PATCH 2/2] fpga: lattice machxo2: Add Lattice MachXO2 support
On Fri, Mar 23, 2018 at 7:27 AM, Paolo Pisati <[email protected]> wrote:
Hi Paolo,
Looking good. A few things below. Also, when you post, please use the '-v' parameter of 'git format-patch' to add the version to the subject line such as [PATCH v9].
Quoted 17 lines
This patch adds support to the FPGA manager for programming MachXO2 device’s internal flash memory, via slave SPI.
Signed-off-by Paolo Pisati <[email protected]>--- drivers/fpga/Kconfig | 8 + drivers/fpga/Makefile | 1 + drivers/fpga/machxo2-spi.c | 388 +++++++++++++++++++++++++++++++++++++++++++++ 3 files changed, 397 insertions(+) create mode 100644 drivers/fpga/machxo2-spi.cdrivers/fpga/Kconfig +0
-0@@ -38,6 +38,14 @@ config FPGA_MGR_ALTERA_PS_SPI FPGA manager driver support for Altera Arria/Cyclone/Stratix
drivers/fpga/Kconfig config FPGA_MGR_ALTERA_PS_SPI · line 48
Quoted 10 lines
drivers/fpga/Kconfig +8
-0using the passive serial interface over SPI. +config FPGA_MGR_MACHXO2_SPI + tristate "Lattice MachXO2 SPI" + depends on SPI + help + FPGA manager driver support for Lattice MachXO2 configuration + over slave SPI interface. + +
Please move this section after FPGA_MGR_ICE40_SPI. That keeps the two lattice drivers together in the Kconfig. You may need to rebase your driver to go onto linux-next.
Also, please delete the extra line feed above.
Quoted 21 lines
drivers/fpga/Kconfig +0
-0config FPGA_MGR_SOCFPGA tristate "Altera SOCFPGA FPGA Manager" depends on ARCH_SOCFPGA || COMPILE_TESTdrivers/fpga/Makefile +1
-0@@ -10,6 +10,7 @@ obj-$(CONFIG_FPGA) += fpga-mgr.o obj-$(CONFIG_FPGA_MGR_ALTERA_CVP) += altera-cvp.o obj-$(CONFIG_FPGA_MGR_ALTERA_PS_SPI) += altera-ps-spi.o obj-$(CONFIG_FPGA_MGR_ICE40_SPI) += ice40-spi.o +obj-$(CONFIG_FPGA_MGR_MACHXO2_SPI) += machxo2-spi.o obj-$(CONFIG_FPGA_MGR_SOCFPGA) += socfpga.o obj-$(CONFIG_FPGA_MGR_SOCFPGA_A10) += socfpga-a10.o obj-$(CONFIG_FPGA_MGR_TS73XX) += ts73xx-fpga.odrivers/fpga/machxo2-spi.c new file +0
-0
drivers/fpga/machxo2-spi.c line 1
Quoted 1 line
drivers/fpga/machxo2-spi.c new file +1
-0@@ -0,0 +1,388 @@ +// SPDX-License-Identifier: GPL-2.0
The above line is good but...
drivers/fpga/machxo2-spi.c line 7
Quoted 6 lines
drivers/fpga/machxo2-spi.c new file +6
-0+// Lattice MachXO2 Slave SPI Driver +// +// Manage Lattice FPGA firmware that is loaded over SPI using +// the slave serial configuration interface. +// +// Copyright (C) 2018 Paolo Pisati <[email protected]>
This stuff should be as:
/*
* some stuff
* more stuff
*/
drivers/fpga/machxo2-spi.c line 9
Quoted 2 lines
drivers/fpga/machxo2-spi.c new file +2
-0+ +//#define DEBUG
Please remove the //#define DEBUG
Quoted 15 lines
drivers/fpga/machxo2-spi.c new file +15
-0+ +#include <linux/delay.h> +#include <linux/fpga/fpga-mgr.h> +#include <linux/gpio/consumer.h> +#include <linux/module.h> +#include <linux/of.h> +#include <linux/spi/spi.h> + +/* MachXO2 Programming Guide - sysCONFIG Programming Commands */ +#define IDCODE_PUB {0xe0, 0x00, 0x00, 0x00} +#define ISC_ENABLE {0xc6, 0x08, 0x00, 0x00} +#define ISC_ERASE {0x0e, 0x04, 0x00, 0x00} +#define ISC_PROGRAMDONE {0x5e, 0x00, 0x00, 0x00} +#define LSC_INITADDRESS {0x46, 0x00, 0x00, 0x00} +#define LSC_PROGINCRNV {0x70, 0x00, 0x00, 0x01}
drivers/fpga/machxo2-spi.c line 34
Quoted 10 lines
drivers/fpga/machxo2-spi.c new file +10
-0+#define LSC_READ_STATUS {0x3c, 0x00, 0x00, 0x00} +#define LSC_REFRESH {0x79, 0x00, 0x00, 0x00} + +/* +* Max CCLK in Slave SPI mode according to 'MachXO2 Family Data +* Sheet' sysCONFIG Port Timing Specifications (3-36) +*/ +#define MACHXO2_MAX_SPEED 66000000 + +#define MACHXO2_LOW_DELAY 5/* us */
I'd prefer MACHXO2_LOW_DELAY_USEC so that you don't need the comment here and it's clear in the rest of the driver. Same for the next two.
Quoted 42 lines
drivers/fpga/machxo2-spi.c new file +42
-0+#define MACHXO2_HIGH_DELAY 200/* us */ +#define MACHXO2_REFRESH 4800/* us */ +#define MACHXO2_MAX_BUSY_LOOP 128 +#define MACHXO2_MAX_REFRESH_LOOP 16 + +#define MACHXO2_PAGE_SIZE 16 +#define MACHXO2_BUF_SIZE (MACHXO2_PAGE_SIZE + 4) + +/* Status register bits, errors and error mask */ +#define BUSY 12 +#define DONE 8 +#define DVER 27 +#define ENAB 9 +#define ERRBITS 23 +#define ERRMASK 7 +#define FAIL 13 + +#define ENOERR 0/* no error */ +#define EID 1 +#define ECMD 2 +#define ECRC 3 +#define EPREAM 4/* preamble error */ +#define EABRT 5/* abort error */ +#define EOVERFL 6/* overflow error */ +#define ESDMEOF 7/* SDM EOF */ + +static inline u8 get_err (unsigned long *status) +{ +return (*status >>ERRBITS ) &ERRMASK ; +} + +static int get_status (struct spi_device *spi,unsigned long *status) +{ +struct spi_message msg; +struct spi_transfer rx, tx; +static const u8 cmd[] =LSC_READ_STATUS ; +int ret; + +memset (&rx,0 ,sizeof (rx)); +memset (&tx,0 ,sizeof (tx)); + tx.tx_buf = cmd; + tx.len =sizeof (cmd);
drivers/fpga/machxo2-spi.c line 86
Quoted 10 lines
drivers/fpga/machxo2-spi.c new file +10
-0+ rx.rx_buf = status; + rx.len =4 ; +spi_message_init (&msg); +spi_message_add_tail (&tx, &msg); +spi_message_add_tail (&rx, &msg); + ret =spi_sync (spi, &msg); +if (ret) +return ret; + + *status =be32_to_cpu (*status);
Add a line feed before returning.
Quoted 6 lines
drivers/fpga/machxo2-spi.c new file +6
-0+return 0 ; +} + +#ifdef DEBUG +static const char *get_err_string (u8 err) +{
drivers/fpga/machxo2-spi.c line 102
Quoted 10 lines
drivers/fpga/machxo2-spi.c new file +10
-0+switch (err) { +case ENOERR :return "No Error" ; +case EID :return "ID ERR" ; +case ECMD :return "CMD ERR" ; +case ECRC :return "CRC ERR" ; +case EPREAM :return "Preamble ERR" ; +case EABRT :return "Abort ERR" ; +case EOVERFL :return "Overflow ERR" ; +case ESDMEOF :return "SDM EOF" ; + }
Line feed before returning
drivers/fpga/machxo2-spi.c line 113
Quoted 11 lines
drivers/fpga/machxo2-spi.c new file +11
-0+ return "Default switch case"; +} +#endif + +static void dump_status_reg (unsigned long *status) +{ +#ifdef DEBUG +pr_debug ("machxo2 status: 0x%08lX - done=%d, cfgena=%d, busy=%d, fail=%d, devver=%d, err=%s\n" , + *status,test_bit (DONE , status),test_bit (ENAB , status), +test_bit (BUSY , status),test_bit (FAIL , status), +test_bit (DVER , status),get_err_string (get_err (status)));
Please align the the 3 lines after the pr_debug to match the above parentheses.
Quoted 49 lines
drivers/fpga/machxo2-spi.c new file +49
-0+#endif +} + +static int wait_until_not_busy (struct spi_device *spi) +{ +unsigned long status; +int ret, loop =0 ; + +do { + ret =get_status (spi, &status); +if (ret) +return ret; +if (++loop >=MACHXO2_MAX_BUSY_LOOP ) +return -EBUSY ; + }while (test_bit (BUSY , &status)); + +return 0 ; +} + +static int machxo2_cleanup (struct fpga_manager *mgr) +{ +struct spi_device *spi = mgr->priv; +struct spi_message msg; +struct spi_transfer tx[2 ]; +static const u8 erase[] =ISC_ERASE ; +static const u8 refresh[] =LSC_REFRESH ; +int ret; + +memset (tx,0 ,sizeof (tx)); +spi_message_init (&msg); + tx[0 ].tx_buf = &erase; + tx[0 ].len =sizeof (erase); +spi_message_add_tail (&tx[0 ], &msg); + ret =spi_sync (spi, &msg); +if (ret) + goto fail; + + ret =wait_until_not_busy (spi); +if (ret) + goto fail; + +spi_message_init (&msg); + tx[1 ].tx_buf = &refresh; + tx[1 ].len =sizeof (refresh); + tx[1 ].delay_usecs =MACHXO2_REFRESH ; +spi_message_add_tail (&tx[1 ], &msg); + ret =spi_sync (spi, &msg); +if (ret) + goto fail;
drivers/fpga/machxo2-spi.c line 172
Quoted 10 lines
drivers/fpga/machxo2-spi.c new file +10
-0+ +return 0 ; +fail: +dev_err (&mgr->dev,"Cleanup failed\n" ); +return ret; +} + +static enum fpga_mgr_states machxo2_spi_state (struct fpga_manager *mgr) +{ +return FPGA_MGR_STATE_UNKNOWN ;
If you have a way of reporting whether the FPGA is in operating state or some error state, etc here, that would be nice.
Quoted 136 lines
drivers/fpga/machxo2-spi.c new file +136
-0+} + +static int machxo2_write_init (struct fpga_manager *mgr, +struct fpga_image_info *info, +const char *buf,size_t count) +{ +struct spi_device *spi = mgr->priv; +struct spi_message msg; +struct spi_transfer tx[3 ]; +static const u8 enable[] =ISC_ENABLE ; +static const u8 erase[] =ISC_ERASE ; +static const u8 initaddr[] =LSC_INITADDRESS ; +unsigned long status; +int ret; + +if ((info->flags &FPGA_MGR_PARTIAL_RECONFIG )) { +dev_err (&mgr->dev, +"Partial reconfiguration is not supported\n" ); +return -ENOTSUPP ; + } + +get_status (spi, &status); +dump_status_reg (&status); +memset (tx,0 ,sizeof (tx)); +spi_message_init (&msg); + tx[0 ].tx_buf = &enable; + tx[0 ].len =sizeof (enable); + tx[0 ].delay_usecs =MACHXO2_LOW_DELAY ; +spi_message_add_tail (&tx[0 ], &msg); + + tx[1 ].tx_buf = &erase; + tx[1 ].len =sizeof (erase); +spi_message_add_tail (&tx[1 ], &msg); + ret =spi_sync (spi, &msg); +if (ret) + goto fail; + + ret =wait_until_not_busy (spi); +if (ret) + goto fail; + +get_status (spi, &status); +if (test_bit (FAIL , &status)) + goto fail; +dump_status_reg (&status); + +spi_message_init (&msg); + tx[2 ].tx_buf = &initaddr; + tx[2 ].len =sizeof (initaddr); +spi_message_add_tail (&tx[2 ], &msg); + ret =spi_sync (spi, &msg); +if (ret) + goto fail; + +get_status (spi, &status); +dump_status_reg (&status); +return 0 ; +fail: +dev_err (&mgr->dev,"Error during FPGA init.\n" ); +return ret; +} + +static int machxo2_write (struct fpga_manager *mgr,const char *buf, +size_t count) +{ +struct spi_device *spi = mgr->priv; +struct spi_message msg; +struct spi_transfer tx; +static const u8 progincr[] =LSC_PROGINCRNV ; +u8 payload[MACHXO2_BUF_SIZE ]; +unsigned long status; +int i, ret; + +if (count %MACHXO2_PAGE_SIZE !=0 ) { +dev_err (&mgr->dev,"Malformed payload.\n" ); +return -EINVAL ; + } +get_status (spi, &status); +dump_status_reg (&status); +memcpy (payload, &progincr,sizeof (progincr)); +for (i =0 ; i < count; i +=MACHXO2_PAGE_SIZE ) { +memcpy (&payload[sizeof (progincr)], &buf[i],MACHXO2_PAGE_SIZE ); +memset (&tx,0 ,sizeof (tx)); +spi_message_init (&msg); + tx.tx_buf = payload; + tx.len =MACHXO2_BUF_SIZE ; + tx.delay_usecs =MACHXO2_HIGH_DELAY ; +spi_message_add_tail (&tx, &msg); + ret =spi_sync (spi, &msg); +if (ret) { +dev_err (&mgr->dev,"Error loading the bitstream.\n" ); +return ret; + } + } +get_status (spi, &status); +dump_status_reg (&status); + +return 0 ; +} + +static int machxo2_write_complete (struct fpga_manager *mgr, +struct fpga_image_info *info) +{ +struct spi_device *spi = mgr->priv; +struct spi_message msg; +struct spi_transfer tx[2 ]; +static const u8 progdone[] =ISC_PROGRAMDONE ; +static const u8 refresh[] =LSC_REFRESH ; +unsigned long status; +int ret, refreshloop =0 ; + +memset (tx,0 ,sizeof (tx)); +spi_message_init (&msg); + tx[0 ].tx_buf = &progdone; + tx[0 ].len =sizeof (progdone); +spi_message_add_tail (&tx[0 ], &msg); + ret =spi_sync (spi, &msg); +if (ret) + goto fail; + ret =wait_until_not_busy (spi); +if (ret) + goto fail; + +get_status (spi, &status); +dump_status_reg (&status); +if (!test_bit (DONE , &status)) { +machxo2_cleanup (mgr); + goto fail; + } + + do { + spi_message_init(&msg); + tx[1].tx_buf = &refresh; + tx[1].len = sizeof(refresh); + tx[1].delay_usecs = MACHXO2_REFRESH; + spi_message_add_tail(&tx[1], &msg);
drivers/fpga/machxo2-spi.c line 318
Quoted 10 lines
drivers/fpga/machxo2-spi.c new file +10
-0+ ret =spi_sync (spi, &msg); +if (ret) + goto fail; + +/* check refresh status */ +get_status (spi, &status); +dump_status_reg (&status); +if (!test_bit (BUSY , &status) &&test_bit (DONE , &status) && +get_err (&status) ==ENOERR ) +break ;
The get_err should be aligned with the above parentheses. Then you can adjust the indentation of the break.
Quoted 73 lines
drivers/fpga/machxo2-spi.c new file +70
-1+if (++refreshloop ==MACHXO2_MAX_REFRESH_LOOP ) { +machxo2_cleanup (mgr); + goto fail; + } + }while (1 ); + +get_status (spi, &status); +dump_status_reg (&status); +return 0 ; +fail: +dev_err (&mgr->dev,"Refresh failed.\n" ); +return ret; +} + +static const struct fpga_manager_ops machxo2_ops = { + .state = machxo2_spi_state, + .write_init = machxo2_write_init, + .write = machxo2_write, + .write_complete = machxo2_write_complete, +}; + +static int machxo2_spi_probe (struct spi_device *spi) +{ +struct device *dev = &spi->dev; + +if (spi->max_speed_hz >MACHXO2_MAX_SPEED ) { +dev_err (dev,"Speed is too high\n" ); +return -EINVAL ; + } + +return fpga_mgr_register (dev,"Lattice MachXO2 SPI FPGA Manager" , + &machxo2_ops, spi); +} + +static int machxo2_spi_remove (struct spi_device *spi) +{ +struct device *dev = &spi->dev; + +fpga_mgr_unregister (dev); + +return 0 ; +} + +static const struct of_device_id of_match[] = { + { .compatible ="lattice,machxo2-slave-spi" , }, + {} +}; +MODULE_DEVICE_TABLE (of, of_match); + +static const struct spi_device_id lattice_ids[] = { + {"machxo2-slave-spi" ,0 }, + { }, +}; +MODULE_DEVICE_TABLE (spi, lattice_ids); + +static struct spi_driver machxo2_spi_driver = { + .driver = { + .name ="machxo2-slave-spi" , + .of_match_table =of_match_ptr (of_match), + }, + .probe = machxo2_spi_probe, + .remove = machxo2_spi_remove, + .id_table = lattice_ids, +}; + +module_spi_driver(machxo2_spi_driver) + +MODULE_AUTHOR("Paolo Pisati <[email protected]>"); +MODULE_DESCRIPTION ("Load Lattice FPGA firmware over SPI" ); +MODULE_LICENSE ("GPL v2" ); --2.7.4
Re: [PATCH 2/2] fpga: lattice machxo2: Add Lattice MachXO2 support
Quoted 3 lines · Alan Tull
On Fri, Mar 23, 2018 at 7:27 AM, Paolo Pisati <[email protected]> wrote:
Hi Paolo,
One more thing. The api for registering a FPGA manager is changing. It won't be hard to adapt. I've pushed a branch that includes the patches that add this change. Please rebase on top of that branch so your patch can go in on top of that.
My git repo is at:git://git.kernel.org/pub/scm/linux/kernel/git/atull/linux-fpga.git
The branch name is
for-review-next-20180327-drvdata-v3
Alan
Quoted 485 lines
Looking good. A few things below. Also, when you post, please use the '-v' parameter of 'git format-patch' to add the version to the subject line such as [PATCH v9].
Quoted 17 lines
This patch adds support to the FPGA manager for programming MachXO2 device’s internal flash memory, via slave SPI.
Signed-off-by Paolo Pisati <[email protected]>--- drivers/fpga/Kconfig | 8 + drivers/fpga/Makefile | 1 + drivers/fpga/machxo2-spi.c | 388 +++++++++++++++++++++++++++++++++++++++++++++ 3 files changed, 397 insertions(+) create mode 100644 drivers/fpga/machxo2-spi.cdrivers/fpga/Kconfig +0
-0@@ -38,6 +38,14 @@ config FPGA_MGR_ALTERA_PS_SPI FPGA manager driver support for Altera Arria/Cyclone/Stratixdrivers/fpga/Kconfig config FPGA_MGR_ALTERA_PS_SPI · line 48
Quoted 10 lines
drivers/fpga/Kconfig +8
-0using the passive serial interface over SPI. +config FPGA_MGR_MACHXO2_SPI + tristate "Lattice MachXO2 SPI" + depends on SPI + help + FPGA manager driver support for Lattice MachXO2 configuration + over slave SPI interface. + +Please move this section after
FPGA_MGR_ICE40_SPI. That keeps the two lattice drivers together in theKconfig. You may need to rebase your driver to go onto linux-next.Also, please delete the extra line feed above.
Quoted 21 lines
drivers/fpga/Kconfig +0
-0config FPGA_MGR_SOCFPGA tristate "Altera SOCFPGA FPGA Manager" depends on ARCH_SOCFPGA || COMPILE_TESTdrivers/fpga/Makefile +1
-0@@ -10,6 +10,7 @@ obj-$(CONFIG_FPGA) += fpga-mgr.o obj-$(CONFIG_FPGA_MGR_ALTERA_CVP) += altera-cvp.o obj-$(CONFIG_FPGA_MGR_ALTERA_PS_SPI) += altera-ps-spi.o obj-$(CONFIG_FPGA_MGR_ICE40_SPI) += ice40-spi.o +obj-$(CONFIG_FPGA_MGR_MACHXO2_SPI) += machxo2-spi.o obj-$(CONFIG_FPGA_MGR_SOCFPGA) += socfpga.o obj-$(CONFIG_FPGA_MGR_SOCFPGA_A10) += socfpga-a10.o obj-$(CONFIG_FPGA_MGR_TS73XX) += ts73xx-fpga.odrivers/fpga/machxo2-spi.c new file +0
-0drivers/fpga/machxo2-spi.c line 1
Quoted 1 line
drivers/fpga/machxo2-spi.c new file +1
-0@@ -0,0 +1,388 @@ +// SPDX-License-Identifier: GPL-2.0 The above line is good but...
drivers/fpga/machxo2-spi.c line 7
Quoted 6 lines
drivers/fpga/machxo2-spi.c new file +6
-0+// Lattice MachXO2 Slave SPI Driver +// +// Manage Lattice FPGA firmware that is loaded over SPI using +// the slave serial configuration interface. +// +// Copyright (C) 2018 Paolo Pisati <[email protected]> This stuff should be as:
/*
* some stuff
* more stuff
*/drivers/fpga/machxo2-spi.c line 9
Quoted 2 lines
drivers/fpga/machxo2-spi.c new file +2
-0+ +//#define DEBUG Please remove the //#define DEBUG
Quoted 15 lines
drivers/fpga/machxo2-spi.c new file +15
-0+ +#include <linux/delay.h> +#include <linux/fpga/fpga-mgr.h> +#include <linux/gpio/consumer.h> +#include <linux/module.h> +#include <linux/of.h> +#include <linux/spi/spi.h> + +/* MachXO2 Programming Guide - sysCONFIG Programming Commands */ +#define IDCODE_PUB {0xe0, 0x00, 0x00, 0x00} +#define ISC_ENABLE {0xc6, 0x08, 0x00, 0x00} +#define ISC_ERASE {0x0e, 0x04, 0x00, 0x00} +#define ISC_PROGRAMDONE {0x5e, 0x00, 0x00, 0x00} +#define LSC_INITADDRESS {0x46, 0x00, 0x00, 0x00} +#define LSC_PROGINCRNV {0x70, 0x00, 0x00, 0x01}drivers/fpga/machxo2-spi.c line 34
Quoted 10 lines
drivers/fpga/machxo2-spi.c new file +10
-0+#define LSC_READ_STATUS {0x3c, 0x00, 0x00, 0x00} +#define LSC_REFRESH {0x79, 0x00, 0x00, 0x00} + +/* +* Max CCLK in Slave SPI mode according to 'MachXO2 Family Data +* Sheet' sysCONFIG Port Timing Specifications (3-36) +*/ +#define MACHXO2_MAX_SPEED 66000000 + +#define MACHXO2_LOW_DELAY 5/* us */ I'd prefer
MACHXO2_LOW_DELAY_USECso that you don't need the comment here and it's clear in the rest of the driver. Same for the next two.Quoted 42 lines
drivers/fpga/machxo2-spi.c new file +42
-0+#define MACHXO2_HIGH_DELAY 200/* us */ +#define MACHXO2_REFRESH 4800/* us */ +#define MACHXO2_MAX_BUSY_LOOP 128 +#define MACHXO2_MAX_REFRESH_LOOP 16 + +#define MACHXO2_PAGE_SIZE 16 +#define MACHXO2_BUF_SIZE (MACHXO2_PAGE_SIZE + 4) + +/* Status register bits, errors and error mask */ +#define BUSY 12 +#define DONE 8 +#define DVER 27 +#define ENAB 9 +#define ERRBITS 23 +#define ERRMASK 7 +#define FAIL 13 + +#define ENOERR 0/* no error */ +#define EID 1 +#define ECMD 2 +#define ECRC 3 +#define EPREAM 4/* preamble error */ +#define EABRT 5/* abort error */ +#define EOVERFL 6/* overflow error */ +#define ESDMEOF 7/* SDM EOF */ + +static inline u8 get_err (unsigned long *status) +{ +return (*status >>ERRBITS ) &ERRMASK ; +} + +static int get_status (struct spi_device *spi,unsigned long *status) +{ +struct spi_message msg; +struct spi_transfer rx, tx; +static const u8 cmd[] =LSC_READ_STATUS ; +int ret; + +memset (&rx,0 ,sizeof (rx)); +memset (&tx,0 ,sizeof (tx)); + tx.tx_buf = cmd; + tx.len =sizeof (cmd);drivers/fpga/machxo2-spi.c line 86
Quoted 10 lines
drivers/fpga/machxo2-spi.c new file +10
-0+ rx.rx_buf = status; + rx.len =4 ; +spi_message_init (&msg); +spi_message_add_tail (&tx, &msg); +spi_message_add_tail (&rx, &msg); + ret =spi_sync (spi, &msg); +if (ret) +return ret; + + *status =be32_to_cpu (*status);Add a line feed before returning.
Quoted 6 lines
drivers/fpga/machxo2-spi.c new file +6
-0+return 0 ; +} + +#ifdef DEBUG +static const char *get_err_string (u8 err) +{drivers/fpga/machxo2-spi.c line 102
Quoted 10 lines
drivers/fpga/machxo2-spi.c new file +10
-0+switch (err) { +case ENOERR :return "No Error" ; +case EID :return "ID ERR" ; +case ECMD :return "CMD ERR" ; +case ECRC :return "CRC ERR" ; +case EPREAM :return "Preamble ERR" ; +case EABRT :return "Abort ERR" ; +case EOVERFL :return "Overflow ERR" ; +case ESDMEOF :return "SDM EOF" ; + }Line feed before returning
drivers/fpga/machxo2-spi.c line 113
Quoted 11 lines
drivers/fpga/machxo2-spi.c new file +11
-0+ return "Default switch case"; +} +#endif + +static void dump_status_reg (unsigned long *status) +{ +#ifdef DEBUG +pr_debug ("machxo2 status: 0x%08lX - done=%d, cfgena=%d, busy=%d, fail=%d, devver=%d, err=%s\n" , + *status,test_bit (DONE , status),test_bit (ENAB , status), +test_bit (BUSY , status),test_bit (FAIL , status), +test_bit (DVER , status),get_err_string (get_err (status)));Please align the the 3 lines after the
pr_debugto match the above parentheses.Quoted 49 lines
drivers/fpga/machxo2-spi.c new file +49
-0+#endif +} + +static int wait_until_not_busy (struct spi_device *spi) +{ +unsigned long status; +int ret, loop =0 ; + +do { + ret =get_status (spi, &status); +if (ret) +return ret; +if (++loop >=MACHXO2_MAX_BUSY_LOOP ) +return -EBUSY ; + }while (test_bit (BUSY , &status)); + +return 0 ; +} + +static int machxo2_cleanup (struct fpga_manager *mgr) +{ +struct spi_device *spi = mgr->priv; +struct spi_message msg; +struct spi_transfer tx[2 ]; +static const u8 erase[] =ISC_ERASE ; +static const u8 refresh[] =LSC_REFRESH ; +int ret; + +memset (tx,0 ,sizeof (tx)); +spi_message_init (&msg); + tx[0 ].tx_buf = &erase; + tx[0 ].len =sizeof (erase); +spi_message_add_tail (&tx[0 ], &msg); + ret =spi_sync (spi, &msg); +if (ret) + goto fail; + + ret =wait_until_not_busy (spi); +if (ret) + goto fail; + +spi_message_init (&msg); + tx[1 ].tx_buf = &refresh; + tx[1 ].len =sizeof (refresh); + tx[1 ].delay_usecs =MACHXO2_REFRESH ; +spi_message_add_tail (&tx[1 ], &msg); + ret =spi_sync (spi, &msg); +if (ret) + goto fail;drivers/fpga/machxo2-spi.c line 172
Quoted 10 lines
drivers/fpga/machxo2-spi.c new file +10
-0+ +return 0 ; +fail: +dev_err (&mgr->dev,"Cleanup failed\n" ); +return ret; +} + +static enum fpga_mgr_states machxo2_spi_state (struct fpga_manager *mgr) +{ +return FPGA_MGR_STATE_UNKNOWN ;If you have a way of reporting whether the FPGA is in operating state or some error state, etc here, that would be nice.
Quoted 136 lines
drivers/fpga/machxo2-spi.c new file +136
-0+} + +static int machxo2_write_init (struct fpga_manager *mgr, +struct fpga_image_info *info, +const char *buf,size_t count) +{ +struct spi_device *spi = mgr->priv; +struct spi_message msg; +struct spi_transfer tx[3 ]; +static const u8 enable[] =ISC_ENABLE ; +static const u8 erase[] =ISC_ERASE ; +static const u8 initaddr[] =LSC_INITADDRESS ; +unsigned long status; +int ret; + +if ((info->flags &FPGA_MGR_PARTIAL_RECONFIG )) { +dev_err (&mgr->dev, +"Partial reconfiguration is not supported\n" ); +return -ENOTSUPP ; + } + +get_status (spi, &status); +dump_status_reg (&status); +memset (tx,0 ,sizeof (tx)); +spi_message_init (&msg); + tx[0 ].tx_buf = &enable; + tx[0 ].len =sizeof (enable); + tx[0 ].delay_usecs =MACHXO2_LOW_DELAY ; +spi_message_add_tail (&tx[0 ], &msg); + + tx[1 ].tx_buf = &erase; + tx[1 ].len =sizeof (erase); +spi_message_add_tail (&tx[1 ], &msg); + ret =spi_sync (spi, &msg); +if (ret) + goto fail; + + ret =wait_until_not_busy (spi); +if (ret) + goto fail; + +get_status (spi, &status); +if (test_bit (FAIL , &status)) + goto fail; +dump_status_reg (&status); + +spi_message_init (&msg); + tx[2 ].tx_buf = &initaddr; + tx[2 ].len =sizeof (initaddr); +spi_message_add_tail (&tx[2 ], &msg); + ret =spi_sync (spi, &msg); +if (ret) + goto fail; + +get_status (spi, &status); +dump_status_reg (&status); +return 0 ; +fail: +dev_err (&mgr->dev,"Error during FPGA init.\n" ); +return ret; +} + +static int machxo2_write (struct fpga_manager *mgr,const char *buf, +size_t count) +{ +struct spi_device *spi = mgr->priv; +struct spi_message msg; +struct spi_transfer tx; +static const u8 progincr[] =LSC_PROGINCRNV ; +u8 payload[MACHXO2_BUF_SIZE ]; +unsigned long status; +int i, ret; + +if (count %MACHXO2_PAGE_SIZE !=0 ) { +dev_err (&mgr->dev,"Malformed payload.\n" ); +return -EINVAL ; + } +get_status (spi, &status); +dump_status_reg (&status); +memcpy (payload, &progincr,sizeof (progincr)); +for (i =0 ; i < count; i +=MACHXO2_PAGE_SIZE ) { +memcpy (&payload[sizeof (progincr)], &buf[i],MACHXO2_PAGE_SIZE ); +memset (&tx,0 ,sizeof (tx)); +spi_message_init (&msg); + tx.tx_buf = payload; + tx.len =MACHXO2_BUF_SIZE ; + tx.delay_usecs =MACHXO2_HIGH_DELAY ; +spi_message_add_tail (&tx, &msg); + ret =spi_sync (spi, &msg); +if (ret) { +dev_err (&mgr->dev,"Error loading the bitstream.\n" ); +return ret; + } + } +get_status (spi, &status); +dump_status_reg (&status); + +return 0 ; +} + +static int machxo2_write_complete (struct fpga_manager *mgr, +struct fpga_image_info *info) +{ +struct spi_device *spi = mgr->priv; +struct spi_message msg; +struct spi_transfer tx[2 ]; +static const u8 progdone[] =ISC_PROGRAMDONE ; +static const u8 refresh[] =LSC_REFRESH ; +unsigned long status; +int ret, refreshloop =0 ; + +memset (tx,0 ,sizeof (tx)); +spi_message_init (&msg); + tx[0 ].tx_buf = &progdone; + tx[0 ].len =sizeof (progdone); +spi_message_add_tail (&tx[0 ], &msg); + ret =spi_sync (spi, &msg); +if (ret) + goto fail; + ret =wait_until_not_busy (spi); +if (ret) + goto fail; + +get_status (spi, &status); +dump_status_reg (&status); +if (!test_bit (DONE , &status)) { +machxo2_cleanup (mgr); + goto fail; + } + + do { + spi_message_init(&msg); + tx[1].tx_buf = &refresh; + tx[1].len = sizeof(refresh); + tx[1].delay_usecs = MACHXO2_REFRESH; + spi_message_add_tail(&tx[1], &msg);drivers/fpga/machxo2-spi.c line 318
Quoted 10 lines
drivers/fpga/machxo2-spi.c new file +10
-0+ ret =spi_sync (spi, &msg); +if (ret) + goto fail; + +/* check refresh status */ +get_status (spi, &status); +dump_status_reg (&status); +if (!test_bit (BUSY , &status) &&test_bit (DONE , &status) && +get_err (&status) ==ENOERR ) +break ;The
get_errshould be aligned with the above parentheses. Then you can adjust the indentation of the break.Quoted 73 lines
drivers/fpga/machxo2-spi.c new file +70
-1+if (++refreshloop ==MACHXO2_MAX_REFRESH_LOOP ) { +machxo2_cleanup (mgr); + goto fail; + } + }while (1 ); + +get_status (spi, &status); +dump_status_reg (&status); +return 0 ; +fail: +dev_err (&mgr->dev,"Refresh failed.\n" ); +return ret; +} + +static const struct fpga_manager_ops machxo2_ops = { + .state = machxo2_spi_state, + .write_init = machxo2_write_init, + .write = machxo2_write, + .write_complete = machxo2_write_complete, +}; + +static int machxo2_spi_probe (struct spi_device *spi) +{ +struct device *dev = &spi->dev; + +if (spi->max_speed_hz >MACHXO2_MAX_SPEED ) { +dev_err (dev,"Speed is too high\n" ); +return -EINVAL ; + } + +return fpga_mgr_register (dev,"Lattice MachXO2 SPI FPGA Manager" , + &machxo2_ops, spi); +} + +static int machxo2_spi_remove (struct spi_device *spi) +{ +struct device *dev = &spi->dev; + +fpga_mgr_unregister (dev); + +return 0 ; +} + +static const struct of_device_id of_match[] = { + { .compatible ="lattice,machxo2-slave-spi" , }, + {} +}; +MODULE_DEVICE_TABLE (of, of_match); + +static const struct spi_device_id lattice_ids[] = { + {"machxo2-slave-spi" ,0 }, + { }, +}; +MODULE_DEVICE_TABLE (spi, lattice_ids); + +static struct spi_driver machxo2_spi_driver = { + .driver = { + .name ="machxo2-slave-spi" , + .of_match_table =of_match_ptr (of_match), + }, + .probe = machxo2_spi_probe, + .remove = machxo2_spi_remove, + .id_table = lattice_ids, +}; + +module_spi_driver(machxo2_spi_driver) + +MODULE_AUTHOR("Paolo Pisati <[email protected]>"); +MODULE_DESCRIPTION ("Load Lattice FPGA firmware over SPI" ); +MODULE_LICENSE ("GPL v2" ); --2.7.4
Re: [PATCH 2/2] fpga: lattice machxo2: Add Lattice MachXO2 support
Quoted 10 lines · Alan Tull
Quoted 3 lines · Alan Tull
On Fri, Mar 23, 2018 at 7:27 AM, Paolo Pisati <[email protected]> wrote:
Hi Paolo,
One more thing. The api for registering a FPGA manager is changing. It won't be hard to adapt. I've pushed a branch that includes the patches that add this change. Please rebase on top of that branch so your patch can go in on top of that.
Quoted 5 lines
My git repo is at:
git://git.kernel.org/pub/scm/linux/kernel/git/atull/linux-fpga.gitThe branch name is
for-review-next-20180327-drvdata-v3
Ok, i'll fix all the issues you raised in the previous email, rebase on this branch and resend.
Signature · 2 lines
bye, p.