chmousset [Fri, 12 Jul 2019 14:37:23 +0000 (16:37 +0200)]
[fix] Slave interface HP0 clk name
Florent Kermarrec [Fri, 12 Jul 2019 07:52:40 +0000 (09:52 +0200)]
soc_zynq: use zynq fabric reset as sys reset
Florent Kermarrec [Wed, 10 Jul 2019 14:51:08 +0000 (16:51 +0200)]
soc_zynq: add missing axi hp0 clock
Florent Kermarrec [Wed, 10 Jul 2019 14:50:06 +0000 (16:50 +0200)]
soc_zynq: move axi gp0 clock connection to add_gp0 method
Florent Kermarrec [Wed, 10 Jul 2019 08:37:32 +0000 (10:37 +0200)]
soc_core: use fixed 16MB CSR address space
Using too small CSR address space cause a regression on PCIe SoC, this would
need to be understood if we want to reduce CSR address space under 16MB.
Florent Kermarrec [Tue, 9 Jul 2019 10:14:50 +0000 (12:14 +0200)]
soc_sdram: limit main_ram to 512MB for now
Otherwise breaks linux-on-litex-vexriscv for targets with 1GB of ram, could
be removed when mem_map will be reworked on linux-on-litex-vexriscv.
Florent Kermarrec [Mon, 8 Jul 2019 21:02:43 +0000 (23:02 +0200)]
compiler-rt: update to new location, fixes #209
Florent Kermarrec [Mon, 8 Jul 2019 20:58:07 +0000 (22:58 +0200)]
soc_core: declare csr address size when registering csr, fixes #212
Florent Kermarrec [Mon, 8 Jul 2019 20:56:14 +0000 (22:56 +0200)]
soc_cores: fix typos
enjoy-digital [Mon, 8 Jul 2019 17:03:28 +0000 (19:03 +0200)]
Merge pull request #214 from gsomlo/gls-alignment-fixup
soc_core: additional csr_alignment follow-up fixes
Gabriel L. Somlo [Mon, 8 Jul 2019 13:43:40 +0000 (09:43 -0400)]
soc_core: additional csr_alignment follow-up fixes
- Update a few additional places to use DFII_ADDR_SHIFT instead of
a hard-coded 4, which assumed 32-bit alignment.
- Force 64-bit alignment Rocket -- the only supported configuration!
This is a fixup for commit
f4770219, tested on Rocket and 64bit Linux.
Signed-off-by: Gabriel Somlo <gsomlo@gmail.com>
Florent Kermarrec [Mon, 8 Jul 2019 07:53:52 +0000 (09:53 +0200)]
soc_core: add csr_alignment to allow 64-bit alignment with 64-bit CPUs
Florent Kermarrec [Mon, 8 Jul 2019 06:57:05 +0000 (08:57 +0200)]
soc/integration: uniformize configuration constants declaration in SoCs (use self.config instead self.add_constant)
Florent Kermarrec [Sat, 6 Jul 2019 15:18:34 +0000 (17:18 +0200)]
software/libbase/id: update code (length is now fixed to 256)
Florent Kermarrec [Fri, 5 Jul 2019 17:38:58 +0000 (19:38 +0200)]
cores: add simple PWM (Pulse Width Modulation) module
Florent Kermarrec [Fri, 5 Jul 2019 17:18:52 +0000 (19:18 +0200)]
core/spi: make cs_n optional (sometimes managed externally)
Florent Kermarrec [Fri, 5 Jul 2019 17:01:55 +0000 (19:01 +0200)]
cores/spi_flash: add non-memory mapped S7SPIFlash modules based on SPIMaster (for design were we only want to re-program the bistream)
Florent Kermarrec [Fri, 5 Jul 2019 16:30:34 +0000 (18:30 +0200)]
cores: add ICAP core (tested with reconfiguration commands)
Florent Kermarrec [Fri, 5 Jul 2019 13:49:17 +0000 (15:49 +0200)]
cores: add simple and minimal hardware SPI Master with CPOL=0, CPHA=0 and build time configurable data_width and frequency.
Florent Kermarrec [Fri, 5 Jul 2019 12:37:46 +0000 (14:37 +0200)]
soc/cores/spi: remove too complicated and does not seem reliable in all cases.
Florent Kermarrec [Fri, 5 Jul 2019 12:26:10 +0000 (14:26 +0200)]
cores: add bitbang class with minimal hardware for I2C/SPI software bit-banging
Florent Kermarrec [Fri, 5 Jul 2019 11:13:31 +0000 (13:13 +0200)]
cores: remove nor_flash_16 (obsolete, most of the boards are now using SPI flash)
Florent Kermarrec [Fri, 5 Jul 2019 11:09:21 +0000 (13:09 +0200)]
cores/gpio: remove Blinker
Tim Ansell [Thu, 4 Jul 2019 00:23:36 +0000 (17:23 -0700)]
Merge pull request #210 from DurandA/master
Add verilog submodule from CPU cores to manifest
Arnaud Durand [Wed, 3 Jul 2019 22:58:26 +0000 (00:58 +0200)]
Add verilog submodule from CPU cores to manifest
Florent Kermarrec [Wed, 3 Jul 2019 11:43:34 +0000 (13:43 +0200)]
csr: add assert to ensure CSR size < busword (thanks tweakoz)
Florent Kermarrec [Fri, 28 Jun 2019 21:40:01 +0000 (23:40 +0200)]
soc_core: update default RocketChip mem_map
Florent Kermarrec [Fri, 28 Jun 2019 21:27:23 +0000 (23:27 +0200)]
soc_core: rearrange default mem_map
Florent Kermarrec [Fri, 28 Jun 2019 20:42:02 +0000 (22:42 +0200)]
bios/main: fix #ifdefs for fw command
Florent Kermarrec [Fri, 28 Jun 2019 20:32:45 +0000 (22:32 +0200)]
libnet/tftp: fix compilation warning
Florent Kermarrec [Fri, 28 Jun 2019 20:18:24 +0000 (22:18 +0200)]
bios/main: fix spiflash compilation warnings
Florent Kermarrec [Thu, 27 Jun 2019 21:32:23 +0000 (23:32 +0200)]
soc_sdram: allow main_ram_size > 256MB (limitation no longer exists)
Florent Kermarrec [Thu, 27 Jun 2019 21:28:12 +0000 (23:28 +0200)]
targets: use new prefered way to add wishbone slave
Florent Kermarrec [Thu, 27 Jun 2019 21:20:12 +0000 (23:20 +0200)]
soc_core: use new way to add wisbone slave (now prefered)
Florent Kermarrec [Thu, 27 Jun 2019 21:07:26 +0000 (23:07 +0200)]
soc_core: remove 256MB mem_map limitation
mem_map was limited to 8 256MB for simplicity but has become an issue for
complex SoCs. Default mem_map size is still 256MB (retro-compatibility) but
size can now be specified.
Florent Kermarrec [Fri, 28 Jun 2019 19:37:52 +0000 (21:37 +0200)]
soc/core: remove #!/usr/bin/env python3
enjoy-digital [Thu, 27 Jun 2019 15:02:29 +0000 (17:02 +0200)]
Merge pull request #206 from gsomlo/gls-tftp-spinner
BIOS: TFTP: ASCII spinner progress indicator (cosmetic)
Gabriel L. Somlo [Thu, 27 Jun 2019 14:31:33 +0000 (10:31 -0400)]
BIOS: TFTP: ASCII spinner progress indicator (cosmetic)
enjoy-digital [Tue, 25 Jun 2019 17:10:17 +0000 (19:10 +0200)]
Merge pull request #204 from antmicro/write_to_flash
fw (flash write) command
Florent Kermarrec [Tue, 25 Jun 2019 17:09:30 +0000 (19:09 +0200)]
core/spi_flash: re-integrate bitbang write support
Mateusz Holenko [Tue, 25 Jun 2019 09:59:22 +0000 (11:59 +0200)]
bios: add fw (flash write) command
Florent Kermarrec [Mon, 24 Jun 2019 13:40:32 +0000 (15:40 +0200)]
README: remove LiteUSB (deprecated)
Florent Kermarrec [Mon, 24 Jun 2019 10:05:02 +0000 (12:05 +0200)]
boards: community supported boards are now located at https://github.com/litex-hub/litex-boards
Florent Kermarrec [Mon, 24 Jun 2019 09:44:41 +0000 (11:44 +0200)]
liteeth: update mac imports (olds still works, but that's now the prefered way)
Florent Kermarrec [Mon, 24 Jun 2019 08:58:36 +0000 (10:58 +0200)]
soc/cores: add usb_fifo with FT245 USB FIFO PHY from LiteUSB, deprecate LiteUSB
LiteUSB was not up to date was not a real USB PHY but was just providing USB FIFO PHYs.
New true USB cores are now available: Daisho, ValentyUSB, so it's better using
then for true USB support. We only keep the FT245 FIFO PHY in LiteX that can be
useful to interface with USB2/USB3 USB FIFOs.
Florent Kermarrec [Mon, 24 Jun 2019 07:59:10 +0000 (09:59 +0200)]
README: update Intro
Florent Kermarrec [Mon, 24 Jun 2019 05:29:24 +0000 (07:29 +0200)]
make sure #!/usr/bin/env python3 is before copyright header
Florent Kermarrec [Sun, 23 Jun 2019 21:31:11 +0000 (23:31 +0200)]
test: add copyright header
Florent Kermarrec [Sun, 23 Jun 2019 20:36:00 +0000 (22:36 +0200)]
add CONTRIBUTORS file and add copyright header to all files
Florent Kermarrec [Sat, 22 Jun 2019 08:53:12 +0000 (10:53 +0200)]
bios/sdram: set init_done/error when DDRCTRL is present (litedram_gen)
Tim 'mithro' Ansell [Fri, 21 Jun 2019 19:03:30 +0000 (12:03 -0700)]
Convert top level comment to a docstring.
enjoy-digital [Fri, 21 Jun 2019 08:26:07 +0000 (10:26 +0200)]
Merge pull request #202 from xobs/add-up5kspram
soc: cores: add up5kspram module
William D. Jones [Thu, 20 Jun 2019 18:12:46 +0000 (11:12 -0700)]
soc: cores: add up5kspram module
The ICE40UP5K has 128 kB of SPRAM that's designed to be used
as memory for a softcore. This memory is actually 4 16-bit
chunks that we can gang together to give us either 64 kB or
128 kB.
Add a module that will allow us to use this memory in an ICE40.
Signed-off-by: Sean Cross <sean@xobs.io>
Florent Kermarrec [Thu, 20 Jun 2019 07:02:08 +0000 (09:02 +0200)]
cores/frequency_meter: allow passing clk to be measured as a parameter
enjoy-digital [Wed, 19 Jun 2019 06:32:38 +0000 (08:32 +0200)]
Merge pull request #201 from gsomlo/gls-fix-initmem
tools/litex_sim: fix default endianness for mem_init
Gabriel L. Somlo [Tue, 18 Jun 2019 20:29:23 +0000 (16:29 -0400)]
tools/litex_sim: fix default endianness for mem_init
Initializing ROM and/or RAM content requires knowing the CPU
endianness before the SimSoC->SoCSDRAM->SoCCore constructor
sequence is invoked (before the SoC's self.cpu.endianness
could be accessed). Given that the majority of supported CPU
models use "little", set it as the new default, and override
only for the two models that use "big" endianness.
enjoy-digital [Tue, 18 Jun 2019 11:15:30 +0000 (13:15 +0200)]
Merge pull request #200 from gsomlo/gls-rocket-variants
cpu/rocket: add "linux" (MMU) and "full" (MMU & FPU) variants
Gabriel L. Somlo [Tue, 18 Jun 2019 10:42:40 +0000 (06:42 -0400)]
cpu/rocket: add "linux" (MMU) and "full" (MMU & FPU) variants
Florent Kermarrec [Tue, 18 Jun 2019 07:44:09 +0000 (09:44 +0200)]
cpu/rocket: update submodule
Florent Kermarrec [Mon, 17 Jun 2019 07:55:27 +0000 (09:55 +0200)]
integration/soc_core: move cpu_variant checks/formating to cpu
Florent Kermarrec [Mon, 17 Jun 2019 07:54:17 +0000 (09:54 +0200)]
cpu/vexriscv: add "linux+no-dsp" variant
Florent Kermarrec [Mon, 17 Jun 2019 07:24:57 +0000 (09:24 +0200)]
cpu/vexriscv: update
Florent Kermarrec [Mon, 17 Jun 2019 07:20:21 +0000 (09:20 +0200)]
targets/ulx3s: use CAS latency of 3 to be compatible with production boards
enjoy-digital [Thu, 13 Jun 2019 05:14:03 +0000 (07:14 +0200)]
Merge pull request #199 from ambrop72/no-ethmac-fix
bios: Fix build when ethphy is present but ethmac is not.
Ambroz Bizjak [Wed, 12 Jun 2019 23:02:22 +0000 (01:02 +0200)]
bios: Fix build when ethphy is present but ethmac is not.
While testing my Ethernet DMA, I renamed the `ethmac` module to `ethmac_dma` so that it wouldn't be used from the BIOS, but I got an undefined reference to `eth_init` because `bios.c` checks different CSR defines than the code that defines `eth_init`.
Florent Kermarrec [Wed, 12 Jun 2019 09:28:06 +0000 (11:28 +0200)]
test/test_axi: remove litex.gen.sim import (was only useful for debug)
Florent Kermarrec [Wed, 12 Jun 2019 09:26:57 +0000 (11:26 +0200)]
setup.py: add migen to install_requires
enjoy-digital [Tue, 11 Jun 2019 13:50:02 +0000 (15:50 +0200)]
Merge pull request #198 from TomKeddie/tomk_20190610_artyspi
boards/arty : Add directly connected spi clk pin
Florent Kermarrec [Mon, 10 Jun 2019 16:53:30 +0000 (18:53 +0200)]
test/test_code8b10b: add test_coding
Tom Keddie [Mon, 10 Jun 2019 15:33:02 +0000 (08:33 -0700)]
boards/arty : Add directly connected spi clk pin to avoid need for STARTUPE2
Florent Kermarrec [Mon, 10 Jun 2019 14:05:53 +0000 (16:05 +0200)]
test/test_prbs: add PRBSGenerator/Checker tests
Florent Kermarrec [Mon, 10 Jun 2019 14:05:36 +0000 (16:05 +0200)]
soc/cores: add PRBS (Pseudo Random Binary Sequence) Generator/Checker
Imported from LiteICLink. PRBS can be useful for different purposes, so is
better integrated in LiteX.
Florent Kermarrec [Mon, 10 Jun 2019 13:06:57 +0000 (15:06 +0200)]
tools/litex_term: exit on 2 consecutive CTRL-C
When running OS with LiteX and when LiteXTerm is use, we want to be able to
send CTRl-C to the OS. Ensure a specific sequence is sent to close the terminal.
Florent Kermarrec [Mon, 10 Jun 2019 10:57:10 +0000 (12:57 +0200)]
cpu/vexriscv: update submodule
Florent Kermarrec [Sun, 9 Jun 2019 17:36:09 +0000 (19:36 +0200)]
doc: add litex-hub logo
Florent Kermarrec [Sat, 8 Jun 2019 22:36:46 +0000 (00:36 +0200)]
doc: redesign new logo
Florent Kermarrec [Fri, 7 Jun 2019 22:45:30 +0000 (00:45 +0200)]
doc: add new logo
Florent Kermarrec [Fri, 7 Jun 2019 16:36:46 +0000 (18:36 +0200)]
cpu/vexriscv: update submodule
Florent Kermarrec [Fri, 7 Jun 2019 10:28:20 +0000 (12:28 +0200)]
build/sim: allow configuring verilator optimization level
Florent Kermarrec [Fri, 7 Jun 2019 09:50:57 +0000 (11:50 +0200)]
build/sim: allow defining start/end cycles for tracing
Florent Kermarrec [Fri, 7 Jun 2019 09:16:39 +0000 (11:16 +0200)]
build/sim: use -O0 for verilator compilation
In most of the case, execution speed is already fast enough with -O0 and
with complex design -O0 is a lost faster to compile than -O3. In the future
we could add a switch to choose which optimization we want.
Florent Kermarrec [Fri, 7 Jun 2019 09:14:36 +0000 (11:14 +0200)]
soc/integration/soc_core: list rocket as supported CPU
Florent Kermarrec [Fri, 7 Jun 2019 09:10:04 +0000 (11:10 +0200)]
software/bios: change prompt to "litex" in green.
Florent Kermarrec [Wed, 5 Jun 2019 21:43:16 +0000 (23:43 +0200)]
integration/soc_core: improve readibility (add separators/comments)
Florent Kermarrec [Wed, 5 Jun 2019 18:03:19 +0000 (20:03 +0200)]
test/test_targets: add de10lite
enjoy-digital [Wed, 5 Jun 2019 17:44:54 +0000 (19:44 +0200)]
Merge pull request #196 from msloniewski/de10lite_support
De10lite support
enjoy-digital [Wed, 5 Jun 2019 17:20:15 +0000 (19:20 +0200)]
Merge pull request #195 from antmicro/extend_generated_headers
Extend generated headers & csv
msloniewski [Wed, 5 Jun 2019 16:53:49 +0000 (18:53 +0200)]
boards/targets: add target for de10lite platform
msloniewski [Wed, 5 Jun 2019 16:53:30 +0000 (18:53 +0200)]
boards/platforms: add de10lite Terasic platform support
msloniewski [Wed, 5 Jun 2019 16:52:40 +0000 (18:52 +0200)]
build/altera: Add possibility to turn off generation of .rbf file
For some FPGAs (e.g. MAX10) .rbf file cannot be generated.
Add possibility to turn off that feature for those chips.
Mateusz Holenko [Wed, 5 Jun 2019 15:27:27 +0000 (17:27 +0200)]
integration/builder: generate flash_boot address to csv
Mateusz Holenko [Wed, 5 Jun 2019 15:35:47 +0000 (17:35 +0200)]
integration/builder: generate shadow_base address to mem.h and csv
enjoy-digital [Tue, 4 Jun 2019 19:49:18 +0000 (21:49 +0200)]
Merge pull request #193 from gsomlo/gls-memcpy-fix
software/libbase: memcpy: simple, arch-width agnostic implementation
Gabriel L. Somlo [Tue, 4 Jun 2019 18:42:54 +0000 (14:42 -0400)]
software/libbase: memcpy: simple, arch-width agnostic implementation
Remove optimizations targeted specifically at rv32 architecture,
allowing memcpy to work on all word sizes.
Since this is "only" the BIOS, it is also arguably better to
optimize for size rather than performance, given that control
will be quickly handed over to some other program being loaded.
Signed-off-by: Gabriel Somlo <gsomlo@gmail.com>
Tim Ansell [Sun, 2 Jun 2019 23:54:20 +0000 (16:54 -0700)]
Merge pull request #192 from sutajiokousagi/pr_c99_types
fix signed char type to be explicitly signed
bunnie [Mon, 3 Jun 2019 06:00:27 +0000 (06:00 +0000)]
fix signed char type to be explicitly signed
bunnie [Sun, 2 Jun 2019 22:26:18 +0000 (22:26 +0000)]
update stdint.h to include c99 types
needed for some third party libraries to compile
Tim Ansell [Sun, 2 Jun 2019 20:00:20 +0000 (13:00 -0700)]
Merge pull request #191 from sergachev/master
Fix interrupt_name in soc_core/add_interrupt
Ilia Sergachev [Sun, 2 Jun 2019 18:56:02 +0000 (20:56 +0200)]
fix csr_name in add_csr()
Ilia Sergachev [Sun, 2 Jun 2019 18:48:08 +0000 (20:48 +0200)]
fix interrupt_name