f3f89ed8d1
Default to HW control for sim
...
(needs corresponding sdram.c fix in litex)
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2020-06-01 23:44:59 +10:00
f23cb8056d
litedram_gen: revert builder.build(..., regular_comb=False).
2020-05-27 19:56:00 +02:00
d1db115d6c
litedram_gen: review/simplify #197 .
2020-05-27 19:42:40 +02:00
a8e281f7c5
Merge pull request #197 from ozbenh/standalone-sim
...
Allow generation of a standalone sim model
2020-05-27 19:16:00 +02:00
83b9a1d798
Merge pull request #199 from antmicro/jboc/spd-read
...
modules: add function for parsing SPD EEPROM dumps from BIOS firmware
2020-05-27 14:57:16 +02:00
d0f0c94652
phy/model: Don't generate empty mem_*.init files
...
When using the SDRAM PHY model without specified init data, the
generator still generates a bunch of $readmemh for each bank
reading mem.init, mem_1.init etc... all of which are 0-sized files.
This is cumbersome especially when using a standalone model in
an external project.
This is fixed by having the default bank_init be set to a list
of "None" rather than a list of empty lists.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2020-05-22 18:09:11 +10:00
b8d6da534f
gen: Allow generation of a standalone sim model
...
If the --sim argument is passed, this generates a model using
the simulation platform and simulated PHY.
Note: The data_width is set to 16 for now, which matches the
Arty. I haven't yet figured out how to extract that info from
the data we have in the .yml
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2020-05-22 18:09:06 +10:00
cbe91bcd41
modules: add function for parsing SPD EEPROM dumps from BIOS firmware
2020-05-21 16:18:34 +02:00
639a31fdd2
test/test_timing: update test_txxd_controller.
2020-05-20 23:40:01 +02:00
3c1ab76bbc
litedram/common/tXXDController: only set reset to 1 when txxd is None.
...
This avoids triggering a warning/error with Yosys.
2020-05-19 13:10:32 +02:00
e95af3f15b
Merge pull request #195 from enjoy-digital/bios-libs
...
Init: Generate DFII defines in sdram_phy.h
2020-05-19 08:18:38 +02:00
fe48a9290c
test/reference: update.
2020-05-19 08:16:11 +02:00
c30910a9d2
init: generate DFII_CONTROL flags in sdram_phy.h instead of defining them in the BIOS.
2020-05-18 23:06:33 +02:00
5078b19bff
core/crossbar: remove retro-compat > 6 months old.
2020-05-18 18:51:56 +02:00
3b105d512b
modules: fix SDRAMRegisteredModule.
2020-05-15 21:39:00 +02:00
b2a5685396
Merge pull request #189 from daveshah1/ddr4_rdimm_init
...
Add support for DDR4 RDIMMs
2020-05-15 21:34:43 +02:00
7ae4ad5b7d
modules: add SDR/DDR/DDR2/DDR3/DDR4 SDRAMModule (and Registered versions).
2020-05-15 21:27:43 +02:00
1f7d9eb0b9
litedram_gen: pass FPGA speedgrade to iodelay_pll.
2020-05-14 11:44:32 +02:00
22bd01c014
frontend/wishbone: simplify by reusing LiteDRAMNativePortConverter
2020-05-13 17:14:42 +02:00
f4871b9f13
litedram_gen: use default settings on wb_bus.
2020-05-13 14:50:06 +02:00
b0bde294c0
frontend/wishbone: fix wb2native missing wdata.ready when wb/port data widths differ
2020-05-13 10:14:00 +02:00
6fb8396d8e
litedram_gen: remove csr_base (no longer needed since CPUNone type will automatically set csr mapping to 0x00000000) and create a use bus with the same address_width as the main bus of the SoC.
...
For some use cases, we will want to have the CPU + wb_ctrl interface.
2020-05-13 09:41:48 +02:00
79314f9549
frontend/wishbone: fix wdata.valid being high with old data, use cmd.last=1
2020-05-12 17:15:18 +02:00
000a352079
frontend/adaptation: delay sending write commands to prevent data loss during up-conversion
2020-05-12 17:15:18 +02:00
84fb7d338a
frontend/adaptation: refactor up-converter logic to use FSM for clarity
2020-05-12 17:14:58 +02:00
94c215e852
litedram_gen: review/simplify #193 , always add ddrctrl.
2020-05-12 16:21:04 +02:00
f036ec2c26
Merge pull request #193 from ozbenh/standalone-cores
...
Improve standalone cores
2020-05-12 14:56:40 +02:00
04717b478b
gen: Rename standalone core wishbone
...
Name it wb_ctrl rather than just wb, which makes the resulting
core signal names a bit more descriptive. IE. The DRAM control
bus (by opposition to the use/data buss(es).
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2020-05-12 21:59:01 +10:00
b0838f70e3
gen: Add option to specify CSR alignment
...
On some standalone core implementations, such with Microwatt,
the main system bus is 64-bit, but the wishbone to access the
CSRs is 32-bit.
To avoid extra logic & muxes and just wire these together, it's
useful to be able to specify a larger alignemnt (64-bit) for the
CSRs so that the generated csr.h contains the right offsets.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2020-05-12 21:58:13 +10:00
d5a03b3d89
gen: Add option to generate DDRCTL on standalone cores
...
Microwatt will want that as it uses init_done to select whether
to run the SDRAM init code or the user code at reset.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2020-05-12 21:56:33 +10:00
efad6b3ca5
gen: Add option to specify CSR base for standalone cores
...
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2020-05-12 21:55:05 +10:00
c91cbb597d
gen: Remove obsolete bus_expose config option
...
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2020-05-12 21:54:47 +10:00
4e539ad76e
litedram_gen: switch to SoCCore.
2020-05-12 09:21:36 +02:00
ac33d29727
litedram_gen: simplify and expose bus when CPU is set to None.
2020-05-12 09:07:59 +02:00
fe478382e1
litedram_gen: expose a Bus Slave port instead of a CSR port.
...
The logic overhead is minimal and it makes things easier with more flexibility:
- since the main Bus is arbitrated, CPU and Bus Slave can coexist.
- integration is easier in LiteX.
- bridging to APB/AXI is easier.
2020-05-11 22:47:09 +02:00
efe9a44c93
frontend/adaptation: clean up LiteDRAMNativePortUpConverter code
2020-05-11 16:47:43 +02:00
2f35e9714d
frontend/adaptation: fix error when read follows write to the same address
2020-05-11 16:11:40 +02:00
1587ee3611
frontend/adaptation: use port.cmd.last instead of port.flush in up-converter
2020-05-11 15:28:32 +02:00
35fa91c055
test/crossbar: up-conversion with mode="both" should be working now
2020-05-11 14:56:39 +02:00
9b90a56e07
frontend/adaptation: combine read/write port up-converters and extend tests
2020-05-11 14:56:39 +02:00
762cd6d0f1
test/adaptation: add port converter tests with mode="both"
2020-05-11 14:56:39 +02:00
7a0f7a7ead
test/common: fix error in test data
2020-05-11 14:56:39 +02:00
1cc9656a2d
test/crossbar: improve NativePortDriver to use separate generatos on data paths
2020-05-11 14:25:06 +02:00
025e280804
test/crossbar: fix test that was not being run
2020-05-11 14:25:06 +02:00
52b49fb80e
test/reference: update.
2020-05-09 18:02:42 +02:00
52ca3936fe
modules: add MT41J512M16/MT41K512M16.
2020-05-09 16:37:24 +02:00
589957f115
phy: extend Bitslip capability to 2 sys_clk cycles.
2020-05-08 13:12:17 +02:00
5c0231d929
common/BitSlip: add cycles parameter to extend bitstlip to multiple system clock cycles.
2020-05-08 13:09:54 +02:00
ed0810a1af
gen: Optionally pass cpu_variant from YAML to SoC
...
This allows the right gcc options to be set when using for
example VexRiscv_Min.v
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
2020-05-06 21:42:37 +02:00
70054bacdb
Add support for DDR4 RDIMMs
...
Signed-off-by: David Shah <dave@ds0.me>
2020-04-29 12:34:41 +01:00
Benjamin Herrenschmidt
Florent Kermarrec
Jędrzej Boczar
David Shah