milkv-pioneer / issues

5 stars 0 forks source link

Pioneer 1.3 bare board not booting #33

Open MrDrMcCoy opened 7 months ago

MrDrMcCoy commented 7 months ago

I purchased a Pioneer 1.3 bare board, and I am at a loss for how to get this thing to boot.

I have tried various combinations of the following RAM and boot media:

These all with the fedora-disk-gnome-workstation_livecd-f38-20231010-033114.n.0-fix.raw.xz image, which most people report working:

I have also tried these things with and without a Radeon R5 240 GPU and no other PCIe devices attached.

Upon powering on, similarly to others here, there is a solid green light near the front panel header and a flashing green light near the USB3 header. With a serial adapter attached to each of the MCU and RISC-V headers, I do get some output on the MCU and am able to issue various commands successfully:

# poweron
MILKV_ATX_CTL [O]
EN_VDD_3V3 [O]
EN_VDDIO18 [O]
EN_VDDC [O]
PG_VDDC [O]
EN_VDDIO33 [O]
EN_DDR_VDD_0V8 [O]
EN_VDD_PCIE_D_0V8 [O]
EN_VDD_PLL_0V8 [O]
EN_VDD_PCIE_H_1V8 [O]
PG_DDR_VDD_0V8 [O]
EN_DDR01_VPP_2V5 [O]
EN_DDR23_VPP_2V5 [O]
EN_DDR01_VDDQ_1V2 [O]
EN_DDR23_VDDQ_1V2 [O]
PG_DDR01_VDDQ_1V2 [O]
PG_DDR23_VDDQ_1V2 [O]
EN_DDR01_VTT_0V6 [O]
EN_DDR23_VTT_0V6 [O]
EN_VQPS18 [O]
SYS_RST_ASSERT [O]
SYS_RST_DEASSERT [O]
MILKV-PIONEER POWER ON
# poweron_rv
MILKV_ATX_CTL [O]
EN_VDD_3V3 [O]
EN_VDDIO18 [O]
EN_VDDC [O]
PG_VDDC [O]
EN_VDDIO33 [O]
EN_DDR_VDD_0V8 [O]
EN_VDD_PCIE_D_0V8 [O]
EN_VDD_PLL_0V8 [O]
EN_VDD_PCIE_H_1V8 [O]
PG_DDR_VDD_0V8 [O]
EN_DDR01_VPP_2V5 [O]
EN_DDR23_VPP_2V5 [O]
EN_DDR01_VDDQ_1V2 [O]
EN_DDR23_VDDQ_1V2 [O]
PG_DDR01_VDDQ_1V2 [O]
PG_DDR23_VDDQ_1V2 [O]
EN_DDR01_VTT_0V6 [O]
EN_DDR23_VTT_0V6 [O]
EN_VQPS18 [O]
SYS_RST_ASSERT [O]
SYS_RST_DEASSERT [O]
PWRON RV OK
# current
V_5V = 483(mV)
V_DDR_VDD_0V8 = 552(mV)
V_DDR01_VDDQ_1V2 = 571(mV)
V_DDR23_VDDQ_1V2 = 547(mV)
V_VDD_12V = 611(mV)
V_VDD_EMMC_1V8 = 509(mV)
V_VDD_EMMC_3V3 = 528(mV)
V_VDD_PCIE_C_0V8 = 580(mV)
V_VDD_PCIE_D_0V8 = 583(mV)
V_VDD_PCIE_H_1V8 = 601(mV)
V_VDD_PLL_0V8 = 589(mV)
V_VDD_RGMII_1V8 = 546(mV)
V_VDDC = 564(mV)
V_VDDIO18 = 622(mV)
V_VQPS18 = 0(mV)
temp:soc:39 Cel board:31 Cel

However, absolutely nothing ever shows up on the RISC-V header in any of these combinations.

Do I have a defective unit? Is there anything else I can try to get something to happen on the RISC-V side?

Firsttime13 commented 7 months ago

I am going through the same issue. Any fix for this situation yet? Thank you

silvanshade commented 7 months ago

Upon powering on, similarly to others here, there is a solid green light near the front panel header and a flashing green light near the USB3 header. With a serial adapter attached to each of the MCU and RISC-V headers, I do get some output on the MCU and am able to issue various commands successfully:

The fact that you are able to get some output from the MCU and issue commands is at least a good first sign.

You mention connection to the MCU and RISC-V consoles. Have you tried connecting to the main board console? I'm not sure if it's even documented anywhere, but I initially had similar problems and was given this image:

serial

This will show output before the RISC-V console, and the RISC-V console will only start printing output after it completes. A normal log of it's output should look like this:

NOTICE:  BOOT: 0x7000140000/0x1/0x5
NOTICE:  Booting Trusted Firmware
NOTICE:  BL1: v2.7(release):83c4f9e3e
NOTICE:  BL1: Built : 10:39:27, Jul 12 2022
INFO:    BL1: RAM 0x7010002000 - 0x7010011000
NOTICE:  SD initializing 100000000Hz
Hit key to stop autoboot: 00
NOTICE:  BOOT: 0x7000140000/0x1/0x5
NOTICE:  Booting Trusted Firmware
NOTICE:  BL1: v2.7(release):83c4f9e3e
NOTICE:  BL1: Built : 10:39:27, Jul 12 2022
INFO:    BL1: RAM 0x7010002000 - 0x7010011000
NOTICE:  SD initializing 100000000Hz
NOTICE:  boot from SD
INFO:    BL1: Loading BL2
NOTICE:  Locate FIP in SD FAT
INFO:    Loading image id=1 at address 0x7010020000
INFO:    Image id=1 loaded: 0x7010020000 - 0x701005ca41
NOTICE:  BL1: Booting BL2
INFO:    Entry point address = 0x7010020000
INFO:    SPSR = 0x3c5
NOTICE:  BL2: v2.7(release):bab7a07f0
NOTICE:  BL2: Built : 20:23:04, Jun 14 2023
INFO:    BL2: Doing platform setup
NOTICE:  SD initializing 100000000Hz
NOTICE:  Chip0 working at single socket, multi spi mode
NOTICE:  ******************DIMM0******************
NOTICE:  DIMM0: Unknown Manufacturer RDIMM DDR4 SDRAM RAW_CARD_E 2Rx8 3200(3200) 32GB ECC(ON)
NOTICE:  ******************DIMM1******************
NOTICE:  DIMM1: Unknown Manufacturer RDIMM DDR4 SDRAM RAW_CARD_E 2Rx8 3200(3200) 32GB ECC(ON)
NOTICE:  ******************DIMM2******************
NOTICE:  DIMM2: Unknown Manufacturer RDIMM DDR4 SDRAM RAW_CARD_E 2Rx8 3200(3200) 32GB ECC(ON)
NOTICE:  ******************DIMM3******************
NOTICE:  DIMM3: Unknown Manufacturer RDIMM DDR4 SDRAM RAW_CARD_E 2Rx8 3200(3200) 32GB ECC(ON)
NOTICE:  Waiting DDR0 ECC done
NOTICE:  Waiting DDR1 ECC done
NOTICE:  Waiting DDR2 ECC done
NOTICE:  Waiting DDR3 ECC done
NOTICE:  Socket id 0, online chip number 1
NOTICE:  Socket0 is present
NOTICE:      Port0.0 RC gen4 16GT/s
NOTICE:      Port0.1 RC gen4 16GT/s
NOTICE:      Port1.0 RC gen4 16GT/s
NOTICE:      Port1.1 RC gen4 16GT/s
NOTICE:  RISC-V map: O O O O O O O O O O O O O O O O
NOTICE:  Socket1 is not present
NOTICE:  Setup CCIX
NOTICE:  Setup CMN
NOTICE:  Maximum compatibility mode
NOTICE:  DRAM0 [0000000000 - 0800000000] 32GB
NOTICE:  DRAM1 [0800000000 - 1000000000] 32GB
NOTICE:  DRAM2 [1000000000 - 1800000000] 32GB
NOTICE:  DRAM3 [1800000000 - 2000000000] 32GB
NOTICE:  Setup PLIC
NOTICE:  Setup PCIe
NOTICE:  PCIe1 Link1 linked with gen3 x8 RC mode
NOTICE:  Setup chip2chip link done
NOTICE:  RISC-V only mode

When I was having problems, initially with incompatible RAM, my output looked like this:

Hit key to stop autoboot: 00 
NOTICE:  BOOT: 0x7000140000/0x1/0x5
NOTICE:  Booting Trusted Firmware
NOTICE:  BL1: v2.7(release):83c4f9e3e
NOTICE:  BL1: Built : 10:39:27, Jul 12 2022
INFO:    BL1: RAM 0x7010002000 - 0x7010011000
NOTICE:  SD initializing 100000000Hz
NOTICE:  boot from SD
INFO:    BL1: Loading BL2
NOTICE:  Locate FIP in SD FAT
INFO:    Loading image id=1 at address 0x7010020000
INFO:    Image id=1 loaded: 0x7010020000 - 0x701005ba81
NOTICE:  BL1: Booting BL2
INFO:    Entry point address = 0x7010020000
INFO:    SPSR = 0x3c5
NOTICE:  BL2: v2.7(release):f2d025ca1
NOTICE:  BL2: Built : 10:09:05, May  9 2023
INFO:    BL2: Doing platform setup
NOTICE:  MILKV Pioneer v1
NOTICE:  SD initializing 100000000Hz
NOTICE:  ******************DIMM0******************
NOTICE:  MANGO_SOPHGO_X8EVB don't support X4 RDIMM!
NOTICE:  De-Skew PLL failed to lock
NOTICE:  lck_0 = 0NOTICE:  lck_1 = 0NOTICE:  Deskew PLL Failed
NOTICE:  Controller initialized
NOTICE:  PI initialized
NOTICE:  DIMM0: Samsung RDIMM DDR4 SDRAM RAW_CARD_B 2Rx4 3200(3200) 32GB ECC(ON)
 time out! clear memory fail
NOTICE:  Chip0 working at single socket, multi spi mode
NOTICE:  Socket id 0, online chip number 1
NOTICE:  Socket0 is present
NOTICE:      Port0.0 RC gen4 16GT/s
NOTICE:      Port0.1 RC gen4 16GT/s
NOTICE:      Port1.0 RC gen4 16GT/s
NOTICE:      Port1.1 RC gen4 16GT/s
NOTICE:  RISC-V map: O O O O O O O O O O O O O O O O
NOTICE:  Socket1 is not present
NOTICE:  Setup CCIX
NOTICE:  Setup CMN
NOTICE:  Maximum compatibility mode
NOTICE:  DRAM0 [0000000000 - 0800000000] 32GB
NOTICE:  Setup PLIC
NOTICE:  Setup PCIe
NOTICE:  PCIe0 Link0 linked with gen4 x8 RC mode
NOTICE:  PCIe1 Link1 linked with gen3 x8 RC mode
NOTICE:  Setup chip2chip link done

See the part time out! clear memory fail. The issue for me was using incompatible 2Rx4 memory.

However, if you're using the official recommended memory then it's hard to say. But at least you should potentially get some more clues as to what might be going wrong.

Firsttime13 commented 7 months ago

My apologies, I should be more specific. My lights on the board are as the original person posted. To add to my issue, I am not even posting or coming on.

Thank you for your response. Just new and trying to learn

On Wed, Jan 31, 2024, 19:39 Darin Morrison @.***> wrote:

Upon powering on, similarly to others here, there is a solid green light near the front panel header and a flashing green light near the USB3 header. With a serial adapter attached to each of the MCU and RISC-V headers, I do get some output on the MCU and am able to issue various commands successfully:

The fact that you are able to get some output from the MCU and issue commands is at least a good first sign.

You mention connection to the MCU and RISC-V consoles. Have you tried connecting to the main board console? I'm not sure if it's even documented anywhere, but I initially had similar problems and was given this image:

serial.png (view on web) https://github.com/milkv-pioneer/issues/assets/11022302/4453e1d7-2436-487a-9f43-7616699d641c

This will show output before the RISC-V console, and the RISC-V console will only start printing output after it completes. A normal log of it's output should look like this:

NOTICE: BOOT: 0x7000140000/0x1/0x5 NOTICE: Booting Trusted Firmware NOTICE: BL1: v2.7(release):83c4f9e3e NOTICE: BL1: Built : 10:39:27, Jul 12 2022 INFO: BL1: RAM 0x7010002000 - 0x7010011000 NOTICE: SD initializing 100000000Hz Hit key to stop autoboot: 00 NOTICE: BOOT: 0x7000140000/0x1/0x5 NOTICE: Booting Trusted Firmware NOTICE: BL1: v2.7(release):83c4f9e3e NOTICE: BL1: Built : 10:39:27, Jul 12 2022 INFO: BL1: RAM 0x7010002000 - 0x7010011000 NOTICE: SD initializing 100000000Hz NOTICE: boot from SD INFO: BL1: Loading BL2 NOTICE: Locate FIP in SD FAT INFO: Loading image id=1 at address 0x7010020000 INFO: Image id=1 loaded: 0x7010020000 - 0x701005ca41 NOTICE: BL1: Booting BL2 INFO: Entry point address = 0x7010020000 INFO: SPSR = 0x3c5 NOTICE: BL2: v2.7(release):bab7a07f0 NOTICE: BL2: Built : 20:23:04, Jun 14 2023 INFO: BL2: Doing platform setup NOTICE: SD initializing 100000000Hz NOTICE: Chip0 working at single socket, multi spi mode NOTICE: **DIMM0** NOTICE: DIMM0: Unknown Manufacturer RDIMM DDR4 SDRAM RAW_CARD_E 2Rx8 3200(3200) 32GB ECC(ON) NOTICE: **DIMM1** NOTICE: DIMM1: Unknown Manufacturer RDIMM DDR4 SDRAM RAW_CARD_E 2Rx8 3200(3200) 32GB ECC(ON) NOTICE: **DIMM2** NOTICE: DIMM2: Unknown Manufacturer RDIMM DDR4 SDRAM RAW_CARD_E 2Rx8 3200(3200) 32GB ECC(ON) NOTICE: **DIMM3** NOTICE: DIMM3: Unknown Manufacturer RDIMM DDR4 SDRAM RAW_CARD_E 2Rx8 3200(3200) 32GB ECC(ON) NOTICE: Waiting DDR0 ECC done NOTICE: Waiting DDR1 ECC done NOTICE: Waiting DDR2 ECC done NOTICE: Waiting DDR3 ECC done NOTICE: Socket id 0, online chip number 1 NOTICE: Socket0 is present NOTICE: Port0.0 RC gen4 16GT/s NOTICE: Port0.1 RC gen4 16GT/s NOTICE: Port1.0 RC gen4 16GT/s NOTICE: Port1.1 RC gen4 16GT/s NOTICE: RISC-V map: O O O O O O O O O O O O O O O O NOTICE: Socket1 is not present NOTICE: Setup CCIX NOTICE: Setup CMN NOTICE: Maximum compatibility mode NOTICE: DRAM0 [0000000000 - 0800000000] 32GB NOTICE: DRAM1 [0800000000 - 1000000000] 32GB NOTICE: DRAM2 [1000000000 - 1800000000] 32GB NOTICE: DRAM3 [1800000000 - 2000000000] 32GB NOTICE: Setup PLIC NOTICE: Setup PCIe NOTICE: PCIe1 Link1 linked with gen3 x8 RC mode NOTICE: Setup chip2chip link done NOTICE: RISC-V only mode

When I was having problems, initially with incompatible RAM, my output looked like this:

Hit key to stop autoboot: 00 NOTICE: BOOT: 0x7000140000/0x1/0x5 NOTICE: Booting Trusted Firmware NOTICE: BL1: v2.7(release):83c4f9e3e NOTICE: BL1: Built : 10:39:27, Jul 12 2022 INFO: BL1: RAM 0x7010002000 - 0x7010011000 NOTICE: SD initializing 100000000Hz NOTICE: boot from SD INFO: BL1: Loading BL2 NOTICE: Locate FIP in SD FAT INFO: Loading image id=1 at address 0x7010020000 INFO: Image id=1 loaded: 0x7010020000 - 0x701005ba81 NOTICE: BL1: Booting BL2 INFO: Entry point address = 0x7010020000 INFO: SPSR = 0x3c5 NOTICE: BL2: v2.7(release):f2d025ca1 NOTICE: BL2: Built : 10:09:05, May 9 2023 INFO: BL2: Doing platform setup NOTICE: MILKV Pioneer v1 NOTICE: SD initializing 100000000Hz NOTICE: **DIMM0** NOTICE: MANGO_SOPHGO_X8EVB don't support X4 RDIMM! NOTICE: De-Skew PLL failed to lock NOTICE: lck_0 = 0NOTICE: lck_1 = 0NOTICE: Deskew PLL Failed NOTICE: Controller initialized NOTICE: PI initialized NOTICE: DIMM0: Samsung RDIMM DDR4 SDRAM RAW_CARD_B 2Rx4 3200(3200) 32GB ECC(ON) time out! clear memory fail NOTICE: Chip0 working at single socket, multi spi mode NOTICE: Socket id 0, online chip number 1 NOTICE: Socket0 is present NOTICE: Port0.0 RC gen4 16GT/s NOTICE: Port0.1 RC gen4 16GT/s NOTICE: Port1.0 RC gen4 16GT/s NOTICE: Port1.1 RC gen4 16GT/s NOTICE: RISC-V map: O O O O O O O O O O O O O O O O NOTICE: Socket1 is not present NOTICE: Setup CCIX NOTICE: Setup CMN NOTICE: Maximum compatibility mode NOTICE: DRAM0 [0000000000 - 0800000000] 32GB NOTICE: Setup PLIC NOTICE: Setup PCIe NOTICE: PCIe0 Link0 linked with gen4 x8 RC mode NOTICE: PCIe1 Link1 linked with gen3 x8 RC mode NOTICE: Setup chip2chip link done

See the part time out! clear memory fail. The issue for me was using incompatible 2Rx4 memory.

However, if you're using the official recommended memory then it's hard to say. But at least you should potentially get some more clues as to what might be going wrong.

— Reply to this email directly, view it on GitHub https://github.com/milkv-pioneer/issues/issues/33#issuecomment-1919711235, or unsubscribe https://github.com/notifications/unsubscribe-auth/AS5EBIQEM7NKOBROHLWQVKDYRKFWLAVCNFSM6AAAAABCNDS4JSVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTSMJZG4YTCMRTGU . You are receiving this because you commented.Message ID: @.***>

MrDrMcCoy commented 7 months ago

@silvanshade Thanks for the tip on the undocumented serial headers. I just tried connecting to that on one of my serial interfaces with the other on the MCU. I got the usual output on the MCU, but garbled text on the main board console, and neither responded to any of my input.

Upon resetting and swapping the serial adapters, I got the MCU output again on the other interface, but nothing from the main board and the MCU still is not responding to my input. I'm hoping this thing isn't dead...

Some others have reported that the board is really picky about power supplies, so I ordered one that was reported working and have to wait a few days for it to arrive. Hopefully that does the trick.

If I may ask, how did you come across the main board header info?

silvanshade commented 7 months ago

I got the usual output on the MCU, but garbled text on the main board console, and neither responded to any of my input.

@MrDrMcCoy That's interesting. I think the very first couple times I tried to connect to the main board console I also got garbled text, but never really figured out why. The problem seemed to go away after I connected to all 3 consoles at once (main board, MCU, RISC-V) and rebooted a few times. I can also confirm it works at 115200 baud.

FWIW, the power supply I am using is the ASUS ROG Loki SFX-L 850W Platinum White.

If I may ask, how did you come across the main board header info?

I just emailed Milk-V and described my problems getting the board to boot and asked if there was any way to diagnose the problem and they emailed me that image. Not a super detailed response but it helped a little.

Anyway, hope you manage to get things booting.

rtcn2 commented 7 months ago

I have memory that appears on the supported list here: https://github.com/milkv-pioneer/pioneer-files/blob/main/hardware/sg2042_ddr_support_list.md

'HMA84GR7DJR4N-XN'

That appears to incompatible as per the serial output:

NOTICE:  MILKV Pioneer v1
NOTICE:  SD initializing 100000000Hz
NOTICE:  ******************DIMM0******************
NOTICE:  MANGO_SOPHGO_X8EVB don't support X4 RDIMM!
NOTICE:  De-Skew PLL failed to lock
NOTICE:  lck_0 = 0NOTICE:  lck_1 = 0NOTICE:  Deskew PLL Failed
NOTICE:  Controller initialized
NOTICE:  PI initialized
NOTICE:  DIMM0: SK Hynix (Hyundai) RDIMM DDR4 SDRAM RAW_CARD_B 2Rx4 3200(3200) 32GB ECC(ON)
 time out! clear memory fail
NOTICE:  Chip0 working at single socket, multi spi mode
NOTICE:  Socket id 0, online chip number 1
NOTICE:  Socket0 is present
NOTICE:      Port0.0 RC gen4 16GT/s
NOTICE:      Port0.1 RC gen4 16GT/s
NOTICE:      Port1.0 RC gen4 16GT/s
NOTICE:      Port1.1 RC gen4 16GT/s
NOTICE:  RISC-V map: O O O O O O O O O O O O O O O O
NOTICE:  Socket1 is not present
NOTICE:  Setup CCIX
NOTICE:  Setup CMN
NOTICE:  Maximum compatibility mode
NOTICE:  DRAM0 [0000000000 - 0800000000] 32GB

Any other things to check?

silvanshade commented 7 months ago

Any other things to check?

Despite the fact that the memory may be listed as compatible with the SG2042 on that list, as far as I know the Pioneer will not boot with 2Rx4. I tried a number of different modules, which also should have been supported according to that list, but none worked for me.

I would try switching to a 2Rx8 module if you can.

jonmasters commented 7 months ago

My guess is OP has a x4 RDIMM. Most of those NEMIX ones I've seen tend to be. I had one myself that failed to boot at all in a similar fashion. I recommend getting a x8 RDIMM. You can pick them up for a reasonable amount. Here's one I know to be working: https://www.newegg.com/micron-32gb/p/N82E16820156365?Item=N82E16820156365

The total lack of useful documentation about any of this is, of course, quite shameful for a board that advertises as being a workstation or server, and not an embedded YOLO dev board. If you want to be more than a random dev board for the very keen, you need to have more than bargain basement documentation.

MrDrMcCoy commented 7 months ago

@jonmasters Both of the RAM models I've tried are dual-rank, one is x8:

I have finally gotten my hands on an alternate power supply that has worked for others (EVGA 500 W1 100-W1-0500-KR 80+ 500W Power Supply), but I am still in the same state where the MCU prints text, but no longer accepts input, and the RISC-V and mainboard headers print nothing at all.

I guess all that's left is to email support and hope they're responsive, unless anyone has other ideas.

jonmasters commented 7 months ago

Did the MCU respond to any commands to power on?

MrDrMcCoy commented 7 months ago

Before I tried the mainboard serial port, the MCU responded to poweron and poweron_rv, but nothing showed on the RISC-V header. Since I tried attaching to the mainboard header, the MCU does not accept any input at all.

ppokorny commented 7 months ago

I placed two different orders from NEMIX for RAM "MR25600-628K01-G"

But received slightly different parts. Both Micron branded and the same Micron part number. But Micron documents 4 different variants of that part number with different PCB, Die revs and PLL Registers. I have two of the four. One works ending in "R1" and the other does not which ends with "E1" So even though they are both 16GB , 3200MHz 2Rx8, it seems the Pioneer is very sensitive to either the PLL/Register or the DRAM+PCB.

rtcn2 commented 7 months ago

Thanks for the suggestions.

I have tried some other ram with limited success. It seems to boot now with either of the new ram I've tried, but I haven't been able to boot on an SD Card, I have output on all 3 'Serial' ports, I have tried various makes but most of the time it boots from the OpenSpi and stop after initializing the usb devices. I then started stripping out all the devices like the gpu etc. Nothing changed until I removed my SSD. It then booted into Fedora. So now I need to work out how to get it to boot with the SSD in...

Memory Details and risc-v console output: HMAA4GR7CJR8N - XN

NOTICE:  Chip0 working at single socket, multi spi mode
NOTICE:  ******************DIMM0******************
NOTICE:  DIMM0: SK Hynix (Hyundai) RDIMM DDR4 SDRAM RAW_CARD_E 2Rx8 3200(3200) 32GB ECC(ON)
NOTICE:  Waiting DDR0 ECC done
NOTICE:  Socket id 0, online chip number 1
NOTICE:  Socket0 is present
NOTICE:      Port0.0 RC gen4 16GT/s
NOTICE:      Port0.1 RC gen4 16GT/s
NOTICE:      Port1.0 RC gen4 16GT/s
NOTICE:      Port1.1 RC gen4 16GT/s
NOTICE:  RISC-V map: O O O O O O O O O O O O O O O O
NOTICE:  Socket1 is not present
NOTICE:  Setup CCIX
NOTICE:  Setup CMN
NOTICE:  Maximum compatibility mode
NOTICE:  DRAM0 [0000000000 - 0800000000] 32GB

MTA9ASF1G72PZ-3G2J3VI

NOTICE:  Chip0 working at single socket, multi spi mode
NOTICE:  ******************DIMM0******************
NOTICE:  DIMM0: Micron Technology RDIMM DDR4 SDRAM RAW_CARD_D 1Rx8 3200(3200) 8GB ECC(ON)
NOTICE:  Waiting DDR0 ECC done
NOTICE:  Socket id 0, online chip number 1
NOTICE:  Socket0 is present
NOTICE:      Port0.0 RC gen4 16GT/s
NOTICE:      Port0.1 RC gen4 16GT/s
NOTICE:      Port1.0 RC gen4 16GT/s
NOTICE:      Port1.1 RC gen4 16GT/s
NOTICE:  RISC-V map: O O O O O O O O O O O O O O O O
NOTICE:  Socket1 is not present
NOTICE:  Setup CCIX
NOTICE:  Setup CMN
NOTICE:  Maximum compatibility mode
NOTICE:  DRAM0 [0000000000 - 0200000000] 8GB
AstxTyrol commented 4 months ago

My board was only booting sometimes with this RAM: KVR32N22S8/16 (in the SG2042 supported ddr list is KVR32N22S8/8 listed). First after many boot retries the board started with a single KVR32N22S8/16 and even Fedora booted OK showing 16GB RAM. After adding a second KVR32N22S8/16 the boot failed. The intervalls between success boots are increasing dramatically and after several hours the board didn't boot anymore with a single KVR32N22S8/16 Here the output log of the UART1 debug port (115200,8n1) BTW: found the reason why only garbage was seen at the debug UART1 port - see pictures below.

# Hit key to stop autoboot: 00 
NOTICE:  BOOT: 0x7000140000/0x1/0x5
Hit key to stop autoboot: 00 
NOTICE:  BOOT: 0x7000140000/0x1/0x5
NOTICE:  Booting Trusted Firmware
NOTICE:  BL1: v2.7(release):83c4f9e3e
NOTICE:  BL1: Built : 10:39:27, Jul 12 2022
INFO:    BL1: RAM 0x7010002000 - 0x7010011000
NOTICE:  SD initializing 100000000Hz
NOTICE:  boot from SD
INFO:    BL1: Loading BL2
NOTICE:  Locate FIP in SD FAT
INFO:    Loading image id=1 at address 0x7010020000
INFO:    Image id=1 loaded: 0x7010020000 - 0x701005ca51
NOTICE:  BL1: Booting BL2
INFO:    Entry point address = 0x7010020000
INFO:    SPSR = 0x3c5
NOTICE:  BL2: v2.7(release):b2aa55dc5-dirty
NOTICE:  BL2: Built : 16:54:08, Jan 17 2024
INFO:    BL2: Doing platform setup
NOTICE:  SD initializing 100000000Hz
NOTICE:  Chip0 working at single socket, multi spi mode
NOTICE:  ******************DIMM0******************
NOTICE:  DIMM0: Kingston UDIMM DDR4 SDRAM RAW_CARD_A 1Rx8 3200(3200) 16GB
NOTICE:  .
NOTICE:  DIMM0 check failed
NOTICE:  Waiting DDR0 ECC done
NOTICE:  Socket id 0, online chip number 1
NOTICE:  Socket0 is present
NOTICE:      Port0.0 RC gen4 16GT/s
NOTICE:      Port0.1 RC gen4 16GT/s
NOTICE:      Port1.0 RC gen4 16GT/s
NOTICE:      Port1.1 RC gen4 16GT/s
NOTICE:  RISC-V map: O O O O O O O O O O O O O O O O
NOTICE:  Socket1 is not present
NOTICE:  Setup CCIX
NOTICE:  Setup CMN
NOTICE:  Maximum compatibility mode
NOTICE:  DRAM0 [0000000000 - 0400000000] 16GB
NOTICE:  Setup PLIC
NOTICE:  Setup PCIe
NOTICE:  PCIe1 Link1 linked with gen3 x8 RC mode
NOTICE:  Setup chip2chip link done
ERROR:   DIMM0 software test failed
NOTICE:  RISC-V only mode

BTW: found the reason why only garbage was seen at the debug UART port: IMG_20240420_124109 IMG_20240420_132209

The UART1 output is very noisy - a permanent 20MHz 800mVpp signal poisoned the output around 0/1 schmitt trigger detection of the attached usb2rs232 converter. A simple RC filter (4k resistor and 1nF capacitor) helped to de-noise the UART stream. Never seen such HF problems on the output of an UART. IMHO: Mainboard quality is not near V1.3 - would give it a V0.5RC1 or so ...

AstxTyrol commented 4 months ago

Tested Kingston modules from Sophgo ddr list: KVR32N22S8/8

Simply not working - I will send the board back if even the supported RAM doesn't work

Hit key to stop autoboot: 00 
NOTICE:  BOOT: 0x7000140000/0x1/0x5
NOTICE:  Booting Trusted Firmware
NOTICE:  BL1: v2.7(release):83c4f9e3e
NOTICE:  BL1: Built : 10:39:27, Jul 12 2022
INFO:    BL1: RAM 0x7010002000 - 0x7010011000
NOTICE:  SD initializing 100000000Hz
NOTICE:  boot from SD
INFO:    BL1: Loading BL2
NOTICE:  Locate FIP in SD FAT
INFO:    Loading image id=1 at address 0x7010020000
INFO:    Image id=1 loaded: 0x7010020000 - 0x701005ba81
NOTICE:  BL1: Booting BL2
INFO:    Entry point address = 0x7010020000
INFO:    SPSR = 0x3c5
NOTICE:  BL2: v2.7(release):f2d025ca1
NOTICE:  BL2: Built : 10:09:05, May  9 2023
INFO:    BL2: Doing platform setup
NOTICE:  MILKV Pioneer v1
NOTICE:  SD initializing 100000000Hz
NOTICE:  ******************DIMM0******************
NOTICE:  DIMM0: Kingston UDIMM DDR4 SDRAM RAW_CARD_A 1Rx8 3200(3200) 8GB
NOTICE:  .
NOTICE:  DIMM0 check failed
NOTICE:  ******************DIMM1******************
NOTICE:  DIMM1: Kingston UDIMM DDR4 SDRAM RAW_CARD_A 1Rx8 3200(3200) 8GB
NOTICE:  .
NOTICE:  DIMM1 check failed
NOTICE:  ******************DIMM2******************
NOTICE:  DIMM2: Kingston UDIMM DDR4 SDRAM RAW_CARD_A 1Rx8 3200(3200) 8GB
NOTICE:  .
NOTICE:  DIMM2 check failed
NOTICE:  ******************DIMM3******************
NOTICE:  DIMM3: Kingston UDIMM DDR4 SDRAM RAW_CARD_A 1Rx8 3200(3200) 8GB
NOTICE:  .
NOTICE:  DIMM3 check failed
NOTICE:  Chip0 working at single socket, multi spi mode
NOTICE:  Socket id 0, online chip number 1
NOTICE:  Socket0 is present
NOTICE:      Port0.0 RC gen4 16GT/s
NOTICE:      Port0.1 RC gen4 16GT/s
NOTICE:      Port1.0 RC gen4 16GT/s
NOTICE:      Port1.1 RC gen4 16GT/s
NOTICE:  RISC-V map: O O O O O O O O O O O O O O O O
NOTICE:  Socket1 is not present
NOTICE:  Setup CCIX
NOTICE:  Setup CMN
NOTICE:  Maximum compatibility mode
NOTICE:  DRAM0 [0000000000 - 0200000000] 8GB
NOTICE:  DRAM1 [0200000000 - 0400000000] 8GB
NOTICE:  DRAM2 [0400000000 - 0600000000] 8GB
NOTICE:  DRAM3 [0600000000 - 0800000000] 8GB
NOTICE:  Setup PLIC
NOTICE:  Setup PCIe
NOTICE:  PCIe0 Link0 linked with gen1 x1 RC mode
NOTICE:  PCIe1 Link1 linked with gen3 x8 RC mode
NOTICE:  Setup chip2chip link done
ERROR:   DIMM0 software test failed
NOTICE:  RISC-V only mode
u0076 commented 4 months ago

Tested Kingston modules from Sophgo ddr list: KVR32N22S8/8

Some DDR have been tested on EVB, but not on Pioneer. We should advise users to buy a DDR that has passed the test in Pioneer. This list does not explicitly state this, which is our problem and we will update it.

We will try to make KVR32N22S8/8 work on Pioneer, which will take some time.

u0076 commented 4 months ago

The UART1 output is very noisy - a permanent 20MHz 800mVpp signal poisoned the output around 0/1 schmitt trigger detection of the attached usb2rs232 converter. A simple RC filter (4k resistor and 1nF capacitor) helped to de-noise the UART stream.

Thank you for pointing out the reason for SCP UART garbled characters, we will fix it in the next revision.

ad-zsolt-imre commented 1 month ago

@u0076 Regarding the SCP UART, was it not tested? The USB-to-UART converter I received with the board produces garbled characters. Is there a recommended, ready-made converter that works with the board? @AstxTyrol solved it, but I'd prefer something I just have to plug in and works fine.

To be honest, I'm kind of losing confidence, so I have to ask: if I invested in the supported RAMs and a supported video card and the board still cannot be made to work, do you have a reimbursement program?

Regarding the limited video card support, just to know for sure: is it a limitation of the board or it's a kernel driver limitation the Fedora image comes with?