It seems like all the work is already done for the CI to upload the release assets.
But there are no release assets to download.
What am I missing?

Would be great if the debs would be available for download.

A number of important changes have accumulated since 3.2.0, including support for additional ethernet boards and a fix for a problem accessing PCI devices on some platforms.

Hi, I am the proud owner of a shiny new 7i96s card and i have gotten recommendations to build this tool from source, as the tool currently on Debian repos is outdated. I see that your code identifies itself as 3.4.6, but in the Debian control files there is no changelog after 3.4.2. Would a PR be accepted to update that so that the ball can get rolling on the packaging side?

Thanks!

I have a 7C81 that I flashed with 7c81epp.bit and now needs to be flashed via EPP mode.

I was having trouble so I uncommented some of the debugging printfs in the epp .c file for mesaflash.

With the device unpowered but attached the following happens:

root@mora:~/mesaflash# ./mesaflash --device 7C81 --epp --verbose --addr 0x378 --readhmid
wrote control 0x04
read status 0x78
selected address 0x8000
read data 0x80808080
read status 0x78
selected address 0x8100
read data 0x81818181
read status 0x78
No 7C81 board found

And then powering it on:

root@mora:~/mesaflash# ./mesaflash --device 7C81 --epp --verbose --addr 0x378 --readhmid
wrote control 0x04
read status 0xF8
selected address 0x8000
read data 0xFFFFFFFF
read status 0xF8
selected address 0x8100
read data 0x55AACAFE
read status 0xF8
selected address 0x8000
read data 0xFF
read data 0xFF
read data 0xFF
read data 0xFF
read status 0xF8
selected address 0x8100
read data 0xFE
read data 0xCA
read data 0xAA
read data 0x55
read status 0xF8
selected address 0x810C
read data 0x00
read data 0x04
read data 0x00
read data 0x00
selected address 0x840C
read data 0x4D
read data 0x45
read data 0x53
read data 0x41
read data 0x37
read data 0x43
read data 0x38
read data 0x31
No 7C81 board found

It clearly sees the card... the hex string 4D 45 53 41 37 43 38 31 is "MESA7C81" which is the fingerprint for the card.

I'm looking at the code to see if it's something simple I can "fix" but I figured I should report this.

mesaflash does this mapping in multiple ways:

• mod_names[]
• mod_names_xml[]
• hm2_get_general_function_name()

Seems like that could be simplified.

seems to be a x86 vs amd64 issue - same card:

amd64:

root@amd64:/home/mah/external-clocking# mesaflash --device 5i25 --verbose --verify 7i76x2d.bit
Checking file... OK
  File type: BIT file
  Design name: TopPCIHostMot2.ncd;UserID=0xFFFFFFFF
  Part name: 6slx9tqg144
  Design date: 2015/06/16
  Design time: 17:20:45
  Config Length: 340604
Error: BootSector is invalid

x86:

root@dumpster:/home/mah/external-clocking# mesaflash --device 5i25 --verbose --write 7i76x2d.bit
Checking file... OK
  File type: BIT file
  Design name: TopPCIHostMot2.ncd;UserID=0xFFFFFFFF
  Part name: 6slx9tqg144
  Design date: 2015/06/16
  Design time: 17:20:45
  Config Length: 340604
Boot sector OK
EEPROM sectors to write: 6, max sectors in area: 8
Erasing EEPROM sectors starting from 0x80000...
  |EEEEEE
  Erasing time: 3.43 seconds
Programming EEPROM sectors starting from 0x80000...
  |WWWWWWWWWWWWWWWWWWWWWWWWWWWWWWWWWWWWWWWWWW
  Programming time: 1.616 seconds
Board configuration updated successfully.

You must power cycle board or use --reload command to load updated firmware.

There are branches, not tags, for the 3.1.0 and 3.2.0 releases. Please add annotated or signed tags so that they appear as "releases" on github. Feel free to hit me up on IRC if you are having trouble making and pushing tags.

"No 7I90 board found" on versions 3.4.3 forward:

mesaflash-3.4.2$ sudo mesaflash --device 7i90 --epp --addr 0x378 --readhmid

Configuration Name: HOSTMOT2
General configuration information:
BoardName : MESA7I90
FPGA Size: 9 KGates

mesaflash-3.4.3$ sudo mesaflash --device 7i90 --epp --addr 0x378 --readhmid

No 7I90 board found

mesaflash --device 7i96 --addr 10.10.10.10 --readhmid results in:

10     30   IOPort       SSerial          0        RXData1         (In)
11     31   IOPort       SSerial          0        TXData1         (Out)
12     32   IOPort       SSerial          0        TXEn1           (Out)
Segmentation fault (core dumped)

Hi. Any chance of making this buildable on BSD (i.e. macOS)? Thanks!

While debugging issue #4 I got a worrisome message:

Erasing sector 0 for boot block
BootBlock installed
Failed to write valid boot sector

If a valid boot block is not present, it's a sign that something is already quite wrong. But trying to write the boot block unrequested might even make things worse, (if there really was a valid boot block but a bug made it appear to be invalid)

I worry that if I had power cycled after "failed to write valid boot sector" I might have been in a situation where I had to use JTAG to rescue my board, which I don't have immediately on hand and many users won't have it either.

Linux2.8.0 .iso installation includes gcc but not make.
It would be helpful to include directions to install build-essential prior to the make instruction.
sudo apt install build-essential

Hallo.

If I look through encoder_module.c, I see that it works only with HM2_GTAG_MUXED_ENCODER* .

Would it be enough to substitute any occurrence of HM2_GTAG_MUXED_ENCODER* through
HM2_GTAG_ENCODER* to make it work with a "normal" encoder?

Thanks.

The current Makefile provides partial support for the sys/io.h stubs implementation, missing is an obvious way to enable it.

Please update the Makefile with the following:

#enable for sys/io.h stubs implementation
#USE_STUBS = 1

A Raspberry Pi build note in the README.md file:

To build on a Raspberry Pi:
   Edit the Makefile to enable for sys/io.h stubs implementation.