Current Server Time: 15:22 (Central Europe)

#apertus IRC Channel Logs

2021/01/12

Timezone: UTC


23:03
mumptai
left the channel
23:49
eppisai
left the channel
04:38
eppisai
joined the channel
05:07
BAndiT1983|away
changed nick to: BAndiT1983
05:10
Bertl_oO
off to bed now ... have a good one everyone!
05:11
Bertl_oO
changed nick to: Bertl_zZ
06:30
RexOrCine
joined the channel
06:31
RexOrCine1
left the channel
07:12
Spirit532
left the channel
07:12
Spirit532
joined the channel
07:39
mumptai
joined the channel
09:25
eppisai
left the channel
09:28
eppisai
joined the channel
09:33
eppisai
left the channel
11:27
Bertl_zZ
changed nick to: Bertl
11:27
Bertl
morning folks!
11:27
BAndiT1983
hi
13:14
anuejn
morning :)
14:02
eppisai
joined the channel
14:17
eppisai
left the channel
14:18
eppisai
joined the channel
14:31
EmilJ
hi there!
14:33
EmilJ
What's supposed to be present in the flash of a zynq running axiom-firmware?
14:37
vup2
it doesn't care about the flash (if you mean no th sdcard)
15:18
Bertl
where 'no th' is short for 'not the' :)
15:21
BAndiT1983
needs to be added to urban dictionary
15:22
BAndiT1983
https://www.urbandictionary.com/define.php?term=no%20th
15:22
Bertl
yeah, well, maybe I'm wrong and it means something completely different ...
15:23
vup2
hehe
15:24
eppisai
left the channel
15:33
EmilJ
Wait, the zynq is booting directly from SD card?
15:33
EmilJ
In that case, how the heck does it refuse to boot
15:35
BAndiT1983
https://www.bastibl.net/zync-boot/#:~:text=When%20powering%20up%20the%20device,dip%20switches%20on%20the%20board.
15:36
BAndiT1983
https://www.centennialsoftwaresolutions.com/post/zynq-7000-boot-process
16:09
vup2
EmilJ: it depends on how the boot is configured, we have configured it such that it tries to boot from SD card first
16:10
vup2
usually there are some jumpers that you can use to select the boot mode
16:11
vup2
The biggest reason why it would not work when booting from SD-card using a axiom-firmware image would probably be the MIO / RAM configuration, as that is hardcoded in a register init sequence generated by vivado and then imported into u-boot
16:11
vup2
this is on of the main reason there is a micro and a beta firmware image, they are incompatible because they need to be initialized differently
16:12
vup2
the beta image does not boot on the micro and vice versa
16:12
EmilJ
Yeah, I do have the jumper in the SD position. That's interesting. You mean the RAM would be remapped on my board compared to for example zybo? I'll check the schematics
16:16
vup2
well not only that, also the training / timings might be different
16:16
vup2
or some clocking has to be different
16:18
vup2
(you can find semi plain text versions of the init stuff here: https://files.niemo.de/zturn_lite.dump https://files.niemo.de/microzed_ps7_init.dump if you are interested what registers are set)
16:20
vup2
in the ps7_ddrc_DRAM_param* registers for example you can see the RAM timings are hardcoded
16:21
EmilJ
I kinda thought the DDR gets retrained on boot
16:22
vup2
maybe the fsbl can do that
16:22
vup2
but i don't see any u-boot code doing that
16:24
vup2
(fyi the zynq boots first from a internal bootrom, which in turn copies the BOOT.bin from the selected boot medium to the on-chip ram and basically then jumps into it, then the u-boot spl (packaged in the BOOT.bin) starts out executing from the on-chip ram, initializes the DDR and other stuff, copies the full u-boot to DDR and hands over to that)
16:25
vup2
if you use the fsbl it replaces the u-boot spl (second program loader) in this flow
16:25
vup2
so maybe the fsbl can to dynamic ddr configuration
16:25
vup2
but iirc it did not look like it
16:25
vup2
it basically uses the same file with register initialization as the u-boot spl
16:31
EmilJ
so then, I should figure out how to dump the DDR configuration data from the pynq image and transplant that as a patch into, hm, boot.bin?
16:33
vup2
well that could maybe work
16:33
vup2
but I think it would be easier to try and figure out how BOOT.bin is built for the pynq
16:36
Obsdark
joined the channel
16:36
Obsdark
left the channel
16:36
Obsdark
joined the channel
16:37
vup2
EmilJ: do you have a Pynq-Z1 or Pynq-Z2?
16:37
EmilJ
Z2
16:38
vup2
I think everything need should be here https://github.com/Xilinx/PYNQ
16:38
vup2
(or should be possible to generate from the files there)
16:39
EmilJ
I'm going through the repo right now
16:43
vup2
right
16:43
vup2
I think we want a older branch
16:43
EmilJ
Why?
16:43
vup2
v2.2 seems nice, as it does not use peta linux yet
16:43
EmilJ
Ohhh!
17:06
vup2
combining vivado 2019.2 with the SDK from 2018.2, what could go wrong...
18:35
vup2
EmilJ: https://files.niemo.de/pynqz2_ps7_init_gpl.dump
18:35
vup2
comparing that to the zturn dump certainly shows some differences
18:35
vup2
(in the ddr controller configuration)
18:43
vup2
EmilJ: lets see if this runs through: https://github.com/apertus-open-source-cinema/axiom-firmware/pull/181
19:24
vup2
EmilJ: https://api.cirrus-ci.com/v1/artifact/task/5951953832247296/image_upload/build/axiom-pynqz2-pr-181-a45f2783.img.xz
20:17
EmilJ
oh wow thanks for the quick fixes - any idea why the CI test fails?
21:02
vup2
well it fails because qemu spews nothing on the serial when trying to boot
21:02
vup2
but that colud be due to the pynq using a different serial port then the others
21:03
vup2
(although that should be automatically handled by qemu as it loads the same devicetree as the kernel...)
21:26
Obsdark
left the channel
21:45
se6ast1an
off to bed, good night
22:06
EmilJ
Yeaahhhh, I get nothing on the "prog uart"
22:06
EmilJ
with the image you set up
22:30
mumptai
left the channel
22:37
vup2
what is the prog uart?
22:45
BAndiT1983
changed nick to: BAndiT1983|away
22:54
Bertl
The PYNQ-Z2 includes an FTDI FT2232HLUSB-UART bridge (attached to connector J8 PROG UART) that supports USB-JTAG, USB-UART. The PYNQ-Z2 can also be powered from the MicroUSB port.
22:55
vup2
hmm is the /proc/cmdline on a working system?
22:57
mumptai
joined the channel