AGX Wireless bug fix #259

emrahbillur · 2023-09-11T09:13:45Z

Separation of netvmExtraModules caused a bug in agx wireless, where this was fixed by adding appropriate corrections. Wireless condition for agx is improved. A small fix was added for docker (where it is not included if it is not enabled) in case it is planned to add for support.

emrahbillur · 2023-09-11T10:42:32Z

https://ssrc.atlassian.net/browse/SP-2906 Bug was introduced by netvmExtraModules in hardware was overriding the netvmExtraModules in target. Fixed that issue. There were unnecessary if statements in som type checking for enabling wireless, fixed. Added a fix for docker if somehow being added but this will not introduce any delay as if docker support is not included the fix is also not included.

targets/nvidia-jetson-orin.nix

vilvo · 2023-09-12T06:33:05Z

Added a fix for docker if somehow being added but this will not introduce any delay as if docker support is not included the fix is also not included.

I'd prefer the WiFi fix and Docker fix as separate commits/PRs as they are not related to each other. There's practical outcome from this - if the Docker overlay was a separate commit, it's easier to revert the overlay commit when we eventually get the fix from the upstream. The whole docker daemon enabling and the docker group are also separate from overlaying the Docker.

vilvo · 2023-09-12T09:29:09Z

nix build .#packages.aarch64-linux.nvidia-jetson-orin-agx-debug (using a remote builder) fails to build - it gets stuck to (with --debug):
[2/0/234 built] building ell-0.5.6 on ssh://[email protected]: debug1: pledge: fork

pledge: fork seems to come from ssh clientloop.c - http://bxr.su/OpenBSD/usr.bin/ssh/clientloop.c#960

I already rebooted the remote builder, but it repeats. I'll check if it's build host related.

vilvo · 2023-09-12T09:33:49Z

I'll check if it's build host related.

Nope, still the same. Ends up stuck in pledge: fork immediately.

emrahbillur · 2023-09-12T09:44:39Z

I'll seperate both sides and recommit with only WiFi fix.

emrahbillur · 2023-09-12T09:46:18Z

I'll check if it's build host related.

Nope, still the same. Ends up stuck in pledge: fork immediately.

I checked the builds for many times for both flash-script and image flakes of agx and image flakes of nx of but I'll try a rebuild switch to check again.

emrahbillur · 2023-09-12T09:49:39Z

And a small note to add. The issue in docker was due to a bug in Go compiler. So that fix will also solve issues in many other Go compiler based applications. It is a docker fix but I'll rename the commit of docker fix as "Docker fix due to Go issue".

vilvo · 2023-09-12T09:49:52Z

I'll check if it's build host related.

Nope, still the same. Ends up stuck in pledge: fork immediately.

I checked the builds for many times for both flash-script and image flakes of agx and image flakes of nx of but I'll try a rebuild switch to check again.

Ok, thanks for confirming. I'm building it now directly on AGX without remote ssh builder connection.

emrahbillur · 2023-09-12T09:54:03Z

I'll check if it's build host related.

Nope, still the same. Ends up stuck in pledge: fork immediately.

I checked the builds for many times for both flash-script and image flakes of agx and image flakes of nx of but I'll try a rebuild switch to check again.

Ok, thanks for confirming. I'm building it now directly on AGX without remote ssh builder connection.

I've removed the docker fix from my local. After confirming build is correct I'll commit and rename the pull request for Wifi Fix. Then I'll later add a PR with Docker fix due to Go compiler issue.

vilvo · 2023-09-12T11:33:29Z

I'm building it now directly on AGX without remote ssh builder connection.

Gets stuck on:

[nix-shell:~/ghaf]$ nix build .#packages.aarch64-linux.nvidia-jetson-orin-agx-debug
[0/235 built, 0.0 MiB DL] waiting for lock on '/nix/store/28lpb19hzlw9v5a0alckln2dgfr03wmk-ell-0.56','/n`

This happens also on local build on aarch64 hardware.

vilvo · 2023-09-12T11:37:47Z

Gets stuck on:

[nix-shell:~/ghaf]$ nix build .#packages.aarch64-linux.nvidia-jetson-orin-agx-debug
[0/235 built, 0.0 MiB DL] waiting for lock on '/nix/store/28lpb19hzlw9v5a0alckln2dgfr03wmk-ell-0.56','/n`

Apparently I'm facing this NixOS/nix#2029

emrahbillur · 2023-09-12T12:08:18Z

Gets stuck on:

[nix-shell:~/ghaf]$ nix build .#packages.aarch64-linux.nvidia-jetson-orin-agx-debug
[0/235 built, 0.0 MiB DL] waiting for lock on '/nix/store/28lpb19hzlw9v5a0alckln2dgfr03wmk-ell-0.56','/n`

Apparently I'm facing this NixOS/nix#2029

I'll commit a Wireless fix only PR just after check. So you could try on that.

Signed-off-by: Emrah Billur <[email protected]> Remove docker Signed-off-by: Emrah Billur <[email protected]>

emrahbillur · 2023-09-12T12:14:08Z

Replaced the PR into only wireless for agx fix.

vilvo · 2023-09-12T12:22:06Z

I'll commit a Wireless fix only PR just after check. So you could try on that.

Still the same:

[nix-shell:~/ghaf]$ git log --oneline -1
54bca8e (HEAD, emrah/main) AGX Wireless bug fix

[nix-shell:~/ghaf]$ nix build .#packages.aarch64-linux.nvidia-jetson-orin-agx-debug
[0/235 built] waiting for lock on ...

emrahbillur · 2023-09-12T12:24:27Z

I'll commit a Wireless fix only PR just after check. So you could try on that.

Still the same:
[nix-shell:~/ghaf]$ git log --oneline -1
54bca8e (HEAD, emrah/main) AGX Wireless bug fix

[nix-shell:~/ghaf]$ nix build .#packages.aarch64-linux.nvidia-jetson-orin-agx-debug
[0/235 built] waiting for lock on ...

I'll commit a Wireless fix only PR just after check. So you could try on that.

Still the same:

[nix-shell:~/ghaf]$ git log --oneline -1
54bca8e (HEAD, emrah/main) AGX Wireless bug fix

[nix-shell:~/ghaf]$ nix build .#packages.aarch64-linux.nvidia-jetson-orin-agx-debug
[0/235 built] waiting for lock on ...

I'll commit a Wireless fix only PR just after check. So you could try on that.

Still the same:

[nix-shell:~/ghaf]$ git log --oneline -1
54bca8e (HEAD, emrah/main) AGX Wireless bug fix

[nix-shell:~/ghaf]$ nix build .#packages.aarch64-linux.nvidia-jetson-orin-agx-debug
[0/235 built] waiting for lock on ...

I'll commit a Wireless fix only PR just after check. So you could try on that.

Still the same:

[nix-shell:~/ghaf]$ git log --oneline -1
54bca8e (HEAD, emrah/main) AGX Wireless bug fix

[nix-shell:~/ghaf]$ nix build .#packages.aarch64-linux.nvidia-jetson-orin-agx-debug
[0/235 built] waiting for lock on ...

root@emrah-ThinkPad-P14s-Gen-3:/home/emrah/ghaf-support-for-nx# nix build .#packages.aarch64-linux.nvidia-jetson-orin-nx-debug
warning: Git tree '/home/emrah/ghaf-support-for-nx' is dirty
Then builds with no issues. Same for nix build .#packages.aarch64-linux.nvidia-jetson-orin-agx-debug

vilvo · 2023-09-12T12:33:51Z

I get following lines to the kernel log - multiple times:

[  846.797982] INFO: task test-cipher:15875 blocked for more than 724 seconds.
[  846.798207]       Tainted: G           O      5.10.104 #1-NixOS
[  846.798376] "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
[  846.798595] task:test-cipher     state:D stack:    0 pid:15875 ppid: 15836 flags:0x00000804
[  846.798599] Call trace:
[  846.798608]  __switch_to+0xc8/0x120
[  846.798615]  __schedule+0x2bc/0x824
[  846.798616]  schedule+0x50/0xd0
[  846.798619]  schedule_timeout+0x160/0x1a0
[  846.798620]  __wait_for_common+0xcc/0x220
[  846.798622]  wait_for_completion+0x34/0x40
[  846.798627]  skcipher_recvmsg+0x314/0x3ec
[  846.798632]  sock_read_iter+0xf0/0x120
[  846.798637]  new_sync_read+0x18c/0x1a0
[  846.798639]  vfs_read+0x134/0x1c0
[  846.798640]  ksys_read+0xec/0x10c
[  846.798641]  __arm64_sys_read+0x28/0x34
[  846.798645]  el0_svc_common.constprop.0+0x80/0x1c4
[  846.798647]  do_el0_svc+0x38/0xac
[  846.798650]  el0_svc+0x1c/0x30
[  846.798652]  el0_sync_handler+0xe0/0x10c
[  846.798654]  el0_sync+0x16c/0x180

targets/nvidia-jetson-orin.nix

emrahbillur · 2023-09-13T11:06:55Z

INFO: task test-cipher:15875 blocked for more than 724 seconds

I'm trying to reproduce the same error but could not manage to get this. I remember in some cross compilation configuration cipher part caused errors but still not sure. I'll also try cross build now.

mikatammi · 2023-09-13T11:17:37Z

Are we ready to merge? I'm just testing this and if everything goes ok I will push the merge button

mikatammi · 2023-09-13T11:21:39Z

In my tests everything go ok, and the NetVM connects to the WLAN network

vilvo · 2023-09-13T11:38:14Z

INFO: task test-cipher:15875 blocked for more than 724 seconds

I'm trying to reproduce the same error but could not manage to get this. I remember in some cross compilation configuration cipher part caused errors but still not sure. I'll also try cross build now.

I changed my remote builder from Orin AGX (ghaf/nixox) to M1 (nixos) and it passed building the problematic package - it's still building the chromium-unwrapped.

emrahbillur force-pushed the main branch from b91ef32 to d1d823b Compare September 11, 2023 10:37

emrahbillur requested review from leivos-unikie and vilvo September 11, 2023 10:39

vilvo reviewed Sep 12, 2023

View reviewed changes

targets/nvidia-jetson-orin.nix Outdated Show resolved Hide resolved

vilvo reviewed Sep 12, 2023

View reviewed changes

targets/nvidia-jetson-orin.nix Show resolved Hide resolved

emrahbillur force-pushed the main branch from d1d823b to 1719a77 Compare September 12, 2023 12:11

AGX Wireless bug fix

54bca8e

Signed-off-by: Emrah Billur <[email protected]> Remove docker Signed-off-by: Emrah Billur <[email protected]>

emrahbillur force-pushed the main branch from 1719a77 to 54bca8e Compare September 12, 2023 12:12

emrahbillur requested a review from mikatammi September 12, 2023 12:14

mikatammi approved these changes Sep 13, 2023

View reviewed changes

targets/nvidia-jetson-orin.nix Show resolved Hide resolved

mikatammi approved these changes Sep 13, 2023

View reviewed changes

targets/nvidia-jetson-orin.nix Show resolved Hide resolved

emrahbillur changed the title ~~AGX Wireless bug and Docker bug fix~~ AGX Wireless bug fix Sep 13, 2023

mikatammi merged commit 3372f53 into tiiuae:main Sep 13, 2023
3 checks passed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

AGX Wireless bug fix #259

AGX Wireless bug fix #259

emrahbillur commented Sep 11, 2023

emrahbillur commented Sep 11, 2023

vilvo commented Sep 12, 2023

vilvo commented Sep 12, 2023

vilvo commented Sep 12, 2023 •

edited

Loading

emrahbillur commented Sep 12, 2023

emrahbillur commented Sep 12, 2023

emrahbillur commented Sep 12, 2023

vilvo commented Sep 12, 2023

emrahbillur commented Sep 12, 2023

vilvo commented Sep 12, 2023

vilvo commented Sep 12, 2023

emrahbillur commented Sep 12, 2023

emrahbillur commented Sep 12, 2023

vilvo commented Sep 12, 2023

emrahbillur commented Sep 12, 2023

vilvo commented Sep 12, 2023

emrahbillur commented Sep 13, 2023

mikatammi commented Sep 13, 2023

mikatammi commented Sep 13, 2023

vilvo commented Sep 13, 2023

AGX Wireless bug fix #259

AGX Wireless bug fix #259

Conversation

emrahbillur commented Sep 11, 2023

emrahbillur commented Sep 11, 2023

vilvo commented Sep 12, 2023

vilvo commented Sep 12, 2023

vilvo commented Sep 12, 2023 • edited Loading

emrahbillur commented Sep 12, 2023

emrahbillur commented Sep 12, 2023

emrahbillur commented Sep 12, 2023

vilvo commented Sep 12, 2023

emrahbillur commented Sep 12, 2023

vilvo commented Sep 12, 2023

vilvo commented Sep 12, 2023

emrahbillur commented Sep 12, 2023

emrahbillur commented Sep 12, 2023

vilvo commented Sep 12, 2023

emrahbillur commented Sep 12, 2023

vilvo commented Sep 12, 2023

emrahbillur commented Sep 13, 2023

mikatammi commented Sep 13, 2023

mikatammi commented Sep 13, 2023

vilvo commented Sep 13, 2023

vilvo commented Sep 12, 2023 •

edited

Loading