Flint 2 (GL-MT6000 ) - bug reports - collective thread

Installing NFS v4 nfs-kernel-server does not work. NFS is crucial for me. Should I report this on a separate thread ?


root@GL-MT6000:~# opkg install nfs-kernel-server
Installing nfs-kernel-server (2.5.2-1) to root...
Downloading https://fw.gl-inet.com/releases/mt798x/packages-4.5/aarch64_cortex-a53/packages/nfs-kernel-server_2.5.2-1_aarch64_cortex-a53.ipk
Collected errors:
 * pkg_hash_check_unresolved: cannot find dependency kmod-dm for libdevmapper-selinux
 * pkg_hash_check_unresolved: cannot find dependency kmod-dm for libdevmapper
 * pkg_hash_fetch_best_installation_candidate: Packages for libdevmapper found, but incompatible with the architectures configured
 * satisfy_dependencies_for: Cannot satisfy the following dependencies for nfs-kernel-server:
 *      kmod-dm
 * opkg_install_cmd: Cannot install package nfs-kernel-server.

Did you try this already?

This is already the existing content of distfeeds.conf

Anyone got DDNS test failed all the time error ? (4.6.0 snapshot 20240521 )

Not here, 4.5.8 -> 4.6 snapshot caused my ping times to jump about all over the place from 1 - 100ms and WAN is constantly crashing.

Tue May 21 10:08:18 2024 kern.info kernel: [244763.073032] QDMA Tx Info
Tue May 21 10:08:18 2024 kern.info kernel: [244763.075659] err_cnt = 4
Tue May 21 10:08:18 2024 kern.info kernel: [244763.075661] is_qfsm_hang = 1
Tue May 21 10:08:18 2024 kern.info kernel: [244763.081136] is_qfwd_hang = 1
Tue May 21 10:08:18 2024 kern.info kernel: [244763.084092] -- -- -- -- -- -- --
Tue May 21 10:08:18 2024 kern.info kernel: [244763.087396] MTK_QDMA_FSM = 0x100
Tue May 21 10:08:18 2024 kern.info kernel: [244763.090699] MTK_QDMA_FWD_CNT = 0x0
Tue May 21 10:08:18 2024 kern.info kernel: [244763.094177] MTK_QDMA_FQ_CNT = 0x44e0800
Tue May 21 10:08:18 2024 kern.info kernel: [244763.098087] ==============================
Tue May 21 10:08:19 2024 kern.info kernel: [244764.129038] QDMA Tx Info
Tue May 21 10:08:19 2024 kern.info kernel: [244764.131656] err_cnt = 5
Tue May 21 10:08:19 2024 kern.info kernel: [244764.131657] is_qfsm_hang = 1
Tue May 21 10:08:20 2024 kern.info kernel: [244764.137146] is_qfwd_hang = 1
Tue May 21 10:08:20 2024 kern.info kernel: [244764.140099] -- -- -- -- -- -- --
Tue May 21 10:08:20 2024 kern.info kernel: [244764.143405] MTK_QDMA_FSM = 0x100
Tue May 21 10:08:20 2024 kern.info kernel: [244764.146711] MTK_QDMA_FWD_CNT = 0x0
Tue May 21 10:08:20 2024 kern.info kernel: [244764.150190] MTK_QDMA_FQ_CNT = 0x44e0800
Tue May 21 10:08:20 2024 kern.info kernel: [244764.154102] ==============================
Tue May 21 10:08:21 2024 kern.info kernel: [244765.185021] QDMA Tx Info
Tue May 21 10:08:21 2024 kern.info kernel: [244765.187647] err_cnt = 6
Tue May 21 10:08:21 2024 kern.info kernel: [244765.187648] is_qfsm_hang = 1
Tue May 21 10:08:21 2024 kern.info kernel: [244765.193122] is_qfwd_hang = 1
Tue May 21 10:08:21 2024 kern.info kernel: [244765.196077] -- -- -- -- -- -- --
Tue May 21 10:08:21 2024 kern.info kernel: [244765.199379] MTK_QDMA_FSM = 0x100
Tue May 21 10:08:21 2024 kern.info kernel: [244765.202679] MTK_QDMA_FWD_CNT = 0x0
Tue May 21 10:08:21 2024 kern.info kernel: [244765.206155] MTK_QDMA_FQ_CNT = 0x44e0800
Tue May 21 10:08:21 2024 kern.info kernel: [244765.210064] ==============================
Tue May 21 10:08:22 2024 kern.info kernel: [244766.241015] QDMA Tx Info
Tue May 21 10:08:22 2024 kern.info kernel: [244766.243638] err_cnt = 7
Tue May 21 10:08:22 2024 kern.info kernel: [244766.243639] is_qfsm_hang = 1
Tue May 21 10:08:22 2024 kern.info kernel: [244766.249113] is_qfwd_hang = 1
Tue May 21 10:08:22 2024 kern.info kernel: [244766.252068] -- -- -- -- -- -- --
Tue May 21 10:08:22 2024 kern.info kernel: [244766.255369] MTK_QDMA_FSM = 0x100
Tue May 21 10:08:22 2024 kern.info kernel: [244766.258670] MTK_QDMA_FWD_CNT = 0x0
Tue May 21 10:08:22 2024 kern.info kernel: [244766.262145] MTK_QDMA_FQ_CNT = 0x44e0800
Tue May 21 10:08:22 2024 kern.info kernel: [244766.266053] ==============================
Tue May 21 10:08:22 2024 kern.err kernel: [244766.625020] mtk_soc_eth 15100000.ethernet eth1: transmit timed out
Tue May 21 10:08:22 2024 kern.info kernel: [244766.631299] [mtk_pending_work] No need to do FE reset !
Tue May 21 10:08:25 2024 user.notice root: mtk-sdk eth halted detected, now reset it...
Tue May 21 10:08:25 2024 kern.info kernel: [244769.614585] mtk_soc_eth 15100000.ethernet eth1: Link is Down
1 Like

Hiya, on OpenWrt 21.02-SNAPSHOT r15812+1073-46b6ee7ffc, I keep getting

Tue May 21 08:32:05 2024 kern.warn kernel: [55969.760744] 7986@C23L2,DfsSanityCheck() 7082: Dereferencing NULL pointer
Tue May 21 08:32:05 2024 kern.warn kernel: [55969.767432] 7986@C23L2,DfsSanityCheck() 7082: Dereferencing NULL pointer
Tue May 21 08:32:05 2024 kern.warn kernel: [55969.774120] 7986@C23L2,DfsSanityCheck() 7082: Dereferencing NULL pointer
Tue May 21 08:32:05 2024 kern.warn kernel: [55969.780807] 7986@C23L2,DfsSanityCheck() 7082: Dereferencing NULL pointer
Tue May 21 08:32:05 2024 kern.warn kernel: [55969.787495] 7986@C23L2,DfsSanityCheck() 7082: Dereferencing NULL pointer
Tue May 21 08:32:05 2024 kern.warn kernel: [55969.794183] 7986@C23L2,DfsSanityCheck() 7082: Dereferencing NULL pointer
Tue May 21 08:32:05 2024 kern.warn kernel: [55969.800871] 7986@C23L2,DfsSanityCheck() 7082: Dereferencing NULL pointer
Tue May 21 08:32:05 2024 kern.warn kernel: [55969.807560] 7986@C23L2,DfsSanityCheck() 7082: Dereferencing NULL pointer

also I know upnp isn't secure but that doesn't work either:

Mon May 20 18:00:34 2024 daemon.err miniupnpd[4418]: refresh_nft_cache: mnl_cb_run returned -1
Mon May 20 18:00:34 2024 daemon.err miniupnpd[4418]: refresh_nft_cache: mnl_cb_run returned -1
Mon May 20 18:00:35 2024 daemon.err miniupnpd[4418]: refresh_nft_cache: mnl_cb_run returned -1
Mon May 20 18:00:59 2024 daemon.err miniupnpd[4418]: refresh_nft_cache: mnl_cb_run returned -1
Mon May 20 18:01:04 2024 daemon.err miniupnpd[4418]: refresh_nft_cache: mnl_cb_run returned -1
Mon May 20 18:01:04 2024 daemon.err miniupnpd[4418]: refresh_nft_cache: mnl_cb_run returned -1
Mon May 20 18:01:04 2024 daemon.err miniupnpd[4418]: refresh_nft_cache: mnl_cb_run returned -1
Mon May 20 18:01:05 2024 daemon.err miniupnpd[4418]: refresh_nft_cache: mnl_cb_run returned -1
Mon May 20 18:01:29 2024 daemon.err miniupnpd[4418]: refresh_nft_cache: mnl_cb_run returned -1
Mon May 20 18:01:34 2024 daemon.err miniupnpd[4418]: refresh_nft_cache: mnl_cb_run returned -1
Mon May 20 18:01:34 2024 daemon.err miniupnpd[4418]: refresh_nft_cache: mnl_cb_run returned -1
Mon May 20 18:01:34 2024 daemon.err miniupnpd[4418]: refresh_nft_cache: mnl_cb_run returned -1

Thanks

Might have solved this by disabling Packet Steering in Luci, does anyone know if this is the same as the Network Acceleration setting in the GL.iNet interface?

Does Packet Steering really make a performance difference? I never noticed one.

Network acceleration is normally only the NAT acceleration.

Not sure, it seems to be on by default after a clean install. NAT acceleration reduces CPU at 500Mb from 7% to 1%

Yes, I have had enough of this daily testing and building my hopes up of finally getting some sort of internet stability but my VPN client connection kept on slowing down and disconnecting and I have decided to dust off my trusty Asus again and have some peace and quiet for some time before family also disowns me for ruining their internet access on daily basis..

5 Likes

Try the OpenWrt snapshots, Wifi has been very stable, no PPPoE crashes and whole network seems blisteringly fast since upgrading shortly after my message.

If you do, use the sys upgrade image and enable NAT acceleration after (according to wrt forum there are crashes if this is off).

1 Like

I pulled the trigger yesterday on an Asus GT-AX6000, I am so done with being a product tester for Gl. iNet.
It's such a shame, because on paper the specs of the Flint 2 are amazing, it's just let down by really bad software.

It feels like they need to completely scrap the firmware they are 'tinkering' with and start from scratch. It does not feel like they are serious at building a good product right now.

5 Likes

I use my flint 2 with latest openwrt snapshot, and my ax86u takes care of wifi. Works great.

@bonez I went back to using my Asus router running Merlin. I had been making my own OpenWRT snapshot builds for GL-MT6000 with the online firmware builder (including 6RD for IPv6, LUCI, material theme, stubby and some other functions I use) but due to the kernel version being so new on the OpenWRT snapshots the online firmware builder errors out most of the time as the packages are not built for such a new kernel version. The version 4.5.6 from GL-iNet is dated now, but at least it is fairly stable and runs on a modern OpenWRT version 23.05. Whoever is deciding things at GL-iNet and deciding to use the closed source drivers from MediaTek and the ancient and unsupported 21.02 version of Open WRT is very misguided. They have spent many man-hours of time trying to build something in version 4.6.0 than won't meet the needs of most of their users, and that I certainly would not want. The attraction to this router was that the firmware was based on the latest OpenWRT base 23.05 which meant SW support, stability and ability to customize. I am not impressed with 4.6.0. I am surprised that GL-iNet is actually repeating their mistakes made with Flint 1. Clearly they did not achieve any organizational learning from Flint 1.

6 Likes

@admon I have been running OpenWRT snapshot builds for a number of weeks and the WIFI performance is actually pretty good. Definitely better than the GL-iNet builds with the closed source drivers.

1 Like

That router gave me no better performance than the Flint 2, especially on the 2.4ghz channel, for Ā£300 GBP its an ASUS added tax rip off, sent it back to amazon after trying it for 2 days.

Hmm i seem to have no issues compiling OpenWrt, but in some cases i had to use rm -r tmp/ and make dirclean then update and install all feeds again with ./scripts/feeds update -a ./scripts/feeds install -a and update the make config by using make menuconfig.

I had only one issue when they enforced a newer version for gcc, and g++ basicly the linux equivalent for .net sdk (this is how i see it).

The Ubuntu wsl images inside the Microsoft stores are kinda dated, i had to add another apt repo and then manual fix the symlinks so the OpenWrt build tools work again, but due this change all the pre compiled tools need to recompiled cleanly.

It took me more than a day to compile :yum:, but yea... I'd agree with these type of updates compiling also becomes no fun, i already spotted some references to gcc14 in the commits that is again another apt repo i need to find and hoping it does not break.

On the other hand somehow these compiled versions work faster with the Arm architecture for me.

As for wifi... there is still some nasty wifi issues but this is very device specific or traffic specific, also the asus one seem to suffer from this type of issue.

Best is to avoid multicast and if the end user device has a option for transmission power set it on lowest modus, (not on the router.)

^ tested this with AX210 and gta online traffic on multi psk environment which plagued me with unpredictable crashing.

I moved to the Qualcomm chipset WIFI network adapter using Qualcomm QCNCM865. It is a way better chipset than Intel AX210. I am getting a full 2.4 Gbps connection on a 2 stream connection at a long distance which is twice the speed I could get from AX210.

2 Likes

You do realise that Asus stopped support for the GT-AX6000 early last year right? It's basically EOL.

Let's hope you don't have to send it in for warranty! Could be a couple thousand dollar repair bill if you scratch the antennas.

Do you get all of your facts from youtube? Is this how you run your life? Where did you get that from?

https://www.asus.com/event/network/eol-product/

Also Eric (AsusMerlin) is still completely committed to the GT-AX6000 which remains one of the best routers in its class.

2 Likes