Release Asuswrt-Merlin 388.2 is now available for select models

[Jumpstarter]

Updated to 388.2 over the weekend, and it seems I got some type of driver error early this morning - it generated a dump to screen and didn't reboot the router.

May 3 01:41:56 irongate kernel: bcm63xx_nand ff801800.nand: program failed at 8340600
May 3 01:41:56 irongate kernel: ubi0 error: ubi_io_write: error -5 while writing 2048 bytes to PEB 1034:0, written 0 bytes
May 3 01:41:56 irongate kernel: CPU: 2 PID: 118 Comm: ubi_bgt0d Tainted: P O 4.19.183 #1
May 3 01:41:56 irongate kernel: Hardware name: GTAX6000_50991 (DT)
May 3 01:41:56 irongate kernel: Call trace:
May 3 01:41:56 irongate kernel: dump_backtrace+0x0/0x150
May 3 01:41:56 irongate kernel: show_stack+0x14/0x20
May 3 01:41:56 irongate kernel: dump_stack+0x94/0xc4
May 3 01:41:56 irongate kernel: ubi_io_write+0x574/0x690
May 3 01:41:56 irongate kernel: ubi_io_write_ec_hdr+0xc4/0x110
May 3 01:41:56 irongate kernel: sync_erase.isra.0+0x11c/0x1f0
May 3 01:41:56 irongate kernel: __erase_worker+0x34/0x460
May 3 01:41:56 irongate kernel: erase_worker+0x18/0x80
May 3 01:41:56 irongate kernel: do_work+0x98/0x120
May 3 01:41:56 irongate kernel: ubi_thread+0x108/0x190
May 3 01:41:56 irongate kernel: kthread+0x118/0x150
May 3 01:41:56 irongate kernel: ret_from_fork+0x10/0x24
May 3 01:41:56 irongate kernel: ubi0: dumping 2048 bytes of data from PEB 1034, offset 0

Can anyone tell me what is at issue here?

I thought I saw something similar in my logs, Now trendmirco wont remain turned on. It turns itself off on reboots.

Does any one here know how to clear Trendmicros history?

 

[Weblee2407]

If you remove spdMerlin and stop hammering constantly your ISP connection for no reason your wife may be happier. You may even move out of the couch. The most Internet traffic generated by your household is perhaps to speedtest servers. Let your family members use some of this Internet connection.

lol

 

[Jumpstarter]

If you remove spdMerlin and stop hammering constantly your ISP connection for no reason your wife may be happier. You may even move out of the couch. The most Internet traffic generated by your household is perhaps to speedtest servers. Let your family members use some of this Internet connection.

What if the Wife is the ISP, doesn't that mean the constant hammering will make them both happier?

 

[Jbennett360]

Always had it on 36. Settings are same as one previous firmware, and I never had this kind of problem with any of the firmwares before just on this one. And again this morning wifi just stopped working and get back again on my S23 Ultra with exclamation mark, and I turn it off and on again, and it was back, really strange problem.

That'll be the S23. It has WiFi issues. Not sure if they're fully fixed yet. Samsung are aware

 

[JGrana]

Updated to 388.2 over the weekend, and it seems I got some type of driver error early this morning - it generated a dump to screen and didn't reboot the router.

May 3 01:41:56 irongate kernel: bcm63xx_nand ff801800.nand: program failed at 8340600
May 3 01:41:56 irongate kernel: ubi0 error: ubi_io_write: error -5 while writing 2048 bytes to PEB 1034:0, written 0 bytes
May 3 01:41:56 irongate kernel: CPU: 2 PID: 118 Comm: ubi_bgt0d Tainted: P O 4.19.183 #1
May 3 01:41:56 irongate kernel: Hardware name: GTAX6000_50991 (DT)
May 3 01:41:56 irongate kernel: Call trace:
May 3 01:41:56 irongate kernel: dump_backtrace+0x0/0x150
May 3 01:41:56 irongate kernel: show_stack+0x14/0x20
May 3 01:41:56 irongate kernel: dump_stack+0x94/0xc4
May 3 01:41:56 irongate kernel: ubi_io_write+0x574/0x690
May 3 01:41:56 irongate kernel: ubi_io_write_ec_hdr+0xc4/0x110
May 3 01:41:56 irongate kernel: sync_erase.isra.0+0x11c/0x1f0
May 3 01:41:56 irongate kernel: __erase_worker+0x34/0x460
May 3 01:41:56 irongate kernel: erase_worker+0x18/0x80
May 3 01:41:56 irongate kernel: do_work+0x98/0x120
May 3 01:41:56 irongate kernel: ubi_thread+0x108/0x190
May 3 01:41:56 irongate kernel: kthread+0x118/0x150
May 3 01:41:56 irongate kernel: ret_from_fork+0x10/0x24
May 3 01:41:56 irongate kernel: ubi0: dumping 2048 bytes of data from PEB 1034, offset 0

Can anyone tell me what is at issue here?

It looks like you had a bad block in your on-board nand flash device. A write error by the looks of the early message.
With your AX6000, Asus has been moving more and more to utilizing UBI. As the Linux mtd website best explains

It is a volume management system for raw flash devices which manages multiple logical volumes on a single physical flash device and spreads the I/O load

From what I have read, it's replacing the older JFFS2 filesystem with the UBI version - ubifs.
The kernel panic is coming from ubi-bgt0d - the background UBI thread that is doing eraseblock management, wear leveling, etc for the ubifs.

Hopefully it's just a one time bad block on the flash device - it happens and it is why there are device drivers (mtd) to manage flash chips. If I were you I would note the block it failed on (1034) and keep an eye out for any more of these errors - watching the block number. If you see more and more (and likely different), time to back up and give Asus a call. You have a degrading flash chip and at some point UBI/UBIFS won't be able to fix or manage...

BTW, you might want to run "ubinfo -a" and capture the output to a file.
I have an AX88U Pro (also using ubifs) and here is the info about my flash chip:

Count of UBI devices: 1
UBI control device major/minor: 10:59
Present UBI devices: ubi0

ubi0
Volumes count: 9
Logical eraseblock size: 126976 bytes, 124.0 KiB
Total amount of logical eraseblocks: 2016 (255983616 bytes, 244.1 MiB)
Amount of available logical eraseblocks: 158 (20062208 bytes, 19.1 MiB)
Maximum count of volumes 128
Count of bad physical eraseblocks: 0
Count of reserved physical eraseblocks: 40
Current maximum erase counter value: 14
Minimum input/output unit size: 2048 bytes
Character device major/minor: 249:0
Present volumes: 1, 2, 3, 4, 5, 6, 10, 11, 13

I would keep an eye out for "Count of bad physical eraseblocks" along with a declining "Count of reserved physical eraseblocks". You likely have 1 or 2 bad physical eraseblocks based on the above panic.
Again, run this command once in a while and see if this number goes bigger over time and reserved physical eraseblocks declining. Again, not unusual for some bad blocks.

 

[Tech9]

What if the Wife is the ISP, doesn't that mean the constant hammering will make them both happier?

Not sure what exactly you're saying, but sure speedtest is not a good idea for this connection.

 

[Weblee2407]

Thanks for the information. I will review my logs and see where the other 3 blocks were marked bad and see how close/far the events are.

UPDATE: Logs start about March 5th - one bad block marked April 2nd and now - May 3rd.

I will call Asus tomorrow and see if I have warranty coverage. It's one year old today.

 

[aex.perez]

Not sure what exactly you're saying, but sure speedtest is not a good idea for this connection.

Actualy, SpeedTests helps show when a particular issue occurrs. The results would be cut by 1/2 on the download and by a 1/3 on the upload until a reboot. Since then I've captured other events where a function fails and restarts. DNS, OpenVPN, HTTP being others Always the the same scenario just different functions. But only when Skynet or the native firewall fails and restarts does it cause Flow Cache to dump and when it does is when the SpeedTest results are cut and packet lossess start to occur. It not that Speedtests are off it's that it's off because of a failed and restarted service and what the dependencies of that function are determine what happens next. When DNSmasq failed and restarted last night I could only get to the router via its IP, and nothing was resolving even though DNSMasq restarted (scMerlin). These are all router functions, nothing in the logs of the nodes correlate to the failure/restart of the router.

These 388 releases have been very good with keeping the memory (RAM) in the 50%-75% range, prior families it would be normal to see the RAM meter showing used in the 85%-95% range. I'm starting to think, but don't have way to verify, that memory management is at the root of what is behind the oddball errors across different functions.

This setup, I've had for quite a while and I'm not doing anything differently, or running anything new since early in th 386 release cycle. If all these components have been working well, and CPU utlization at the router is not any different. My bandwidth is 1Gb Fiber (1Gb in both directions) and Speedtests are in the 920-940 range typically no reason to change anything on the router. Never had complanied about Speestest results, but when the Firewall would fail and restart, flowcache dumps and that how it was easy to tell something was off. Becase I run spdMerlin at :12 / :42 (as I always have / every hour / best available server) it made it easy to pinpoint when Skynet/native Firewall would restart and cause the issue with the Speedtest reults.

For DNS and the Skynet / Native Firewall - FlowCache dumps, the only solution to get functionality back is to resart the router. OpenVPN (router as a client) and HTTP have been able to restart and recover.

As I suspect that's closed source, I need to revert back to stock (same GPL as this 388 release is based on) to see if the same happens there. If it does, share it with ASUS directly and let them sort it out.

 

[Tech9]

Actualy, SpeedTests helps show when a particular issue occurrs.

Actually, the post you quoted is perhaps for different type connections. Speedtest-ing there means bad things.

For DNS and the Skynet / Native Firewall

Skynet is not a firewall. Skynet is an IP-based blocker. And I'm not talking to you unless you clean up your signature. A lot of issues there.

 

[modestyB]

That'll be the S23. It has WiFi issues. Not sure if they're fully fixed yet. Samsung are aware

Most likely the phone.
I have same problem with S23+, even after the may-update installed yesterday.
Did not have any problem with S21.

 

[DiGz_Au]

Most likely the phone.
I have same problem with S23+, even after the may-update installed yesterday.
Did not have any problem with S21.

You know I use an s23u as my daily driver and on merlin releases I see the wifi issues but on stock it's 10x more stable with no drop-outs. Both running standard wifi settings aswell. How strange. Repetable to the point I just stay on stock for my ax86u due to this issue.

 

[Calkulin]

My S23U stopped having Wi-Fi issues once I turned off "Switch to better Wi-Fi networks" a few months back. I would check if you have that on and turn it off to see if it helps your wireless issues

 

[Dedel66]

Always had it on 36. Settings are same as one previous firmware, and I never had this kind of problem with any of the firmwares before just on this one. And again this morning wifi just stopped working and get back again on my S23 Ultra with exclamation mark, and I turn it off and on again, and it was back, really strange problem.

I also saw that with my S23 ultra. I'm now back to 388.1 and have given up 160GHz. I'll switch back to 388.1 with 160GHz tomorrow and see what happens.

 

[Dedel66]

My S23U stopped having Wi-Fi issues once I turned off "Switch to better Wi-Fi networks" a few months back. I would check if you have that on and turn it off to see if it helps your wireless issues

Thanks for that hint, I will test that at weekend. Now on 388.1 with 80MHz it´s stable since this morning...

 

[gk802]

Not sure of others' experience, but I definitely saw a step-change in the Traffic Analyzer's data records after upgrading from 388.1 to 388.2 (RT-AX86U). It was my earlier understanding that the daily and monthly data included only internet packets. I didn't see a change in reception volume, but on the transmission side, it now seems to be including a substantial amount of volume that is internal to the network.

 

[modestyB]

Thanks for that hint, I will test that at weekend. Now on 388.1 with 80MHz it´s stable since this morning...

Have 20/40/80/160 set. S23+ still drops wifi sometimes. Might tweak that intelligent wifi settings on the phone.

 

[Henkan]

Have 20/40/80/160 set. S23+ still drops wifi sometimes. Might tweak that intelligent wifi settings on the phone.

Turn off Randomised MAC address in all devices. It's a setting per SSID so it doesn't affect general security. Who needs random MACs at home.
I used to have strange wifi drops and rebooting nodes but now everything works great and I have working AX

 

[Dedel66]

Have 20/40/80/160 set. S23+ still drops wifi sometimes. Might tweak that intelligent wifi settings on the phone.

Switching of "Intelligent WiFi" Tweaks does not work for me. Just "driving" at 80MHz is fixing this problem for me.

I have additionally observed that the s23u is not able to get my ax88u to switch to 160MHz. Only my PCs/laptops with intel wireless cards do this. If these have triggered the 160MHz, I have to restart the s23u's wifi so that it connects with 2.4 Gbit. But then with interruptions up to the exclamation mark...

I´ve always turned off Randomised MAC address in all devices.

AND: It is the same on 388.1 as on 388.2 - so it´s a Samsung caused Problem, NOT a 388.2 problem!

 

[Jbennett360]

Switching of "Intelligent WiFi" Tweaks does not work for me. Just "driving" at 80MHz is fixing this problem for me.

I have additionally observed that the s23u is not able to get my ax88u to switch to 160MHz. Only my PCs/laptops with intel wireless cards do this. If these have triggered the 160MHz, I have to restart the s23u's wifi so that it connects with 2.4 Gbit. But then with interruptions up to the exclamation mark...

I´ve always turned off Randomised MAC address in all devices.

AND: It is the same on 388.1 as on 388.2 - so it´s a Samsung caused Problem, NOT a 388.2 problem!

Yep. All people need to do is report the issue to Samsung via the members app. The more that report it, the better.

 

[gerardr]

I need some help with a 5Ghz wireless issue I have on my RT-AX58U router. When I use the 388.2 firmware and set the region to Australia under Wireless > Professional, the 5Ghz channel has no throughput. This was not a problem on the 388.1 firmware, where the 5Ghz channel worked fine with Auto channel and DFS enabled.

These are the settings I used for testing, changing only the region. The 5Ghz channel works when the region is set to ‘Asia (Default)’, but not when it is set to ‘Australia’.

I have tried a hard WPS reset on both 388.1 and 388.2 firmware, and also restored factory default settings, but the issue remains. I think this is more than a coincidence that the 5Ghz channel stops working when I switch to the Australian region.

Hi! Check that the channel you'd picked earlier is available in Australia. Here's a helpful Wikipedia page: List of WLAN channels.