Posted: Wed Nov 01, 2023 5:03 Post subject: [SOLVED] EA8500 & other models not booting after 53562
UPDATE 2024-04-11: The next public release finally fixes all these issues. If you are new to DD-WRT and are planning to flash a build onto your EA8500 router, do NOT flash any builds between r53616 through r55723. They will not work! All these builds are broken in one way or another on this particular router. Always start with a build later than r55723 or earlier than r53616 and you should be fine. A test build released today (r55776) seems to fix all in the aforementioned builds. The new build should be publicly available soon.
A BIG Thank You to everybody who helped troubleshoot these issues over the past several months and especially a HUGE shoutout to the Head Developer of DD-WRT, @Brainslayer!
Router: EA8500
Issue: Not able to boot any release after 53562 with WebUI or CLI flashing.
I am wondering, especially after looking closely at the boot logs of these failing builds, if there is a way to completely erase the flash rom memory, and perhaps even NVRAM even though it is suppose to erase when TFTPing a build. I am not so sure it's actually doing this, at least in the correct places.
Getting into the boot loader offers many options, including erasing the flash memory. I do not dare try any of them as I have no idea what they actually do and how it would affect uboot in itself. Not much info I can find on it. Disregarding the latest release (53833) which is un-bootable on the EA8500, everything since 53562 goes through some very odd actions.
It seems that something is either not getting written to, or not being erased. I don't have enough data to discern which. I also am wondering if the mtd addresses have moved in different router versions (some with built in serial pins, many without. That is a HW version change for sure). but what else changed? Maybe nothing. Maybe something that k6.1 turns on/off.
There is absolutely no reason that swconfig should cause a kernel panic on some EA8500s and not others. Why is there no boot device available when initially rebooting after TFTP flashing? It does somewhat fix itself after the second try.
Could this be evidence of bad blocks? If so, why do builds 53562 and below flash and boot fine?
We have been having these issues on every build since 53562. I only opened this thread to be more focused on it, as the build threads are somewhat hard to follow. It's not just the EA8500 either affected by this.
Please, if able, provide serial boot logs. _________________ Linksys EA8500 (Internet Gateway, AP/VAP) - DD-WRT r53562
Features in use: WDS-AP, Multiple VLANs, Samba, WireGuard, Entware: mqtt, mlocate
Wireless 5ghz only
Netgear R7800 (WDS-AP, WAP, VAP) - DD-WRT r55779
Features in use: multiple VLANs over single trunk port
Linksys EA8500 WDS Station x2 - DD-WRT r55799
Netgear R6400v2 WAP, VAP 2.4ghz only w/VLANs over single trunk port. DD-WRT r55779
OSes: Fedora 38, 9 RPis (2,3,4,5), 20 ESP8266s: Straight from Amiga to Linux in '94, never having owned a Windows PC.
Forum member #248
Last edited by lexridge on Thu Apr 11, 2024 22:23; edited 2 times in total
After many hours of testing, flashing and re-flashing mixed partitions, I have finally got a handle on what may be going on. Partition 1 seems to be the "go to" partition in most cases. Since most of us are using Part 1 all the time, flashing an update almost always ends up on Part 2. Any builds after 53562 cannot boot for some reason from Part 2.
With this in mind, I flashed 53562 to both partitions and force booted it from Part 2. Then ran the webUI flash for all builds 53694 through 53843, which in this test ALL ended up intentionally on Part 1. All booted and worked as they should.
The issue seems to be with Part 2 on builds after 53562. This only happens whenever the update is flashed to Part 2. It ALWAYS works when these builds are flashed to Part 1. Now the question of the day is, WHY? What is different about Part 2?
If users having these issues with late builds could run this simple command
Code:
# ubootenv get boot_part
It will tell you which partition you are booted from. You can force boot to the opposite partition by issuing the command
Code:
ubootenv set boot_part (number 1 or 2)
_________________ Linksys EA8500 (Internet Gateway, AP/VAP) - DD-WRT r53562
Features in use: WDS-AP, Multiple VLANs, Samba, WireGuard, Entware: mqtt, mlocate
Wireless 5ghz only
Netgear R7800 (WDS-AP, WAP, VAP) - DD-WRT r55779
Features in use: multiple VLANs over single trunk port
Linksys EA8500 WDS Station x2 - DD-WRT r55799
Netgear R6400v2 WAP, VAP 2.4ghz only w/VLANs over single trunk port. DD-WRT r55779
OSes: Fedora 38, 9 RPis (2,3,4,5), 20 ESP8266s: Straight from Amiga to Linux in '94, never having owned a Windows PC.
I have two ea8500s and both are using partition 1. One is on r53562 and the other is on r53708. I am an armchair expert on this, but one thing I will point out is that the size of the firmware shot up after r53562.
I have two ea8500s and both are using partition 1. One is on r53562 and the other is on r53708. I am an armchair expert on this, but one thing I will point out is that the size of the firmware shot up after r53562.
I will try getting r53843 on partition 1.
Thanks.
It's interesting you were able to get 53708 onto partition 1. Do you remember whether either router had factory serial header pins on them when you originally flashed them?
EDIT: Also curious if you force boot the router with 53708 to part 2 what version is there? I would guess it's something after 53562 unless you were originally booting from part 2. _________________ Linksys EA8500 (Internet Gateway, AP/VAP) - DD-WRT r53562
Features in use: WDS-AP, Multiple VLANs, Samba, WireGuard, Entware: mqtt, mlocate
Wireless 5ghz only
Netgear R7800 (WDS-AP, WAP, VAP) - DD-WRT r55779
Features in use: multiple VLANs over single trunk port
Linksys EA8500 WDS Station x2 - DD-WRT r55799
Netgear R6400v2 WAP, VAP 2.4ghz only w/VLANs over single trunk port. DD-WRT r55779
OSes: Fedora 38, 9 RPis (2,3,4,5), 20 ESP8266s: Straight from Amiga to Linux in '94, never having owned a Windows PC.
Forum member #248
Last edited by lexridge on Wed Nov 01, 2023 22:22; edited 1 time in total
my r53562 has always been on part 2 so reckon that is why most k6.1.x have booted after flash
NO telling what BS or new kernel has changed-up
Size IS NOT THE problem.
part 1 is a bit larger but the new k6.1.x builds are NOT even close to part 2 limit
----
EDIT:
then again don't really see the Marvelly people having same problem.
AFAIK all their Linksys routers are a dual boot like the EA8500 ...but they are a different beast so just dunno
I know this is a really crazy question, but is it possible to mount partition 2 while booted to partition 1, in order to do a side by side comparison of the boot parts? Not sure if that would do any good however unless it's something really obvious. _________________ Linksys EA8500 (Internet Gateway, AP/VAP) - DD-WRT r53562
Features in use: WDS-AP, Multiple VLANs, Samba, WireGuard, Entware: mqtt, mlocate
Wireless 5ghz only
Netgear R7800 (WDS-AP, WAP, VAP) - DD-WRT r55779
Features in use: multiple VLANs over single trunk port
Linksys EA8500 WDS Station x2 - DD-WRT r55799
Netgear R6400v2 WAP, VAP 2.4ghz only w/VLANs over single trunk port. DD-WRT r55779
OSes: Fedora 38, 9 RPis (2,3,4,5), 20 ESP8266s: Straight from Amiga to Linux in '94, never having owned a Windows PC.
EDIT:
then again don't really see the Marvelly people having same problem.
AFAIK all their Linksys routers are a dual boot like the EA8500 ...but they are a different beast so just dunno
There are a few WRT1900ACv1s experiencing the issue. _________________ Linksys EA8500 (Internet Gateway, AP/VAP) - DD-WRT r53562
Features in use: WDS-AP, Multiple VLANs, Samba, WireGuard, Entware: mqtt, mlocate
Wireless 5ghz only
Netgear R7800 (WDS-AP, WAP, VAP) - DD-WRT r55779
Features in use: multiple VLANs over single trunk port
Linksys EA8500 WDS Station x2 - DD-WRT r55799
Netgear R6400v2 WAP, VAP 2.4ghz only w/VLANs over single trunk port. DD-WRT r55779
OSes: Fedora 38, 9 RPis (2,3,4,5), 20 ESP8266s: Straight from Amiga to Linux in '94, never having owned a Windows PC.
EDIT:
then again don't really see the Marvelly people having same problem.
AFAIK all their Linksys routers are a dual boot like the EA8500 ...but they are a different beast so just dunno
There are a few WRT1900ACv1s experiencing the issue.
ok --- I scoot thru there sometmes but can't keep tack of it all
Your catch is prolly right on the money ...nice work lexridge
Thanks. I love troubleshooting. Especially when as challenging as this issue. I had a lot of free time lately and I've spent about 30 (probably more) hours on this already. Now the hard part is up to the devs to figure out why this is happening.
I have figured out a funky workaround/fix, but hesitate to post it for fear that it may soft brick some routers and require a TFTP to fix 'er up again. lol Would not be good for those without the ability to TFTP flash their router. Also, randomly the router sometimes comes back up with a full reset. _________________ Linksys EA8500 (Internet Gateway, AP/VAP) - DD-WRT r53562
Features in use: WDS-AP, Multiple VLANs, Samba, WireGuard, Entware: mqtt, mlocate
Wireless 5ghz only
Netgear R7800 (WDS-AP, WAP, VAP) - DD-WRT r55779
Features in use: multiple VLANs over single trunk port
Linksys EA8500 WDS Station x2 - DD-WRT r55799
Netgear R6400v2 WAP, VAP 2.4ghz only w/VLANs over single trunk port. DD-WRT r55779
OSes: Fedora 38, 9 RPis (2,3,4,5), 20 ESP8266s: Straight from Amiga to Linux in '94, never having owned a Windows PC.
funky workaround/fix, but hesitate to post it for fear that it may soft brick some routers and require a TFTP to fix .
My EA8500 is stuck on r53562.
My WRT1900ACv1 is stuck on r53633.
But my WRT3200ACM is now on r53843.
The EA8500 is busy, so I tried this out on my WRT1900ACv1
It was booting on Part 2
I switched it to Part 1, rebooted and tried to upgrade to r53843
Nope - back to r53633.
I switched Part to 1, but this time did not reboot it and did an update to r53843.
And now it is in a boot loop.
I've tried several times to turn on/off 3 times to switch partitions - it's stuck in its loop.
So, not great.
Not sure what to do now - not keen to TFTP it - had no luck last time trying.
My WRT3200ACM is on duty now, as AP. _________________ ------------------------------------
Linksys WRT3200ACM DD-WRT r56029 (Gateway)
Linksys EA8500 OpenWrt-23.05.3 (Wired AP)
Netgear WNDR3700v4 DD-WRT r55819 (spare Gateway)
Photos: https://www.flickr.com/photos/nickant44/albums
funky workaround/fix, but hesitate to post it for fear that it may soft brick some routers and require a TFTP to fix .
My EA8500 is stuck on r53562.
My WRT1900ACv1 is stuck on r53633.
But my WRT3200ACM is now on r53843.
The EA8500 is busy, so I tried this out on my WRT1900ACv1
It was booting on Part 2
I switched it to Part 1, rebooted and tried to upgrade to r53843
Nope - back to r53633.
I switched Part to 1, but this time did not reboot it and did an update to r53843.
And now it is in a boot loop.
I've tried several times to turn on/off 3 times to switch partitions - it's stuck in its loop.
So, not great.
Not sure what to do now - not keen to TFTP it - had no luck last time trying.
My WRT3200ACM is on duty now, as AP.
Sorry to hear that. I have ended up in the same scenario many times. You may eventually get it to reboot to the good partition by power cycling at just the right moments, as you know but it's hard to get it right. I usually count 3 seconds between power cycles to get it to finally change back over which at one point took around 9 attempts. But since I already have the serial port attached, it was just way faster to tftp 53562 back onto both partitions so I never messed with power cycling again. Too random and I am too impatient for that.
This is the situation too that once you do get it booted back to the good partition, I think you will find your router has been reset to defaults. I never tried switching the boot partition then flashing without a reboot in between. Not sure where the firmware would end up in this scenario. _________________ Linksys EA8500 (Internet Gateway, AP/VAP) - DD-WRT r53562
Features in use: WDS-AP, Multiple VLANs, Samba, WireGuard, Entware: mqtt, mlocate
Wireless 5ghz only
Netgear R7800 (WDS-AP, WAP, VAP) - DD-WRT r55779
Features in use: multiple VLANs over single trunk port
Linksys EA8500 WDS Station x2 - DD-WRT r55799
Netgear R6400v2 WAP, VAP 2.4ghz only w/VLANs over single trunk port. DD-WRT r55779
OSes: Fedora 38, 9 RPis (2,3,4,5), 20 ESP8266s: Straight from Amiga to Linux in '94, never having owned a Windows PC.
Lex:
I'm wondering if it now has a good partition at all.
My WRT3200ACM was my most recent find.
I flashed factory and OpenWRT and DD-WRT on either partition and all still good.
My WRT1900ACv1 has had many years of use and many flashes of DD-WRT.
This is not its first bad flash, and I'm now thinking it only had 1 good partition left today.
The same may or may not be true of my EA8500 - but I'm not touching that! _________________ ------------------------------------
Linksys WRT3200ACM DD-WRT r56029 (Gateway)
Linksys EA8500 OpenWrt-23.05.3 (Wired AP)
Netgear WNDR3700v4 DD-WRT r55819 (spare Gateway)
Photos: https://www.flickr.com/photos/nickant44/albums
@nickant I am not at all familiar with the WRT1900ACv1, and maybe it's a completely different issue with those routers. Without a serial log, it's impossible to say. There may be a way to flash it back with a USB stick in the boot loader, but again, I do not know that router. I think it's possible on the EA8500, but yeah, you are right. Don't touch that since it's your production router. I am lucky enough to have a few spares to play with, otherwise I would not be trying this. _________________ Linksys EA8500 (Internet Gateway, AP/VAP) - DD-WRT r53562
Features in use: WDS-AP, Multiple VLANs, Samba, WireGuard, Entware: mqtt, mlocate
Wireless 5ghz only
Netgear R7800 (WDS-AP, WAP, VAP) - DD-WRT r55779
Features in use: multiple VLANs over single trunk port
Linksys EA8500 WDS Station x2 - DD-WRT r55799
Netgear R6400v2 WAP, VAP 2.4ghz only w/VLANs over single trunk port. DD-WRT r55779
OSes: Fedora 38, 9 RPis (2,3,4,5), 20 ESP8266s: Straight from Amiga to Linux in '94, never having owned a Windows PC.
reboot to the good partition by power cycling at just the right moments
yea that is best bet
it was fairly easy last time it got in boot loop I had serial attached
you see 'Hit any key to stop autoboot: 2 1 0'
the countdown 2 . 1 . 0 . I don't think is actually seconds ...anyways soon as I seen the 2 I hit power switch off --back on within 1 second --- same couple more times and it switched boot parts.
Best guess is power on for NO longer than 2.5 seconds before hitting back off