Build for Netgear R7800

Sure, -j 4 would speed the downloads up especially at the initialization of the buildroot environment. But as I keep most the source downloads locally (in dl/ ), only the new stuff is downloaded at each build. So the improvement would be rather small.

I have wanted to keep the basic preparation steps simple and easy to debug, so I have not bothered with that optimization.

2 Likes

Looks like the network config in master and 21.02 will deviate quite a lot from 19.07. The handling of bridges, devices & ports is changing quite a lot.

I will likely stop building 19.07 from R7800 quite soon, as downgrading from master or 21.02 to 19.07 with the new style of settings will not be possible.

4 Likes

Does this mean we need to redo the network setup or will it be fixed automatically on firmware update?

Just check the commits. Mainly base-files & netifd & ...

There is a migration. First it was in master, and now backported to 21.02.
https://git.openwrt.org/?p=openwrt/openwrt.git;a=shortlog;h=refs/heads/openwrt-21.02

Actually, also the old config still works for now, but the first edit in LuCI will migrate it at least partially, and all new configs automatically generated by master or 21.02 will be the "new" style.

It is unlikely that the changes would be applied to the soon-to-be-deprecated 19.07, so downgrading to 19.07 will be problematic.

Looks like wireless is changing in master too. hwmode becoming band.

First, let me thank you hnyman for all of the work you've done for the community :smiley:
Second, could please anyone tell me if anything regarding hostapd/hostapd_cli changed between stable openwrt-19.07 and stable openwrt-21.02?

Why I'm asking? Well, after upgrading to 21.02 some time ago I've gotten to setting up my smarthome automations, including device detection. For this purpose I've used hostapd_cli with -a to launch a script every time hostapd event happens.
The issue is that on 21.02 the script is never called.

For the test purposes I've flashed back vanilla 19.07 (clean install) and used this simplest example I could find: [SOLVED] Assoc/Disassoc event/trigger - #2 by undef - and it worked just as expected.
Next, I've flashed hnymans 21.02 (also clean install), setup the WiFi network the same exact way as on 19.07, added the script and...nothing.

Does anyone have any idea why is that/how I can troubleshoot this further? (I'll test hnymans 19.07 tomorrow just in case - I've assumed that the hostapd_cli -a not working wouldn't go unnoticed for so long)

@Edit
Ok, I've installed hnymans 19.07 downloaded just few minutes ago and the hostapd_cli -a is also not working here. The only difference between vanilla 19.07 and this one is hostapd_common being in version ..2-5 and ..2-7

If it's not that (and I think there is no way to test since opkg does not allow downgrades) then I have no other ideas :\

Any way to force the config migration to the new format from the command line?

What is the preferred branch to use? Master? @hnyman have you messed around with the NSS HW offloading yet?

Master is probably the best at the moment.

Not all new network syntax things have yet been backported to 21.02, so master is probably slightly more feature-complete.

I have not messed with nss.

2 Likes

Am I the only one who is facing issues after using recent 21 master build?

I upgraded, like I do for many years once in a while, to the master 21 build of June 17th. Since this day I face many issues:
Internet speed dropped to below 50Mbps (was able to get 500Mbps and more);
Internet is very unreliable, DNS lookups seems to be gone every few minutes. However it's not just DNS, it's whole traffic which stops every few minutes for multiple minutes.
In short: unusable.

So I thought: probably it's the old settings, let's give it a factory reset. Still the same issues.

I flashed back to v19 build, expecting that the issues would be gone: still happening.

I checked with htop in the router, load is low.
Happens already on LAN, however WIFI same issues.
Turned off all my devices to rule out device creating conflicts or high load/requests. Even with single PC still the same issues.
Checked with htop in router, load is low.
After reset enabled software and hardware acceleration, however speed and stability still equally bad.

Multiple builds and factory resets later, it seems my device is still having big issues. It started after june 17th upgrade to v21.

Anything I am missing? Anyone else with these issues?

Use TFTP to flash a full factory build.

1 Like

Thanks. Also tried a TFTP flash now, however unfortunately it did not solve the problem.

TFTP "factory" image has no extra content compared to the sysupgrade image (if you sysupgrade without settings). The factory image just has an compatibility header for OEM flashing routines.

There is no hardware accleration in the normal R7800 images, yet. (there are some experimental NSS builds, but not in my builds)

As reverting to older builds did not help, it sounds like some hardware problem. Bad cable somewhere?

There is no "21 master".

  • There is the development master
  • There is the 21.02 branch, of which the 21.02 release builds are made.

My suggestions would be:

  • look at the system logs and kernel logs. anything visible there?

  • use the master builds, not 21.02. The OpenWrt network config is undergoing some changes, and 21.02 may have some quircks. Master will get fixed more quickly

  • Sysupgrade without keeping settings (or do a reset from LuCI after sysupgrading)

  • Try falling back to the 19.07 builds, but again, without settings.

1 Like

Thanks hnyman.

Thanks for explaining the build. I am indeed using latest development master. Currently R7800-master-r16972-3dabb62581-20210620-1954-sysupgrade.bin

In system logs I see:

Mon Jun 21 11:43:39 2021 kern.warn kernel: [  729.086471] print_req_error: 2 callbacks suppressed
Mon Jun 21 11:43:39 2021 kern.err kernel: [  729.086480] blk_update_request: I/O error, dev mtdblock0, sector 0 op 0x0:(READ) flags 0x80700 phys_seg 2 prio class 0
Mon Jun 21 11:43:39 2021 kern.err kernel: [  729.116521] blk_update_request: I/O error, dev mtdblock0, sector 8 op 0x0:(READ) flags 0x80700 phys_seg 2 prio class 0
Mon Jun 21 11:43:39 2021 kern.err kernel: [  729.117661] blk_update_request: I/O error, dev mtdblock0, sector 16 op 0x0:(READ) flags 0x80700 phys_seg 1 prio class 0
Mon Jun 21 11:43:39 2021 kern.err kernel: [  729.127667] blk_update_request: I/O error, dev mtdblock0, sector 24 op 0x0:(READ) flags 0x80700 phys_seg 1 prio class 0
Mon Jun 21 11:43:39 2021 kern.err kernel: [  729.176226] blk_update_request: I/O error, dev mtdblock0, sector 0 op 0x0:(READ) flags 0x0 phys_seg 1 prio class 0
Mon Jun 21 11:43:39 2021 kern.err kernel: [  729.176269] Buffer I/O error on dev mtdblock0, logical block 0, async page read
Mon Jun 21 11:43:39 2021 kern.err kernel: [  729.190374] blk_update_request: I/O error, dev mtdblock0, sector 0 op 0x0:(READ) flags 0x0 phys_seg 1 prio class 0
Mon Jun 21 11:43:39 2021 kern.err kernel: [  729.192673] Buffer I/O error on dev mtdblock0, logical block 0, async page read
Mon Jun 21 11:43:39 2021 kern.err kernel: [  729.204770] blk_update_request: I/O error, dev mtdblock1, sector 0 op 0x0:(READ) flags 0x80700 phys_seg 4 prio class 0
Mon Jun 21 11:43:39 2021 kern.err kernel: [  729.211649] blk_update_request: I/O error, dev mtdblock1, sector 8 op 0x0:(READ) flags 0x80700 phys_seg 3 prio class 0
Mon Jun 21 11:43:39 2021 kern.err kernel: [  729.222395] blk_update_request: I/O error, dev mtdblock1, sector 16 op 0x0:(READ) flags 0x80700 phys_seg 2 prio class 0
Mon Jun 21 11:43:39 2021 kern.err kernel: [  729.233076] blk_update_request: I/O error, dev mtdblock1, sector 24 op 0x0:(READ) flags 0x80700 phys_seg 1 prio class 0
Mon Jun 21 11:43:39 2021 kern.err kernel: [  729.243772] Buffer I/O error on dev mtdblock1, logical block 0, async page read
Mon Jun 21 11:43:39 2021 kern.err kernel: [  729.262870] Buffer I/O error on dev mtdblock1, logical block 0, async page read
> Blockquote
[  729.086471] print_req_error: 2 callbacks suppressed
[  729.086480] blk_update_request: I/O error, dev mtdblock0, sector 0 op 0x0:(READ) flags 0x80700 phys_seg 2 prio class 0
[  729.116521] blk_update_request: I/O error, dev mtdblock0, sector 8 op 0x0:(READ) flags 0x80700 phys_seg 2 prio class 0
[  729.117661] blk_update_request: I/O error, dev mtdblock0, sector 16 op 0x0:(READ) flags 0x80700 phys_seg 1 prio class 0
[  729.127667] blk_update_request: I/O error, dev mtdblock0, sector 24 op 0x0:(READ) flags 0x80700 phys_seg 1 prio class 0
[  729.176226] blk_update_request: I/O error, dev mtdblock0, sector 0 op 0x0:(READ) flags 0x0 phys_seg 1 prio class 0
[  729.176269] Buffer I/O error on dev mtdblock0, logical block 0, async page read
[  729.190374] blk_update_request: I/O error, dev mtdblock0, sector 0 op 0x0:(READ) flags 0x0 phys_seg 1 prio class 0
[  729.192673] Buffer I/O error on dev mtdblock0, logical block 0, async page read
[  729.204770] blk_update_request: I/O error, dev mtdblock1, sector 0 op 0x0:(READ) flags 0x80700 phys_seg 4 prio class 0
[  729.211649] blk_update_request: I/O error, dev mtdblock1, sector 8 op 0x0:(READ) flags 0x80700 phys_seg 3 prio class 0
[  729.222395] blk_update_request: I/O error, dev mtdblock1, sector 16 op 0x0:(READ) flags 0x80700 phys_seg 2 prio class 0
[  729.233076] blk_update_request: I/O error, dev mtdblock1, sector 24 op 0x0:(READ) flags 0x80700 phys_seg 1 prio class 0
[  729.243772] Buffer I/O error on dev mtdblock1, logical block 0, async page read
[  729.262870] Buffer I/O error on dev mtdblock1, logical block 0, async page read

Sysupgrade without settings: tried multiple times, different build. No effect, also not with v19 build.
Luci reset: tried multiple times. No effect.

Will unplug all cables except one and continue debugging. The issues started after upgrade to latest master on 17th with 17th build. Before the upgrade it was stable. No other changes (software nor hardware) as far as I am aware.

Those are normal messages. (first few NAND sectors without ECC)

I meant that what do you see while you experience difficulties later. Like your claim here:

DNS lookups seems to be gone every few minutes. However it's not just DNS, it's whole traffic which stops every few minutes for multiple minutes.

2 Likes

Thanks a lot. I might have found the issue!

The only clear possible "error" I found in the syslog was about Max concurrent DNS (150) connections reached a couple of times. Not sure yet if this was related, but could be.

However, I likely found the culprit:
My provider... So on the day I flashed the June 17th master firmware, just a few minutes or hours later, the trouble started. I checked the provider status page and postal code check. No issues. So the flash was always the suspect.

However today I checked again and I saw they now raised an issue during night. I did reboot my modem (also tried during weekend without effect). Now it worked! The line-speed is much faster again and the disconnections seem to be gone. Need to test a little bit longer, but probably the internet-connection itself was the issue. It does not have issues often, but this weekend it seems it had. And the tricky thing is: I was lead into the wrong direction.

Besides that: it looks like the master build actually seems to run fine. Need to do more and longer testing though.

Thanks a lot you all for the help. And lessons learned:

  • Do not always trust the ISP initial (no) known issues report;
  • Even if you flashed a new firmware and ISP issues are rare, there can be the situation where your ISP starts having issues after you just flashed. Unlucky timing.
2 Likes

I've only been a follower of this thread but I've always liked keeping up with what's going on.

26 days since last update. Has this build been abandoned or are things just going so well there's no more discussion needed? :grinning:

I'm assuming hnyman has been busy.

If you're looking for newer builds, you can always get the snapshots from openwrt's site https://firmware-selector.openwrt.org/?version=21.02.0-rc3&target=ipq806x%2Fgeneric&id=netgear_r7800

Not sure how you calculate that 26 days.
Last update in master is from 20210704, 13 days ago...

I have only updated the master during the summer holidays (as I have only been a few hours per week at home).

I will be back at home after next week, and then there will be also new 21.02 builds.

1 Like

Sorry, I meant 26 days since last post in the thread, not the last build. Sorry for the confusion.