Yes.
The tftp recovery is in uboot bootloader and does not get modified by OpenWrt images. It will work.
best to ask @ansuel. I believe he ran into this and recovered. From personal experience on my r7500v2, its pretty tough to completely brick it... see @hnyman's response.
thanks for the quick answer. Will give it some thought.
Yep I restored mine with normal recovery process so...
bit of a necro post here...can you provide a link or something to this comment in the qsdk? i'm afraid i'm a bit new to all this...where can one get QSDK?
if anyone would like to test / have their device supported by my "usb-builder"... ( r7800 / nbg6817 / other )
feel free to drop in: Synology rt2600ac support thread
if you don't have ssh to OEM, then serial will be needed.
what will this provide to R7800? ability to boot it off USB image rather than the NAND? without serial?
if someone with serial contributes and tests it, yes, that could be possible....
the scripts also provide other beta features;
-automatic seed diff ( whats been added )
-auto patching
-multi seedfile selection
-targetdir / k-config / ./patches / auto-diffconfig backup for development
etc. etc. best to discuss over there, don't wanna spam this thread.
1 Like
on a related note, i have been trying to fix the L2 scaling: R7800 cache scaling issue. would be much appreciated if others could test it too.
i have not accounted for this hardware constraint in my patch, but it seems to be working even with the minimum frequency at 384. i can try pinning one core to 384 and one to 1725 later and see what happens 
. i must admit i haven't had to deal with traffic bursting to 4Gbps, so haven't seen the router locking up.
i have a quick mess around yesterday but have not tested exhaustively - was installing a new GPU which is more important than some router nonsense 
i pinned one core to 384 and one to 1725 and nothing exploded, so i am not sure what exactly 'does not work'. however i believe QSDK is authoritative (it comes from QCA direct, right?) so i will try to work in some logic to ensure that if either CPU is non-idle (above 600MHz) then both are set to that at a minimum.
well most of the times using qsdk source doesn't mean have a 100% stable system... as they mostly make things work but not make things works the right way...
Just watch how much work they put to push something from qsdk to mainline kernel... (and also how much work they dropped)
well that's the question i suppose 
! when they say
CPU1 min frequency also has to be increased because there is a hardware constraint
kraits cannot operate at 384MHz when L2 is at 1GHz.
do they mean
- there is a caveat (documented eg in the datasheet) that if L2 is at 1G then the cores cannot be running in the idle bin (384Mhz); or
- when blackbox experimenting with the cpu/L2 frequencies this is one way we found of making the L2 operate at the correct speed?
It's probably worth noting that the changes in question were originally from 2015, so we don't even know what hardware they were running against. I don't know, but suspect, that the comments about the CPU getting stuck in softirq and unable to run cpufreq are no longer true in the current kernel's because of https://lwn.net/Articles/732740/ (if i'm reading it correctly).
1 Like
Sorry, I am willing but I can't, I go-to another city for working , and I don't have routers now
1 Like
Pull from yesterday gives
Collected errors:
* satisfy_dependencies_for: Cannot satisfy the following dependencies for luci-app-samba4:
* libpam
* opkg_install_cmd: Cannot install package luci-app-samba4.
package/Makefile:65: recipe for target 'package/install' failed
type or paste code here
during compile. After setting
CONFIG_PACKAGE_libpam=y
everything was fine.
Nothing to do with R7800 specifically. This is not a generic thread about everything.
Ps. Most likely just due to either some packages getting updated (like iproute2) causing some dependent packages to wait for suitable re-compilation, or buildbot being just updating a new batch of packages. Right now libpam is downloadable. (also remember to update package lists)
Pps.Are you talking about real compilation or about composing new image from already compiled package via imagebuilder...
Sorry for the inconvenience. I'm far away from knowing about everything regarding the compile process of openwrt. 
ATM just learning...
I'm not using the imagebuilder, everything was done with make commands
My R7800 is running 17.01.4, r3560-79f57e422d version. this version is customized version to enable 80M space on 17.01.4 build.
Now the latest is 18.06.4, do I need update to the latest build? what benefit I can get if I upgrade to?
I read the instalation guidelines, seems I can only use TFTP method? after TFTP method, I need reinstall all the software I have now, does any backup makes it work so that I not need reinstall one by one.
@avbohemen from a wireless laptop to a wired pc on local lan would not reach speeds above 100Mbps, where previous builds (kernel 4.14) could easily do 300Mpbs and up. Does this have anything to do with the "SWBA overrun on vdev 0, skipped old beacon" messages/bug?
@hnyman At least partially. When monitoring CPU utilisation for a longer period, the mainline ath10k with "SWBA overrun" errors seemed to cause somewhat higher idle CPU load than the -ct version. Likely the "SWBA overrun" version also causes higher CPU load (and lower throughput) during higher traffic loads, too.
I tried a new build today, r11219. Looks much better than my first attempt at the 4.19 kernel. Software offloading is fixed, and moreover, the wireless speed issues I had are also gone. Even better: there are no "SWBA overrun" messages anymore! Maybe the R7800 is not so lucky, but my Archer C7 v2 is. I suspect the speed fixes come from https://git.openwrt.org/?p=openwrt/openwrt.git;a=commit;h=97c37f8dd067fd4750a64bcaa639e6d08462060d and maybe https://git.openwrt.org/?p=openwrt/openwrt.git;a=commit;h=9c033242b1a27628195278a3ce37a4a994ce9a60.
At the moment, things look great again for OpenWRT.
am running into a weird issue with the 2.4ghz interface on one of my r7800s.
it seems fully functional from the routers perspective, but no SSID can be seen by any client.
I tried reflashing it with the latest snapshot, wiping the wireless config to default, powercycling, etc, but no dice. it does however pickup other SSIDS fine if scan for them via this interface, so it seems functional somehow.
Could it be the amplifier is busted? anybody as any other ideas to troubleshoot this
That doesn't really belong into this topic, as it's 99.8% hardware agnostic, but would be better suited in a dedicated one.
Is there a particular reason why you don't want to install and run tcpdump on your r7800, to catch and store the data from eth2 on the r7800 itself (you can then import that dump into wireshark for easier access)?
1 Like