Qualcomm Fast Path For LEDE


#691

Can you help me with a compiled fast path lede 17.01.5 for a D-Link DHP-1565 rev. A1?
Thank you.


#692

I do not see any benefits of this patch, bandwidth is same, negligible differences.
Used iperf -s and iperf -c serveripaddress in LAN-LAN and WAN-LAN


#693

Dlink Dhp-1565
openwrt 18.06.1 up 362.46 mbps / down 350.52 mbps
lede 17.01.5 up 294.90 mbps / down 246.02 mbps
DD-WRT v3.0-r37012 up 876.23 mbps / down 489.39 mbps


#694

this is because DD-WRT has very unsecure firewall, if you optimize OpenWRT firewall or turn it off just for testing you can see same or even better speeds


#695

Can you teach me how? I'm still learning how to work with Linux and these kinds ok things :laughing::laughing::laughing:


#696

I've already tried that. If I disable the firewall I no longer have Internet connection.
But I'm open to any test given that I use lede-openwrt for a few years and dd-wrt for 1 day.
But I think the lack of hardware NAT is to blame.


#697

obviously you did not disable firewall, try iptables to make sure it is disabled, should be like:
iptables -nvL
Chain INPUT (policy ACCEPT 0 packets, 0 bytes)
pkts bytes target prot opt in out source destination

Chain FORWARD (policy ACCEPT 0 packets, 0 bytes)
pkts bytes target prot opt in out source destination

Chain OUTPUT (policy ACCEPT 0 packets, 0 bytes)
pkts bytes target prot opt in out source destination

Should be ACCEPT on all tables nat, mangle etc and no rules
And also check ipv6 tables.


#698

If you disable firewall, NAT will be disabled as well


#699

The official LEDE build for Ubiquiti's ER-X has, in the firewall page, both "Software flow offloading" and "Hardware flo offloading" (the later specific to the mt7621 chipset).

Is this the same patch?


#700

I have followed thes steps to make a build for PowerPC mpc85xx (TL-WDR4900)

I run ./patch_LEDE.sh and get the folowing error:

mins@ubuntumate:~/devel/lede$ ./patch_LEDE.sh 
Add QCA Repo
--2018-10-03 19:22:59--  https://source.codeaurora.org/quic/qsdk/oss/system/openwrt/plain/include/local-development.mk
Resolving source.codeaurora.org (source.codeaurora.org)... 35.154.29.226
Connecting to source.codeaurora.org (source.codeaurora.org)|35.154.29.226|:443... connected.
HTTP request sent, awaiting response... 200 OK
Length: 440 [text/plain]
Saving to: ‘./include/local-development.mk’

local-development.mk            100%[======================================================>]     440  --.-KB/s    in 0s      

2018-10-03 19:23:00 (98,9 MB/s) - ‘./include/local-development.mk’ saved [440/440]

Updating feed 'packages' from 'https://git.openwrt.org/feed/packages.git' ...
Cloning into './feeds/packages'...
remote: Enumerating objects: 4419, done.
remote: Counting objects: 100% (4419/4419), done.
remote: Compressing objects: 100% (3722/3722), done.
remote: Total 4419 (delta 199), reused 3335 (delta 126)
Receiving objects: 100% (4419/4419), 2.62 MiB | 5.12 MiB/s, done.
Resolving deltas: 100% (199/199), done.
Create index file './feeds/packages.index' 
fatal: Invalid revision range ee53a240ac902dc83209008a2671e7fdcf55957a..HEAD
fatal: Invalid revision range ee53a240ac902dc83209008a2671e7fdcf55957a..2cc821e7edeafecc57ed9f3be46af4c322d58560
Checking 'working-make'... ok.
Checking 'case-sensitive-fs'... ok.
Checking 'proper-umask'... ok.
Checking 'gcc'... ok.
Checking 'working-gcc'... ok.
Checking 'g++'... ok.
Checking 'working-g++'... ok.
Checking 'ncurses'... ok.
Checking 'perl-thread-queue'... ok.
Checking 'tar'... ok.
Checking 'find'... ok.
Checking 'bash'... ok.
Checking 'patch'... ok.
Checking 'diff'... ok.
Checking 'cp'... ok.
Checking 'seq'... ok.
Checking 'awk'... ok.
Checking 'grep'... ok.
Checking 'getopt'... ok.
Checking 'stat'... ok.
Checking 'unzip'... ok.
Checking 'bzip2'... ok.
Checking 'wget'... ok.
Checking 'perl'... ok.
Checking 'python'... ok.
Checking 'git'... ok.
Checking 'file'... ok.
Checking 'ldconfig-stub'... ok.
fatal: Invalid revision range ee53a240ac902dc83209008a2671e7fdcf55957a..HEAD
fatal: Invalid revision range ee53a240ac902dc83209008a2671e7fdcf55957a..2cc821e7edeafecc57ed9f3be46af4c322d58560
Collecting package info: done
fatal: Invalid revision range ee53a240ac902dc83209008a2671e7fdcf55957a..HEAD
fatal: Invalid revision range ee53a240ac902dc83209008a2671e7fdcf55957a..2cc821e7edeafecc57ed9f3be46af4c322d58560

#701

I tried your SFE patch, it seems to have trouble offloading UDP connections


#702

I've been finding that my router's a bit unstable -- wifi stops working properly (after a day or two, the wifi networks become inaccessible to new connections, and devices already connected to them appear to be connected but get no internet access). It's very odd. I have to clear it by power-cycling the router.

I tried reversing the overclock to see if that fixed the issue (I dropped the overclock from 1GHz down to 800MHz) but I'm still running into the same wifi radio or network instability.

I'm not sure how to track down the problem. Next time it happens I'll see if it's only affecting wireless connections or is also affecting wired clients. And if wire clients can still connect to the router I'll poke around and see if there's anything apparently wrong.

For what it's worth, I'm running the same fast path build, and am also using 802.11r to do the fast handoff with another openwrt router configured as an access point with no DHCP.


#703

Hi folks
Is this feature already integrated to snapshot OpenWrt builds or even since any specific version ?

Mentioning specifically some well known hardware does it come on vanilla builds for TP-LInk WDR3600, WDR4300 and Archer C7 (all versions) ?


#704

i doubt it will ever be integrated as there is really no benefit, 4.14 has some offloading


#705

Thanks for the information.