That's the site's copyright, not that page's... You are aware private citizens cannot copyright information put forth by a governmental agency, such as the FCC, right?
You be you, friend. I'm out of this.
The regulatory bodies set the maximum allowable transmit power, they do not mandate that the device has to use that, and only that, transmit power. The device driver should be allowed to use a lower power, but mwlwifi does not allow this. IC (CA), and FCC (US), pretty much dictate the same regulations (something to do with a shared ~8900 km border), and from a rango:
root@bsaedgy:~# iw reg get
global
country CA: DFS-FCC
(2402 - 2472 @ 40), (N/A, 30), (N/A)
(5150 - 5250 @ 80), (N/A, 23), (N/A), NO-OUTDOOR, AUTO-BW
(5250 - 5350 @ 80), (N/A, 24), (0 ms), DFS, AUTO-BW
(5470 - 5600 @ 80), (N/A, 24), (0 ms), DFS
(5650 - 5730 @ 80), (N/A, 24), (0 ms), DFS
(5735 - 5835 @ 80), (N/A, 30), (N/A)
phy#1
country CA: DFS-FCC
(2402 - 2472 @ 40), (N/A, 30), (N/A)
(5150 - 5250 @ 80), (N/A, 23), (N/A), NO-OUTDOOR, AUTO-BW
(5250 - 5350 @ 80), (N/A, 24), (0 ms), DFS, AUTO-BW
(5470 - 5600 @ 80), (N/A, 24), (0 ms), DFS
(5650 - 5730 @ 80), (N/A, 24), (0 ms), DFS
(5735 - 5835 @ 80), (N/A, 30), (N/A)
phy#0
country CA: DFS-FCC
(2402 - 2472 @ 40), (N/A, 30), (N/A)
(5150 - 5250 @ 80), (N/A, 23), (N/A), NO-OUTDOOR, AUTO-BW
(5250 - 5350 @ 80), (N/A, 24), (0 ms), DFS, AUTO-BW
(5470 - 5600 @ 80), (N/A, 24), (0 ms), DFS
(5650 - 5730 @ 80), (N/A, 24), (0 ms), DFS
(5735 - 5835 @ 80), (N/A, 30), (N/A)
what I get from mwlwifi is dictated by the channel I choose. For example choosing 149, yields 30 dBm TX-power(and 30 dBm == xxxx mW , and only that is allowed. There is no reason that the device should be shouting at something 2 floors down, when that device cannot shout back.
This is of course true on all wrtpac devices with the baked in eeprom power tables, rango, and V2 flavours of the other units. This discussion keeps on happening.
2 Likes
Exceeding the FCC regulation was never my intention of bringing this up. Want to be able to reduce the power output because I don't need the full power. Just a like a CPU would scale down as well.
Scaling works fine on the R7000 and R7800. The WRT3200ACM for example, is fixed at a higher limit at the driver level, I would like to reduce it as needed.
Not sure what the actual WRT series is fixed at (doubt it's 1000mW that would be absurd), but I'm running the R7000 right now at 25mW and it works great, low power output for a smaller apartment, etc. WRT routers should be able to do this as well, that is all.
2 Likes
Yes, I understood from your original response to the Frequency Scaling link above that your desire was to lower the TX power from the mwlwifi default.
My post was in response to the misconception that the reason behind mwlwifi fixing the TX power at the maximum was due to a dictate by some regulatory body, and not due to a design decision behind mwlwifi.
fwiw, the driver can be modified to accept a DTS powertable rather than hardcoded (quite trivially, I’ve got a patch lying around somewhere) so maybe it can be modded to allow dropped power levels as well.
I never actually used a spectrum analyser to check if it was obeying what I set though.
So, using the latest build (with slabinfo enabled so I can actually use slabtop and /proc/slabinfo), it looks like skbuff_head_cache is leaking memory, although even that doesn't account for most of the memory leaking, which I still haven't been able to track down.
I have no idea what this means practically, but it's at least something I can start looking in to
Near crashing slabtop shows:
OBJS ACTIVE USE OBJ SIZE SLABS OBJ/SLAB CACHE SIZE NAME
425592 100046 0% 0.31K 35466 12 141864K skbuff_head_cache
At boot time it's more around
32844 32813 0% 0.31K 2737 12 10948K skbuff_head_cache
1 Like
Anyone with Oneplus 5, 5t or 6? Or some recent phone with Qualcomm chipset and Android 8.1.
Can you please test what is the upload speed after re-association? For me it is about 1Mbit/s.
So setup two APs with same SSID, roam back to Linksys so that it reassociates on 5GHz channel.
I have wrote more details here (as bbogey)
https://github.com/kaloz/mwlwifi/issues/295#issuecomment-391610963
Edit: issue will be fixed by OnePlus
I have a PixelXL w/ Android 8.1 and a Qualcomm MSM8996 chipset. I don't see any limiting of wifi. I move things to and from it at about 150-200 Mbps on 'AC', 75-100Mbps on 'N'.
My wife's Pixel2 has about the same speeds.
So it is the upload speed after reassociation?
Let me put my results here also
Initial association
Connecting to host 192.168.1.33, port 5201
[ 4] local 192.168.1.181 port 46364 connected to 192.168.1.33 port 5201
[ ID] Interval Transfer Bandwidth
[ 4] 0.00-1.01 sec 77.7 MBytes 642 Mbits/sec
[ 4] 1.01-2.00 sec 76.2 MBytes 646 Mbits/sec
[ 4] 2.00-3.00 sec 75.0 MBytes 629 Mbits/sec
[ 4] 3.00-4.02 sec 75.0 MBytes 622 Mbits/sec
[ 4] 4.02-5.01 sec 77.5 MBytes 655 Mbits/sec
[ 4] 5.01-6.00 sec 76.2 MBytes 644 Mbits/sec
[ 4] 6.00-7.01 sec 75.0 MBytes 623 Mbits/sec
[ 4] 7.01-8.01 sec 71.2 MBytes 596 Mbits/sec
[ 4] 8.01-9.00 sec 60.0 MBytes 509 Mbits/sec
[ 4] 9.00-10.02 sec 76.2 MBytes 627 Mbits/sec
- - - - - - - - - - - - - - - - - - - - - - - - -
[ ID] Interval Transfer Bandwidth
[ 4] 0.00-10.02 sec 740 MBytes 620 Mbits/sec sender
[ 4] 0.00-10.02 sec 740 MBytes 619 Mbits/sec receiver
iperf Done.After reassociation
Connecting to host 192.168.1.33, port 5201
[ 4] local 192.168.1.181 port 46300 connected to 192.168.1.33 port 5201
[ ID] Interval Transfer Bandwidth
[ 4] 0.00-1.01 sec 947 KBytes 7.67 Mbits/sec
[ 4] 1.01-2.04 sec 0.00 Bytes 0.00 bits/sec
[ 4] 2.04-3.04 sec 368 KBytes 3.01 Mbits/sec
[ 4] 3.04-4.04 sec 0.00 Bytes 0.00 bits/sec
[ 4] 4.04-5.02 sec 0.00 Bytes 0.00 bits/sec
[ 4] 5.02-6.04 sec 382 KBytes 3.08 Mbits/sec
[ 4] 6.04-7.04 sec 0.00 Bytes 0.00 bits/sec
[ 4] 7.04-8.02 sec 0.00 Bytes 0.00 bits/sec
[ 4] 8.02-9.02 sec 0.00 Bytes 0.00 bits/sec
[ 4] 9.02-10.04 sec 382 KBytes 3.07 Mbits/sec
- - - - - - - - - - - - - - - - - - - - - - - - -
[ ID] Interval Transfer Bandwidth
[ 4] 0.00-10.04 sec 2.03 MBytes 1.70 Mbits/sec sender
[ 4] 0.00-10.04 sec 1.10 MBytes 920 Kbits/sec receiver
iperf Done.
Just installed r7093 on my Rango and appears DarkMatter theme is a bit broken.
Had the same issue with my WRT32X. Darkmatter Theme was jumbled. The Bootstrap and Material Themes were fine.
Also, Advanced Reboot doesn't recognize the dual-partitions.
1 Like
Advanced Reboot working fine here @r7093.... edit: but i'm using a WRT3200ACM.
This is where in Github the theme is pulled from --> https://github.com/apollo-ng/luci-theme-darkmatter
Appears there are two contributors to the project.
They only recently got Advanced Reboot working on the WRT32000ACM with help from hnyman and stangri. Post about the WRT32X from link below. Provide them necessary info so they can fix it so it works as well:
1 Like
There were no cahnges on github for half a year, so something was broken elsewhere.
Does anyone else here have issues with some Android devices being randomly being kicked off the wifi? The only solution to this problem is to restart the router. Restarting the wifi via Luci or re-authenticating on the phone itself doesn't work.
Have you read the mwlwifi driver bug tracker?
2 Likes
Running davidc's latest build on my WRT3200acm and i have zero problems so far ! Uptime: 5d 22h
I'm using SQM, dnscrypt-proxy, adblock, vlan's, ebtables and everything is working fine.
Wifi performance and range could be abit better but that's not related to his build quality...
btw if you use SQM and your ISP is enforcing a 24h disconnect@pppoe like my ISP, check out the following thread: [Solved/hotfixed] SQM does not reinitialize after ifdown/ifup
1 Like
I wonder if the only difference is setup or maybe my router is screwed up somehow, latest build on my 3200acm starts having the wireless not work after a couple hours, and then after 8 or so hours runs out of memory 