New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
FS#3765 - RT5592 initializes but does not function when enabled #8780
Comments
mcpratt: hexdump of eeprom:
00000000 92 55 07 01 88 dc 96 aa bb cc 92 55 14 18 01 80 |.U.....#...U....|
00000010 00 00 92 55 14 18 00 00 01 00 6a ff 13 02 ff ff |...U......j.....|
00000020 ff ff c1 ff 01 92 ff ff ff ff ff ff ff ff ff ff |................|
00000030 8e 75 01 43 22 ff 27 00 ff ff 17 01 ff ff d9 aa |.u.C".'.........|
00000040 cc 88 ff ff 08 07 00 00 00 07 00 00 00 07 ff ff |................|
00000050 ff 00 05 05 05 05 05 05 05 05 05 05 05 05 05 05 |................|
00000060 00 00 00 00 00 00 00 00 00 00 00 00 00 00 42 73 |..............Bs|
00000070 f7 00 86 0f 00 00 72 fc 0c 0c 0c 0c 0c 0c 0c 0c |......r.........|
00000080 0c 0c 0d 0d 0e 0e 0e 0e 0e 0e 0e 0e 0e 0e 0e 0e |................|
00000090 0e 0e 0f 0f 0d 0d 0d 0e 0e 0f 0f ff ff ff ff ff |................|
000000a0 ff ff ff ff ff ff 0c 0c 0d 0d 0e 0f 0f 0f 0f 0f |................|
000000b0 10 10 10 10 10 0f 0f 0e 0e 0d 0d 0d 0e 0e 0f 0f |................|
000000c0 10 0e 0f 0f 0f 0f 0f 0f 10 ff ff ff ff ff ff ff |................|
000000d0 ff ff ff ff 43 43 f3 00 21 fa 00 00 72 e1 66 66 |....CC..!...r.ff|
000000e0 aa aa 88 66 aa aa 88 44 aa aa 88 44 aa aa 88 44 |...f...D...D...D|
000000f0 ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff |................|
00000100 ff ff ff ff ff ff ff ff ff ff ff ff 00 00 ff ff |................|
00000110 ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff |................|
00000120 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 |................|
*
00010000
|
nbriggs: I can confirm that this issue is present on the ESR600. |
I confirm the issue on a mikrotik MIPS device with a CSL 27395 USB device with an RT5592 chip connected via USB.
It looks fine initially. I also manage to get a bit of connection (a client connect showing up in LuCI), but things break after that. Dmesg suggests that the USB connection or the device itself die … maybe there's not enough power. Maybe totally unrelated, but maybe not. An Atheros USB device works nicely (but only 1x1 setup with 72 MBit/s, not 144). |
Red herring. I tested with a USB switch with external power supply, ruling out flaky contacts/not enough power. Also without device resets, there is no real communication going on. The activity LED on the device doesn't light up once, apparently, while it does flash all the time when used on a client laptop. So is there something that just broke AP operation of RT5592? Are other rt2800-style devices working fine? |
This bug also affects the Asus RP-N53 Repeater/Access Point:
|
This bug also affects Buffalo WSR-600DHP. It makes the 5Ghz wifi completely unusable, i.e., the 5Ghz ssid is visible, but does not connect.
|
@sobukus to be clear, you do not see any kernel messages like |
Yes. I just tried again. All I get (if the USB connection is not loose) is
This was swtiching from one USB wifi device to another (stopping wifi and then
No proper connection establishing … |
what is the model of the usb device? or is there a similar one I can get? the mikrotik is an mt7620 or ath79? |
The Mikrotik is an RB750Gr3, with the wifi connected to the USB port.
The USB wifi is a CSL 27395. It's a 300 MBit/s dual-antenna wifi not needing USB 3 that I hoped would work well, based on the Ralink chip. Wondering now … the product page of $store says RT5572, dmesg says RT5592. The label on the chip says RT5572. The board carries the name MT-WN813N-D-V2.0. That may help to identify an identical product. Having said that … there's not much more besides the two antenna ports and the chip on it, so I'd expect any RT5572 device behaving the same. |
can you install usbutils and run |
Sure.
For good measure, the working USB wifi:
Edit: With 'working', I mean that it normally works, but has occasional interruptions in the connection. During video conferences, this is rather annoying, but it's mostly OK, for a year now. I had hoped for better performance using the dual-antenna device with another chip (it's just a few meters, so it is not really about the antennas, I guess). |
it might be missing kernel config symbol
|
nevermind I forgot its made as a module https://github.com/openwrt/openwrt/blob/openwrt-21.02/package/kernel/mac80211/ralink.mk#L31 |
@sobukus I got my hands on a RT5572 usb and it's misidentified as a RT5592 in kernel log like you said, However, I plugged it into an ath79 device (specifically QCA9558), not an mt762x device. going to try an mt7620 device next. have you made sure to install the right driver and reboot? https://openwrt.org/packages/pkgdata/kmod-rt2800-usb If you still have trouble, I would suspect the USB driver as the problem relevant log output
|
I am pretty sure we got the same driver being active, loading the same firmware file, etc. If you could try a mt7620 and verify yourself (or find a setup that actually works;-), that would be great! A somewhat flaky USB would explain that the other device also does not work perfectly stable. And sure, routers with onboard wifi do exist and are known to work better, as would a separate access point. But putting that USB port to some use would be more fun, and in my case conserve resources (power and device production). |
in the meantime would you happen to have an ath79 device to test it with? |
Nope, I only played with openwrt on this one mikrotik device. Maybe I can ask around. |
With the latest OpenWRT (2023-03-20 snapshot, firmware 0.40) there are no more "unrecognized signal" or "wrong frame size" errors in dmesg (on the RP-N53), but the behaviour is the same (clients cannot connect) |
@mpratt14 since you added this, care to comment on the following changes?
seems to be copy/paste from the other engenius device where the pci reg values are a bit different:
|
if you're trying to solve this issue, that won't make a difference I believe there's something wrong with the PCI driver that makes it incompatible with the rt5592 |
eh no I'm trying to fix dtc warnings. Addresses are wrong here. The pci id is as well. |
what's the warning? seems really strict...I don't see anything wrong with the way it is... |
full commit: neheb@af1d5d3 the reg values are wrong. that's it. I copied what other dts files do to fix. |
What I put in DTS is just from what I saw in dmesg. I don't know that much about PCI so maybe the driver problems are causing a nonsense configuration to be displayed... I can't tell what's wrong here... I use the class value Side note on naming the "wifi" node: I've been told that the numbers after
|
Another very old issue also reports the PCI ID discrepancy |
mcpratt:
this issue is also present on master branch
board is EPG600 (not added yet)
this probably also affects ESR600 which is the same hardware
and other boards with RT5592
5 GHz wireless initializes and SSID can be seen on scan
however, clients cannot connect and traffic fails
init:
iw list (phy0):
errors:
The text was updated successfully, but these errors were encountered: