| 29 May 2025 |
 K900 | Do it | 02:55:12 |
| K900 | If you have a flowtable | 02:55:17 |
 emily | yeah it seemed to | 02:55:45 |
| emily | working with mini jumbo frames too | 02:55:53 |
| emily | also 6 GHz WiFi seems to work pretty well, since I guess the txpower data for those doesn't get messed up on the messed up units for some reason?? | 02:56:24 |
| emily | I have no idea how they managed that | 02:56:30 |
| emily | like the EEPROM isn't garbage, it's just… missing the data for 2.4 and 5 GHz? | 02:56:40 |
| K900 | I don't know | 02:57:20 |
| K900 | Mine is perfectly fine | 02:57:43 |
| emily | I get the impression that at some point they broke it and now all new units ship out bad. | 02:58:46 |
 @uep:matrix.org | pppoe hardware offload.. mini jumbo frames..
they have taken us for absolute fools, etc | 02:58:47 |
| emily | I wouldn't be doing PPP if I could avoid it :) | 02:59:01 |
| emily | I need to figure out how to enable WED too. | 03:02:30 |
| emily | even if it's kinda broken it'd be interesting to see. | 03:02:41 |
| emily | my phone can't quite max out my WAN over WiFi seemingly | 03:03:03 |
| K900 | OpenWRT should have all the firmware needed | 03:03:19 |
| K900 | So you should be able to just set mem=2G and then the module knob that I forget | 03:03:56 |
| K900 | And supposedly it works until it doesn't | 03:03:58 |
| emily | I still don't understand why you'd have to limit your RAM. | 03:04:22 |
| K900 | Because the firmware can't handle large addresses, presumably | 03:07:15 |
| emily | so gross. | 03:07:57 |
| emily | well there was a new firmware release this month so maybe it'll work better. | 03:08:04 |
| emily | I hoped it would solve the txpower thing but it seems to be working no worse (and no better) | 03:08:19 |
| K900 | There is at least one IP block on there that operates on 31-bit addresses | 03:08:47 |
| K900 | I don't know why | 03:08:52 |
| K900 | In reply to @emilazy:matrix.org
well there was a new firmware release this month so maybe it'll work better. That's the wrong firmware | 03:08:57 |
| K900 | There's a separate firmware running on the Ethernet end and that's the one that's broken | 03:09:11 |
| K900 | That one is mt7988_wo_*.bin IIRC | 03:10:11 |
 matthewcroughan | I couldn't, you have to "take pictures" of the device in its broken state, and then they let you speak to the vendor, but it times out after a while, you would also have to ship it back to them and disassemble | 10:05:22 |
| matthewcroughan | From 8088940ffd047a6d282a95af829562e8038f5b2d Mon Sep 17 00:00:00 2001
From: Yukari Yakumo <mistelinn@gmail.com>
Date: Wed, 5 Feb 2025 09:33:47 +0300
Subject: [PATCH] Add tx_power check
---
drivers/net/wireless/mediatek/mt76/mt7996/eeprom.c | 7 +++++++
1 file changed, 7 insertions(+)
diff --git a/drivers/net/wireless/mediatek/mt76/mt7996/eeprom.c b/drivers/net/wireless/mediatek/mt76/mt7996/eeprom.c
index 53dfac02f..a699668ca 100644
--- a/drivers/net/wireless/mediatek/mt76/mt7996/eeprom.c
+++ b/drivers/net/wireless/mediatek/mt76/mt7996/eeprom.c
@@ -206,6 +206,13 @@ static int mt7996_eeprom_load(struct mt7996_dev *dev)
goto out;
}
}
+
+ /* read tx_power values from fw */
+ u8 *eeprom = dev->mt76.eeprom.data;
+ if (!eeprom[MT_EE_TX0_POWER_2G] || !eeprom[MT_EE_TX0_POWER_5G] || !eeprom[MT_EE_TX0_POWER_6G] ) {
+ use_default = true;
+ goto out;
+ }
}
out:
| 10:07:23 |
