Mt7621, SNAPSHOT not Enabling Wireless

For Developers
1

Hi,

I have an older build of master (r16019-3ffc30f05a), and wireless is working fine, but in recent builds it seems like wireless is not working. From the older build, in the kernel log,

dmesg | grep "1e140000.pcie"
[    1.421872] mt7621-pci 1e140000.pcie: Parsing DT failed
[    1.541085] mt7621-pci 1e140000.pcie: Error applying setting, reverse things back
[    1.774326] mt7621-pci 1e140000.pcie: pcie2 no card, disable it (RST & CLK)
[    1.781280] mt7621-pci 1e140000.pcie: PCIE0 enabled
[    1.786157] mt7621-pci 1e140000.pcie: PCIE1 enabled
[    1.791033] mt7621-pci 1e140000.pcie: PCI coherence region base: 0x60000000, mask/settings: 0xf0000002
[    1.800514] mt7621-pci 1e140000.pcie: PCI host bridge to bus 0000:00
[   12.526022] mt7621-pci 1e140000.pcie: bus=1 slot=0 irq=23
[   13.401136] mt7621-pci 1e140000.pcie: bus=2 slot=1 irq=24

But in recent builds, the last two rows are missing (device not found / recognized?), and as a result wireless is not working.

Is this a known issue? Is there a fix / workaround?

Thanks!

2

Would be helpful if you'd mention your specific board. FWIW, I'm on 21.02 HEAD with mt 76 2021-03-26 backported and no issues.

Full log instead of a filtered one might help as well.

3

LOL - true! :laughing:. It's an EA7300 V2. I did capture a full log, hopefully this helps. And I just tried rebuilding, 21.02 HEAD like you mention - no difference.

One very odd thing ... I would think the lastest master HEAD would be kernel 5.4.110, but it still seems to be 5.4.101? I did make clean, just to be sure, but nope ... still older? That could be me, no argument there.

Thanks!

4

No definitely something wrong with the kernel version there... you should be on .110. 3ffc30f05a is from February 28th, so it's about six weeks of commits to bisect.

I just went through your log and nothing that catches the eye, unfortunately.

1 Like
5

Thanks! Really appreciate it - and yes, not seeing it here either.

Thinking the same - let me dig. I did pull from upstream, clean (and forced clean Linux). Very odd. Will check it.

6

OK, figured out my stupidity (was on me!). I forgot about my own PR (https://github.com/openwrt/openwrt/pull/3957), and had flashed 21.02 on top of it (to the alternative partition). So then, further updates failed (alternate is r-only without this PR). Manually flashed (from CFE), all updated now - but the problem still exists, no Wi-Fi in recent builds.

1 Like
7

OK, not sure what the cause was (but I have seen this happen before, changing branches, etc.) => did a full make distclean, and now wireless is working again. Thanks for the pointers! Knowing that it was working for you was helpful, pushed me to keep digging :grinning_face_with_smiling_eyes:. Appreciate it!

1 Like
8

I just keep the branches in separate directories, I found too much breaks if you compile e.g. 19.07 / pre-21.02 / master all in the same directory.

1 Like
9

That makes sense, thanks!

10

Its good practise to do a make clean or make dirclean, if changing branches or major patch changes.

1 Like
11

That makes sense, thanks! I did make clean, but really needed to do make dirclean it seems.

Appreciate it!

12

This topic was automatically closed 10 days after the last reply. New replies are no longer allowed.