Issues Upgrading from 24.03 to 24.11 SG-1100 Atheros 9280
-
Hmm, what card is that?
I've seen that, or similar, in early 24.08 builds on aarch64.
Steve
-
@stephenw10 Compex WLE200NX 2x2
-
Ah, OK. Looks like an upstream bug in aarch64 PCI unfortunately.
-
@stephenw10 so should be patched in the next development cycle. Do you need me to fill out a red mine bug report?
-
We have an internal bug open already since it's aarch hardware specific. But opening one on the public redmine can't hurt.
It doesn't look like an easy bug to fix. A couple of our devs here looked at it and anticipated it would require some extremely deep diving into the architecture specific code.
-
@stephenw10 Does the DEV team have an idea what may have changed between the two versions to break this?
-
Not specifically. We tested reverting a seemingly likely upstream commit but it didn't change anything.
-
https://redmine.pfsense.org/issues/15472
It has been as issue for a while
-
@LostInIgnorance can you add to this one?
https://redmine.pfsense.org/issues/15472
I had crashes also same card 9280
Chipset -
@JonathanLee said in Issues Upgrading from 24.03 to 24.11 SG-1100 Atheros 9280:
https://redmine.pfsense.org/issues/15472
That's a different crash. In FreeBSD 15 based builds (24.03, 24.11 and any beta rc in between) it will not boot at all with an Ath base card present. It panics trying to initialise the driver.
-
@stephenw10 so if I have the same card in a 2100 it will do the same thing?
-
Yes, I would expect so.
-
@stephenw10 Do you know if the fix for this will be included into the new version for 25?
-
I don't know. But I can test it against current internal builds....hold on....
-
Ok well it's impossible to test because some new bug in the PCIe system prevents the NIC being seen.
pcib0: <Marvell Armada 3700 PCIe Bus Controller> mem 0xd0070000-0xd008ffff irq 5 on simplebus0 pcib0: link up pci0: <OFW PCI bus> on pcib0 pci0: failed to allocate bus number device_attach: pci0 attach returned 6Let me see what I can find....
-
@stephenw10 Do you know if this was added into the list to be fixed in 25.03?
-
Unlikely I would say. The base code is frozen at this point.
-
@LostInIgnorance you could defrost it , maybe fork it and fix it …
-
If if gets fixed it will be in 25.07.
