Hi 2023. június 29., csütörtök 20:46 keltezéssel, Greg KH <gregkh@xxxxxxxxxxxxxxxxxxx> írta: > On Thu, Jun 29, 2023 at 05:30:34PM +0000, Barnabás Pőcze wrote: > > Hi > > > > > > 2023. június 29., csütörtök 18:50 keltezéssel, Linux regression tracking (Thorsten Leemhuis) <regressions@xxxxxxxxxxxxx> írta: > > > > > On 29.06.23 17:28, Barnabás Pőcze wrote: > > > > 2023. június 29., csütörtök 13:31 keltezéssel, Thorsten Leemhuis <regressions@xxxxxxxxxxxxx> írta: > > > >> On 27.06.23 23:55, Barnabás Pőcze wrote: > > > >>> 2023. május 26., péntek 19:55 keltezéssel, Barnabás Pőcze <pobrn@xxxxxxxxxxxxxx> írta: > > > >>>> I have an Acer Aspire A315-58 laptop[0], which, according to lspci has the following intel wireless chip: > > > >>>> > > > >>>> Device: 00:14.3 > > > >>>> Class: Network controller [0280] > > > >>>> Vendor: Intel Corporation [8086] > > > >>>> Device: Wi-Fi 6 AX201 [a0f0] > > > >>>> SVendor: Intel Corporation [8086] > > > >>>> SDevice: Wi-Fi 6 AX201 [0244] > > > >>>> Rev: 20 > > > >>>> ProgIf: 00 > > > >>>> Driver: iwlwifi > > > >>>> Module: iwlwifi > > > >>>> > > > >>>> However, in the kernel message buffer, the following message is written: > > > >>>> > > > >>>> [ 3.633490] iwlwifi 0000:00:14.3: Detected Intel(R) Wi-Fi 6 AX101, REV=0x351 > > > > > > Gregory who is CCed since the start of the thread might care to look > > > into this. According to a quick search lspci is right, but this needs > > > expects. Maybe it's just a cosmetic problem. > > > > > > >>> [...] > > > >>> For the record, I have done a git-bisect[1], and that turned up > > > >>> d2ccc5c1526396aef5196e4dd2ec8538994e5072 > > > >>> > > > >>> as the first good commit[0]. > > > >>> > > > >>> Sadly, that patch (or even the series) does not appear to be easily backportable, > > > >>> so it seems likely that current stable kernels won't be able to support this device. > > > >> > > > >> Thx for bisecting. > > > >> > > > >> Thing is: 6.3 will be EOL in about two or three weeks anyway. And Arch > > > >> will likely switch to 6.4 (which works -- or did I get this wrong?) > > > >> within a week or two. Makes be wonder if it's the best solution for > > > >> everyone involved to not resolve this and move on. > > > > > > > > Yes, 6.4 seems to work. > > > > > > #regzbot inconclusive: unlikely to be fixed in 6.3 before EOL, but 6.4 works > > > > > > > I mainly meant LTS releases. Nonetheless, lspci > > > > and iwlwifi still disagree on what device it actually is. > > > > > > Not sure if I properly understand this. > > > > > > Is this broken in 6.1.y as well, but working in earlier kernels? Then it > > > needs a separate report. > > > > There are two issues: > > > > * lspci and iwlwifi disagreeing about the device (AX201 vs AX101), and > > * the device not working until the aforementioned commit. > > > > This email was originally intended to be about the first one since the device > > does work with 6.4. > > > > But out of curiosity I wanted to find the commit that made the device work to > > see if maybe it could be easily backported to 6.3, 6.1, etc. > > > > In conclusion: > > > > device does not work device is misdetected > > 6.1 present present > > 6.3 present present > > 6.4 fixed present > > > > Since the commits that fixed it don't appear to be easily backportable, > > the only thing remains to be seen is why lspci and iwlwifi disagree about the > > name of the device. That issue is still present in 6.4. > > This isn't a regression then, it's never worked, but now it works in > 6.4, so yeah! And we don't generally backport new hardware enablement > to stable kernels unless it's a simple device id table update. I was hoping the change is just a couple lines of bug fixing, which I imagine, would've been a good candidate for backporting. To be fair, the commit that made it work is in fact a couple lines, but it has many dependencies... > > As for naming, realize that iwlwifi and lspci get the names of the > device from different places, so perhaps those userspace databases just > are not in sync yet? > [...] Yes, I am aware of that. Another thing I have realized is that I could just open the laptop to see what's written on the chip to determine who is right. As it turns out, I did not even have to do that because it says right on the back that The Device Contains WLAN + Bluetooth Module AX101NGW I really should've thought of this before. So if this label is to be believed, then iwlwifi is right, and lspci is incorrect. So sorry everyone for this waste of time. For some reasons I really thought that iwlwifi was in the wrong. I submitted the info to the pci-ids project: https://pci-ids.ucw.cz/read/PC/8086/a0f0 Regards, Barnabás Pőcze
