Bug 101651
Summary: | ath10k QCA6174 2.1 firmware files not available in linux-firmware | ||
---|---|---|---|
Product: | Drivers | Reporter: | Enrico Tagliavini (enrico.tagliavini) |
Component: | network-wireless | Assignee: | drivers_network-wireless (drivers_network-wireless) |
Status: | RESOLVED CODE_FIX | ||
Severity: | normal | CC: | ath9k-devel, calamandrei, dedanna1029, linville |
Priority: | P1 | ||
Hardware: | All | ||
OS: | Linux | ||
Kernel Version: | 4.0, 4.1, 4.2-rc | Subsystem: | |
Regression: | No | Bisected commit-id: |
Description
Enrico Tagliavini
2015-07-17 09:16:48 UTC
I've just got a new Lenovo Z70-80 laptop, have been through distribution after distribution of Linux trying to find a supported driver for the wireless on this chipset. I've found that github has the (still in development)files board.bin and kernelfirmware-5.bin files here: https://github.com/kvalo/ath10k-firmware, but there is literally zero support on how to install these files, i.e. where to, and the proper installation of them. I am currently in the position to where I have to manually install the drivers for the QCA6174, and hope and pray it works. This is the first time I've seen Linux drivers drag Windows ones. I've tried ndiswrapper with the .inf file that I had backed up for the driver on my Lenovo Windows backup; it goes nowhere. I don't think that the driver for this chipset should be held back any longer, properly installed with Linux, as it was first seen as a device on the wikidevi here: https://wikidevi.com/wiki/List_of_Wi-Fi_Device_IDs_in_Linux in April of 2015. I ask, what is the hold up? I can not install Linux without this support. There are distros who are testing with kernel 4.2 already, but they can not support this chip until the drivers for it are released officially, such as Fedora 23. They are on the 4.2 kernel with F23; however, I am finding no real support for my wifi within F23, because they can't contain the support for the QCA6174 yet, since the kernel doesn't officially support it yet. What are we to do??? I want Linux, not Windows; however, I have no choice but to stick with Windows for now, as it supports this chipset. Qualcomm has been "under development" for years now; how about we get it rolling for real for those with new computers (and they stand as many)? As an addendum, I'm not sure as to whether mine is a "Killer", which is doubtful, or as just what it is, the Qualcomm QCA6174. Supporting documentation for my claim(s)can be found at: http://bjoernvold.com/forum/viewtopic.php?f=20&t=3814&start=24 and https://forums.mageia.org/en/viewtopic.php?f=25&t=10272 Thank you for any attention that can be brought to this matter. If we wish to keep up with Windows and surpass it, then we should support the latest, as Linux has always done. "more than a month without activity on the matter" tells me that this is on the back burner; however, there will be many more flood you on this bug if not dealt with. Well that's really not the way you will get attention. What you'll get is, at most, a ban. FYI, there is documentation explaining where to put the firmware files you can find on github in the ath10k-firmware git repo. It is available here https://wireless.wiki.kernel.org/en/users/drivers/ath10k/firmware I can understand you are frustrated, however threatening of flooding the bugzilla if this issue is not dealt with is just a, not effective, black mail. You are just loosing people time for dealing with you instead of dealing with the problem. If you want to join the discussion and be constructive you are welcome on the mailing list. If you don't like the way Qualcomm support its hardware under Linux simply avoid Qualcomm hardware. If you want the best out of the box experience for wireless I would recommend to swap you NIC for an Intel one, 7260 series. 30 euro, worth the frustration free experience I would say. Firmware is released well ahead of hardware. With this, please no more off topics. This bug is meant just as a reminder the firmware is still missing in linux-firmware.git and nothing more than this. Discussion should happen elsewhere (mainly the mailing list), so I will not add any more discussion comments on this bug. Best regards. I do apologise; I did not mean to come off as threatening at least. I was voicing opinion however. You are correct in stating that I am quite frustrated, I've been working on this for some three weeks now, and so far no joy. I am not in the EU, I'm in the states, and unable to get a new nic at this time due to finances. I have a Lenovo Z70-80 laptop with 8GBs of RAM, a 2.4GHz processor. I got this computer just over a month ago. I've kept Windows on it solely so I can be able to connect to the internet, and have dual-booted trying three distributions of Linux so far, one being openSUSE Leap, a testing distro, which ships with kernel 4.2. It did not yet however, have drivers for the QCA6174 from the link you gave, and I downloaded them manually to a local source; however, so far I've had no luck in getting them to work even with distributions that have the 4.2 kernel, as at this time they are testing distros and do not include the driver for the wireless on my machine. Another distro I've tried is Fedora 23, which also ships with the 4.2 kernel, and I've had no luck there either. The distro I have installed right now is Mageia 5, but it shipped with the 3.19 kernel. I reinstalled it three times for various reasons, but the main one is the following, of which I will explain: There appears to be an interface issue with these drivers, or something(?). The first time I installed Mageia 5, I had gone via Windows to one of their update mirrors, downloaded the needed kernel (in Mageia's case, kernel 4.1.8), dependencies for it, etc. When I updated the kernel manually from the local source, it removed btrfs-progs and os-prober, which I believe is a bug of some sort. I went with it anyway, and placed the board.bin and firmware-4.bin in the appropriate folders in /lib/firmware. I ran "echo "options ath10k_core skip_otp=y" >/etc/modprobe.d/ath10k.conf" (without the quotes), and I then rebooted. When I was back in the Mageia desktop, I then started Mageia's Control Center, and tried to create a connection for my wireless. When I did, it came up as unlisted, and asked for me to create it manually (my question right now is why wasn't it detected? It's my ISP connection). I chose to do so, and the only option for WPA that I was given was WPA/WPA2 Enterprise. My connection is not Enterprise. It's simple WPA2-PSK [AES], or "regular" WPA/WPA2 802.11, but there is no option to choose it. On another install, I did try the Windows driver with ndiswrapper, but in Mageia, it seems that ndiswrapper "forgets to load" and I got an interface error with it as well, "Unable to load the ndiswrapper interface." Instead of ndiswrapper, I've decided to stick with using the latest board.bin and firmware-4.bin. I have filed Bug #16915 at http://bugzilla.mageia.org, that the kernel does not support the QCA6174, and the gentleman there has directed me through much of this. However, due to the btrfs/os-prober "bug" I have not reinstalled their 4.1.8 kernel, until I find out for sure if there a patch hiding for it anywhere at their updates mirror. Three distributions however, and I've still got nowhere with this. Frustrating, as I've said, is quite accurate for this. I would like to pursue this further however, and see if anything further can be done. I'll report back when I can get further with Mageia's updates (if there is a patch for the "bug"?), on the current installation of Mageia 5, I went with reiserfs for both / and /home, so I don't understand this with btrfs-progs and os-prober. It still is sitting on the 3.19 kernel this time around, removing the two packages messed up more than one thing. Thank you, and again, I apologise. Oops, I'm sorry, the bug is at https://bugs.mageia.org/. Good news! Few days ago commit aab28b04bcae237f871f4594bbc654da0cf36b50 [0] added the needed firmware! I think this complete the puzzle. I was running already kernel 4.2 and got the firmware from the github ath10k-firmware. Everything works for me now. So I mark my own bug as FIXED. Thank you! [0] https://git.kernel.org/cgit/linux/kernel/git/firmware/linux-firmware.git/commit/?id=aab28b04bcae237f871f4594bbc654da0cf36b50 For me, it is not quite fixed but it's very close. I've been working with Thomas Backlund at the Mageia bugzilla. My hardware is now initializing, and sees my connection but it's seeing it as a WPA/WPA2 Pre-Shared key connection, when it's a WPA2-PSK [AES] connection. Still, I'm going to try it on it with the encryption key and see if it will fly. The bug is here: https://bugs.mageia.org/show_bug.cgi?id=16915 Although give me a bit please before you look, I'm about to add a comment to let him know the latest of what's going on. I've just got done closing Mageia bug #16915. They got it working and made some security fixes as well. Thanks. |