Hi there, Using DELL XPS 9700 i9 and Killer(TM) AX500 WiFi card. I was able to use it with kernel 5.10.1 but not with 5.10.2 and 5.10.3 There the the errors from the log: dmesg|grep ath11 [ 2.081002] ath11k_pci 0000:05:00.0: WARNING: ath11k PCI support is experimental! [ 2.081038] ath11k_pci 0000:05:00.0: BAR 0: assigned [mem 0xd2100000-0xd21fffff 64bit] [ 2.081067] ath11k_pci 0000:05:00.0: enabling device (0000 -> 0002) [ 2.081110] ath11k_pci 0000:05:00.0: failed to get 32 MSI vectors, only -28 available [ 2.081111] ath11k_pci 0000:05:00.0: failed to enable msi: -28 [ 2.081122] ath11k_pci: probe of 0000:05:00.0 failed with error -28 Let me know if you need any more info. Thanks Ivo
Some more info =============== Intel(R) Core(TM) i9-10885H CPU @ 2.40GHz 32GB RAM Artix Linux release LSB_VERSION=1.4 DISTRIB_ID=Artix DISTRIB_RELEASE=rolling DISTRIB_DESCRIPTION="Artix Linux" VERSION_ID=20201207 VARIANT=cinnamon-openrc NAME="Artix Linux" PRETTY_NAME="Artix Linux" ID=artix BUILD_ID=rolling after modprobe ath11k - no messages in the kernel log after modprobe ath11k_pci - the following is displayed. ================= [ 522.906063] ath11k_pci 0000:05:00.0: WARNING: ath11k PCI support is experimental! [ 522.906213] ath11k_pci 0000:05:00.0: BAR 0: assigned [mem 0xd2100000-0xd21fffff 64bit] [ 522.906296] ath11k_pci 0000:05:00.0: failed to get 32 MSI vectors, only -28 available [ 522.906297] ath11k_pci 0000:05:00.0: failed to enable msi: -28 [ 522.906310] ath11k_pci: probe of 0000:05:00.0 failed with error -28 ===================== lsmod|grep ath11k ath11k_pci 24576 0 mhi 81920 1 ath11k_pci ath11k 454656 1 ath11k_pci qmi_helpers 32768 1 ath11k mac80211 1110016 1 ath11k cfg80211 987136 2 ath11k,mac80211 lspci 05:00.0 Network controller: Qualcomm Device 1101 (rev 01) Subsystem: Bigfoot Networks, Inc. Device a501 Flags: fast devsel Memory at d2100000 (64-bit, non-prefetchable) [size=1M] Capabilities: <access denied> Kernel modules: ath11k_pci
kernel 5.10.4 also fails withe same error.
After enabling Vt-D in BIOS and upgrading it to the latest version got a different error now [ 2.267396] ath11k_pci 0000:05:00.0: WARNING: ath11k PCI support is experimental! [ 2.267435] ath11k_pci 0000:05:00.0: BAR 0: assigned [mem 0xd2100000-0xd21fffff 64bit] [ 2.267455] ath11k_pci 0000:05:00.0: enabling device (0000 -> 0002) [ 2.727171] ath11k_pci 0000:05:00.0: Respond mem req failed, result: 1, err: 0 [ 2.727172] ath11k_pci 0000:05:00.0: qmi failed to respond fw mem req:-22 [ 7.880850] ath11k_pci 0000:05:00.0: qmi failed memory request, err = -110 [ 7.880859] ath11k_pci 0000:05:00.0: qmi failed to respond fw mem req:-110 Not sure what is causing this, Probably not connected to the MSI vectors
(In reply to Ivo from comment #3) > After enabling Vt-D in BIOS and upgrading it to the latest version got a > different error now > > [ 2.267396] ath11k_pci 0000:05:00.0: WARNING: ath11k PCI support is > experimental! > [ 2.267435] ath11k_pci 0000:05:00.0: BAR 0: assigned [mem > 0xd2100000-0xd21fffff 64bit] > [ 2.267455] ath11k_pci 0000:05:00.0: enabling device (0000 -> 0002) > [ 2.727171] ath11k_pci 0000:05:00.0: Respond mem req failed, result: 1, > err: 0 > [ 2.727172] ath11k_pci 0000:05:00.0: qmi failed to respond fw mem req:-22 > [ 7.880850] ath11k_pci 0000:05:00.0: qmi failed memory request, err = -110 > [ 7.880859] ath11k_pci 0000:05:00.0: qmi failed to respond fw mem req:-110 > > Not sure what is causing this, Probably not connected to the MSI vectors Firmware WLAN.HST.1.0.1-01740-QCAHSTSWPLZ_V2_TO_X86-1 has issues if host uses memory under 32M limit, I suspect that's the issue here. Try adding "memmap=12M\$20M" to your kernel: https://lore.kernel.org/ath11k/056b74d48fb735c69a6268b18184d7e8@codeaurora.org/
Reserved some ranged of unassigned memory for use by emulated persistent memory and added it to the grub boot options with "memmap=12M\\\$20M" as with one backslash the grub interpreted the dollar sign as a shell command. Looks OK now. I can use the Wi-Fi card and the kernel log messages are: [ 2.092063] ath11k_pci 0000:05:00.0: WARNING: ath11k PCI support is experimental! [ 2.092100] ath11k_pci 0000:05:00.0: BAR 0: assigned [mem 0xd2100000-0xd21fffff 64bit] [ 2.092120] ath11k_pci 0000:05:00.0: enabling device (0000 -> 0002) [ 2.609765] ath11k_pci 0000:05:00.0: chip_id 0x0 chip_family 0xb board_id 0xff soc_id 0xffffffff [ 2.609766] ath11k_pci 0000:05:00.0: fw_version 0x101c06cc fw_build_timestamp 2020-06-24 19:50 fw_build_id Thank you!
I'll keep this bug open until there's a fix for the firmware. Also changing the title to reflect the issue is better.
I'm having similar problems with this Wi-Fi card. I added the "memmap=12M\$20M" line to the boot, and now the error I am getting is this: https://owenlynch.org/static/ath11k_error.jpeg Any ideas on what might be causing this?
For some reason my webserver doesn't like jpegs? Here's a png: https://owenlynch.org/static/ath11k_error.png
(In reply to Owen Lynch from comment #7) > I'm having similar problems with this Wi-Fi card. I added the > "memmap=12M\$20M" line to the boot, and now the error I am getting is this: > > https://owenlynch.org/static/ath11k_error.jpeg > > Any ideas on what might be causing this? I recommend double checking from /proc/cmdline that memmap parameter is correctly set. It should look like this: $ cat /proc/cmdline [...] memmap=12M$20M [...] Next I would make sure the firmware files are not corrupted: $ md5sum /lib/firmware/ath11k/QCA6390/hw2.0/* a101dc90f8e876f39383b60c9da64ec4 /lib/firmware/ath11k/QCA6390/hw2.0/amss.bin 4c0781f659d2b7d6bef10a2e3d457728 /lib/firmware/ath11k/QCA6390/hw2.0/board-2.bin d4c912a3501a3694a3f460d13de06d28 /lib/firmware/ath11k/QCA6390/hw2.0/m3.bin ddc7afcb118def559bf131c3ec2563af /lib/firmware/ath11k/QCA6390/hw2.0/Notice.txt If that doesn't help, please file a new bug. Let's keep one issue per report, please.
As I'm receiving more reports, just a little more background of this issue. This was first seen in v5.10-rc1 as there was a change in mm: 7fef431be9c9 mm/page_alloc: place pages to tail in __free_pages_core() https://git.kernel.org/linus/7fef431be9c9 Reverting that commit would also workaround this firmware issue. In some platforms the memmap above is not enough. For example, in one x86 box I had to use memmap=14M\\\$18M. To make it easier to detect this issue there's now a qmi debug message which prints the buffer addresses (will be in v5.12-rc1): https://git.kernel.org/pub/scm/linux/kernel/git/kvalo/ath.git/commit/?h=ath-next&id=d5395a548659
I'm seeing a similar but possibly different issue on Dell XPS 15 9500 with latest Bios (1.8.1). $ uname -a Linux xps15 5.13.11 #1-NixOS SMP Sun Aug 15 12:01:57 UTC 2021 x86_64 GNU/Linux $ cat /proc/cmdline initrd=\efi\nixos\7nqynfr6b72q6ycivac4lwfs4ibvfm0v-initrd-linux-5.13.11-initrd.efi init=/nix/store/1kcalf57sypcvdz7a73ld7v6lvqhpai2-nixos-system-xps15-21.05.2501.94c989365d5/init [...] memmap=12M$20M loglevel=4 $ modinfo ath11k_pci filename: /run/current-system/kernel-modules/lib/modules/5.13.11/kernel/drivers/net/wireless/ath/ath11k/ath11k_pci.ko.xz firmware: ath11k/QCA6390/hw2.0/m3.bin firmware: ath11k/QCA6390/hw2.0/amss.bin firmware: ath11k/QCA6390/hw2.0/board-2.bin license: Dual BSD/GPL description: Driver support for Qualcomm Technologies 802.11ax WLAN PCIe devices alias: pci:v000017CBd00001101sv*sd*bc*sc*i* depends: mhi,ath11k retpoline: Y intree: Y name: ath11k_pci vermagic: 5.13.11 SMP mod_unload Here is the log: https://pastebin.com/YrMEictS Should I open a new issue for this? Thank you!
Just upgraded to kernel 5.14.4 Updated to the latest BIOS as well. Problem reappeared with different messges: failed to register fw indication: -110 failed to send qmi firmware indication: -110 Thanks Ivo
This is a regression which was caused by commit in 5.14.4 and 5.13.16: 3f38cc1ba1e289306f52cb7238406f38ecbfa21a net: qrtr: make checks in qrtr_endpoint_post() stricter Here is the bugreport: https://bugzilla.kernel.org/show_bug.cgi?id=214455
The 32M memory limit bug is now fixed in this firmware version: https://github.com/kvalo/ath11k-firmware/tree/master/QCA6390/hw2.0/1.0.1/WLAN.HST.1.0.1-05266-QCAHSTSWPLZ_V2_TO_X86-1 With that firmware release memmap workaround is not needed anymore. Closing the bug.
As this broke in 5.10 and worked in earlier releases, marking as a REGRESSION. As the fix requires new firmware, rather than a kernel "CODE_FIX" patch, marking as closed/DOCUMENTED.
*** Bug 215998 has been marked as a duplicate of this bug. ***
Is there another reason this message could be triggered? I started seeing this in 6.3.7. ~> sudo dmesg | grep ath11k 17:31:57 [ 11.441679] ath11k_pci 0000:6c:00.0: BAR 0: assigned [mem 0xb4200000-0xb42fffff 64bit] [ 11.441715] ath11k_pci 0000:6c:00.0: enabling device (0000 -> 0002) [ 11.442189] ath11k_pci 0000:6c:00.0: MSI vectors: 32 [ 11.442199] ath11k_pci 0000:6c:00.0: qca6390 hw2.0 [ 14.574034] ath11k_pci 0000:6c:00.0: ignore reset dev flags 0x8000 [ 21.940059] ath11k_pci 0000:6c:00.0: failed to wait qmi memory request: -110 [ 21.940066] ath11k_pci 0000:6c:00.0: qmi failed to respond fw mem req: -110 Firmware seems to be up to date: modinfo ath11k_pci 17:35:26 filename: /lib/modules/6.3.7-arch1-1/kernel/drivers/net/wireless/ath/ath11k/ath11k_pci.ko.zst firmware: ath11k/QCA6390/hw2.0/m3.bin firmware: ath11k/QCA6390/hw2.0/amss.bin firmware: ath11k/QCA6390/hw2.0/board-2.bin license: Dual BSD/GPL description: Driver support for Qualcomm Technologies 802.11ax WLAN PCIe devices srcversion: FFC77E7DB9B98E6850D2BDA alias: pci:v000017CBd00001104sv*sd*bc*sc*i* alias: pci:v000017CBd00001103sv*sd*bc*sc*i* alias: pci:v000017CBd00001101sv*sd*bc*sc*i* depends: mhi,ath11k retpoline: Y intree: Y name: ath11k_pci vermagic: 6.3.7-arch1-1 SMP preempt mod_unload sig_id: PKCS#7 signer: Build time autogenerated kernel key sig_key: 1A:AF:97:2E:13:77:23:DE:23:E6:E7:80:AD:98:5E:5C:F1:2D:C8:D7 sig_hashalgo: sha512 signature: 30:66:02:31:00:D4:FB:48:5F:5A:4A:76:81:9E:6D:72:B7:39:54:59: D9:CF:FB:62:E3:AA:5C:DE:0D:0C:CF:08:44:46:A6:12:67:6D:8E:B6: 20:B1:D8:57:FB:0B:D7:BB:72:BE:DA:E1:54:02:31:00:85:4E:F8:B6: E3:F1:4E:66:DF:B1:B8:57:01:FC:5E:D9:F6:45:24:08:EF:59:87:80: F0:91:F0:4A:A3:8B:49:B5:46:66:30:76:52:5E:5C:89:DE:13:65:05: 2D:1F:8B:E8
> Is there another reason this message could be triggered? I started seeing > this > in 6.3.7. One issue per report, please. Please file a new bug.