Kernel tainted after resume (QCA9377/hw1.0/TF2.1-firmware-v6)

[Tobias Predel]

Hello everyone,

this happens on 4.20.7 (Arch Linux) with 
ath10k/QCA9377/hw1.0/TF2.1/firmware-6.bin (WLAN.TF.2.1-00021-QCARMSWP-1) while 
suspending to RAM and resuming. It doesn't break functionality (reconnection 
works), but it doesn't look good in the logs.

Regards,
Tobias Predel

[  107.302390] ath10k_pci :02:00.0: unsupported HTC service id: 1536
[  107.317978] ath10k_pci :02:00.0: Unknown eventid: 118809
[  107.320886] ath10k_pci :02:00.0: Unknown eventid: 90118
[  107.346304] WARNING: CPU: 1 PID: 1425 at 
drivers/net/wireless/ath/ath10k/mac.c:5650 ath10k_bss_info_changed+0xfb7/0x1120 
[ath10k_core]
[  107.346310] Modules linked in: dm_crypt algif_skcipher af_alg dm_mod rfcomm 
fuse ccm cmac arc4 bnep snd_hda_codec_hdmi snd_hda_codec_realtek 
snd_hda_codec_generic hid_multitouch ofpart cmdlinepart hid_generic joydev 
mousedev intel_spi_platform intel_spi spi_nor asus_nb_wmi asus_wmi mtd btusb 
btrtl iTCO_wdt sparse_keymap btbcm iTCO_vendor_support btintel intel_rapl 
bluetooth intel_powerclamp ath10k_pci coretemp kvm_intel ath10k_core usbhid 
crct10dif_pclmul crc32_pclmul ecdh_generic ghash_clmulni_intel ath i915 
nls_iso8859_1 mac80211 uas nls_cp437 vfat fat squashfs aesni_intel loop 
aes_x86_64 kvmgt vfio_mdev mdev crypto_simd vfio_iommu_type1 cryptd vfio 
glue_helper intel_cstate intel_xhci_usb_role_switch kvm roles input_leds pcspkr 
ak8975 irqbypass i2c_algo_bit drm_kms_helper cfg80211 snd_hda_intel 
snd_hda_codec drm snd_hda_core snd_hwdep snd_pcm snd_timer intel_gtt agpgart 
snd syscopyarea mei_txe sysfillrect processor_thermal_device sysimgblt rfkill 
int340x_thermal_zone i2c_i801 i2c_hid
[  107.346467]  fb_sys_fops mei lpc_ich soundcore intel_soc_dts_iosf 
inv_mpu6050_i2c uvcvideo battery wmi inv_mpu6050 hid 
industrialio_triggered_buffer kfifo_buf i2c_mux videobuf2_vmalloc tpm_crb 
tpm_tis industrialio tpm_tis_core videobuf2_memops ac tpm asus_wireless 
pwm_lpss_platform videobuf2_v4l2 evdev pwm_lpss elan_i2c rng_core pcc_cpufreq 
int3400_thermal acpi_thermal_rel sd_mod mac_hid intel_int0002_vgpio 
videobuf2_common videodev media sg crypto_user ip_tables x_tables usb_storage 
scsi_mod ext4 crc32c_generic crc16 mbcache jbd2 fscrypto mmc_block serio_raw 
atkbd libps2 crc32c_intel xhci_pci xhci_hcd i8042 serio sdhci_acpi sdhci 
mmc_core
[  107.346597] CPU: 1 PID: 1425 Comm: kworker/u8:12 Tainted: GW 
4.20.7-arch1-1-ARCH #1
[  107.346602] Hardware name: ASUSTeK COMPUTER INC. E205SA/E205SA, BIOS 
E205SA.302 02/10/2017
[  107.346615] Workqueue: events_unbound async_run_entry_fn
[  107.346654] RIP: 0010:ath10k_bss_info_changed+0xfb7/0x1120 [ath10k_core]
[  107.346661] Code: 4f 67 00 00 e9 a5 f1 ff ff 0f 0b 4c 89 f7 e8 00 ca d3 dd 
e9 ff f3 ff ff b8 40 00 00 00 e9 65 f5 ff ff 8b 95 70 01 00 00 eb c4 <0f> 0b e9 
02 f3 ff ff 89 c2 48 c7 c6 80 72 f4 c0 4c 89 cf 89 04 24
[  107.346665] RSP: 0018:ab0b40613cd0 EFLAGS: 00010282
[  107.346674] RAX: fffe RBX: 9e85ee0d1580 RCX: 
[  107.346679] RDX: 9e85ee0d18f8 RSI: ab0b40613cf0 RDI: 9e85e4d9b3a0
[  107.346683] RBP: 9e85e4d9b3a0 R08: 0001 R09: 
[  107.346687] R10: 001f R11:  R12: 020674ff
[  107.346692] R13: 9e85e4d9b3a8 R14: 9e85ee0d2558 R15: 9e85ee0d1580
[  107.346697] FS:  () GS:9e85f5c8() 
knlGS:
[  107.346702] CS:  0010 DS:  ES:  CR0: 80050033
[  107.346706] CR2: 55a5eba9a8e0 CR3: 2ee0a000 CR4: 001006e0
[  107.346710] Call Trace:
[  107.346757]  ? ath10k_conf_tx+0x7a/0x4b0 [ath10k_core]
[  107.346840]  ieee80211_bss_info_change_notify+0xb5/0x1c0 [mac80211]
[  107.346926]  ieee80211_reconfig+0x6c3/0x14d0 [mac80211]
[  107.346940]  ? preempt_count_add+0x79/0xb0
[  107.347018]  wiphy_resume+0x7f/0x150 [cfg80211]
[  107.347088]  ? wiphy_namespace+0x10/0x10 [cfg80211]
[  107.347097]  dpm_run_callback+0x59/0x180
[  107.347107]  device_resume+0xac/0x1f0
[  107.347116]  async_resume+0x19/0x30
[  107.347123]  async_run_entry_fn+0x37/0x140
[  107.347135]  process_one_work+0x1eb/0x410
[  107.347145]  worker_thread+0x2d/0x3d0
[  107.347153]  ? process_one_work+0x410/0x410
[  107.347161]  kthread+0x112/0x130
[  107.347169]  ? kthread_park+0x80/0x80
[  107.347181]  ret_from_fork+0x35/0x40
[  107.347194] ---[ end trace 023720a07093ea74 ]---
[  107.347286] WARNING: CPU: 1 PID: 1425 at 
drivers/net/wireless/ath/ath10k/mac.c:5688 ath10k_bss_info_changed+0xf96/0x1120 
[ath10k_core]
[  107.347289] Modules linked in: dm_crypt algif_skcipher af_alg dm_mod rfcomm 
fuse ccm cmac arc4 bnep snd_hda_codec_hdmi snd_hda_codec_realtek 
snd_hda_codec_generic hid_multitouch ofpart cmdlinepart hid_generic joydev 
mousedev intel_spi_platform intel_spi spi_nor asus_nb_wmi asus_wmi mtd btusb 
btrtl iTCO_wdt sparse_keymap btbcm iTCO_vendor_support btintel intel_rapl 
bluetooth intel_powerclamp ath10k_pci coretemp kvm_intel ath10k_core 

[no subject]

[Tobias Predel]

Hello,

I want to apologize for my recent mail that was very rude. It needed new
AP hardware. 802.11ac is okay now.

Sorry for the inconvenience.

Regard,
Tobias

___
ath10k mailing list
ath10k@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/ath10k

QCA9377 stops working operating on 802.11g when WMM disabled

[Tobias Predel]

Hello,

playing with my QCA9377 card and my 802.11 b+g (g++) router (Fritz!Box 7170 
doing still a great job)
in order to check interoperability I get this:

When I disable WMM (IEEE 802.11e) on the web interface of my router that is 
operating
on 2,4 GHz channel with IEEE 802.11g, ath10k can't manage to establish a 
reliable
connection any more (no VHT and HT, so no WMM/QoS). Instead, after seemingly 
having established a connection (authenticated/associated), it disconnects 
directly after that giving reason=1 UNSPECIFIED (following deauthentification). 

The product website of QCA9377 says that the chipset supports the complete 
range of 
commonly used WLAN standards (802.11a/b/g/n/ac). Other devices continue
working properly (without WMM) even on my Linux Android phone so I think they 
implement support for IEEE 802.11 (g) standard in the right way while
QCA9377 screws up already at key negotiation state on my laptop here
running Linux 4.12.4 (the router log says wrong passphrase but it's the 
same as on the other devices I can guarantee). 

And even with WMM enabled, the connection is ruther lousy with periods 
of very high transfer rate following periods of drought. Is it 
possible that 
(1) WMM and/or 802.11g isn't implemented in the firmware or somewhere else in a 
standard-compilant way 
or 
(2) is it that the firmware requires WMM to be activated in order to
operate on 802.11g (1) given?

Or more generally speaking, is g/b and so on supported by those devices
as they focus on 802.11 ac support? I am currently using 

driver: ath10k_pci
version: 4.12.4-1-ARCH
firmware-version: WLAN.TF.1.0-00267-1
expansion-rom-version: 
bus-info: :02:00.0
supports-statistics: yes
supports-test: no
supports-eeprom-access: no
supports-register-dump: no
supports-priv-flags: no

as ethtool says and

[4.673694] ath10k_pci :02:00.0: pci irq msi oper_irq_mode 2 irq_mode 0 
reset_mode 0
[4.957885] ath10k_pci :02:00.0: qca9377 hw1.0 target 0x0502 chip_id 
0x003820ff sub 1a3b:2231
[4.957890] ath10k_pci :02:00.0: kconfig debug 0 debugfs 1 tracing 0 dfs 
0 testmode 0
[4.958574] ath10k_pci :02:00.0: firmware ver WLAN.TF.1.0-00267-1 api 5 
features ignore-otp crc32 79cea2c7
[5.030545] ath10k_pci :02:00.0: board_file api 2 bmi_id N/A crc32 
8aedfa4a
[6.838415] ath10k_pci :02:00.0: htt-ver 3.1 wmi-op 4 htt-op 3 cal otp 
max-sta 32 raw 0 hwcrypto 1
[6.860542] ath10k_pci :02:00.0 wlp2s0: renamed from wlan0
[   15.923945] ath10k_pci :02:00.0 wlp2s0: disabling HT as WMM/QoS is not 
supported by the AP
[   15.923960] ath10k_pci :02:00.0 wlp2s0: disabling VHT as WMM/QoS is not 
supported by the AP

as dmesg says. I think that I'm using the latest binary blobs of Qualcomm 
regarding firmware and EEPROM board. 
I mean QCNFA435 -> QCA9377? Would it be possible to provide all firmware and 
board BLOBs (different versions etc.) regarding
ath10k instead of squeezing each file out of the Qualcomm staff?

Regards,
Tobias

___
ath10k mailing list
ath10k@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/ath10k

Re: Recent driver changes destabilized QCA9377 connection quality

[Tobias Predel]

Hello,

The driver recognizes my hardware as "qca9377 hw1.0 (0x0502, 0x003820ff sub 
1a3b:2231)" with firmware details "fw WLAN.TF.1.0-00267-1 fwapi 5 bdapi 2 
htt-ver 3.1 wmi-op 4 htt-op 3 cal otp max-sta 32 raw 0 hwcrypto 1 features 
ignore-otp". Firmware file is saved as firmware-5.bin.

I did remove the board.bin file while keeping board-2.bin for testing but that 
didn't help either. With or without that board.bin no more error messages are 
thrown up but connection losses (sometimes).

Cheers,

Tobias
-- 
Tobias Predel
currently studying Transportation engineering B.Sc.
at Stuttgart University

Mails to amne...@hotmail.com (and st108...@stud.uni-stuttgart.de) are 
automatically forwarded to tobias.pre...@gmail.com

___
ath10k mailing list
ath10k@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/ath10k

Recent driver changes destabilized QCA9377 connection quality

[Tobias Predel]

Hello,

I just want to report a negative development I have been experiencing for a 
week now.

Recent changes in ath10k driver code seem to destabilize QCA9377 connection 
quality in 802.11b/g networks, as hardware and software configuration 
(wpa_supplicant) hasn't changed. Before, I have managed to have a fair 
connection quality for around three months. At the moment I encounter multiple 
connection losses.

dmesg reports usual losses and timeouts on both linux-lts 4.4.44 and linux 
4.8.13 (Arch Linux). 

Please restore former functionality.

Sincerely yours,

Tobias Predel
-- 
Tobias Predel
currently studying Transportation engineering B.Sc.
at Stuttgart University

___
ath10k mailing list
ath10k@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/ath10k

Re: ath10k and QCA3977 - random transfer rate losses

[Tobias Predel]

Hello vagner,

I'm sorry but I don't believe that I can tell a solution instantly :(

But nevertheless that makes me wonder since the bluetooth of my model 
doesn't require additional files for Bluetooth - works 'out of the box'. 

On my Arch Linux I had to enable and start the systemd service Bluetooth (starts
the Bluez bluetoothd daemon) and load the btusb kernel module with 
# modprobe btusb

Maybe it's (also) soft blocked by rfkill?
# rfkill unblock bluetooth 
should then help, after that Bluetooth may appear in NetworkManager applet 
again.

In the bluetooth driver packages for Windows there are files named 
ramps_0x**_**_*.dfu, AtherosBT.bin and AthrBT_0x**.dfu.
But most of these files are firmware updates files that can be written
to the chipset with the dfu-util. I don't think that they are 
really needed for this (just for information). 

You may post the message buffer of the kernel with 'dmesg'? Maybe
much more skilled people than me can help. I'll try my best :)

Best regards,
Tobias

On Mon, Feb 22, 2016 at 11:12:55AM +0200, vagner_ wrote:
> Hi Tobias,
> 
> Do you think that aplying the right board.bin makes the Bluetooth work
> properly?
> My wireless card
> 
> Network controller: Qualcomm Atheros Device 0042 (rev 30)
> 
> Has got bluethooth capabilities, but does not work. On boot time it
> complains about a file called ranpatch (...).bin related to it  :(
> 
> ++
> 
> On 02/21/2016 11:54 PM, Tobias Predel wrote:
> >Hello Adrian,
> >
> >bug report can be found here 
> ><https://bugzilla.kernel.org/show_bug.cgi?id=112451#c10>
> >I added other firmware files. I don't know any other place.
> >
> >Tobias
> >
> >On Sun, Feb 21, 2016 at 11:25:52AM -0800, Adrian Chadd wrote:
> >>ok, well, we should file a bug report so this isn't forgotten.
> >>
> >>Where's the right place to create bug reports? is there an ath10k
> >>staging branch in github somewhere? or bugzilla.kernel.org ?
> >>
> >>
> >>-a
> >>
> >>
> >>On 21 February 2016 at 11:20, Tobias Predel <tobias.pre...@gmail.com> wrote:
> >>>Yes. I confirm full functionality now.
> >>>
> >>>On Sun, Feb 21, 2016 at 10:33:26AM -0800, Adrian Chadd wrote:
> >>>>Right, so just to be clear - it's just the board file updates that
> >>>>make it all work fine?
> >>>>
> >>>>
> >>>>-a
> >>>>
> >>>>
> >>>>On 21 February 2016 at 06:44, Tobias Predel <tobias.pre...@gmail.com> 
> >>>>wrote:
> >>>>>Hello again,
> >>>>>
> >>>>>one last word about the the board files:
> >>>>>
> >>>>>I extracted the board files from an AsusTek Windows Driver installer
> >>>>>that included the Qualcomm driver files for diverse variants of QCA9377
> >>>>>built in Samsung, Asus (Askey), Acer, Lenovo notebooks.
> >>>>>
> >>>>>An archive with all these files can be found at the comment 10 of my bug 
> >>>>>report
> >>>>><https://bugzilla.kernel.org/show_bug.cgi?id=112451#c10>
> >>>>>including board files (eeprom_*.bin) for NFA425, NFA435 (A), NFA455.
> >>>>>
> >>>>>If some people experience trouble, those might look at the netathr10x.inf
> >>>>>to find out which eeprom board file is needed or try out by installing
> >>>>>it to /lib/firmware/ath10k/QCA9377/hw1.0/board.bin.
> >>>>>
> >>>>>Those board files are using api 1 and must be named board.bin.
> >>>>>firmware-5.bin is still to be pulled from kvalo's github repo.
> >>>>>
> >>>>>Thank you to all for your support. I wish the best!
> >>>>>
> >>>>>Tobias
> >>>>>
> >>>>>
> >>>>>On Fri, Feb 19, 2016 at 05:55:26PM +, Hsu, Ryan wrote:
> >>>>>>Glad that help!
> >>>>>>
> >>>>>>I've cc that file to Kalle, hope we can see that get merged soon.
> >>>>>>
> >>>>>>Ryan
> >>>>>>
> >>>>>>>-Original Message-
> >>>>>>>From: Tobias Predel [mailto:tobias.pre...@gmail.com]
> >>>>>>>Sent: Friday, February 19, 2016 09:23 AM
> >>>>>>>To: Adrian Chadd
> >>>>>>>Cc: Hsu, Ryan; ath10k@lists.infradead.org; vagner_
> >>>>>>>Subject: Re: ath10k and QCA3977 - random transfer rate losses
> >>>>>>>
> >>>>>>>On Fri, Feb 19, 2016 at 08:07:49AM -0800, Adrian Chadd wrote:
> >>>>>>>>Ok, so it's a board file problem? this can be solved by telling kvalo@
> >>>>>>>>about it, right?
> >>>>>>>>
> >>>>>>>>
> >>>>>>>>-a
> >>>>>>>Yes, they differ and the board-2.bin file isn't included neither in 
> >>>>>>>linux-
> >>>>>>>firmware nor in kvalo's repository.
> >>>>>>>
> >>>>>>>MD5 checksums:
> >>>>>>>73cca12b9398f0cecaa3cfa35c8961b7  board-2.bin
> >>>>>>>44349f6ae93cf9ceffb164f58d4fb124  board.bin
> >>>>>>>
> >>>>>>>I'll try to contact him and send the board file.
> >>>>>>>
> >>>>>>>Tobias
> 

___
ath10k mailing list
ath10k@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/ath10k

Re: ath10k and QCA3977 - random transfer rate losses

[Tobias Predel]

Hello Adrian,

bug report can be found here 
<https://bugzilla.kernel.org/show_bug.cgi?id=112451#c10>
I added other firmware files. I don't know any other place.

Tobias

On Sun, Feb 21, 2016 at 11:25:52AM -0800, Adrian Chadd wrote:
> ok, well, we should file a bug report so this isn't forgotten.
> 
> Where's the right place to create bug reports? is there an ath10k
> staging branch in github somewhere? or bugzilla.kernel.org ?
> 
> 
> -a
> 
> 
> On 21 February 2016 at 11:20, Tobias Predel <tobias.pre...@gmail.com> wrote:
> > Yes. I confirm full functionality now.
> >
> > On Sun, Feb 21, 2016 at 10:33:26AM -0800, Adrian Chadd wrote:
> >> Right, so just to be clear - it's just the board file updates that
> >> make it all work fine?
> >>
> >>
> >> -a
> >>
> >>
> >> On 21 February 2016 at 06:44, Tobias Predel <tobias.pre...@gmail.com> 
> >> wrote:
> >> > Hello again,
> >> >
> >> > one last word about the the board files:
> >> >
> >> > I extracted the board files from an AsusTek Windows Driver installer
> >> > that included the Qualcomm driver files for diverse variants of QCA9377
> >> > built in Samsung, Asus (Askey), Acer, Lenovo notebooks.
> >> >
> >> > An archive with all these files can be found at the comment 10 of my bug 
> >> > report
> >> > <https://bugzilla.kernel.org/show_bug.cgi?id=112451#c10>
> >> > including board files (eeprom_*.bin) for NFA425, NFA435 (A), NFA455.
> >> >
> >> > If some people experience trouble, those might look at the netathr10x.inf
> >> > to find out which eeprom board file is needed or try out by installing
> >> > it to /lib/firmware/ath10k/QCA9377/hw1.0/board.bin.
> >> >
> >> > Those board files are using api 1 and must be named board.bin.
> >> > firmware-5.bin is still to be pulled from kvalo's github repo.
> >> >
> >> > Thank you to all for your support. I wish the best!
> >> >
> >> > Tobias
> >> >
> >> >
> >> > On Fri, Feb 19, 2016 at 05:55:26PM +, Hsu, Ryan wrote:
> >> >> Glad that help!
> >> >>
> >> >> I've cc that file to Kalle, hope we can see that get merged soon.
> >> >>
> >> >> Ryan
> >> >>
> >> >> > -Original Message-
> >> >> > From: Tobias Predel [mailto:tobias.pre...@gmail.com]
> >> >> > Sent: Friday, February 19, 2016 09:23 AM
> >> >> > To: Adrian Chadd
> >> >> > Cc: Hsu, Ryan; ath10k@lists.infradead.org; vagner_
> >> >> > Subject: Re: ath10k and QCA3977 - random transfer rate losses
> >> >> >
> >> >> > On Fri, Feb 19, 2016 at 08:07:49AM -0800, Adrian Chadd wrote:
> >> >> > > Ok, so it's a board file problem? this can be solved by telling 
> >> >> > > kvalo@
> >> >> > > about it, right?
> >> >> > >
> >> >> > >
> >> >> > > -a
> >> >> >
> >> >> > Yes, they differ and the board-2.bin file isn't included neither in 
> >> >> > linux-
> >> >> > firmware nor in kvalo's repository.
> >> >> >
> >> >> > MD5 checksums:
> >> >> > 73cca12b9398f0cecaa3cfa35c8961b7  board-2.bin
> >> >> > 44349f6ae93cf9ceffb164f58d4fb124  board.bin
> >> >> >
> >> >> > I'll try to contact him and send the board file.
> >> >> >
> >> >> > Tobias

___
ath10k mailing list
ath10k@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/ath10k

Re: ath10k and QCA3977 - random transfer rate losses

[Tobias Predel]

Hello again,

one last word about the the board files:

I extracted the board files from an AsusTek Windows Driver installer
that included the Qualcomm driver files for diverse variants of QCA9377
built in Samsung, Asus (Askey), Acer, Lenovo notebooks.

An archive with all these files can be found at the comment 10 of my bug report
<https://bugzilla.kernel.org/show_bug.cgi?id=112451#c10>
including board files (eeprom_*.bin) for NFA425, NFA435 (A), NFA455.

If some people experience trouble, those might look at the netathr10x.inf 
to find out which eeprom board file is needed or try out by installing
it to /lib/firmware/ath10k/QCA9377/hw1.0/board.bin. 

Those board files are using api 1 and must be named board.bin.
firmware-5.bin is still to be pulled from kvalo's github repo.

Thank you to all for your support. I wish the best!

Tobias


On Fri, Feb 19, 2016 at 05:55:26PM +, Hsu, Ryan wrote:
> Glad that help!
> 
> I've cc that file to Kalle, hope we can see that get merged soon.
> 
> Ryan
> 
> > -Original Message-
> > From: Tobias Predel [mailto:tobias.pre...@gmail.com]
> > Sent: Friday, February 19, 2016 09:23 AM
> > To: Adrian Chadd
> > Cc: Hsu, Ryan; ath10k@lists.infradead.org; vagner_
> > Subject: Re: ath10k and QCA3977 - random transfer rate losses
> > 
> > On Fri, Feb 19, 2016 at 08:07:49AM -0800, Adrian Chadd wrote:
> > > Ok, so it's a board file problem? this can be solved by telling kvalo@
> > > about it, right?
> > >
> > >
> > > -a
> > 
> > Yes, they differ and the board-2.bin file isn't included neither in linux-
> > firmware nor in kvalo's repository.
> > 
> > MD5 checksums:
> > 73cca12b9398f0cecaa3cfa35c8961b7  board-2.bin
> > 44349f6ae93cf9ceffb164f58d4fb124  board.bin
> > 
> > I'll try to contact him and send the board file.
> > 
> > Tobias

___
ath10k mailing list
ath10k@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/ath10k

Re: ath10k and QCA3977 - random transfer rate losses

[Tobias Predel]

On Fri, Feb 19, 2016 at 08:07:49AM -0800, Adrian Chadd wrote:
> Ok, so it's a board file problem? this can be solved by telling kvalo@
> about it, right?
> 
> 
> -a

Yes, they differ and the board-2.bin file isn't included neither in 
linux-firmware nor in kvalo's repository. 

MD5 checksums:
73cca12b9398f0cecaa3cfa35c8961b7  board-2.bin
44349f6ae93cf9ceffb164f58d4fb124  board.bin

I'll try to contact him and send the board file. 

Tobias

___
ath10k mailing list
ath10k@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/ath10k

Re: ath10k and QCA3977 - random transfer rate losses

[Tobias Predel]

Hello Ryan,

thank you very much for providing the correct board file. :)
Now it works without problem!

Inside the /lib/firmware/ath10k/QCA9377/hw1.0 folder are now:
- firmware-5.bin from kvalo's github repository
- board-2.bin you provided

The dmesg and lspci output are attached.

Tobias

--

[3.321927] ath10k_pci :02:00.0: pci irq msi interrupts 1 irq_mode 0 
reset_mode 0
[3.552539] ath10k_pci :02:00.0: Direct firmware load for 
ath10k/cal-pci-:02:00.0.bin failed with error -2
[3.564800] ath10k_pci :02:00.0: qca9377 hw1.0 target 0x0502 chip_id 
0x003820ff sub 1a3b:2231
[3.564804] ath10k_pci :02:00.0: kconfig debug 1 debugfs 1 tracing 0 dfs 
0 testmode 0
[3.565451] ath10k_pci :02:00.0: firmware ver WLAN.TF.1.0-00267-1 api 5 
features ignore-otp crc32 79cea2c7
[3.635299] ath10k_pci :02:00.0: board_file api 2 bmi_id N/A crc32 
93da0176
[5.444342] ath10k_pci :02:00.0: htt-ver 3.1 wmi-op 4 htt-op 3 cal otp 
max-sta 32 raw 0 hwcrypto 1
[5.446063] ath: EEPROM regdomain: 0x6a
[5.446068] ath: EEPROM indicates we should expect a direct regpair map
[5.446071] ath: Country alpha2 being used: 00
[5.446073] ath: Regpair used: 0x6a
[5.461059] ath10k_pci :02:00.0 wlp2s0: renamed from wlan0


02:00.0 Network controller: Qualcomm Atheros Device 0042 (rev 30)
Subsystem: AzureWave Device 2231
Control: I/O- Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- 
Stepping- SERR- FastB2B- DisINTx+
Status: Cap+ 66MHz- UDF- FastB2B- ParErr- DEVSEL=fast >TAbort- SERR-  Tobias,
> 
> It seems to be like the RF tx/rx capability issue, usually due to the 
> incorrect board file is using.
> 
> I can't find exactly the same module you've to try.
> Can you help to verify this attached board-2.bin file to see if that help the 
> issue you've?
> 
> cp board-2.bin.9377 /lib/firmware/ath10k/QCA9377/hw1.0/board-2.bin
> 
> If this is not helping, would be helpful if you can share the commands and 
> scenario you're running.
> Also the dump of the kernel logs.
> 
> Ryan
> 
> > -Original Message-
> > From: ath10k [mailto:ath10k-boun...@lists.infradead.org] On Behalf Of
> > Tobias Predel
> > Sent: Sunday, February 14, 2016 01:40 PM
> > To: vagner_
> > Cc: ath10k@lists.infradead.org
> > Subject: Re: ath10k and QCA3977 - random transfer rate losses
> > 
> > Hi Vagner,
> > 
> > turned out that it didn't work well over the long term. Issues appear again.
> > 
> > I gave up trying workarounds and decided to file an official bug report
> > (https://bugzilla.kernel.org/show_bug.cgi?id=112451) and keep quiet,
> > patiently waiting for a bug fix (I'm not a programmer) to prevent this from
> > becoming too frustating.
> > 
> > Thanks for your help anyway!
> > 
> > Best wishes
> > Tobias
> > 
> > On Sun, Feb 14, 2016 at 08:02:23PM +0200, vagner_ wrote:
> > > Great,
> > >
> > > You can write to him < prof.ajaybha...@gmail.com >  saying thx. So
> > > that, he can get motivated and improve the drive...
> > > Good luck,
> > >
> > > Vagner
> > >
> > > +++
> > >
> > >
> > > On 02/14/2016 05:46 PM, Tobias Predel wrote:
> > > >Hi Vagner,
> > > >
> > > >thanks for your quick reply! I tried out your proposal and together
> > > >with power management switched off (iwconfig wlp2s0 power off) the
> > chipset works like a charm!
> > > >
> > > >A big thank you to you and the professor from India! You made my day
> > > >:)
> > > >
> > > >I just keep wondering now whether the disabled power management will
> > > >increase the battery drain but maybe the firmware does handle those
> > > >things by itself...
> > > >
> > > >So have a nice day and cheers,
> > > >Tobias
> > > >
> > > >On Sat, Feb 13, 2016 at 02:06:32PM +0200, vagner_ wrote:
> > > >>Hi Tobias,
> > > >>
> > > >>I also bought a laptop some weeks ago with an atheros chipset that
> > > >>required
> > > >>qca9377 - exactly like yours
> > > >>
> > > >>Network controller: Qualcomm Atheros Device 0042 (rev 30) -
> > > >>
> > > >>and it took me 2 or 3 weeks trying all kinds of solutions, including
> > > >>compiling a new kernel, kvalo drives and others. None of them worked.
> > > >>Finally, I found a drive that a professor from India, I think, wrote
> > > >>then I installed a new kernel 4.0 and the

Re: ath10k and QCA3977 - random transfer rate losses

[Tobias Predel]

Hi Vagner,

turned out that it didn't work well over the long term. Issues appear again.

I gave up trying workarounds and decided to file an official bug report 
(https://bugzilla.kernel.org/show_bug.cgi?id=112451) and keep quiet, 
patiently waiting for a bug fix (I'm not a programmer) to prevent this 
from becoming too frustating. 

Thanks for your help anyway!

Best wishes
Tobias

On Sun, Feb 14, 2016 at 08:02:23PM +0200, vagner_ wrote:
> Great,
> 
> You can write to him < prof.ajaybha...@gmail.com >  saying thx. So that, he
> can get motivated and improve the drive...
> Good luck,
> 
> Vagner
> 
> +++
> 
> 
> On 02/14/2016 05:46 PM, Tobias Predel wrote:
> >Hi Vagner,
> >
> >thanks for your quick reply! I tried out your proposal and together with 
> >power
> >management switched off (iwconfig wlp2s0 power off) the chipset works like a 
> >charm!
> >
> >A big thank you to you and the professor from India! You made my day :)
> >
> >I just keep wondering now whether the disabled power management
> >will increase the battery drain but maybe the firmware does handle
> >those things by itself...
> >
> >So have a nice day and cheers,
> >Tobias
> >
> >On Sat, Feb 13, 2016 at 02:06:32PM +0200, vagner_ wrote:
> >>Hi Tobias,
> >>
> >>I also bought a laptop some weeks ago with an atheros chipset that required
> >>qca9377 - exactly like yours
> >>
> >>Network controller: Qualcomm Atheros Device 0042 (rev 30) -
> >>
> >>and it took me 2 or 3 weeks trying all kinds of solutions, including
> >>compiling a new kernel, kvalo drives and others. None of them worked.
> >>Finally, I found a drive that a professor from India, I think, wrote then I
> >>installed a new kernel 4.0 and the drive and since then it has been working
> >>marvelously.
> >>I even wrote to him to say thanks and got a good reply.
> >>I run Siduction distro with their very updated kernel.
> >>
> >>Try this:
> >>
> >>Remove your drive and install this one. It is very easy. Just run the
> >>script. Then tell us if that worked for you.
> >>Cheers and good luck...
> >>
> >>https://github.com/ajaybhatia/Qualcomm-Atheros-QCA9377-Wifi-Linux
> >>
> >>Vagner
> >>
> >>+++
> >>
> >>On 02/12/2016 10:30 PM, Tobias Predel wrote:
> >>>Hello,
> >>>
> >>>I bought a new laptop from ASUS, Product Name E205SA and installed Arch 
> >>>Linux. Unfortunately, the included WLAN chipset from Qualcomm  (QCA 9377) 
> >>>doesn't seem to work well with Linux. Beside of being recognized, the 
> >>>driver/device functions normally for the first minutes after boot, but 
> >>>then the transfer rate is suddenly lowering from around 700 kbs to 200 kbs 
> >>>and less, making Internet unusable. After a while, it's recovering a bit 
> >>>but it seems to be random (like it's trying to sleep/power save mode?). 
> >>>This behavior occures for my home network as well as at my university's 
> >>>network. As wifi access wasn't a issue with the Windows 10 OEM 
> >>>installation, I guess that this might a driver/firmware/configuration 
> >>>problem.
> >>>
> >>>dmesg indicates
> >>>[   13.401917] ath: EEPROM regdomain: 0x6a
> >>>[   13.403870] ath: EEPROM indicates we should expect a direct regpair map
> >>>[   13.405919] ath: Country alpha2 being used: 00
> >>>[   13.407837] ath: Regpair used: 0x6a
> >>>and
> >>>[   11.267403] ath10k_pci :02:00.0: pci irq msi interrupts 1 irq_mode 
> >>>0 reset_mode 0
> >>>[   11.500220] ath10k_pci :02:00.0: Direct firmware load for 
> >>>ath10k/cal-pci-:02:00.0.bin failed with error -2
> >>>[   11.583900] ath10k_pci :02:00.0: Direct firmware load for 
> >>>ath10k/QCA9377/hw1.0/board-2.bin failed with error -2
> >>>[   13.394121] ath10k_pci :02:00.0: qca9377 hw1.0 (0x0502, 
> >>>0x003820ff sub 1a3b:2231) fw WLAN.TF.1.0-00267-1 fwapi 5 bdapi 1 htt-ver 
> >>>3.1 wmi-op 4 htt-op 3 cal otp max-sta 32 raw 0 hwcrypto 1 features 
> >>>ignore-otp
> >>>[   13.398452] ath10k_pci :02:00.0: debug 1 debugfs 1 tracing 0 dfs 0 
> >>>testmode 0
> >>>[   13.422666] ath10k_pci :02:00.0 wlp2s0: renamed from wlan0
> >>>
> >>>The firmware files were downloaded from the git repository 
> &

Re: ath10k and QCA3977 - random transfer rate losses

[Tobias Predel]

No not all. Same firmware files that can be found in kvalo's git repository.


On Sun, Feb 14, 2016 at 08:35:04AM -0800, Adrian Chadd wrote:
> Does it have updated firmware or something?
> 
> 
> -a
> 

___
ath10k mailing list
ath10k@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/ath10k

Re: ath10k and QCA3977 - random transfer rate losses

[Tobias Predel]

Hi Vagner,

thanks for your quick reply! I tried out your proposal and together with power 
management switched off (iwconfig wlp2s0 power off) the chipset works like a 
charm! 

A big thank you to you and the professor from India! You made my day :)

I just keep wondering now whether the disabled power management
will increase the battery drain but maybe the firmware does handle
those things by itself...

So have a nice day and cheers,
Tobias

On Sat, Feb 13, 2016 at 02:06:32PM +0200, vagner_ wrote:
> Hi Tobias,
> 
> I also bought a laptop some weeks ago with an atheros chipset that required
> qca9377 - exactly like yours
> 
> Network controller: Qualcomm Atheros Device 0042 (rev 30) -
> 
> and it took me 2 or 3 weeks trying all kinds of solutions, including
> compiling a new kernel, kvalo drives and others. None of them worked.
> Finally, I found a drive that a professor from India, I think, wrote then I
> installed a new kernel 4.0 and the drive and since then it has been working
> marvelously.
> I even wrote to him to say thanks and got a good reply.
> I run Siduction distro with their very updated kernel.
> 
> Try this:
> 
> Remove your drive and install this one. It is very easy. Just run the
> script. Then tell us if that worked for you.
> Cheers and good luck...
> 
> https://github.com/ajaybhatia/Qualcomm-Atheros-QCA9377-Wifi-Linux
> 
> Vagner
> 
> +++
> 
> On 02/12/2016 10:30 PM, Tobias Predel wrote:
> >Hello,
> >
> >I bought a new laptop from ASUS, Product Name E205SA and installed Arch 
> >Linux. Unfortunately, the included WLAN chipset from Qualcomm  (QCA 9377) 
> >doesn't seem to work well with Linux. Beside of being recognized, the 
> >driver/device functions normally for the first minutes after boot, but then 
> >the transfer rate is suddenly lowering from around 700 kbs to 200 kbs and 
> >less, making Internet unusable. After a while, it's recovering a bit but it 
> >seems to be random (like it's trying to sleep/power save mode?). This 
> >behavior occures for my home network as well as at my university's network. 
> >As wifi access wasn't a issue with the Windows 10 OEM installation, I guess 
> >that this might a driver/firmware/configuration problem.
> >
> >dmesg indicates
> >[   13.401917] ath: EEPROM regdomain: 0x6a
> >[   13.403870] ath: EEPROM indicates we should expect a direct regpair map
> >[   13.405919] ath: Country alpha2 being used: 00
> >[   13.407837] ath: Regpair used: 0x6a
> >and
> >[   11.267403] ath10k_pci :02:00.0: pci irq msi interrupts 1 irq_mode 0 
> >reset_mode 0
> >[   11.500220] ath10k_pci :02:00.0: Direct firmware load for 
> >ath10k/cal-pci-:02:00.0.bin failed with error -2
> >[   11.583900] ath10k_pci :02:00.0: Direct firmware load for 
> >ath10k/QCA9377/hw1.0/board-2.bin failed with error -2
> >[   13.394121] ath10k_pci :02:00.0: qca9377 hw1.0 (0x0502, 
> >0x003820ff sub 1a3b:2231) fw WLAN.TF.1.0-00267-1 fwapi 5 bdapi 1 htt-ver 3.1 
> >wmi-op 4 htt-op 3 cal otp max-sta 32 raw 0 hwcrypto 1 features ignore-otp
> >[   13.398452] ath10k_pci :02:00.0: debug 1 debugfs 1 tracing 0 dfs 0 
> >testmode 0
> >[   13.422666] ath10k_pci :02:00.0 wlp2s0: renamed from wlan0
> >
> >The firmware files were downloaded from the git repository 
> ><https://github.com/kvalo/ath10k-firmware>. In addition to that iwconfig 
> >gives me a high number of invalid miscs. I use kernel 4.4.1 (Arch Linux). 
> >ethtools -i wlp2s0 prints out
> >
> >driver: ath10k_pci
> >version: 4.4.1-2-ARCH
> >firmware-version: WLAN.TF.1.0-00267-1
> >expansion-rom-version:
> >bus-info: :02:00.0
> >supports-statistics: yes
> >supports-test: no
> >supports-eeprom-access: no
> >supports-register-dump: no
> >supports-priv-flags: no
> >
> >and lspci provides
> >
> >02:00.0 Network controller: Qualcomm Atheros Device 0042 (rev 30)
> > Subsystem: AzureWave Device 2231
> > Control: I/O- Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- 
> > Stepping- SERR- FastB2B- DisINTx+
> > Status: Cap+ 66MHz- UDF- FastB2B- ParErr- DEVSEL=fast >TAbort-  > SERR-  > Latency: 0, Cache Line Size: 64 bytes
> > Interrupt: pin A routed to IRQ 315
> > Region 0: Memory at 8100 (64-bit, non-prefetchable) [size=2M]
> > Capabilities: [40] Power Management version 3
> > Flags: PMEClk- DSI- D1- D2- AuxCurrent=375mA 
> > PME(D0+,D1-,D2-,D3hot+,D3cold+)
> > Status: D0 NoSoftRst- PME-Enable- DSel=0 DScale=0 PME-
> > Capabilities: [50] MSI: Enable+

Re: ath10k and QCA3977 - random transfer rate losses

[Tobias Predel]

Hello,

I just looked into it and it seems to be nothing special.

A recursive diff just showed me that the author just placed some additional 
defconfig files
that don't exist in the backport tree from 20151218. The defconfig-ath10k just 
switches debug on.

Sorry for the confusion.

Tobias

On Sun, Feb 14, 2016 at 08:00:46AM -0800, Adrian Chadd wrote:
> ... what has the professor done that isn't yet in ath10k?
> 
> 
> -a
> 
> 
> On 14 February 2016 at 07:46, Tobias Predel <tobias.pre...@gmail.com> wrote:
> > Hi Vagner,
> >
> > thanks for your quick reply! I tried out your proposal and together with 
> > power
> > management switched off (iwconfig wlp2s0 power off) the chipset works like 
> > a charm!
> >
> > A big thank you to you and the professor from India! You made my day :)
> >
> > I just keep wondering now whether the disabled power management
> > will increase the battery drain but maybe the firmware does handle
> > those things by itself...
> >
> > So have a nice day and cheers,
> > Tobias
> >
> > On Sat, Feb 13, 2016 at 02:06:32PM +0200, vagner_ wrote:
> >> Hi Tobias,
> >>
> >> I also bought a laptop some weeks ago with an atheros chipset that required
> >> qca9377 - exactly like yours
> >>
> >> Network controller: Qualcomm Atheros Device 0042 (rev 30) -
> >>
> >> and it took me 2 or 3 weeks trying all kinds of solutions, including
> >> compiling a new kernel, kvalo drives and others. None of them worked.
> >> Finally, I found a drive that a professor from India, I think, wrote then I
> >> installed a new kernel 4.0 and the drive and since then it has been working
> >> marvelously.
> >> I even wrote to him to say thanks and got a good reply.
> >> I run Siduction distro with their very updated kernel.
> >>
> >> Try this:
> >>
> >> Remove your drive and install this one. It is very easy. Just run the
> >> script. Then tell us if that worked for you.
> >> Cheers and good luck...
> >>
> >> https://github.com/ajaybhatia/Qualcomm-Atheros-QCA9377-Wifi-Linux
> >>
> >> Vagner
> >>
> >> +++
> >>
> >> On 02/12/2016 10:30 PM, Tobias Predel wrote:
> >> >Hello,
> >> >
> >> >I bought a new laptop from ASUS, Product Name E205SA and installed Arch 
> >> >Linux. Unfortunately, the included WLAN chipset from Qualcomm  (QCA 9377) 
> >> >doesn't seem to work well with Linux. Beside of being recognized, the 
> >> >driver/device functions normally for the first minutes after boot, but 
> >> >then the transfer rate is suddenly lowering from around 700 kbs to 200 
> >> >kbs and less, making Internet unusable. After a while, it's recovering a 
> >> >bit but it seems to be random (like it's trying to sleep/power save 
> >> >mode?). This behavior occures for my home network as well as at my 
> >> >university's network. As wifi access wasn't a issue with the Windows 10 
> >> >OEM installation, I guess that this might a driver/firmware/configuration 
> >> >problem.
> >> >
> >> >dmesg indicates
> >> >[   13.401917] ath: EEPROM regdomain: 0x6a
> >> >[   13.403870] ath: EEPROM indicates we should expect a direct regpair map
> >> >[   13.405919] ath: Country alpha2 being used: 00
> >> >[   13.407837] ath: Regpair used: 0x6a
> >> >and
> >> >[   11.267403] ath10k_pci :02:00.0: pci irq msi interrupts 1 irq_mode 
> >> >0 reset_mode 0
> >> >[   11.500220] ath10k_pci :02:00.0: Direct firmware load for 
> >> >ath10k/cal-pci-:02:00.0.bin failed with error -2
> >> >[   11.583900] ath10k_pci :02:00.0: Direct firmware load for 
> >> >ath10k/QCA9377/hw1.0/board-2.bin failed with error -2
> >> >[   13.394121] ath10k_pci :02:00.0: qca9377 hw1.0 (0x0502, 
> >> >0x003820ff sub 1a3b:2231) fw WLAN.TF.1.0-00267-1 fwapi 5 bdapi 1 htt-ver 
> >> >3.1 wmi-op 4 htt-op 3 cal otp max-sta 32 raw 0 hwcrypto 1 features 
> >> >ignore-otp
> >> >[   13.398452] ath10k_pci :02:00.0: debug 1 debugfs 1 tracing 0 dfs 0 
> >> >testmode 0
> >> >[   13.422666] ath10k_pci :02:00.0 wlp2s0: renamed from wlan0
> >> >
> >> >The firmware files were downloaded from the git repository 
> >> ><https://github.com/kvalo/ath10k-firmware>. In addition to that iwconfig 
> >> >give

ath10k and QCA3977 - random transfer rate losses

[Tobias Predel]

Hello,

I bought a new laptop from ASUS, Product Name E205SA and installed Arch Linux. 
Unfortunately, the included WLAN chipset from Qualcomm  (QCA 9377) doesn't seem 
to work well with Linux. Beside of being recognized, the driver/device 
functions normally for the first minutes after boot, but then the transfer rate 
is suddenly lowering from around 700 kbs to 200 kbs and less, making Internet 
unusable. After a while, it's recovering a bit but it seems to be random (like 
it's trying to sleep/power save mode?). This behavior occures for my home 
network as well as at my university's network. As wifi access wasn't a issue 
with the Windows 10 OEM installation, I guess that this might a 
driver/firmware/configuration problem. 

dmesg indicates
[   13.401917] ath: EEPROM regdomain: 0x6a
[   13.403870] ath: EEPROM indicates we should expect a direct regpair map
[   13.405919] ath: Country alpha2 being used: 00
[   13.407837] ath: Regpair used: 0x6a
and
[   11.267403] ath10k_pci :02:00.0: pci irq msi interrupts 1 irq_mode 0 
reset_mode 0
[   11.500220] ath10k_pci :02:00.0: Direct firmware load for 
ath10k/cal-pci-:02:00.0.bin failed with error -2
[   11.583900] ath10k_pci :02:00.0: Direct firmware load for 
ath10k/QCA9377/hw1.0/board-2.bin failed with error -2
[   13.394121] ath10k_pci :02:00.0: qca9377 hw1.0 (0x0502, 0x003820ff 
sub 1a3b:2231) fw WLAN.TF.1.0-00267-1 fwapi 5 bdapi 1 htt-ver 3.1 wmi-op 4 
htt-op 3 cal otp max-sta 32 raw 0 hwcrypto 1 features ignore-otp
[   13.398452] ath10k_pci :02:00.0: debug 1 debugfs 1 tracing 0 dfs 0 
testmode 0
[   13.422666] ath10k_pci :02:00.0 wlp2s0: renamed from wlan0

The firmware files were downloaded from the git repository 
. In addition to that iwconfig gives 
me a high number of invalid miscs. I use kernel 4.4.1 (Arch Linux). ethtools -i 
wlp2s0 prints out 

driver: ath10k_pci
version: 4.4.1-2-ARCH
firmware-version: WLAN.TF.1.0-00267-1
expansion-rom-version: 
bus-info: :02:00.0
supports-statistics: yes
supports-test: no
supports-eeprom-access: no
supports-register-dump: no
supports-priv-flags: no

and lspci provides

02:00.0 Network controller: Qualcomm Atheros Device 0042 (rev 30)
Subsystem: AzureWave Device 2231
Control: I/O- Mem+ BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- 
Stepping- SERR- FastB2B- DisINTx+
Status: Cap+ 66MHz- UDF- FastB2B- ParErr- DEVSEL=fast >TAbort- SERR- http://lists.infradead.org/mailman/listinfo/ath10k