On Tue, 13 Feb 2018 12:00:29 +0100, Jan-Marek Glogowski wrote:
Hi Takashi,
Am 13.02.2018 um 07:04 schrieb Takashi Iwai:
On Fri, 26 Jan 2018 16:12:53 +0100, Jan-Marek Glogowski wrote:
I got new HW from our supplier, a Fujitsu U727 and a U757 laptop, which have a headset jack. Attaching a headset, the microphone doesn't work. After some reading and trying various tools to understand the problem, I found hdajacksensetest, which showed a sense change for a seemingly wrong unconnected pin. Overriding this pins' config has fixed the headsets microphone, so now I have the attached kernel patch with a SND_HDA_PIN_QUIRK.
The pin config override itself looks OK, but the way to identify the machine is an open question. You've added the entries in alc269_pin_fixup_tbl[]. This table is for matching with a group of machines that have the same "fingerprint" by pin configs. Usually this is used for Dell laptops and such.
OTOH, if one quirk is needed only for one or two machines, using the PCI SSID (or codec SSID) is the easier way to go, and it's done by putting into alc269_fixup_tbl[].
So I re-check the hardware:
$ cat /proc/asound/card0/codec#0 | head Codec: Realtek ALC255 Address: 0 AFG Function Id: 0x1 (unsol 1) Vendor Id: 0x10ec0255 Subsystem Id: 0x10cf1629 Revision Id: 0x100002 No Modem Function Group found Default PCM: rates [0x560]: 44100 48000 96000 192000 bits [0xe]: 16 20 24
$ sudo lspci -vns 00:1f.3 00:1f.3 0403: 8086:9d70 (rev 21) Subsystem: 10cf:193e Flags: bus master, fast devsel, latency 32, IRQ 280 Memory at c1240000 (64-bit, non-prefetchable) [size=16K] Memory at c1230000 (64-bit, non-prefetchable) [size=64K] Capabilities: [50] Power Management version 3 Capabilities: [60] MSI: Enable+ Count=1/1 Maskable- 64bit+ Kernel driver in use: snd_hda_intel
I ran and diff'ed the output on both laptops. It's the same for both notebook types, so AFAIK there is no other way to distinguish them, except for the Jack pin being on the left or right side. The only difference I see is in the DMI datas' product name, but that data is not used in the sound driver at all.
$ sudo dmidecode -t 1 Handle 0x0001, DMI type 1, 27 bytes System Information Manufacturer: FUJITSU Product Name: LIFEBOOK U757 Family: LIFEBOOK-FTS
The driver doesn't check any difference of the position for identifying the jack detection capability, so you can apply the same quirk to both models. Then I prefer putting with SND_PCI_QUIRK().
After compiling the patched module (for 4.4), everything works as expected, but hdajackretask still shows the pin as unconnected, so I'm wondering, if I need to do something in addition?
If everything is working, why any more changes? :)
Ok. I'm not familiar with the HDA codecs, and I was just wondering, if this should / can be patched.
Is this normal behaviour?
Jan-Marek
P.S. checkpatch.pl warned about too long lines for the SND_HDA_PIN_QUIRK. Should I fix that?
No, it's not strictly needed. It's a warning, not an error.
thanks,
Takashi