[alsa-devel] Regression in kernel 3.19-rc5 for usb2 scarlett 18i20
Hi, I noticed a regression in 3.19-rc5 kernel (as shipped by fedora rawhide) with usb2 focusrite scarlett 18i20. I get a lot of xruns with jack, even at 1024 frame buffer and 2 periods. In 3.17.8 it works quite well without xruns at the same settings.
I use a fedora 20 (64bit) laptop with 8GB ram, I only installed the rawhide kernel to try the new included scarlett mixer, no other software components changed.
I can try to do some other test, but compiling the kernel on this laptop is time consuming.
Thanks Guido
At Sun, 25 Jan 2015 12:42:22 +0100, Guido Aulisi wrote:
Hi, I noticed a regression in 3.19-rc5 kernel (as shipped by fedora rawhide) with usb2 focusrite scarlett 18i20. I get a lot of xruns with jack, even at 1024 frame buffer and 2 periods. In 3.17.8 it works quite well without xruns at the same settings.
I use a fedora 20 (64bit) laptop with 8GB ram, I only installed the rawhide kernel to try the new included scarlett mixer, no other software components changed.
I can try to do some other test, but compiling the kernel on this laptop is time consuming.
Thanks Guido
Looks like a regression in the recent rework for Scarlett devices. Chris, do you have any clue?
In anyway, Guido, could you give lsusb -v output (for the relevant part)?
thanks,
Takashi
I will ASAP, now I'm not at home. IMHO I think this is something about usb than the mixer, because I didn't use it, I only noticed the xruns even without playing audio. I didn't try 3.18, because it isn't shipped with fedora yet.
Ciao
2015-01-26 14:04 GMT+01:00 Takashi Iwai tiwai@suse.de:
At Sun, 25 Jan 2015 12:42:22 +0100, Guido Aulisi wrote:
Hi, I noticed a regression in 3.19-rc5 kernel (as shipped by fedora rawhide) with usb2 focusrite scarlett 18i20. I get a lot of xruns with jack, even at 1024 frame buffer and 2 periods. In 3.17.8 it works quite well without xruns at the same settings.
I use a fedora 20 (64bit) laptop with 8GB ram, I only installed the rawhide kernel to try the new included scarlett mixer, no other software components changed.
I can try to do some other test, but compiling the kernel on this laptop is time consuming.
Thanks Guido
Looks like a regression in the recent rework for Scarlett devices. Chris, do you have any clue?
In anyway, Guido, could you give lsusb -v output (for the relevant part)?
thanks,
Takashi
At Mon, 26 Jan 2015 14:28:14 +0100, Guido Aulisi wrote:
I will ASAP, now I'm not at home. IMHO I think this is something about usb than the mixer, because I didn't use it, I only noticed the xruns even without playing audio. I didn't try 3.18, because it isn't shipped with fedora yet.
Well, then we might need the git bisection, I'm afraid...
thanks,
Takashi
Ciao
2015-01-26 14:04 GMT+01:00 Takashi Iwai tiwai@suse.de:
At Sun, 25 Jan 2015 12:42:22 +0100, Guido Aulisi wrote:
Hi, I noticed a regression in 3.19-rc5 kernel (as shipped by fedora rawhide) with usb2 focusrite scarlett 18i20. I get a lot of xruns with jack, even at 1024 frame buffer and 2 periods. In 3.17.8 it works quite well without xruns at the same settings.
I use a fedora 20 (64bit) laptop with 8GB ram, I only installed the rawhide kernel to try the new included scarlett mixer, no other software components changed.
I can try to do some other test, but compiling the kernel on this laptop is time consuming.
Thanks Guido
Looks like a regression in the recent rework for Scarlett devices. Chris, do you have any clue?
In anyway, Guido, could you give lsusb -v output (for the relevant part)?
thanks,
Takashi
On 01/26/2015 07:28 AM, Guido Aulisi wrote:
I will ASAP, now I'm not at home. IMHO I think this is something about usb than the mixer, because I didn't use it, I only noticed the xruns even without playing audio. I didn't try 3.18, because it isn't shipped with fedora yet.
Ciao
2015-01-26 14:04 GMT+01:00 Takashi Iwai tiwai@suse.de:
At Sun, 25 Jan 2015 12:42:22 +0100, Guido Aulisi wrote:
Hi, I noticed a regression in 3.19-rc5 kernel (as shipped by fedora rawhide) with usb2 focusrite scarlett 18i20. I get a lot of xruns with jack, even at 1024 frame buffer and 2 periods. In 3.17.8 it works quite well without xruns at the same settings.
Guido, Takashi,
I tried to reproduce this and failed on my Scarlett 18i8 (I don't have any other hardware).
I've just tested Ubuntu's 3.18.2 and daily mainline 3.19-r5 as follows: 1) Setup Scarlett 18i8 w/ jack (RT, 1024 frames/period, sample rate 48k, periods/buffer 2). 2) Launch Ardour3 and create/arm 8 tracks 3) Hit record
Using this I was unable to see any XRuns even when using Scarlett's mixer with alsamixer in the background on 3.19.
Let me know more details. Thanks, --chris j arges
I use a fedora 20 (64bit) laptop with 8GB ram, I only installed the rawhide kernel to try the new included scarlett mixer, no other software components changed.
I can try to do some other test, but compiling the kernel on this laptop is time consuming.
Thanks Guido
Looks like a regression in the recent rework for Scarlett devices. Chris, do you have any clue?
In anyway, Guido, could you give lsusb -v output (for the relevant part)?
thanks,
Takashi
I tried 3.19-rc6 and it seems to work well... I compiled it myself using fedora config from http://pkgs.fedoraproject.org/cgit/kernel.git/
Also the mixer is present and I could switch headphones on and off (great thing). I still have to understand most of the controls... Thank you for the support. If I have some time the next days, I'll try to bisect from rc5 to rc6, to see if some commit fixed this issue, or if more probably I did something wrong. I always used the same usb port on my laptop, these are some info about it:
Linux 3.19.0-rc6 #1 SMP Mon Jan 26 22:21:16 CET 2015 x86_64 x86_64 x86_64 GNU/Linux
Bus 002 Device 003: ID 1235:800c Focusrite-Novation Scarlett 18i20 Bus 002 Device 002: ID 8087:0024 Intel Corp. Integrated Rate Matching Hub Bus 002 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub Bus 001 Device 004: ID 0cf3:3005 Atheros Communications, Inc. AR3011 Bluetooth Bus 001 Device 003: ID 0c45:6436 Microdia Bus 001 Device 002: ID 8087:0024 Intel Corp. Integrated Rate Matching Hub Bus 001 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub
System Information Manufacturer: Sony Corporation Product Name: VPCEH2H1E Version: C105Z1UX Serial Number: UUID: Wake-up Type: Power Switch SKU Number: N/A Family: VAIO
Il giorno lun, 26/01/2015 alle 08.30 -0600, Chris J Arges ha scritto:
On 01/26/2015 07:28 AM, Guido Aulisi wrote:
I will ASAP, now I'm not at home. IMHO I think this is something about usb than the mixer, because I didn't use it, I only noticed the xruns even without playing audio. I didn't try 3.18, because it isn't shipped with fedora yet.
Ciao
2015-01-26 14:04 GMT+01:00 Takashi Iwai tiwai@suse.de:
At Sun, 25 Jan 2015 12:42:22 +0100, Guido Aulisi wrote:
Hi, I noticed a regression in 3.19-rc5 kernel (as shipped by fedora rawhide) with usb2 focusrite scarlett 18i20. I get a lot of xruns with jack, even at 1024 frame buffer and 2 periods. In 3.17.8 it works quite well without xruns at the same settings.
Guido, Takashi,
I tried to reproduce this and failed on my Scarlett 18i8 (I don't have any other hardware).
I've just tested Ubuntu's 3.18.2 and daily mainline 3.19-r5 as follows:
- Setup Scarlett 18i8 w/ jack (RT, 1024 frames/period, sample rate 48k,
periods/buffer 2). 2) Launch Ardour3 and create/arm 8 tracks 3) Hit record
Using this I was unable to see any XRuns even when using Scarlett's mixer with alsamixer in the background on 3.19.
Let me know more details. Thanks, --chris j arges
I use a fedora 20 (64bit) laptop with 8GB ram, I only installed the rawhide kernel to try the new included scarlett mixer, no other software components changed.
I can try to do some other test, but compiling the kernel on this laptop is time consuming.
Thanks Guido
Looks like a regression in the recent rework for Scarlett devices. Chris, do you have any clue?
In anyway, Guido, could you give lsusb -v output (for the relevant part)?
thanks,
Takashi
participants (3)
-
Chris J Arges -
Guido Aulisi -
Takashi Iwai