Small Regression in midi for M-Audio Fast Track Ultra.

Small Regression in midi for M-Audio Fast Track Ultra.

[Grant Diffey]

In 2.6.33/34 when support originally was put in alsa mainline the MIDI ports
were functional today they're not for me and I get the following in the
kernel log when using this card. (Audio all works acceptably.)

[ 1846.736062] usb 2-1: new high speed USB device using ehci_hcd and address
4
[ 1846.869790] usb 2-1: config 1 interface 3 altsetting 0 bulk endpoint 0x7
has invalid maxpacket 8
[ 1846.869802] usb 2-1: config 1 interface 3 altsetting 0 bulk endpoint 0x87
has invalid maxpacket 8
[ 1846.870897] usb 2-1: New USB device found, idVendor=0763, idProduct=2080
[ 1846.870905] usb 2-1: New USB device strings: Mfr=1, Product=2,
SerialNumber=0
[ 1846.870912] usb 2-1: Product: Fast Track Ultra
[ 1846.870917] usb 2-1: Manufacturer: M-Audio

This issue may be unrelated to the data below but it's included in case it's
useful.

The midi port is interface 3 and has a comment in quirk-table.h like this

                        /* interface 3 (MIDI) is standard compliant */
                        {
                               .ifnum = -1
                        }

Which is as I understand it a noop to the quirk stuff so that it just get's
processed normally. Which could also be completely wrong.

I've verified in lsusb -v of the midi interface (3) of the card is below
    Interface Descriptor:
      bLength                 9
      bDescriptorType         4
      bInterfaceNumber        3
      bAlternateSetting       0
      bNumEndpoints           2
      bInterfaceClass         1 Audio
      bInterfaceSubClass      3 MIDI Streaming
      bInterfaceProtocol      0
      iInterface              5
      MIDIStreaming Interface Descriptor:
        bLength                 7
        bDescriptorType        36
        bDescriptorSubtype      1 (HEADER)
        bcdADC               1.00
        wTotalLength           63
      MIDIStreaming Interface Descriptor:
        bLength                 6
        bDescriptorType        36
        bDescriptorSubtype      2 (MIDI_IN_JACK)
        bJackType               1 Embedded
        bJackID                 1
        iJack                   0
      MIDIStreaming Interface Descriptor:
        bLength                 6
        bDescriptorType        36
        bDescriptorSubtype      2 (MIDI_IN_JACK)
        bJackType               2 External
        bJackID                 2
        iJack                   0
      MIDIStreaming Interface Descriptor:
        bLength                 8
        bDescriptorType        36
        bDescriptorSubtype      3 (MIDI_OUT_JACK)
      Warning: Descriptor too short
        bJackType               1 Embedded
        bJackID                 3
        bNrInputPins            1
        baSourceID( 0)          2
        BaSourcePin( 0)         1
        iJack                   8
      MIDIStreaming Interface Descriptor:
        bLength                 8
        bDescriptorType        36
        bDescriptorSubtype      3 (MIDI_OUT_JACK)
      Warning: Descriptor too short
        bJackType               2 External
        bJackID                 4
        bNrInputPins            1
        baSourceID( 0)          1
        BaSourcePin( 0)         1
        iJack                  97
      Endpoint Descriptor:
        bLength                 7
        bDescriptorType         5
        bEndpointAddress     0x07  EP 7 OUT
        bmAttributes            2
          Transfer Type            Bulk
          Synch Type               None
          Usage Type               Data
        wMaxPacketSize     0x0008  1x 8 bytes
        bInterval               8
        MIDIStreaming Endpoint Descriptor:
          bLength                 5
          bDescriptorType        37
          bDescriptorSubtype      1 (GENERAL)
          bNumEmbMIDIJack         1
          baAssocJackID( 0)       1
      Endpoint Descriptor:
        bLength                 7
        bDescriptorType         5
        bEndpointAddress     0x87  EP 7 IN
        bmAttributes            2
          Transfer Type            Bulk
          Synch Type               None
          Usage Type               Data
        wMaxPacketSize     0x0008  1x 8 bytes
        bInterval               0
        MIDIStreaming Endpoint Descriptor:
          bLength                 5
          bDescriptorType        37
          bDescriptorSubtype      1 (GENERAL)
          bNumEmbMIDIJack         1
          baAssocJackID( 0)       3

Re: USB MIDI Regression? - - Was: Small Regression in midi for M-Audio Fast Track Ultra.

[Felix Homann]

Hi,

Am 16.04.2011 04:29, schrieb Grant Diffey:
> In 2.6.33/34 when support originally was put in alsa mainline the MIDI ports
> were functional today they're not for me and I get the following in the
> kernel log when using this card. (Audio all works acceptably.)

I can confirm Grant's issue.

They are not related to any code specific to the M-Audio Fast Track 
Ultra (FTU) series:

Even after removing the FTU patches from current ALSA (git) the MIDI 
ports of the device will not be recognized.

Switching back to my oldest installed kernel (2.6.32) the FTU will be 
recognized as a MIDI interface (but as expected the PCM ports will not 
be recognized.)

Just to make sure, that I'm not missing anything:

* I've used 'amidi - l' to look for MIDI ports
* Due to the deprecation of OSS in the kernel my Alsa configure line now 
includes '--with-oss=no --with-pcm-oss-plugins=no' which shouldn't have 
an effect but I mention it anyway

Any ideas?

Regards,

Felix

Re: Small Regression in midi for M-Audio Fast Track Ultra.

[Clemens Ladisch]

Grant Diffey wrote:
> In 2.6.33/34 when support originally was put in alsa mainline the MIDI ports
> were functional today they're not [...]
> The midi port is interface 3 and has a comment in quirk-table.h like this
> 
>                         /* interface 3 (MIDI) is standard compliant */

Apparently, having no entry for an interface prevents the driver from
looking at this interface at all.

Try replacing the comment with:
			{
				.iface = 3,
				.type = QUIRK_MIDI_STANDARD_INTERFACE
			},


Regards,
Clemens

Re: Small Regression in midi for M-Audio Fast Track Ultra.

[Felix Homann]

Thanks Clemens,

for your quick reply.

Am 18.04.2011 11:16, schrieb Clemens Ladisch:
>
> Try replacing the comment with:
> 			{
> 				.iface = 3,
> 				.type = QUIRK_MIDI_STANDARD_INTERFACE
> 			},
>

I already tried that yesterday. It didn't change anything.

Remember, on my 2.6.32 kernel the MIDI part of that device will be 
recognized without any quirk for the FTU. If I remove all FTU related 
patches from current alsa-driver/alsa-kernel  (over the Ubuntu 
2.6.35-28-generic kernel) the device will not be recognized as a MIDI 
interface.

Regards,

Felix

Re: Small Regression in midi for M-Audio Fast Track Ultra.

[Felix Homann]

[-- Attachment #1: Type: text/plain, Size: 747 bytes --]

Sorry, Clemens,

I must have messed around with too many Alsa versions on my machine, again.
Your suggested fix _does_ fix the issue!

Would you mind adding this to the git tree? Or shall I prepare a patch?

Am 18.04.2011 11:16, schrieb Clemens Ladisch:
>> Try replacing the comment with:
>> 			{
>> 				.iface = 3,
>> 				.type = QUIRK_MIDI_STANDARD_INTERFACE
>> 			},
>>

In the meantime I started a bisect session on the mainline kernel. The 
cause of this issue lies in the refactoring done last year. I couldn't 
identify the exact patch introducing the regression, though, since 
starting at step 10 the kernels would not boot. See the corresponding 
git bisect log (the last revision is the first one that would not boot).

Regards,

Felix

[-- Attachment #2: bisect-log-step10.log --]
[-- Type: text/x-log, Size: 1829 bytes --]

git bisect start
# bad: [9fe6206f400646a2322096b56c59891d530e8d51] Linux 2.6.35
git bisect bad 9fe6206f400646a2322096b56c59891d530e8d51
# good: [e40152ee1e1c7a63f4777791863215e3faa37a86] Linus 2.6.34
git bisect good e40152ee1e1c7a63f4777791863215e3faa37a86
# bad: [e0bc5d4a54938eedcde14005210e6c08aa9727e4] Merge branch 'i2c-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jdelvare/staging
git bisect bad e0bc5d4a54938eedcde14005210e6c08aa9727e4
# bad: [04afb40593f9a3007e5ea817d009529ef10fb685] Merge branch 'acpica' of git://git.kernel.org/pub/scm/linux/kernel/git/lenb/linux-acpi-2.6
git bisect bad 04afb40593f9a3007e5ea817d009529ef10fb685
# good: [e4e47eb15b7884963efe7f98231009c5770a2c3d] Merge branch 'msm-core' of git://codeaurora.org/quic/kernel/dwalker/linux-msm
git bisect good e4e47eb15b7884963efe7f98231009c5770a2c3d
# good: [e8d0416796d43a950ec7b65629e53419b2e22453] V4L/DVB (13554a): v4l: Use the video_drvdata function in drivers
git bisect good e8d0416796d43a950ec7b65629e53419b2e22453
# good: [f39d01be4c59a61a08d0cb53f615e7016b85d339] Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jikos/trivial
git bisect good f39d01be4c59a61a08d0cb53f615e7016b85d339
# good: [ad8332c1302bcb4f80d593fd3eb477be9d7f5604] ASoC: tpa6130a2: Remove CPVSS and HPVdd supplies
git bisect good ad8332c1302bcb4f80d593fd3eb477be9d7f5604
# bad: [19008bdacb9f7841166ebafe0aef361ee582ffbf] Merge branch 'topic/hda' into for-linus
git bisect bad 19008bdacb9f7841166ebafe0aef361ee582ffbf
# bad: [550a8b691ca67761bbf382d98fbd81d215f1d7f0] ALSA: sound/pci/asihpi: Use kzalloc
git bisect bad 550a8b691ca67761bbf382d98fbd81d215f1d7f0
# bad: [ca4c2adaf2d7efcc43e16c8010cf4c30def75058] ALSA: usb/mixer - use get_iface_desc() rather than direct structure
git bisect bad ca4c2adaf2d7efcc43e16c8010cf4c30def75058

[-- Attachment #3: Type: text/plain, Size: 0 bytes --]

Re: USB MIDI Regression? - - Was: Small Regression in midi for M-Audio Fast Track Ultra.

[Daniel Mack]

Hi Felix,

On Mon, Apr 18, 2011 at 10:37 AM, Felix Homann <linuxaudio@showlabor.de> wrote:
> Am 16.04.2011 04:29, schrieb Grant Diffey:
>> In 2.6.33/34 when support originally was put in alsa mainline the MIDI ports
>> were functional today they're not for me and I get the following in the
>> kernel log when using this card. (Audio all works acceptably.)
>
> I can confirm Grant's issue.
>
> They are not related to any code specific to the M-Audio Fast Track
> Ultra (FTU) series:
>
> Even after removing the FTU patches from current ALSA (git) the MIDI
> ports of the device will not be recognized.
>
> Switching back to my oldest installed kernel (2.6.32) the FTU will be
> recognized as a MIDI interface (but as expected the PCM ports will not
> be recognized.)

Hmm, between these version, there was my huge refactoring of the
usb-audio code, so this could indeed be related. However, most of my
patches were for the audio part, and not for MIDI.

> Just to make sure, that I'm not missing anything:
>
> * I've used 'amidi - l' to look for MIDI ports
> * Due to the deprecation of OSS in the kernel my Alsa configure line now
> includes '--with-oss=no --with-pcm-oss-plugins=no' which shouldn't have
> an effect but I mention it anyway

I would rather start off with two source trees of the kernel (one for
2.6.32 and one for any newer version that doesn't work anymore) and
then put lots of printk() lines to trace the functions. If you build
the driver as modules in both cases, you can easily add more debug
output on the fly without rebooting (some sort of "sudo rmmod ...;
make SUBDIRS=sound/usb; sudo insmod ..." is what I usually do).

I would like to help more, but it's difficult without access to the
hardware. Please let me know if you find out anything.

Daniel

Re: USB MIDI Regression? - - Was: Small Regression in midi for M-Audio Fast Track Ultra.

[Felix Homann]

[-- Attachment #1: Type: text/plain, Size: 712 bytes --]

Hi Daniel,

nice to "see" you again :-)


Am 23.04.2011 11:09, schrieb Daniel Mack:
>
> Hmm, between these version, there was my huge refactoring of the
> usb-audio code, so this could indeed be related. However, most of my
> patches were for the audio part, and not for MIDI.

It happened on the transition from 2.6.34 to 2.6.35, i.e. exactly 
whithin the refactoring phase. I've sent a bisect log to the list some 
posts before. I've attached it to this mail again, in case you didn't 
see it. The last iteration would not result in a bootable kernel, so 
I've stopped there. (Actually I didn't stop there, but even guessing 
good/bad would not yield a bootable kernel in the next step).

Kind regards,

Felix

[-- Attachment #2: bisect-log-step10.log --]
[-- Type: text/x-log, Size: 1829 bytes --]

git bisect start
# bad: [9fe6206f400646a2322096b56c59891d530e8d51] Linux 2.6.35
git bisect bad 9fe6206f400646a2322096b56c59891d530e8d51
# good: [e40152ee1e1c7a63f4777791863215e3faa37a86] Linus 2.6.34
git bisect good e40152ee1e1c7a63f4777791863215e3faa37a86
# bad: [e0bc5d4a54938eedcde14005210e6c08aa9727e4] Merge branch 'i2c-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jdelvare/staging
git bisect bad e0bc5d4a54938eedcde14005210e6c08aa9727e4
# bad: [04afb40593f9a3007e5ea817d009529ef10fb685] Merge branch 'acpica' of git://git.kernel.org/pub/scm/linux/kernel/git/lenb/linux-acpi-2.6
git bisect bad 04afb40593f9a3007e5ea817d009529ef10fb685
# good: [e4e47eb15b7884963efe7f98231009c5770a2c3d] Merge branch 'msm-core' of git://codeaurora.org/quic/kernel/dwalker/linux-msm
git bisect good e4e47eb15b7884963efe7f98231009c5770a2c3d
# good: [e8d0416796d43a950ec7b65629e53419b2e22453] V4L/DVB (13554a): v4l: Use the video_drvdata function in drivers
git bisect good e8d0416796d43a950ec7b65629e53419b2e22453
# good: [f39d01be4c59a61a08d0cb53f615e7016b85d339] Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jikos/trivial
git bisect good f39d01be4c59a61a08d0cb53f615e7016b85d339
# good: [ad8332c1302bcb4f80d593fd3eb477be9d7f5604] ASoC: tpa6130a2: Remove CPVSS and HPVdd supplies
git bisect good ad8332c1302bcb4f80d593fd3eb477be9d7f5604
# bad: [19008bdacb9f7841166ebafe0aef361ee582ffbf] Merge branch 'topic/hda' into for-linus
git bisect bad 19008bdacb9f7841166ebafe0aef361ee582ffbf
# bad: [550a8b691ca67761bbf382d98fbd81d215f1d7f0] ALSA: sound/pci/asihpi: Use kzalloc
git bisect bad 550a8b691ca67761bbf382d98fbd81d215f1d7f0
# bad: [ca4c2adaf2d7efcc43e16c8010cf4c30def75058] ALSA: usb/mixer - use get_iface_desc() rather than direct structure
git bisect bad ca4c2adaf2d7efcc43e16c8010cf4c30def75058

[-- Attachment #3: Type: text/plain, Size: 0 bytes --]

Re: Small Regression in midi for M-Audio Fast Track Ultra.

[Grant Diffey]

So I finally got around to building a new kernel today.

and after rebooting I get:

[  131.832084] usb 2-1: new high speed USB device using ehci_hcd and address
3
[  131.969795] usb 2-1: config 1 interface 3 altsetting 0 bulk endpoint 0x7
has invalid maxpacket 8
[  131.969800] usb 2-1: config 1 interface 3 altsetting 0 bulk endpoint 0x87
has invalid maxpacket 8
[  131.970886] usb 2-1: New USB device found, idVendor=0763, idProduct=2080
[  131.970890] usb 2-1: New USB device strings: Mfr=1, Product=2,
SerialNumber=0
[  131.970892] usb 2-1: Product: Fast Track Ultra
[  131.970894] usb 2-1: Manufacturer: M-Audio
[  132.060121] usbcore: registered new interface driver snd-usb-audio
nevyn@cetacea:~$ amidi -l
Dir Device    Name
IO  hw:1,0,0  Fast Track Ultra MIDI 1

Which is excellent and solves my problem.

now to find and apply that mixer patch.

Grant.


On Thu, Apr 21, 2011 at 6:25 AM, Felix Homann <linuxaudio@showlabor.de>wrote:

> Sorry, Clemens,
>
> I must have messed around with too many Alsa versions on my machine, again.
> Your suggested fix _does_ fix the issue!
>
> Would you mind adding this to the git tree? Or shall I prepare a patch?
>
>
> Am 18.04.2011 11:16, schrieb Clemens Ladisch:
>
>> Try replacing the comment with:
>>>                        {
>>>                                .iface = 3,
>>>                                .type = QUIRK_MIDI_STANDARD_INTERFACE
>>>                        },
>>>
>>>
> In the meantime I started a bisect session on the mainline kernel. The
> cause of this issue lies in the refactoring done last year. I couldn't
> identify the exact patch introducing the regression, though, since starting
> at step 10 the kernels would not boot. See the corresponding git bisect log
> (the last revision is the first one that would not boot).
>
> Regards,
>
> Felix
>

[PATCH] Re: Small Regression in midi for M-Audio Fast Track Ultra.

[Grant Diffey]

[-- Attachment #1: Type: text/plain, Size: 2021 bytes --]

oh here's a git patch if that helps

On Tue, Apr 26, 2011 at 7:17 PM, Grant Diffey <gdiffey@gmail.com> wrote:

> So I finally got around to building a new kernel today.
>
> and after rebooting I get:
>
> [  131.832084] usb 2-1: new high speed USB device using ehci_hcd and
> address 3
> [  131.969795] usb 2-1: config 1 interface 3 altsetting 0 bulk endpoint 0x7
> has invalid maxpacket 8
> [  131.969800] usb 2-1: config 1 interface 3 altsetting 0 bulk endpoint
> 0x87 has invalid maxpacket 8
> [  131.970886] usb 2-1: New USB device found, idVendor=0763, idProduct=2080
> [  131.970890] usb 2-1: New USB device strings: Mfr=1, Product=2,
> SerialNumber=0
> [  131.970892] usb 2-1: Product: Fast Track Ultra
> [  131.970894] usb 2-1: Manufacturer: M-Audio
> [  132.060121] usbcore: registered new interface driver snd-usb-audio
> nevyn@cetacea:~$ amidi -l
> Dir Device    Name
> IO  hw:1,0,0  Fast Track Ultra MIDI 1
>
> Which is excellent and solves my problem.
>
> now to find and apply that mixer patch.
>
> Grant.
>
>
>
> On Thu, Apr 21, 2011 at 6:25 AM, Felix Homann <linuxaudio@showlabor.de>wrote:
>
>> Sorry, Clemens,
>>
>> I must have messed around with too many Alsa versions on my machine,
>> again.
>> Your suggested fix _does_ fix the issue!
>>
>> Would you mind adding this to the git tree? Or shall I prepare a patch?
>>
>>
>> Am 18.04.2011 11:16, schrieb Clemens Ladisch:
>>
>>> Try replacing the comment with:
>>>>                        {
>>>>                                .iface = 3,
>>>>                                .type = QUIRK_MIDI_STANDARD_INTERFACE
>>>>                        },
>>>>
>>>>
>> In the meantime I started a bisect session on the mainline kernel. The
>> cause of this issue lies in the refactoring done last year. I couldn't
>> identify the exact patch introducing the regression, though, since starting
>> at step 10 the kernels would not boot. See the corresponding git bisect log
>> (the last revision is the first one that would not boot).
>>
>> Regards,
>>
>> Felix
>>
>
>

[-- Attachment #2: FastTrackUltra_MIDI_Quirk.patch --]
[-- Type: text/x-diff, Size: 615 bytes --]

diff --git a/sound/usb/quirks-table.h b/sound/usb/quirks-table.h
index 921a86f..6a25101 100644
--- a/sound/usb/quirks-table.h
+++ b/sound/usb/quirks-table.h
@@ -1972,7 +1972,10 @@ YAMAHA_DEVICE(0x7010, "UB99"),
 					}
 				}
 			},
-			/* interface 3 (MIDI) is standard compliant */
+			{
+				.ifnum = 3,
+				.type = QUIRK_MIDI_STANDARD_INTERFACE
+			},
 			{
 				.ifnum = -1
 			}
@@ -2039,7 +2042,10 @@ YAMAHA_DEVICE(0x7010, "UB99"),
 					}
 				}
 			},
-			/* interface 3 (MIDI) is standard compliant */
+			{
+				.ifnum = 3,
+				.type = QUIRK_MIDI_STANDARD_INTERFACE
+			},
 			{
 				.ifnum = -1
 			}

[-- Attachment #3: Type: text/plain, Size: 0 bytes --]

Re: [PATCH] Re: Small Regression in midi for M-Audio Fast Track Ultra.

[Daniel Mack]

On Tue, Apr 26, 2011 at 11:42 AM, Grant Diffey <gdiffey@gmail.com> wrote:
> oh here's a git patch if that helps
>
> On Tue, Apr 26, 2011 at 7:17 PM, Grant Diffey <gdiffey@gmail.com> wrote:
>>
>> So I finally got around to building a new kernel today.
>>
>> and after rebooting I get:
>>
>> [  131.832084] usb 2-1: new high speed USB device using ehci_hcd and
>> address 3
>> [  131.969795] usb 2-1: config 1 interface 3 altsetting 0 bulk endpoint
>> 0x7 has invalid maxpacket 8
>> [  131.969800] usb 2-1: config 1 interface 3 altsetting 0 bulk endpoint
>> 0x87 has invalid maxpacket 8
>> [  131.970886] usb 2-1: New USB device found, idVendor=0763,
>> idProduct=2080
>> [  131.970890] usb 2-1: New USB device strings: Mfr=1, Product=2,
>> SerialNumber=0
>> [  131.970892] usb 2-1: Product: Fast Track Ultra
>> [  131.970894] usb 2-1: Manufacturer: M-Audio
>> [  132.060121] usbcore: registered new interface driver snd-usb-audio
>> nevyn@cetacea:~$ amidi -l
>> Dir Device    Name
>> IO  hw:1,0,0  Fast Track Ultra MIDI 1
>>
>> Which is excellent and solves my problem.

Sorry everyone for the long absence - I've been travelling and it took
me some weeks to get all my setup up and ready again.

Is there anything on this topic that is yet to be solved, anything I
can help with?

Thanks,
Daniel