[9fans] usb problem of Asus P5K-VM motherboard

[9fans] usb problem of Asus P5K-VM motherboard

[kokamoto]

I'm now being forced to use this motherboard with C2D 2.66GHZ for
a new fossil+venti server of my new lab with a SATA HDD and a SATA
CDROM drive.   I know this is not suitable to Plan 9, however forced to...

As you know, it has no PS/2 mouse port, but only has usb (12 ports!).
As I faced usb problem (uhci), I put a serial RS232C port to this mother
board, and using serial mouse now (not so happy though).

This machine reports only 6 usb hubs (not 8 hubs I have), why?
Anyone using this boad' usb succefull?

Thanks in advance

Kenji

Re: [9fans] usb problem of Asus P5K-VM motherboard

[erik quanstrom]

> I'm now being forced to use this motherboard with C2D 2.66GHZ for
> a new fossil+venti server of my new lab with a SATA HDD and a SATA
> CDROM drive.   I know this is not suitable to Plan 9, however forced to...
>
> As you know, it has no PS/2 mouse port, but only has usb (12 ports!).
> As I faced usb problem (uhci), I put a serial RS232C port to this mother
> board, and using serial mouse now (not so happy though).
>
> This machine reports only 6 usb hubs (not 8 hubs I have), why?
> Anyone using this boad' usb succefull?
>
> Thanks in advance

could you send the output of pci?

i have some patches to the uart code that will allow you to
use a non-onboard uart as a console if you're interested.

- erik

Re: [9fans] usb problem of Asus P5K-VM motherboard

[kokamoto]

> i have some patches to the uart code that will allow you to
> use a non-onboard uart as a console if you're interested.

Thanks for your reply, erick.
This motherboard has on board serial connector but not
the cable to the rear panel.   I hand-maded that's cable, and
using serial mouse.    However, serial mouse is not so good
for a large display area...

The output of pci will be reported later (I'm now in home)

Kenji

Re: [9fans] usb problem of Asus P5K-VM motherboard

[erik quanstrom]

On Thu Jan 29 06:17:11 EST 2009, kokamoto@hera.eonet.ne.jp wrote:
> I'm now being forced to use this motherboard with C2D 2.66GHZ for
> a new fossil+venti server of my new lab with a SATA HDD and a SATA
> CDROM drive.   I know this is not suitable to Plan 9, however forced to...
>
> As you know, it has no PS/2 mouse port, but only has usb (12 ports!).
> As I faced usb problem (uhci), I put a serial RS232C port to this mother
> board, and using serial mouse now (not so happy though).
>
> This machine reports only 6 usb hubs (not 8 hubs I have), why?
> Anyone using this boad' usb succefull?
>
> Thanks in advance

sorry about the double post, but since you mention
the benefits of ken's fs and this motherboard together,
i should mention that you can use ken's fs with sata
drives if you have a hba that supports ahci, ide
emulation on sata (many do), or the marvell 88sx chipset:

/n/sources/contrib/quanstro/src/myfs

- erik

Re: [9fans] usb problem of Asus P5K-VM motherboard

[kokamoto]

> i should mention that you can use ken's fs with sata
> drives if you have a hba that supports ahci, ide
> emulation on sata (many do), or the marvell 88sx chipset:

Wao, good news!

Yes, it has marvell 88... chip.   Later I'll confirm it...
I have already one Ken's fs (geof's 64bit update version), and this time
it'll be used as fossil+venti server.   The sata drives are now
being used as compatible mode (IDE), sick☺

Kenji

Re: [9fans] usb problem of Asus P5K-VM motherboard

[kokamoto]

my P5K-VM machine's pci output is as follows:
0.0.0:	brg  06.00.00 8086/29c0   0
0.2.0:	vid  03.00.00 8086/29c2  10 0:fea00000 524288 1:0000cc01 16 2:d0000008 268435456 3:fe900000 1048576
0.2.1:	vid  03.80.00 8086/29c3   0 0:fea80000 524288
0.26.0:	usb  0c.03.00 8086/2937  10 4:0000c481 32
0.26.1:	usb  0c.03.00 8086/2938   3 4:0000c801 32
0.26.2:	usb  0c.03.00 8086/2939   5 4:0000c881 32
0.27.0:	aud  04.03.00 8086/293e  11 0:fe8f8004 16384 1:00000000 16
0.28.0:	brg  06.04.00 8086/2940  11
0.28.4:	brg  06.04.00 8086/2948  11
0.29.0:	usb  0c.03.00 8086/2934   5 4:0000c001 32
0.29.1:	usb  0c.03.00 8086/2935  10 4:0000c081 32
0.29.2:	usb  0c.03.00 8086/2936   5 4:0000c401 32
0.30.0:	brg  06.04.01 8086/244e 255
0.31.0:	brg  06.01.00 8086/2918   0
0.31.2:	disk 01.01.8a 8086/2921   0 0:00000001 16 1:00000001 16 2:00000001 16 3:00000001 16 4:0000ff91 16 5:0000ffa1 16
0.31.3:	smb  0c.05.00 8086/2930  10 0:fe8f7c04 256 1:00000000 16 4:00000401 32
0.31.5:	disk 01.01.85 8086/2926  11 0:0000bc01 16 1:0000b881 16 2:0000b801 16 3:0000b481 16 4:0000b401 16 5:0000b081 16
1.0.0:	disk 01.01.85 197b/2368  10 0:0000dc01 16 1:0000d881 16 2:0000d801 16 3:0000d481 16 4:0000d401 16
3.1.0:	net  02.00.00 8086/1229  11 0:febff000 4096 1:0000ec01 64 2:febc0000 131072
3.2.0:	---  0c.00.10 1106/3044   5 0:febfe800 2048 1:0000e881 128

I disabled onboard nic(Marvell 88E8056), and using Intel's i82557.
A curious poit is there are only six usb ports here.
When I plugin any USB device on this board, system hungs.

Kenji

Re: [9fans] usb problem of Asus P5K-VM motherboard

[erik quanstrom]

> I disabled onboard nic(Marvell 88E8056), and using Intel's i82557.
> A curious poit is there are only six usb ports here.
> When I plugin any USB device on this board, system hungs.
>
> Kenji

i'm don't know very much about plan 9's usb stuff but
the motherboard claims more ports (12) than UsbMax
(10).  i wonder if that could cause some trouble.
see pc/devusb/^usbinit.

- erik

Re: [9fans] usb problem of Asus P5K-VM motherboard

[kokamoto]

> my P5K-VM machine's pci output is as follows:
> 0.26.0:	usb  0c.03.00 8086/2937  10 4:0000c481 32
> 0.26.1:	usb  0c.03.00 8086/2938   3 4:0000c801 32
> 0.26.2:	usb  0c.03.00 8086/2939   5 4:0000c881 32
> 0.29.0:	usb  0c.03.00 8086/2934   5 4:0000c001 32
> 0.29.1:	usb  0c.03.00 8086/2935  10 4:0000c081 32
> 0.29.2:	usb  0c.03.00 8086/2936   5 4:0000c401 32

I debugged the devusv.c usbuhci.c in /sys/src/9/pc, and got the
following result, where I'm confused why the b=0x0 value of the last line.
If this is true it would write to 0!, and system hungs.   It's very
reasonable, but why we get b=0x0 at dumptd()...

Anyway the debugging outputs after I plugined a usb disk are:

e: 95
r: 95
r2: 280
r3: 80
r: 95 0
e: 95
speed 1
out [8] 80 06 00 01 00 00 08 00
queuetd qxmit: t=f0047080 lt=f0047080 q=f0048080 first=0 last=0 entries=00000001
	t=f0047080 q=f0048080 first=f0047080 last=f0047080 entries=00047080
qh 0xf0048080: 0x480c2 0x1
qh 0xf00480c0: 0x480e2 0x1
qh 0xf00480e0: 0x480a2 0x1
qh 0xf00480a0: 0x1 0x47060
td 0xf0047060: l=0x47060 s=0x0 d=0x0 b=0x0 0x0 f=0x0
	s=,ep=0,d=0,D=0

here stalls.

The functions used are:
  portenable()-->portreset()-->portenable()-->write()-->qxmit()-->queuetd()
	-->dumpqh()-->dumptd()

I touched some on usb device long time ago, and now forgot all
because of my aging efect.

If someone point me the place reporting wrong values, please
let me know it.

Thanks inadvance

Kenji

Re: [9fans] usb problem of Asus P5K-VM motherboard

[Francisco J Ballesteros]

control transfers may read/write 0 bytes.
That's not the issue probably.

IMHO, it's likely that at some point the driver missed a td and
from that point on the endpoint stalled.


On Wed, Feb 4, 2009 at 11:57 AM,  <kokamoto@hera.eonet.ne.jp> wrote:
>> my P5K-VM machine's pci output is as follows:
>> 0.26.0:       usb  0c.03.00 8086/2937  10 4:0000c481 32
>> 0.26.1:       usb  0c.03.00 8086/2938   3 4:0000c801 32
>> 0.26.2:       usb  0c.03.00 8086/2939   5 4:0000c881 32
>> 0.29.0:       usb  0c.03.00 8086/2934   5 4:0000c001 32
>> 0.29.1:       usb  0c.03.00 8086/2935  10 4:0000c081 32
>> 0.29.2:       usb  0c.03.00 8086/2936   5 4:0000c401 32
>
> I debugged the devusv.c usbuhci.c in /sys/src/9/pc, and got the
> following result, where I'm confused why the b=0x0 value of the last line.
> If this is true it would write to 0!, and system hungs.   It's very
> reasonable, but why we get b=0x0 at dumptd()...
>
> Anyway the debugging outputs after I plugined a usb disk are:
>
> e: 95
> r: 95
> r2: 280
> r3: 80
> r: 95 0
> e: 95
> speed 1
> out [8] 80 06 00 01 00 00 08 00
> queuetd qxmit: t=f0047080 lt=f0047080 q=f0048080 first=0 last=0 entries=00000001
>        t=f0047080 q=f0048080 first=f0047080 last=f0047080 entries=00047080
> qh 0xf0048080: 0x480c2 0x1
> qh 0xf00480c0: 0x480e2 0x1
> qh 0xf00480e0: 0x480a2 0x1
> qh 0xf00480a0: 0x1 0x47060
> td 0xf0047060: l=0x47060 s=0x0 d=0x0 b=0x0 0x0 f=0x0
>        s=,ep=0,d=0,D=0
>
> here stalls.
>
> The functions used are:
>  portenable()-->portreset()-->portenable()-->write()-->qxmit()-->queuetd()
>        -->dumpqh()-->dumptd()
>
> I touched some on usb device long time ago, and now forgot all
> because of my aging efect.
>
> If someone point me the place reporting wrong values, please
> let me know it.
>
> Thanks inadvance
>
> Kenji
>
>
>

Re: [9fans] usb problem of Asus P5K-VM motherboard

[kokamoto]

> control transfers may read/write 0 bytes.
> That's not the issue probably.

No, I meant kernel tried to write to fd 0, not 0 byte.

Kenji

Re: [9fans] usb problem of Asus P5K-VM motherboard

[Francisco J Ballesteros]

If I understood your mail, the problem is that buf is 0.
That's not always a problem for a usb td. It could be a 0-bytes transfer.
Such thing is ok.

On Wed, Feb 4, 2009 at 3:50 PM,  <kokamoto@hera.eonet.ne.jp> wrote:
>> control transfers may read/write 0 bytes.
>> That's not the issue probably.
>
> No, I meant kernel tried to write to fd 0, not 0 byte.
>
> Kenji
>
>
>