Skip to contentSkip to conversation

kaweth fails to work on 2.6.12-rc[56]

DISCUSSION Wakko Warner 3 participants 11 messages Jun 8, 2005 – Jun 11, 2005

lore
Share & tools
lore
#0 lore raw

Seems that it is unable to send packets however I can see packets coming in. I downgraded to 2.6.12-rc2 which works.

I have not tested rc3 or rc4 yet. I am preparing to try rc4.

Device ID: 0846:1001

Info from /sys:

bConfigurationValue: 1
bDeviceClass: 00
bDeviceProtocol: 00
bDeviceSubClass: 00
bMaxPower:  90mA
bNumConfigurations: 1
bNumInterfaces:  1
bcdDevice: 0202
bmAttributes: 80
configuration: detach_state: 0
devnum: 8
idProduct: 1001
idVendor: 0846
manufacturer: NETGEAR
maxchild: 0
product: NETGEAR EA101 USB Ethernet Adapter
speed: 12
version:  1.00
Signature · 1 line
 Lab tests show that use of micro$oft causes cancer in lab animals
#1 lore raw
Quoted 4 lines · Wakko Warner

Seems that it is unable to send packets however I can see packets coming in. I downgraded to 2.6.12-rc2 which works.

I have not tested rc3 or rc4 yet. I am preparing to try rc4.

Just finished testing.
rc2 works
rc3 works (fails on aic7xxx with my scsi hardware but rc5 works with my
scsi hardware)
rc4 through rc6 do not work.

I am currently running on 2.6.12-rc2

Quoted 30 lines

Device ID: 0846:1001

Info from /sys:

bConfigurationValue: 1
bDeviceClass: 00
bDeviceProtocol: 00
bDeviceSubClass: 00
bMaxPower:  90mA
bNumConfigurations: 1
bNumInterfaces:  1
bcdDevice: 0202
bmAttributes: 80
configuration: detach_state: 0
devnum: 8
idProduct: 1001
idVendor: 0846
manufacturer: NETGEAR
maxchild: 0
product: NETGEAR EA101 USB Ethernet Adapter
speed: 12
version:  1.00
Signature · 6 lines
 Lab tests show that use of micro$oft causes cancer in lab animals
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/
Signature · 1 line
 Lab tests show that use of micro$oft causes cancer in lab animals
#2 lore raw

Am Mittwoch, 8. Juni 2005 05:11 schrieb Wakko Warner:

Quoted 11 lines
Quoted 4 lines · Wakko Warner

Seems that it is unable to send packets however I can see packets coming in. I downgraded to 2.6.12-rc2 which works.

I have not tested rc3 or rc4 yet.  I am preparing to try rc4.

Just finished testing.
rc2 works
rc3 works (fails on aic7xxx with my scsi hardware but rc5 works with my scsi hardware)
rc4 through rc6 do not work.

Hi,

have you tried recompiling with debug enabled? It does show the status codes in the interrupt handler.

Regards
Oliver

#3 lore raw
Quoted 4 lines · Oliver Neukum

Am Mittwoch, 8. Juni 2005 05:11 schrieb Wakko Warner:

Quoted 3 lines
Quoted 2 lines · Wakko Warner

Seems that it is unable to send packets however I can see packets coming in. I downgraded to 2.6.12-rc2 which works.

Don't forget to report which host controller(s) it fails with, and which it works with ...

#4 lore raw
Quoted 8 lines · David Brownell
Quoted 4 lines · Oliver Neukum

Am Mittwoch, 8. Juni 2005 05:11 schrieb Wakko Warner:

Quoted 3 lines
Quoted 2 lines · Wakko Warner

Seems that it is unable to send packets however I can see packets coming in. I downgraded to 2.6.12-rc2 which works.

Don't forget to report which host controller(s) it fails with, and which it works with ...

See the last message I sent. The host controller did not change, only the kernel version. The .config really didn't change except for the new stuff between rc2 and newer.

Signature · 1 line
 Lab tests show that use of micro$oft causes cancer in lab animals
#5 lore raw
In reply to Oliver Neukum
Quoted 13 lines · Oliver Neukum

Am Mittwoch, 8. Juni 2005 05:11 schrieb Wakko Warner:

Quoted 11 lines
Quoted 4 lines · Wakko Warner

Seems that it is unable to send packets however I can see packets coming in. I downgraded to 2.6.12-rc2 which works.

I have not tested rc3 or rc4 yet. ?I am preparing to try rc4.

Just finished testing.
rc2 works
rc3 works (fails on aic7xxx with my scsi hardware but rc5 works with my ????????scsi hardware)
rc4 through rc6 do not work.

Quoted 4 lines

Hi,

have you tried recompiling with debug enabled? It does show the status codes in the interrupt handler.

I have not. My keyboard and mouse (on a hub) are plugged in beside the kaweth device so they would be on the same interrupt.

I did notice that the kaweth file itself hasn't changed, but something is different between rc3 and rc4.

Someone else asked what host controllers. I tried it on 2 systems. The device itself is USB1. One machine running rc5 used Intel 440bx chipset. I'm not sure what controller it has but I was using the onboard USB which uses uhci. I don't have that machine anymore, it was temporary. I'm currently using this on an Intel E7505 chipset which uses both ehci and uhci (loaded in that order)

I used the same configfile from rc2 when I compiled rc3 and rc4 (doing make old config of course). I'll see about testing this later on. I don't have time this morning to look at this.

Signature · 1 line
 Lab tests show that use of micro$oft causes cancer in lab animals
#6 lore raw

Am Mittwoch, 8. Juni 2005 12:42 schrieb Wakko Warner:

Quoted 5 lines
Quoted 2 lines

have you tried recompiling with debug enabled? It does show the status codes in the interrupt handler.

I have not.  My keyboard and mouse (on a hub) are plugged in beside the kaweth device so they would be on the same interrupt.

Sorry, I should be more specific. It will print out the error codes internal to the USB layer which are meaningful even if interrupts are shared. Also, are you seeing tx errors in the error count?

Regards
Oliver

#7 lore raw
Quoted 10 lines · Oliver Neukum

Am Mittwoch, 8. Juni 2005 12:42 schrieb Wakko Warner:

Quoted 5 lines
Quoted 2 lines

have you tried recompiling with debug enabled? It does show the status codes in the interrupt handler.

I have not. ?My keyboard and mouse (on a hub) are plugged in beside the kaweth device so they would be on the same interrupt.

Sorry, I should be more specific. It will print out the error codes internal to the USB layer which are meaningful even if interrupts are shared. Also, are you seeing tx errors in the error count?

Unless I did something wrong, it's apparently not in the USB subsystem.

I have 2.6.12-rc3 and rc4 compiled. Same config file on a test system. again, rc3 works, rc4 does not. I booted rc3 and force loaded the usb modules from rc4 into rc3. It tainted the kernel, but it worked.

I booted into rc4. I do have ACPI compiled so I added acpi=off. This did not work, same problem as before. I left a ping running, the usb controller was on IRQ 9. The test system is an old NEC notebook with an Intel chipset (thus uhci-hcd usb1.1). I plugged in a Belkin USB2.0 cardbus card (USB1.1 is serviced by ohci-hcd) and plugged the ethernet adapter into that. Same problem.

As far as controllers, ALL except 1 (The cardbus USB2.0) has been some version of Intel. I have tested this on 4 different systems with the same results. An MSI 6163 board, a Supermicro X5DA8 board, a Dell Inspiron 8100, and the NEC Versa FX. The Supermicro is running SMP, but I doubt that matters.

At this time, it's late. I'll see about reverting files from rc4 back to rc3 to see which one causes it to work again.

Signature · 1 line
 Lab tests show that use of micro$oft causes cancer in lab animals
#8 lore raw

Am Samstag, 11. Juni 2005 06:11 schrieb Wakko Warner:

Quoted 18 lines
Quoted 10 lines · Oliver Neukum

Am Mittwoch, 8. Juni 2005 12:42 schrieb Wakko Warner:

Quoted 5 lines
Quoted 2 lines

have you tried recompiling with debug enabled? It does show the status codes in the interrupt handler.

I have not. ?My keyboard and mouse (on a hub) are plugged in beside the kaweth device so they would be on the same interrupt.

Sorry, I should be more specific. It will print out the error codes internal to the USB layer which are meaningful even if interrupts are shared. Also, are you seeing tx errors in the error count?

Unless I did something wrong, it's apparently not in the USB subsystem.

I have 2.6.12-rc3 and rc4 compiled. Same config file on a test system. again, rc3 works, rc4 does not. I booted rc3 and force loaded the usb modules from rc4 into rc3. It tainted the kernel, but it worked.

Quoted 6 lines

I booted into rc4. I do have ACPI compiled so I added acpi=off. This did not work, same problem as before. I left a ping running, the usb controller was on IRQ 9. The test system is an old NEC notebook with an Intel chipset (thus uhci-hcd usb1.1). I plugged in a Belkin USB2.0 cardbus card (USB1.1 is serviced by ohci-hcd) and plugged the ethernet adapter into that. Same problem.

Do you see tx errors?

Regards
Oliver

#9 lore raw
Quoted 12 lines · Oliver Neukum

Am Samstag, 11. Juni 2005 06:11 schrieb Wakko Warner:

Quoted 11 lines
Quoted 3 lines · Oliver Neukum

Sorry, I should be more specific. It will print out the error codes internal to the USB layer which are meaningful even if interrupts are shared. Also, are you seeing tx errors in the error count?

Unless I did something wrong, it's apparently not in the USB subsystem.

I have 2.6.12-rc3 and rc4 compiled. Same config file on a test system. again, rc3 works, rc4 does not. I booted rc3 and force loaded the usb modules from rc4 into rc3. It tainted the kernel, but it worked.

Quoted 8 lines
Quoted 6 lines

I booted into rc4. I do have ACPI compiled so I added acpi=off. This did not work, same problem as before. I left a ping running, the usb controller was on IRQ 9. The test system is an old NEC notebook with an Intel chipset (thus uhci-hcd usb1.1). I plugged in a Belkin USB2.0 cardbus card (USB1.1 is serviced by ohci-hcd) and plugged the ethernet adapter into that. Same problem.

Do you see tx errors?

No. Oh oops I forgot to say the ping caused the interrupts to increase. ifconfig showed 1 packet sent but never went up.

Signature · 1 line
 Lab tests show that use of micro$oft causes cancer in lab animals
#10 lore raw

Am Samstag, 11. Juni 2005 14:52 schrieb Wakko Warner:

Quoted 20 lines
Quoted 19 lines · Oliver Neukum

Am Samstag, 11. Juni 2005 06:11 schrieb Wakko Warner:

Quoted 17 lines
Quoted 3 lines · Oliver Neukum

Sorry, I should be more specific. It will print out the error codes internal to the USB layer which are meaningful even if interrupts are shared. Also, are you seeing tx errors in the error count?

Unless I did something wrong, it's apparently not in the USB subsystem.

I have 2.6.12-rc3 and rc4 compiled. Same config file on a test system. again, rc3 works, rc4 does not. I booted rc3 and force loaded the usb modules from rc4 into rc3. It tainted the kernel, but it worked.

I booted into rc4. I do have ACPI compiled so I added acpi=off. This did not work, same problem as before. I left a ping running, the usb controller was on IRQ 9. The test system is an old NEC notebook with an Intel chipset (thus uhci-hcd usb1.1). I plugged in a Belkin USB2.0 cardbus card (USB1.1 is serviced by ohci-hcd) and plugged the ethernet adapter into that. Same problem.

Quoted 4 lines
Quoted 1 line

Do you see tx errors?

No. Oh oops I forgot to say the ping caused the interrupts to increase. ifconfig showed 1 packet sent but never went up.

This might very well indicate that the queue is not reenabled. The interrupt handler should do that. If you enable debug in kaweth.c, you'll learn more. If it doesn't help, I'll send you a patch which produces more output.

Regards
Oliver