netdev.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* VLAN regression caused by: e1000: do vlan cleanup (799d531).
@ 2012-03-12 18:17 Ben Greear
  2012-03-13  9:59 ` Jiri Pirko
  0 siblings, 1 reply; 7+ messages in thread
From: Ben Greear @ 2012-03-12 18:17 UTC (permalink / raw)
  To: e1000-devel list, Jiri Pirko; +Cc: netdev

[-- Attachment #1: Type: text/plain, Size: 3605 bytes --]

This commit breaks 802.1Q VLANs on my e1000 NIC, and it remains broken in top-of-tree
(as of 2 days ago, at least).

commit 5622e4044a916de1af84bfcc4d437ce0c799d531
Author: Jiri Pirko <jpirko@redhat.com>
Date:   Thu Jul 21 03:26:31 2011 +0000

     e1000: do vlan cleanup

     - unify vlan and nonvlan rx path
     - kill adapter->vlgrp and e1000_vlan_rx_register
     - allow to turn on/off rx/tx vlan accel via ethtool (set_features)

     Signed-off-by: Jiri Pirko <jpirko@redhat.com>
     Signed-off-by: David S. Miller <davem@davemloft.net>


My OS is RH 8, 32-bit.

My e1000 hardware is:

05:01.0 Ethernet controller: Intel Corporation 82546GB Gigabit Ethernet Controller (rev 03)
	Subsystem: Intel Corporation PRO/1000 MT Dual Port Server Adapter
	Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV+ VGASnoop- ParErr+ Stepping- SERR+ FastB2B- DisINTx-
	Status: Cap+ 66MHz+ UDF- FastB2B- ParErr- DEVSEL=medium >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx-
	Latency: 52 (63750ns min), Cache Line Size: 32 bytes
	Interrupt: pin A routed to IRQ 24
	Region 0: Memory at d8100000 (64-bit, non-prefetchable) [size=128K]
	Region 4: I/O ports at 3000 [size=64]
	Capabilities: [dc] Power Management version 2
		Flags: PMEClk- DSI+ D1- D2- AuxCurrent=0mA PME(D0+,D1-,D2-,D3hot+,D3cold+)
		Status: D0 PME-Enable- DSel=0 DScale=1 PME-
	Capabilities: [e4] PCI-X non-bridge device
		Command: DPERE- ERO+ RBC=512 OST=1
		Status: Dev=05:01.0 64bit+ 133MHz+ SCD- USC- DC=simple DMMRBC=2048 DMOST=1 DMCRS=16 RSCEM- 266MHz- 533MHz-
	Capabilities: [f0] Message Signalled Interrupts: Mask- 64bit+ Queue=0/0 Enable-
		Address: 0000000000000000  Data: 0000
	Kernel driver in use: e1000
	Kernel modules: e1000

05:01.1 Ethernet controller: Intel Corporation 82546GB Gigabit Ethernet Controller (rev 03)
	Subsystem: Intel Corporation PRO/1000 MT Dual Port Server Adapter
	Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV+ VGASnoop- ParErr+ Stepping- SERR+ FastB2B- DisINTx-
	Status: Cap+ 66MHz+ UDF- FastB2B- ParErr- DEVSEL=medium >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx-
	Latency: 52 (63750ns min), Cache Line Size: 32 bytes
	Interrupt: pin B routed to IRQ 25
	Region 0: Memory at d8120000 (64-bit, non-prefetchable) [size=128K]
	Region 4: I/O ports at 3040 [size=64]
	Capabilities: [dc] Power Management version 2
		Flags: PMEClk- DSI+ D1- D2- AuxCurrent=0mA PME(D0+,D1-,D2-,D3hot+,D3cold+)
		Status: D0 PME-Enable- DSel=0 DScale=1 PME-
	Capabilities: [e4] PCI-X non-bridge device
		Command: DPERE- ERO+ RBC=512 OST=1
		Status: Dev=05:01.1 64bit+ 133MHz+ SCD- USC- DC=simple DMMRBC=2048 DMOST=1 DMCRS=16 RSCEM- 266MHz- 533MHz-
	Capabilities: [f0] Message Signalled Interrupts: Mask- 64bit+ Queue=0/0 Enable-
		Address: 0000000000000000  Data: 0000
	Kernel driver in use: e1000
	Kernel modules: e1000


I'm attaching ethregs output from good and bad kernels, in case that helps.

The symptom is that pkts show up as 'ISL' instead of Ethernet frames.  I'm
including a capture of that as well.  I am sending UDP pkts from one NIC to the other,
on vlan 8, though because ARP fails, no UDP is ever sent.  The sniff shows rcvd pkts
as looking funky, but it could be they are transmitted badly once they hit the NIC,
so I don't know if it's actually rx or tx (or both) logic that is busted at this point.

I'm using routing rules to make send-to-self work, but the kernel is un-patched.

Please let me know if you need additional info.  I'll be happy to test patches,
I have a good setup to reproduce this easily.

Thanks,
Ben

-- 
Ben Greear <greearb@candelatech.com>
Candela Technologies Inc  http://www.candelatech.com


[-- Attachment #2: ethregs-e1000-bad-vlans.txt.gz --]
[-- Type: application/x-gzip, Size: 5997 bytes --]

[-- Attachment #3: ethregs-e1000-good-vlans.txt.gz --]
[-- Type: application/x-gzip, Size: 6038 bytes --]

[-- Attachment #4: e1000-bad-vlans.pcap --]
[-- Type: application/x-pcap, Size: 170 bytes --]

^ permalink raw reply	[flat|nested] 7+ messages in thread
* Re: VLAN regression caused by: e1000: do vlan cleanup (799d531).
  2012-03-12 18:17 VLAN regression caused by: e1000: do vlan cleanup (799d531) Ben Greear
@ 2012-03-13  9:59 ` Jiri Pirko
  2012-03-13 16:18   ` Ben Greear
  2012-03-13 18:44   ` Ben Greear
  0 siblings, 2 replies; 7+ messages in thread
From: Jiri Pirko @ 2012-03-13  9:59 UTC (permalink / raw)
  To: Ben Greear; +Cc: e1000-devel list, netdev

Mon, Mar 12, 2012 at 07:17:51PM CET, greearb@candelatech.com wrote:
>This commit breaks 802.1Q VLANs on my e1000 NIC, and it remains broken in top-of-tree
>(as of 2 days ago, at least).
>
>commit 5622e4044a916de1af84bfcc4d437ce0c799d531
>Author: Jiri Pirko <jpirko@redhat.com>
>Date:   Thu Jul 21 03:26:31 2011 +0000
>
>    e1000: do vlan cleanup
>
>    - unify vlan and nonvlan rx path
>    - kill adapter->vlgrp and e1000_vlan_rx_register
>    - allow to turn on/off rx/tx vlan accel via ethtool (set_features)
>
>    Signed-off-by: Jiri Pirko <jpirko@redhat.com>
>    Signed-off-by: David S. Miller <davem@davemloft.net>
>
>
>My OS is RH 8, 32-bit.
>
>My e1000 hardware is:
>
>05:01.0 Ethernet controller: Intel Corporation 82546GB Gigabit Ethernet Controller (rev 03)
>	Subsystem: Intel Corporation PRO/1000 MT Dual Port Server Adapter
>	Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV+ VGASnoop- ParErr+ Stepping- SERR+ FastB2B- DisINTx-
>	Status: Cap+ 66MHz+ UDF- FastB2B- ParErr- DEVSEL=medium >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx-
>	Latency: 52 (63750ns min), Cache Line Size: 32 bytes
>	Interrupt: pin A routed to IRQ 24
>	Region 0: Memory at d8100000 (64-bit, non-prefetchable) [size=128K]
>	Region 4: I/O ports at 3000 [size=64]
>	Capabilities: [dc] Power Management version 2
>		Flags: PMEClk- DSI+ D1- D2- AuxCurrent=0mA PME(D0+,D1-,D2-,D3hot+,D3cold+)
>		Status: D0 PME-Enable- DSel=0 DScale=1 PME-
>	Capabilities: [e4] PCI-X non-bridge device
>		Command: DPERE- ERO+ RBC=512 OST=1
>		Status: Dev=05:01.0 64bit+ 133MHz+ SCD- USC- DC=simple DMMRBC=2048 DMOST=1 DMCRS=16 RSCEM- 266MHz- 533MHz-
>	Capabilities: [f0] Message Signalled Interrupts: Mask- 64bit+ Queue=0/0 Enable-
>		Address: 0000000000000000  Data: 0000
>	Kernel driver in use: e1000
>	Kernel modules: e1000
>
>05:01.1 Ethernet controller: Intel Corporation 82546GB Gigabit Ethernet Controller (rev 03)
>	Subsystem: Intel Corporation PRO/1000 MT Dual Port Server Adapter
>	Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV+ VGASnoop- ParErr+ Stepping- SERR+ FastB2B- DisINTx-
>	Status: Cap+ 66MHz+ UDF- FastB2B- ParErr- DEVSEL=medium >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx-
>	Latency: 52 (63750ns min), Cache Line Size: 32 bytes
>	Interrupt: pin B routed to IRQ 25
>	Region 0: Memory at d8120000 (64-bit, non-prefetchable) [size=128K]
>	Region 4: I/O ports at 3040 [size=64]
>	Capabilities: [dc] Power Management version 2
>		Flags: PMEClk- DSI+ D1- D2- AuxCurrent=0mA PME(D0+,D1-,D2-,D3hot+,D3cold+)
>		Status: D0 PME-Enable- DSel=0 DScale=1 PME-
>	Capabilities: [e4] PCI-X non-bridge device
>		Command: DPERE- ERO+ RBC=512 OST=1
>		Status: Dev=05:01.1 64bit+ 133MHz+ SCD- USC- DC=simple DMMRBC=2048 DMOST=1 DMCRS=16 RSCEM- 266MHz- 533MHz-
>	Capabilities: [f0] Message Signalled Interrupts: Mask- 64bit+ Queue=0/0 Enable-
>		Address: 0000000000000000  Data: 0000
>	Kernel driver in use: e1000
>	Kernel modules: e1000
>
>
>I'm attaching ethregs output from good and bad kernels, in case that helps.
>
>The symptom is that pkts show up as 'ISL' instead of Ethernet frames.  I'm
>including a capture of that as well.  I am sending UDP pkts from one NIC to the other,
>on vlan 8, though because ARP fails, no UDP is ever sent.  The sniff shows rcvd pkts
>as looking funky, but it could be they are transmitted badly once they hit the NIC,
>so I don't know if it's actually rx or tx (or both) logic that is busted at this point.
>
>I'm using routing rules to make send-to-self work, but the kernel is un-patched.
>
>Please let me know if you need additional info.  I'll be happy to test patches,
>I have a good setup to reproduce this easily.

Thanks for the report Ben.

It would help it you try to enable/disable vlan hw accel by ethtool to
see if it helps.

I have suspition that vlan filter must be on/offed along with vlan
accel.

Thanks.

Jirka

>
>Thanks,
>Ben
>
>-- 
>Ben Greear <greearb@candelatech.com>
>Candela Technologies Inc  http://www.candelatech.com
>

^ permalink raw reply	[flat|nested] 7+ messages in thread
* Re: VLAN regression caused by: e1000: do vlan cleanup (799d531).
  2012-03-13  9:59 ` Jiri Pirko
@ 2012-03-13 16:18   ` Ben Greear
  2012-03-13 18:44   ` Ben Greear
  1 sibling, 0 replies; 7+ messages in thread
From: Ben Greear @ 2012-03-13 16:18 UTC (permalink / raw)
  To: Jiri Pirko; +Cc: e1000-devel list, netdev

On 03/13/2012 02:59 AM, Jiri Pirko wrote:
> Mon, Mar 12, 2012 at 07:17:51PM CET, greearb@candelatech.com wrote:

>> Please let me know if you need additional info.  I'll be happy to test patches,
>> I have a good setup to reproduce this easily.
>
> Thanks for the report Ben.
>
> It would help it you try to enable/disable vlan hw accel by ethtool to
> see if it helps.
>
> I have suspition that vlan filter must be on/offed along with vlan
> accel.

Can that even be done in the 3.0 kernel?  Or do you want me to just
try the 3.3-rxX kernel?

Ben

-- 
Ben Greear <greearb@candelatech.com>
Candela Technologies Inc  http://www.candelatech.com


------------------------------------------------------------------------------
Keep Your Developer Skills Current with LearnDevNow!
The most comprehensive online learning library for Microsoft developers
is just $99.99! Visual Studio, SharePoint, SQL - plus HTML5, CSS3, MVC3,
Metro Style Apps, more. Free future releases when you subscribe now!
http://p.sf.net/sfu/learndevnow-d2d
_______________________________________________
E1000-devel mailing list
E1000-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/e1000-devel
To learn more about Intel&#174; Ethernet, visit http://communities.intel.com/community/wired

^ permalink raw reply	[flat|nested] 7+ messages in thread
* Re: VLAN regression caused by: e1000: do vlan cleanup (799d531).
  2012-03-13  9:59 ` Jiri Pirko
  2012-03-13 16:18   ` Ben Greear
@ 2012-03-13 18:44   ` Ben Greear
  2012-03-19 13:08     ` Jiri Pirko
  1 sibling, 1 reply; 7+ messages in thread
From: Ben Greear @ 2012-03-13 18:44 UTC (permalink / raw)
  To: Jiri Pirko; +Cc: e1000-devel list, netdev

On 03/13/2012 02:59 AM, Jiri Pirko wrote:
> Mon, Mar 12, 2012 at 07:17:51PM CET, greearb@candelatech.com wrote:
>> This commit breaks 802.1Q VLANs on my e1000 NIC, and it remains broken in top-of-tree
>> (as of 2 days ago, at least).
>>
>> commit 5622e4044a916de1af84bfcc4d437ce0c799d531
>> Author: Jiri Pirko<jpirko@redhat.com>
>> Date:   Thu Jul 21 03:26:31 2011 +0000
>>
>>     e1000: do vlan cleanup
>>
>>     - unify vlan and nonvlan rx path
>>     - kill adapter->vlgrp and e1000_vlan_rx_register
>>     - allow to turn on/off rx/tx vlan accel via ethtool (set_features)
>>
>>     Signed-off-by: Jiri Pirko<jpirko@redhat.com>
>>     Signed-off-by: David S. Miller<davem@davemloft.net>
>>
>>
>> My OS is RH 8, 32-bit.
>>
>> My e1000 hardware is:
>>
>> 05:01.0 Ethernet controller: Intel Corporation 82546GB Gigabit Ethernet Controller (rev 03)
>> 	Subsystem: Intel Corporation PRO/1000 MT Dual Port Server Adapter
>> 	Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV+ VGASnoop- ParErr+ Stepping- SERR+ FastB2B- DisINTx-
>> 	Status: Cap+ 66MHz+ UDF- FastB2B- ParErr- DEVSEL=medium>TAbort-<TAbort-<MAbort->SERR-<PERR- INTx-
>> 	Latency: 52 (63750ns min), Cache Line Size: 32 bytes
>> 	Interrupt: pin A routed to IRQ 24
>> 	Region 0: Memory at d8100000 (64-bit, non-prefetchable) [size=128K]
>> 	Region 4: I/O ports at 3000 [size=64]
>> 	Capabilities: [dc] Power Management version 2
>> 		Flags: PMEClk- DSI+ D1- D2- AuxCurrent=0mA PME(D0+,D1-,D2-,D3hot+,D3cold+)
>> 		Status: D0 PME-Enable- DSel=0 DScale=1 PME-
>> 	Capabilities: [e4] PCI-X non-bridge device
>> 		Command: DPERE- ERO+ RBC=512 OST=1
>> 		Status: Dev=05:01.0 64bit+ 133MHz+ SCD- USC- DC=simple DMMRBC=2048 DMOST=1 DMCRS=16 RSCEM- 266MHz- 533MHz-
>> 	Capabilities: [f0] Message Signalled Interrupts: Mask- 64bit+ Queue=0/0 Enable-
>> 		Address: 0000000000000000  Data: 0000
>> 	Kernel driver in use: e1000
>> 	Kernel modules: e1000
>>
>> 05:01.1 Ethernet controller: Intel Corporation 82546GB Gigabit Ethernet Controller (rev 03)
>> 	Subsystem: Intel Corporation PRO/1000 MT Dual Port Server Adapter
>> 	Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV+ VGASnoop- ParErr+ Stepping- SERR+ FastB2B- DisINTx-
>> 	Status: Cap+ 66MHz+ UDF- FastB2B- ParErr- DEVSEL=medium>TAbort-<TAbort-<MAbort->SERR-<PERR- INTx-
>> 	Latency: 52 (63750ns min), Cache Line Size: 32 bytes
>> 	Interrupt: pin B routed to IRQ 25
>> 	Region 0: Memory at d8120000 (64-bit, non-prefetchable) [size=128K]
>> 	Region 4: I/O ports at 3040 [size=64]
>> 	Capabilities: [dc] Power Management version 2
>> 		Flags: PMEClk- DSI+ D1- D2- AuxCurrent=0mA PME(D0+,D1-,D2-,D3hot+,D3cold+)
>> 		Status: D0 PME-Enable- DSel=0 DScale=1 PME-
>> 	Capabilities: [e4] PCI-X non-bridge device
>> 		Command: DPERE- ERO+ RBC=512 OST=1
>> 		Status: Dev=05:01.1 64bit+ 133MHz+ SCD- USC- DC=simple DMMRBC=2048 DMOST=1 DMCRS=16 RSCEM- 266MHz- 533MHz-
>> 	Capabilities: [f0] Message Signalled Interrupts: Mask- 64bit+ Queue=0/0 Enable-
>> 		Address: 0000000000000000  Data: 0000
>> 	Kernel driver in use: e1000
>> 	Kernel modules: e1000
>>
>>
>> I'm attaching ethregs output from good and bad kernels, in case that helps.
>>
>> The symptom is that pkts show up as 'ISL' instead of Ethernet frames.  I'm
>> including a capture of that as well.  I am sending UDP pkts from one NIC to the other,
>> on vlan 8, though because ARP fails, no UDP is ever sent.  The sniff shows rcvd pkts
>> as looking funky, but it could be they are transmitted badly once they hit the NIC,
>> so I don't know if it's actually rx or tx (or both) logic that is busted at this point.
>>
>> I'm using routing rules to make send-to-self work, but the kernel is un-patched.
>>
>> Please let me know if you need additional info.  I'll be happy to test patches,
>> I have a good setup to reproduce this easily.
>
> Thanks for the report Ben.
>
> It would help it you try to enable/disable vlan hw accel by ethtool to
> see if it helps.
>
> I have suspition that vlan filter must be on/offed along with vlan
> accel.

Ok, if I disable rx-vlan-hw-parse on both NICs, traffic starts working.

And, if I re-enable rx-vlan-hw-parse on both NICs, it continues to work.

Maybe the initialization logic needs some fixing?

This is on 3.3.0-rc7+.

Thanks,
Ben


-- 
Ben Greear <greearb@candelatech.com>
Candela Technologies Inc  http://www.candelatech.com


------------------------------------------------------------------------------
Keep Your Developer Skills Current with LearnDevNow!
The most comprehensive online learning library for Microsoft developers
is just $99.99! Visual Studio, SharePoint, SQL - plus HTML5, CSS3, MVC3,
Metro Style Apps, more. Free future releases when you subscribe now!
http://p.sf.net/sfu/learndevnow-d2d
_______________________________________________
E1000-devel mailing list
E1000-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/e1000-devel
To learn more about Intel&#174; Ethernet, visit http://communities.intel.com/community/wired

^ permalink raw reply	[flat|nested] 7+ messages in thread
* Re: VLAN regression caused by: e1000: do vlan cleanup (799d531).
  2012-03-13 18:44   ` Ben Greear
@ 2012-03-19 13:08     ` Jiri Pirko
  2012-03-19 23:24       ` Ben Greear
  0 siblings, 1 reply; 7+ messages in thread
From: Jiri Pirko @ 2012-03-19 13:08 UTC (permalink / raw)
  To: Ben Greear; +Cc: e1000-devel list, netdev

Tue, Mar 13, 2012 at 07:44:18PM CET, greearb@candelatech.com wrote:
>On 03/13/2012 02:59 AM, Jiri Pirko wrote:
>>Mon, Mar 12, 2012 at 07:17:51PM CET, greearb@candelatech.com wrote:
>>>This commit breaks 802.1Q VLANs on my e1000 NIC, and it remains broken in top-of-tree
>>>(as of 2 days ago, at least).
>>>
>>>commit 5622e4044a916de1af84bfcc4d437ce0c799d531
>>>Author: Jiri Pirko<jpirko@redhat.com>
>>>Date:   Thu Jul 21 03:26:31 2011 +0000
>>>
>>>    e1000: do vlan cleanup
>>>
>>>    - unify vlan and nonvlan rx path
>>>    - kill adapter->vlgrp and e1000_vlan_rx_register
>>>    - allow to turn on/off rx/tx vlan accel via ethtool (set_features)
>>>
>>>    Signed-off-by: Jiri Pirko<jpirko@redhat.com>
>>>    Signed-off-by: David S. Miller<davem@davemloft.net>
>>>
>>>
>>>My OS is RH 8, 32-bit.
>>>
>>>My e1000 hardware is:
>>>
>>>05:01.0 Ethernet controller: Intel Corporation 82546GB Gigabit Ethernet Controller (rev 03)
>>>	Subsystem: Intel Corporation PRO/1000 MT Dual Port Server Adapter
>>>	Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV+ VGASnoop- ParErr+ Stepping- SERR+ FastB2B- DisINTx-
>>>	Status: Cap+ 66MHz+ UDF- FastB2B- ParErr- DEVSEL=medium>TAbort-<TAbort-<MAbort->SERR-<PERR- INTx-
>>>	Latency: 52 (63750ns min), Cache Line Size: 32 bytes
>>>	Interrupt: pin A routed to IRQ 24
>>>	Region 0: Memory at d8100000 (64-bit, non-prefetchable) [size=128K]
>>>	Region 4: I/O ports at 3000 [size=64]
>>>	Capabilities: [dc] Power Management version 2
>>>		Flags: PMEClk- DSI+ D1- D2- AuxCurrent=0mA PME(D0+,D1-,D2-,D3hot+,D3cold+)
>>>		Status: D0 PME-Enable- DSel=0 DScale=1 PME-
>>>	Capabilities: [e4] PCI-X non-bridge device
>>>		Command: DPERE- ERO+ RBC=512 OST=1
>>>		Status: Dev=05:01.0 64bit+ 133MHz+ SCD- USC- DC=simple DMMRBC=2048 DMOST=1 DMCRS=16 RSCEM- 266MHz- 533MHz-
>>>	Capabilities: [f0] Message Signalled Interrupts: Mask- 64bit+ Queue=0/0 Enable-
>>>		Address: 0000000000000000  Data: 0000
>>>	Kernel driver in use: e1000
>>>	Kernel modules: e1000
>>>
>>>05:01.1 Ethernet controller: Intel Corporation 82546GB Gigabit Ethernet Controller (rev 03)
>>>	Subsystem: Intel Corporation PRO/1000 MT Dual Port Server Adapter
>>>	Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV+ VGASnoop- ParErr+ Stepping- SERR+ FastB2B- DisINTx-
>>>	Status: Cap+ 66MHz+ UDF- FastB2B- ParErr- DEVSEL=medium>TAbort-<TAbort-<MAbort->SERR-<PERR- INTx-
>>>	Latency: 52 (63750ns min), Cache Line Size: 32 bytes
>>>	Interrupt: pin B routed to IRQ 25
>>>	Region 0: Memory at d8120000 (64-bit, non-prefetchable) [size=128K]
>>>	Region 4: I/O ports at 3040 [size=64]
>>>	Capabilities: [dc] Power Management version 2
>>>		Flags: PMEClk- DSI+ D1- D2- AuxCurrent=0mA PME(D0+,D1-,D2-,D3hot+,D3cold+)
>>>		Status: D0 PME-Enable- DSel=0 DScale=1 PME-
>>>	Capabilities: [e4] PCI-X non-bridge device
>>>		Command: DPERE- ERO+ RBC=512 OST=1
>>>		Status: Dev=05:01.1 64bit+ 133MHz+ SCD- USC- DC=simple DMMRBC=2048 DMOST=1 DMCRS=16 RSCEM- 266MHz- 533MHz-
>>>	Capabilities: [f0] Message Signalled Interrupts: Mask- 64bit+ Queue=0/0 Enable-
>>>		Address: 0000000000000000  Data: 0000
>>>	Kernel driver in use: e1000
>>>	Kernel modules: e1000
>>>
>>>
>>>I'm attaching ethregs output from good and bad kernels, in case that helps.
>>>
>>>The symptom is that pkts show up as 'ISL' instead of Ethernet frames.  I'm
>>>including a capture of that as well.  I am sending UDP pkts from one NIC to the other,
>>>on vlan 8, though because ARP fails, no UDP is ever sent.  The sniff shows rcvd pkts
>>>as looking funky, but it could be they are transmitted badly once they hit the NIC,
>>>so I don't know if it's actually rx or tx (or both) logic that is busted at this point.
>>>
>>>I'm using routing rules to make send-to-self work, but the kernel is un-patched.
>>>
>>>Please let me know if you need additional info.  I'll be happy to test patches,
>>>I have a good setup to reproduce this easily.
>>
>>Thanks for the report Ben.
>>
>>It would help it you try to enable/disable vlan hw accel by ethtool to
>>see if it helps.
>>
>>I have suspition that vlan filter must be on/offed along with vlan
>>accel.
>
>Ok, if I disable rx-vlan-hw-parse on both NICs, traffic starts working.
>
>And, if I re-enable rx-vlan-hw-parse on both NICs, it continues to work.
>
>Maybe the initialization logic needs some fixing?

Could you please try following patch?

diff --git a/drivers/net/ethernet/intel/e1000/e1000_main.c b/drivers/net/ethernet/intel/e1000/e1000_main.c
index 6419a88..b06c31c 100644
--- a/drivers/net/ethernet/intel/e1000/e1000_main.c
+++ b/drivers/net/ethernet/intel/e1000/e1000_main.c
@@ -164,6 +164,8 @@ static int e1000_82547_fifo_workaround(struct e1000_adapter *adapter,
 static bool e1000_vlan_used(struct e1000_adapter *adapter);
 static void e1000_vlan_mode(struct net_device *netdev,
 			    netdev_features_t features);
+static void e1000_vlan_filter_on_off(struct e1000_adapter *adapter,
+				     bool filter_on);
 static int e1000_vlan_rx_add_vid(struct net_device *netdev, u16 vid);
 static int e1000_vlan_rx_kill_vid(struct net_device *netdev, u16 vid);
 static void e1000_restore_vlan(struct e1000_adapter *adapter);
@@ -1214,7 +1216,7 @@ static int __devinit e1000_probe(struct pci_dev *pdev,
 	if (err)
 		goto err_register;
 
-	e1000_vlan_mode(netdev, netdev->features);
+	e1000_vlan_filter_on_off(adapter, false);
 
 	/* print bus type/speed/width info */
 	e_info(probe, "(PCI%s:%dMHz:%d-bit) %pM\n",
@@ -4772,6 +4774,22 @@ static bool e1000_vlan_used(struct e1000_adapter *adapter)
 	return false;
 }
 
+static void __e1000_vlan_mode(struct e1000_adapter *adapter,
+			      netdev_features_t features)
+{
+	struct e1000_hw *hw = &adapter->hw;
+	u32 ctrl;
+
+	ctrl = er32(CTRL);
+	if (features & NETIF_F_HW_VLAN_RX) {
+		/* enable VLAN tag insert/strip */
+		ctrl |= E1000_CTRL_VME;
+	} else {
+		/* disable VLAN tag insert/strip */
+		ctrl &= ~E1000_CTRL_VME;
+	}
+	ew32(CTRL, ctrl);
+}
 static void e1000_vlan_filter_on_off(struct e1000_adapter *adapter,
 				     bool filter_on)
 {
@@ -4781,6 +4799,7 @@ static void e1000_vlan_filter_on_off(struct e1000_adapter *adapter,
 	if (!test_bit(__E1000_DOWN, &adapter->flags))
 		e1000_irq_disable(adapter);
 
+	__e1000_vlan_mode(adapter, adapter->netdev->features);
 	if (filter_on) {
 		/* enable VLAN receive filtering */
 		rctl = er32(RCTL);
@@ -4801,24 +4820,14 @@ static void e1000_vlan_filter_on_off(struct e1000_adapter *adapter,
 }
 
 static void e1000_vlan_mode(struct net_device *netdev,
-	netdev_features_t features)
+			    netdev_features_t features)
 {
 	struct e1000_adapter *adapter = netdev_priv(netdev);
-	struct e1000_hw *hw = &adapter->hw;
-	u32 ctrl;
 
 	if (!test_bit(__E1000_DOWN, &adapter->flags))
 		e1000_irq_disable(adapter);
 
-	ctrl = er32(CTRL);
-	if (features & NETIF_F_HW_VLAN_RX) {
-		/* enable VLAN tag insert/strip */
-		ctrl |= E1000_CTRL_VME;
-	} else {
-		/* disable VLAN tag insert/strip */
-		ctrl &= ~E1000_CTRL_VME;
-	}
-	ew32(CTRL, ctrl);
+	__e1000_vlan_mode(adapter, features);
 
 	if (!test_bit(__E1000_DOWN, &adapter->flags))
 		e1000_irq_enable(adapter);

------------------------------------------------------------------------------
This SF email is sponsosred by:
Try Windows Azure free for 90 days Click Here 
http://p.sf.net/sfu/sfd2d-msazure
_______________________________________________
E1000-devel mailing list
E1000-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/e1000-devel
To learn more about Intel&#174; Ethernet, visit http://communities.intel.com/community/wired

^ permalink raw reply related	[flat|nested] 7+ messages in thread
* Re: VLAN regression caused by: e1000: do vlan cleanup (799d531).
  2012-03-19 13:08     ` Jiri Pirko
@ 2012-03-19 23:24       ` Ben Greear
  2012-03-20  6:26         ` Jiri Pirko
  0 siblings, 1 reply; 7+ messages in thread
From: Ben Greear @ 2012-03-19 23:24 UTC (permalink / raw)
  To: Jiri Pirko; +Cc: e1000-devel list, netdev

On 03/19/2012 06:08 AM, Jiri Pirko wrote:
> Tue, Mar 13, 2012 at 07:44:18PM CET, greearb@candelatech.com wrote:

>>> I have suspition that vlan filter must be on/offed along with vlan
>>> accel.
>>
>> Ok, if I disable rx-vlan-hw-parse on both NICs, traffic starts working.
>>
>> And, if I re-enable rx-vlan-hw-parse on both NICs, it continues to work.
>>
>> Maybe the initialization logic needs some fixing?
>
> Could you please try following patch?

That patches fixes the problem for me.  Only lightly tested
at this point, but I have verified VLAN traffic works as desired.

Thanks,
Ben

>
> diff --git a/drivers/net/ethernet/intel/e1000/e1000_main.c b/drivers/net/ethernet/intel/e1000/e1000_main.c
> index 6419a88..b06c31c 100644
> --- a/drivers/net/ethernet/intel/e1000/e1000_main.c
> +++ b/drivers/net/ethernet/intel/e1000/e1000_main.c
> @@ -164,6 +164,8 @@ static int e1000_82547_fifo_workaround(struct e1000_adapter *adapter,
>   static bool e1000_vlan_used(struct e1000_adapter *adapter);
>   static void e1000_vlan_mode(struct net_device *netdev,
>   			    netdev_features_t features);
> +static void e1000_vlan_filter_on_off(struct e1000_adapter *adapter,
> +				     bool filter_on);
>   static int e1000_vlan_rx_add_vid(struct net_device *netdev, u16 vid);
>   static int e1000_vlan_rx_kill_vid(struct net_device *netdev, u16 vid);
>   static void e1000_restore_vlan(struct e1000_adapter *adapter);
> @@ -1214,7 +1216,7 @@ static int __devinit e1000_probe(struct pci_dev *pdev,
>   	if (err)
>   		goto err_register;
>
> -	e1000_vlan_mode(netdev, netdev->features);
> +	e1000_vlan_filter_on_off(adapter, false);
>
>   	/* print bus type/speed/width info */
>   	e_info(probe, "(PCI%s:%dMHz:%d-bit) %pM\n",
> @@ -4772,6 +4774,22 @@ static bool e1000_vlan_used(struct e1000_adapter *adapter)
>   	return false;
>   }
>
> +static void __e1000_vlan_mode(struct e1000_adapter *adapter,
> +			      netdev_features_t features)
> +{
> +	struct e1000_hw *hw =&adapter->hw;
> +	u32 ctrl;
> +
> +	ctrl = er32(CTRL);
> +	if (features&  NETIF_F_HW_VLAN_RX) {
> +		/* enable VLAN tag insert/strip */
> +		ctrl |= E1000_CTRL_VME;
> +	} else {
> +		/* disable VLAN tag insert/strip */
> +		ctrl&= ~E1000_CTRL_VME;
> +	}
> +	ew32(CTRL, ctrl);
> +}
>   static void e1000_vlan_filter_on_off(struct e1000_adapter *adapter,
>   				     bool filter_on)
>   {
> @@ -4781,6 +4799,7 @@ static void e1000_vlan_filter_on_off(struct e1000_adapter *adapter,
>   	if (!test_bit(__E1000_DOWN,&adapter->flags))
>   		e1000_irq_disable(adapter);
>
> +	__e1000_vlan_mode(adapter, adapter->netdev->features);
>   	if (filter_on) {
>   		/* enable VLAN receive filtering */
>   		rctl = er32(RCTL);
> @@ -4801,24 +4820,14 @@ static void e1000_vlan_filter_on_off(struct e1000_adapter *adapter,
>   }
>
>   static void e1000_vlan_mode(struct net_device *netdev,
> -	netdev_features_t features)
> +			    netdev_features_t features)
>   {
>   	struct e1000_adapter *adapter = netdev_priv(netdev);
> -	struct e1000_hw *hw =&adapter->hw;
> -	u32 ctrl;
>
>   	if (!test_bit(__E1000_DOWN,&adapter->flags))
>   		e1000_irq_disable(adapter);
>
> -	ctrl = er32(CTRL);
> -	if (features&  NETIF_F_HW_VLAN_RX) {
> -		/* enable VLAN tag insert/strip */
> -		ctrl |= E1000_CTRL_VME;
> -	} else {
> -		/* disable VLAN tag insert/strip */
> -		ctrl&= ~E1000_CTRL_VME;
> -	}
> -	ew32(CTRL, ctrl);
> +	__e1000_vlan_mode(adapter, features);
>
>   	if (!test_bit(__E1000_DOWN,&adapter->flags))
>   		e1000_irq_enable(adapter);


-- 
Ben Greear <greearb@candelatech.com>
Candela Technologies Inc  http://www.candelatech.com


------------------------------------------------------------------------------
This SF email is sponsosred by:
Try Windows Azure free for 90 days Click Here 
http://p.sf.net/sfu/sfd2d-msazure
_______________________________________________
E1000-devel mailing list
E1000-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/e1000-devel
To learn more about Intel&#174; Ethernet, visit http://communities.intel.com/community/wired

^ permalink raw reply	[flat|nested] 7+ messages in thread
* Re: VLAN regression caused by: e1000: do vlan cleanup (799d531).
  2012-03-19 23:24       ` Ben Greear
@ 2012-03-20  6:26         ` Jiri Pirko
  0 siblings, 0 replies; 7+ messages in thread
From: Jiri Pirko @ 2012-03-20  6:26 UTC (permalink / raw)
  To: Ben Greear; +Cc: e1000-devel list, netdev

Tue, Mar 20, 2012 at 12:24:04AM CET, greearb@candelatech.com wrote:
>On 03/19/2012 06:08 AM, Jiri Pirko wrote:
>>Tue, Mar 13, 2012 at 07:44:18PM CET, greearb@candelatech.com wrote:
>
>>>>I have suspition that vlan filter must be on/offed along with vlan
>>>>accel.
>>>
>>>Ok, if I disable rx-vlan-hw-parse on both NICs, traffic starts working.
>>>
>>>And, if I re-enable rx-vlan-hw-parse on both NICs, it continues to work.
>>>
>>>Maybe the initialization logic needs some fixing?
>>
>>Could you please try following patch?
>
>That patches fixes the problem for me.  Only lightly tested
>at this point, but I have verified VLAN traffic works as desired.

Thank you. I'm going to do proper post of this.

Jirka

>
>Thanks,
>Ben
>
>>
>>diff --git a/drivers/net/ethernet/intel/e1000/e1000_main.c b/drivers/net/ethernet/intel/e1000/e1000_main.c
>>index 6419a88..b06c31c 100644
>>--- a/drivers/net/ethernet/intel/e1000/e1000_main.c
>>+++ b/drivers/net/ethernet/intel/e1000/e1000_main.c
>>@@ -164,6 +164,8 @@ static int e1000_82547_fifo_workaround(struct e1000_adapter *adapter,
>>  static bool e1000_vlan_used(struct e1000_adapter *adapter);
>>  static void e1000_vlan_mode(struct net_device *netdev,
>>  			    netdev_features_t features);
>>+static void e1000_vlan_filter_on_off(struct e1000_adapter *adapter,
>>+				     bool filter_on);
>>  static int e1000_vlan_rx_add_vid(struct net_device *netdev, u16 vid);
>>  static int e1000_vlan_rx_kill_vid(struct net_device *netdev, u16 vid);
>>  static void e1000_restore_vlan(struct e1000_adapter *adapter);
>>@@ -1214,7 +1216,7 @@ static int __devinit e1000_probe(struct pci_dev *pdev,
>>  	if (err)
>>  		goto err_register;
>>
>>-	e1000_vlan_mode(netdev, netdev->features);
>>+	e1000_vlan_filter_on_off(adapter, false);
>>
>>  	/* print bus type/speed/width info */
>>  	e_info(probe, "(PCI%s:%dMHz:%d-bit) %pM\n",
>>@@ -4772,6 +4774,22 @@ static bool e1000_vlan_used(struct e1000_adapter *adapter)
>>  	return false;
>>  }
>>
>>+static void __e1000_vlan_mode(struct e1000_adapter *adapter,
>>+			      netdev_features_t features)
>>+{
>>+	struct e1000_hw *hw =&adapter->hw;
>>+	u32 ctrl;
>>+
>>+	ctrl = er32(CTRL);
>>+	if (features&  NETIF_F_HW_VLAN_RX) {
>>+		/* enable VLAN tag insert/strip */
>>+		ctrl |= E1000_CTRL_VME;
>>+	} else {
>>+		/* disable VLAN tag insert/strip */
>>+		ctrl&= ~E1000_CTRL_VME;
>>+	}
>>+	ew32(CTRL, ctrl);
>>+}
>>  static void e1000_vlan_filter_on_off(struct e1000_adapter *adapter,
>>  				     bool filter_on)
>>  {
>>@@ -4781,6 +4799,7 @@ static void e1000_vlan_filter_on_off(struct e1000_adapter *adapter,
>>  	if (!test_bit(__E1000_DOWN,&adapter->flags))
>>  		e1000_irq_disable(adapter);
>>
>>+	__e1000_vlan_mode(adapter, adapter->netdev->features);
>>  	if (filter_on) {
>>  		/* enable VLAN receive filtering */
>>  		rctl = er32(RCTL);
>>@@ -4801,24 +4820,14 @@ static void e1000_vlan_filter_on_off(struct e1000_adapter *adapter,
>>  }
>>
>>  static void e1000_vlan_mode(struct net_device *netdev,
>>-	netdev_features_t features)
>>+			    netdev_features_t features)
>>  {
>>  	struct e1000_adapter *adapter = netdev_priv(netdev);
>>-	struct e1000_hw *hw =&adapter->hw;
>>-	u32 ctrl;
>>
>>  	if (!test_bit(__E1000_DOWN,&adapter->flags))
>>  		e1000_irq_disable(adapter);
>>
>>-	ctrl = er32(CTRL);
>>-	if (features&  NETIF_F_HW_VLAN_RX) {
>>-		/* enable VLAN tag insert/strip */
>>-		ctrl |= E1000_CTRL_VME;
>>-	} else {
>>-		/* disable VLAN tag insert/strip */
>>-		ctrl&= ~E1000_CTRL_VME;
>>-	}
>>-	ew32(CTRL, ctrl);
>>+	__e1000_vlan_mode(adapter, features);
>>
>>  	if (!test_bit(__E1000_DOWN,&adapter->flags))
>>  		e1000_irq_enable(adapter);
>
>
>-- 
>Ben Greear <greearb@candelatech.com>
>Candela Technologies Inc  http://www.candelatech.com
>

------------------------------------------------------------------------------
This SF email is sponsosred by:
Try Windows Azure free for 90 days Click Here 
http://p.sf.net/sfu/sfd2d-msazure
_______________________________________________
E1000-devel mailing list
E1000-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/e1000-devel
To learn more about Intel&#174; Ethernet, visit http://communities.intel.com/community/wired

^ permalink raw reply	[flat|nested] 7+ messages in thread
end of thread, other threads:[~2012-03-20  6:26 UTC | newest]

Thread overview: 7+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2012-03-12 18:17 VLAN regression caused by: e1000: do vlan cleanup (799d531) Ben Greear
2012-03-13  9:59 ` Jiri Pirko
2012-03-13 16:18   ` Ben Greear
2012-03-13 18:44   ` Ben Greear
2012-03-19 13:08     ` Jiri Pirko
2012-03-19 23:24       ` Ben Greear
2012-03-20  6:26         ` Jiri Pirko

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).