diff mbox

Documentation: document the gianfar tx vlan offload issue

Message ID 1396580685-17156-1-git-send-email-roy.qing.li@gmail.com
State Changes Requested, archived
Delegated to: David Miller
Headers show

Commit Message

Li RongQing April 4, 2014, 3:04 a.m. UTC
From: Li RongQing <roy.qing.li@gmail.com>

The commit b852b72087[gianfar: fix bug caused by 87c288c6e9] and e2c53be223
[gianfar: fix default tx vlan offload feature flag] disabled tx vlan offload
from dev->feature, since enabling vlan tag insertion leads to unusable
connections on some configurations.

But this feature is in dev->hw_feature still, and a user can enable it by
"ethtool", so document the tx vlan offload issue.

Signed-off-by: Li RongQing <roy.qing.li@gmail.com>
Signed-off-by: Zhu Yanjun <Yanjun.Zhu@windriver.com>
Cc: Claudiu Manoil <claudiu.manoil@freescale.com>
---
 Documentation/networking/gianfar.txt |   12 ++++++++++++
 1 file changed, 12 insertions(+)

Comments

yzhu1 April 4, 2014, 3:07 a.m. UTC | #1
On 04/04/2014 11:04 AM, roy.qing.li@gmail.com wrote:
> From: Li RongQing <roy.qing.li@gmail.com>
>
> The commit b852b72087[gianfar: fix bug caused by 87c288c6e9] and e2c53be223
> [gianfar: fix default tx vlan offload feature flag] disabled tx vlan offload
> from dev->feature, since enabling vlan tag insertion leads to unusable
> connections on some configurations.
>
> But this feature is in dev->hw_feature still, and a user can enable it by
> "ethtool", so document the tx vlan offload issue.
>
> Signed-off-by: Li RongQing <roy.qing.li@gmail.com>
> Signed-off-by: Zhu Yanjun <Yanjun.Zhu@windriver.com>
Nice!

Zhu Yanjun
> Cc: Claudiu Manoil <claudiu.manoil@freescale.com>
> ---
>   Documentation/networking/gianfar.txt |   12 ++++++++++++
>   1 file changed, 12 insertions(+)
>
> diff --git a/Documentation/networking/gianfar.txt b/Documentation/networking/gianfar.txt
> index ad474ea..2d3dbe6 100644
> --- a/Documentation/networking/gianfar.txt
> +++ b/Documentation/networking/gianfar.txt
> @@ -51,6 +51,18 @@ configuring VLANs.  The gianfar driver supports hardware insertion and
>   extraction of VLAN headers, but not filtering.  Filtering will be
>   done by the kernel.
>   
> +But if tx vlan tag hardware insertion is enabled, and a vlan-device is
> +created, the physical device which vlan device is on will be unable to
> +work, like:
> +
> +	$ethtool -K eth0 tx-vlan-stag-hw-insert on
> +	$ifconfig eth0 192.168.1.13/24 up
> +
> +	$vconfig add eth0 4091
> +	$ifconfig eth0.4091 10.1.1.13/24 up
> +eth0.4091 can work, but eth0 can not work, since eth0 sends the insane
> +packets out.
> +
>   MULTICASTING
>   
>   The gianfar driver supports using the group hash table on the

--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Claudiu Manoil April 4, 2014, 7:07 a.m. UTC | #2
On 4/4/2014 6:04 AM, roy.qing.li@gmail.com wrote:
> From: Li RongQing <roy.qing.li@gmail.com>
>
> The commit b852b72087[gianfar: fix bug caused by 87c288c6e9] and e2c53be223
> [gianfar: fix default tx vlan offload feature flag] disabled tx vlan offload
> from dev->feature, since enabling vlan tag insertion leads to unusable
> connections on some configurations.
>
> But this feature is in dev->hw_feature still, and a user can enable it by
> "ethtool", so document the tx vlan offload issue.
>
> Signed-off-by: Li RongQing <roy.qing.li@gmail.com>
> Signed-off-by: Zhu Yanjun <Yanjun.Zhu@windriver.com>
> Cc: Claudiu Manoil <claudiu.manoil@freescale.com>
> ---
>   Documentation/networking/gianfar.txt |   12 ++++++++++++
>   1 file changed, 12 insertions(+)
>
> diff --git a/Documentation/networking/gianfar.txt b/Documentation/networking/gianfar.txt
> index ad474ea..2d3dbe6 100644
> --- a/Documentation/networking/gianfar.txt
> +++ b/Documentation/networking/gianfar.txt
> @@ -51,6 +51,18 @@ configuring VLANs.  The gianfar driver supports hardware insertion and
>   extraction of VLAN headers, but not filtering.  Filtering will be
>   done by the kernel.
>   
> +But if tx vlan tag hardware insertion is enabled, and a vlan-device is
> +created, the physical device which vlan device is on will be unable to
> +work, like:
> +
> +	$ethtool -K eth0 tx-vlan-stag-hw-insert on
> +	$ifconfig eth0 192.168.1.13/24 up
> +
> +	$vconfig add eth0 4091
> +	$ifconfig eth0.4091 10.1.1.13/24 up
> +eth0.4091 can work, but eth0 can not work, since eth0 sends the insane
> +packets out.
> +
I don't really like this, this is hardly a technical explanation. 
("insane packets"?)
Please have a look at the HW reference manual.  There you have the 
description
of the VLAN extraction feature (see VLEX).

Regards,
Claudiu


--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Claudiu Manoil April 4, 2014, 7:10 a.m. UTC | #3
On 4/4/2014 10:07 AM, Claudiu Manoil wrote:
> On 4/4/2014 6:04 AM, roy.qing.li@gmail.com wrote:
>> From: Li RongQing <roy.qing.li@gmail.com>
>>
>> The commit b852b72087[gianfar: fix bug caused by 87c288c6e9] and
>> e2c53be223
>> [gianfar: fix default tx vlan offload feature flag] disabled tx vlan
>> offload
>> from dev->feature, since enabling vlan tag insertion leads to unusable
>> connections on some configurations.
>>
>> But this feature is in dev->hw_feature still, and a user can enable it by
>> "ethtool", so document the tx vlan offload issue.
>>
>> Signed-off-by: Li RongQing <roy.qing.li@gmail.com>
>> Signed-off-by: Zhu Yanjun <Yanjun.Zhu@windriver.com>
>> Cc: Claudiu Manoil <claudiu.manoil@freescale.com>
>> ---
>>   Documentation/networking/gianfar.txt |   12 ++++++++++++
>>   1 file changed, 12 insertions(+)
>>
>> diff --git a/Documentation/networking/gianfar.txt
>> b/Documentation/networking/gianfar.txt
>> index ad474ea..2d3dbe6 100644
>> --- a/Documentation/networking/gianfar.txt
>> +++ b/Documentation/networking/gianfar.txt
>> @@ -51,6 +51,18 @@ configuring VLANs.  The gianfar driver supports
>> hardware insertion and
>>   extraction of VLAN headers, but not filtering.  Filtering will be
>>   done by the kernel.
>> +But if tx vlan tag hardware insertion is enabled, and a vlan-device is
>> +created, the physical device which vlan device is on will be unable to
>> +work, like:
>> +
>> +    $ethtool -K eth0 tx-vlan-stag-hw-insert on
>> +    $ifconfig eth0 192.168.1.13/24 up
>> +
>> +    $vconfig add eth0 4091
>> +    $ifconfig eth0.4091 10.1.1.13/24 up
>> +eth0.4091 can work, but eth0 can not work, since eth0 sends the insane
>> +packets out.
>> +
> I don't really like this, this is hardly a technical explanation.
> ("insane packets"?)
> Please have a look at the HW reference manual.  There you have the
> description
> of the VLAN extraction feature (see VLEX).
>
And VLAN insertion (VLINS) as well, for that matter.


--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Li RongQing April 4, 2014, 7:39 a.m. UTC | #4
>>
>> I don't really like this, this is hardly a technical explanation.
>> ("insane packets"?)
>> Please have a look at the HW reference manual.  There you have the
>> description
>> of the VLAN extraction feature (see VLEX).
>>
> And VLAN insertion (VLINS) as well, for that matter.
>
>

Not all users will have time to read HW reference manual, if we can give them
more explanation, I think it maybe save their time to debug, to see code, and
search manual

-Roy
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
David Miller April 4, 2014, 1:56 p.m. UTC | #5
From: roy.qing.li@gmail.com
Date: Fri,  4 Apr 2014 11:04:45 +0800

> +But if tx vlan tag hardware insertion is enabled, and a vlan-device is
> +created, the physical device which vlan device is on will be unable to
> +work, like:
> +
> +	$ethtool -K eth0 tx-vlan-stag-hw-insert on
> +	$ifconfig eth0 192.168.1.13/24 up
> +
> +	$vconfig add eth0 4091
> +	$ifconfig eth0.4091 10.1.1.13/24 up
> +eth0.4091 can work, but eth0 can not work, since eth0 sends the insane
> +packets out.

This is a bit of a joke.

If a configuration doesn't work properly, it should be disallowed
by the driver, rather than "documented".
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
diff mbox

Patch

diff --git a/Documentation/networking/gianfar.txt b/Documentation/networking/gianfar.txt
index ad474ea..2d3dbe6 100644
--- a/Documentation/networking/gianfar.txt
+++ b/Documentation/networking/gianfar.txt
@@ -51,6 +51,18 @@  configuring VLANs.  The gianfar driver supports hardware insertion and
 extraction of VLAN headers, but not filtering.  Filtering will be
 done by the kernel.
 
+But if tx vlan tag hardware insertion is enabled, and a vlan-device is
+created, the physical device which vlan device is on will be unable to
+work, like:
+
+	$ethtool -K eth0 tx-vlan-stag-hw-insert on
+	$ifconfig eth0 192.168.1.13/24 up
+
+	$vconfig add eth0 4091
+	$ifconfig eth0.4091 10.1.1.13/24 up
+eth0.4091 can work, but eth0 can not work, since eth0 sends the insane
+packets out.
+
 MULTICASTING
 
 The gianfar driver supports using the group hash table on the