[PATCH v2] vxlan: restore dev->mtu setting based on lower device

[PATCH v2] vxlan: restore dev->mtu setting based on lower device

[Alexey Kodanev]

Stefano Brivio says:
    Commit a985343ba906 ("vxlan: refactor verification and
    application of configuration") introduced a change in the
    behaviour of initial MTU setting: earlier, the MTU for a link
    created on top of a given lower device, without an initial MTU
    specification, was set to the MTU of the lower device minus
    headroom as a result of this path in vxlan_dev_configure():

	if (!conf->mtu)
		dev->mtu = lowerdev->mtu -
			   (use_ipv6 ? VXLAN6_HEADROOM : VXLAN_HEADROOM);

    which is now gone. Now, the initial MTU, in absence of a
    configured value, is simply set by ether_setup() to ETH_DATA_LEN
    (1500 bytes).

    This breaks userspace expectations in case the MTU of
    the lower device is higher than 1500 bytes minus headroom.

This patch restores the previous behaviour on newlink operation. Since
max_mtu can be negative and we update dev->mtu directly, also check it
for valid minimum.

Reported-by: Junhan Yan <juyan@redhat.com>
Fixes: a985343ba906 ("vxlan: refactor verification and application of configuration")
Signed-off-by: Alexey Kodanev <alexey.kodanev@oracle.com>
---

v2: Based on initial patch from Stefano Brivio:
    "vxlan: Restore initial MTU setting based on lower device"

    In this version, 'dev->mtu' is set from 'max_mtu' that already has the
    the needed initial MTU value based on the lower device. Plus, it is adding
    extra check for valid minimum of 'max_mtu'.

 drivers/net/vxlan.c |    5 +++++
 1 files changed, 5 insertions(+), 0 deletions(-)

diff --git a/drivers/net/vxlan.c b/drivers/net/vxlan.c
index 19b9cc5..1000b0e 100644
--- a/drivers/net/vxlan.c
+++ b/drivers/net/vxlan.c
@@ -3103,6 +3103,11 @@ static void vxlan_config_apply(struct net_device *dev,
 
 		max_mtu = lowerdev->mtu - (use_ipv6 ? VXLAN6_HEADROOM :
 					   VXLAN_HEADROOM);
+		if (max_mtu < ETH_MIN_MTU)
+			max_mtu = ETH_MIN_MTU;
+
+		if (!changelink && !conf->mtu)
+			dev->mtu = max_mtu;
 	}
 
 	if (dev->mtu > max_mtu)
-- 
1.7.1

Re: [PATCH v2] vxlan: restore dev->mtu setting based on lower device

[Stefano Brivio]

On Thu, 14 Dec 2017 20:20:00 +0300
Alexey Kodanev <alexey.kodanev@oracle.com> wrote:

> Stefano Brivio says:
>     Commit a985343ba906 ("vxlan: refactor verification and
>     application of configuration") introduced a change in the
>     behaviour of initial MTU setting: earlier, the MTU for a link
>     created on top of a given lower device, without an initial MTU
>     specification, was set to the MTU of the lower device minus
>     headroom as a result of this path in vxlan_dev_configure():
> 
> 	if (!conf->mtu)
> 		dev->mtu = lowerdev->mtu -
> 			   (use_ipv6 ? VXLAN6_HEADROOM : VXLAN_HEADROOM);
> 
>     which is now gone. Now, the initial MTU, in absence of a
>     configured value, is simply set by ether_setup() to ETH_DATA_LEN
>     (1500 bytes).
> 
>     This breaks userspace expectations in case the MTU of
>     the lower device is higher than 1500 bytes minus headroom.
> 
> This patch restores the previous behaviour on newlink operation. Since
> max_mtu can be negative and we update dev->mtu directly, also check it
> for valid minimum.
> 
> Reported-by: Junhan Yan <juyan@redhat.com>
> Fixes: a985343ba906 ("vxlan: refactor verification and application of configuration")
> Signed-off-by: Alexey Kodanev <alexey.kodanev@oracle.com>

Thanks for re-posting this.

Acked-by: Stefano Brivio <sbrivio@redhat.com>

Or even (but I don't know what's the current netdev practice for these
particular cases):

Signed-off-by: Stefano Brivio <sbrivio@redhat.com>

-- 
Stefano

Re: [PATCH v2] vxlan: restore dev->mtu setting based on lower device

[David Miller]

From: Alexey Kodanev <alexey.kodanev@oracle.com>
Date: Thu, 14 Dec 2017 20:20:00 +0300

> Stefano Brivio says:
>     Commit a985343ba906 ("vxlan: refactor verification and
>     application of configuration") introduced a change in the
>     behaviour of initial MTU setting: earlier, the MTU for a link
>     created on top of a given lower device, without an initial MTU
>     specification, was set to the MTU of the lower device minus
>     headroom as a result of this path in vxlan_dev_configure():
> 
> 	if (!conf->mtu)
> 		dev->mtu = lowerdev->mtu -
> 			   (use_ipv6 ? VXLAN6_HEADROOM : VXLAN_HEADROOM);
> 
>     which is now gone. Now, the initial MTU, in absence of a
>     configured value, is simply set by ether_setup() to ETH_DATA_LEN
>     (1500 bytes).
> 
>     This breaks userspace expectations in case the MTU of
>     the lower device is higher than 1500 bytes minus headroom.
> 
> This patch restores the previous behaviour on newlink operation. Since
> max_mtu can be negative and we update dev->mtu directly, also check it
> for valid minimum.
> 
> Reported-by: Junhan Yan <juyan@redhat.com>
> Fixes: a985343ba906 ("vxlan: refactor verification and application of configuration")
> Signed-off-by: Alexey Kodanev <alexey.kodanev@oracle.com>

Applied and queued up for -stable, thanks.