| Age | Commit message (Collapse) | Author |
|---|
|
Rename PHY_INTERFACE_MODE_SGMII_2500 to PHY_INTERFACE_MODE_2500SGMII
Convention is to put the number(2500) first and then the
interface mode(SGMII)
Signed-off-by: Bhaskar Upadhaya <Bhaskar.Upadhaya@nxp.com>
|
|
Signed-off-by: Camelia Groza <camelia.groza@nxp.com>
|
|
Signed-off-by: Camelia Groza <camelia.groza@nxp.com>
|
|
When the egress CQ is congested, drop the frames instead of enqueueing
them. This is more efficient than enqueueing and receiving them back on
the ERN queue.
We also can't stop the netdev queues because that would affect all the CQs
and would hinder prioritization.
Signed-off-by: Camelia Groza <camelia.groza@nxp.com>
|
|
Configure the CEETM egress congestion thresholds independently from the
default Ethernet driver's Work Queues. Allow the user to edit the
thresholds through menuconfig.
Signed-off-by: Camelia Groza <camelia.groza@nxp.com>
|
|
The errata prevents us from transmitting S/G frames. Instead of
linearizing them ourselves, stop advertising S/G support and have the
stack do it for us.
Signed-off-by: Camelia Groza <camelia.groza@nxp.com>
|
|
Signed-off-by: Vicentiu Galanopulo <vicentiu.galanopulo@nxp.com>
|
|
Signed-off-by: Madalin Bucur <madalin.bucur@nxp.com>
|
|
Documentation/networking/netdevices.txt mentions that interfaces must
be able to receive frames at least the size of the configured MTU. The
behavior for received frames larger than the MTU is unspecified. We have
been dropping these frames in software. Remove this behavior and accept
them.
Signed-off-by: Camelia Groza <camelia.groza@nxp.com>
|
|
We can not allow for Jumbo frames and large MTU values on LS1043A due to
the A-010022 FMan errata. All outgoing frames larger than 4K bytes are dropped.
Signed-off-by: Camelia Groza <camelia.groza@nxp.com>
|
|
When creating a new skb for the errata workaround, maintain the socket
and timestamp configurations for timestamp hardware offloading.
Signed-off-by: Camelia Groza <camelia.groza@nxp.com>
|
|
[dpaa part]
Update txq0's trans_start in order to prevent the netdev watchdog from
triggering too quickly. Since we set the LLTX flag, the stack won't update
the jiffies for other tx queues. Prevent the watchdog from checking the
other tx queues by adding the NETIF_HW_ACCEL_MQ flag.
Signed-off-by: Camelia Groza <camelia.groza@nxp.com>
|
|
Signed-off-by: Camelia Groza <camelia.groza@nxp.com>
|
|
Signed-off-by: Camelia Groza <camelia.groza@nxp.com>
|
|
Signed-off-by: Camelia Groza <camelia.groza@nxp.com>
|
|
If the skb's headroom isn't aligned to 16 bytes, reallocate the entire
skb and resize its headroom to priv->tx_headroom. Update the pointers
to the network and transport headers accordingly.
Signed-off-by: Camelia Groza <camelia.groza@nxp.com>
|
|
Signed-off-by: Camelia Groza <camelia.groza@nxp.com>
|
|
Allocate a new page and copy the skb's contents to it in order to
guarantee that 4k boundary crossings do not occur.
Signed-off-by: Camelia Groza <camelia.groza@nxp.com>
|
|
Signed-off-by: Camelia Groza <camelia.groza@nxp.com>
|
|
The tcf_destroy_chain call prototype has been moved.
Signed-off-by: Camelia Groza <camelia.groza@nxp.com>
|
|
Signed-off-by: Camelia Groza <camelia.groza@nxp.com>
|
|
The tc_classify and tc_classify_compat calls have been merged. Adapt to
the new API.
Signed-off-by: Camelia Groza <camelia.groza@nxp.com>
|
|
This patch is to move global variable 'clock' for DPAA PTP
clock pointer into ptp_priv_s struct.
Signed-off-by: Yangbo Lu <yangbo.lu@nxp.com>
|
|
Layerscape DPAA platforms have updated dts to use ptp-timer phandle
instead of ptimer-handle for Fman RTC node. This patch is to
update it in driver.
Signed-off-by: Yangbo Lu <yangbo.lu@nxp.com>
|
|
Signed-off-by: Camelia Groza camelia.groza@nxp.com
|
|
Signed-off-by: Zhao Qiang <qiang.zhao@nxp.com>
|
|
Signed-off-by: Madalin Bucur <madalin.bucur@nxp.com>
|
|
commit 3399e34ee929b3498b7ae53181e29c4000e2dfa7
[sdk_dpaa part]
Signed-off-by: Madalin Bucur <madalin.bucur@nxp.com>
Integrated-by: Zhao Qiang <qiang.zhao@nxp.com>
|
