Separate TX and RX DMA rings to make their handling slightly clearer.
This is a preparatory patch for bulk RX descriptor flushing.
Signed-off-by: Marek Vasut <marex@denx.de>
Reviewed-by: Patrice Chotard <patrice.chotard@foss.st.com>
Reviewed-by: Ramon Fried <rfried.dev@gmail.com>
Reorder include files in the U-Boot expected order.
Signed-off-by: Patrick Delaunay <patrick.delaunay@foss.st.com>
Reviewed-by: Ramon Fried <rfried.dev@gmail.com>
Only two boards in the tree set the macro DWC_NET_PHYADDR. Both have
CONFIG_DM_ETH_PHY=y, so should set the phy address in DT if necessary.
The imx8mp_evk does set the correct address in device tree.
The other board seems to be a copy-paste-adapt from an old
version of the imx8mp_evk config header, given the "#ifdef
CONFIG_DWC_ETH_QOS" block that has been removed from imx8mp_evk header
in commit 127fb45495. Its device tree doesn't even enable (i.e., set
'status = "okay"') the &eqos node. But the other ethernet device,
&fec, does get enabled, and does have a phy sitting at address 4 (and
it also has a corresponding legacy #define CONFIG_FEC_MXC_PHYADDR
4). So I believe it should be completely safe to remove it from there
as well.
Signed-off-by: Rasmus Villemoes <rasmus.villemoes@prevas.dk>
Reviewed-by: Ramon Fried <rfried.dev@gmail.com>
[trini: Re-apply to top of tree, update imx93_evk.h]
Signed-off-by: Tom Rini <trini@konsulko.com>
I have an iMX8MP with a ti,dp83867 phy in front of the eqos
interface. The phy is Gbit capable - however, the C and D differential
pairs are not physically routed to the RJ45 connector. So I need to
prevent the phy from advertising 1000Mbps.
The necessary code is almost already there in the form of a
phy_set_supported() call in eqos_start(), but the max-speed DT
property is currently only parsed in
eqos_probe_resources_stm32(). Lift that parsing to eqos_probe().
Signed-off-by: Rasmus Villemoes <rasmus.villemoes@prevas.dk>
Reviewed-by: Ramon Fried <rfried.dev@gmail.com>
Not only does eqos_remove() fail to free the buffers that have been
allocated by eqos_probe_resources_core(), it repeats those allocations
and thus drops twice as much memory on the floor.
Signed-off-by: Rasmus Villemoes <rasmus.villemoes@prevas.dk>
Reviewed-by: Ramon Fried <rfried.dev@gmail.com>
Since new atheros PHY driver needs to access its PHY node through
phy device, we have to assign the phy node in ethernet controller
driver. Otherwise the PHY driver will fail to get some nodes
and properties.
Signed-off-by: Ye Li <ye.li@nxp.com>
Signed-off-by: Peng Fan <peng.fan@nxp.com>
Move i.MX code to a standalone file to make it easy for adding new
platform support
Reviewed-by: Ramon Fried <rfried.dev@gmail.com>
Signed-off-by: Peng Fan <peng.fan@nxp.com>
Move macros and structures to header file and make some functions
public, so that could used by other files, this is to
prepare split platform specific config to one file.
Signed-off-by: Peng Fan <peng.fan@nxp.com>
When CONFIG_CLK is not enabled, there will be buil break:
"error: ‘eqos’ undeclared (first use in this function)"
Take eqos definition out the CONFIG_CLK ifdef.
Signed-off-by: Peng Fan <peng.fan@nxp.com>
- DM9000 DM support
- tftp server bug fix
- mdio ofnode support functions
- Various phy fixes and improvements.
[trini: Fixup merge conflicts in drivers/net/phy/ethernet_id.c
drivers/net/phy/phy.c include/phy.h]
The correct compatible string for i.MX8MP variant of DWC EQoS MAC
is "nxp,imx8mp-dwmac-eqos", use it. Drop the two current users of
the current wrong compatible string to avoid breaking them.
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Fabio Estevam <festevam@denx.de>
Cc: Marcel Ziswiler <marcel.ziswiler@toradex.com>
Cc: Peng Fan <peng.fan@nxp.com>
Cc: Stefano Babic <sbabic@denx.de>
Reviewed-by: Fabio Estevam <festevam@denx.de>
Tested-by: Marcel Ziswiler <marcel.ziswiler@toradex.com>
Reviewed-by: Ramon Fried <rfried.dev@gmail.com>
Rename constant PHY_INTERFACE_MODE_NONE to PHY_INTERFACE_MODE_NA to make
it compatible with Linux' naming.
Signed-off-by: Marek Behún <marek.behun@nic.cz>
Reviewed-by: Stefan Roese <sr@denx.de>
Reviewed-by: Ramon Fried <rfried.dev@gmail.com>
Reviewed-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Add helpers ofnode_read_phy_mode() and dev_read_phy_mode() to parse the
"phy-mode" / "phy-connection-type" property. Add corresponding UT test.
Use them treewide.
This allows us to inline the phy_get_interface_by_name() into
ofnode_read_phy_mode(), since the former is not used anymore.
Signed-off-by: Marek Behún <marek.behun@nic.cz>
Reviewed-by: Ramon Fried <rfried.dev@gmail.com>
Tested-by: Patrice Chotard <patrice.chotard@foss.st.com>
Enable DWC IP clock in driver probe, so the MII access is possible even
outside of active network transfers. This is particularly useful when
using 'mii' or 'mdio' commands to explore PHY state, neither of which
works with DWMAC currently due to the disabled clock.
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Joe Hershberger <joe.hershberger@ni.com>
Cc: Patrice Chotard <patrice.chotard@foss.st.com>
Cc: Patrick Delaunay <patrick.delaunay@foss.st.com>
Cc: Ramon Fried <rfried.dev@gmail.com>
Cc: Stephen Warren <swarren@nvidia.com>
Reviewed-by: Patrice Chotard <patrice.chotard@foss.st.com>
Reviewed-by: Ramon Fried <rfried.dev@gmail.com>
Define LOG_CATEGORY to allow filtering with log command.
Signed-off-by: Patrick Delaunay <patrick.delaunay@foss.st.com>
Reviewed-by: Ramon Fried <rfried.dev@gmail.com>
Add a common empty ops: eqos_null_ops() to remove the duplicated empty
functions and reduce the driver size for stm32 and imx config.
This patch also aligns the prototype of ops 'eqos_stop_clks' with other
eqos ops by adding return value.
Reviewed-by: Ramon Fried <rfried.dev@gmail.com>
Signed-off-by: Patrick Delaunay <patrick.delaunay@foss.st.com>
Use the generic ethernet phy which already manages the correct binding
for gpio reset, including the assert an deassert delays.
Reviewed-by: Ramon Fried <rfried.dev@gmail.com>
Signed-off-by: Patrick Delaunay <patrick.delaunay@foss.st.com>
Since the commit commit 6a895d039b ("net: Update eQos driver and FEC
driver to use eth phy interfaces") the field phyaddr of driver private data
struct eqos_priv is no more used in eqos_start() for the phy_connect()
parameter.
Now this variable is only initialized in eqos_probe_resources_stm32()
it can be removed.
Reviewed-by: Ramon Fried <rfried.dev@gmail.com>
Signed-off-by: Patrick Delaunay <patrick.delaunay@foss.st.com>
Remove the define EQOS_DESCRIPTOR_ALIGN unused since the
commit 6f1e668d96 ("net: dwc_eth_qos: Pad descriptors to cacheline size")
Signed-off-by: Patrick Delaunay <patrick.delaunay@foss.st.com>
Acked-by: Marek Vasut <marex@denx.de>
Reviewed-by: Ramon Fried <rfried.dev@gmail.com>
Revert some changes of commit 3a97da12ee ("net: dwc_eth_qos: add dwc
eqos for imx support") that were probably added by mistake.
One of these changes can lead to received data corruption (enabling
FUP and FEP bits). Another causes invalid register rxq_ctrl0 settings
for some platforms. And another makes some writes at unknown memory
location.
Fixes: 3a97da12ee ("net: dwc_eth_qos: add dwc eqos for imx support")
Signed-off-by: Daniil Stas <daniil.stas@posteo.net>
Cc: Ye Li <ye.li@nxp.com>
Cc: Fugang Duan <fugang.duan@nxp.com>
Cc: Peng Fan <peng.fan@nxp.com>
Cc: Ramon Fried <rfried.dev@gmail.com>
Cc: Joe Hershberger <joe.hershberger@ni.com>
Cc: Patrice Chotard <patrice.chotard@foss.st.com>
Cc: Patrick Delaunay <patrick.delaunay@foss.st.com>
Reviewed-by: Ramon Fried <rfried.dev@gmail.com>
Disabling clk_ck clock leads to link up status loss in phy, which
leads to auto-negotiation restart before each network command
execution.
This issue is especially big for PXE boot protocol because of
auto-negotiation restarts before each configuration filename trial.
To avoid this issue don't disable clk_ck clock after it was enabled.
Signed-off-by: Daniil Stas <daniil.stas@posteo.net>
Cc: Ramon Fried <rfried.dev@gmail.com>
Cc: Joe Hershberger <joe.hershberger@ni.com>
Cc: Patrick Delaunay <patrick.delaunay@foss.st.com>
Cc: Patrice Chotard <patrice.chotard@foss.st.com>
Reviewed-by: Ramon Fried <rfried.dev@gmail.com>
Current MDIO wait time is too long, which introduce long delay when
PHY negotiation register checking. Reduce it to 10us
Signed-off-by: Ye Li <ye.li@nxp.com>
Reviewed-by: Fugang Duan <Fugang.duan@nxp.com>
Signed-off-by: Peng Fan <peng.fan@nxp.com>
Reviewed-by: Priyanka Jain <priyanka.jain@nxp.com>
The DWMAC4 IP has the possibility to skip up to 7 AXI bus width size words
after the descriptor. Use this to pad the descriptors to cacheline size and
remove the need for noncached memory altogether. Moreover, this lets Tegra
use the generic cache flush / invalidate operations.
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Joe Hershberger <joe.hershberger@ni.com>
Cc: Patrice Chotard <patrice.chotard@st.com>
Cc: Patrick Delaunay <patrick.delaunay@st.com>
Cc: Ramon Fried <rfried.dev@gmail.com>
Cc: Stephen Warren <swarren@nvidia.com>
Tested-by: Stephen Warren <swarren@nvidia.com>
Reviewed-by: Stephen Warren <swarren@nvidia.com>
Tested-by: Patrice Chotard <patrice.chotard@foss.st.com>
Checking for seq == -1 is effectively checking that the device is
activated. The new sequence numbers are never -1 for a bound device, so
update the check.
Also drop the note about valid sequence numbers so it is accurate with the
new approach.
Signed-off-by: Simon Glass <sjg@chromium.org>
We use 'priv' for private data but often use 'platdata' for platform data.
We can't really use 'pdata' since that is ambiguous (it could mean private
or platform data).
Rename some of the latter variables to end with 'plat' for consistency.
Signed-off-by: Simon Glass <sjg@chromium.org>
This construct is quite long-winded. In earlier days it made some sense
since auto-allocation was a strange concept. But with driver model now
used pretty universally, we can shorten this to 'auto'. This reduces
verbosity and makes it easier to read.
Coincidentally it also ensures that every declaration is on one line,
thus making dtoc's job easier.
Signed-off-by: Simon Glass <sjg@chromium.org>
Use ofnode_ or dev_ APIs instead of fdt_ and fdtdec_ APIs so that the
driver can support live DT.
Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
Add configuration flag to select the supported dwc driver configuration:
- CONFIG_DWC_ETH_QOS_TEGRA186
- CONFIG_DWC_ETH_QOS_IMX
- CONFIG_DWC_ETH_QOS_STM32
See Linux driver ethernet/stmicro/stmmac and associated glue layers
for other configuration examples.
This patch removes the not-selected compatibles and lets the linker remove
the unused functions to reduce the size of the driver.
Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
When you enable CONFIG_OF_LIVE, you will end up with a lot of
conversions.
To generate this commit, I used coccinelle excluding drivers/core/,
include/dm/, and test/
The semantic patch that makes this change is as follows:
<smpl>
@@
expression dev;
@@
-devfdt_get_addr(dev)
+dev_read_addr(dev)
</smpl>
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
When you enable CONFIG_OF_LIVE, you will end up with a lot of
conversions.
To generate this commit, I used coccinelle excluding drivers/core/,
include/dm/, and test/
The semantic patch that makes this change is as follows:
<smpl>
@@
expression dev;
@@
-devfdt_get_addr(dev)
+dev_read_addr(dev)
</smpl>
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Update the compatible associated with the STM32 MPU glue
in the DWC ethernet driver.
The supported compatible is the specific "st,stm32mp1-dwmac"
as indicated in Linux binding
Documentation/devicetree/bindings/net/stm32-dwmac.txt
and not the "snps,dwmac-4.20a" only used to the select IP
version.
This glue is implemented in Linux kernel in:
drivers/net/ethernet/stmicro/stmmac/dwmac-stm32.c
For information in stm32mp151.dtsi, the 2 compatibles are
supported:
ethernet0: ethernet@5800a000 {
compatible = "st,stm32mp1-dwmac", "snps,dwmac-4.20a";
...
};
Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
Reviewed-by: Patrice Chotard <patrice.chotard@st.com>
Move this header out of the common header. Network support is used in
quite a few places but it still does not warrant blanket inclusion.
Note that this net.h header itself has quite a lot in it. It could be
split into the driver-mode support, functions, structures, checksumming,
etc.
Signed-off-by: Simon Glass <sjg@chromium.org>
Implement the read_rom_hwaddr callback to load MAC address from fuse
for imx8m platforms.
Reviewed-by: Fugang Duan <fugang.duan@nxp.com>
Signed-off-by: Ye Li <ye.li@nxp.com>
Signed-off-by: Peng Fan <peng.fan@nxp.com>
Implement the callbacks to get phy mode interface and txclk
rate configuration.
Reviewed-by: Ye Li <ye.li@nxp.com>
Signed-off-by: Fugang Duan <fugang.duan@nxp.com>
Signed-off-by: Peng Fan <peng.fan@nxp.com>
Add dwc eqos for imx support.
Reviewed-by: Ye Li <ye.li@nxp.com>
Signed-off-by: Fugang Duan <fugang.duan@nxp.com>
Signed-off-by: Peng Fan <peng.fan@nxp.com>
The DMA may attempt to write a DMA descriptor in the ring while it is
being updated. By writing the DMA descriptor buffer address to 0, it
is assured the DMA will not use such a buffer and the buffer can be
updated without any interference.
Reviewed-by: Patrick Delaunay <patrick.delaunay@st.com>
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Joe Hershberger <joe.hershberger@ni.com>
Cc: Patrice Chotard <patrice.chotard@st.com>
Cc: Patrick Delaunay <patrick.delaunay@st.com>
Cc: Ramon Fried <rfried.dev@gmail.com>
Cc: Stephen Warren <swarren@nvidia.com>
This patch prevents an issue where the RX packet might have been
accessed by the CPU, which now has cached data from the packet in
the caches and possibly various write buffers, and these data may
be evicted from the caches into the DRAM while the buffer is also
written by the DMA.
By invalidating the buffer after the CPU accessed it and before the
DMA populates the buffer, it is assured that the buffer will not be
corrupted.
Reviewed-by: Patrick Delaunay <patrick.delaunay@st.com>
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Joe Hershberger <joe.hershberger@ni.com>
Cc: Patrice Chotard <patrice.chotard@st.com>
Cc: Patrick Delaunay <patrick.delaunay@st.com>
Cc: Ramon Fried <rfried.dev@gmail.com>
Cc: Stephen Warren <swarren@nvidia.com>
The current code polls the RX desciptor ring for new packets by reading
the RX descriptor status. This works by accident, as the RX descriptors
are often in non-cacheable memory. However, the driver does support use
of RX descriptors in cacheable memory.
This patch adds a missing RX descriptor invalidation, which assures the
CPU will read a fresh copy of the RX descriptor instead of a cached one.
Reviewed-by: Patrick Delaunay <patrick.delaunay@st.com>
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Joe Hershberger <joe.hershberger@ni.com>
Cc: Patrice Chotard <patrice.chotard@st.com>
Cc: Patrick Delaunay <patrick.delaunay@st.com>
Cc: Ramon Fried <rfried.dev@gmail.com>
Cc: Stephen Warren <swarren@nvidia.com>
Currently the code only flushes the first RX descriptor, not every entry
in the RX descriptor ring. Fix this, to make sure the DMA engine can pick
the RX descriptors correctly.
Reviewed-by: Patrick Delaunay <patrick.delaunay@st.com>
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Joe Hershberger <joe.hershberger@ni.com>
Cc: Patrice Chotard <patrice.chotard@st.com>
Cc: Patrick Delaunay <patrick.delaunay@st.com>
Cc: Ramon Fried <rfried.dev@gmail.com>
Cc: Stephen Warren <swarren@nvidia.com>
This code programs the next descriptor in the TX descriptor ring into
the hardware as the last valid TX descriptor. The problem is that if
the currenty descriptor is the last one in the array, the code will
not wrap around correctly and use TX descriptor 0 again, but instead
will use TX descriptor at address right past the TX descriptor ring,
which is the first descriptor in the RX ring.
Fix this by adding the necessary wrap-around.
Reviewed-by: Patrick Delaunay <patrick.delaunay@st.com>
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Joe Hershberger <joe.hershberger@ni.com>
Cc: Patrice Chotard <patrice.chotard@st.com>
Cc: Patrick Delaunay <patrick.delaunay@st.com>
Cc: Ramon Fried <rfried.dev@gmail.com>
Cc: Stephen Warren <swarren@nvidia.com>
The RX descriptor field 3 should contain only OWN and BUF1V bits before
being used for receiving data by the DMA engine. However, right now, if
the descriptor was already used for receiving data and is being cleared,
the field 3 is only modified and the aforementioned two bits are ORRed
into the field. This could lead to a residual dirty bits being left in
the field 3 from previous transfer, and it generally does. Fully set the
field 3 instead to clear those residual dirty bits.
Reviewed-by: Patrick Delaunay <patrick.delaunay@st.com>
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Joe Hershberger <joe.hershberger@ni.com>
Cc: Patrice Chotard <patrice.chotard@st.com>
Cc: Patrick Delaunay <patrick.delaunay@st.com>
Cc: Ramon Fried <rfried.dev@gmail.com>
Cc: Stephen Warren <swarren@nvidia.com>
