| Age | Commit message (Collapse) | Author |
|---|
|
Airoha switch mdio maybe used not only by GDM1, but also by other GDM
ports (ex: as21xxx phy connected to GDM2 port). So it's better probe
airoha switch mdio a bit early in the airoha_eth_probe() code.
Also remove useless eth_phy_set_mdio_bus() call and related code.
Signed-off-by: Mikhail Kshevetskiy <[email protected]>
|
|
Current code just bind mt7531 mdio with it's driver, so mdio device may
not be probed and hense not usable.
This patch:
* Forces probing of mt7531 mdio for GDM1 port
* Renames the mt7531 mdio bus interface to 'mt7531-mdio'. We may have
multiple available MDIO, so the name 'mdio' isn't descriptive enough.
* Sets mdio bus for the GDM port device
Signed-off-by: Mikhail Kshevetskiy <[email protected]>
|
|
It's not possible to disable PCS support just now, an7581 u-boot will not
compile. This patch fixes an issue.
Signed-off-by: Mikhail Kshevetskiy <[email protected]>
|
|
Add required changes to call PCS function to configure the Serdes Port.
The Ethernet driver is adapted following Upstream Kernel node structure.
Function calling order is the same of Phylink upstream kernel.
With the PCS support, also add support for attaching PHY. With
"in-band-status" set in DT for the managed property, a rudimental
support for SFP module is present.
Signed-off-by: Christian Marangi <[email protected]>
|
|
Call airoha_switch_init() before creating GDM instances, so if
allocation of GDM port fails, early created GDM instances will work
normally.
Signed-off-by: Mikhail Kshevetskiy <[email protected]>
|
|
We should not call airoha_fe_init() from GDM port independent code,
because it do a GDM specific things.
Makes airoha_fe_maccr_init() and airoha_fe_init() port dependent
and call them from airoha_eth_port_probe()
Signed-off-by: Mikhail Kshevetskiy <[email protected]>
|
|
Declare airoha_eth_port as U_BOOT_DRIVER(), fix airoha_alloc_gdm_port()
to lookup a driver instead of direct airoha_eth_port usage.
Signed-off-by: Mikhail Kshevetskiy <[email protected]>
|
|
In the case of an7581 possible GDM port id are: 1, 2 and 4.
Initialization of port GDM4 will lead to out of boundary writing
to gdm_port_str[] array.
Let's increase the array size by 1 to avoid it.
Signed-off-by: Mikhail Kshevetskiy <[email protected]>
|
|
Rework the driver to support multiple GDM port. The driver is split to
main driver as a MISC driver with forced probe (by using the
DM_FLAG_PROBE_AFTER_BIND) and each GDM port register a ETH driver.
This permit a 1:1 implementation with the linux kernel driver and permit
to use the same exact DT nodes.
Signed-off-by: Christian Marangi <[email protected]>
|
|
Commit d2145a89bcf6 ("net: airoha: bind MDIO controller on Ethernet load")
uses "airoha,en7581-switch" dts node for finding MDIO childs. This is wrong
for EN7523 SoC. The correct node name should be used instead.
Fixes: d2145a89bcf6 ("net: airoha: bind MDIO controller on Ethernet load")
Signed-off-by: Mikhail Kshevetskiy <[email protected]>
|
|
Commit d2145a89bcf6 ("net: airoha: bind MDIO controller on Ethernet load")
refers to non-present CONFIG_MDIO_MT7531 and non-present "mt7531-mdio"
driver. It should use CONFIG_MDIO_MT7531_MMIO and "mt7531-mdio-mmio"
instead.
Fixes: d2145a89bcf6 ("net: airoha: bind MDIO controller on Ethernet load")
Signed-off-by: Mikhail Kshevetskiy <[email protected]>
|
|
Bind MDIO controller on Ethernet Controller load. The Airoha AN7581 SoC
have an integrated Switch based on MT7531 (or more saying MT7988).
Attach it to the mdio node in the switch node to support scanning for
MDIO devices on the BUS with DM API.
Signed-off-by: Christian Marangi <[email protected]>
|
|
Add support for Ethernet controller present in Airoha en7523/en7529/en7562.
Signed-off-by: Mikhail Kshevetskiy <[email protected]>
|
|
This allow us remove some an7581/an7583 specific code and use a common
code instead.
Signed-off-by: Mikhail Kshevetskiy <[email protected]>
|
|
The commit 997786bbf473 ("drivers/net/airoha_eth: fix stalling in package
receiving") can be improved. Instead of returning previous descriptor
it's possible:
* do nothing in even descriptor case
* return 2 descriptor to the queue (current and previous) in the odd
descriptor case.
This patch:
* implements above approach
* remove logic not required within new approach
* adds note that PKTBUFSRX must be even and larger than 7
for reliable driver operations
Signed-off-by: Mikhail Kshevetskiy <[email protected]>
|
|
Transmission of short packets does not work good for XFI (GDM2) and
HSGMII (GDM3) interfaces. The issue can be solved with:
- padding of short packets to 60 bytes
- setting of PAD_EN bit in the corresponding REG_GDM_FWD_CFG(n)
register.
The issue should present for the lan switch (GDM1) as well, but it does
does not appear due to unknown reason.
This patch set PAD_EN bit for the used GDM.
Signed-off-by: Mikhail Kshevetskiy <[email protected]>
|
|
ARCH_DMA_MINALIGN is 64 for ARMv7a/ARMv8a architectures, but RX/TX
descriptors are 32 bytes long. So they may not be aligned on an
ARCH_DMA_MINALIGN boundary. In case of RX path, this may cause the
following problem
1) Assume that a packet has arrived and the EVEN rx descriptor has been
updated with the incoming data. The driver will invalidate and check
the corresponding rx descriptor.
2) Now suppose the next descriptor (ODD) has not yet completed.
Please note that all even descriptors starts on 64-byte boundary,
and the odd ones are NOT aligned on 64-byte boundary.
Inspecting even descriptor, we will read the entire CPU cache line
(64 bytes). So we read and sore in CPU cache also the next (odd)
descriptor.
3) Now suppose the next packet (for the odd rx descriptor) arrived
while the first packet was being processed. So we have new data
in memory but old data in cache.
4) After packet processing (in arht_eth_free_pkt() function) we will
cleanup the descriptor and put it back to rx queue.
This will call flush_dcache_range() function for the even descriptor,
so the odd one will be flushed as well (it is in the same cache line).
So the old data will be written to the next rx descriptor.
5) We get a freeze. The next descriptor is empty (so the driver is
waiting for packets), but the hardware will continue to receive
packets on other available descriptors. This will continue until
the last available rx descriptor is full. Then the hardware will
also freeze.
The problem will be solved if the previous descriptor will be put back
to the queue instead of the current one.
If the current descriptor is even (starts on a 64-byte boundary),
then putting the previous descriptor to the rx queue will affect
the previous cache line. To be 100% ok, we must make sure that the
previous and the one before the previous descriptor cannot be used
for receiving at this moment.
If the current descriptor is odd, then the previous descriptor is on
the same cache line. Both (current and previous) descriptors are not
currently in use, so issue will not arrise.
WARNING: The following restrictions on PKTBUFSRX must be held:
* PKTBUFSRX is even,
* PKTBUFSRX >= 4.
The bug appears on 32-bit airoha platform, but should be present on
64-bit as well.
The code was tested both on 32-bit and 64-bit airoha boards.
Signed-off-by: Mikhail Kshevetskiy <[email protected]>
|
|
The dma_map_single() function calls one of the functions
* invalidate_dcache_range(),
* flush_dcache_range().
Both of them expect that 'vaddr' is aligned to the ARCH_DMA_MINALIGN
boundary. Unfortunately, RX/TX descriptors are 32-byte long. Thus they
might not be aligned to the ARCH_DMA_MINALIGN boundary. Data flushing
(or invalidating) might do nothing in this case.
The same applies to dma_unmap_single() function.
In the TX path case the issue might prevent package transmission (filled
TX descriptor was not flushed).
To fix an issue a special wrappers for
* dma_map_single(),
* dma_unmap_single()
functions were created. The patch fix flushing/invalidatiog for the
RX path as well.
The bug appears on 32-bit airoha platform, but should be present on
64-bit as well.
The code was tested both on 32-bit and 64-bit airoha boards.
Signed-off-by: Mikhail Kshevetskiy <[email protected]>
|
|
Compatible device list must have a terminator. If terminator is missed
the u-boot driver subsystem will access random data placed after the
list in the memory.
The issue can be observed with the "dm compat" command.
Signed-off-by: Mikhail Kshevetskiy <[email protected]>
|
|
Add airoha Ethernet driver for Airoha AN7581 SoC. This is a majorly
rewritten and simplified version of the Linux airoha_eth.c driver.
It's has been modified to support a single RX/TX ring to reflect U-Boot
implementation with recv and send API.
The struct and the define are kept as similar as possible to upstream
one to not diverge too much.
The AN7581 SoC include an Ethernet Switch based on the Mediatek MT753x
but doesn't require any modification aside from setting the CPU port and
applying the Flood configuration hence it can be handled entirely in the
Ethernet driver.
Signed-off-by: Christian Marangi <[email protected]>
|
