| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
|
|
|
| |
The driver supports older MACB-compatible NICs as well as the newer
GEM-compatible NICs. Both use DMA, but the older MACB support failed to
allocate the single receive buffer with suitable alignment.
Use net_alloc_packet() to fix that.
Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
Link: https://lore.barebox.org/20240424064058.3608016-15-a.fatoum@pengutronix.de
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
| |
Instead of directly reading from memory addresses and inserting
sw barriers to be sure that the compiler will not move loads/stores
behind this point, just use proper volatile writel/readl accesses.
Signed-off-by: Steffen Trumtrar <s.trumtrar@pengutronix.de>
Link: https://lore.barebox.org/20231201-v2023-08-0-topic-macb-v4-2-b4f5d64de1ad@pengutronix.de
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
rx_buffer gets dma_alloc'ed but is never dma_map'ed and therefor not
flushed before it is initially used.
Map the rx_buffer when the macb is initialized and unmap it on ether_halt.
While at it, cleanup the dma_alloc_coherent rx_ring/tx_ring, too.
Signed-off-by: Steffen Trumtrar <s.trumtrar@pengutronix.de>
Link: https://lore.barebox.org/20231201-v2023-08-0-topic-macb-v4-1-b4f5d64de1ad@pengutronix.de
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|\ |
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Otherwise, it leads to a bunch of warnings on passing those pointers to
dma_*() functions. As every other driver seem to be having non-const dev
pointers in their private structs, do the same here.
Signed-off-by: Denis Orlov <denorl2009@gmail.com>
Link: https://lore.barebox.org/20230604215002.20240-3-denorl2009@gmail.com
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Currently, a lot of code handles dma_addr_t values as if they actually
hold CPU addresses. However, this is not always true. For example, MIPS
architecture requires an explicit conversion from the physical address
space to some virtual address space segment to get a valid CPU-side
pointer. Another issue is that DMA ranges that may be specified in a
device tree will not work this way. To get from a virtual address to a
dma handle and vice versa we need to add/subtract some offset, which is
calculated from "dma-ranges" property. Only dma_map_single() was doing
this, but dma_sync_single_for_*() also should.
Improve the interface by adding 'struct device' as the first argument to
the dma_sync_single_for_*(). This allows to do cpu_to_dma/dma_to_cpu()
conversions in common code and call into arch-specific code with proper
cpu-side addresses. To make things more clear, make the virtual address
argument of those arch-side functions be properly represented with a
void* type.
Apply the required changes in device drivers that use the affected
functions, making them pass the appropriate device pointer.
Signed-off-by: Denis Orlov <denorl2009@gmail.com>
Link: https://lore.barebox.org/20230604215002.20240-2-denorl2009@gmail.com
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|/
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Syncing device trees with Linux upstream can lead to breakage, when
the device trees are switched to newer bindings, which are not yet
supported in barebox. To make it easier to spot such issues, we want to
start applying some heuristics to flag possibly problematic DT changes.
One step towards being able to do that is to know what nodes barebox
actually consumes. Most of the nodes have a compatible entry, which is
matched by an array of of_device_id, so let's have MODULE_DEVICE_TABLE
point at it for future extraction.
Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
Link: https://lore.barebox.org/20230612125908.1087340-1-a.fatoum@pengutronix.de
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
|
|
|
| |
Linux v6.3-rc1 renamed the zynq macb compatibles from cdns,zynq-gem to
xlnx,zynq-gem and cdns,zynqmp-gem to xlnx,zynqmp-gem. Adjust the
compatibles in the driver accordingly.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
| |
After adjustments for SAMA5D2, driver readily works on SAMA5D4 if we
have the appropriate of_device_id entry. Add it.
Acked-by: Sam Ravnborg <sam@ravnborg.org>
Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
Link: https://lore.barebox.org/20230109130822.1657470-5-a.fatoum@pengutronix.de
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
|
| |
The '_d' suffix was originally meant to distinguish barebox struct
names from Linux struct names. struct driver doesn't exist in Linux,
so we can rename it and remove the meaningless suffix.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Reviewed-by: Marco Felsch <m.felsch@pengutronix.de>
Link: https://lore.barebox.org/20221214123512.189688-4-s.hauer@pengutronix.de
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The '_d' suffix was originally introduced in case we want to import
Linux struct device as a separate struct into barebox. Over time it
became clear that this won't happen, instead barebox struct device_d
is basically the same as Linux struct device. Rename the struct name
accordingly to make porting Linux code easier.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Reviewed-by: Marco Felsch <m.felsch@pengutronix.de>
Link: https://lore.barebox.org/20221214123512.189688-3-s.hauer@pengutronix.de
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Linux struct device has the member of_node for the device_node pointer.
Rename this in barebox accordingly to minimize the necessary changes
when porting Linux code. This was done with the semantic patch:
@@ struct device_d E; @@
- E.device_node
+ E.of_node
@@ struct device_d *E; @@
- E->device_node
+ E->of_node
Plus some manual adjustments.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Reviewed-by: Marco Felsch <m.felsch@pengutronix.de>
Link: https://lore.barebox.org/20221214123512.189688-2-s.hauer@pengutronix.de
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
|
|
|
|
| |
Use a driver private buffer as network receive buffer rather than the
globally allocated ones which will be removed soon.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Link: https://lore.barebox.org/20220919080133.877651-3-s.hauer@pengutronix.de
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The Ethernet controller on the SiFive SoCs needs some special TX clock
setup. Port the relevant Linux v5.12 bits to enable it.
As the macb driver is used for some old at91 boards that aren't yet
ported to the common clock framework, that new setup code has to be
in an #ifdef.
Tested on qemu-system-riscv64 -M sifive_u with dhcp.
Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
Link: https://lore.barebox.org/20210531065815.18641-1-a.fatoum@pengutronix.de
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
|
| |
If the phy reports a 10 MBit link, which can happen during link
negotiation, the macb prints a warning, because it does not know the
clock rate for the TX clock.
Implement setting the TX clock rate for 10 MBit to avoid the warnings.
Signed-off-by: Michael Tretter <m.tretter@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
On arm64 the compiler prints the following warning, when the macb driver
is enabled:
warning: cast from pointer to integer of different size
Add the same explicit cast as implemented for all other dma addresses in
the macb driver.
Fixes: befd110b5922 ("net: macb: init multiple dummy TX queues")
Signed-off-by: Michael Tretter <m.tretter@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Changes of the link speed might require an adjustment of the clk_tx.
Read the clk_tx from the device tree and change the rate if the link
speed changes.
If the clk_tx rate is already correct, changing the clock is not
required and, thus, not being able to get the clock rate is not fatal.
Signed-off-by: Michael Tretter <m.tretter@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
|
|
| |
The macb debugging output printed the function entry for various
functions. Especially for *_recv() this debugging flooded the serial
output while conveying very little information.
Remove printing of the function calls to make enabling debugging for the
macb driver more useful.
Signed-off-by: Michael Tretter <m.tretter@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
|
|
|
| |
Coccinelle detected that the IS_ERR and ERR_PTR are mismatched.
Fix it.
Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
|
|
| |
Add compatible for the GEM on the Zynq 7000 device.
Signed-off-by: Lucas Stach <dev@lynxeye.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
|
|
|
|
| |
Both pclk and hclk are required clocks in the DT binding.
rx_clk and tx_clk are optional, but must be enabled if a system
has separate gates for them.
Signed-off-by: Lucas Stach <dev@lynxeye.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
|
|
|
| |
Tested on Microchip SAMA5D27 SOM1 EK1
Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
|
|
|
|
| |
Microchip SAMA5D27 has more then one TX queue. So it will
go in to TX timeout if only one was initialized.
Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de>
Tested-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
|
|
|
|
| |
Receive buffers are properly synchronized only if Cadence is
GEM. Fix it for MACB as well.
Fixes: 86dc5259e25d (net: macb: no need for coherent memory for receive buffer)
Signed-off-by: Ladislav Michl <ladis@linux-mips.org>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|\ |
|
| |
| |
| |
| |
| | |
Signed-off-by: Thomas Haemmerle <thomas.haemmerle@wolfvision.net>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Provide descriptors for second priority rx and tx queues and disable the
the queues if hardware is GEM. Otherwise the function macb_send() will run
into a timeout.
Signed-off-by: Thomas Haemmerle <thomas.haemmerle@wolfvision.net>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Remove memory allocation of rx buffer in function
macb_init_rx_buffer_size, which caused a memory leak since it also is
alocated in macb_probe().
Signed-off-by: Thomas Haemmerle <thomas.haemmerle@wolfvision.net>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
| |
| |
| |
| |
| |
| |
| |
| | |
Fix check for peripheral version in MACB_MID register to treat Xilinx
ZynqMP as GEM. All MIDs >= 2 indicate a GEM not only MID == 2.
Signed-off-by: Thomas Haemmerle <thomas.haemmerle@wolfvision.net>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|/
|
|
|
|
|
|
| |
Fixes compiler warning "format '%d' expects argument of type 'int', but
argument 4 has type 'size_t {aka long unsigned int}' [-Wformat=]".
Signed-off-by: Thomas Haemmerle <thomas.haemmerle@wolfvision.net>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
|
| |
Signed-off-by: Oleksij Rempel <linux@rempel-privat.de>
Reviewed-by: Roland Hieber <r.hieber@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
|
|
| |
The macb driver works with the gigabit MAC found on Atmel sama5d3 SoCs.
Add a compatible for it.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
|
|
| |
The macb driver does DMA and thus should be quiesced on shutdown. Add
the remove callback.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
|
|
|
| |
Instead of allocating struct macb_device and a struct eth_device
separately just use the eth_device structure that is already embedded
in struct macb_device but currently unused.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
|
| |
tx_tail is unused. Remove it.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
|
|
| |
The receive buffers are properly synchronized with dma_sync_*, no need
to use coherent memory for them.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
With DT enabled the following was logged:
macb fffbc000.ethernet: no macb_clk
macb fffbc000.ethernet: probe failed: Invalid argument
The clock probed was with the name used in a non-DT setup.
Fix so we use a proper name ("pclk") with DT, and keep
current fuctionality without DT.
Signed-off-by: Sam Ravnborg <sam@ravnborg.org>
Acked-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Signed-off-by: Lucas Stach <l.stach@pengutronix.de>
|
|
|
|
|
|
| |
Acked-by: Sam Ravnborg <sam@ravnborg.org>
Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
dev_request_mem_region doesn't work properly one some SoCs on which
PTR_ERR() values clash with valid return values from dev_request_mem_region.
Replace them with dev_request_mem_resource where possible.
This patch has been generated with the following semantic patch:
// <smpl>
@@
expression d;
expression n;
expression io;
identifier func;
@@
func(...) {
+struct resource *iores;
<+...
-io = dev_request_mem_region(d, n);
-if (IS_ERR(io)) {
+iores = dev_request_mem_resource(d, n);
+if (IS_ERR(iores)) {
...
- return PTR_ERR(io);
-}
+ return PTR_ERR(iores);
+}
+io = IOMEM(iores->start);
...+>
}
@@
expression d;
expression n;
expression io;
identifier func;
@@
func(...) {
+struct resource *iores;
<+...
-io = dev_request_mem_region(d, n);
-if (IS_ERR(io)) {
+iores = dev_request_mem_resource(d, n);
+if (IS_ERR(iores))
- return PTR_ERR(io);
-}
+ return PTR_ERR(iores);
+io = IOMEM(iores->start);
...+>
}
@@
expression d;
expression n;
expression io;
identifier func;
@@
func(...) {
+struct resource *iores;
<+...
-io = dev_request_mem_region(d, n);
-if (IS_ERR(io)) {
- ret = PTR_ERR(io);
+iores = dev_request_mem_resource(d, n);
+if (IS_ERR(iores)) {
+ ret = PTR_ERR(iores);
...
}
+io = IOMEM(iores->start);
...+>
}
@@
expression d;
expression n;
expression io;
identifier func;
@@
func(...) {
+struct resource *iores;
<+...
-io = dev_request_mem_region(d, n);
+iores = dev_request_mem_resource(d, n);
+if (IS_ERR(iores))
+ return PTR_ERR(iores);
+io = IOMEM(iores->start);
...+>
}
@@
identifier func;
@@
func(...) {
<+...
struct resource *iores;
-struct resource *iores;
...+>
}
// </smpl>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
|
|
| |
The set_ethaddr callback should not modify the MAC address passed to it, so
make it const.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
|
|
|
|
| |
Move to the common streaming DMA ops in order to get rid of
the direct usage of the ARM MMU functions for the cache
maintenance.
Signed-off-by: Lucas Stach <dev@lynxeye.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
| |
This better separates the DMA from the MMU functionality.
Also move all drivers that only depends on asm/mmu.h for the alloc
functions over to the common header.
Signed-off-by: Lucas Stach <dev@lynxeye.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
|
|
|
|
| |
As a lot drivers currently rely on the 1:1 virt->phys mapping on ARM
we define DMA_ADDRESS_BROKEN to mark them. In order to use them on
other architectures with a different mapping they need proper fixing.
Signed-off-by: Lucas Stach <dev@lynxeye.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|\ |
|
| |
| |
| |
| |
| |
| | |
For all users fix or add the error check.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|/
|
|
|
|
|
|
|
|
|
|
| |
If the GEM is used, however only FE featuren enabled, then it can not
support FBLDO configured as 16, and no RGMII/GMII configuratin bit.
So, in order to support GEM only with FE feature enabled, we do:
- using default value for FBLDO.
- using same bit for RMII/MII configuration as GEM for RGMII/GMII.
Signed-off-by: Bo Shen <voice.shen@atmel.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
|
|
| |
So that barebox has the information which interface a packet
came from.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
|
|
| |
check the 4 mac address register and return at the first valid
Signed-off-by: Jean-Christophe PLAGNIOL-VILLARD <plagnioj@jcrosoft.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
| |
The core has a bit for swapping packet data endianism.
Reset default from Cadence is off. Xilinx however, that uses this core on the
Zynq SoCs, opted for on. Turn it off for all devices.
Signed-off-by: Steffen Trumtrar <s.trumtrar@pengutronix.de>
Cc: Jean-Christophe PLAGNIOL-VILLARD <plagnioj@jcrosoft.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
| |
The macb/gem core is used by the Zynq SoC. In preparation of sharing
the macb driver between at91 and Zynq, rename the platform data to
'struct macb_platform_data', and move the definition to a common
location.
Signed-off-by: Josh Cartwright <joshc@eso.teric.us>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|