| Age | Commit message (Collapse) | Author |
|---|
|
This bit is necessary to receive packets from the internal PHY.
Without this bit set, no activity occurs on the interface.
Normally u-boot sets this bit, but if u-boot is compiled without
net support, the interface will be up but without any activity.
The vendor SDK sets this bit along with the PHY_ID bits.
Ported from the Linux change at [1] from Da Xu merged in
commit [2].
[1] https://lore.kernel.org/all/[email protected]/
[2] b23285e93bef ("net: mdio: mux-meson-gxl: set reversed bit when using internal phy")
Suggested-by: Da Xue <[email protected]>
Link: https://lore.kernel.org/r/20250502-u-boot-topic-mdio-mux-gxl-bit28-v1-1-399f6c3db154@linaro.org
Signed-off-by: Neil Armstrong <[email protected]>
|
|
Prepare v2025.10-rc5
|
|
ethernet PHY (v4)"
Weijie Gao <[email protected]> says:
This patch adds PHY driver for MediaTek MT7987/MT7988 built-in 2.5Gb
ethernet PHY.
[trini: Change 'tristate' Kconfig to 'bool']
Link: https://lore.kernel.org/r/[email protected]
|
|
The MediaTek MT7987/MT7988 SoCs features a built-in 2.5Gb PHY
connected to GMAC1. The PHY supports 10/100/1000/2500 Mbps
full-duplex only.
The PHY requires one or two firmware files. Firmware for MT7988 has
already been added to upstream: mediatek/mt7988/i2p5ge-phy-pmb.bin.
MT7987 has two firmware files which will be add to upstream later:
i2p5ge-phy-pmb.bin and i2p5ge-phy-DSPBitTb.bin.
Environment variable can be set for firmware data loading:
mt7987_i2p5ge_load_pmb_firmware for i2p5ge-phy-pmb.bin
mt7987_i2p5ge_load_dspbit_firmware for i2p5ge-phy-DSPBitTb.bin
mt7988_i2p5ge_load_pmb_firmware for i2p5ge-phy-pmb.bin
This driver allows dedicated weak functions to be overridden by
board to provide the firmware data:
mt7987_i2p5ge_get_fw() for MT7987
mt7988_i2p5ge_get_fw() for MT7988
To enable the PHY, add the following not to device tree:
ð1 {
status = "okay";
phy-mode = "xgmii";
phy-handle = <&phy15>;
phy15: ethernet-phy@15 {
compatible = "ethernet-phy-ieee802.3-c45";
reg = <15>;
phy-mode = "xgmii";
};
};
Signed-off-by: Sky Huang <[email protected]>
Signed-off-by: Weijie Gao <[email protected]>
|
|
Associate PHY device with its device node specified by phy-handle
property. This makes it possible for PHY drivers to read dedicated
information to configure the PHY device.
Signed-off-by: Weijie Gao <[email protected]>
|
|
Greg Malysa <[email protected]> says:
Between 2025.07 and 2025.10 many header dependency chains were improved,
but this exposed implicit header usage in several of our drivers. This
wasn't discovered before or included in the original fixes because our
drivers are not yet used by any mainline-supported boards, so build
tests did not find them. This series addresses the two build failures
I've encountered while rebasing our work onto 2025.10 and continuing to
prepare the next submission of our board files.
Link: https://lore.kernel.org/r/[email protected]
|
|
Following header dependency cleanups, an implicit dependence on env.h
was exposed in dwc_eth_qos_adi. However because this driver is not (yet)
enabled in any defconfigs, build tests did not identify the missing
header. This adds the missing #include so that the driver builds
correctly when enabled.
Signed-off-by: Greg Malysa <[email protected]>
|
|
The layout of SYSCFG_ETHnCR on STM32MP25xx is slightly different yet again.
Add missing swizzling to program the correct register with the correct content.
Fixes: 20afca89ed53 ("net: dwc_eth_qos: add support of stm32mp2 platform")
Signed-off-by: Marek Vasut <[email protected]>
Reviewed-by: Christophe ROULLIER<[email protected]>
|
|
Prepare v2025.10-rc4
|
|
Add a MAINTAINERS entry for the DesignWare XGMAC network driver to
ensure future patches are properly routed for review and support.
Signed-off-by: Boon Khai Ng <[email protected]>
Reviewed-by: Tom Rini <[email protected]>
|
|
In mc_fixup_dpc_mac_addr noff is assigned the return value from
fdt_add_subnode so that is the variable that should be passed to
fdt_strerror and returned when negative.
This issue was found by Smatch.
Signed-off-by: Andrew Goodbody <[email protected]>
Signed-off-by: Peng Fan <[email protected]>
|
|
In dpni_exit there is a NULL check for dflt_dpni after it is
dereferenced a number of times. Instead move the NULL check to early in
the function. Also assign NULL to dflt_dpni after free in both dpni_init
and dpni_exit.
This issue was found by Smatch.
Signed-off-by: Andrew Goodbody <[email protected]>
Signed-off-by: Peng Fan <[email protected]>
|
|
In dpbp_exit there is a NULL check for dflt_dpbp after it is
dereferenced a number of times. Instead move the NULL check to early in
the function. Also assign NULL to dflt_dpbp after free in both dpbp_init
and dpbp_exit.
This issue was found by Smatch.
Signed-off-by: Andrew Goodbody <[email protected]>
Signed-off-by: Peng Fan <[email protected]>
|
|
In dpio_exit there is a NULL check for dflt_dpio but it happens after
dpio_dflt has been dereferenced a number of times already. Instead move
the NULL check to first thing in the function. Also assign NULL to
dflt_dpio after free in both dpio_init and dpio_exit.
This issue was found by Smatch.
Signed-off-by: Andrew Goodbody <[email protected]>
Signed-off-by: Peng Fan <[email protected]>
|
|
In netc_blk_ctrl_probe the test for failure of the function
clk_prepare_enable should not return PTR_ERR(ipg_clk) as it does not
check IS_ERR(ipg_clk) instead it should return err as that is what is
holding the error code in this case.
This issue was found by Smatch.
Signed-off-by: Andrew Goodbody <[email protected]>
Signed-off-by: Peng Fan <[email protected]>
|
|
In memac_wait_until_free and memac_wait_until_done the use of
post-decrement on the test in the while loop for a timeout means that
timeout will be equal to -1 on exit in that case. Adjust the test for
this expected value.
This issue was found by Smatch.
Signed-off-by: Andrew Goodbody <[email protected]>
Signed-off-by: Peng Fan <[email protected]>
|
|
In fm_eth_bind there is a dereference of dev before it is NULL checked.
Add a NULL check before the first dereference and remove a later NULL
check that is now redundant.
This issue was found by Smatch.
Signed-off-by: Andrew Goodbody <[email protected]>
Signed-off-by: Peng Fan <[email protected]>
|
|
In bmi_rx_port_disable and bmi_tx_port_disable the use of post-decrement
on the test in the while loop for a timeout means that timeout will be
equal to -1 on exit in that case. Adjust the test for this expected
value.
This issue was found by Smatch.
Signed-off-by: Andrew Goodbody <[email protected]>
Signed-off-by: Peng Fan <[email protected]>
|
|
Prepare v2025.10-rc3
|
|
The timeout test in axi_dma_init is not correct due to the
post-decrement used on the timeout variable which will mean timeout is
not 0 if the timeout occurs. Make the timeout variable an int instead of
a u32 and then test for timeout being -1.
This issue was found by Smatch.
Signed-off-by: Andrew Goodbody <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Michal Simek <[email protected]>
|
|
Chintan Vankar <[email protected]> says:
This series adds bind method for CPSW to avoid explicit probing, removes
explicit probing of CPSW, adds support for Ethernet boot on SK-AM68,
SK-AM62P-LP, J722S, SK-AM69.
Link: https://lore.kernel.org/r/[email protected]
|
|
TI's Ethernet switch needs system controllers enabled in R5SPL stage while
booting via Ethernet. Enable SPL_SYSCON config for
CONFIG_TI_AM65_CPSW_NUSS.
Reviewed-by: Alexander Sverdlin <[email protected]>
Signed-off-by: Chintan Vankar <[email protected]>
|
|
CPSW driver is defined as UCLASS_MISC driver which needs to be probed
explicitly. Define bind method for CPSW driver to scan and bind
ethernet-ports with UCLASS_ETH driver which will eventually probe CPSW
driver and avoid probing CPSW driver explicitly.
Signed-off-by: Chintan Vankar <[email protected]>
|
|
In ks8851_mll_detect_chip the if..else code detects the case of (val &
0xfff0) != CIDER_ID and returns if found. So testing for this again will
always fail and the code is unreachable. Just remove the test and code
block.
This issue was found by Smatch.
Signed-off-by: Andrew Goodbody <[email protected]>
|
|
In ca_phy_probe when checking for an external phy it uses a field from
the internal phy due to what is assumed to be a copy/paste typo. Make
the obvious fix to use the field from the external phy.
This issue was found by Smatch.
Signed-off-by: Andrew Goodbody <[email protected]>
|
|
In rvu_af_init if the code fails to allocate memory for nix_af it will
take the error path with nix_af == NULL which will dereference nix_af.
Add the appropriate NULL check.
This issue was found by Smatch.
Signed-off-by: Andrew Goodbody <[email protected]>
|
|
In nix_lf_setup there is a default value assigned to err in case an
error is detected. However this default value will be overwritten in the
for loop so that later code does not return an error code from the
function. Add a new assignment to restore err to the default error code.
This issue was found by Smatch.
Signed-off-by: Andrew Goodbody <[email protected]>
|
|
In mvpp2_probe the code attempts to get a value for "gop-port-id" and
assigns it to port->gop_id but it then tests port->id for being equal to
-1. That is an impossible test as port->id is a field of type u8 so
cannot be negative. Change the test to port->gop_id.
This issue was found by Smatch.
Signed-off-by: Andrew Goodbody <[email protected]>
|
|
In phy_info_parse all uses of the variable phyaddr are as an int so
declaring as u32 is not useful and prevents the test for an error return
from fdtdec_get_int ever detecting an error. Change phyaddr to be an
int.
This issue was found by Smatch.
Signed-off-by: Andrew Goodbody <[email protected]>
|
|
Instead of returning -1 on a failed alloc, return -ENOMEM.
This issue was found by Smatch.
Signed-off-by: Andrew Goodbody <[email protected]>
|
|
You cannot test an unsigned char to be >= 256. Instead make the
variables start and end to be ints.
This issue was found by Smatch.
Signed-off-by: Andrew Goodbody <[email protected]>
|
|
In mv88e6xxx_port_enable when attempting to mask out the previous
settings of two bits a logical operator was used instead of a bitwise
operator. Fix this.
This issue was found by Smatch.
Signed-off-by: Andrew Goodbody <[email protected]>
|
|
In mtk_eth_of_to_plat, the last error check has the value in
'priv->phy_addr' but returns ret. Correct to return 'priv->phy_addr'
instead.
This issue was found by Smatch.
Signed-off-by: Andrew Goodbody <[email protected]>
|
|
In ldpaa_eth_open a memset is used to initialise a struct to 0 but the
size passed is that of a different struct. Correct to pass the sizeof
the struct that is being initialised.
This issue was found by Smatch.
Signed-off-by: Andrew Goodbody <[email protected]>
|
|
It's Broadcom PHY simply described as single-port
RGMII 10/100/1000BASE-T PHY.
Signed-off-by: Jim Liu <[email protected]>
|
|
get_timer returns a ulong value representing system time in ms.
On a 64-bit system, this ulong value is 64 bits long. However,
the driver stores it in a 32-bit unsigned integer, which overflows
after 49 days up time, causing the driver to get an incorrect time.
Replace the unsigned int variable with a ulong type to properly store
the value returned by get_timer.
Signed-off-by: Stanley Chu <[email protected]>
Signed-off-by: Jim Liu <[email protected]>
|
|
In vsc8514_config there is a while loop for detecting a config failure
using a timeout counter with a post-decrement. In the case of a timeout
this will result in the loop exiting with timeout == -1 so use that as
the test below the loop to detect that the timeout occurred.
This issue was found by Smatch.
Signed-off-by: Andrew Goodbody <[email protected]>
Reviewed-by: Quentin Schulz <[email protected]>
|
|
In octeontx_smi_probe if an error is detected then memory that was
allocated is not freed. Small refactor of the code to use a common
return and free memory. Also return -ENOMEM for an allocation failure.
This issue was found by Smatch.
Signed-off-by: Andrew Goodbody <[email protected]>
|
|
In nicvf_cq_handler there is a test for !cqe_count which will return if
true so it is guaranteed that cqe_count will true after that point. This
makes the later test for cqe_count redundant so it can be removed.
This issue was found by Smatch.
Signed-off-by: Andrew Goodbody <[email protected]>
|
|
In nicvf_rcv_pkt_handler there is no need to initialise err as it is
assigned to immediately after. Also the test for !pkt will return if
true meaning that pkt is guaranteed to be true after that code block and
so no need to test for it and the redundant test can be removed.
This issue was found by Smatch.
Signed-off-by: Andrew Goodbody <[email protected]>
|
|
port_id is an unsigned variable so cannot be negative. Remove the test
checking for port_id being less than 0.
This issue was found by Smatch.
Signed-off-by: Andrew Goodbody <[email protected]>
|
|
In am65_cpsw_phy_init it is not certain that ret will be assigned to
before it reaches the 'return ret' statement. Initialise ret to ensure
that ret is valid.
This issue was found by Smatch.
Signed-off-by: Andrew Goodbody <[email protected]>
|
|
In cs4340_upload_firmware a buffer is allocated with malloc but this is
never freed. The pointer to this buffer, addr, is not even kept
unchanged. But in some cases addr is not a buffer allocated by malloc.
Introduce the use of another pointer to keep track of the buffer and to
know if it needs to be freed.
This issue was found by Smatch.
Signed-off-by: Andrew Goodbody <[email protected]>
|
|
In do_e1000_spi_checksum a temporary buffer is allocated but never
freed. Add code to free on exit. Also refactor the code to make the exit
code common.
This issue found by Smatch.
Signed-off-by: Andrew Goodbody <[email protected]>
|
|
https://source.denx.de/u-boot/custodians/u-boot-socfpga
This pull request introduces initial U-Boot support for Agilex7 M-series, along
with several enhancements and cleanups across existing Agilex platforms. Key
changes include new board support, DDR driver additions, updated device trees,
and broader SoCFPGA SPL improvements.
Highlights:
- Agilex7 M-series bring-up:
- Basic DT support and board initialization for Agilex7 M-series SoC and
SoCDK.
- New sdram_agilex7m DDR driver with UIBSSM mailbox support and HBM support.
- Clock driver support for Agilex7 M-series.
- New defconfig: socfpga_agilex7m_defconfig.
- Agilex and Agilex5 enhancements:
- Improved SPL support: ASYNC interrupt enabling, system manager init
refactor, and cold scratch register usage.
- Updated firewall probing and watchdog support in SPL.
- Cleaned up DDR code, added secure region support for ATF, and improved warm
reset handling.
- Device Tree and config updates:
- Migration to upstream Linux DT layout for Agilex platforms.
- Consolidated socfpga_agilex_defconfig and removed deprecated configs.
- Platform-specific environment variables for Distro Boot added.
- Driver fixes and cleanups:
- dwc_eth_xgmac and clk-agilex cleanup and improvements.
- Several coverity and style fixes.
Contributions in this PR are from Alif Zakuan Yuslaimi, Tingting Meng, and
Andrew Goodbody. This patch set has been tested on Agilex 5 devkit, Agilex
devkit and Agilex7m devkit.
Passing all pipeline tests at SoCFPGA U-boot custodian
https://source.denx.de/u-boot/custodians/u-boot-socfpga/-/pipelines/27318
|
|
In dwxgmac_of_get_mac_mode there is a test for mac_mode which will
return if false. After this point mac_mode is guaranteed to be true so
there is no need to test for this. Remove that test.
This issue was found by Smatch.
Signed-off-by: Andrew Goodbody <[email protected]>
Reviewed-by: Tien Fong Chee <[email protected]>
|
|
In xgmac_probe there is a direct return after the point where unwind
gotos start to be used to undo actions performed by earlier code. Use
the appropriate unwind goto instead.
This issue was found by Smatch.
Signed-off-by: Andrew Goodbody <[email protected]>
Reviewed-by: Tien Fong Chee <[email protected]>
|
|
In fecmxc_probe if a timeout is detected when resetting the chip no
error code is set before taking the error exit. This could lead to a
silent failure. Instead set an error code.
This issue was found by Smatch.
Signed-off-by: Andrew Goodbody <[email protected]>
|
|
In rswitch_probe the error detection after the call to devm_clk_get is
very wrong. It checks the value of ret which is uninitialised at that
point. Instead it should be using the macros for including errors into
pointers.
This issue was found by Smatch.
Signed-off-by: Andrew Goodbody <[email protected]>
Reviewed-by: Marek Vasut <[email protected]>
|
|
The Auto Negotiation procedure between two Ethernet PHYs consists of
determining the best commonly supported parameters among Speed,
Duplex Mode and Flow Control.
The time taken for this procedure is not only dependent on the local
PHY used, but also on the link-partner PHY.
While a timeout can be specified in the form of a "CONFIG" on the basis
of the local PHY present on the device, since the timeout also depends
on the link-partner PHY, it might be necessary to modify the timeout.
To avoid rebuilding the bootloader for a given device, just because it
may be connected to various link-partner PHYs, each with a different
timeout, introduce an environment variable named "phy_aneg_timeout" and
override "CONFIG_PHY_ANEG_TIMEOUT" with "phy_aneg_timeout".
Signed-off-by: Siddharth Vadapalli <[email protected]>
Acked-by: Jerome Forissier <[email protected]>
[jf: add missing #include <env.h>]
Signed-off-by: Jerome Forissier <[email protected]>
|
