[PATCH v2] net: dwc_eth_qos: Fix support for stm32mp2 platform
Christophe ROULLIER
christophe.roullier at foss.st.com
Tue Sep 16 08:53:58 CEST 2025
On 9/15/25 02:53, Marek Vasut wrote:
> 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<marek.vasut at mailbox.org>
> ---
> Cc: Christophe Roullier<christophe.roullier at foss.st.com>
> Cc: Joe Hershberger<joe.hershberger at ni.com>
> Cc: Patrice Chotard<patrice.chotard at foss.st.com>
> Cc: Patrick Delaunay<patrick.delaunay at foss.st.com>
> Cc: Ramon Fried<rfried.dev at gmail.com>
> Cc: Tom Rini<trini at konsulko.com>
> Cc:u-boot at lists.denx.de
> Cc:uboot-stm32 at st-md-mailman.stormreply.com
> ---
> V2: Gate off SELMII bit on both MP13 and MP2
> ---
> drivers/net/dwc_eth_qos_stm32.c | 44 ++++++++++++++++++++++++++-------
> 1 file changed, 35 insertions(+), 9 deletions(-)
>
> diff --git a/drivers/net/dwc_eth_qos_stm32.c b/drivers/net/dwc_eth_qos_stm32.c
> index 7a28f2a3e83..57ac528ebf5 100644
> --- a/drivers/net/dwc_eth_qos_stm32.c
> +++ b/drivers/net/dwc_eth_qos_stm32.c
> @@ -137,13 +137,15 @@ static int eqos_probe_syscfg_stm32(struct udevice *dev,
> const bool eth_ref_clk_sel = dev_read_bool(dev, "st,eth-ref-clk-sel");
> /* SoC is STM32MP13xx with two ethernet MACs */
> const bool is_mp13 = device_is_compatible(dev, "st,stm32mp13-dwmac");
> + /* SoC is STM32MP25xx with two ethernet MACs */
> + const bool is_mp2 = device_is_compatible(dev, "st,stm32mp25-dwmac");
> /* Gigabit Ethernet 125MHz clock selection. */
> const bool eth_clk_sel = dev_read_bool(dev, "st,eth-clk-sel");
> /* Ethernet clock source is RCC. */
> const bool ext_phyclk = dev_read_bool(dev, "st,ext-phyclk");
> struct regmap *regmap;
> u32 regmap_mask;
> - u32 value;
> + u32 reg, value;
>
> regmap = syscon_regmap_lookup_by_phandle(dev, "st,syscon");
> if (IS_ERR(regmap))
> @@ -163,7 +165,7 @@ static int eqos_probe_syscfg_stm32(struct udevice *dev,
> * acts as a selector between 0:GMII and 1:MII. As STM32MP13xx
> * supports only MII, ETH_SELMII is not present.
> */
> - if (!is_mp13) /* Select MII mode on STM32MP15xx */
> + if (!is_mp13 && !is_mp2) /* Select MII mode on STM32MP15xx */
> value |= SYSCFG_PMCSETR_ETH_SELMII;
> break;
> case PHY_INTERFACE_MODE_GMII: /* STM32MP15xx only */
> @@ -213,15 +215,39 @@ static int eqos_probe_syscfg_stm32(struct udevice *dev,
> return -EINVAL;
> }
>
> - /* Shift value at correct ethernet MAC offset in SYSCFG_PMCSETR */
> - value <<= ffs(regmap_mask) - ffs(SYSCFG_PMCSETR_ETH1_MASK);
> + if (is_mp2) {
> + /*
> + * STM32MP25xx SYSCFG EthernetN control register
> + * has ETHn_CLK_SEL and ETHn_REF_CLK_SEL bits
> + * swapped, swizzle them back in place. Bitfield
> + * ETHn_SEL is shifted by 1 right, fix this up.
> + */
> + value = ((value & SYSCFG_PMCSETR_ETH_SEL_MASK) >> 1) |
> + ((value & SYSCFG_PMCSETR_ETH_CLK_SEL) << 1) |
> + ((value & SYSCFG_PMCSETR_ETH_REF_CLK_SEL) >> 1);
> +
> + /*
> + * STM32MP25xx SYSCFG EthernetN control register
> + * bits always start at bit shift 0 and there is
> + * one register for each MAC, shift the register
> + * content in place.
> + */
> + value >>= ffs(SYSCFG_PMCSETR_ETH1_MASK) - 1;
> +
> + reg = dev_read_u32_index_default(dev, "st,syscon", 1, 0);
>
> - /* Update PMCCLRR (clear register) */
> - regmap_write(regmap, is_mp13 ?
> - SYSCFG_PMCCLRR_MP13 : SYSCFG_PMCCLRR_MP15,
> - regmap_mask);
> + return regmap_write(regmap, reg, value);
> + } else {
> + /* Shift value at correct ethernet MAC offset in SYSCFG_PMCSETR */
> + value <<= ffs(regmap_mask) - ffs(SYSCFG_PMCSETR_ETH1_MASK);
>
> - return regmap_update_bits(regmap, SYSCFG_PMCSETR, regmap_mask, value);
> + /* Update PMCCLRR (clear register) */
> + regmap_write(regmap, is_mp13 ?
> + SYSCFG_PMCCLRR_MP13 : SYSCFG_PMCCLRR_MP15,
> + regmap_mask);
> +
> + return regmap_update_bits(regmap, SYSCFG_PMCSETR, regmap_mask, value);
> + }
> }
>
> static int eqos_probe_resources_stm32(struct udevice *dev)
Reviewed-by: Christophe ROULLIER<christophe.roullier at foss.st.com>
More information about the U-Boot
mailing list