--- /dev/null
+From 25ce45fe40b574e5d7ffa407f7f2db03e7d5a910 Mon Sep 17 00:00:00 2001
+From: Daniel Golle <daniel@makrotopia.org>
+Date: Tue, 22 Aug 2023 17:32:54 +0100
+Subject: [PATCH 112/250] net: ethernet: mtk_eth_soc: add support for in-SoC
+ SRAM
+
+MT7981, MT7986 and MT7988 come with in-SoC SRAM dedicated for Ethernet
+DMA rings. Support using the SRAM without breaking existing device tree
+bindings, ie. only new SoC starting from MT7988 will have the SRAM
+declared as additional resource in device tree. For MT7981 and MT7986
+an offset on top of the main I/O base is used.
+
+Signed-off-by: Daniel Golle <daniel@makrotopia.org>
+Link: https://lore.kernel.org/r/e45e0f230c63ad58869e8fe35b95a2fb8925b625.1692721443.git.daniel@makrotopia.org
+Signed-off-by: Jakub Kicinski <kuba@kernel.org>
+---
+ drivers/net/ethernet/mediatek/mtk_eth_soc.c | 88 ++++++++++++++++-----
+ drivers/net/ethernet/mediatek/mtk_eth_soc.h | 12 ++-
+ 2 files changed, 78 insertions(+), 22 deletions(-)
+
+--- a/drivers/net/ethernet/mediatek/mtk_eth_soc.c
++++ b/drivers/net/ethernet/mediatek/mtk_eth_soc.c
+@@ -1075,10 +1075,13 @@ static int mtk_init_fq_dma(struct mtk_et
+ dma_addr_t dma_addr;
+ int i;
+
+- eth->scratch_ring = dma_alloc_coherent(eth->dma_dev,
+- cnt * soc->txrx.txd_size,
+- ð->phy_scratch_ring,
+- GFP_KERNEL);
++ if (MTK_HAS_CAPS(eth->soc->caps, MTK_SRAM))
++ eth->scratch_ring = eth->sram_base;
++ else
++ eth->scratch_ring = dma_alloc_coherent(eth->dma_dev,
++ cnt * soc->txrx.txd_size,
++ ð->phy_scratch_ring,
++ GFP_KERNEL);
+ if (unlikely(!eth->scratch_ring))
+ return -ENOMEM;
+
+@@ -2376,8 +2379,14 @@ static int mtk_tx_alloc(struct mtk_eth *
+ if (!ring->buf)
+ goto no_tx_mem;
+
+- ring->dma = dma_alloc_coherent(eth->dma_dev, ring_size * sz,
+- &ring->phys, GFP_KERNEL);
++ if (MTK_HAS_CAPS(soc->caps, MTK_SRAM)) {
++ ring->dma = eth->sram_base + ring_size * sz;
++ ring->phys = eth->phy_scratch_ring + ring_size * (dma_addr_t)sz;
++ } else {
++ ring->dma = dma_alloc_coherent(eth->dma_dev, ring_size * sz,
++ &ring->phys, GFP_KERNEL);
++ }
++
+ if (!ring->dma)
+ goto no_tx_mem;
+
+@@ -2476,8 +2485,7 @@ static void mtk_tx_clean(struct mtk_eth
+ kfree(ring->buf);
+ ring->buf = NULL;
+ }
+-
+- if (ring->dma) {
++ if (!MTK_HAS_CAPS(soc->caps, MTK_SRAM) && ring->dma) {
+ dma_free_coherent(eth->dma_dev,
+ ring->dma_size * soc->txrx.txd_size,
+ ring->dma, ring->phys);
+@@ -2496,9 +2504,14 @@ static int mtk_rx_alloc(struct mtk_eth *
+ {
+ const struct mtk_reg_map *reg_map = eth->soc->reg_map;
+ struct mtk_rx_ring *ring;
+- int rx_data_len, rx_dma_size;
++ int rx_data_len, rx_dma_size, tx_ring_size;
+ int i;
+
++ if (MTK_HAS_CAPS(eth->soc->caps, MTK_QDMA))
++ tx_ring_size = MTK_QDMA_RING_SIZE;
++ else
++ tx_ring_size = MTK_DMA_SIZE;
++
+ if (rx_flag == MTK_RX_FLAGS_QDMA) {
+ if (ring_no)
+ return -EINVAL;
+@@ -2533,9 +2546,20 @@ static int mtk_rx_alloc(struct mtk_eth *
+ ring->page_pool = pp;
+ }
+
+- ring->dma = dma_alloc_coherent(eth->dma_dev,
+- rx_dma_size * eth->soc->txrx.rxd_size,
+- &ring->phys, GFP_KERNEL);
++ if (!MTK_HAS_CAPS(eth->soc->caps, MTK_SRAM) ||
++ rx_flag != MTK_RX_FLAGS_NORMAL) {
++ ring->dma = dma_alloc_coherent(eth->dma_dev,
++ rx_dma_size * eth->soc->txrx.rxd_size,
++ &ring->phys, GFP_KERNEL);
++ } else {
++ struct mtk_tx_ring *tx_ring = ð->tx_ring;
++
++ ring->dma = tx_ring->dma + tx_ring_size *
++ eth->soc->txrx.txd_size * (ring_no + 1);
++ ring->phys = tx_ring->phys + tx_ring_size *
++ eth->soc->txrx.txd_size * (ring_no + 1);
++ }
++
+ if (!ring->dma)
+ return -ENOMEM;
+
+@@ -2618,7 +2642,7 @@ static int mtk_rx_alloc(struct mtk_eth *
+ return 0;
+ }
+
+-static void mtk_rx_clean(struct mtk_eth *eth, struct mtk_rx_ring *ring)
++static void mtk_rx_clean(struct mtk_eth *eth, struct mtk_rx_ring *ring, bool in_sram)
+ {
+ int i;
+
+@@ -2641,7 +2665,7 @@ static void mtk_rx_clean(struct mtk_eth
+ ring->data = NULL;
+ }
+
+- if (ring->dma) {
++ if (!in_sram && ring->dma) {
+ dma_free_coherent(eth->dma_dev,
+ ring->dma_size * eth->soc->txrx.rxd_size,
+ ring->dma, ring->phys);
+@@ -3001,7 +3025,7 @@ static void mtk_dma_free(struct mtk_eth
+ for (i = 0; i < MTK_MAX_DEVS; i++)
+ if (eth->netdev[i])
+ netdev_reset_queue(eth->netdev[i]);
+- if (eth->scratch_ring) {
++ if (!MTK_HAS_CAPS(soc->caps, MTK_SRAM) && eth->scratch_ring) {
+ dma_free_coherent(eth->dma_dev,
+ MTK_QDMA_RING_SIZE * soc->txrx.txd_size,
+ eth->scratch_ring, eth->phy_scratch_ring);
+@@ -3009,13 +3033,13 @@ static void mtk_dma_free(struct mtk_eth
+ eth->phy_scratch_ring = 0;
+ }
+ mtk_tx_clean(eth);
+- mtk_rx_clean(eth, ð->rx_ring[0]);
+- mtk_rx_clean(eth, ð->rx_ring_qdma);
++ mtk_rx_clean(eth, ð->rx_ring[0], MTK_HAS_CAPS(soc->caps, MTK_SRAM));
++ mtk_rx_clean(eth, ð->rx_ring_qdma, false);
+
+ if (eth->hwlro) {
+ mtk_hwlro_rx_uninit(eth);
+ for (i = 1; i < MTK_MAX_RX_RING_NUM; i++)
+- mtk_rx_clean(eth, ð->rx_ring[i]);
++ mtk_rx_clean(eth, ð->rx_ring[i], false);
+ }
+
+ kfree(eth->scratch_head);
+@@ -4585,7 +4609,7 @@ static int mtk_sgmii_init(struct mtk_eth
+
+ static int mtk_probe(struct platform_device *pdev)
+ {
+- struct resource *res = NULL;
++ struct resource *res = NULL, *res_sram;
+ struct device_node *mac_np;
+ struct mtk_eth *eth;
+ int err, i;
+@@ -4605,6 +4629,20 @@ static int mtk_probe(struct platform_dev
+ if (MTK_HAS_CAPS(eth->soc->caps, MTK_SOC_MT7628))
+ eth->ip_align = NET_IP_ALIGN;
+
++ if (MTK_HAS_CAPS(eth->soc->caps, MTK_SRAM)) {
++ /* SRAM is actual memory and supports transparent access just like DRAM.
++ * Hence we don't require __iomem being set and don't need to use accessor
++ * functions to read from or write to SRAM.
++ */
++ if (mtk_is_netsys_v3_or_greater(eth)) {
++ eth->sram_base = (void __force *)devm_platform_ioremap_resource(pdev, 1);
++ if (IS_ERR(eth->sram_base))
++ return PTR_ERR(eth->sram_base);
++ } else {
++ eth->sram_base = (void __force *)eth->base + MTK_ETH_SRAM_OFFSET;
++ }
++ }
++
+ spin_lock_init(ð->page_lock);
+ spin_lock_init(ð->tx_irq_lock);
+ spin_lock_init(ð->rx_irq_lock);
+@@ -4668,6 +4706,18 @@ static int mtk_probe(struct platform_dev
+ err = -EINVAL;
+ goto err_destroy_sgmii;
+ }
++ if (MTK_HAS_CAPS(eth->soc->caps, MTK_SRAM)) {
++ if (mtk_is_netsys_v3_or_greater(eth)) {
++ res_sram = platform_get_resource(pdev, IORESOURCE_MEM, 1);
++ if (!res_sram) {
++ err = -EINVAL;
++ goto err_destroy_sgmii;
++ }
++ eth->phy_scratch_ring = res_sram->start;
++ } else {
++ eth->phy_scratch_ring = res->start + MTK_ETH_SRAM_OFFSET;
++ }
++ }
+ }
+
+ if (eth->soc->offload_version) {
+--- a/drivers/net/ethernet/mediatek/mtk_eth_soc.h
++++ b/drivers/net/ethernet/mediatek/mtk_eth_soc.h
+@@ -139,6 +139,9 @@
+ #define MTK_GDMA_MAC_ADRH(x) ({ typeof(x) _x = (x); (_x == MTK_GMAC3_ID) ? \
+ 0x54C : 0x50C + (_x * 0x1000); })
+
++/* Internal SRAM offset */
++#define MTK_ETH_SRAM_OFFSET 0x40000
++
+ /* FE global misc reg*/
+ #define MTK_FE_GLO_MISC 0x124
+
+@@ -935,6 +938,7 @@ enum mkt_eth_capabilities {
+ MTK_RSTCTRL_PPE1_BIT,
+ MTK_RSTCTRL_PPE2_BIT,
+ MTK_U3_COPHY_V2_BIT,
++ MTK_SRAM_BIT,
+
+ /* MUX BITS*/
+ MTK_ETH_MUX_GDM1_TO_GMAC1_ESW_BIT,
+@@ -970,6 +974,7 @@ enum mkt_eth_capabilities {
+ #define MTK_RSTCTRL_PPE1 BIT_ULL(MTK_RSTCTRL_PPE1_BIT)
+ #define MTK_RSTCTRL_PPE2 BIT_ULL(MTK_RSTCTRL_PPE2_BIT)
+ #define MTK_U3_COPHY_V2 BIT_ULL(MTK_U3_COPHY_V2_BIT)
++#define MTK_SRAM BIT_ULL(MTK_SRAM_BIT)
+
+ #define MTK_ETH_MUX_GDM1_TO_GMAC1_ESW \
+ BIT_ULL(MTK_ETH_MUX_GDM1_TO_GMAC1_ESW_BIT)
+@@ -1045,14 +1050,14 @@ enum mkt_eth_capabilities {
+ #define MT7981_CAPS (MTK_GMAC1_SGMII | MTK_GMAC2_SGMII | MTK_GMAC2_GEPHY | \
+ MTK_MUX_GMAC12_TO_GEPHY_SGMII | MTK_QDMA | \
+ MTK_MUX_U3_GMAC2_TO_QPHY | MTK_U3_COPHY_V2 | \
+- MTK_RSTCTRL_PPE1)
++ MTK_RSTCTRL_PPE1 | MTK_SRAM)
+
+ #define MT7986_CAPS (MTK_GMAC1_SGMII | MTK_GMAC2_SGMII | \
+ MTK_MUX_GMAC12_TO_GEPHY_SGMII | MTK_QDMA | \
+- MTK_RSTCTRL_PPE1)
++ MTK_RSTCTRL_PPE1 | MTK_SRAM)
+
+ #define MT7988_CAPS (MTK_GDM1_ESW | MTK_QDMA | MTK_RSTCTRL_PPE1 | \
+- MTK_RSTCTRL_PPE2)
++ MTK_RSTCTRL_PPE2 | MTK_SRAM)
+
+ struct mtk_tx_dma_desc_info {
+ dma_addr_t addr;
+@@ -1212,6 +1217,7 @@ struct mtk_eth {
+ struct device *dev;
+ struct device *dma_dev;
+ void __iomem *base;
++ void *sram_base;
+ spinlock_t page_lock;
+ spinlock_t tx_irq_lock;
+ spinlock_t rx_irq_lock;
projects / openwrt / staging / dangole.git / blobdiff