| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Message-Id: <20241120105625.22508-5-Frank.Sae@motor-comm.com>
Date: Wed, 20 Nov 2024 18:56:08 +0800
From: Frank Sae <Frank.Sae@...or-comm.com>
To: davem@...emloft.net,
edumazet@...gle.com,
kuba@...nel.org,
pabeni@...hat.com
Cc: netdev@...r.kernel.org,
linux-kernel@...r.kernel.org,
xiaogang.fan@...or-comm.com,
fei.zhang@...or-comm.com,
hua.sun@...or-comm.com,
Frank.Sae@...or-comm.com
Subject: [PATCH net-next v2 04/21] motorcomm:yt6801: Implement the .ndo_open function
Implement the .ndo_open function to Calculate the Rx buffer sizeļ¼ allocate the
channels and rings.
Signed-off-by: Frank Sae <Frank.Sae@...or-comm.com>
---
.../ethernet/motorcomm/yt6801/yt6801_desc.c | 273 ++++++++++++++++++
.../ethernet/motorcomm/yt6801/yt6801_net.c | 99 +++++++
2 files changed, 372 insertions(+)
diff --git a/drivers/net/ethernet/motorcomm/yt6801/yt6801_desc.c b/drivers/net/ethernet/motorcomm/yt6801/yt6801_desc.c
index 476cf6633..2edf53d9b 100644
--- a/drivers/net/ethernet/motorcomm/yt6801/yt6801_desc.c
+++ b/drivers/net/ethernet/motorcomm/yt6801/yt6801_desc.c
@@ -47,3 +47,276 @@ void fxgmac_desc_data_unmap(struct fxgmac_pdata *pdata,
desc_data->mapped_as_page = 0;
}
+
+static void fxgmac_ring_free(struct fxgmac_pdata *pdata,
+ struct fxgmac_ring *ring)
+{
+ if (!ring)
+ return;
+
+ if (ring->desc_data_head) {
+ for (u32 i = 0; i < ring->dma_desc_count; i++)
+ fxgmac_desc_data_unmap(pdata,
+ FXGMAC_GET_DESC_DATA(ring, i));
+
+ kfree(ring->desc_data_head);
+ ring->desc_data_head = NULL;
+ }
+
+ if (ring->rx_hdr_pa.pages) {
+ dma_unmap_page(pdata->dev, ring->rx_hdr_pa.pages_dma,
+ ring->rx_hdr_pa.pages_len, DMA_FROM_DEVICE);
+ put_page(ring->rx_hdr_pa.pages);
+
+ ring->rx_hdr_pa.pages = NULL;
+ ring->rx_hdr_pa.pages_len = 0;
+ ring->rx_hdr_pa.pages_offset = 0;
+ ring->rx_hdr_pa.pages_dma = 0;
+ }
+
+ if (ring->rx_buf_pa.pages) {
+ dma_unmap_page(pdata->dev, ring->rx_buf_pa.pages_dma,
+ ring->rx_buf_pa.pages_len, DMA_FROM_DEVICE);
+ put_page(ring->rx_buf_pa.pages);
+
+ ring->rx_buf_pa.pages = NULL;
+ ring->rx_buf_pa.pages_len = 0;
+ ring->rx_buf_pa.pages_offset = 0;
+ ring->rx_buf_pa.pages_dma = 0;
+ }
+ if (ring->dma_desc_head) {
+ dma_free_coherent(pdata->dev, (sizeof(struct fxgmac_dma_desc) *
+ ring->dma_desc_count), ring->dma_desc_head,
+ ring->dma_desc_head_addr);
+ ring->dma_desc_head = NULL;
+ }
+}
+
+static int fxgmac_ring_init(struct fxgmac_pdata *pdata,
+ struct fxgmac_ring *ring,
+ unsigned int dma_desc_count)
+{
+ /* Descriptors */
+ ring->dma_desc_count = dma_desc_count;
+ ring->dma_desc_head =
+ dma_alloc_coherent(pdata->dev, (sizeof(struct fxgmac_dma_desc) *
+ dma_desc_count),
+ &ring->dma_desc_head_addr, GFP_KERNEL);
+ if (!ring->dma_desc_head)
+ return -ENOMEM;
+
+ /* Array of descriptor data */
+ ring->desc_data_head = kcalloc(dma_desc_count,
+ sizeof(struct fxgmac_desc_data),
+ GFP_KERNEL);
+ if (!ring->desc_data_head)
+ return -ENOMEM;
+
+ yt_dbg(pdata,
+ "dma_desc_head=%p, dma_desc_head_addr=%pad, desc_data_head=%p\n",
+ ring->dma_desc_head, &ring->dma_desc_head_addr,
+ ring->desc_data_head);
+ return 0;
+}
+
+static void fxgmac_rings_free(struct fxgmac_pdata *pdata)
+{
+ struct fxgmac_channel *channel = pdata->channel_head;
+
+ fxgmac_ring_free(pdata, channel->tx_ring);
+
+ for (u32 i = 0; i < pdata->channel_count; i++, channel++)
+ fxgmac_ring_free(pdata, channel->rx_ring);
+}
+
+static int fxgmac_rings_alloc(struct fxgmac_pdata *pdata)
+{
+ struct fxgmac_channel *channel = pdata->channel_head;
+ int ret;
+
+ yt_dbg(pdata, "%s - Tx ring:\n", channel->name);
+ ret = fxgmac_ring_init(pdata, channel->tx_ring,
+ pdata->tx_desc_count);
+ if (ret < 0) {
+ yt_err(pdata, "error initializing Tx ring");
+ goto err_init_ring;
+ }
+ if (netif_msg_drv(pdata))
+ yt_dbg(pdata,
+ "%s, ch=%u,tx_desc_cnt=%u\n",
+ __func__, 0, pdata->tx_desc_count);
+
+ for (u32 i = 0; i < pdata->channel_count; i++, channel++) {
+ yt_dbg(pdata, "%s - Rx ring:\n", channel->name);
+ ret = fxgmac_ring_init(pdata, channel->rx_ring,
+ pdata->rx_desc_count);
+ if (ret < 0) {
+ yt_err(pdata, "error initializing Rx ring\n");
+ goto err_init_ring;
+ }
+ if (netif_msg_drv(pdata))
+ yt_dbg(pdata,
+ "%s, ch=%u,rx_desc_cnt=%u\n",
+ __func__, i,
+ pdata->rx_desc_count);
+ }
+
+ if (netif_msg_drv(pdata))
+ yt_dbg(pdata, "%s ok\n", __func__);
+
+ return 0;
+
+err_init_ring:
+ fxgmac_rings_free(pdata);
+ return ret;
+}
+
+static void fxgmac_channels_free(struct fxgmac_pdata *pdata)
+{
+ struct fxgmac_channel *channel = pdata->channel_head;
+
+ if (netif_msg_drv(pdata))
+ yt_dbg(pdata, "free_channels,tx_ring=%p\n", channel->tx_ring);
+
+ kfree(channel->tx_ring);
+ channel->tx_ring = NULL;
+ if (netif_msg_drv(pdata))
+ yt_dbg(pdata, "free_channels,rx_ring=%p\n", channel->rx_ring);
+
+ kfree(channel->rx_ring);
+ channel->rx_ring = NULL;
+ if (netif_msg_drv(pdata))
+ yt_dbg(pdata, "free_channels,channel=%p\n", channel);
+
+ kfree(channel);
+ pdata->channel_head = NULL;
+}
+
+#ifdef CONFIG_PCI_MSI
+static void fxgmac_set_msix_tx_irq(struct fxgmac_pdata *pdata,
+ struct fxgmac_channel *channel, u32 i)
+{
+ if (!FXGMAC_IS_CHANNEL_WITH_TX_IRQ(i))
+ return;
+
+ pdata->channel_irq[FXGMAC_MAX_DMA_RX_CHANNELS] =
+ pdata->msix_entries[FXGMAC_MAX_DMA_RX_CHANNELS].vector;
+ channel->dma_irq_tx = pdata->channel_irq[FXGMAC_MAX_DMA_RX_CHANNELS];
+
+ yt_dbg(pdata, "%s, for MSIx, channel %d dma_irq_tx=%u\n", __func__, i,
+ channel->dma_irq_tx);
+}
+#endif
+
+static int fxgmac_channels_alloc(struct fxgmac_pdata *pdata)
+{
+ struct fxgmac_channel *channel_head, *channel;
+ struct fxgmac_ring *tx_ring, *rx_ring;
+ int ret = -ENOMEM;
+
+ channel_head = kcalloc(pdata->channel_count,
+ sizeof(struct fxgmac_channel), GFP_KERNEL);
+ if (netif_msg_drv(pdata))
+ yt_dbg(pdata, "alloc_channels,channel_head=%p,size=%d*%ld\n",
+ channel_head, pdata->channel_count,
+ sizeof(struct fxgmac_channel));
+
+ if (!channel_head)
+ return ret;
+
+ tx_ring = kcalloc(FXGMAC_TX_1_RING, sizeof(struct fxgmac_ring),
+ GFP_KERNEL);
+ if (!tx_ring)
+ goto err_tx_ring;
+
+ if (netif_msg_drv(pdata))
+ yt_dbg(pdata, "alloc_channels,tx_ring=%p,size=%d*%ld\n",
+ tx_ring, FXGMAC_TX_1_RING,
+ sizeof(struct fxgmac_ring));
+
+ rx_ring = kcalloc(pdata->rx_ring_count, sizeof(struct fxgmac_ring),
+ GFP_KERNEL);
+ if (!rx_ring)
+ goto err_rx_ring;
+
+ if (netif_msg_drv(pdata))
+ yt_dbg(pdata, "alloc_channels,rx_ring=%p,size=%d*%ld\n",
+ rx_ring, pdata->rx_ring_count,
+ sizeof(struct fxgmac_ring));
+
+ channel = channel_head;
+ for (u32 i = 0; i < pdata->channel_count; i++, channel++) {
+ snprintf(channel->name, sizeof(channel->name), "channel-%u", i);
+ channel->pdata = pdata;
+ channel->queue_index = i;
+ channel->dma_reg_offset = DMA_CH_BASE + (DMA_CH_INC * i);
+
+ if (pdata->per_channel_irq) {
+ pdata->channel_irq[i] = pdata->msix_entries[i].vector;
+#ifdef CONFIG_PCI_MSI
+ fxgmac_set_msix_tx_irq(pdata, channel, i);
+#endif
+
+ /* Get the per DMA rx interrupt */
+ ret = pdata->channel_irq[i];
+ if (ret < 0) {
+ yt_err(pdata, "get_irq %u err\n", i + 1);
+ goto err_irq;
+ }
+
+ channel->dma_irq_rx = ret;
+ yt_dbg(pdata,
+ "%s, for MSIx, channel %d dma_irq_rx=%u\n",
+ __func__, i, channel->dma_irq_rx);
+ }
+
+ if (i < FXGMAC_TX_1_RING)
+ channel->tx_ring = tx_ring++;
+
+ if (i < pdata->rx_ring_count)
+ channel->rx_ring = rx_ring++;
+ }
+
+ pdata->channel_head = channel_head;
+
+ if (netif_msg_drv(pdata))
+ yt_dbg(pdata, "%s ok\n", __func__);
+ return 0;
+
+err_irq:
+ kfree(rx_ring);
+
+err_rx_ring:
+ kfree(tx_ring);
+
+err_tx_ring:
+ kfree(channel_head);
+
+ yt_err(pdata, "%s err:%d\n", __func__, ret);
+ return ret;
+}
+
+void fxgmac_channels_rings_free(struct fxgmac_pdata *pdata)
+{
+ fxgmac_rings_free(pdata);
+ fxgmac_channels_free(pdata);
+}
+
+int fxgmac_channels_rings_alloc(struct fxgmac_pdata *pdata)
+{
+ int ret;
+
+ ret = fxgmac_channels_alloc(pdata);
+ if (ret < 0)
+ goto err_alloc;
+
+ ret = fxgmac_rings_alloc(pdata);
+ if (ret < 0)
+ goto err_alloc;
+
+ return 0;
+
+err_alloc:
+ fxgmac_channels_rings_free(pdata);
+ return ret;
+}
diff --git a/drivers/net/ethernet/motorcomm/yt6801/yt6801_net.c b/drivers/net/ethernet/motorcomm/yt6801/yt6801_net.c
index 0cb2808b7..39b91cc26 100644
--- a/drivers/net/ethernet/motorcomm/yt6801/yt6801_net.c
+++ b/drivers/net/ethernet/motorcomm/yt6801/yt6801_net.c
@@ -11,6 +11,26 @@
#include "yt6801_desc.h"
#include "yt6801_net.h"
+static int fxgmac_calc_rx_buf_size(struct fxgmac_pdata *pdata, unsigned int mtu)
+{
+ u32 rx_buf_size, max_mtu;
+
+ max_mtu = FXGMAC_JUMBO_PACKET_MTU - ETH_HLEN;
+ if (mtu > max_mtu) {
+ yt_err(pdata, "MTU exceeds maximum supported value\n");
+ return -EINVAL;
+ }
+
+ rx_buf_size = mtu + ETH_HLEN + ETH_FCS_LEN + VLAN_HLEN;
+ rx_buf_size =
+ clamp_val(rx_buf_size, FXGMAC_RX_MIN_BUF_SIZE, PAGE_SIZE * 4);
+
+ rx_buf_size = (rx_buf_size + FXGMAC_RX_BUF_ALIGN - 1) &
+ ~(FXGMAC_RX_BUF_ALIGN - 1);
+
+ return rx_buf_size;
+}
+
#define FXGMAC_NAPI_ENABLE 0x1
#define FXGMAC_NAPI_DISABLE 0x0
static void fxgmac_napi_disable(struct fxgmac_pdata *pdata)
@@ -197,6 +217,36 @@ void fxgmac_stop(struct fxgmac_pdata *pdata)
netdev_tx_reset_queue(txq);
}
+void fxgmac_restart(struct fxgmac_pdata *pdata)
+{
+ int ret;
+
+ /* If not running, "restart" will happen on open */
+ if (!netif_running(pdata->netdev) &&
+ pdata->dev_state != FXGMAC_DEV_START)
+ return;
+
+ mutex_lock(&pdata->mutex);
+ fxgmac_stop(pdata);
+ fxgmac_free_tx_data(pdata);
+ fxgmac_free_rx_data(pdata);
+ ret = fxgmac_start(pdata);
+ if (ret < 0)
+ yt_err(pdata, "%s err.\n", __func__);
+
+ mutex_unlock(&pdata->mutex);
+}
+
+static void fxgmac_restart_work(struct work_struct *work)
+{
+ struct fxgmac_pdata *pdata =
+ container_of(work, struct fxgmac_pdata, restart_work);
+
+ rtnl_lock();
+ fxgmac_restart(pdata);
+ rtnl_unlock();
+}
+
int fxgmac_net_powerdown(struct fxgmac_pdata *pdata, bool wake_en)
{
struct fxgmac_hw_ops *hw_ops = &pdata->hw_ops;
@@ -250,6 +300,46 @@ int fxgmac_net_powerdown(struct fxgmac_pdata *pdata, bool wake_en)
return 0;
}
+static int fxgmac_open(struct net_device *netdev)
+{
+ struct fxgmac_pdata *pdata = netdev_priv(netdev);
+ int ret;
+
+ mutex_lock(&pdata->mutex);
+ pdata->dev_state = FXGMAC_DEV_OPEN;
+
+ /* Calculate the Rx buffer size before allocating rings */
+ ret = fxgmac_calc_rx_buf_size(pdata, netdev->mtu);
+ if (ret < 0)
+ goto unlock;
+ pdata->rx_buf_size = ret;
+
+ /* Allocate the channels and rings */
+ ret = fxgmac_channels_rings_alloc(pdata);
+ if (ret < 0)
+ goto unlock;
+
+ INIT_WORK(&pdata->restart_work, fxgmac_restart_work);
+
+ ret = fxgmac_start(pdata);
+ if (ret < 0)
+ goto err_channels_and_rings;
+
+ mutex_unlock(&pdata->mutex);
+
+ if (netif_msg_drv(pdata))
+ yt_dbg(pdata, "%s ok\n", __func__);
+
+ return 0;
+
+err_channels_and_rings:
+ fxgmac_channels_rings_free(pdata);
+ yt_dbg(pdata, "%s, channel alloc err\n", __func__);
+unlock:
+ mutex_unlock(&pdata->mutex);
+ return ret;
+}
+
#ifdef CONFIG_PCI_MSI
static void fxgmac_init_interrupt_scheme(struct fxgmac_pdata *pdata)
{
@@ -442,3 +532,12 @@ int fxgmac_drv_probe(struct device *dev, struct fxgmac_resources *res)
free_netdev(netdev);
return ret;
}
+
+static const struct net_device_ops fxgmac_netdev_ops = {
+ .ndo_open = fxgmac_open,
+};
+
+const struct net_device_ops *fxgmac_get_netdev_ops(void)
+{
+ return &fxgmac_netdev_ops;
+}
--
2.34.1
Powered by blists - more mailing lists