nv-tegra-mirror/linux-nv-oot
1
0
Fork 0
mirror of git://nv-tegra.nvidia.com/linux-nv-oot.git synced 2025-12-23 17:55:05 +03:00
Code Issues Packages Projects Releases Wiki Activity

nvethernet: fix MC filter corruption issue

Browse Source 
Issue:
OSD Driver maintains a local MC address list to avoid the programming
old addresses again into HW when new MC address list passed from the
network stack.
Its observed MC address index for a specific address is different once a
MC address deleted from the list. Because of this new MC address which
is passed in new list is overwriting the old MC address.

Fix:
Program the entire list every time when network stack passes it instead
of maintaining local list at driver.

Bug 3609583

Change-Id: I97cb3ea4d35641ad156d16c54c89e8593ad64459
Signed-off-by: Bhadram Varka <vbhadram@nvidia.com>
Reviewed-on: https://git-master.nvidia.com/r/c/linux-nvidia/+/2702851
Reviewed-by: svc-mobile-coverity <svc-mobile-coverity@nvidia.com>
Reviewed-by: svc-mobile-cert <svc-mobile-cert@nvidia.com>
Reviewed-by: Narayan Reddy <narayanr@nvidia.com>
Reviewed-by: Sachin Nikam <snikam@nvidia.com>
Reviewed-by: svc_kernel_abi <svc_kernel_abi@nvidia.com>
GVS: Gerrit_Virtual_Submit
This commit is contained in:
Bhadram Varka
2022-04-28 14:42:05 +05:30
committed by Revanth Kumar Uppala
parent 9d509e2313
commit 8d2e3a6161
2 changed files with 56 additions and 186 deletions
Download Patch File Download Diff File Expand all files Collapse all files

228
drivers/net/ethernet/nvidia/nvethernet/ether_linux.c
View File

@@ -101,141 +101,6 @@ update_skb:
}
}
/** @brief invalidate local l2 address list
*
* Algorithm: Invalidate all nodes in local address link list
*
* @param[in] pdata: OSD private data.
*
*/
static inline void ether_invalidate_mac_addr_list(struct ether_priv_data *pdata)
{
struct ether_mac_addr_list *pnode;
struct list_head *head_node, *temp_head_node;
if (list_empty(&pdata->mac_addr_list_head)) {
return;
}
list_for_each_safe(head_node, temp_head_node,
&pdata->mac_addr_list_head) {
pnode = list_entry(head_node,
struct ether_mac_addr_list,
list_head);
if (pnode->index != ETHER_MAC_ADDRESS_INDEX &&
pnode->index != ETHER_BC_ADDRESS_INDEX) {
pnode->is_valid_addr = OSI_DISABLE;
}
}
}
/**
* @brief find index and add l2 address
*
* Algorithm: Find index for already added address or empty index to add filter
* address. If new L2 address, add it in local link list and call OSI
* API.
* @param[in] pdata: OSD private data.
* @param[in] filter: OSI filter structure pointer
*/
static int ether_update_mac_addr(struct ether_priv_data *pdata,
struct osi_ioctl *ioctl_data)
{
struct ether_mac_addr_list *pnode = NULL;
struct list_head *head_node, *temp_head_node;
struct osi_core_priv_data *osi_core = pdata->osi_core;
int ret = 0;
list_for_each_safe(head_node, temp_head_node,
&pdata->mac_addr_list_head) {
pnode = list_entry(head_node,
struct ether_mac_addr_list,
list_head);
if (memcmp(ioctl_data->l2_filter.mac_address, pnode->addr,
OSI_ETH_ALEN) == 0) {
pnode->is_valid_addr = OSI_ENABLE;
return ret;
}
}
pnode = kmalloc(sizeof(*pnode), GFP_KERNEL);
if (!pnode) {
dev_err(pdata->dev, "kmalloc failed %s()\n", __func__);
return -1;
}
memcpy(pnode->addr, ioctl_data->l2_filter.mac_address, OSI_ETH_ALEN);
pnode->is_valid_addr = OSI_ENABLE;
pnode->dma_chan = ioctl_data->l2_filter.dma_chan;
pnode->index = ioctl_data->l2_filter.index;
ret = osi_handle_ioctl(osi_core, ioctl_data);
if (ret < 0) {
kfree(pnode);
return ret;
}
list_add(&pnode->list_head,
&pdata->mac_addr_list_head);
return ret;
}
/**
* @brief remove invalid l2 address
*
* Algorithm: call OSI call for addresses which are not valid with new list
* from network stack.
*
* @param[in] pdata: OSD private data.
* @param[in] ioctl_data: OSI IOCTL data structure.
*
*/
static inline int ether_remove_invalid_mac_addr(struct ether_priv_data *pdata,
struct osi_ioctl *ioctl_data)
{
struct ether_mac_addr_list *pnode;
struct list_head *head_node, *temp_head_node;
struct osi_core_priv_data *osi_core = pdata->osi_core;
int ret = 0;
if (list_empty(&pdata->mac_addr_list_head)) {
return ret;
}
list_for_each_safe(head_node, temp_head_node,
&pdata->mac_addr_list_head) {
pnode = list_entry(head_node,
struct ether_mac_addr_list,
list_head);
if (pnode->is_valid_addr == OSI_DISABLE) {
memset(&ioctl_data->l2_filter, 0x0,
sizeof(struct osi_filter));
ioctl_data->l2_filter.dma_routing = OSI_ENABLE;
ioctl_data->l2_filter.addr_mask = OSI_AMASK_DISABLE;
ioctl_data->l2_filter.src_dest = OSI_DA_MATCH;
ioctl_data->l2_filter.oper_mode = OSI_OPER_ADDR_DEL;
memcpy(ioctl_data->l2_filter.mac_address, pnode->addr,
OSI_ETH_ALEN);
ioctl_data->l2_filter.dma_chan = pnode->dma_chan;
ioctl_data->l2_filter.index = pnode->index;
ret = osi_handle_ioctl(osi_core, ioctl_data);
if (ret < 0) {
dev_err(pdata->dev,
"%s failed to remove address\n",
__func__);
return ret;
}
list_del(head_node);
kfree(pnode);
}
}
return ret;
}
#ifdef HSI_SUPPORT
static inline u64 rdtsc(void)
{
@@ -2900,9 +2765,6 @@ static inline void ether_reset_stats(struct ether_priv_data *pdata)
static inline void ether_delete_l2_filter(struct ether_priv_data *pdata)
{
struct osi_core_priv_data *osi_core = pdata->osi_core;
struct osi_dma_priv_data *osi_dma = pdata->osi_dma;
struct ether_mac_addr_list *pnode;
struct list_head *head_node, *temp_head_node;
struct osi_ioctl ioctl_data = {};
int ret, i;
@@ -2921,38 +2783,30 @@ static inline void ether_delete_l2_filter(struct ether_priv_data *pdata)
}
/* Loop to delete l2 address list filters */
for (i = ETHER_MAC_ADDRESS_INDEX + 1; i <= pdata->last_filter_index;
for (i = ETHER_MAC_ADDRESS_INDEX + 1; i < pdata->last_filter_index;
i++) {
/* Reset the filter structure to avoid any old value */
memset(&ioctl_data.l2_filter, 0x0, sizeof(struct osi_filter));
ioctl_data.l2_filter.oper_mode = OSI_OPER_ADDR_DEL;
ioctl_data.l2_filter.index = i;
ioctl_data.l2_filter.dma_routing = OSI_ENABLE;
if (osi_dma->num_dma_chans > 1) {
ioctl_data.l2_filter.dma_chan = osi_dma->dma_chans[1];
} else {
ioctl_data.l2_filter.dma_chan = osi_dma->dma_chans[0];
}
memcpy(ioctl_data.l2_filter.mac_address,
pdata->mac_addr[i].addr, ETH_ALEN);
ioctl_data.l2_filter.dma_chan = pdata->mac_addr[i].dma_chan;
ioctl_data.l2_filter.addr_mask = OSI_AMASK_DISABLE;
ioctl_data.l2_filter.src_dest = OSI_DA_MATCH;
ioctl_data.cmd = OSI_CMD_L2_FILTER;
ret = osi_handle_ioctl(osi_core, &ioctl_data);
if (ret < 0) {
dev_err(pdata->dev,
"failed to delete L2 filter index = %d\n", i);
mutex_unlock(&pdata->rx_mode_lock);
return;
}
}
if (!list_empty(&pdata->mac_addr_list_head)) {
list_for_each_safe(head_node, temp_head_node,
&pdata->mac_addr_list_head) {
pnode = list_entry(head_node,
struct ether_mac_addr_list,
list_head);
list_del(head_node);
kfree(pnode);
}
}
pdata->last_filter_index = 0;
}
/**
@@ -3555,13 +3409,14 @@ static int ether_start_xmit(struct sk_buff *skb, struct net_device *ndev)
* @retval "negative value" on failure.
*/
static int ether_prepare_mc_list(struct net_device *dev,
struct osi_ioctl *ioctl_data)
struct osi_ioctl *ioctl_data,
unsigned int *mac_addr_idx)
{
struct ether_priv_data *pdata = netdev_priv(dev);
struct osi_core_priv_data *osi_core = pdata->osi_core;
struct osi_dma_priv_data *osi_dma = pdata->osi_dma;
struct netdev_hw_addr *ha;
unsigned int i = ETHER_MAC_ADDRESS_INDEX + 1;
unsigned int i = *mac_addr_idx;
int ret = -1;
if (ioctl_data == NULL) {
@@ -3619,21 +3474,23 @@ static int ether_prepare_mc_list(struct net_device *dev,
ioctl_data->l2_filter.addr_mask = OSI_AMASK_DISABLE;
ioctl_data->l2_filter.src_dest = OSI_DA_MATCH;
ioctl_data->cmd = OSI_CMD_L2_FILTER;
ret = ether_update_mac_addr(pdata, ioctl_data);
ret = osi_handle_ioctl(pdata->osi_core, ioctl_data);
if (ret < 0) {
dev_err(pdata->dev, "issue in creating mc list\n");
pdata->last_filter_index = i - 1;
*mac_addr_idx = i;
return ret;
}
memcpy(pdata->mac_addr[i].addr, ha->addr, ETH_ALEN);
pdata->mac_addr[i].dma_chan = ioctl_data->l2_filter.dma_chan;
if (i == EQOS_MAX_MAC_ADDRESS_FILTER - 1) {
dev_err(pdata->dev, "Configured max number of supported MAC, ignoring it\n");
break;
}
i++;
}
/* preserve last filter index to pass on to UC */
pdata->last_filter_index = i - 1;
*mac_addr_idx = i;
}
return ret;
@@ -3654,13 +3511,14 @@ static int ether_prepare_mc_list(struct net_device *dev,
* @retval "negative value" on failure.
*/
static int ether_prepare_uc_list(struct net_device *dev,
struct osi_ioctl *ioctl_data)
struct osi_ioctl *ioctl_data,
unsigned int *mac_addr_idx)
{
struct ether_priv_data *pdata = netdev_priv(dev);
struct osi_core_priv_data *osi_core = pdata->osi_core;
struct osi_dma_priv_data *osi_dma = pdata->osi_dma;
/* last valid MC/MAC DA + 1 should be start of UC addresses */
unsigned int i = pdata->last_filter_index + 1;
unsigned int i = *mac_addr_idx;
struct netdev_hw_addr *ha;
int ret = -1;
@@ -3717,21 +3575,24 @@ static int ether_prepare_uc_list(struct net_device *dev,
ioctl_data->l2_filter.src_dest = OSI_DA_MATCH;
ioctl_data->cmd = OSI_CMD_L2_FILTER;
ret = ether_update_mac_addr(pdata, ioctl_data);
ret = osi_handle_ioctl(pdata->osi_core, ioctl_data);
if (ret < 0) {
dev_err(pdata->dev,
"issue in creating uc list\n");
pdata->last_filter_index = i - 1;
*mac_addr_idx = i;
return ret;
}
memcpy(pdata->mac_addr[i].addr, ha->addr, ETH_ALEN);
pdata->mac_addr[i].dma_chan = ioctl_data->l2_filter.dma_chan;
if (i == EQOS_MAX_MAC_ADDRESS_FILTER - 1) {
dev_err(pdata->dev, "Already MAX MAC added\n");
break;
}
i++;
}
pdata->last_filter_index = i - 1;
*mac_addr_idx = i;
}
return ret;
@@ -3752,6 +3613,7 @@ static inline void set_rx_mode_work_func(struct work_struct *work)
/* store last call last_uc_filter_index in temporary variable */
struct osi_ioctl ioctl_data = {};
struct net_device *dev = pdata->ndev;
unsigned int mac_addr_idx = ETHER_MAC_ADDRESS_INDEX + 1U, i;
int ret = -1;
mutex_lock(&pdata->rx_mode_lock);
@@ -3790,9 +3652,7 @@ static inline void set_rx_mode_work_func(struct work_struct *work)
mutex_unlock(&pdata->rx_mode_lock);
return;
} else if (!netdev_mc_empty(dev)) {
/*MC list will be always there, invalidate list only once*/
ether_invalidate_mac_addr_list(pdata);
if (ether_prepare_mc_list(dev, &ioctl_data) != 0) {
if (ether_prepare_mc_list(dev, &ioctl_data, &mac_addr_idx) != 0) {
dev_err(pdata->dev, "Setting MC address failed\n");
}
} else {
@@ -3801,19 +3661,36 @@ static inline void set_rx_mode_work_func(struct work_struct *work)
}
if (!netdev_uc_empty(dev)) {
if (ether_prepare_uc_list(dev, &ioctl_data) != 0) {
if (ether_prepare_uc_list(dev, &ioctl_data, &mac_addr_idx) != 0) {
dev_err(pdata->dev, "Setting UC address failed\n");
}
}
ret = ether_remove_invalid_mac_addr(pdata, &ioctl_data);
if (ret < 0) {
dev_err(pdata->dev,
"Invalidating expired L2 filter failed\n");
mutex_unlock(&pdata->rx_mode_lock);
return;
if (pdata->last_filter_index > mac_addr_idx) {
for (i = mac_addr_idx; i < pdata->last_filter_index; i++) {
/* Reset the filter structure to avoid any old value */
memset(&ioctl_data.l2_filter, 0x0, sizeof(struct osi_filter));
ioctl_data.l2_filter.oper_mode = OSI_OPER_ADDR_DEL;
ioctl_data.l2_filter.index = i;
ioctl_data.l2_filter.dma_routing = OSI_ENABLE;
memcpy(ioctl_data.l2_filter.mac_address,
pdata->mac_addr[i].addr, ETH_ALEN);
ioctl_data.l2_filter.dma_chan = pdata->mac_addr[i].dma_chan;
ioctl_data.l2_filter.addr_mask = OSI_AMASK_DISABLE;
ioctl_data.l2_filter.src_dest = OSI_DA_MATCH;
ioctl_data.cmd = OSI_CMD_L2_FILTER;
ret = osi_handle_ioctl(osi_core, &ioctl_data);
if (ret < 0) {
dev_err(pdata->dev,
"failed to delete L2 filter index = %d\n", i);
mutex_unlock(&pdata->rx_mode_lock);
return;
}
}
}
pdata->last_filter_index = mac_addr_idx;
/* Set default MAC configuration because if this path is called
* only when flag for promiscuous or all_multi is not set.
*/
@@ -6659,7 +6536,6 @@ static int ether_probe(struct platform_device *pdev)
/* Initialization of set speed workqueue */
INIT_DELAYED_WORK(&pdata->set_speed_work, set_speed_work_func);
osi_core->hw_feature = &pdata->hw_feat;
INIT_LIST_HEAD(&pdata->mac_addr_list_head);
INIT_LIST_HEAD(&pdata->tx_ts_skb_head);
INIT_DELAYED_WORK(&pdata->tx_ts_work, ether_get_tx_ts);
pdata->rx_m_enabled = false;

14
drivers/net/ethernet/nvidia/nvethernet/ether_linux.h
View File

@@ -225,7 +225,7 @@
/**
* @brief Broadcast and MAC address macros
*/
#define ETHER_MAC_ADDRESS_INDEX 1
#define ETHER_MAC_ADDRESS_INDEX 1U
#define ETHER_BC_ADDRESS_INDEX 0
#define ETHER_ADDRESS_MAC 1
#define ETHER_ADDRESS_BC 0
@@ -368,17 +368,11 @@ struct ether_ivc_ctxt {
/**
* @brief local L2 filter table structure
*/
struct ether_mac_addr_list {
/** Link list node head */
struct list_head list_head;
struct ether_mac_addr {
/** L2 address */
unsigned char addr[ETH_ALEN];
/** Flag represent is address valid(1) or not (0) */
char is_valid_addr;
/** DMA channel to route packets */
unsigned int dma_chan;
/** index number at the time of add */
unsigned int index;
};
/**
@@ -503,7 +497,7 @@ struct ether_priv_data {
/** max address register count, 2*mac_addr64_sel */
int num_mac_addr_regs;
/** Last address reg filter index added in last call*/
int last_filter_index;
unsigned int last_filter_index;
/** vlan hash filter 1: hash, 0: perfect */
unsigned int vlan_hash_filtering;
/** L2 filter mode */
@@ -565,7 +559,7 @@ struct ether_priv_data {
struct macsec_priv_data *macsec_pdata;
#endif /* MACSEC_SUPPORT */
/** local L2 filter address list head pointer */
struct list_head mac_addr_list_head;
struct ether_mac_addr mac_addr[ETHER_ADDR_REG_CNT_128];
/** skb tx timestamp update work queue */
struct delayed_work tx_ts_work;
/** local skb list head */