* RE: [PATCH net-next 5/5] r8152: change rx_frag_head_sz and rx_max_agg_num dynamically
From: Hayes Wang @ 2019-08-09 3:38 UTC (permalink / raw)
To: Jakub Kicinski
Cc: Maciej Fijalkowski, netdev@vger.kernel.org, nic_swsd,
linux-kernel@vger.kernel.org, linux-usb@vger.kernel.org
In-Reply-To: <20190808114325.5c346d3a@cakuba.netronome.com>
Jakub Kicinski [jakub.kicinski@netronome.com]
[..]> The kernel could support it. And I has finished it.
> > However, when I want to test it by ethtool, I couldn't find suitable command.
> > I couldn't find relative feature in the source code of ethtool, either.
> It's possible it's not implemented in the user space tool 🤔
>
> Looks like it got posted here:
>
> https://www.spinics.net/lists/netdev/msg299877.html
>
> But perhaps never finished?
May I implement both sysfs and set_tunalbe for copybreak first
before the user space tool is ready? Otherwise, the user couldn't
change the copybreak now.
Best Regards,
Hayes
^ permalink raw reply
* [PATCH net-next 02/12] net: hns3: fix interrupt clearing error for VF
From: Huazhong Tan @ 2019-08-09 2:31 UTC (permalink / raw)
To: davem
Cc: netdev, linux-kernel, salil.mehta, yisen.zhuang, linuxarm,
Huazhong Tan
In-Reply-To: <1565317878-31806-1-git-send-email-tanhuazhong@huawei.com>
Currently, VF driver has two kinds of interrupts, reset & CMDQ RX.
For revision 0x21, according to the UM, each interrupt should be
cleared by write 0 to the corresponding bit, but the implementation
writes 0 to the whole register in fact, it will clear other
interrupt at the same time, then the VF will loss the interrupt.
But for revision 0x20, this interrupt clear register is a read &
write register, for compatible, we just keep the old implementation
for 0x20.
This patch fixes it, also, adds a new register for reading the interrupt
status according to hardware user manual.
Fixes: e2cb1dec9779 ("net: hns3: Add HNS3 VF HCL(Hardware Compatibility Layer) Support")
Fixes: b90fcc5bd904 ("net: hns3: add reset handling for VF when doing Core/Global/IMP reset")
Signed-off-by: Huazhong Tan <tanhuazhong@huawei.com>
Reviewed-by: Yunsheng Lin <linyunsheng@huawei.com>
---
.../ethernet/hisilicon/hns3/hns3vf/hclgevf_main.c | 28 +++++++++++++++-------
.../ethernet/hisilicon/hns3/hns3vf/hclgevf_main.h | 2 ++
2 files changed, 21 insertions(+), 9 deletions(-)
diff --git a/drivers/net/ethernet/hisilicon/hns3/hns3vf/hclgevf_main.c b/drivers/net/ethernet/hisilicon/hns3/hns3vf/hclgevf_main.c
index ce82b2b..d8b8281 100644
--- a/drivers/net/ethernet/hisilicon/hns3/hns3vf/hclgevf_main.c
+++ b/drivers/net/ethernet/hisilicon/hns3/hns3vf/hclgevf_main.c
@@ -1889,21 +1889,20 @@ static void hclgevf_clear_event_cause(struct hclgevf_dev *hdev, u32 regclr)
static enum hclgevf_evt_cause hclgevf_check_evt_cause(struct hclgevf_dev *hdev,
u32 *clearval)
{
- u32 val, cmdq_src_reg, rst_ing_reg;
+ u32 val, cmdq_stat_reg, rst_ing_reg;
/* fetch the events from their corresponding regs */
- cmdq_src_reg = hclgevf_read_dev(&hdev->hw,
- HCLGEVF_VECTOR0_CMDQ_SRC_REG);
+ cmdq_stat_reg = hclgevf_read_dev(&hdev->hw,
+ HCLGEVF_VECTOR0_CMDQ_STAT_REG);
- if (BIT(HCLGEVF_VECTOR0_RST_INT_B) & cmdq_src_reg) {
+ if (BIT(HCLGEVF_VECTOR0_RST_INT_B) & cmdq_stat_reg) {
rst_ing_reg = hclgevf_read_dev(&hdev->hw, HCLGEVF_RST_ING);
dev_info(&hdev->pdev->dev,
"receive reset interrupt 0x%x!\n", rst_ing_reg);
set_bit(HNAE3_VF_RESET, &hdev->reset_pending);
set_bit(HCLGEVF_RESET_PENDING, &hdev->reset_state);
set_bit(HCLGEVF_STATE_CMD_DISABLE, &hdev->state);
- cmdq_src_reg &= ~BIT(HCLGEVF_VECTOR0_RST_INT_B);
- *clearval = cmdq_src_reg;
+ *clearval = ~(1U << HCLGEVF_VECTOR0_RST_INT_B);
hdev->rst_stats.vf_rst_cnt++;
/* set up VF hardware reset status, its PF will clear
* this status when PF has initialized done.
@@ -1915,9 +1914,20 @@ static enum hclgevf_evt_cause hclgevf_check_evt_cause(struct hclgevf_dev *hdev,
}
/* check for vector0 mailbox(=CMDQ RX) event source */
- if (BIT(HCLGEVF_VECTOR0_RX_CMDQ_INT_B) & cmdq_src_reg) {
- cmdq_src_reg &= ~BIT(HCLGEVF_VECTOR0_RX_CMDQ_INT_B);
- *clearval = cmdq_src_reg;
+ if (BIT(HCLGEVF_VECTOR0_RX_CMDQ_INT_B) & cmdq_stat_reg) {
+ /* for revision 0x21, clearing interrupt is writing bit 0
+ * to the clear register, writing bit 1 means to keep the
+ * old value.
+ * for revision 0x20, the clear register is a read & write
+ * register, so we should just write 0 to the bit we are
+ * handling, and keep other bits as cmdq_stat_reg.
+ */
+ if (hdev->pdev->revision >= 0x21)
+ *clearval = ~(1U << HCLGEVF_VECTOR0_RX_CMDQ_INT_B);
+ else
+ *clearval = cmdq_stat_reg &
+ ~BIT(HCLGEVF_VECTOR0_RX_CMDQ_INT_B);
+
return HCLGEVF_VECTOR0_EVENT_MBX;
}
diff --git a/drivers/net/ethernet/hisilicon/hns3/hns3vf/hclgevf_main.h b/drivers/net/ethernet/hisilicon/hns3/hns3vf/hclgevf_main.h
index f0736b0..4ccf107 100644
--- a/drivers/net/ethernet/hisilicon/hns3/hns3vf/hclgevf_main.h
+++ b/drivers/net/ethernet/hisilicon/hns3/hns3vf/hclgevf_main.h
@@ -87,6 +87,8 @@
/* Vector0 interrupt CMDQ event source register(RW) */
#define HCLGEVF_VECTOR0_CMDQ_SRC_REG 0x27100
+/* Vector0 interrupt CMDQ event status register(RO) */
+#define HCLGEVF_VECTOR0_CMDQ_STAT_REG 0x27104
/* CMDQ register bits for RX event(=MBX event) */
#define HCLGEVF_VECTOR0_RX_CMDQ_INT_B 1
/* RST register bits for RESET event */
--
2.7.4
^ permalink raw reply related
* [PATCH net-next 09/12] net: hns3: add function display NCL_CONFIG info
From: Huazhong Tan @ 2019-08-09 2:31 UTC (permalink / raw)
To: davem
Cc: netdev, linux-kernel, salil.mehta, yisen.zhuang, linuxarm,
Yufeng Mo, Huazhong Tan
In-Reply-To: <1565317878-31806-1-git-send-email-tanhuazhong@huawei.com>
From: Yufeng Mo <moyufeng@huawei.com>
This adds a new function hclge_ncl_config_data_print()
to print the data of NCL_CONFIG, to make the code more
readable. Also, using macro replaces some magic number.
Signed-off-by: Yufeng Mo <moyufeng@huawei.com>
Reviewed-by: Peng Li <lipeng321@huawei.com>
Signed-off-by: Huazhong Tan <tanhuazhong@huawei.com>
---
.../ethernet/hisilicon/hns3/hns3pf/hclge_debugfs.c | 52 +++++++++++++---------
1 file changed, 30 insertions(+), 22 deletions(-)
diff --git a/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_debugfs.c b/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_debugfs.c
index 933dec5..f0295d1 100644
--- a/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_debugfs.c
+++ b/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_debugfs.c
@@ -995,6 +995,33 @@ void hclge_dbg_get_m7_stats_info(struct hclge_dev *hdev)
kfree(desc_src);
}
+#define HCLGE_CMD_NCL_CONFIG_BD_NUM 5
+
+static void hclge_ncl_config_data_print(struct hclge_dev *hdev,
+ struct hclge_desc *desc, int *offset,
+ int *length)
+{
+#define HCLGE_CMD_DATA_NUM 6
+
+ int i;
+ int j;
+
+ for (i = 0; i < HCLGE_CMD_NCL_CONFIG_BD_NUM; i++) {
+ for (j = 0; j < HCLGE_CMD_DATA_NUM; j++) {
+ if (i == 0 && j == 0)
+ continue;
+
+ dev_info(&hdev->pdev->dev, "0x%04x | 0x%08x\n",
+ *offset,
+ le32_to_cpu(desc[i].data[j]));
+ *offset += sizeof(u32);
+ *length -= sizeof(u32);
+ if (*length <= 0)
+ return;
+ }
+ }
+}
+
/* hclge_dbg_dump_ncl_config: print specified range of NCL_CONFIG file
* @hdev: pointer to struct hclge_dev
* @cmd_buf: string that contains offset and length
@@ -1004,17 +1031,13 @@ static void hclge_dbg_dump_ncl_config(struct hclge_dev *hdev,
{
#define HCLGE_MAX_NCL_CONFIG_OFFSET 4096
#define HCLGE_MAX_NCL_CONFIG_LENGTH (20 + 24 * 4)
-#define HCLGE_CMD_DATA_NUM 6
- struct hclge_desc desc[5];
- u32 byte_offset;
- int bd_num = 5;
+ struct hclge_desc desc[HCLGE_CMD_NCL_CONFIG_BD_NUM];
+ int bd_num = HCLGE_CMD_NCL_CONFIG_BD_NUM;
int offset;
int length;
int data0;
int ret;
- int i;
- int j;
ret = sscanf(cmd_buf, "%x %x", &offset, &length);
if (ret != 2 || offset >= HCLGE_MAX_NCL_CONFIG_OFFSET ||
@@ -1040,22 +1063,7 @@ static void hclge_dbg_dump_ncl_config(struct hclge_dev *hdev,
if (ret)
return;
- byte_offset = offset;
- for (i = 0; i < bd_num; i++) {
- for (j = 0; j < HCLGE_CMD_DATA_NUM; j++) {
- if (i == 0 && j == 0)
- continue;
-
- dev_info(&hdev->pdev->dev, "0x%04x | 0x%08x\n",
- byte_offset,
- le32_to_cpu(desc[i].data[j]));
- byte_offset += sizeof(u32);
- length -= sizeof(u32);
- if (length <= 0)
- return;
- }
- }
- offset += HCLGE_MAX_NCL_CONFIG_LENGTH;
+ hclge_ncl_config_data_print(hdev, desc, &offset, &length);
}
}
--
2.7.4
^ permalink raw reply related
* [PATCH net-next 11/12] net: hns3: add handshake with VF for PF reset
From: Huazhong Tan @ 2019-08-09 2:31 UTC (permalink / raw)
To: davem
Cc: netdev, linux-kernel, salil.mehta, yisen.zhuang, linuxarm,
Huazhong Tan
In-Reply-To: <1565317878-31806-1-git-send-email-tanhuazhong@huawei.com>
Before PF asserting function reset, it should make sure
that all its VFs have been ready, otherwise, it will cause
some hardware errors.
So this patch adds function hclge_func_reset_sync_vf() to
synchronize VF before asserting PF function reset. For new
firmware which supports command HCLGE_OPC_QUERY_VF_RST_RDY,
we will try to query VFs' ready status within 30 seconds.
And keep the old implementation for compatible with firmware
which does not support this command.
Signed-off-by: Huazhong Tan <tanhuazhong@huawei.com>
Reviewed-by: Yunsheng Lin <linyunsheng@huawei.com>
---
.../net/ethernet/hisilicon/hns3/hns3pf/hclge_cmd.h | 7 +++
.../ethernet/hisilicon/hns3/hns3pf/hclge_main.c | 56 ++++++++++++++++++----
2 files changed, 55 insertions(+), 8 deletions(-)
diff --git a/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_cmd.h b/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_cmd.h
index dade20a..29979be 100644
--- a/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_cmd.h
+++ b/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_cmd.h
@@ -87,6 +87,7 @@ enum hclge_opcode_type {
HCLGE_OPC_QUERY_VF_RSRC = 0x0024,
HCLGE_OPC_GET_CFG_PARAM = 0x0025,
HCLGE_OPC_PF_RST_DONE = 0x0026,
+ HCLGE_OPC_QUERY_VF_RST_RDY = 0x0027,
HCLGE_OPC_STATS_64_BIT = 0x0030,
HCLGE_OPC_STATS_32_BIT = 0x0031,
@@ -588,6 +589,12 @@ struct hclge_config_mac_mode_cmd {
u8 rsv[20];
};
+struct hclge_pf_rst_sync_cmd {
+#define HCLGE_PF_RST_ALL_VF_RDY_B 0
+ u8 all_vf_ready;
+ u8 rsv[23];
+};
+
#define HCLGE_CFG_SPEED_S 0
#define HCLGE_CFG_SPEED_M GENMASK(5, 0)
diff --git a/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c b/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c
index 1315275..d207dac 100644
--- a/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c
+++ b/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c
@@ -35,6 +35,9 @@
#define BUF_RESERVE_PERCENT 90
#define HCLGE_RESET_MAX_FAIL_CNT 5
+#define HCLGE_RESET_SYNC_TIME 100
+#define HCLGE_PF_RESET_SYNC_TIME 20
+#define HCLGE_PF_RESET_SYNC_CNT 1500
/* Get DFX BD number offset */
#define HCLGE_DFX_BIOS_BD_OFFSET 1
@@ -3184,6 +3187,39 @@ static int hclge_set_all_vf_rst(struct hclge_dev *hdev, bool reset)
return 0;
}
+int hclge_func_reset_sync_vf(struct hclge_dev *hdev)
+{
+ struct hclge_pf_rst_sync_cmd *req;
+ struct hclge_desc desc;
+ int cnt = 0;
+ int ret;
+
+ req = (struct hclge_pf_rst_sync_cmd *)desc.data;
+ hclge_cmd_setup_basic_desc(&desc, HCLGE_OPC_QUERY_VF_RST_RDY, true);
+
+ do {
+ ret = hclge_cmd_send(&hdev->hw, &desc, 1);
+ /* for compatible with old firmware, wait
+ * 100 ms for VF to stop IO
+ */
+ if (ret == -EOPNOTSUPP) {
+ msleep(HCLGE_RESET_SYNC_TIME);
+ return 0;
+ } else if (ret) {
+ dev_err(&hdev->pdev->dev, "sync with VF fail %d!\n",
+ ret);
+ return ret;
+ } else if (req->all_vf_ready) {
+ return 0;
+ }
+ msleep(HCLGE_PF_RESET_SYNC_TIME);
+ hclge_cmd_reuse_desc(&desc, true);
+ } while (cnt++ < HCLGE_PF_RESET_SYNC_CNT);
+
+ dev_err(&hdev->pdev->dev, "sync with VF timeout!\n");
+ return -ETIME;
+}
+
int hclge_func_reset_cmd(struct hclge_dev *hdev, int func_id)
{
struct hclge_desc desc;
@@ -3350,17 +3386,18 @@ static void hclge_reset_handshake(struct hclge_dev *hdev, bool enable)
static int hclge_reset_prepare_wait(struct hclge_dev *hdev)
{
-#define HCLGE_RESET_SYNC_TIME 100
-
u32 reg_val;
int ret = 0;
switch (hdev->reset_type) {
case HNAE3_FUNC_RESET:
- /* There is no mechanism for PF to know if VF has stopped IO
- * for now, just wait 100 ms for VF to stop IO
+ /* to confirm whether all running VF is ready
+ * before request PF reset
*/
- msleep(HCLGE_RESET_SYNC_TIME);
+ ret = hclge_func_reset_sync_vf(hdev);
+ if (ret)
+ return ret;
+
ret = hclge_func_reset_cmd(hdev, 0);
if (ret) {
dev_err(&hdev->pdev->dev,
@@ -3377,10 +3414,13 @@ static int hclge_reset_prepare_wait(struct hclge_dev *hdev)
hdev->rst_stats.pf_rst_cnt++;
break;
case HNAE3_FLR_RESET:
- /* There is no mechanism for PF to know if VF has stopped IO
- * for now, just wait 100 ms for VF to stop IO
+ /* to confirm whether all running VF is ready
+ * before request PF reset
*/
- msleep(HCLGE_RESET_SYNC_TIME);
+ ret = hclge_func_reset_sync_vf(hdev);
+ if (ret)
+ return ret;
+
set_bit(HCLGE_STATE_CMD_DISABLE, &hdev->state);
set_bit(HNAE3_FLR_DOWN, &hdev->flr_state);
hdev->rst_stats.flr_rst_cnt++;
--
2.7.4
^ permalink raw reply related
* [PATCH net-next 12/12] net: hns3: refine some macro definitions
From: Huazhong Tan @ 2019-08-09 2:31 UTC (permalink / raw)
To: davem
Cc: netdev, linux-kernel, salil.mehta, yisen.zhuang, linuxarm,
Guojia Liao, Guangbin Huang, Huazhong Tan
In-Reply-To: <1565317878-31806-1-git-send-email-tanhuazhong@huawei.com>
From: Guojia Liao <liaoguojia@huawei.com>
Macro arguments should be enclosed in parentheses, in case of
expression argument, but parentheses of pure number in macro
definition should be removed for simplicity.
Signed-off-by: Guojia Liao <liaoguojia@huawei.com>
Signed-off-by: Guangbin Huang <huangguangbin2@huawei.com>
Signed-off-by: Huazhong Tan <tanhuazhong@huawei.com>
---
drivers/net/ethernet/hisilicon/hns3/hnae3.h | 4 ++--
drivers/net/ethernet/hisilicon/hns3/hns3vf/hclgevf_main.h | 2 +-
2 files changed, 3 insertions(+), 3 deletions(-)
diff --git a/drivers/net/ethernet/hisilicon/hns3/hnae3.h b/drivers/net/ethernet/hisilicon/hns3/hnae3.h
index 43740ee..6c9fd58 100644
--- a/drivers/net/ethernet/hisilicon/hns3/hnae3.h
+++ b/drivers/net/ethernet/hisilicon/hns3/hnae3.h
@@ -58,10 +58,10 @@
BIT(HNAE3_DEV_SUPPORT_ROCE_B))
#define hnae3_dev_roce_supported(hdev) \
- hnae3_get_bit(hdev->ae_dev->flag, HNAE3_DEV_SUPPORT_ROCE_B)
+ hnae3_get_bit((hdev)->ae_dev->flag, HNAE3_DEV_SUPPORT_ROCE_B)
#define hnae3_dev_dcb_supported(hdev) \
- hnae3_get_bit(hdev->ae_dev->flag, HNAE3_DEV_SUPPORT_DCB_B)
+ hnae3_get_bit((hdev)->ae_dev->flag, HNAE3_DEV_SUPPORT_DCB_B)
#define hnae3_dev_fd_supported(hdev) \
hnae3_get_bit((hdev)->ae_dev->flag, HNAE3_DEV_SUPPORT_FD_B)
diff --git a/drivers/net/ethernet/hisilicon/hns3/hns3vf/hclgevf_main.h b/drivers/net/ethernet/hisilicon/hns3/hns3vf/hclgevf_main.h
index 4ccf107..bdde3af 100644
--- a/drivers/net/ethernet/hisilicon/hns3/hns3vf/hclgevf_main.h
+++ b/drivers/net/ethernet/hisilicon/hns3/hns3vf/hclgevf_main.h
@@ -125,7 +125,7 @@
#define HCLGEVF_S_IP_BIT BIT(3)
#define HCLGEVF_V_TAG_BIT BIT(4)
-#define HCLGEVF_STATS_TIMER_INTERVAL (36)
+#define HCLGEVF_STATS_TIMER_INTERVAL 36U
enum hclgevf_evt_cause {
HCLGEVF_VECTOR0_EVENT_RST,
--
2.7.4
^ permalink raw reply related
* [PATCH net-next 04/12] net: hns3: add input length check for debugfs write function
From: Huazhong Tan @ 2019-08-09 2:31 UTC (permalink / raw)
To: davem
Cc: netdev, linux-kernel, salil.mehta, yisen.zhuang, linuxarm,
Yufeng Mo, Huazhong Tan
In-Reply-To: <1565317878-31806-1-git-send-email-tanhuazhong@huawei.com>
From: Yufeng Mo <moyufeng@huawei.com>
If the input length reaches the maximum value of size_t, the reverse is
triggered when 1 is added. In addition, there is no need to have such a
large length. Therefore, the input length should be checked and the value
should be less than or equal to 1024.
Signed-off-by: Yufeng Mo <moyufeng@huawei.com>
Reviewed-by: Peng Li <lipeng321@huawei.com>
Signed-off-by: Huazhong Tan <tanhuazhong@huawei.com>
---
drivers/net/ethernet/hisilicon/hns3/hns3_debugfs.c | 4 ++++
1 file changed, 4 insertions(+)
diff --git a/drivers/net/ethernet/hisilicon/hns3/hns3_debugfs.c b/drivers/net/ethernet/hisilicon/hns3/hns3_debugfs.c
index a4b9372..7996dcc 100644
--- a/drivers/net/ethernet/hisilicon/hns3/hns3_debugfs.c
+++ b/drivers/net/ethernet/hisilicon/hns3/hns3_debugfs.c
@@ -8,6 +8,7 @@
#include "hns3_enet.h"
#define HNS3_DBG_READ_LEN 256
+#define HNS3_DBG_WRITE_LEN 1024
static struct dentry *hns3_dbgfs_root;
@@ -322,6 +323,9 @@ static ssize_t hns3_dbg_cmd_write(struct file *filp, const char __user *buffer,
test_bit(HNS3_NIC_STATE_RESETTING, &priv->state))
return 0;
+ if (count > HNS3_DBG_WRITE_LEN)
+ return -ENOSPC;
+
cmd_buf = kzalloc(count + 1, GFP_KERNEL);
if (!cmd_buf)
return count;
--
2.7.4
^ permalink raw reply related
* [PATCH net-next 08/12] net: hns3: add check for max TX BD num for tso and non-tso case
From: Huazhong Tan @ 2019-08-09 2:31 UTC (permalink / raw)
To: davem
Cc: netdev, linux-kernel, salil.mehta, yisen.zhuang, linuxarm,
Yunsheng Lin, Huazhong Tan
In-Reply-To: <1565317878-31806-1-git-send-email-tanhuazhong@huawei.com>
From: Yunsheng Lin <linyunsheng@huawei.com>
Hardware supports up to 8 TX BD for non-TSO skb and 63 TX
BD for TSO skb. Currently hns3 driver does not check the max
BD num that required by a skb before filling desc, which may
cause the hardware to issue a RAS error throug PCIe AER.
This patch adds the max BD num check before filling desc,
if the bd num is not within the hardware limit, it will
record the error by ring->stats.sw_err_cnt counter and
free the skb.
This patch also cleans up the hns3_nic_bd_num function by
changing the return type and removing an unnecessary check.
Signed-off-by: Yunsheng Lin <linyunsheng@huawei.com>
Reviewed-by: Peng Li <lipeng321@huawei.com>
Signed-off-by: Huazhong Tan <tanhuazhong@huawei.com>
---
drivers/net/ethernet/hisilicon/hns3/hns3_enet.c | 38 ++++++++++---------------
drivers/net/ethernet/hisilicon/hns3/hns3_enet.h | 3 +-
2 files changed, 17 insertions(+), 24 deletions(-)
diff --git a/drivers/net/ethernet/hisilicon/hns3/hns3_enet.c b/drivers/net/ethernet/hisilicon/hns3/hns3_enet.c
index b2a668d..df08f9e 100644
--- a/drivers/net/ethernet/hisilicon/hns3/hns3_enet.c
+++ b/drivers/net/ethernet/hisilicon/hns3/hns3_enet.c
@@ -1186,28 +1186,20 @@ static int hns3_fill_desc(struct hns3_enet_ring *ring, void *priv,
return 0;
}
-static int hns3_nic_bd_num(struct sk_buff *skb)
+static unsigned int hns3_nic_bd_num(struct sk_buff *skb)
{
- int size = skb_headlen(skb);
- int i, bd_num;
+ unsigned int bd_num;
+ int i;
/* if the total len is within the max bd limit */
if (likely(skb->len <= HNS3_MAX_BD_SIZE))
return skb_shinfo(skb)->nr_frags + 1;
- bd_num = hns3_tx_bd_count(size);
+ bd_num = hns3_tx_bd_count(skb_headlen(skb));
for (i = 0; i < skb_shinfo(skb)->nr_frags; i++) {
skb_frag_t *frag = &skb_shinfo(skb)->frags[i];
- int frag_bd_num;
-
- size = skb_frag_size(frag);
- frag_bd_num = hns3_tx_bd_count(size);
-
- if (unlikely(frag_bd_num > HNS3_MAX_BD_PER_FRAG))
- return -ENOMEM;
-
- bd_num += frag_bd_num;
+ bd_num += hns3_tx_bd_count(skb_frag_size(frag));
}
return bd_num;
@@ -1228,7 +1220,7 @@ static unsigned int hns3_gso_hdr_len(struct sk_buff *skb)
*/
static bool hns3_skb_need_linearized(struct sk_buff *skb)
{
- int bd_limit = HNS3_MAX_BD_PER_FRAG - 1;
+ int bd_limit = HNS3_MAX_BD_NUM_NORMAL - 1;
unsigned int tot_len = 0;
int i;
@@ -1258,21 +1250,16 @@ static int hns3_nic_maybe_stop_tx(struct hns3_enet_ring *ring,
struct sk_buff **out_skb)
{
struct sk_buff *skb = *out_skb;
- int bd_num;
+ unsigned int bd_num;
bd_num = hns3_nic_bd_num(skb);
- if (bd_num < 0)
- return bd_num;
-
- if (unlikely(bd_num > HNS3_MAX_BD_PER_FRAG)) {
+ if (unlikely(bd_num > HNS3_MAX_BD_NUM_NORMAL)) {
struct sk_buff *new_skb;
- if (skb_is_gso(skb) && !hns3_skb_need_linearized(skb))
+ if (skb_is_gso(skb) && bd_num <= HNS3_MAX_BD_NUM_TSO &&
+ !hns3_skb_need_linearized(skb))
goto out;
- bd_num = hns3_tx_bd_count(skb->len);
- if (unlikely(ring_space(ring) < bd_num))
- return -EBUSY;
/* manual split the send packet */
new_skb = skb_copy(skb, GFP_ATOMIC);
if (!new_skb)
@@ -1280,6 +1267,11 @@ static int hns3_nic_maybe_stop_tx(struct hns3_enet_ring *ring,
dev_kfree_skb_any(skb);
*out_skb = new_skb;
+ bd_num = hns3_nic_bd_num(new_skb);
+ if ((skb_is_gso(new_skb) && bd_num > HNS3_MAX_BD_NUM_TSO) ||
+ (!skb_is_gso(new_skb) && bd_num > HNS3_MAX_BD_NUM_NORMAL))
+ return -ENOMEM;
+
u64_stats_update_begin(&ring->syncp);
ring->stats.tx_copy++;
u64_stats_update_end(&ring->syncp);
diff --git a/drivers/net/ethernet/hisilicon/hns3/hns3_enet.h b/drivers/net/ethernet/hisilicon/hns3/hns3_enet.h
index a76712c..5b0ee1f 100644
--- a/drivers/net/ethernet/hisilicon/hns3/hns3_enet.h
+++ b/drivers/net/ethernet/hisilicon/hns3/hns3_enet.h
@@ -195,7 +195,8 @@ enum hns3_nic_state {
#define HNS3_VECTOR_INITED 1
#define HNS3_MAX_BD_SIZE 65535
-#define HNS3_MAX_BD_PER_FRAG 8
+#define HNS3_MAX_BD_NUM_NORMAL 8
+#define HNS3_MAX_BD_NUM_TSO 63
#define HNS3_MAX_BD_PER_PKT MAX_SKB_FRAGS
#define HNS3_VECTOR_GL0_OFFSET 0x100
--
2.7.4
^ permalink raw reply related
* [PATCH net-next 03/12] net: hns3: clean up for vlan handling in hns3_fill_desc_vtags
From: Huazhong Tan @ 2019-08-09 2:31 UTC (permalink / raw)
To: davem
Cc: netdev, linux-kernel, salil.mehta, yisen.zhuang, linuxarm,
Yunsheng Lin, Huazhong Tan
In-Reply-To: <1565317878-31806-1-git-send-email-tanhuazhong@huawei.com>
From: Yunsheng Lin <linyunsheng@huawei.com>
This patch refactors the hns3_fill_desc_vtags function
by avoiding passing too many parameters, reducing indent
level and some other clean up.
This patch also adds the hns3_fill_skb_desc function to
fill the first desc of a skb.
Signed-off-by: Yunsheng Lin <linyunsheng@huawei.com>
Reviewed-by: Peng Li <lipeng321@huawei.com>
Signed-off-by: Huazhong Tan <tanhuazhong@huawei.com>
---
drivers/net/ethernet/hisilicon/hns3/hns3_enet.c | 167 +++++++++++++-----------
1 file changed, 89 insertions(+), 78 deletions(-)
diff --git a/drivers/net/ethernet/hisilicon/hns3/hns3_enet.c b/drivers/net/ethernet/hisilicon/hns3/hns3_enet.c
index ed05fb9..fd6a3d5 100644
--- a/drivers/net/ethernet/hisilicon/hns3/hns3_enet.c
+++ b/drivers/net/ethernet/hisilicon/hns3/hns3_enet.c
@@ -45,6 +45,9 @@ MODULE_PARM_DESC(debug, " Network interface message level setting");
#define DEFAULT_MSG_LEVEL (NETIF_MSG_PROBE | NETIF_MSG_LINK | \
NETIF_MSG_IFDOWN | NETIF_MSG_IFUP)
+#define HNS3_INNER_VLAN_TAG 1
+#define HNS3_OUTER_VLAN_TAG 2
+
/* hns3_pci_tbl - PCI Device ID Table
*
* Last entry must be all 0s
@@ -961,16 +964,16 @@ static void hns3_set_txbd_baseinfo(u16 *bdtp_fe_sc_vld_ra_ri, int frag_end)
hns3_set_field(*bdtp_fe_sc_vld_ra_ri, HNS3_TXD_VLD_B, 1U);
}
-static int hns3_fill_desc_vtags(struct sk_buff *skb,
- struct hns3_enet_ring *tx_ring,
- u32 *inner_vlan_flag,
- u32 *out_vlan_flag,
- u16 *inner_vtag,
- u16 *out_vtag)
+static int hns3_handle_vtags(struct hns3_enet_ring *tx_ring,
+ struct sk_buff *skb)
{
-#define HNS3_TX_VLAN_PRIO_SHIFT 13
-
struct hnae3_handle *handle = tx_ring->tqp->handle;
+ struct vlan_ethhdr *vhdr;
+ int rc;
+
+ if (!(skb->protocol == htons(ETH_P_8021Q) ||
+ skb_vlan_tag_present(skb)))
+ return 0;
/* Since HW limitation, if port based insert VLAN enabled, only one VLAN
* header is allowed in skb, otherwise it will cause RAS error.
@@ -981,8 +984,7 @@ static int hns3_fill_desc_vtags(struct sk_buff *skb,
return -EINVAL;
if (skb->protocol == htons(ETH_P_8021Q) &&
- !(tx_ring->tqp->handle->kinfo.netdev->features &
- NETIF_F_HW_VLAN_CTAG_TX)) {
+ !(handle->kinfo.netdev->features & NETIF_F_HW_VLAN_CTAG_TX)) {
/* When HW VLAN acceleration is turned off, and the stack
* sets the protocol to 802.1q, the driver just need to
* set the protocol to the encapsulated ethertype.
@@ -992,45 +994,92 @@ static int hns3_fill_desc_vtags(struct sk_buff *skb,
}
if (skb_vlan_tag_present(skb)) {
- u16 vlan_tag;
-
- vlan_tag = skb_vlan_tag_get(skb);
- vlan_tag |= (skb->priority & 0x7) << HNS3_TX_VLAN_PRIO_SHIFT;
-
/* Based on hw strategy, use out_vtag in two layer tag case,
* and use inner_vtag in one tag case.
*/
- if (skb->protocol == htons(ETH_P_8021Q)) {
- if (handle->port_base_vlan_state ==
- HNAE3_PORT_BASE_VLAN_DISABLE){
- hns3_set_field(*out_vlan_flag,
- HNS3_TXD_OVLAN_B, 1);
- *out_vtag = vlan_tag;
- } else {
- hns3_set_field(*inner_vlan_flag,
- HNS3_TXD_VLAN_B, 1);
- *inner_vtag = vlan_tag;
- }
- } else {
- hns3_set_field(*inner_vlan_flag, HNS3_TXD_VLAN_B, 1);
- *inner_vtag = vlan_tag;
- }
- } else if (skb->protocol == htons(ETH_P_8021Q)) {
- struct vlan_ethhdr *vhdr;
- int rc;
+ if (skb->protocol == htons(ETH_P_8021Q) &&
+ handle->port_base_vlan_state ==
+ HNAE3_PORT_BASE_VLAN_DISABLE)
+ rc = HNS3_OUTER_VLAN_TAG;
+ else
+ rc = HNS3_INNER_VLAN_TAG;
- rc = skb_cow_head(skb, 0);
- if (unlikely(rc < 0))
- return rc;
- vhdr = (struct vlan_ethhdr *)skb->data;
- vhdr->h_vlan_TCI |= cpu_to_be16((skb->priority & 0x7)
- << HNS3_TX_VLAN_PRIO_SHIFT);
+ skb->protocol = vlan_get_protocol(skb);
+ return rc;
}
+ rc = skb_cow_head(skb, 0);
+ if (unlikely(rc < 0))
+ return rc;
+
+ vhdr = (struct vlan_ethhdr *)skb->data;
+ vhdr->h_vlan_TCI |= cpu_to_be16((skb->priority << VLAN_PRIO_SHIFT)
+ & VLAN_PRIO_MASK);
+
skb->protocol = vlan_get_protocol(skb);
return 0;
}
+static int hns3_fill_skb_desc(struct hns3_enet_ring *ring,
+ struct sk_buff *skb, struct hns3_desc *desc)
+{
+ u32 ol_type_vlan_len_msec = 0;
+ u32 type_cs_vlan_tso = 0;
+ u32 paylen = skb->len;
+ u16 inner_vtag = 0;
+ u16 out_vtag = 0;
+ u16 mss = 0;
+ int ret;
+
+ ret = hns3_handle_vtags(ring, skb);
+ if (unlikely(ret < 0)) {
+ return ret;
+ } else if (ret == HNS3_INNER_VLAN_TAG) {
+ inner_vtag = skb_vlan_tag_get(skb);
+ inner_vtag |= (skb->priority << VLAN_PRIO_SHIFT) &
+ VLAN_PRIO_MASK;
+ hns3_set_field(type_cs_vlan_tso, HNS3_TXD_VLAN_B, 1);
+ } else if (ret == HNS3_OUTER_VLAN_TAG) {
+ out_vtag = skb_vlan_tag_get(skb);
+ out_vtag |= (skb->priority << VLAN_PRIO_SHIFT) &
+ VLAN_PRIO_MASK;
+ hns3_set_field(ol_type_vlan_len_msec, HNS3_TXD_OVLAN_B,
+ 1);
+ }
+
+ if (skb->ip_summed == CHECKSUM_PARTIAL) {
+ u8 ol4_proto, il4_proto;
+
+ skb_reset_mac_len(skb);
+
+ ret = hns3_get_l4_protocol(skb, &ol4_proto, &il4_proto);
+ if (unlikely(ret))
+ return ret;
+
+ ret = hns3_set_l2l3l4(skb, ol4_proto, il4_proto,
+ &type_cs_vlan_tso,
+ &ol_type_vlan_len_msec);
+ if (unlikely(ret))
+ return ret;
+
+ ret = hns3_set_tso(skb, &paylen, &mss,
+ &type_cs_vlan_tso);
+ if (unlikely(ret))
+ return ret;
+ }
+
+ /* Set txbd */
+ desc->tx.ol_type_vlan_len_msec =
+ cpu_to_le32(ol_type_vlan_len_msec);
+ desc->tx.type_cs_vlan_tso_len = cpu_to_le32(type_cs_vlan_tso);
+ desc->tx.paylen = cpu_to_le32(paylen);
+ desc->tx.mss = cpu_to_le16(mss);
+ desc->tx.vlan_tag = cpu_to_le16(inner_vtag);
+ desc->tx.outer_vlan_tag = cpu_to_le16(out_vtag);
+
+ return 0;
+}
+
static int hns3_fill_desc(struct hns3_enet_ring *ring, void *priv,
unsigned int size, int frag_end,
enum hns_desc_type type)
@@ -1045,50 +1094,12 @@ static int hns3_fill_desc(struct hns3_enet_ring *ring, void *priv,
if (type == DESC_TYPE_SKB) {
struct sk_buff *skb = (struct sk_buff *)priv;
- u32 ol_type_vlan_len_msec = 0;
- u32 type_cs_vlan_tso = 0;
- u32 paylen = skb->len;
- u16 inner_vtag = 0;
- u16 out_vtag = 0;
- u16 mss = 0;
int ret;
- ret = hns3_fill_desc_vtags(skb, ring, &type_cs_vlan_tso,
- &ol_type_vlan_len_msec,
- &inner_vtag, &out_vtag);
+ ret = hns3_fill_skb_desc(ring, skb, desc);
if (unlikely(ret))
return ret;
- if (skb->ip_summed == CHECKSUM_PARTIAL) {
- u8 ol4_proto, il4_proto;
-
- skb_reset_mac_len(skb);
-
- ret = hns3_get_l4_protocol(skb, &ol4_proto, &il4_proto);
- if (unlikely(ret))
- return ret;
-
- ret = hns3_set_l2l3l4(skb, ol4_proto, il4_proto,
- &type_cs_vlan_tso,
- &ol_type_vlan_len_msec);
- if (unlikely(ret))
- return ret;
-
- ret = hns3_set_tso(skb, &paylen, &mss,
- &type_cs_vlan_tso);
- if (unlikely(ret))
- return ret;
- }
-
- /* Set txbd */
- desc->tx.ol_type_vlan_len_msec =
- cpu_to_le32(ol_type_vlan_len_msec);
- desc->tx.type_cs_vlan_tso_len = cpu_to_le32(type_cs_vlan_tso);
- desc->tx.paylen = cpu_to_le32(paylen);
- desc->tx.mss = cpu_to_le16(mss);
- desc->tx.vlan_tag = cpu_to_le16(inner_vtag);
- desc->tx.outer_vlan_tag = cpu_to_le16(out_vtag);
-
dma = dma_map_single(dev, skb->data, size, DMA_TO_DEVICE);
} else {
frag = (skb_frag_t *)priv;
--
2.7.4
^ permalink raw reply related
* [PATCH net-next 10/12] net: hns3: refine MAC pause statistics querying function
From: Huazhong Tan @ 2019-08-09 2:31 UTC (permalink / raw)
To: davem
Cc: netdev, linux-kernel, salil.mehta, yisen.zhuang, linuxarm,
Yufeng Mo, Huazhong Tan
In-Reply-To: <1565317878-31806-1-git-send-email-tanhuazhong@huawei.com>
From: Yufeng Mo <moyufeng@huawei.com>
This patch refines the interface for querying MAC pause
statistics, and adds structure hns3_mac_stats to keep the
count of TX & RX.
Signed-off-by: Yufeng Mo <moyufeng@huawei.com>
Reviewed-by: Peng Li <lipeng321@huawei.com>
Signed-off-by: Huazhong Tan <tanhuazhong@huawei.com>
---
drivers/net/ethernet/hisilicon/hns3/hnae3.h | 11 +++++++++--
drivers/net/ethernet/hisilicon/hns3/hns3_enet.c | 11 ++++-------
drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c | 12 +++++++-----
3 files changed, 20 insertions(+), 14 deletions(-)
diff --git a/drivers/net/ethernet/hisilicon/hns3/hnae3.h b/drivers/net/ethernet/hisilicon/hns3/hnae3.h
index a4624db..43740ee 100644
--- a/drivers/net/ethernet/hisilicon/hns3/hnae3.h
+++ b/drivers/net/ethernet/hisilicon/hns3/hnae3.h
@@ -91,6 +91,11 @@ struct hnae3_queue {
u16 rx_desc_num;/* total number of rx desc */
};
+struct hns3_mac_stats {
+ u64 tx_pause_cnt;
+ u64 rx_pause_cnt;
+};
+
/*hnae3 loop mode*/
enum hnae3_loop {
HNAE3_LOOP_APP,
@@ -298,6 +303,8 @@ struct hnae3_ae_dev {
* Remove multicast address from mac table
* update_stats()
* Update Old network device statistics
+ * get_mac_stats()
+ * get mac pause statistics including tx_cnt and rx_cnt
* get_ethtool_stats()
* Get ethtool network device statistics
* get_strings()
@@ -426,8 +433,8 @@ struct hnae3_ae_ops {
void (*update_stats)(struct hnae3_handle *handle,
struct net_device_stats *net_stats);
void (*get_stats)(struct hnae3_handle *handle, u64 *data);
- void (*get_mac_pause_stats)(struct hnae3_handle *handle, u64 *tx_cnt,
- u64 *rx_cnt);
+ void (*get_mac_stats)(struct hnae3_handle *handle,
+ struct hns3_mac_stats *mac_stats);
void (*get_strings)(struct hnae3_handle *handle,
u32 stringset, u8 *data);
int (*get_sset_count)(struct hnae3_handle *handle, int stringset);
diff --git a/drivers/net/ethernet/hisilicon/hns3/hns3_enet.c b/drivers/net/ethernet/hisilicon/hns3/hns3_enet.c
index df08f9e..1750f80 100644
--- a/drivers/net/ethernet/hisilicon/hns3/hns3_enet.c
+++ b/drivers/net/ethernet/hisilicon/hns3/hns3_enet.c
@@ -1726,15 +1726,12 @@ static bool hns3_get_tx_timeo_queue_info(struct net_device *ndev)
/* When mac received many pause frames continuous, it's unable to send
* packets, which may cause tx timeout
*/
- if (h->ae_algo->ops->update_stats &&
- h->ae_algo->ops->get_mac_pause_stats) {
- u64 tx_pause_cnt, rx_pause_cnt;
+ if (h->ae_algo->ops->get_mac_stats) {
+ struct hns3_mac_stats mac_stats;
- h->ae_algo->ops->update_stats(h, &ndev->stats);
- h->ae_algo->ops->get_mac_pause_stats(h, &tx_pause_cnt,
- &rx_pause_cnt);
+ h->ae_algo->ops->get_mac_stats(h, &mac_stats);
netdev_info(ndev, "tx_pause_cnt: %llu, rx_pause_cnt: %llu\n",
- tx_pause_cnt, rx_pause_cnt);
+ mac_stats.tx_pause_cnt, mac_stats.rx_pause_cnt);
}
hw_head = readl_relaxed(tx_ring->tqp->io_base +
diff --git a/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c b/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c
index 7d7ab9e..1315275 100644
--- a/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c
+++ b/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c
@@ -750,14 +750,16 @@ static void hclge_get_stats(struct hnae3_handle *handle, u64 *data)
p = hclge_tqps_get_stats(handle, p);
}
-static void hclge_get_mac_pause_stat(struct hnae3_handle *handle, u64 *tx_cnt,
- u64 *rx_cnt)
+static void hclge_get_mac_stat(struct hnae3_handle *handle,
+ struct hns3_mac_stats *mac_stats)
{
struct hclge_vport *vport = hclge_get_vport(handle);
struct hclge_dev *hdev = vport->back;
- *tx_cnt = hdev->hw_stats.mac_stats.mac_tx_mac_pause_num;
- *rx_cnt = hdev->hw_stats.mac_stats.mac_rx_mac_pause_num;
+ hclge_update_stats(handle, NULL);
+
+ mac_stats->tx_pause_cnt = hdev->hw_stats.mac_stats.mac_tx_mac_pause_num;
+ mac_stats->rx_pause_cnt = hdev->hw_stats.mac_stats.mac_rx_mac_pause_num;
}
static int hclge_parse_func_status(struct hclge_dev *hdev,
@@ -9798,7 +9800,7 @@ static const struct hnae3_ae_ops hclge_ops = {
.set_mtu = hclge_set_mtu,
.reset_queue = hclge_reset_tqp,
.get_stats = hclge_get_stats,
- .get_mac_pause_stats = hclge_get_mac_pause_stat,
+ .get_mac_stats = hclge_get_mac_stat,
.update_stats = hclge_update_stats,
.get_strings = hclge_get_strings,
.get_sset_count = hclge_get_sset_count,
--
2.7.4
^ permalink raw reply related
* [PATCH net-next 05/12] net: hns3: modify how pause options is displayed
From: Huazhong Tan @ 2019-08-09 2:31 UTC (permalink / raw)
To: davem
Cc: netdev, linux-kernel, salil.mehta, yisen.zhuang, linuxarm,
Yonglong Liu, Huazhong Tan
In-Reply-To: <1565317878-31806-1-git-send-email-tanhuazhong@huawei.com>
From: Yonglong Liu <liuyonglong@huawei.com>
Currently, the pause options of HNS3 shown like this:
"RX/TX" is always the same with "RX negotiated/TX negotiated".
Because of the driver covered the value of "RX/TX" with the value
of "RX negotiated/TX negotiated" after adjust link.
This patch records the pause configurations of the user, and never
covered them in adjust link.
Signed-off-by: Yonglong Liu <liuyonglong@huawei.com>
Reviewed-by: Yunsheng Lin <linyunsheng@huawei.com>
Signed-off-by: Huazhong Tan <tanhuazhong@huawei.com>
---
.../ethernet/hisilicon/hns3/hns3pf/hclge_main.c | 38 ++++++++++++----------
1 file changed, 21 insertions(+), 17 deletions(-)
diff --git a/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c b/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c
index c0feae3a..381f195 100644
--- a/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c
+++ b/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c
@@ -8207,28 +8207,15 @@ static int hclge_cfg_pauseparam(struct hclge_dev *hdev, u32 rx_en, u32 tx_en)
{
int ret;
- if (rx_en && tx_en)
- hdev->fc_mode_last_time = HCLGE_FC_FULL;
- else if (rx_en && !tx_en)
- hdev->fc_mode_last_time = HCLGE_FC_RX_PAUSE;
- else if (!rx_en && tx_en)
- hdev->fc_mode_last_time = HCLGE_FC_TX_PAUSE;
- else
- hdev->fc_mode_last_time = HCLGE_FC_NONE;
-
if (hdev->tm_info.fc_mode == HCLGE_FC_PFC)
return 0;
ret = hclge_mac_pause_en_cfg(hdev, tx_en, rx_en);
- if (ret) {
- dev_err(&hdev->pdev->dev, "configure pauseparam error, ret = %d.\n",
- ret);
- return ret;
- }
-
- hdev->tm_info.fc_mode = hdev->fc_mode_last_time;
+ if (ret)
+ dev_err(&hdev->pdev->dev,
+ "configure pauseparam error, ret = %d.\n", ret);
- return 0;
+ return ret;
}
int hclge_cfg_flowctrl(struct hclge_dev *hdev)
@@ -8293,6 +8280,21 @@ static void hclge_get_pauseparam(struct hnae3_handle *handle, u32 *auto_neg,
}
}
+static void hclge_record_user_pauseparam(struct hclge_dev *hdev,
+ u32 rx_en, u32 tx_en)
+{
+ if (rx_en && tx_en)
+ hdev->fc_mode_last_time = HCLGE_FC_FULL;
+ else if (rx_en && !tx_en)
+ hdev->fc_mode_last_time = HCLGE_FC_RX_PAUSE;
+ else if (!rx_en && tx_en)
+ hdev->fc_mode_last_time = HCLGE_FC_TX_PAUSE;
+ else
+ hdev->fc_mode_last_time = HCLGE_FC_NONE;
+
+ hdev->tm_info.fc_mode = hdev->fc_mode_last_time;
+}
+
static int hclge_set_pauseparam(struct hnae3_handle *handle, u32 auto_neg,
u32 rx_en, u32 tx_en)
{
@@ -8318,6 +8320,8 @@ static int hclge_set_pauseparam(struct hnae3_handle *handle, u32 auto_neg,
hclge_set_flowctrl_adv(hdev, rx_en, tx_en);
+ hclge_record_user_pauseparam(hdev, rx_en, tx_en);
+
if (!auto_neg)
return hclge_cfg_pauseparam(hdev, rx_en, tx_en);
--
2.7.4
^ permalink raw reply related
* [PATCH net-next 06/12] net: hns3: add DFX registers information for ethtool -d
From: Huazhong Tan @ 2019-08-09 2:31 UTC (permalink / raw)
To: davem
Cc: netdev, linux-kernel, salil.mehta, yisen.zhuang, linuxarm,
Guangbin Huang, Huazhong Tan
In-Reply-To: <1565317878-31806-1-git-send-email-tanhuazhong@huawei.com>
From: Guangbin Huang <huangguangbin2@huawei.com>
Now we can use ethtool -d command to dump some registers. However,
these registers information is not enough to find out where the problem is.
This patch adds DFX registers information after original registers
when use ethtool -d commmand to dump registers. Also, using macro
replaces some related magic number.
Signed-off-by: Guangbin Huang <huangguangbin2@huawei.com>
Reviewed-by: Peng Li <lipeng321@huawei.com>
Signed-off-by: Huazhong Tan <tanhuazhong@huawei.com>
---
.../ethernet/hisilicon/hns3/hns3pf/hclge_debugfs.c | 12 +-
.../ethernet/hisilicon/hns3/hns3pf/hclge_main.c | 342 ++++++++++++++++++---
.../ethernet/hisilicon/hns3/hns3pf/hclge_main.h | 2 +
3 files changed, 301 insertions(+), 55 deletions(-)
diff --git a/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_debugfs.c b/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_debugfs.c
index f16bfc6..933dec5 100644
--- a/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_debugfs.c
+++ b/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_debugfs.c
@@ -14,16 +14,8 @@ static int hclge_dbg_get_dfx_bd_num(struct hclge_dev *hdev, int offset)
struct hclge_desc desc[4];
int ret;
- hclge_cmd_setup_basic_desc(&desc[0], HCLGE_OPC_DFX_BD_NUM, true);
- desc[0].flag |= cpu_to_le16(HCLGE_CMD_FLAG_NEXT);
- hclge_cmd_setup_basic_desc(&desc[1], HCLGE_OPC_DFX_BD_NUM, true);
- desc[1].flag |= cpu_to_le16(HCLGE_CMD_FLAG_NEXT);
- hclge_cmd_setup_basic_desc(&desc[2], HCLGE_OPC_DFX_BD_NUM, true);
- desc[2].flag |= cpu_to_le16(HCLGE_CMD_FLAG_NEXT);
- hclge_cmd_setup_basic_desc(&desc[3], HCLGE_OPC_DFX_BD_NUM, true);
-
- ret = hclge_cmd_send(&hdev->hw, desc, 4);
- if (ret != HCLGE_CMD_EXEC_SUCCESS) {
+ ret = hclge_query_bd_num_cmd_send(hdev, desc);
+ if (ret) {
dev_err(&hdev->pdev->dev,
"get dfx bdnum fail, status is %d.\n", ret);
return ret;
diff --git a/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c b/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c
index 381f195..7d7ab9e 100644
--- a/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c
+++ b/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c
@@ -36,6 +36,20 @@
#define HCLGE_RESET_MAX_FAIL_CNT 5
+/* Get DFX BD number offset */
+#define HCLGE_DFX_BIOS_BD_OFFSET 1
+#define HCLGE_DFX_SSU_0_BD_OFFSET 2
+#define HCLGE_DFX_SSU_1_BD_OFFSET 3
+#define HCLGE_DFX_IGU_BD_OFFSET 4
+#define HCLGE_DFX_RPU_0_BD_OFFSET 5
+#define HCLGE_DFX_RPU_1_BD_OFFSET 6
+#define HCLGE_DFX_NCSI_BD_OFFSET 7
+#define HCLGE_DFX_RTC_BD_OFFSET 8
+#define HCLGE_DFX_PPP_BD_OFFSET 9
+#define HCLGE_DFX_RCB_BD_OFFSET 10
+#define HCLGE_DFX_TQP_BD_OFFSET 11
+#define HCLGE_DFX_SSU_2_BD_OFFSET 12
+
static int hclge_set_mac_mtu(struct hclge_dev *hdev, int new_mps);
static int hclge_init_vlan_config(struct hclge_dev *hdev);
static void hclge_sync_vlan_filter(struct hclge_dev *hdev);
@@ -317,6 +331,36 @@ static const u8 hclge_hash_key[] = {
0x6A, 0x42, 0xB7, 0x3B, 0xBE, 0xAC, 0x01, 0xFA
};
+static const u32 hclge_dfx_bd_offset_list[] = {
+ HCLGE_DFX_BIOS_BD_OFFSET,
+ HCLGE_DFX_SSU_0_BD_OFFSET,
+ HCLGE_DFX_SSU_1_BD_OFFSET,
+ HCLGE_DFX_IGU_BD_OFFSET,
+ HCLGE_DFX_RPU_0_BD_OFFSET,
+ HCLGE_DFX_RPU_1_BD_OFFSET,
+ HCLGE_DFX_NCSI_BD_OFFSET,
+ HCLGE_DFX_RTC_BD_OFFSET,
+ HCLGE_DFX_PPP_BD_OFFSET,
+ HCLGE_DFX_RCB_BD_OFFSET,
+ HCLGE_DFX_TQP_BD_OFFSET,
+ HCLGE_DFX_SSU_2_BD_OFFSET
+};
+
+static const enum hclge_opcode_type hclge_dfx_reg_opcode_list[] = {
+ HCLGE_OPC_DFX_BIOS_COMMON_REG,
+ HCLGE_OPC_DFX_SSU_REG_0,
+ HCLGE_OPC_DFX_SSU_REG_1,
+ HCLGE_OPC_DFX_IGU_EGU_REG,
+ HCLGE_OPC_DFX_RPU_REG_0,
+ HCLGE_OPC_DFX_RPU_REG_1,
+ HCLGE_OPC_DFX_NCSI_REG,
+ HCLGE_OPC_DFX_RTC_REG,
+ HCLGE_OPC_DFX_PPP_REG,
+ HCLGE_OPC_DFX_RCB_REG,
+ HCLGE_OPC_DFX_TQP_REG,
+ HCLGE_OPC_DFX_SSU_REG_2
+};
+
static int hclge_mac_update_stats_defective(struct hclge_dev *hdev)
{
#define HCLGE_MAC_CMD_NUM 21
@@ -9332,106 +9376,314 @@ static int hclge_get_64_bit_regs(struct hclge_dev *hdev, u32 regs_num,
}
#define MAX_SEPARATE_NUM 4
-#define SEPARATOR_VALUE 0xFFFFFFFF
+#define SEPARATOR_VALUE 0xFDFCFBFA
#define REG_NUM_PER_LINE 4
#define REG_LEN_PER_LINE (REG_NUM_PER_LINE * sizeof(u32))
+#define REG_SEPARATOR_LINE 1
+#define REG_NUM_REMAIN_MASK 3
+#define BD_LIST_MAX_NUM 30
-static int hclge_get_regs_len(struct hnae3_handle *handle)
+int hclge_query_bd_num_cmd_send(struct hclge_dev *hdev, struct hclge_desc *desc)
{
- int cmdq_lines, common_lines, ring_lines, tqp_intr_lines;
- struct hnae3_knic_private_info *kinfo = &handle->kinfo;
- struct hclge_vport *vport = hclge_get_vport(handle);
- struct hclge_dev *hdev = vport->back;
- u32 regs_num_32_bit, regs_num_64_bit;
+ /*prepare 4 commands to query DFX BD number*/
+ hclge_cmd_setup_basic_desc(&desc[0], HCLGE_OPC_DFX_BD_NUM, true);
+ desc[0].flag |= cpu_to_le16(HCLGE_CMD_FLAG_NEXT);
+ hclge_cmd_setup_basic_desc(&desc[1], HCLGE_OPC_DFX_BD_NUM, true);
+ desc[1].flag |= cpu_to_le16(HCLGE_CMD_FLAG_NEXT);
+ hclge_cmd_setup_basic_desc(&desc[2], HCLGE_OPC_DFX_BD_NUM, true);
+ desc[2].flag |= cpu_to_le16(HCLGE_CMD_FLAG_NEXT);
+ hclge_cmd_setup_basic_desc(&desc[3], HCLGE_OPC_DFX_BD_NUM, true);
+
+ return hclge_cmd_send(&hdev->hw, desc, 4);
+}
+
+static int hclge_get_dfx_reg_bd_num(struct hclge_dev *hdev,
+ int *bd_num_list,
+ u32 type_num)
+{
+#define HCLGE_DFX_REG_BD_NUM 4
+
+ u32 entries_per_desc, desc_index, index, offset, i;
+ struct hclge_desc desc[HCLGE_DFX_REG_BD_NUM];
int ret;
- ret = hclge_get_regs_num(hdev, ®s_num_32_bit, ®s_num_64_bit);
+ ret = hclge_query_bd_num_cmd_send(hdev, desc);
if (ret) {
dev_err(&hdev->pdev->dev,
- "Get register number failed, ret = %d.\n", ret);
- return -EOPNOTSUPP;
+ "Get dfx bd num fail, status is %d.\n", ret);
+ return ret;
}
- cmdq_lines = sizeof(cmdq_reg_addr_list) / REG_LEN_PER_LINE + 1;
- common_lines = sizeof(common_reg_addr_list) / REG_LEN_PER_LINE + 1;
- ring_lines = sizeof(ring_reg_addr_list) / REG_LEN_PER_LINE + 1;
- tqp_intr_lines = sizeof(tqp_intr_reg_addr_list) / REG_LEN_PER_LINE + 1;
+ entries_per_desc = ARRAY_SIZE(desc[0].data);
+ for (i = 0; i < type_num; i++) {
+ offset = hclge_dfx_bd_offset_list[i];
+ index = offset % entries_per_desc;
+ desc_index = offset / entries_per_desc;
+ bd_num_list[i] = le32_to_cpu(desc[desc_index].data[index]);
+ }
- return (cmdq_lines + common_lines + ring_lines * kinfo->num_tqps +
- tqp_intr_lines * (hdev->num_msi_used - 1)) * REG_LEN_PER_LINE +
- regs_num_32_bit * sizeof(u32) + regs_num_64_bit * sizeof(u64);
+ return ret;
}
-static void hclge_get_regs(struct hnae3_handle *handle, u32 *version,
- void *data)
+static int hclge_dfx_reg_cmd_send(struct hclge_dev *hdev,
+ struct hclge_desc *desc_src, int bd_num,
+ enum hclge_opcode_type cmd)
{
- struct hnae3_knic_private_info *kinfo = &handle->kinfo;
- struct hclge_vport *vport = hclge_get_vport(handle);
- struct hclge_dev *hdev = vport->back;
- u32 regs_num_32_bit, regs_num_64_bit;
- int i, j, reg_um, separator_num;
+ struct hclge_desc *desc = desc_src;
+ int i, ret;
+
+ hclge_cmd_setup_basic_desc(desc, cmd, true);
+ for (i = 0; i < bd_num - 1; i++) {
+ desc->flag |= cpu_to_le16(HCLGE_CMD_FLAG_NEXT);
+ desc++;
+ hclge_cmd_setup_basic_desc(desc, cmd, true);
+ }
+
+ desc = desc_src;
+ ret = hclge_cmd_send(&hdev->hw, desc, bd_num);
+ if (ret)
+ dev_err(&hdev->pdev->dev,
+ "Query dfx reg cmd(0x%x) send fail, status is %d.\n",
+ cmd, ret);
+
+ return ret;
+}
+
+static int hclge_dfx_reg_fetch_data(struct hclge_desc *desc_src, int bd_num,
+ void *data)
+{
+ int entries_per_desc, reg_num, separator_num, desc_index, index, i;
+ struct hclge_desc *desc = desc_src;
u32 *reg = data;
+
+ entries_per_desc = ARRAY_SIZE(desc->data);
+ reg_num = entries_per_desc * bd_num;
+ separator_num = REG_NUM_PER_LINE - (reg_num & REG_NUM_REMAIN_MASK);
+ for (i = 0; i < reg_num; i++) {
+ index = i % entries_per_desc;
+ desc_index = i / entries_per_desc;
+ *reg++ = le32_to_cpu(desc[desc_index].data[index]);
+ }
+ for (i = 0; i < separator_num; i++)
+ *reg++ = SEPARATOR_VALUE;
+
+ return reg_num + separator_num;
+}
+
+static int hclge_get_dfx_reg_len(struct hclge_dev *hdev, int *len)
+{
+ u32 dfx_reg_type_num = ARRAY_SIZE(hclge_dfx_bd_offset_list);
+ int data_len_per_desc, data_len, bd_num, i;
+ int bd_num_list[BD_LIST_MAX_NUM];
int ret;
- *version = hdev->fw_version;
+ ret = hclge_get_dfx_reg_bd_num(hdev, bd_num_list, dfx_reg_type_num);
+ if (ret) {
+ dev_err(&hdev->pdev->dev,
+ "Get dfx reg bd num fail, status is %d.\n", ret);
+ return ret;
+ }
- ret = hclge_get_regs_num(hdev, ®s_num_32_bit, ®s_num_64_bit);
+ data_len_per_desc = FIELD_SIZEOF(struct hclge_desc, data);
+ *len = 0;
+ for (i = 0; i < dfx_reg_type_num; i++) {
+ bd_num = bd_num_list[i];
+ data_len = data_len_per_desc * bd_num;
+ *len += (data_len / REG_LEN_PER_LINE + 1) * REG_LEN_PER_LINE;
+ }
+
+ return ret;
+}
+
+static int hclge_get_dfx_reg(struct hclge_dev *hdev, void *data)
+{
+ u32 dfx_reg_type_num = ARRAY_SIZE(hclge_dfx_bd_offset_list);
+ int bd_num, bd_num_max, buf_len, i;
+ int bd_num_list[BD_LIST_MAX_NUM];
+ struct hclge_desc *desc_src;
+ u32 *reg = data;
+ int ret;
+
+ ret = hclge_get_dfx_reg_bd_num(hdev, bd_num_list, dfx_reg_type_num);
if (ret) {
dev_err(&hdev->pdev->dev,
- "Get register number failed, ret = %d.\n", ret);
- return;
+ "Get dfx reg bd num fail, status is %d.\n", ret);
+ return ret;
+ }
+
+ bd_num_max = bd_num_list[0];
+ for (i = 1; i < dfx_reg_type_num; i++)
+ bd_num_max = max_t(int, bd_num_max, bd_num_list[i]);
+
+ buf_len = sizeof(*desc_src) * bd_num_max;
+ desc_src = kzalloc(buf_len, GFP_KERNEL);
+ if (!desc_src) {
+ dev_err(&hdev->pdev->dev, "%s kzalloc failed\n", __func__);
+ return -ENOMEM;
}
+ for (i = 0; i < dfx_reg_type_num; i++) {
+ bd_num = bd_num_list[i];
+ ret = hclge_dfx_reg_cmd_send(hdev, desc_src, bd_num,
+ hclge_dfx_reg_opcode_list[i]);
+ if (ret) {
+ dev_err(&hdev->pdev->dev,
+ "Get dfx reg fail, status is %d.\n", ret);
+ break;
+ }
+
+ reg += hclge_dfx_reg_fetch_data(desc_src, bd_num, reg);
+ }
+
+ kfree(desc_src);
+ return ret;
+}
+
+static int hclge_fetch_pf_reg(struct hclge_dev *hdev, void *data,
+ struct hnae3_knic_private_info *kinfo)
+{
+#define HCLGE_RING_REG_OFFSET 0x200
+#define HCLGE_RING_INT_REG_OFFSET 0x4
+
+ int i, j, reg_num, separator_num;
+ int data_num_sum;
+ u32 *reg = data;
+
/* fetching per-PF registers valus from PF PCIe register space */
- reg_um = sizeof(cmdq_reg_addr_list) / sizeof(u32);
- separator_num = MAX_SEPARATE_NUM - reg_um % REG_NUM_PER_LINE;
- for (i = 0; i < reg_um; i++)
+ reg_num = ARRAY_SIZE(cmdq_reg_addr_list);
+ separator_num = MAX_SEPARATE_NUM - (reg_num & REG_NUM_REMAIN_MASK);
+ for (i = 0; i < reg_num; i++)
*reg++ = hclge_read_dev(&hdev->hw, cmdq_reg_addr_list[i]);
for (i = 0; i < separator_num; i++)
*reg++ = SEPARATOR_VALUE;
+ data_num_sum = reg_num + separator_num;
- reg_um = sizeof(common_reg_addr_list) / sizeof(u32);
- separator_num = MAX_SEPARATE_NUM - reg_um % REG_NUM_PER_LINE;
- for (i = 0; i < reg_um; i++)
+ reg_num = ARRAY_SIZE(common_reg_addr_list);
+ separator_num = MAX_SEPARATE_NUM - (reg_num & REG_NUM_REMAIN_MASK);
+ for (i = 0; i < reg_num; i++)
*reg++ = hclge_read_dev(&hdev->hw, common_reg_addr_list[i]);
for (i = 0; i < separator_num; i++)
*reg++ = SEPARATOR_VALUE;
+ data_num_sum += reg_num + separator_num;
- reg_um = sizeof(ring_reg_addr_list) / sizeof(u32);
- separator_num = MAX_SEPARATE_NUM - reg_um % REG_NUM_PER_LINE;
+ reg_num = ARRAY_SIZE(ring_reg_addr_list);
+ separator_num = MAX_SEPARATE_NUM - (reg_num & REG_NUM_REMAIN_MASK);
for (j = 0; j < kinfo->num_tqps; j++) {
- for (i = 0; i < reg_um; i++)
+ for (i = 0; i < reg_num; i++)
*reg++ = hclge_read_dev(&hdev->hw,
ring_reg_addr_list[i] +
- 0x200 * j);
+ HCLGE_RING_REG_OFFSET * j);
for (i = 0; i < separator_num; i++)
*reg++ = SEPARATOR_VALUE;
}
+ data_num_sum += (reg_num + separator_num) * kinfo->num_tqps;
- reg_um = sizeof(tqp_intr_reg_addr_list) / sizeof(u32);
- separator_num = MAX_SEPARATE_NUM - reg_um % REG_NUM_PER_LINE;
+ reg_num = ARRAY_SIZE(tqp_intr_reg_addr_list);
+ separator_num = MAX_SEPARATE_NUM - (reg_num & REG_NUM_REMAIN_MASK);
for (j = 0; j < hdev->num_msi_used - 1; j++) {
- for (i = 0; i < reg_um; i++)
+ for (i = 0; i < reg_num; i++)
*reg++ = hclge_read_dev(&hdev->hw,
tqp_intr_reg_addr_list[i] +
- 4 * j);
+ HCLGE_RING_INT_REG_OFFSET * j);
for (i = 0; i < separator_num; i++)
*reg++ = SEPARATOR_VALUE;
}
+ data_num_sum += (reg_num + separator_num) * (hdev->num_msi_used - 1);
+
+ return data_num_sum;
+}
+
+static int hclge_get_regs_len(struct hnae3_handle *handle)
+{
+ int cmdq_lines, common_lines, ring_lines, tqp_intr_lines;
+ struct hnae3_knic_private_info *kinfo = &handle->kinfo;
+ struct hclge_vport *vport = hclge_get_vport(handle);
+ struct hclge_dev *hdev = vport->back;
+ int regs_num_32_bit, regs_num_64_bit, dfx_regs_len;
+ int regs_lines_32_bit, regs_lines_64_bit;
+ int ret;
+
+ ret = hclge_get_regs_num(hdev, ®s_num_32_bit, ®s_num_64_bit);
+ if (ret) {
+ dev_err(&hdev->pdev->dev,
+ "Get register number failed, ret = %d.\n", ret);
+ return ret;
+ }
+
+ ret = hclge_get_dfx_reg_len(hdev, &dfx_regs_len);
+ if (ret) {
+ dev_err(&hdev->pdev->dev,
+ "Get dfx reg len failed, ret = %d.\n", ret);
+ return ret;
+ }
+
+ cmdq_lines = sizeof(cmdq_reg_addr_list) / REG_LEN_PER_LINE +
+ REG_SEPARATOR_LINE;
+ common_lines = sizeof(common_reg_addr_list) / REG_LEN_PER_LINE +
+ REG_SEPARATOR_LINE;
+ ring_lines = sizeof(ring_reg_addr_list) / REG_LEN_PER_LINE +
+ REG_SEPARATOR_LINE;
+ tqp_intr_lines = sizeof(tqp_intr_reg_addr_list) / REG_LEN_PER_LINE +
+ REG_SEPARATOR_LINE;
+ regs_lines_32_bit = regs_num_32_bit * sizeof(u32) / REG_LEN_PER_LINE +
+ REG_SEPARATOR_LINE;
+ regs_lines_64_bit = regs_num_64_bit * sizeof(u64) / REG_LEN_PER_LINE +
+ REG_SEPARATOR_LINE;
+
+ return (cmdq_lines + common_lines + ring_lines * kinfo->num_tqps +
+ tqp_intr_lines * (hdev->num_msi_used - 1) + regs_lines_32_bit +
+ regs_lines_64_bit) * REG_LEN_PER_LINE + dfx_regs_len;
+}
+
+static void hclge_get_regs(struct hnae3_handle *handle, u32 *version,
+ void *data)
+{
+ struct hnae3_knic_private_info *kinfo = &handle->kinfo;
+ struct hclge_vport *vport = hclge_get_vport(handle);
+ struct hclge_dev *hdev = vport->back;
+ u32 regs_num_32_bit, regs_num_64_bit;
+ int i, reg_num, separator_num, ret;
+ u32 *reg = data;
+
+ *version = hdev->fw_version;
+
+ ret = hclge_get_regs_num(hdev, ®s_num_32_bit, ®s_num_64_bit);
+ if (ret) {
+ dev_err(&hdev->pdev->dev,
+ "Get register number failed, ret = %d.\n", ret);
+ return;
+ }
+
+ reg += hclge_fetch_pf_reg(hdev, reg, kinfo);
- /* fetching PF common registers values from firmware */
ret = hclge_get_32_bit_regs(hdev, regs_num_32_bit, reg);
if (ret) {
dev_err(&hdev->pdev->dev,
"Get 32 bit register failed, ret = %d.\n", ret);
return;
}
+ reg_num = regs_num_32_bit;
+ reg += reg_num;
+ separator_num = MAX_SEPARATE_NUM - (reg_num & REG_NUM_REMAIN_MASK);
+ for (i = 0; i < separator_num; i++)
+ *reg++ = SEPARATOR_VALUE;
- reg += regs_num_32_bit;
ret = hclge_get_64_bit_regs(hdev, regs_num_64_bit, reg);
- if (ret)
+ if (ret) {
dev_err(&hdev->pdev->dev,
"Get 64 bit register failed, ret = %d.\n", ret);
+ return;
+ }
+ reg_num = regs_num_64_bit * 2;
+ reg += reg_num;
+ separator_num = MAX_SEPARATE_NUM - (reg_num & REG_NUM_REMAIN_MASK);
+ for (i = 0; i < separator_num; i++)
+ *reg++ = SEPARATOR_VALUE;
+
+ ret = hclge_get_dfx_reg(hdev, reg);
+ if (ret)
+ dev_err(&hdev->pdev->dev,
+ "Get dfx register failed, ret = %d.\n", ret);
}
static int hclge_set_led_status(struct hclge_dev *hdev, u8 locate_led_status)
diff --git a/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.h b/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.h
index c9b9867f..f6d9b57 100644
--- a/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.h
+++ b/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.h
@@ -1029,4 +1029,6 @@ int hclge_push_vf_port_base_vlan_info(struct hclge_vport *vport, u8 vfid,
u16 state, u16 vlan_tag, u16 qos,
u16 vlan_proto);
void hclge_task_schedule(struct hclge_dev *hdev, unsigned long delay_time);
+int hclge_query_bd_num_cmd_send(struct hclge_dev *hdev,
+ struct hclge_desc *desc);
#endif
--
2.7.4
^ permalink raw reply related
* [PATCH net-next 01/12] net: hns3: fix GFP flag error in hclge_mac_update_stats()
From: Huazhong Tan @ 2019-08-09 2:31 UTC (permalink / raw)
To: davem
Cc: netdev, linux-kernel, salil.mehta, yisen.zhuang, linuxarm,
Zhongzhu Liu, Huazhong Tan
In-Reply-To: <1565317878-31806-1-git-send-email-tanhuazhong@huawei.com>
From: Zhongzhu Liu <liuzhongzhu@huawei.com>
When CONFIG_DEBUG_ATOMIC_SLEEP on, calling kzalloc with
GFP_KERNEL in hclge_mac_update_stats() will get below warning:
[ 52.514677] BUG: sleeping function called from invalid context at mm/slab.h:501
[ 52.522051] in_atomic(): 0, irqs_disabled(): 0, pid: 1015, name: ifconfig
[ 52.528827] 2 locks held by ifconfig/1015:
[ 52.532921] #0: (____ptrval____) (&p->lock){....}, at: seq_read+0x54/0x748
[ 52.539878] #1: (____ptrval____) (rcu_read_lock){....}, at: dev_seq_start+0x0/0x140
[ 52.547610] CPU: 16 PID: 1015 Comm: ifconfig Not tainted 5.3.0-rc3-00697-g20b80be #98
[ 52.555408] Hardware name: Huawei TaiShan 2280 V2/BC82AMDC, BIOS 2280-V2 CS V3.B050.01 08/08/2019
[ 52.564242] Call trace:
[ 52.566687] dump_backtrace+0x0/0x1f8
[ 52.570338] show_stack+0x14/0x20
[ 52.573646] dump_stack+0xb4/0xec
[ 52.576950] ___might_sleep+0x178/0x198
[ 52.580773] __might_sleep+0x74/0xe0
[ 52.584338] __kmalloc+0x244/0x2d8
[ 52.587744] hclge_mac_update_stats+0xc8/0x1f8 [hclge]
[ 52.592870] hclge_update_stats+0xe0/0x170 [hclge]
[ 52.597651] hns3_nic_get_stats64+0xa0/0x458 [hns3]
[ 52.602514] dev_get_stats+0x58/0x138
[ 52.606165] dev_seq_printf_stats+0x8c/0x280
[ 52.610420] dev_seq_show+0x14/0x40
[ 52.613898] seq_read+0x574/0x748
[ 52.617205] proc_reg_read+0xb4/0x108
[ 52.620857] __vfs_read+0x54/0xa8
[ 52.624162] vfs_read+0xa0/0x190
[ 52.627380] ksys_read+0xc8/0x178
[ 52.630685] __arm64_sys_read+0x40/0x50
[ 52.634509] el0_svc_common.constprop.0+0x120/0x1e0
[ 52.639369] el0_svc_handler+0x50/0x90
[ 52.643106] el0_svc+0x8/0xc
So this patch uses GFP_ATOMIC instead of GFP_KERNEL to fix it.
Fixes: d174ea75c96a ("net: hns3: add statistics for PFC frames and MAC control frames")
Signed-off-by: Zhongzhu Liu <liuzhongzhu@huawei.com>
Reviewed-by: Yunsheng Lin <linyunsheng@huawei.com>
Reviewed-by: Peng Li <lipeng321@huawei.com>
Signed-off-by: Huazhong Tan <tanhuazhong@huawei.com>
---
drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c | 6 +++++-
1 file changed, 5 insertions(+), 1 deletion(-)
diff --git a/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c b/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c
index b7399f5..c0feae3a 100644
--- a/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c
+++ b/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c
@@ -364,9 +364,13 @@ static int hclge_mac_update_stats_complete(struct hclge_dev *hdev, u32 desc_num)
u16 i, k, n;
int ret;
- desc = kcalloc(desc_num, sizeof(struct hclge_desc), GFP_KERNEL);
+ /* This may be called inside atomic sections,
+ * so GFP_ATOMIC is more suitalbe here
+ */
+ desc = kcalloc(desc_num, sizeof(struct hclge_desc), GFP_ATOMIC);
if (!desc)
return -ENOMEM;
+
hclge_cmd_setup_basic_desc(&desc[0], HCLGE_OPC_STATS_MAC_ALL, true);
ret = hclge_cmd_send(&hdev->hw, desc, desc_num);
if (ret) {
--
2.7.4
^ permalink raw reply related
* [PATCH net-next 07/12] net: hns3: add some statitics info to tx process
From: Huazhong Tan @ 2019-08-09 2:31 UTC (permalink / raw)
To: davem
Cc: netdev, linux-kernel, salil.mehta, yisen.zhuang, linuxarm,
Yunsheng Lin, Huazhong Tan
In-Reply-To: <1565317878-31806-1-git-send-email-tanhuazhong@huawei.com>
From: Yunsheng Lin <linyunsheng@huawei.com>
This patch adds tx_vlan_err, tx_l4_proto_err, tx_l2l3l4_err
and tx_tso_err counter to tx process, in order to better
debug the desc filling error.
This patch also adds a missing u64_stats_update_* around
ring->stats.sw_err_cnt and adds hns3_rl_err to limit the
error printing in the IO patch.
Signed-off-by: Yunsheng Lin <linyunsheng@huawei.com>
Reviewed-by: Peng Li <lipeng321@huawei.com>
Signed-off-by: Huazhong Tan <tanhuazhong@huawei.com>
---
drivers/net/ethernet/hisilicon/hns3/hns3_enet.c | 58 ++++++++++++++++------
drivers/net/ethernet/hisilicon/hns3/hns3_enet.h | 4 ++
drivers/net/ethernet/hisilicon/hns3/hns3_ethtool.c | 4 ++
3 files changed, 52 insertions(+), 14 deletions(-)
diff --git a/drivers/net/ethernet/hisilicon/hns3/hns3_enet.c b/drivers/net/ethernet/hisilicon/hns3/hns3_enet.c
index fd6a3d5..b2a668d 100644
--- a/drivers/net/ethernet/hisilicon/hns3/hns3_enet.c
+++ b/drivers/net/ethernet/hisilicon/hns3/hns3_enet.c
@@ -28,6 +28,12 @@
#define hns3_set_field(origin, shift, val) ((origin) |= ((val) << (shift)))
#define hns3_tx_bd_count(S) DIV_ROUND_UP(S, HNS3_MAX_BD_SIZE)
+#define hns3_rl_err(fmt, ...) \
+ do { \
+ if (net_ratelimit()) \
+ netdev_err(fmt, ##__VA_ARGS__); \
+ } while (0)
+
static void hns3_clear_all_ring(struct hnae3_handle *h, bool force);
static void hns3_remove_hw_addr(struct net_device *netdev);
@@ -1033,6 +1039,9 @@ static int hns3_fill_skb_desc(struct hns3_enet_ring *ring,
ret = hns3_handle_vtags(ring, skb);
if (unlikely(ret < 0)) {
+ u64_stats_update_begin(&ring->syncp);
+ ring->stats.tx_vlan_err++;
+ u64_stats_update_end(&ring->syncp);
return ret;
} else if (ret == HNS3_INNER_VLAN_TAG) {
inner_vtag = skb_vlan_tag_get(skb);
@@ -1053,19 +1062,31 @@ static int hns3_fill_skb_desc(struct hns3_enet_ring *ring,
skb_reset_mac_len(skb);
ret = hns3_get_l4_protocol(skb, &ol4_proto, &il4_proto);
- if (unlikely(ret))
+ if (unlikely(ret)) {
+ u64_stats_update_begin(&ring->syncp);
+ ring->stats.tx_l4_proto_err++;
+ u64_stats_update_end(&ring->syncp);
return ret;
+ }
ret = hns3_set_l2l3l4(skb, ol4_proto, il4_proto,
&type_cs_vlan_tso,
&ol_type_vlan_len_msec);
- if (unlikely(ret))
+ if (unlikely(ret)) {
+ u64_stats_update_begin(&ring->syncp);
+ ring->stats.tx_l2l3l4_err++;
+ u64_stats_update_end(&ring->syncp);
return ret;
+ }
ret = hns3_set_tso(skb, &paylen, &mss,
&type_cs_vlan_tso);
- if (unlikely(ret))
+ if (unlikely(ret)) {
+ u64_stats_update_begin(&ring->syncp);
+ ring->stats.tx_tso_err++;
+ u64_stats_update_end(&ring->syncp);
return ret;
+ }
}
/* Set txbd */
@@ -1107,7 +1128,9 @@ static int hns3_fill_desc(struct hns3_enet_ring *ring, void *priv,
}
if (unlikely(dma_mapping_error(dev, dma))) {
+ u64_stats_update_begin(&ring->syncp);
ring->stats.sw_err_cnt++;
+ u64_stats_update_end(&ring->syncp);
return -ENOMEM;
}
@@ -1330,9 +1353,7 @@ netdev_tx_t hns3_nic_net_xmit(struct sk_buff *skb, struct net_device *netdev)
u64_stats_update_end(&ring->syncp);
}
- if (net_ratelimit())
- netdev_err(netdev, "xmit error: %d!\n", buf_num);
-
+ hns3_rl_err(netdev, "xmit error: %d!\n", buf_num);
goto out_err_tx_ok;
}
@@ -1498,7 +1519,15 @@ static void hns3_nic_get_stats64(struct net_device *netdev,
tx_bytes += ring->stats.tx_bytes;
tx_pkts += ring->stats.tx_pkts;
tx_drop += ring->stats.sw_err_cnt;
+ tx_drop += ring->stats.tx_vlan_err;
+ tx_drop += ring->stats.tx_l4_proto_err;
+ tx_drop += ring->stats.tx_l2l3l4_err;
+ tx_drop += ring->stats.tx_tso_err;
tx_errors += ring->stats.sw_err_cnt;
+ tx_errors += ring->stats.tx_vlan_err;
+ tx_errors += ring->stats.tx_l4_proto_err;
+ tx_errors += ring->stats.tx_l2l3l4_err;
+ tx_errors += ring->stats.tx_tso_err;
} while (u64_stats_fetch_retry_irq(&ring->syncp, start));
/* fetch the rx stats */
@@ -2382,8 +2411,9 @@ static void hns3_nic_alloc_rx_buffers(struct hns3_enet_ring *ring,
ring->stats.sw_err_cnt++;
u64_stats_update_end(&ring->syncp);
- netdev_err(ring->tqp->handle->kinfo.netdev,
- "hnae reserve buffer map failed.\n");
+ hns3_rl_err(ring->tqp_vector->napi.dev,
+ "alloc rx buffer failed: %d\n",
+ ret);
break;
}
hns3_replace_buffer(ring, ring->next_to_use, &res_cbs);
@@ -2468,9 +2498,9 @@ static int hns3_gro_complete(struct sk_buff *skb, u32 l234info)
th->check = ~tcp_v6_check(skb->len - depth, &iph->saddr,
&iph->daddr, 0);
} else {
- netdev_err(skb->dev,
- "Error: FW GRO supports only IPv4/IPv6, not 0x%04x, depth: %d\n",
- be16_to_cpu(type), depth);
+ hns3_rl_err(skb->dev,
+ "Error: FW GRO supports only IPv4/IPv6, not 0x%04x, depth: %d\n",
+ be16_to_cpu(type), depth);
return -EFAULT;
}
@@ -2612,7 +2642,7 @@ static int hns3_alloc_skb(struct hns3_enet_ring *ring, unsigned int length,
ring->skb = napi_alloc_skb(&ring->tqp_vector->napi, HNS3_RX_HEAD_SIZE);
skb = ring->skb;
if (unlikely(!skb)) {
- netdev_err(netdev, "alloc rx skb fail\n");
+ hns3_rl_err(netdev, "alloc rx skb fail\n");
u64_stats_update_begin(&ring->syncp);
ring->stats.sw_err_cnt++;
@@ -2687,8 +2717,8 @@ static int hns3_add_frag(struct hns3_enet_ring *ring, struct hns3_desc *desc,
new_skb = napi_alloc_skb(&ring->tqp_vector->napi,
HNS3_RX_HEAD_SIZE);
if (unlikely(!new_skb)) {
- netdev_err(ring->tqp->handle->kinfo.netdev,
- "alloc rx skb frag fail\n");
+ hns3_rl_err(ring->tqp_vector->napi.dev,
+ "alloc rx fraglist skb fail\n");
return -ENXIO;
}
ring->frag_num = 0;
diff --git a/drivers/net/ethernet/hisilicon/hns3/hns3_enet.h b/drivers/net/ethernet/hisilicon/hns3/hns3_enet.h
index 0a970f5..a76712c 100644
--- a/drivers/net/ethernet/hisilicon/hns3/hns3_enet.h
+++ b/drivers/net/ethernet/hisilicon/hns3/hns3_enet.h
@@ -377,6 +377,10 @@ struct ring_stats {
u64 restart_queue;
u64 tx_busy;
u64 tx_copy;
+ u64 tx_vlan_err;
+ u64 tx_l4_proto_err;
+ u64 tx_l2l3l4_err;
+ u64 tx_tso_err;
};
struct {
u64 rx_pkts;
diff --git a/drivers/net/ethernet/hisilicon/hns3/hns3_ethtool.c b/drivers/net/ethernet/hisilicon/hns3/hns3_ethtool.c
index 02f46c7..185ff32 100644
--- a/drivers/net/ethernet/hisilicon/hns3/hns3_ethtool.c
+++ b/drivers/net/ethernet/hisilicon/hns3/hns3_ethtool.c
@@ -30,6 +30,10 @@ static const struct hns3_stats hns3_txq_stats[] = {
HNS3_TQP_STAT("wake", restart_queue),
HNS3_TQP_STAT("busy", tx_busy),
HNS3_TQP_STAT("copy", tx_copy),
+ HNS3_TQP_STAT("vlan_err", tx_vlan_err),
+ HNS3_TQP_STAT("l4_proto_err", tx_l4_proto_err),
+ HNS3_TQP_STAT("l2l3l4_err", tx_l2l3l4_err),
+ HNS3_TQP_STAT("tso_err", tx_tso_err),
};
#define HNS3_TXQ_STATS_COUNT ARRAY_SIZE(hns3_txq_stats)
--
2.7.4
^ permalink raw reply related
* [PATCH net-next 00/12] net: hns3: add some bugfixes & optimizations & cleanups for HNS3 driver
From: Huazhong Tan @ 2019-08-09 2:31 UTC (permalink / raw)
To: davem
Cc: netdev, linux-kernel, salil.mehta, yisen.zhuang, linuxarm,
Huazhong Tan
This patch-set includes code optimizations, bugfixes and cleanups for
the HNS3 ethernet controller driver.
[patch 01/12] fixes a GFP flag error.
[patch 02/12] fixes a VF interrupt error.
[patch 03/12] adds a cleanup for VLAN handling.
[patch 04/12] fixes a bug in debugfs.
[patch 05/12] modifies pause displaying format.
[patch 06/12] adds more DFX information for ethtool -d.
[patch 07/12] adds more TX statistics information.
[patch 08/12] adds a check for TX BD number.
[patch 09/12] adds a cleanup for dumping NCL_CONFIG.
[patch 10/12] refines function for querying MAC pause statistics.
[patch 11/12] adds a handshake with VF when doing PF reset.
[patch 12/12] refines some macro definitions.
Guangbin Huang (1):
net: hns3: add DFX registers information for ethtool -d
Guojia Liao (1):
net: hns3: refine some macro definitions
Huazhong Tan (2):
net: hns3: fix interrupt clearing error for VF
net: hns3: add handshake with VF for PF reset
Yonglong Liu (1):
net: hns3: modify how pause options is displayed
Yufeng Mo (3):
net: hns3: add input length check for debugfs write function
net: hns3: add function display NCL_CONFIG info
net: hns3: refine MAC pause statistics querying function
Yunsheng Lin (3):
net: hns3: clean up for vlan handling in hns3_fill_desc_vtags
net: hns3: add some statitics info to tx process
net: hns3: add check for max TX BD num for tso and non-tso case
Zhongzhu Liu (1):
net: hns3: fix GFP flag error in hclge_mac_update_stats()
drivers/net/ethernet/hisilicon/hns3/hnae3.h | 15 +-
drivers/net/ethernet/hisilicon/hns3/hns3_debugfs.c | 4 +
drivers/net/ethernet/hisilicon/hns3/hns3_enet.c | 268 ++++++------
drivers/net/ethernet/hisilicon/hns3/hns3_enet.h | 7 +-
drivers/net/ethernet/hisilicon/hns3/hns3_ethtool.c | 4 +
.../net/ethernet/hisilicon/hns3/hns3pf/hclge_cmd.h | 7 +
.../ethernet/hisilicon/hns3/hns3pf/hclge_debugfs.c | 64 +--
.../ethernet/hisilicon/hns3/hns3pf/hclge_main.c | 454 +++++++++++++++++----
.../ethernet/hisilicon/hns3/hns3pf/hclge_main.h | 2 +
.../ethernet/hisilicon/hns3/hns3vf/hclgevf_main.c | 28 +-
.../ethernet/hisilicon/hns3/hns3vf/hclgevf_main.h | 4 +-
11 files changed, 615 insertions(+), 242 deletions(-)
--
2.7.4
^ permalink raw reply
* [PATCH net v2] hv_netvsc: Fix a warning of suspicious RCU usage
From: Dexuan Cui @ 2019-08-09 1:58 UTC (permalink / raw)
To: netdev@vger.kernel.org, David S. Miller, Haiyang Zhang,
Stephen Hemminger, Jakub Kicinski
Cc: sashal@kernel.org, KY Srinivasan, Michael Kelley,
linux-hyperv@vger.kernel.org, linux-kernel@vger.kernel.org,
olaf@aepfle.de, apw@canonical.com, jasowang@redhat.com, vkuznets,
marcelo.cerri@canonical.com
This fixes a warning of "suspicious rcu_dereference_check() usage"
when nload runs.
Fixes: 776e726bfb34 ("netvsc: fix RCU warning in get_stats")
Signed-off-by: Dexuan Cui <decui@microsoft.com>
---
Changes in v2:
Made the minimal required change.
Added a Fixes tag.
Removed Stephen H.'s Signed-off-by since this is somewhat different from the
v1 from him; if there is any bug in v2, it's all my fault. :-)
drivers/net/hyperv/netvsc_drv.c | 9 +++++++--
1 file changed, 7 insertions(+), 2 deletions(-)
diff --git a/drivers/net/hyperv/netvsc_drv.c b/drivers/net/hyperv/netvsc_drv.c
index f9209594624b..b6357a75712c 100644
--- a/drivers/net/hyperv/netvsc_drv.c
+++ b/drivers/net/hyperv/netvsc_drv.c
@@ -1240,12 +1240,15 @@ static void netvsc_get_stats64(struct net_device *net,
struct rtnl_link_stats64 *t)
{
struct net_device_context *ndev_ctx = netdev_priv(net);
- struct netvsc_device *nvdev = rcu_dereference_rtnl(ndev_ctx->nvdev);
+ struct netvsc_device *nvdev;
struct netvsc_vf_pcpu_stats vf_tot;
int i;
+ rcu_read_lock();
+
+ nvdev = rcu_dereference(ndev_ctx->nvdev);
if (!nvdev)
- return;
+ goto out;
netdev_stats_to_stats64(t, &net->stats);
@@ -1284,6 +1287,8 @@ static void netvsc_get_stats64(struct net_device *net,
t->rx_packets += packets;
t->multicast += multicast;
}
+out:
+ rcu_read_unlock();
}
static int netvsc_set_mac_addr(struct net_device *ndev, void *p)
--
2.19.1
^ permalink raw reply related
* [PATCH v2 net-next] taprio: remove unused variable 'entry_list_policy'
From: YueHaibing @ 2019-08-09 1:49 UTC (permalink / raw)
To: davem, jhs, xiyou.wangcong, jiri, vinicius.gomes
Cc: linux-kernel, netdev, YueHaibing
In-Reply-To: <20190808142623.69188-1-yuehaibing@huawei.com>
net/sched/sch_taprio.c:680:32: warning:
entry_list_policy defined but not used [-Wunused-const-variable=]
One of the points of commit a3d43c0d56f1 ("taprio: Add support adding
an admin schedule") is that it removes support (it now returns "not
supported") for schedules using the TCA_TAPRIO_ATTR_SCHED_SINGLE_ENTRY
attribute (which were never used), the parsing of those types of schedules
was the only user of this policy. So removing this policy should be fine.
Reported-by: Hulk Robot <hulkci@huawei.com>
Suggested-by: Vinicius Costa Gomes <vinicius.gomes@intel.com>
Signed-off-by: YueHaibing <yuehaibing@huawei.com>
---
v2: respin commit log using Vinicius's explanation.
---
net/sched/sch_taprio.c | 4 ----
1 file changed, 4 deletions(-)
diff --git a/net/sched/sch_taprio.c b/net/sched/sch_taprio.c
index c39db50..046fd2c 100644
--- a/net/sched/sch_taprio.c
+++ b/net/sched/sch_taprio.c
@@ -677,10 +677,6 @@ static const struct nla_policy entry_policy[TCA_TAPRIO_SCHED_ENTRY_MAX + 1] = {
[TCA_TAPRIO_SCHED_ENTRY_INTERVAL] = { .type = NLA_U32 },
};
-static const struct nla_policy entry_list_policy[TCA_TAPRIO_SCHED_MAX + 1] = {
- [TCA_TAPRIO_SCHED_ENTRY] = { .type = NLA_NESTED },
-};
-
static const struct nla_policy taprio_policy[TCA_TAPRIO_ATTR_MAX + 1] = {
[TCA_TAPRIO_ATTR_PRIOMAP] = {
.len = sizeof(struct tc_mqprio_qopt)
--
2.7.4
^ permalink raw reply related
* Re: [PATCH net-next v7 0/6] flow_offload: add indr-block in nf_table_offload
From: David Miller @ 2019-08-09 1:44 UTC (permalink / raw)
To: wenxu; +Cc: jakub.kicinski, pablo, netfilter-devel, netdev
In-Reply-To: <1565140434-8109-1-git-send-email-wenxu@ucloud.cn>
From: wenxu@ucloud.cn
Date: Wed, 7 Aug 2019 09:13:48 +0800
> This series patch make nftables offload support the vlan and
> tunnel device offload through indr-block architecture.
>
> The first four patches mv tc indr block to flow offload and
> rename to flow-indr-block.
> Because the new flow-indr-block can't get the tcf_block
> directly. The fifth patch provide a callback list to get
> flow_block of each subsystem immediately when the device
> register and contain a block.
> The last patch make nf_tables_offload support flow-indr-block.
>
> This version add a mutex lock for add/del flow_indr_block_ing_cb
Series applied, thank you.
^ permalink raw reply
* Re: [PATCH 00/17] Networking driver debugfs cleanups
From: David Miller @ 2019-08-09 1:42 UTC (permalink / raw)
To: gregkh; +Cc: netdev
In-Reply-To: <20190808.183756.2198405327467483431.davem@davemloft.net>
From: David Miller <davem@davemloft.net>
Date: Thu, 08 Aug 2019 18:37:56 -0700 (PDT)
> I applied this without patch #17 which you said you would respin in order
> to get rid of the now unused local variable.
Actually, there is a bunch of fallout still:
drivers/net/wimax/i2400m/debugfs.c: In function ‘i2400m_debugfs_add’:
drivers/net/wimax/i2400m/debugfs.c:192:17: warning: unused variable ‘dev’ [-Wunused-variable]
struct device *dev = i2400m_dev(i2400m);
^~~
drivers/net/wimax/i2400m/usb.c: In function ‘i2400mu_debugfs_add’:
drivers/net/wimax/i2400m/usb.c:375:17: warning: unused variable ‘fd’ [-Wunused-variable]
struct dentry *fd;
^~
drivers/net/wimax/i2400m/usb.c:373:17: warning: unused variable ‘dev’ [-Wunused-variable]
struct device *dev = &i2400mu->usb_iface->dev;
^~~
drivers/net/wimax/i2400m/usb.c:372:6: warning: unused variable ‘result’ [-Wunused-variable]
int result;
^~~~~~
drivers/net/ethernet/intel/i40e/i40e_debugfs.c: In function ‘i40e_dbg_pf_init’:
drivers/net/ethernet/intel/i40e/i40e_debugfs.c:1736:23: warning: unused variable ‘dev’ [-Wunused-variable]
const struct device *dev = &pf->pdev->dev;
^~~
This is with:
[davem@localhost net-next]$ gcc --version
gcc (GCC) 8.3.1 20190223 (Red Hat 8.3.1-2)
Copyright (C) 2018 Free Software Foundation, Inc.
This is free software; see the source for copying conditions. There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
[davem@localhost net-next]$
So I'm reverting.
Please respin the series with this stuff fixed, thanks Greg.
^ permalink raw reply
* Re: [PATCH 00/17] Networking driver debugfs cleanups
From: David Miller @ 2019-08-09 1:37 UTC (permalink / raw)
To: gregkh; +Cc: netdev
In-Reply-To: <20190806161128.31232-1-gregkh@linuxfoundation.org>
From: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Date: Tue, 6 Aug 2019 18:11:11 +0200
> There is no need to test the result of any debugfs call anymore. The
> debugfs core warns the user if something fails, and the return value of
> a debugfs call can always be fed back into another debugfs call with no
> problems.
>
> Also, debugfs is for debugging, so if there are problems with debugfs
> (i.e. the system is out of memory) the rest of the kernel should not
> change behavior, so testing for debugfs calls is pointless and not the
> goal of debugfs at all.
>
> This series cleans up a lot of networking drivers and some wimax code
> that was calling debugfs and trying to do something with the return
> value that it didn't need to. Removing this logic makes the code
> smaller, easier to understand, and use less run-time memory in some
> cases, all good things.
>
> The series is against net-next, and have no dependancies between any of
> them if they want to go through any random tree/order. Or, if wanted,
> I can take them through my driver-core tree where other debugfs cleanups
> are being slowly fed during major merge windows.
I applied this without patch #17 which you said you would respin in order
to get rid of the now unused local variable.
Thanks.
^ permalink raw reply
* Re: [PATCH net-next] taprio: remove unused variable 'entry_list_policy'
From: Yuehaibing @ 2019-08-09 1:33 UTC (permalink / raw)
To: Vinicius Costa Gomes, David Miller
Cc: jhs, xiyou.wangcong, jiri, linux-kernel, netdev
In-Reply-To: <87mugjtmn7.fsf@intel.com>
On 2019/8/9 4:42, Vinicius Costa Gomes wrote:
> Hi,
>
> David Miller <davem@davemloft.net> writes:
>
>> From: YueHaibing <yuehaibing@huawei.com>
>> Date: Thu, 8 Aug 2019 22:26:23 +0800
>>
>>> net/sched/sch_taprio.c:680:32: warning:
>>> entry_list_policy defined but not used [-Wunused-const-variable=]
>>>
>>> It is not used since commit a3d43c0d56f1 ("taprio: Add
>>> support adding an admin schedule")
>>>
>>> Reported-by: Hulk Robot <hulkci@huawei.com>
>>> Signed-off-by: YueHaibing <yuehaibing@huawei.com>
>>
>> This is probably unintentional and a bug, we should be using that
>> policy value to validate that the sched list is indeed a nested
>> attribute.
>
> Removing this policy should be fine.
>
> One of the points of commit (as explained in the commit message)
> a3d43c0d56f1 ("taprio: Add support adding an admin schedule") is that it
> removes support (it now returns "not supported") for schedules using the
> TCA_TAPRIO_ATTR_SCHED_SINGLE_ENTRY attribute (which were never used),
> the parsing of those types of schedules was the only user of this
> policy.
>
>>
>> I'm not applying this without at least a better and clear commit
>> message explaining why we shouldn't be using this policy any more.
>
> YueHaibing may use the text above in the commit message of a new spin of
> this patch if you think it's clear enough.
Thanks, will send v2 with your explanation.
>
>
> Cheers,
> --
> Vinicius
>
> .
>
^ permalink raw reply
* Re: [PATCH v3 38/41] powerpc: convert put_page() to put_user_page*()
From: John Hubbard @ 2019-08-09 1:26 UTC (permalink / raw)
To: Michael Ellerman, Andrew Morton
Cc: Christoph Hellwig, Dan Williams, Dave Chinner, Dave Hansen,
Ira Weiny, Jan Kara, Jason Gunthorpe, Jérôme Glisse,
LKML, amd-gfx, ceph-devel, devel, devel, dri-devel, intel-gfx,
kvm, linux-arm-kernel, linux-block, linux-crypto, linux-fbdev,
linux-fsdevel, linux-media, linux-mm, linux-nfs, linux-rdma,
linux-rpi-kernel, linux-xfs, netdev, rds-devel, sparclinux, x86,
xen-devel, Benjamin Herrenschmidt, Christoph Hellwig,
linuxppc-dev
In-Reply-To: <87k1botdpx.fsf@concordia.ellerman.id.au>
On 8/7/19 10:42 PM, Michael Ellerman wrote:
> Hi John,
>
> john.hubbard@gmail.com writes:
>> diff --git a/arch/powerpc/mm/book3s64/iommu_api.c b/arch/powerpc/mm/book3s64/iommu_api.c
>> index b056cae3388b..e126193ba295 100644
>> --- a/arch/powerpc/mm/book3s64/iommu_api.c
>> +++ b/arch/powerpc/mm/book3s64/iommu_api.c
>> @@ -203,6 +202,7 @@ static void mm_iommu_unpin(struct mm_iommu_table_group_mem_t *mem)
>> {
>> long i;
>> struct page *page = NULL;
>> + bool dirty = false;
>
> I don't think you need that initialisation do you?
>
Nope, it can go. Fixed locally, thanks.
Did you get a chance to look at enough of the other bits to feel comfortable
with the patch, overall?
thanks,
--
John Hubbard
NVIDIA
^ permalink raw reply
* Re: [PATCH v3 net-next 0/3] net: batched receive in GRO path
From: David Miller @ 2019-08-09 1:22 UTC (permalink / raw)
To: ecree; +Cc: netdev, eric.dumazet, linux-net-drivers
In-Reply-To: <c6e2474e-2c8a-5881-86bf-59c66bdfc34f@solarflare.com>
From: Edward Cree <ecree@solarflare.com>
Date: Tue, 6 Aug 2019 14:52:06 +0100
> This series listifies part of GRO processing, in a manner which allows those
> packets which are not GROed (i.e. for which dev_gro_receive returns
> GRO_NORMAL) to be passed on to the listified regular receive path.
> dev_gro_receive() itself is not listified, nor the per-protocol GRO
> callback, since GRO's need to hold packets on lists under napi->gro_hash
> makes keeping the packets on other lists awkward, and since the GRO control
> block state of held skbs can refer only to one 'new' skb at a time.
> Instead, when napi_frags_finish() handles a GRO_NORMAL result, stash the skb
> onto a list in the napi struct, which is received at the end of the napi
> poll or when its length exceeds the (new) sysctl net.core.gro_normal_batch.
>
> Performance figures with this series, collected on a back-to-back pair of
> Solarflare sfn8522-r2 NICs with 120-second NetPerf tests. In the stats,
> sample size n for old and new code is 6 runs each; p is from a Welch t-test.
> Tests were run both with GRO enabled and disabled, the latter simulating
> uncoalesceable packets (e.g. due to IP or TCP options). The receive side
> (which was the device under test) had the NetPerf process pinned to one CPU,
> and the device interrupts pinned to a second CPU. CPU utilisation figures
> (used in cases of line-rate performance) are summed across all CPUs.
> net.core.gro_normal_batch was left at its default value of 8.
...
> The above results are fairly mixed, and in most cases not statistically
> significant. But I think we can roughly conclude that the series
> marginally improves non-GROable throughput, without hurting latency
> (except in the large-payload busy-polling case, which in any case yields
> horrid performance even on net-next (almost triple the latency without
> busy-poll). Also, drivers which, unlike sfc, pass UDP traffic to GRO
> would expect to see a benefit from gaining access to batching.
>
> Changed in v3:
> * gro_normal_batch sysctl now uses SYSCTL_ONE instead of &one
> * removed RFC tags (no comments after a week means no-one objects, right?)
>
> Changed in v2:
> * During busy poll, call gro_normal_list() to receive batched packets
> after each cycle of the napi busy loop. See comments in Patch #3 for
> complications of doing the same in busy_poll_stop().
>
> [1]: Cohen 1959, doi: 10.1080/00401706.1959.10489859
Series applied, thanks Edward.
^ permalink raw reply
* RE: [PATCH net] hv_netvsc: Fix a warning of suspicious RCU usage
From: Dexuan Cui @ 2019-08-09 1:16 UTC (permalink / raw)
To: David Miller
Cc: netdev@vger.kernel.org, Haiyang Zhang, Stephen Hemminger,
sashal@kernel.org, KY Srinivasan, Michael Kelley,
linux-hyperv@vger.kernel.org, linux-kernel@vger.kernel.org,
olaf@aepfle.de, apw@canonical.com, jasowang@redhat.com, vkuznets,
marcelo.cerri@canonical.com
In-Reply-To: <20190808.181350.1331633709956960086.davem@davemloft.net>
> From: David Miller <davem@davemloft.net>
> Sent: Thursday, August 8, 2019 6:14 PM
> To: Dexuan Cui <decui@microsoft.com>
> Cc: netdev@vger.kernel.org; Haiyang Zhang <haiyangz@microsoft.com>;
> Stephen Hemminger <sthemmin@microsoft.com>; sashal@kernel.org; KY
> Srinivasan <kys@microsoft.com>; Michael Kelley <mikelley@microsoft.com>;
> linux-hyperv@vger.kernel.org; linux-kernel@vger.kernel.org; olaf@aepfle.de;
> apw@canonical.com; jasowang@redhat.com; vkuznets
> <vkuznets@redhat.com>; marcelo.cerri@canonical.com
> Subject: Re: [PATCH net] hv_netvsc: Fix a warning of suspicious RCU usage
>
> From: Dexuan Cui <decui@microsoft.com>
> Date: Tue, 6 Aug 2019 05:17:44 +0000
>
> >
> > This fixes a warning of "suspicious rcu_dereference_check() usage"
> > when nload runs.
> >
> > Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
> > Signed-off-by: Dexuan Cui <decui@microsoft.com>
>
> Please resend with appropriate fixes tag.
Will do shortly.
Thanks,
-- Dexuan
^ permalink raw reply
* Re: [net 01/12] net/mlx5e: Use flow keys dissector to parse packets for ARFS
From: Jakub Kicinski @ 2019-08-09 1:15 UTC (permalink / raw)
To: Saeed Mahameed
Cc: David S. Miller, netdev@vger.kernel.org, Maxim Mikityanskiy,
Tariq Toukan
In-Reply-To: <20190808202025.11303-2-saeedm@mellanox.com>
On Thu, 8 Aug 2019 20:22:00 +0000, Saeed Mahameed wrote:
> From: Maxim Mikityanskiy <maximmi@mellanox.com>
>
> The current ARFS code relies on certain fields to be set in the SKB
> (e.g. transport_header) and extracts IP addresses and ports by custom
> code that parses the packet. The necessary SKB fields, however, are not
> always set at that point, which leads to an out-of-bounds access. Use
> skb_flow_dissect_flow_keys() to get the necessary information reliably,
> fix the out-of-bounds access and reuse the code.
The whole series LGTM, FWIW.
I'd be curious to hear which path does not have the skb fully
set up, could you elaborate? (I'm certainly no aRFC expert this
is pure curiosity).
^ permalink raw reply
* Re: [PATCH net-next v2 1/1] qed: Add new ethtool supported port types based on media.
From: David Miller @ 2019-08-09 1:14 UTC (permalink / raw)
To: rahulv; +Cc: netdev, aelior, mkalderon
In-Reply-To: <20190806065950.19073-1-rahulv@marvell.com>
From: Rahul Verma <rahulv@marvell.com>
Date: Mon, 5 Aug 2019 23:59:50 -0700
> Supported ports in ethtool <eth1> are displayed based on media type.
> For media type fibre and twinaxial, port type is "FIBRE". Media type
> Base-T is "TP" and media KR is "Backplane".
>
> V1->V2:
> Corrected the subject.
>
> Signed-off-by: Rahul Verma <rahulv@marvell.com>
> Signed-off-by: Michal Kalderon <michal.kalderon@marvell.com>
Applied.
^ permalink raw reply
page: next (older) | prev (newer) | latest
- recent:[subjects (threaded)|topics (new)|topics (active)]
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox