* [PATCH net-next 4/9] net/mlx4_core: Enhance the catas flow to support device reset
From: Or Gerlitz @ 2015-01-21 14:45 UTC (permalink / raw)
To: David S. Miller
Cc: netdev, Matan Barak, Amir Vadai, Tal Alon, Roland Dreier,
Yishai Hadas, Or Gerlitz
In-Reply-To: <1421851557-12084-1-git-send-email-ogerlitz@mellanox.com>
From: Yishai Hadas <yishaih@mellanox.com>
This includes:
- resetting the chip when a fatal error is detected (the current code
does not do this).
- exposing the ability to enter error state from outside the catas code
by calling its functionality. (E.g. FW Command timeout, AER error).
- managing a persistent device state. This is needed to sync between
reset flow cases.
Signed-off-by: Yishai Hadas <yishaih@mellanox.com>
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
---
drivers/net/ethernet/mellanox/mlx4/catas.c | 122 +++++++++++++++++++++-------
drivers/net/ethernet/mellanox/mlx4/main.c | 6 ++
drivers/net/ethernet/mellanox/mlx4/mlx4.h | 2 +-
include/linux/mlx4/device.h | 7 ++
4 files changed, 108 insertions(+), 29 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx4/catas.c b/drivers/net/ethernet/mellanox/mlx4/catas.c
index 5bb9aa6..588d6b5 100644
--- a/drivers/net/ethernet/mellanox/mlx4/catas.c
+++ b/drivers/net/ethernet/mellanox/mlx4/catas.c
@@ -48,6 +48,83 @@ MODULE_PARM_DESC(internal_err_reset,
"Reset device on internal errors if non-zero"
" (default 1, in SRIOV mode default is 0)");
+static int read_vendor_id(struct mlx4_dev *dev)
+{
+ u16 vendor_id = 0;
+ int ret;
+
+ ret = pci_read_config_word(dev->persist->pdev, 0, &vendor_id);
+ if (ret) {
+ mlx4_err(dev, "Failed to read vendor ID, ret=%d\n", ret);
+ return ret;
+ }
+
+ if (vendor_id == 0xffff) {
+ mlx4_err(dev, "PCI can't be accessed to read vendor id\n");
+ return -EINVAL;
+ }
+
+ return 0;
+}
+
+static int mlx4_reset_master(struct mlx4_dev *dev)
+{
+ int err = 0;
+
+ if (!pci_channel_offline(dev->persist->pdev)) {
+ err = read_vendor_id(dev);
+ /* If PCI can't be accessed to read vendor ID we assume that its
+ * link was disabled and chip was already reset.
+ */
+ if (err)
+ return 0;
+
+ err = mlx4_reset(dev);
+ if (err)
+ mlx4_err(dev, "Fail to reset HCA\n");
+ }
+
+ return err;
+}
+
+void mlx4_enter_error_state(struct mlx4_dev_persistent *persist)
+{
+ int err;
+ struct mlx4_dev *dev;
+
+ if (!internal_err_reset)
+ return;
+
+ mutex_lock(&persist->device_state_mutex);
+ if (persist->state & MLX4_DEVICE_STATE_INTERNAL_ERROR)
+ goto out;
+
+ dev = persist->dev;
+ mlx4_err(dev, "device is going to be reset\n");
+ err = mlx4_reset_master(dev);
+ BUG_ON(err != 0);
+
+ dev->persist->state |= MLX4_DEVICE_STATE_INTERNAL_ERROR;
+ mlx4_err(dev, "device was reset successfully\n");
+ mutex_unlock(&persist->device_state_mutex);
+
+ /* At that step HW was already reset, now notify clients */
+ mlx4_dispatch_event(dev, MLX4_DEV_EVENT_CATASTROPHIC_ERROR, 0);
+ return;
+
+out:
+ mutex_unlock(&persist->device_state_mutex);
+}
+
+static void mlx4_handle_error_state(struct mlx4_dev_persistent *persist)
+{
+ int err = 0;
+
+ mlx4_enter_error_state(persist);
+ err = mlx4_restart_one(persist->pdev);
+ mlx4_info(persist->dev, "mlx4_restart_one was ended, ret=%d\n", err);
+}
+
static void dump_err_buf(struct mlx4_dev *dev)
{
struct mlx4_priv *priv = mlx4_priv(dev);
@@ -66,21 +143,22 @@ static void poll_catas(unsigned long dev_ptr)
struct mlx4_priv *priv = mlx4_priv(dev);
if (readl(priv->catas_err.map)) {
- /* If the device is off-line, we cannot try to recover it */
- if (pci_channel_offline(dev->persist->pdev))
- mod_timer(&priv->catas_err.timer,
- round_jiffies(jiffies + MLX4_CATAS_POLL_INTERVAL));
- else {
- dump_err_buf(dev);
- mlx4_dispatch_event(dev, MLX4_DEV_EVENT_CATASTROPHIC_ERROR, 0);
-
- if (internal_err_reset)
- queue_work(dev->persist->catas_wq,
- &dev->persist->catas_work);
- }
- } else
- mod_timer(&priv->catas_err.timer,
- round_jiffies(jiffies + MLX4_CATAS_POLL_INTERVAL));
+ dump_err_buf(dev);
+ goto internal_err;
+ }
+
+ if (dev->persist->state & MLX4_DEVICE_STATE_INTERNAL_ERROR) {
+ mlx4_warn(dev, "Internal error mark was detected on device\n");
+ goto internal_err;
+ }
+
+ mod_timer(&priv->catas_err.timer,
+ round_jiffies(jiffies + MLX4_CATAS_POLL_INTERVAL));
+ return;
+
+internal_err:
+ if (internal_err_reset)
+ queue_work(dev->persist->catas_wq, &dev->persist->catas_work);
}
static void catas_reset(struct work_struct *work)
@@ -88,20 +166,8 @@ static void catas_reset(struct work_struct *work)
struct mlx4_dev_persistent *persist =
container_of(work, struct mlx4_dev_persistent,
catas_work);
- struct pci_dev *pdev = persist->pdev;
- int ret;
-
- /* If the device is off-line, we cannot reset it */
- if (pci_channel_offline(pdev))
- return;
- ret = mlx4_restart_one(pdev);
- /* 'priv' now is not valid */
- if (ret)
- pr_err("mlx4 %s: Reset failed (%d)\n",
- pci_name(pdev), ret);
- else
- mlx4_dbg(persist->dev, "Reset succeeded\n");
+ mlx4_handle_error_state(persist);
}
void mlx4_start_catas_poll(struct mlx4_dev *dev)
diff --git a/drivers/net/ethernet/mellanox/mlx4/main.c b/drivers/net/ethernet/mellanox/mlx4/main.c
index a61694c..dc2d910 100644
--- a/drivers/net/ethernet/mellanox/mlx4/main.c
+++ b/drivers/net/ethernet/mellanox/mlx4/main.c
@@ -2624,6 +2624,11 @@ static int mlx4_load_one(struct pci_dev *pdev, int pci_dev_data,
}
}
+ /* on load remove any previous indication of internal error,
+ * device is up.
+ */
+ dev->persist->state = MLX4_DEVICE_STATE_UP;
+
slave_start:
err = mlx4_cmd_init(dev);
if (err) {
@@ -3108,6 +3113,7 @@ static int mlx4_init_one(struct pci_dev *pdev, const struct pci_device_id *id)
dev->persist->dev = dev;
pci_set_drvdata(pdev, dev->persist);
priv->pci_dev_data = id->driver_data;
+ mutex_init(&dev->persist->device_state_mutex);
ret = __mlx4_init_one(pdev, id->driver_data, priv);
if (ret) {
diff --git a/drivers/net/ethernet/mellanox/mlx4/mlx4.h b/drivers/net/ethernet/mellanox/mlx4/mlx4.h
index d41af84..aa1ecbc 100644
--- a/drivers/net/ethernet/mellanox/mlx4/mlx4.h
+++ b/drivers/net/ethernet/mellanox/mlx4/mlx4.h
@@ -1178,7 +1178,7 @@ void mlx4_qp_event(struct mlx4_dev *dev, u32 qpn, int event_type);
void mlx4_srq_event(struct mlx4_dev *dev, u32 srqn, int event_type);
-void mlx4_handle_catas_err(struct mlx4_dev *dev);
+void mlx4_enter_error_state(struct mlx4_dev_persistent *persist);
int mlx4_SENSE_PORT(struct mlx4_dev *dev, int port,
enum mlx4_port_type *type);
diff --git a/include/linux/mlx4/device.h b/include/linux/mlx4/device.h
index da425d2..7d5d317 100644
--- a/include/linux/mlx4/device.h
+++ b/include/linux/mlx4/device.h
@@ -411,6 +411,11 @@ enum {
MLX4_EQ_PORT_INFO_MSTR_SM_SL_CHANGE_MASK = 1 << 4,
};
+enum {
+ MLX4_DEVICE_STATE_UP = 1 << 0,
+ MLX4_DEVICE_STATE_INTERNAL_ERROR = 1 << 1,
+};
+
#define MSTR_SM_CHANGE_MASK (MLX4_EQ_PORT_INFO_MSTR_SM_SL_CHANGE_MASK | \
MLX4_EQ_PORT_INFO_MSTR_SM_LID_CHANGE_MASK)
@@ -753,6 +758,8 @@ struct mlx4_dev_persistent {
enum mlx4_port_type curr_port_poss_type[MLX4_MAX_PORTS + 1];
struct work_struct catas_work;
struct workqueue_struct *catas_wq;
+ struct mutex device_state_mutex; /* protect HW state */
+ u8 state;
};
struct mlx4_dev {
--
1.7.1
^ permalink raw reply related
* [PATCH net-next 0/9] mlx4: Fix and enhance the device reset flow
From: Or Gerlitz @ 2015-01-21 14:45 UTC (permalink / raw)
To: David S. Miller
Cc: netdev, Matan Barak, Amir Vadai, Tal Alon, Roland Dreier,
Or Gerlitz
Hi Dave,
This series from Yishai Hadas fixes the device reset flow and adds SRIOV support.
Reset flows are required whenever a device experiences errors, is unresponsive,
or is not in a deterministic state. In such cases, the driver is expected to
reset the HW and continue operation. When SRIOV is enabled, these requirements
apply both to PF and VF devices.
Currently, the mlx4 reset flow doesn't work properly: when a fatal error is
detected on the FW internal buffer the chip is not reset and stays in its
bad state. There are cases that assumed to be fatal such as non-responsive FW,
errors via closing commands but are not handled today.
The AER mechanism should also be fixed:
- It should use mlx4_load_one instead of __mlx4_init_one which is done
upon HCA probing.
- It must be aligned with concurrent catas flow, mark device to be in
an error state, reset chip, etc.
- Port types should be restored to their original values before error occurred.
In addition, there the SRIOV use-case isn't supported.
In above cases when the device state becomes fatal we must act as follows:
1) Reset the chip and mark the HW device state as in fatal error.
2) Wake up any pending commands, preventing new ones to come in.
3) Restart the software stack.
We also address the SRIOV mode as follows: In case the PF detects a fatal error,
it lets VFs know about that, then both itself and VFs are restarted asynchronously.
However, in case only the VF encountered a fatal case or forced to be reset, they
reset the VF stuff and then restart software.
Yishai, Matan and Or.
Yishai Hadas (9):
net/mlx4_core: Maintain a persistent memory for mlx4 device
net/mlx4_core: Set device configuration data to be persistent across reset
net/mlx4_core: Refactor the catas flow to work per device
net/mlx4_core: Enhance the catas flow to support device reset
net/mlx4_core: Activate reset flow upon fatal command cases
net/mlx4_core: Manage interface state for Reset flow cases
net/mlx4_core: Handle AER flow properly
net/mlx4_core: Enable device recovery flow with SRIOV
net/mlx4_core: Reset flow activation upon SRIOV fatal command cases
drivers/infiniband/hw/mlx4/alias_GUID.c | 2 +-
drivers/infiniband/hw/mlx4/mad.c | 3 +-
drivers/infiniband/hw/mlx4/main.c | 17 +-
drivers/infiniband/hw/mlx4/mr.c | 6 +-
drivers/infiniband/hw/mlx4/sysfs.c | 6 +-
drivers/net/ethernet/mellanox/mlx4/alloc.c | 15 +-
drivers/net/ethernet/mellanox/mlx4/catas.c | 294 +++++++++++----
drivers/net/ethernet/mellanox/mlx4/cmd.c | 405 +++++++++++++++-----
drivers/net/ethernet/mellanox/mlx4/en_cq.c | 4 +-
drivers/net/ethernet/mellanox/mlx4/en_ethtool.c | 2 +-
drivers/net/ethernet/mellanox/mlx4/en_main.c | 4 +-
drivers/net/ethernet/mellanox/mlx4/en_netdev.c | 2 +-
drivers/net/ethernet/mellanox/mlx4/en_rx.c | 4 +-
drivers/net/ethernet/mellanox/mlx4/en_tx.c | 4 +-
drivers/net/ethernet/mellanox/mlx4/eq.c | 52 ++-
drivers/net/ethernet/mellanox/mlx4/icm.c | 11 +-
drivers/net/ethernet/mellanox/mlx4/intf.c | 8 +-
drivers/net/ethernet/mellanox/mlx4/main.c | 387 +++++++++++++++----
drivers/net/ethernet/mellanox/mlx4/mcg.c | 6 +
drivers/net/ethernet/mellanox/mlx4/mlx4.h | 27 +-
drivers/net/ethernet/mellanox/mlx4/mr.c | 8 +-
drivers/net/ethernet/mellanox/mlx4/pd.c | 6 +-
drivers/net/ethernet/mellanox/mlx4/port.c | 17 +-
drivers/net/ethernet/mellanox/mlx4/reset.c | 23 +-
.../net/ethernet/mellanox/mlx4/resource_tracker.c | 36 ++-
include/linux/mlx4/cmd.h | 3 +
include/linux/mlx4/device.h | 34 ++-
27 files changed, 1042 insertions(+), 344 deletions(-)
^ permalink raw reply
* [PATCH net-next 8/9] net/mlx4_core: Enable device recovery flow with SRIOV
From: Or Gerlitz @ 2015-01-21 14:45 UTC (permalink / raw)
To: David S. Miller
Cc: netdev, Matan Barak, Amir Vadai, Tal Alon, Roland Dreier,
Yishai Hadas, Or Gerlitz
In-Reply-To: <1421851557-12084-1-git-send-email-ogerlitz@mellanox.com>
From: Yishai Hadas <yishaih@mellanox.com>
In SRIOV, both the PF and the VF may attempt device recovery whenever they
assume that the device is not functioning. When the PF driver resets the
device, the VF should detect this and attempt to reinitialize itself.
The VF must be able to reset itself under all circumstances, even
if the PF is not responsive.
The VF shall reset itself in the following cases:
1. Commands are not processed within reasonable time over the communication channel.
This is done considering device state and the correct return code based on
the command as was done in the native mode, done in the next patch.
2. The VF driver receives an internal error event reported by the PF on the
communication channel. This occurs when the PF driver resets the device or
when VF is out of sync with the PF.
Add 'VF reset' capability, which allows the VF to reinitialize itself even when the
PF is not responsive.
As PF and VF may run their reset flow simulantanisly, there are several cases
that are handled:
- Prevent freeing VF resources upon FLR, when PF is in its unloading stage.
- Prevent PF getting VF commands before it has finished initializing its resources.
- Upon VF startup, check that comm-channel is online before sending
commands to the PF and getting timed-out.
Signed-off-by: Yishai Hadas <yishaih@mellanox.com>
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
---
drivers/net/ethernet/mellanox/mlx4/catas.c | 120 +++++++++++++++++++++----
drivers/net/ethernet/mellanox/mlx4/cmd.c | 75 ++++++++++++----
drivers/net/ethernet/mellanox/mlx4/eq.c | 10 ++-
drivers/net/ethernet/mellanox/mlx4/intf.c | 6 +-
drivers/net/ethernet/mellanox/mlx4/main.c | 135 ++++++++++++++++++++++------
drivers/net/ethernet/mellanox/mlx4/mlx4.h | 7 ++-
include/linux/mlx4/cmd.h | 2 +
include/linux/mlx4/device.h | 5 +
8 files changed, 291 insertions(+), 69 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx4/catas.c b/drivers/net/ethernet/mellanox/mlx4/catas.c
index 3fcf3cf..715de8a 100644
--- a/drivers/net/ethernet/mellanox/mlx4/catas.c
+++ b/drivers/net/ethernet/mellanox/mlx4/catas.c
@@ -45,8 +45,7 @@ enum {
int mlx4_internal_err_reset = 1;
module_param_named(internal_err_reset, mlx4_internal_err_reset, int, 0644);
MODULE_PARM_DESC(internal_err_reset,
- "Reset device on internal errors if non-zero"
- " (default 1, in SRIOV mode default is 0)");
+ "Reset device on internal errors if non-zero (default 1)");
static int read_vendor_id(struct mlx4_dev *dev)
{
@@ -71,6 +70,9 @@ static int mlx4_reset_master(struct mlx4_dev *dev)
{
int err = 0;
+ if (mlx4_is_master(dev))
+ mlx4_report_internal_err_comm_event(dev);
+
if (!pci_channel_offline(dev->persist->pdev)) {
err = read_vendor_id(dev);
/* If PCI can't be accessed to read vendor ID we assume that its
@@ -87,6 +89,81 @@ static int mlx4_reset_master(struct mlx4_dev *dev)
return err;
}
+static int mlx4_reset_slave(struct mlx4_dev *dev)
+{
+#define COM_CHAN_RST_REQ_OFFSET 0x10
+#define COM_CHAN_RST_ACK_OFFSET 0x08
+
+ u32 comm_flags;
+ u32 rst_req;
+ u32 rst_ack;
+ unsigned long end;
+ struct mlx4_priv *priv = mlx4_priv(dev);
+
+ if (pci_channel_offline(dev->persist->pdev))
+ return 0;
+
+ comm_flags = swab32(readl((__iomem char *)priv->mfunc.comm +
+ MLX4_COMM_CHAN_FLAGS));
+ if (comm_flags == 0xffffffff) {
+ mlx4_err(dev, "VF reset is not needed\n");
+ return 0;
+ }
+
+ if (!(dev->caps.vf_caps & MLX4_VF_CAP_FLAG_RESET)) {
+ mlx4_err(dev, "VF reset is not supported\n");
+ return -EOPNOTSUPP;
+ }
+
+ rst_req = (comm_flags & (u32)(1 << COM_CHAN_RST_REQ_OFFSET)) >>
+ COM_CHAN_RST_REQ_OFFSET;
+ rst_ack = (comm_flags & (u32)(1 << COM_CHAN_RST_ACK_OFFSET)) >>
+ COM_CHAN_RST_ACK_OFFSET;
+ if (rst_req != rst_ack) {
+ mlx4_err(dev, "Communication channel isn't sync, fail to send reset\n");
+ return -EIO;
+ }
+
+ rst_req ^= 1;
+ mlx4_warn(dev, "VF is sending reset request to Firmware\n");
+ comm_flags = rst_req << COM_CHAN_RST_REQ_OFFSET;
+ __raw_writel((__force u32)cpu_to_be32(comm_flags),
+ (__iomem char *)priv->mfunc.comm + MLX4_COMM_CHAN_FLAGS);
+ /* Make sure that our comm channel write doesn't
+ * get mixed in with writes from another CPU.
+ */
+ mmiowb();
+
+ end = msecs_to_jiffies(MLX4_COMM_TIME) + jiffies;
+ while (time_before(jiffies, end)) {
+ comm_flags = swab32(readl((__iomem char *)priv->mfunc.comm +
+ MLX4_COMM_CHAN_FLAGS));
+ rst_ack = (comm_flags & (u32)(1 << COM_CHAN_RST_ACK_OFFSET)) >>
+ COM_CHAN_RST_ACK_OFFSET;
+
+ /* Reading rst_req again since the communication channel can
+ * be reset at any time by the PF and all its bits will be
+ * set to zero.
+ */
+ rst_req = (comm_flags & (u32)(1 << COM_CHAN_RST_REQ_OFFSET)) >>
+ COM_CHAN_RST_REQ_OFFSET;
+
+ if (rst_ack == rst_req) {
+ mlx4_warn(dev, "VF Reset succeed\n");
+ return 0;
+ }
+ cond_resched();
+ }
+ mlx4_err(dev, "Fail to send reset over the communication channel\n");
+ return -ETIMEDOUT;
+}
+
+static int mlx4_comm_internal_err(u32 slave_read)
+{
+ return (u32)COMM_CHAN_EVENT_INTERNAL_ERR ==
+ (slave_read & (u32)COMM_CHAN_EVENT_INTERNAL_ERR) ? 1 : 0;
+}
+
void mlx4_enter_error_state(struct mlx4_dev_persistent *persist)
{
int err;
@@ -101,7 +178,10 @@ void mlx4_enter_error_state(struct mlx4_dev_persistent *persist)
dev = persist->dev;
mlx4_err(dev, "device is going to be reset\n");
- err = mlx4_reset_master(dev);
+ if (mlx4_is_slave(dev))
+ err = mlx4_reset_slave(dev);
+ else
+ err = mlx4_reset_master(dev);
BUG_ON(err != 0);
dev->persist->state |= MLX4_DEVICE_STATE_INTERNAL_ERROR;
@@ -148,8 +228,15 @@ static void poll_catas(unsigned long dev_ptr)
{
struct mlx4_dev *dev = (struct mlx4_dev *) dev_ptr;
struct mlx4_priv *priv = mlx4_priv(dev);
-
- if (readl(priv->catas_err.map)) {
+ u32 slave_read;
+
+ if (mlx4_is_slave(dev)) {
+ slave_read = swab32(readl(&priv->mfunc.comm->slave_read));
+ if (mlx4_comm_internal_err(slave_read)) {
+ mlx4_warn(dev, "Internal error detected on the communication channel\n");
+ goto internal_err;
+ }
+ } else if (readl(priv->catas_err.map)) {
dump_err_buf(dev);
goto internal_err;
}
@@ -182,22 +269,21 @@ void mlx4_start_catas_poll(struct mlx4_dev *dev)
struct mlx4_priv *priv = mlx4_priv(dev);
phys_addr_t addr;
- /*If we are in SRIOV the default of the module param must be 0*/
- if (mlx4_is_mfunc(dev))
- mlx4_internal_err_reset = 0;
-
INIT_LIST_HEAD(&priv->catas_err.list);
init_timer(&priv->catas_err.timer);
priv->catas_err.map = NULL;
- addr = pci_resource_start(dev->persist->pdev, priv->fw.catas_bar) +
- priv->fw.catas_offset;
-
- priv->catas_err.map = ioremap(addr, priv->fw.catas_size * 4);
- if (!priv->catas_err.map) {
- mlx4_warn(dev, "Failed to map internal error buffer at 0x%llx\n",
- (unsigned long long) addr);
- return;
+ if (!mlx4_is_slave(dev)) {
+ addr = pci_resource_start(dev->persist->pdev,
+ priv->fw.catas_bar) +
+ priv->fw.catas_offset;
+
+ priv->catas_err.map = ioremap(addr, priv->fw.catas_size * 4);
+ if (!priv->catas_err.map) {
+ mlx4_warn(dev, "Failed to map internal error buffer at 0x%llx\n",
+ (unsigned long long)addr);
+ return;
+ }
}
priv->catas_err.timer.data = (unsigned long) dev;
diff --git a/drivers/net/ethernet/mellanox/mlx4/cmd.c b/drivers/net/ethernet/mellanox/mlx4/cmd.c
index 3895b2b..e3d6f7f 100644
--- a/drivers/net/ethernet/mellanox/mlx4/cmd.c
+++ b/drivers/net/ethernet/mellanox/mlx4/cmd.c
@@ -42,6 +42,7 @@
#include <linux/mlx4/device.h>
#include <linux/semaphore.h>
#include <rdma/ib_smi.h>
+#include <linux/delay.h>
#include <asm/io.h>
@@ -729,7 +730,7 @@ int __mlx4_cmd(struct mlx4_dev *dev, u64 in_param, u64 *out_param,
EXPORT_SYMBOL_GPL(__mlx4_cmd);
-static int mlx4_ARM_COMM_CHANNEL(struct mlx4_dev *dev)
+int mlx4_ARM_COMM_CHANNEL(struct mlx4_dev *dev)
{
return mlx4_cmd(dev, 0, 0, 0, MLX4_CMD_ARM_COMM_CHANNEL,
MLX4_CMD_TIME_CLASS_B, MLX4_CMD_NATIVE);
@@ -1945,8 +1946,11 @@ static void mlx4_master_do_cmd(struct mlx4_dev *dev, int slave, u8 cmd,
break;
case MLX4_COMM_CMD_VHCR_POST:
if ((slave_state[slave].last_cmd != MLX4_COMM_CMD_VHCR_EN) &&
- (slave_state[slave].last_cmd != MLX4_COMM_CMD_VHCR_POST))
+ (slave_state[slave].last_cmd != MLX4_COMM_CMD_VHCR_POST)) {
+ mlx4_warn(dev, "slave:%d is out of sync, cmd=0x%x, last command=0x%x, reset is needed\n",
+ slave, cmd, slave_state[slave].last_cmd);
goto reset_slave;
+ }
mutex_lock(&priv->cmd.slave_cmd_mutex);
if (mlx4_master_process_vhcr(dev, slave, NULL)) {
@@ -1980,7 +1984,18 @@ static void mlx4_master_do_cmd(struct mlx4_dev *dev, int slave, u8 cmd,
reset_slave:
/* cleanup any slave resources */
- mlx4_delete_all_resources_for_slave(dev, slave);
+ if (dev->persist->interface_state & MLX4_INTERFACE_STATE_UP)
+ mlx4_delete_all_resources_for_slave(dev, slave);
+
+ if (cmd != MLX4_COMM_CMD_RESET) {
+ mlx4_warn(dev, "Turn on internal error to force reset, slave=%d, cmd=0x%x\n",
+ slave, cmd);
+ /* Turn on internal error letting slave reset itself immeditaly,
+ * otherwise it might take till timeout on command is passed
+ */
+ reply |= ((u32)COMM_CHAN_EVENT_INTERNAL_ERR);
+ }
+
spin_lock_irqsave(&priv->mfunc.master.slave_state_lock, flags);
if (!slave_state[slave].is_slave_going_down)
slave_state[slave].last_cmd = MLX4_COMM_CMD_RESET;
@@ -2056,16 +2071,27 @@ void mlx4_master_comm_channel(struct work_struct *work)
static int sync_toggles(struct mlx4_dev *dev)
{
struct mlx4_priv *priv = mlx4_priv(dev);
- int wr_toggle;
- int rd_toggle;
+ u32 wr_toggle;
+ u32 rd_toggle;
unsigned long end;
- wr_toggle = swab32(readl(&priv->mfunc.comm->slave_write)) >> 31;
- end = jiffies + msecs_to_jiffies(5000);
+ wr_toggle = swab32(readl(&priv->mfunc.comm->slave_write));
+ if (wr_toggle == 0xffffffff)
+ end = jiffies + msecs_to_jiffies(30000);
+ else
+ end = jiffies + msecs_to_jiffies(5000);
while (time_before(jiffies, end)) {
- rd_toggle = swab32(readl(&priv->mfunc.comm->slave_read)) >> 31;
- if (rd_toggle == wr_toggle) {
+ rd_toggle = swab32(readl(&priv->mfunc.comm->slave_read));
+ if (wr_toggle == 0xffffffff || rd_toggle == 0xffffffff) {
+ /* PCI might be offline */
+ msleep(100);
+ wr_toggle = swab32(readl(&priv->mfunc.comm->
+ slave_write));
+ continue;
+ }
+
+ if (rd_toggle >> 31 == wr_toggle >> 31) {
priv->cmd.comm_toggle = rd_toggle;
return 0;
}
@@ -2172,13 +2198,6 @@ int mlx4_multi_func_init(struct mlx4_dev *dev)
if (mlx4_init_resource_tracker(dev))
goto err_thread;
- err = mlx4_ARM_COMM_CHANNEL(dev);
- if (err) {
- mlx4_err(dev, " Failed to arm comm channel eq: %x\n",
- err);
- goto err_resource;
- }
-
} else {
err = sync_toggles(dev);
if (err) {
@@ -2188,8 +2207,6 @@ int mlx4_multi_func_init(struct mlx4_dev *dev)
}
return 0;
-err_resource:
- mlx4_free_resource_tracker(dev, RES_TR_FREE_ALL);
err_thread:
flush_workqueue(priv->mfunc.master.comm_wq);
destroy_workqueue(priv->mfunc.master.comm_wq);
@@ -2266,6 +2283,27 @@ err:
return -ENOMEM;
}
+void mlx4_report_internal_err_comm_event(struct mlx4_dev *dev)
+{
+ struct mlx4_priv *priv = mlx4_priv(dev);
+ int slave;
+ u32 slave_read;
+
+ /* Report an internal error event to all
+ * communication channels.
+ */
+ for (slave = 0; slave < dev->num_slaves; slave++) {
+ slave_read = swab32(readl(&priv->mfunc.comm[slave].slave_read));
+ slave_read |= (u32)COMM_CHAN_EVENT_INTERNAL_ERR;
+ __raw_writel((__force u32)cpu_to_be32(slave_read),
+ &priv->mfunc.comm[slave].slave_read);
+ /* Make sure that our comm channel write doesn't
+ * get mixed in with writes from another CPU.
+ */
+ mmiowb();
+ }
+}
+
void mlx4_multi_func_cleanup(struct mlx4_dev *dev)
{
struct mlx4_priv *priv = mlx4_priv(dev);
@@ -2281,6 +2319,7 @@ void mlx4_multi_func_cleanup(struct mlx4_dev *dev)
kfree(priv->mfunc.master.slave_state);
kfree(priv->mfunc.master.vf_admin);
kfree(priv->mfunc.master.vf_oper);
+ dev->num_slaves = 0;
}
iounmap(priv->mfunc.comm);
diff --git a/drivers/net/ethernet/mellanox/mlx4/eq.c b/drivers/net/ethernet/mellanox/mlx4/eq.c
index 7538c9c..2f2e606 100644
--- a/drivers/net/ethernet/mellanox/mlx4/eq.c
+++ b/drivers/net/ethernet/mellanox/mlx4/eq.c
@@ -429,8 +429,14 @@ void mlx4_master_handle_slave_flr(struct work_struct *work)
if (MLX4_COMM_CMD_FLR == slave_state[i].last_cmd) {
mlx4_dbg(dev, "mlx4_handle_slave_flr: clean slave: %d\n",
i);
-
- mlx4_delete_all_resources_for_slave(dev, i);
+ /* In case of 'Reset flow' FLR can be generated for
+ * a slave before mlx4_load_one is done.
+ * make sure interface is up before trying to delete
+ * slave resources which weren't allocated yet.
+ */
+ if (dev->persist->interface_state &
+ MLX4_INTERFACE_STATE_UP)
+ mlx4_delete_all_resources_for_slave(dev, i);
/*return the slave to running mode*/
spin_lock_irqsave(&priv->mfunc.master.slave_state_lock, flags);
slave_state[i].last_cmd = MLX4_COMM_CMD_RESET;
diff --git a/drivers/net/ethernet/mellanox/mlx4/intf.c b/drivers/net/ethernet/mellanox/mlx4/intf.c
index fba0b96..68d2bad 100644
--- a/drivers/net/ethernet/mellanox/mlx4/intf.c
+++ b/drivers/net/ethernet/mellanox/mlx4/intf.c
@@ -144,8 +144,7 @@ int mlx4_register_device(struct mlx4_dev *dev)
mlx4_add_device(intf, priv);
mutex_unlock(&intf_mutex);
- if (!mlx4_is_slave(dev))
- mlx4_start_catas_poll(dev);
+ mlx4_start_catas_poll(dev);
return 0;
}
@@ -155,8 +154,7 @@ void mlx4_unregister_device(struct mlx4_dev *dev)
struct mlx4_priv *priv = mlx4_priv(dev);
struct mlx4_interface *intf;
- if (!mlx4_is_slave(dev))
- mlx4_stop_catas_poll(dev);
+ mlx4_stop_catas_poll(dev);
mutex_lock(&intf_mutex);
list_for_each_entry(intf, &intf_list, list)
diff --git a/drivers/net/ethernet/mellanox/mlx4/main.c b/drivers/net/ethernet/mellanox/mlx4/main.c
index 01d12f7..496d6e7 100644
--- a/drivers/net/ethernet/mellanox/mlx4/main.c
+++ b/drivers/net/ethernet/mellanox/mlx4/main.c
@@ -108,6 +108,8 @@ MODULE_PARM_DESC(enable_64b_cqe_eqe,
MLX4_FUNC_CAP_EQE_CQE_STRIDE | \
MLX4_FUNC_CAP_DMFS_A0_STATIC)
+#define RESET_PERSIST_MASK_FLAGS (MLX4_FLAG_SRIOV)
+
static char mlx4_version[] =
DRV_NAME ": Mellanox ConnectX core driver v"
DRV_VERSION " (" DRV_RELDATE ")\n";
@@ -1579,6 +1581,50 @@ static void mlx4_close_fw(struct mlx4_dev *dev)
}
}
+static int mlx4_comm_check_offline(struct mlx4_dev *dev)
+{
+#define COMM_CHAN_OFFLINE_OFFSET 0x09
+
+ u32 comm_flags;
+ u32 offline_bit;
+ unsigned long end;
+ struct mlx4_priv *priv = mlx4_priv(dev);
+
+ end = msecs_to_jiffies(MLX4_COMM_OFFLINE_TIME_OUT) + jiffies;
+ while (time_before(jiffies, end)) {
+ comm_flags = swab32(readl((__iomem char *)priv->mfunc.comm +
+ MLX4_COMM_CHAN_FLAGS));
+ offline_bit = (comm_flags &
+ (u32)(1 << COMM_CHAN_OFFLINE_OFFSET));
+ if (!offline_bit)
+ return 0;
+ /* There are cases as part of AER/Reset flow that PF needs
+ * around 100 msec to load. We therefore sleep for 100 msec
+ * to allow other tasks to make use of that CPU during this
+ * time interval.
+ */
+ msleep(100);
+ }
+ mlx4_err(dev, "Communication channel is offline.\n");
+ return -EIO;
+}
+
+static void mlx4_reset_vf_support(struct mlx4_dev *dev)
+{
+#define COMM_CHAN_RST_OFFSET 0x1e
+
+ struct mlx4_priv *priv = mlx4_priv(dev);
+ u32 comm_rst;
+ u32 comm_caps;
+
+ comm_caps = swab32(readl((__iomem char *)priv->mfunc.comm +
+ MLX4_COMM_CHAN_CAPS));
+ comm_rst = (comm_caps & (u32)(1 << COMM_CHAN_RST_OFFSET));
+
+ if (comm_rst)
+ dev->caps.vf_caps |= MLX4_VF_CAP_FLAG_RESET;
+}
+
static int mlx4_init_slave(struct mlx4_dev *dev)
{
struct mlx4_priv *priv = mlx4_priv(dev);
@@ -1594,6 +1640,12 @@ static int mlx4_init_slave(struct mlx4_dev *dev)
mutex_lock(&priv->cmd.slave_cmd_mutex);
priv->cmd.max_cmds = 1;
+ if (mlx4_comm_check_offline(dev)) {
+ mlx4_err(dev, "PF is not responsive, skipping initialization\n");
+ goto err_offline;
+ }
+
+ mlx4_reset_vf_support(dev);
mlx4_warn(dev, "Sending reset\n");
ret_from_reset = mlx4_comm_cmd(dev, MLX4_COMM_CMD_RESET, 0,
MLX4_COMM_TIME);
@@ -1637,6 +1689,7 @@ static int mlx4_init_slave(struct mlx4_dev *dev)
err:
mlx4_comm_cmd(dev, MLX4_COMM_CMD_RESET, 0, 0);
+err_offline:
mutex_unlock(&priv->cmd.slave_cmd_mutex);
return -EIO;
}
@@ -2494,11 +2547,19 @@ static void mlx4_free_ownership(struct mlx4_dev *dev)
!!((flags) & MLX4_FLAG_MASTER))
static u64 mlx4_enable_sriov(struct mlx4_dev *dev, struct pci_dev *pdev,
- u8 total_vfs, int existing_vfs)
+ u8 total_vfs, int existing_vfs, int reset_flow)
{
u64 dev_flags = dev->flags;
int err = 0;
+ if (reset_flow) {
+ dev->dev_vfs = kcalloc(total_vfs, sizeof(*dev->dev_vfs),
+ GFP_KERNEL);
+ if (!dev->dev_vfs)
+ goto free_mem;
+ return dev_flags;
+ }
+
atomic_inc(&pf_loading);
if (dev->flags & MLX4_FLAG_SRIOV) {
if (existing_vfs != total_vfs) {
@@ -2533,6 +2594,7 @@ static u64 mlx4_enable_sriov(struct mlx4_dev *dev, struct pci_dev *pdev,
disable_sriov:
atomic_dec(&pf_loading);
+free_mem:
dev->persist->num_vfs = 0;
kfree(dev->dev_vfs);
return dev_flags & ~MLX4_FLAG_MASTER;
@@ -2557,7 +2619,8 @@ static int mlx4_check_dev_cap(struct mlx4_dev *dev, struct mlx4_dev_cap *dev_cap
}
static int mlx4_load_one(struct pci_dev *pdev, int pci_dev_data,
- int total_vfs, int *nvfs, struct mlx4_priv *priv)
+ int total_vfs, int *nvfs, struct mlx4_priv *priv,
+ int reset_flow)
{
struct mlx4_dev *dev;
unsigned sum = 0;
@@ -2679,8 +2742,10 @@ slave_start:
goto err_fw;
if (!(dev_cap->flags2 & MLX4_DEV_CAP_FLAG2_SYS_EQS)) {
- u64 dev_flags = mlx4_enable_sriov(dev, pdev, total_vfs,
- existing_vfs);
+ u64 dev_flags = mlx4_enable_sriov(dev, pdev,
+ total_vfs,
+ existing_vfs,
+ reset_flow);
mlx4_cmd_cleanup(dev, MLX4_CMD_CLEANUP_ALL);
dev->flags = dev_flags;
@@ -2722,7 +2787,7 @@ slave_start:
if (dev->flags & MLX4_FLAG_SRIOV) {
if (!existing_vfs)
pci_disable_sriov(pdev);
- if (mlx4_is_master(dev))
+ if (mlx4_is_master(dev) && !reset_flow)
atomic_dec(&pf_loading);
dev->flags &= ~MLX4_FLAG_SRIOV;
}
@@ -2736,7 +2801,8 @@ slave_start:
}
if (mlx4_is_master(dev) && (dev_cap->flags2 & MLX4_DEV_CAP_FLAG2_SYS_EQS)) {
- u64 dev_flags = mlx4_enable_sriov(dev, pdev, total_vfs, existing_vfs);
+ u64 dev_flags = mlx4_enable_sriov(dev, pdev, total_vfs,
+ existing_vfs, reset_flow);
if ((dev->flags ^ dev_flags) & (MLX4_FLAG_MASTER | MLX4_FLAG_SLAVE)) {
mlx4_cmd_cleanup(dev, MLX4_CMD_CLEANUP_VHCR);
@@ -2848,6 +2914,17 @@ slave_start:
goto err_steer;
mlx4_init_quotas(dev);
+ /* When PF resources are ready arm its comm channel to enable
+ * getting commands
+ */
+ if (mlx4_is_master(dev)) {
+ err = mlx4_ARM_COMM_CHANNEL(dev);
+ if (err) {
+ mlx4_err(dev, " Failed to arm comm channel eq: %x\n",
+ err);
+ goto err_steer;
+ }
+ }
for (port = 1; port <= dev->caps.num_ports; port++) {
err = mlx4_init_port_info(dev, port);
@@ -2866,7 +2943,7 @@ slave_start:
priv->removed = 0;
- if (mlx4_is_master(dev) && dev->persist->num_vfs)
+ if (mlx4_is_master(dev) && dev->persist->num_vfs && !reset_flow)
atomic_dec(&pf_loading);
kfree(dev_cap);
@@ -2925,10 +3002,12 @@ err_cmd:
mlx4_cmd_cleanup(dev, MLX4_CMD_CLEANUP_ALL);
err_sriov:
- if (dev->flags & MLX4_FLAG_SRIOV && !existing_vfs)
+ if (dev->flags & MLX4_FLAG_SRIOV && !existing_vfs) {
pci_disable_sriov(pdev);
+ dev->flags &= ~MLX4_FLAG_SRIOV;
+ }
- if (mlx4_is_master(dev) && dev->persist->num_vfs)
+ if (mlx4_is_master(dev) && dev->persist->num_vfs && !reset_flow)
atomic_dec(&pf_loading);
kfree(priv->dev.dev_vfs);
@@ -3073,7 +3152,7 @@ static int __mlx4_init_one(struct pci_dev *pdev, int pci_dev_data,
if (err)
goto err_release_regions;
- err = mlx4_load_one(pdev, pci_dev_data, total_vfs, nvfs, priv);
+ err = mlx4_load_one(pdev, pci_dev_data, total_vfs, nvfs, priv, 0);
if (err)
goto err_catas;
@@ -3131,9 +3210,11 @@ static void mlx4_clean_dev(struct mlx4_dev *dev)
{
struct mlx4_dev_persistent *persist = dev->persist;
struct mlx4_priv *priv = mlx4_priv(dev);
+ unsigned long flags = (dev->flags & RESET_PERSIST_MASK_FLAGS);
memset(priv, 0, sizeof(*priv));
priv->dev.persist = persist;
+ priv->dev.flags = flags;
}
static void mlx4_unload_one(struct pci_dev *pdev)
@@ -3143,7 +3224,6 @@ static void mlx4_unload_one(struct pci_dev *pdev)
struct mlx4_priv *priv = mlx4_priv(dev);
int pci_dev_data;
int p, i;
- int active_vfs = 0;
if (priv->removed)
return;
@@ -3157,14 +3237,6 @@ static void mlx4_unload_one(struct pci_dev *pdev)
pci_dev_data = priv->pci_dev_data;
- /* Disabling SR-IOV is not allowed while there are active vf's */
- if (mlx4_is_master(dev)) {
- active_vfs = mlx4_how_many_lives_vf(dev);
- if (active_vfs) {
- pr_warn("Removing PF when there are active VF's !!\n");
- pr_warn("Will not disable SR-IOV.\n");
- }
- }
mlx4_stop_sense(dev);
mlx4_unregister_device(dev);
@@ -3208,12 +3280,6 @@ static void mlx4_unload_one(struct pci_dev *pdev)
if (dev->flags & MLX4_FLAG_MSI_X)
pci_disable_msix(pdev);
- if (dev->flags & MLX4_FLAG_SRIOV && !active_vfs) {
- mlx4_warn(dev, "Disabling SR-IOV\n");
- pci_disable_sriov(pdev);
- dev->flags &= ~MLX4_FLAG_SRIOV;
- dev->persist->num_vfs = 0;
- }
if (!mlx4_is_slave(dev))
mlx4_free_ownership(dev);
@@ -3235,11 +3301,21 @@ static void mlx4_remove_one(struct pci_dev *pdev)
struct mlx4_dev_persistent *persist = pci_get_drvdata(pdev);
struct mlx4_dev *dev = persist->dev;
struct mlx4_priv *priv = mlx4_priv(dev);
+ int active_vfs = 0;
mutex_lock(&persist->interface_state_mutex);
persist->interface_state |= MLX4_INTERFACE_STATE_DELETION;
mutex_unlock(&persist->interface_state_mutex);
+ /* Disabling SR-IOV is not allowed while there are active vf's */
+ if (mlx4_is_master(dev) && dev->flags & MLX4_FLAG_SRIOV) {
+ active_vfs = mlx4_how_many_lives_vf(dev);
+ if (active_vfs) {
+ pr_warn("Removing PF when there are active VF's !!\n");
+ pr_warn("Will not disable SR-IOV.\n");
+ }
+ }
+
/* device marked to be under deletion running now without the lock
* letting other tasks to be terminated
*/
@@ -3248,6 +3324,11 @@ static void mlx4_remove_one(struct pci_dev *pdev)
else
mlx4_info(dev, "%s: interface is down\n", __func__);
mlx4_catas_end(dev);
+ if (dev->flags & MLX4_FLAG_SRIOV && !active_vfs) {
+ mlx4_warn(dev, "Disabling SR-IOV\n");
+ pci_disable_sriov(pdev);
+ }
+
pci_release_regions(pdev);
pci_disable_device(pdev);
kfree(dev->persist);
@@ -3287,7 +3368,7 @@ int mlx4_restart_one(struct pci_dev *pdev)
memcpy(nvfs, dev->persist->nvfs, sizeof(dev->persist->nvfs));
mlx4_unload_one(pdev);
- err = mlx4_load_one(pdev, pci_dev_data, total_vfs, nvfs, priv);
+ err = mlx4_load_one(pdev, pci_dev_data, total_vfs, nvfs, priv, 1);
if (err) {
mlx4_err(dev, "%s: ERROR: mlx4_load_one failed, pci_name=%s, err=%d\n",
__func__, pci_name(pdev), err);
@@ -3396,7 +3477,7 @@ static pci_ers_result_t mlx4_pci_slot_reset(struct pci_dev *pdev)
mutex_lock(&persist->interface_state_mutex);
if (!(persist->interface_state & MLX4_INTERFACE_STATE_UP)) {
ret = mlx4_load_one(pdev, priv->pci_dev_data, total_vfs, nvfs,
- priv);
+ priv, 1);
if (ret) {
mlx4_err(dev, "%s: mlx4_load_one failed, ret=%d\n",
__func__, ret);
diff --git a/drivers/net/ethernet/mellanox/mlx4/mlx4.h b/drivers/net/ethernet/mellanox/mlx4/mlx4.h
index 5c772ea..2a15b82 100644
--- a/drivers/net/ethernet/mellanox/mlx4/mlx4.h
+++ b/drivers/net/ethernet/mellanox/mlx4/mlx4.h
@@ -85,7 +85,9 @@ enum {
MLX4_CLR_INT_SIZE = 0x00008,
MLX4_SLAVE_COMM_BASE = 0x0,
MLX4_COMM_PAGESIZE = 0x1000,
- MLX4_CLOCK_SIZE = 0x00008
+ MLX4_CLOCK_SIZE = 0x00008,
+ MLX4_COMM_CHAN_CAPS = 0x8,
+ MLX4_COMM_CHAN_FLAGS = 0xc
};
enum {
@@ -120,6 +122,8 @@ enum mlx4_mpt_state {
};
#define MLX4_COMM_TIME 10000
+#define MLX4_COMM_OFFLINE_TIME_OUT 30000
+
enum {
MLX4_COMM_CMD_RESET,
MLX4_COMM_CMD_VHCR0,
@@ -1162,6 +1166,7 @@ enum {
int mlx4_cmd_init(struct mlx4_dev *dev);
void mlx4_cmd_cleanup(struct mlx4_dev *dev, int cleanup_mask);
int mlx4_multi_func_init(struct mlx4_dev *dev);
+int mlx4_ARM_COMM_CHANNEL(struct mlx4_dev *dev);
void mlx4_multi_func_cleanup(struct mlx4_dev *dev);
void mlx4_cmd_event(struct mlx4_dev *dev, u16 token, u8 status, u64 out_param);
int mlx4_cmd_use_events(struct mlx4_dev *dev);
diff --git a/include/linux/mlx4/cmd.h b/include/linux/mlx4/cmd.h
index e754384..c989442 100644
--- a/include/linux/mlx4/cmd.h
+++ b/include/linux/mlx4/cmd.h
@@ -280,6 +280,7 @@ int mlx4_set_vf_link_state(struct mlx4_dev *dev, int port, int vf, int link_stat
int mlx4_config_dev_retrieval(struct mlx4_dev *dev,
struct mlx4_config_dev_params *params);
void mlx4_cmd_wake_completions(struct mlx4_dev *dev);
+void mlx4_report_internal_err_comm_event(struct mlx4_dev *dev);
/*
* mlx4_get_slave_default_vlan -
* return true if VST ( default vlan)
@@ -289,5 +290,6 @@ bool mlx4_get_slave_default_vlan(struct mlx4_dev *dev, int port, int slave,
u16 *vlan, u8 *qos);
#define MLX4_COMM_GET_IF_REV(cmd_chan_ver) (u8)((cmd_chan_ver) >> 8)
+#define COMM_CHAN_EVENT_INTERNAL_ERR (1 << 17)
#endif /* MLX4_CMD_H */
diff --git a/include/linux/mlx4/device.h b/include/linux/mlx4/device.h
index 33f9ca7..5ef54e1 100644
--- a/include/linux/mlx4/device.h
+++ b/include/linux/mlx4/device.h
@@ -208,6 +208,10 @@ enum {
MLX4_QUERY_FUNC_FLAGS_A0_RES_QP = 1LL << 1
};
+enum {
+ MLX4_VF_CAP_FLAG_RESET = 1 << 0
+};
+
/* bit enums for an 8-bit flags field indicating special use
* QPs which require special handling in qp_reserve_range.
* Currently, this only includes QPs used by the ETH interface,
@@ -545,6 +549,7 @@ struct mlx4_caps {
u8 alloc_res_qp_mask;
u32 dmfs_high_rate_qpn_base;
u32 dmfs_high_rate_qpn_range;
+ u32 vf_caps;
};
struct mlx4_buf_list {
--
1.7.1
^ permalink raw reply related
* [PATCH net-next 2/9] net/mlx4_core: Set device configuration data to be persistent across reset
From: Or Gerlitz @ 2015-01-21 14:45 UTC (permalink / raw)
To: David S. Miller
Cc: netdev, Matan Barak, Amir Vadai, Tal Alon, Roland Dreier,
Yishai Hadas, Or Gerlitz
In-Reply-To: <1421851557-12084-1-git-send-email-ogerlitz@mellanox.com>
From: Yishai Hadas <yishaih@mellanox.com>
When an HCA enters an internal error state, this is detected by the driver.
The driver then should reset the HCA and restart the software stack.
Keep ports information and some SRIOV configuration in a persistent area
to have it valid across reset.
Signed-off-by: Yishai Hadas <yishaih@mellanox.com>
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
---
drivers/net/ethernet/mellanox/mlx4/main.c | 45 +++++++++++++++++++++++++++-
include/linux/mlx4/device.h | 2 +
2 files changed, 45 insertions(+), 2 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx4/main.c b/drivers/net/ethernet/mellanox/mlx4/main.c
index abcee61..2c5a555 100644
--- a/drivers/net/ethernet/mellanox/mlx4/main.c
+++ b/drivers/net/ethernet/mellanox/mlx4/main.c
@@ -3109,18 +3109,34 @@ static int mlx4_init_one(struct pci_dev *pdev, const struct pci_device_id *id)
return ret;
}
+static void mlx4_clean_dev(struct mlx4_dev *dev)
+{
+ struct mlx4_dev_persistent *persist = dev->persist;
+ struct mlx4_priv *priv = mlx4_priv(dev);
+
+ memset(priv, 0, sizeof(*priv));
+ priv->dev.persist = persist;
+}
+
static void mlx4_unload_one(struct pci_dev *pdev)
{
struct mlx4_dev_persistent *persist = pci_get_drvdata(pdev);
struct mlx4_dev *dev = persist->dev;
struct mlx4_priv *priv = mlx4_priv(dev);
int pci_dev_data;
- int p;
+ int p, i;
int active_vfs = 0;
if (priv->removed)
return;
+ /* saving current ports type for further use */
+ for (i = 0; i < dev->caps.num_ports; i++) {
+ dev->persist->curr_port_type[i] = dev->caps.port_type[i + 1];
+ dev->persist->curr_port_poss_type[i] = dev->caps.
+ possible_type[i + 1];
+ }
+
pci_dev_data = priv->pci_dev_data;
/* Disabling SR-IOV is not allowed while there are active vf's */
@@ -3191,7 +3207,7 @@ static void mlx4_unload_one(struct pci_dev *pdev)
kfree(dev->caps.qp1_proxy);
kfree(dev->dev_vfs);
- memset(priv, 0, sizeof(*priv));
+ mlx4_clean_dev(dev);
priv->pci_dev_data = pci_dev_data;
priv->removed = 1;
}
@@ -3210,6 +3226,25 @@ static void mlx4_remove_one(struct pci_dev *pdev)
pci_set_drvdata(pdev, NULL);
}
+static int restore_current_port_types(struct mlx4_dev *dev,
+ enum mlx4_port_type *types,
+ enum mlx4_port_type *poss_types)
+{
+ struct mlx4_priv *priv = mlx4_priv(dev);
+ int err, i;
+
+ mlx4_stop_sense(dev);
+
+ mutex_lock(&priv->port_mutex);
+ for (i = 0; i < dev->caps.num_ports; i++)
+ dev->caps.possible_type[i + 1] = poss_types[i];
+ err = mlx4_change_port_types(dev, types);
+ mlx4_start_sense(dev);
+ mutex_unlock(&priv->port_mutex);
+
+ return err;
+}
+
int mlx4_restart_one(struct pci_dev *pdev)
{
struct mlx4_dev_persistent *persist = pci_get_drvdata(pdev);
@@ -3230,6 +3265,12 @@ int mlx4_restart_one(struct pci_dev *pdev)
return err;
}
+ err = restore_current_port_types(dev, dev->persist->curr_port_type,
+ dev->persist->curr_port_poss_type);
+ if (err)
+ mlx4_err(dev, "could not restore original port types (%d)\n",
+ err);
+
return err;
}
diff --git a/include/linux/mlx4/device.h b/include/linux/mlx4/device.h
index 1069ce6..8c3837a 100644
--- a/include/linux/mlx4/device.h
+++ b/include/linux/mlx4/device.h
@@ -749,6 +749,8 @@ struct mlx4_dev_persistent {
struct mlx4_dev *dev;
int nvfs[MLX4_MAX_PORTS + 1];
int num_vfs;
+ enum mlx4_port_type curr_port_type[MLX4_MAX_PORTS + 1];
+ enum mlx4_port_type curr_port_poss_type[MLX4_MAX_PORTS + 1];
};
struct mlx4_dev {
--
1.7.1
^ permalink raw reply related
* [PATCH net-next 3/9] net/mlx4_core: Refactor the catas flow to work per device
From: Or Gerlitz @ 2015-01-21 14:45 UTC (permalink / raw)
To: David S. Miller
Cc: netdev, Matan Barak, Amir Vadai, Tal Alon, Roland Dreier,
Yishai Hadas, Or Gerlitz
In-Reply-To: <1421851557-12084-1-git-send-email-ogerlitz@mellanox.com>
From: Yishai Hadas <yishaih@mellanox.com>
Using a WQ per device instead of a single global WQ, this allows
independent reset handling per device even when SRIOV is used.
This comes as a pre-patch for supporting chip reset
for both native and SRIOV.
Signed-off-by: Yishai Hadas <yishaih@mellanox.com>
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
---
drivers/net/ethernet/mellanox/mlx4/catas.c | 75 +++++++++++++---------------
drivers/net/ethernet/mellanox/mlx4/main.c | 12 ++++-
drivers/net/ethernet/mellanox/mlx4/mlx4.h | 3 +-
include/linux/mlx4/device.h | 2 +
4 files changed, 48 insertions(+), 44 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx4/catas.c b/drivers/net/ethernet/mellanox/mlx4/catas.c
index 1a102c9..5bb9aa6 100644
--- a/drivers/net/ethernet/mellanox/mlx4/catas.c
+++ b/drivers/net/ethernet/mellanox/mlx4/catas.c
@@ -40,10 +40,7 @@ enum {
MLX4_CATAS_POLL_INTERVAL = 5 * HZ,
};
-static DEFINE_SPINLOCK(catas_lock);
-static LIST_HEAD(catas_list);
-static struct work_struct catas_work;
static int internal_err_reset = 1;
module_param(internal_err_reset, int, 0644);
@@ -77,13 +74,9 @@ static void poll_catas(unsigned long dev_ptr)
dump_err_buf(dev);
mlx4_dispatch_event(dev, MLX4_DEV_EVENT_CATASTROPHIC_ERROR, 0);
- if (internal_err_reset) {
- spin_lock(&catas_lock);
- list_add(&priv->catas_err.list, &catas_list);
- spin_unlock(&catas_lock);
-
- queue_work(mlx4_wq, &catas_work);
- }
+ if (internal_err_reset)
+ queue_work(dev->persist->catas_wq,
+ &dev->persist->catas_work);
}
} else
mod_timer(&priv->catas_err.timer,
@@ -92,34 +85,23 @@ static void poll_catas(unsigned long dev_ptr)
static void catas_reset(struct work_struct *work)
{
- struct mlx4_priv *priv, *tmppriv;
- struct mlx4_dev *dev;
- struct mlx4_dev_persistent *persist;
-
- LIST_HEAD(tlist);
+ struct mlx4_dev_persistent *persist =
+ container_of(work, struct mlx4_dev_persistent,
+ catas_work);
+ struct pci_dev *pdev = persist->pdev;
int ret;
- spin_lock_irq(&catas_lock);
- list_splice_init(&catas_list, &tlist);
- spin_unlock_irq(&catas_lock);
-
- list_for_each_entry_safe(priv, tmppriv, &tlist, catas_err.list) {
- struct pci_dev *pdev = priv->dev.persist->pdev;
-
- /* If the device is off-line, we cannot reset it */
- if (pci_channel_offline(pdev))
- continue;
+ /* If the device is off-line, we cannot reset it */
+ if (pci_channel_offline(pdev))
+ return;
- ret = mlx4_restart_one(priv->dev.persist->pdev);
- /* 'priv' now is not valid */
- if (ret)
- pr_err("mlx4 %s: Reset failed (%d)\n",
- pci_name(pdev), ret);
- else {
- persist = pci_get_drvdata(pdev);
- mlx4_dbg(persist->dev, "Reset succeeded\n");
- }
- }
+ ret = mlx4_restart_one(pdev);
+ /* 'priv' now is not valid */
+ if (ret)
+ pr_err("mlx4 %s: Reset failed (%d)\n",
+ pci_name(pdev), ret);
+ else
+ mlx4_dbg(persist->dev, "Reset succeeded\n");
}
void mlx4_start_catas_poll(struct mlx4_dev *dev)
@@ -158,15 +140,26 @@ void mlx4_stop_catas_poll(struct mlx4_dev *dev)
del_timer_sync(&priv->catas_err.timer);
- if (priv->catas_err.map)
+ if (priv->catas_err.map) {
iounmap(priv->catas_err.map);
+ priv->catas_err.map = NULL;
+ }
+}
- spin_lock_irq(&catas_lock);
- list_del(&priv->catas_err.list);
- spin_unlock_irq(&catas_lock);
+int mlx4_catas_init(struct mlx4_dev *dev)
+{
+ INIT_WORK(&dev->persist->catas_work, catas_reset);
+ dev->persist->catas_wq = create_singlethread_workqueue("mlx4_health");
+ if (!dev->persist->catas_wq)
+ return -ENOMEM;
+
+ return 0;
}
-void __init mlx4_catas_init(void)
+void mlx4_catas_end(struct mlx4_dev *dev)
{
- INIT_WORK(&catas_work, catas_reset);
+ if (dev->persist->catas_wq) {
+ destroy_workqueue(dev->persist->catas_wq);
+ dev->persist->catas_wq = NULL;
+ }
}
diff --git a/drivers/net/ethernet/mellanox/mlx4/main.c b/drivers/net/ethernet/mellanox/mlx4/main.c
index 2c5a555..a61694c 100644
--- a/drivers/net/ethernet/mellanox/mlx4/main.c
+++ b/drivers/net/ethernet/mellanox/mlx4/main.c
@@ -3064,11 +3064,19 @@ static int __mlx4_init_one(struct pci_dev *pdev, int pci_dev_data,
}
}
- err = mlx4_load_one(pdev, pci_dev_data, total_vfs, nvfs, priv);
+ err = mlx4_catas_init(&priv->dev);
if (err)
goto err_release_regions;
+
+ err = mlx4_load_one(pdev, pci_dev_data, total_vfs, nvfs, priv);
+ if (err)
+ goto err_catas;
+
return 0;
+err_catas:
+ mlx4_catas_end(&priv->dev);
+
err_release_regions:
pci_release_regions(pdev);
@@ -3219,6 +3227,7 @@ static void mlx4_remove_one(struct pci_dev *pdev)
struct mlx4_priv *priv = mlx4_priv(dev);
mlx4_unload_one(pdev);
+ mlx4_catas_end(dev);
pci_release_regions(pdev);
pci_disable_device(pdev);
kfree(dev->persist);
@@ -3403,7 +3412,6 @@ static int __init mlx4_init(void)
if (mlx4_verify_params())
return -EINVAL;
- mlx4_catas_init();
mlx4_wq = create_singlethread_workqueue("mlx4");
if (!mlx4_wq)
diff --git a/drivers/net/ethernet/mellanox/mlx4/mlx4.h b/drivers/net/ethernet/mellanox/mlx4/mlx4.h
index faa37ab..d41af84 100644
--- a/drivers/net/ethernet/mellanox/mlx4/mlx4.h
+++ b/drivers/net/ethernet/mellanox/mlx4/mlx4.h
@@ -995,7 +995,8 @@ void __mlx4_xrcd_free(struct mlx4_dev *dev, u32 xrcdn);
void mlx4_start_catas_poll(struct mlx4_dev *dev);
void mlx4_stop_catas_poll(struct mlx4_dev *dev);
-void mlx4_catas_init(void);
+int mlx4_catas_init(struct mlx4_dev *dev);
+void mlx4_catas_end(struct mlx4_dev *dev);
int mlx4_restart_one(struct pci_dev *pdev);
int mlx4_register_device(struct mlx4_dev *dev);
void mlx4_unregister_device(struct mlx4_dev *dev);
diff --git a/include/linux/mlx4/device.h b/include/linux/mlx4/device.h
index 8c3837a..da425d2 100644
--- a/include/linux/mlx4/device.h
+++ b/include/linux/mlx4/device.h
@@ -751,6 +751,8 @@ struct mlx4_dev_persistent {
int num_vfs;
enum mlx4_port_type curr_port_type[MLX4_MAX_PORTS + 1];
enum mlx4_port_type curr_port_poss_type[MLX4_MAX_PORTS + 1];
+ struct work_struct catas_work;
+ struct workqueue_struct *catas_wq;
};
struct mlx4_dev {
--
1.7.1
^ permalink raw reply related
* [PATCH net-next 6/9] net/mlx4_core: Manage interface state for Reset flow cases
From: Or Gerlitz @ 2015-01-21 14:45 UTC (permalink / raw)
To: David S. Miller
Cc: netdev, Matan Barak, Amir Vadai, Tal Alon, Roland Dreier,
Yishai Hadas, Or Gerlitz
In-Reply-To: <1421851557-12084-1-git-send-email-ogerlitz@mellanox.com>
From: Yishai Hadas <yishaih@mellanox.com>
We need to manage interface state to sync between reset flow and some other
relative cases such as remove_one. This has to be done to prevent certain
races. For example in case software stack is down as a result of unload call,
the remove_one should skip the unload phase.
Implement the remove_one case, handling AER and other cases comes next.
The interface can be up/down, upon remove_one, the state will include an extra
bit indicating that the device is cleaned-up, forcing other tasks to finish
before the final cleanup.
Signed-off-by: Yishai Hadas <yishaih@mellanox.com>
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
---
drivers/net/ethernet/mellanox/mlx4/catas.c | 13 +++++++++++--
drivers/net/ethernet/mellanox/mlx4/intf.c | 2 ++
drivers/net/ethernet/mellanox/mlx4/main.c | 13 ++++++++++++-
include/linux/mlx4/device.h | 7 +++++++
4 files changed, 32 insertions(+), 3 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx4/catas.c b/drivers/net/ethernet/mellanox/mlx4/catas.c
index 63f14ff..3fcf3cf 100644
--- a/drivers/net/ethernet/mellanox/mlx4/catas.c
+++ b/drivers/net/ethernet/mellanox/mlx4/catas.c
@@ -122,8 +122,14 @@ static void mlx4_handle_error_state(struct mlx4_dev_persistent *persist)
int err = 0;
mlx4_enter_error_state(persist);
- err = mlx4_restart_one(persist->pdev);
- mlx4_info(persist->dev, "mlx4_restart_one was ended, ret=%d\n", err);
+ mutex_lock(&persist->interface_state_mutex);
+ if (persist->interface_state & MLX4_INTERFACE_STATE_UP &&
+ !(persist->interface_state & MLX4_INTERFACE_STATE_DELETION)) {
+ err = mlx4_restart_one(persist->pdev);
+ mlx4_info(persist->dev, "mlx4_restart_one was ended, ret=%d\n",
+ err);
+ }
+ mutex_unlock(&persist->interface_state_mutex);
}
static void dump_err_buf(struct mlx4_dev *dev)
@@ -211,6 +217,9 @@ void mlx4_stop_catas_poll(struct mlx4_dev *dev)
iounmap(priv->catas_err.map);
priv->catas_err.map = NULL;
}
+
+ if (dev->persist->interface_state & MLX4_INTERFACE_STATE_DELETION)
+ flush_workqueue(dev->persist->catas_wq);
}
int mlx4_catas_init(struct mlx4_dev *dev)
diff --git a/drivers/net/ethernet/mellanox/mlx4/intf.c b/drivers/net/ethernet/mellanox/mlx4/intf.c
index 116895a..fba0b96 100644
--- a/drivers/net/ethernet/mellanox/mlx4/intf.c
+++ b/drivers/net/ethernet/mellanox/mlx4/intf.c
@@ -138,6 +138,7 @@ int mlx4_register_device(struct mlx4_dev *dev)
mutex_lock(&intf_mutex);
+ dev->persist->interface_state |= MLX4_INTERFACE_STATE_UP;
list_add_tail(&priv->dev_list, &dev_list);
list_for_each_entry(intf, &intf_list, list)
mlx4_add_device(intf, priv);
@@ -162,6 +163,7 @@ void mlx4_unregister_device(struct mlx4_dev *dev)
mlx4_remove_device(intf, priv);
list_del(&priv->dev_list);
+ dev->persist->interface_state &= ~MLX4_INTERFACE_STATE_UP;
mutex_unlock(&intf_mutex);
}
diff --git a/drivers/net/ethernet/mellanox/mlx4/main.c b/drivers/net/ethernet/mellanox/mlx4/main.c
index dc2d910..d59cae5 100644
--- a/drivers/net/ethernet/mellanox/mlx4/main.c
+++ b/drivers/net/ethernet/mellanox/mlx4/main.c
@@ -3114,6 +3114,7 @@ static int mlx4_init_one(struct pci_dev *pdev, const struct pci_device_id *id)
pci_set_drvdata(pdev, dev->persist);
priv->pci_dev_data = id->driver_data;
mutex_init(&dev->persist->device_state_mutex);
+ mutex_init(&dev->persist->interface_state_mutex);
ret = __mlx4_init_one(pdev, id->driver_data, priv);
if (ret) {
@@ -3232,7 +3233,17 @@ static void mlx4_remove_one(struct pci_dev *pdev)
struct mlx4_dev *dev = persist->dev;
struct mlx4_priv *priv = mlx4_priv(dev);
- mlx4_unload_one(pdev);
+ mutex_lock(&persist->interface_state_mutex);
+ persist->interface_state |= MLX4_INTERFACE_STATE_DELETION;
+ mutex_unlock(&persist->interface_state_mutex);
+
+ /* device marked to be under deletion running now without the lock
+ * letting other tasks to be terminated
+ */
+ if (persist->interface_state & MLX4_INTERFACE_STATE_UP)
+ mlx4_unload_one(pdev);
+ else
+ mlx4_info(dev, "%s: interface is down\n", __func__);
mlx4_catas_end(dev);
pci_release_regions(pdev);
pci_disable_device(pdev);
diff --git a/include/linux/mlx4/device.h b/include/linux/mlx4/device.h
index 7d5d317..33f9ca7 100644
--- a/include/linux/mlx4/device.h
+++ b/include/linux/mlx4/device.h
@@ -416,6 +416,11 @@ enum {
MLX4_DEVICE_STATE_INTERNAL_ERROR = 1 << 1,
};
+enum {
+ MLX4_INTERFACE_STATE_UP = 1 << 0,
+ MLX4_INTERFACE_STATE_DELETION = 1 << 1,
+};
+
#define MSTR_SM_CHANGE_MASK (MLX4_EQ_PORT_INFO_MSTR_SM_SL_CHANGE_MASK | \
MLX4_EQ_PORT_INFO_MSTR_SM_LID_CHANGE_MASK)
@@ -760,6 +765,8 @@ struct mlx4_dev_persistent {
struct workqueue_struct *catas_wq;
struct mutex device_state_mutex; /* protect HW state */
u8 state;
+ struct mutex interface_state_mutex; /* protect SW state */
+ u8 interface_state;
};
struct mlx4_dev {
--
1.7.1
^ permalink raw reply related
* [PATCH net-next 5/9] net/mlx4_core: Activate reset flow upon fatal command cases
From: Or Gerlitz @ 2015-01-21 14:45 UTC (permalink / raw)
To: David S. Miller
Cc: netdev, Matan Barak, Amir Vadai, Tal Alon, Roland Dreier,
Yishai Hadas, Or Gerlitz
In-Reply-To: <1421851557-12084-1-git-send-email-ogerlitz@mellanox.com>
From: Yishai Hadas <yishaih@mellanox.com>
We activate reset flow upon command fatal errors, when the device enters an
erroneous state, and must be reset.
The cases below are assumed to be fatal: FW command timed-out, an error from FW
on closing commands, pci is offline when posting/pending a command.
In those cases we place the device into an error state: chip is reset, pending
commands are awakened and completed immediately. Subsequent commands will
return immediately.
The return code in the above cases will depend on the command. Commands which
free and close resources will return success (because the chip was reset, so
callers may safely free their kernel resources). Other commands will return -EIO.
Since the device's state was marked as error, the catas poller will
detect this and restart the device's software stack (as is done when a FW
internal error is directly detected). The device state is protected by a
persistent mutex lives on its mlx4_dev, as such no need any more for the
hcr_mutex which is removed.
Signed-off-by: Yishai Hadas <yishaih@mellanox.com>
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
---
drivers/net/ethernet/mellanox/mlx4/catas.c | 11 +-
drivers/net/ethernet/mellanox/mlx4/cmd.c | 163 ++++++++++++++++++++++++----
drivers/net/ethernet/mellanox/mlx4/mcg.c | 3 +
drivers/net/ethernet/mellanox/mlx4/mlx4.h | 2 +-
include/linux/mlx4/cmd.h | 1 +
5 files changed, 153 insertions(+), 27 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx4/catas.c b/drivers/net/ethernet/mellanox/mlx4/catas.c
index 588d6b5..63f14ff 100644
--- a/drivers/net/ethernet/mellanox/mlx4/catas.c
+++ b/drivers/net/ethernet/mellanox/mlx4/catas.c
@@ -42,8 +42,8 @@ enum {
-static int internal_err_reset = 1;
-module_param(internal_err_reset, int, 0644);
+int mlx4_internal_err_reset = 1;
+module_param_named(internal_err_reset, mlx4_internal_err_reset, int, 0644);
MODULE_PARM_DESC(internal_err_reset,
"Reset device on internal errors if non-zero"
" (default 1, in SRIOV mode default is 0)");
@@ -92,7 +92,7 @@ void mlx4_enter_error_state(struct mlx4_dev_persistent *persist)
int err;
struct mlx4_dev *dev;
- if (!internal_err_reset)
+ if (!mlx4_internal_err_reset)
return;
mutex_lock(&persist->device_state_mutex);
@@ -110,6 +110,7 @@ void mlx4_enter_error_state(struct mlx4_dev_persistent *persist)
/* At that step HW was already reset, now notify clients */
mlx4_dispatch_event(dev, MLX4_DEV_EVENT_CATASTROPHIC_ERROR, 0);
+ mlx4_cmd_wake_completions(dev);
return;
out:
@@ -157,7 +158,7 @@ static void poll_catas(unsigned long dev_ptr)
return;
internal_err:
- if (internal_err_reset)
+ if (mlx4_internal_err_reset)
queue_work(dev->persist->catas_wq, &dev->persist->catas_work);
}
@@ -177,7 +178,7 @@ void mlx4_start_catas_poll(struct mlx4_dev *dev)
/*If we are in SRIOV the default of the module param must be 0*/
if (mlx4_is_mfunc(dev))
- internal_err_reset = 0;
+ mlx4_internal_err_reset = 0;
INIT_LIST_HEAD(&priv->catas_err.list);
init_timer(&priv->catas_err.timer);
diff --git a/drivers/net/ethernet/mellanox/mlx4/cmd.c b/drivers/net/ethernet/mellanox/mlx4/cmd.c
index 7cd90e6..3895b2b 100644
--- a/drivers/net/ethernet/mellanox/mlx4/cmd.c
+++ b/drivers/net/ethernet/mellanox/mlx4/cmd.c
@@ -182,6 +182,72 @@ static u8 mlx4_errno_to_status(int errno)
}
}
+static int mlx4_internal_err_ret_value(struct mlx4_dev *dev, u16 op,
+ u8 op_modifier)
+{
+ switch (op) {
+ case MLX4_CMD_UNMAP_ICM:
+ case MLX4_CMD_UNMAP_ICM_AUX:
+ case MLX4_CMD_UNMAP_FA:
+ case MLX4_CMD_2RST_QP:
+ case MLX4_CMD_HW2SW_EQ:
+ case MLX4_CMD_HW2SW_CQ:
+ case MLX4_CMD_HW2SW_SRQ:
+ case MLX4_CMD_HW2SW_MPT:
+ case MLX4_CMD_CLOSE_HCA:
+ case MLX4_QP_FLOW_STEERING_DETACH:
+ case MLX4_CMD_FREE_RES:
+ case MLX4_CMD_CLOSE_PORT:
+ return CMD_STAT_OK;
+
+ case MLX4_CMD_QP_ATTACH:
+ /* On Detach case return success */
+ if (op_modifier == 0)
+ return CMD_STAT_OK;
+ return mlx4_status_to_errno(CMD_STAT_INTERNAL_ERR);
+
+ default:
+ return mlx4_status_to_errno(CMD_STAT_INTERNAL_ERR);
+ }
+}
+
+static int mlx4_closing_cmd_fatal_error(u16 op, u8 fw_status)
+{
+ /* Any error during the closing commands below is considered fatal */
+ if (op == MLX4_CMD_CLOSE_HCA ||
+ op == MLX4_CMD_HW2SW_EQ ||
+ op == MLX4_CMD_HW2SW_CQ ||
+ op == MLX4_CMD_2RST_QP ||
+ op == MLX4_CMD_HW2SW_SRQ ||
+ op == MLX4_CMD_SYNC_TPT ||
+ op == MLX4_CMD_UNMAP_ICM ||
+ op == MLX4_CMD_UNMAP_ICM_AUX ||
+ op == MLX4_CMD_UNMAP_FA)
+ return 1;
+ /* Error on MLX4_CMD_HW2SW_MPT is fatal except when fw status equals
+ * CMD_STAT_REG_BOUND.
+ * This status indicates that memory region has memory windows bound to it
+ * which may result from invalid user space usage and is not fatal.
+ */
+ if (op == MLX4_CMD_HW2SW_MPT && fw_status != CMD_STAT_REG_BOUND)
+ return 1;
+ return 0;
+}
+
+static int mlx4_cmd_reset_flow(struct mlx4_dev *dev, u16 op, u8 op_modifier,
+ int err)
+{
+ /* Only if reset flow is really active return code is based on
+ * command, otherwise current error code is returned.
+ */
+ if (mlx4_internal_err_reset) {
+ mlx4_enter_error_state(dev->persist);
+ err = mlx4_internal_err_ret_value(dev, op, op_modifier);
+ }
+
+ return err;
+}
+
static int comm_pending(struct mlx4_dev *dev)
{
struct mlx4_priv *priv = mlx4_priv(dev);
@@ -258,7 +324,7 @@ static int mlx4_comm_cmd_wait(struct mlx4_dev *dev, u8 op,
cmd->free_head = context->next;
spin_unlock(&cmd->context_lock);
- init_completion(&context->done);
+ reinit_completion(&context->done);
mlx4_comm_cmd_post(dev, op, param);
@@ -323,17 +389,21 @@ static int mlx4_cmd_post(struct mlx4_dev *dev, u64 in_param, u64 out_param,
{
struct mlx4_cmd *cmd = &mlx4_priv(dev)->cmd;
u32 __iomem *hcr = cmd->hcr;
- int ret = -EAGAIN;
+ int ret = -EIO;
unsigned long end;
- mutex_lock(&cmd->hcr_mutex);
-
- if (pci_channel_offline(dev->persist->pdev)) {
+ mutex_lock(&dev->persist->device_state_mutex);
+ /* To avoid writing to unknown addresses after the device state was
+ * changed to internal error and the chip was reset,
+ * check the INTERNAL_ERROR flag which is updated under
+ * device_state_mutex lock.
+ */
+ if (pci_channel_offline(dev->persist->pdev) ||
+ (dev->persist->state & MLX4_DEVICE_STATE_INTERNAL_ERROR)) {
/*
* Device is going through error recovery
* and cannot accept commands.
*/
- ret = -EIO;
goto out;
}
@@ -347,7 +417,6 @@ static int mlx4_cmd_post(struct mlx4_dev *dev, u64 in_param, u64 out_param,
* Device is going through error recovery
* and cannot accept commands.
*/
- ret = -EIO;
goto out;
}
@@ -391,7 +460,11 @@ static int mlx4_cmd_post(struct mlx4_dev *dev, u64 in_param, u64 out_param,
ret = 0;
out:
- mutex_unlock(&cmd->hcr_mutex);
+ if (ret)
+ mlx4_warn(dev, "Could not post command 0x%x: ret=%d, in_param=0x%llx, in_mod=0x%x, op_mod=0x%x\n",
+ op, ret, in_param, in_modifier, op_modifier);
+ mutex_unlock(&dev->persist->device_state_mutex);
+
return ret;
}
@@ -464,12 +537,12 @@ static int mlx4_cmd_poll(struct mlx4_dev *dev, u64 in_param, u64 *out_param,
down(&priv->cmd.poll_sem);
- if (pci_channel_offline(dev->persist->pdev)) {
+ if (dev->persist->state & MLX4_DEVICE_STATE_INTERNAL_ERROR) {
/*
* Device is going through error recovery
* and cannot accept commands.
*/
- err = -EIO;
+ err = mlx4_internal_err_ret_value(dev, op, op_modifier);
goto out;
}
@@ -483,7 +556,7 @@ static int mlx4_cmd_poll(struct mlx4_dev *dev, u64 in_param, u64 *out_param,
err = mlx4_cmd_post(dev, in_param, out_param ? *out_param : 0,
in_modifier, op_modifier, op, CMD_POLL_TOKEN, 0);
if (err)
- goto out;
+ goto out_reset;
end = msecs_to_jiffies(timeout) + jiffies;
while (cmd_pending(dev) && time_before(jiffies, end)) {
@@ -493,6 +566,11 @@ static int mlx4_cmd_poll(struct mlx4_dev *dev, u64 in_param, u64 *out_param,
* and cannot accept commands.
*/
err = -EIO;
+ goto out_reset;
+ }
+
+ if (dev->persist->state & MLX4_DEVICE_STATE_INTERNAL_ERROR) {
+ err = mlx4_internal_err_ret_value(dev, op, op_modifier);
goto out;
}
@@ -502,8 +580,8 @@ static int mlx4_cmd_poll(struct mlx4_dev *dev, u64 in_param, u64 *out_param,
if (cmd_pending(dev)) {
mlx4_warn(dev, "command 0x%x timed out (go bit not cleared)\n",
op);
- err = -ETIMEDOUT;
- goto out;
+ err = -EIO;
+ goto out_reset;
}
if (out_is_imm)
@@ -515,10 +593,17 @@ static int mlx4_cmd_poll(struct mlx4_dev *dev, u64 in_param, u64 *out_param,
stat = be32_to_cpu((__force __be32)
__raw_readl(hcr + HCR_STATUS_OFFSET)) >> 24;
err = mlx4_status_to_errno(stat);
- if (err)
+ if (err) {
mlx4_err(dev, "command 0x%x failed: fw status = 0x%x\n",
op, stat);
+ if (mlx4_closing_cmd_fatal_error(op, stat))
+ goto out_reset;
+ goto out;
+ }
+out_reset:
+ if (err)
+ err = mlx4_cmd_reset_flow(dev, op, op_modifier, err);
out:
up(&priv->cmd.poll_sem);
return err;
@@ -565,17 +650,19 @@ static int mlx4_cmd_wait(struct mlx4_dev *dev, u64 in_param, u64 *out_param,
goto out;
}
- init_completion(&context->done);
+ reinit_completion(&context->done);
- mlx4_cmd_post(dev, in_param, out_param ? *out_param : 0,
- in_modifier, op_modifier, op, context->token, 1);
+ err = mlx4_cmd_post(dev, in_param, out_param ? *out_param : 0,
+ in_modifier, op_modifier, op, context->token, 1);
+ if (err)
+ goto out_reset;
if (!wait_for_completion_timeout(&context->done,
msecs_to_jiffies(timeout))) {
mlx4_warn(dev, "command 0x%x timed out (go bit not cleared)\n",
op);
- err = -EBUSY;
- goto out;
+ err = -EIO;
+ goto out_reset;
}
err = context->result;
@@ -592,12 +679,20 @@ static int mlx4_cmd_wait(struct mlx4_dev *dev, u64 in_param, u64 *out_param,
else
mlx4_err(dev, "command 0x%x failed: fw status = 0x%x\n",
op, context->fw_status);
+ if (dev->persist->state & MLX4_DEVICE_STATE_INTERNAL_ERROR)
+ err = mlx4_internal_err_ret_value(dev, op, op_modifier);
+ else if (mlx4_closing_cmd_fatal_error(op, context->fw_status))
+ goto out_reset;
+
goto out;
}
if (out_is_imm)
*out_param = context->out_param;
+out_reset:
+ if (err)
+ err = mlx4_cmd_reset_flow(dev, op, op_modifier, err);
out:
spin_lock(&cmd->context_lock);
context->next = cmd->free_head;
@@ -613,9 +708,12 @@ int __mlx4_cmd(struct mlx4_dev *dev, u64 in_param, u64 *out_param,
u16 op, unsigned long timeout, int native)
{
if (pci_channel_offline(dev->persist->pdev))
- return -EIO;
+ return mlx4_cmd_reset_flow(dev, op, op_modifier, -EIO);
if (!mlx4_is_mfunc(dev) || (native && mlx4_is_master(dev))) {
+ if (dev->persist->state & MLX4_DEVICE_STATE_INTERNAL_ERROR)
+ return mlx4_internal_err_ret_value(dev, op,
+ op_modifier);
if (mlx4_priv(dev)->cmd.use_events)
return mlx4_cmd_wait(dev, in_param, out_param,
out_is_imm, in_modifier,
@@ -2121,7 +2219,6 @@ int mlx4_cmd_init(struct mlx4_dev *dev)
int flags = 0;
if (!priv->cmd.initialized) {
- mutex_init(&priv->cmd.hcr_mutex);
mutex_init(&priv->cmd.slave_cmd_mutex);
sema_init(&priv->cmd.poll_sem, 1);
priv->cmd.use_events = 0;
@@ -2232,6 +2329,11 @@ int mlx4_cmd_use_events(struct mlx4_dev *dev)
for (i = 0; i < priv->cmd.max_cmds; ++i) {
priv->cmd.context[i].token = i;
priv->cmd.context[i].next = i + 1;
+ /* To support fatal error flow, initialize all
+ * cmd contexts to allow simulating completions
+ * with complete() at any time.
+ */
+ init_completion(&priv->cmd.context[i].done);
}
priv->cmd.context[priv->cmd.max_cmds - 1].next = -1;
@@ -2329,6 +2431,25 @@ int mlx4_get_vf_indx(struct mlx4_dev *dev, int slave)
return slave - 1;
}
+void mlx4_cmd_wake_completions(struct mlx4_dev *dev)
+{
+ struct mlx4_priv *priv = mlx4_priv(dev);
+ struct mlx4_cmd_context *context;
+ int i;
+
+ spin_lock(&priv->cmd.context_lock);
+ if (priv->cmd.context) {
+ for (i = 0; i < priv->cmd.max_cmds; ++i) {
+ context = &priv->cmd.context[i];
+ context->fw_status = CMD_STAT_INTERNAL_ERR;
+ context->result =
+ mlx4_status_to_errno(CMD_STAT_INTERNAL_ERR);
+ complete(&context->done);
+ }
+ }
+ spin_unlock(&priv->cmd.context_lock);
+}
+
struct mlx4_active_ports mlx4_get_active_ports(struct mlx4_dev *dev, int slave)
{
struct mlx4_active_ports actv_ports;
diff --git a/drivers/net/ethernet/mellanox/mlx4/mcg.c b/drivers/net/ethernet/mellanox/mlx4/mcg.c
index a3867e7..d22d928 100644
--- a/drivers/net/ethernet/mellanox/mlx4/mcg.c
+++ b/drivers/net/ethernet/mellanox/mlx4/mcg.c
@@ -1318,6 +1318,9 @@ out:
mutex_unlock(&priv->mcg_table.mutex);
mlx4_free_cmd_mailbox(dev, mailbox);
+ if (err && dev->persist->state & MLX4_DEVICE_STATE_INTERNAL_ERROR)
+ /* In case device is under an error, return success as a closing command */
+ err = 0;
return err;
}
diff --git a/drivers/net/ethernet/mellanox/mlx4/mlx4.h b/drivers/net/ethernet/mellanox/mlx4/mlx4.h
index aa1ecbc..5c772ea 100644
--- a/drivers/net/ethernet/mellanox/mlx4/mlx4.h
+++ b/drivers/net/ethernet/mellanox/mlx4/mlx4.h
@@ -235,6 +235,7 @@ do { \
extern int mlx4_log_num_mgm_entry_size;
extern int log_mtts_per_seg;
+extern int mlx4_internal_err_reset;
#define MLX4_MAX_NUM_SLAVES (MLX4_MAX_NUM_PF + MLX4_MAX_NUM_VF)
#define ALL_SLAVES 0xff
@@ -607,7 +608,6 @@ struct mlx4_mgm {
struct mlx4_cmd {
struct pci_pool *pool;
void __iomem *hcr;
- struct mutex hcr_mutex;
struct mutex slave_cmd_mutex;
struct semaphore poll_sem;
struct semaphore event_sem;
diff --git a/include/linux/mlx4/cmd.h b/include/linux/mlx4/cmd.h
index 64d2594..e754384 100644
--- a/include/linux/mlx4/cmd.h
+++ b/include/linux/mlx4/cmd.h
@@ -279,6 +279,7 @@ int mlx4_get_vf_config(struct mlx4_dev *dev, int port, int vf, struct ifla_vf_in
int mlx4_set_vf_link_state(struct mlx4_dev *dev, int port, int vf, int link_state);
int mlx4_config_dev_retrieval(struct mlx4_dev *dev,
struct mlx4_config_dev_params *params);
+void mlx4_cmd_wake_completions(struct mlx4_dev *dev);
/*
* mlx4_get_slave_default_vlan -
* return true if VST ( default vlan)
--
1.7.1
^ permalink raw reply related
* [PATCH net-next 7/9] net/mlx4_core: Handle AER flow properly
From: Or Gerlitz @ 2015-01-21 14:45 UTC (permalink / raw)
To: David S. Miller
Cc: netdev, Matan Barak, Amir Vadai, Tal Alon, Roland Dreier,
Yishai Hadas, Or Gerlitz
In-Reply-To: <1421851557-12084-1-git-send-email-ogerlitz@mellanox.com>
From: Yishai Hadas <yishaih@mellanox.com>
Fix AER callbacks to work properly, it includes:
- Refractoring AER to be aligned with Reset flow support.
- Sync with concurrent catas flow.
In addition, fix the shutdown PCI callback to sync with
concurrent catas flow.
Signed-off-by: Yishai Hadas <yishaih@mellanox.com>
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
---
drivers/net/ethernet/mellanox/mlx4/main.c | 66 +++++++++++++++++++++++++++--
1 files changed, 62 insertions(+), 4 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx4/main.c b/drivers/net/ethernet/mellanox/mlx4/main.c
index d59cae5..01d12f7 100644
--- a/drivers/net/ethernet/mellanox/mlx4/main.c
+++ b/drivers/net/ethernet/mellanox/mlx4/main.c
@@ -3120,7 +3120,10 @@ static int mlx4_init_one(struct pci_dev *pdev, const struct pci_device_id *id)
if (ret) {
kfree(dev->persist);
kfree(priv);
+ } else {
+ pci_save_state(pdev);
}
+
return ret;
}
@@ -3351,7 +3354,17 @@ MODULE_DEVICE_TABLE(pci, mlx4_pci_table);
static pci_ers_result_t mlx4_pci_err_detected(struct pci_dev *pdev,
pci_channel_state_t state)
{
- mlx4_unload_one(pdev);
+ struct mlx4_dev_persistent *persist = pci_get_drvdata(pdev);
+
+ mlx4_err(persist->dev, "mlx4_pci_err_detected was called\n");
+ mlx4_enter_error_state(persist);
+
+ mutex_lock(&persist->interface_state_mutex);
+ if (persist->interface_state & MLX4_INTERFACE_STATE_UP)
+ mlx4_unload_one(pdev);
+
+ mutex_unlock(&persist->interface_state_mutex);
+ pci_disable_device(pdev);
return state == pci_channel_io_perm_failure ?
PCI_ERS_RESULT_DISCONNECT : PCI_ERS_RESULT_NEED_RESET;
@@ -3359,15 +3372,60 @@ static pci_ers_result_t mlx4_pci_err_detected(struct pci_dev *pdev,
static pci_ers_result_t mlx4_pci_slot_reset(struct pci_dev *pdev)
{
- struct mlx4_dev *dev = pci_get_drvdata(pdev);
+ struct mlx4_dev_persistent *persist = pci_get_drvdata(pdev);
+ struct mlx4_dev *dev = persist->dev;
struct mlx4_priv *priv = mlx4_priv(dev);
int ret;
+ int nvfs[MLX4_MAX_PORTS + 1] = {0, 0, 0};
+ int total_vfs;
- ret = __mlx4_init_one(pdev, priv->pci_dev_data, priv);
+ mlx4_err(dev, "mlx4_pci_slot_reset was called\n");
+ ret = pci_enable_device(pdev);
+ if (ret) {
+ mlx4_err(dev, "Can not re-enable device, ret=%d\n", ret);
+ return PCI_ERS_RESULT_DISCONNECT;
+ }
+
+ pci_set_master(pdev);
+ pci_restore_state(pdev);
+ pci_save_state(pdev);
+
+ total_vfs = dev->persist->num_vfs;
+ memcpy(nvfs, dev->persist->nvfs, sizeof(dev->persist->nvfs));
+
+ mutex_lock(&persist->interface_state_mutex);
+ if (!(persist->interface_state & MLX4_INTERFACE_STATE_UP)) {
+ ret = mlx4_load_one(pdev, priv->pci_dev_data, total_vfs, nvfs,
+ priv);
+ if (ret) {
+ mlx4_err(dev, "%s: mlx4_load_one failed, ret=%d\n",
+ __func__, ret);
+ goto end;
+ }
+
+ ret = restore_current_port_types(dev, dev->persist->
+ curr_port_type, dev->persist->
+ curr_port_poss_type);
+ if (ret)
+ mlx4_err(dev, "could not restore original port types (%d)\n", ret);
+ }
+end:
+ mutex_unlock(&persist->interface_state_mutex);
return ret ? PCI_ERS_RESULT_DISCONNECT : PCI_ERS_RESULT_RECOVERED;
}
+static void mlx4_shutdown(struct pci_dev *pdev)
+{
+ struct mlx4_dev_persistent *persist = pci_get_drvdata(pdev);
+
+ mlx4_info(persist->dev, "mlx4_shutdown was called\n");
+ mutex_lock(&persist->interface_state_mutex);
+ if (persist->interface_state & MLX4_INTERFACE_STATE_UP)
+ mlx4_unload_one(pdev);
+ mutex_unlock(&persist->interface_state_mutex);
+}
+
static const struct pci_error_handlers mlx4_err_handler = {
.error_detected = mlx4_pci_err_detected,
.slot_reset = mlx4_pci_slot_reset,
@@ -3377,7 +3435,7 @@ static struct pci_driver mlx4_driver = {
.name = DRV_NAME,
.id_table = mlx4_pci_table,
.probe = mlx4_init_one,
- .shutdown = mlx4_unload_one,
+ .shutdown = mlx4_shutdown,
.remove = mlx4_remove_one,
.err_handler = &mlx4_err_handler,
};
--
1.7.1
^ permalink raw reply related
* [PATCH net-next 9/9] net/mlx4_core: Reset flow activation upon SRIOV fatal command cases
From: Or Gerlitz @ 2015-01-21 14:45 UTC (permalink / raw)
To: David S. Miller
Cc: netdev, Matan Barak, Amir Vadai, Tal Alon, Roland Dreier,
Yishai Hadas, Or Gerlitz
In-Reply-To: <1421851557-12084-1-git-send-email-ogerlitz@mellanox.com>
From: Yishai Hadas <yishaih@mellanox.com>
When SRIOV commands are executed over the comm-channel and get
a fatal error (e.g. timeout, closing command failure) the VF enters
into error state and reset flow is activated.
To be able to recognize whether the failure was on a closing command, the
operational code for the given VHCR command is used. Once the device entered
into an error state we prevent redundant error messages from being printed.
Signed-off-by: Yishai Hadas <yishaih@mellanox.com>
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
---
drivers/net/ethernet/mellanox/mlx4/cmd.c | 125 +++++++++++++++++++++--------
drivers/net/ethernet/mellanox/mlx4/main.c | 16 ++--
drivers/net/ethernet/mellanox/mlx4/mcg.c | 3 +
drivers/net/ethernet/mellanox/mlx4/mlx4.h | 4 +-
4 files changed, 107 insertions(+), 41 deletions(-)
diff --git a/drivers/net/ethernet/mellanox/mlx4/cmd.c b/drivers/net/ethernet/mellanox/mlx4/cmd.c
index e3d6f7f..6e229a2 100644
--- a/drivers/net/ethernet/mellanox/mlx4/cmd.c
+++ b/drivers/net/ethernet/mellanox/mlx4/cmd.c
@@ -257,16 +257,30 @@ static int comm_pending(struct mlx4_dev *dev)
return (swab32(status) >> 31) != priv->cmd.comm_toggle;
}
-static void mlx4_comm_cmd_post(struct mlx4_dev *dev, u8 cmd, u16 param)
+static int mlx4_comm_cmd_post(struct mlx4_dev *dev, u8 cmd, u16 param)
{
struct mlx4_priv *priv = mlx4_priv(dev);
u32 val;
+ /* To avoid writing to unknown addresses after the device state was
+ * changed to internal error and the function was rest,
+ * check the INTERNAL_ERROR flag which is updated under
+ * device_state_mutex lock.
+ */
+ mutex_lock(&dev->persist->device_state_mutex);
+
+ if (dev->persist->state & MLX4_DEVICE_STATE_INTERNAL_ERROR) {
+ mutex_unlock(&dev->persist->device_state_mutex);
+ return -EIO;
+ }
+
priv->cmd.comm_toggle ^= 1;
val = param | (cmd << 16) | (priv->cmd.comm_toggle << 31);
__raw_writel((__force u32) cpu_to_be32(val),
&priv->mfunc.comm->slave_write);
mmiowb();
+ mutex_unlock(&dev->persist->device_state_mutex);
+ return 0;
}
static int mlx4_comm_cmd_poll(struct mlx4_dev *dev, u8 cmd, u16 param,
@@ -286,7 +300,13 @@ static int mlx4_comm_cmd_poll(struct mlx4_dev *dev, u8 cmd, u16 param,
/* Write command */
down(&priv->cmd.poll_sem);
- mlx4_comm_cmd_post(dev, cmd, param);
+ if (mlx4_comm_cmd_post(dev, cmd, param)) {
+ /* Only in case the device state is INTERNAL_ERROR,
+ * mlx4_comm_cmd_post returns with an error
+ */
+ err = mlx4_status_to_errno(CMD_STAT_INTERNAL_ERR);
+ goto out;
+ }
end = msecs_to_jiffies(timeout) + jiffies;
while (comm_pending(dev) && time_before(jiffies, end))
@@ -298,18 +318,23 @@ static int mlx4_comm_cmd_poll(struct mlx4_dev *dev, u8 cmd, u16 param,
* is MLX4_DELAY_RESET_SLAVE*/
if ((MLX4_COMM_CMD_RESET == cmd)) {
err = MLX4_DELAY_RESET_SLAVE;
+ goto out;
} else {
- mlx4_warn(dev, "Communication channel timed out\n");
- err = -ETIMEDOUT;
+ mlx4_warn(dev, "Communication channel command 0x%x timed out\n",
+ cmd);
+ err = mlx4_status_to_errno(CMD_STAT_INTERNAL_ERR);
}
}
+ if (err)
+ mlx4_enter_error_state(dev->persist);
+out:
up(&priv->cmd.poll_sem);
return err;
}
-static int mlx4_comm_cmd_wait(struct mlx4_dev *dev, u8 op,
- u16 param, unsigned long timeout)
+static int mlx4_comm_cmd_wait(struct mlx4_dev *dev, u8 vhcr_cmd,
+ u16 param, u16 op, unsigned long timeout)
{
struct mlx4_cmd *cmd = &mlx4_priv(dev)->cmd;
struct mlx4_cmd_context *context;
@@ -327,32 +352,47 @@ static int mlx4_comm_cmd_wait(struct mlx4_dev *dev, u8 op,
reinit_completion(&context->done);
- mlx4_comm_cmd_post(dev, op, param);
+ if (mlx4_comm_cmd_post(dev, vhcr_cmd, param)) {
+ /* Only in case the device state is INTERNAL_ERROR,
+ * mlx4_comm_cmd_post returns with an error
+ */
+ err = mlx4_status_to_errno(CMD_STAT_INTERNAL_ERR);
+ goto out;
+ }
if (!wait_for_completion_timeout(&context->done,
msecs_to_jiffies(timeout))) {
- mlx4_warn(dev, "communication channel command 0x%x timed out\n",
- op);
- err = -EBUSY;
- goto out;
+ mlx4_warn(dev, "communication channel command 0x%x (op=0x%x) timed out\n",
+ vhcr_cmd, op);
+ goto out_reset;
}
err = context->result;
if (err && context->fw_status != CMD_STAT_MULTI_FUNC_REQ) {
mlx4_err(dev, "command 0x%x failed: fw status = 0x%x\n",
- op, context->fw_status);
- goto out;
+ vhcr_cmd, context->fw_status);
+ if (mlx4_closing_cmd_fatal_error(op, context->fw_status))
+ goto out_reset;
}
-out:
/* wait for comm channel ready
* this is necessary for prevention the race
* when switching between event to polling mode
+ * Skipping this section in case the device is in FATAL_ERROR state,
+ * In this state, no commands are sent via the comm channel until
+ * the device has returned from reset.
*/
- end = msecs_to_jiffies(timeout) + jiffies;
- while (comm_pending(dev) && time_before(jiffies, end))
- cond_resched();
+ if (!(dev->persist->state & MLX4_DEVICE_STATE_INTERNAL_ERROR)) {
+ end = msecs_to_jiffies(timeout) + jiffies;
+ while (comm_pending(dev) && time_before(jiffies, end))
+ cond_resched();
+ }
+ goto out;
+out_reset:
+ err = mlx4_status_to_errno(CMD_STAT_INTERNAL_ERR);
+ mlx4_enter_error_state(dev->persist);
+out:
spin_lock(&cmd->context_lock);
context->next = cmd->free_head;
cmd->free_head = context - cmd->context;
@@ -363,10 +403,13 @@ out:
}
int mlx4_comm_cmd(struct mlx4_dev *dev, u8 cmd, u16 param,
- unsigned long timeout)
+ u16 op, unsigned long timeout)
{
+ if (dev->persist->state & MLX4_DEVICE_STATE_INTERNAL_ERROR)
+ return mlx4_status_to_errno(CMD_STAT_INTERNAL_ERR);
+
if (mlx4_priv(dev)->cmd.use_events)
- return mlx4_comm_cmd_wait(dev, cmd, param, timeout);
+ return mlx4_comm_cmd_wait(dev, cmd, param, op, timeout);
return mlx4_comm_cmd_poll(dev, cmd, param, timeout);
}
@@ -502,8 +545,11 @@ static int mlx4_slave_cmd(struct mlx4_dev *dev, u64 in_param, u64 *out_param,
}
ret = mlx4_status_to_errno(vhcr->status);
}
+ if (ret &&
+ dev->persist->state & MLX4_DEVICE_STATE_INTERNAL_ERROR)
+ ret = mlx4_internal_err_ret_value(dev, op, op_modifier);
} else {
- ret = mlx4_comm_cmd(dev, MLX4_COMM_CMD_VHCR_POST, 0,
+ ret = mlx4_comm_cmd(dev, MLX4_COMM_CMD_VHCR_POST, 0, op,
MLX4_COMM_TIME + timeout);
if (!ret) {
if (out_is_imm) {
@@ -517,9 +563,14 @@ static int mlx4_slave_cmd(struct mlx4_dev *dev, u64 in_param, u64 *out_param,
}
}
ret = mlx4_status_to_errno(vhcr->status);
- } else
- mlx4_err(dev, "failed execution of VHCR_POST command opcode 0x%x\n",
- op);
+ } else {
+ if (dev->persist->state &
+ MLX4_DEVICE_STATE_INTERNAL_ERROR)
+ ret = mlx4_internal_err_ret_value(dev, op,
+ op_modifier);
+ else
+ mlx4_err(dev, "failed execution of VHCR_POST command opcode 0x%x\n", op);
+ }
}
mutex_unlock(&priv->cmd.slave_cmd_mutex);
@@ -1559,8 +1610,10 @@ static int mlx4_master_process_vhcr(struct mlx4_dev *dev, int slave,
ALIGN(sizeof(struct mlx4_vhcr_cmd),
MLX4_ACCESS_MEM_ALIGN), 1);
if (ret) {
- mlx4_err(dev, "%s: Failed reading vhcr ret: 0x%x\n",
- __func__, ret);
+ if (!(dev->persist->state &
+ MLX4_DEVICE_STATE_INTERNAL_ERROR))
+ mlx4_err(dev, "%s: Failed reading vhcr ret: 0x%x\n",
+ __func__, ret);
kfree(vhcr);
return ret;
}
@@ -1599,11 +1652,14 @@ static int mlx4_master_process_vhcr(struct mlx4_dev *dev, int slave,
goto out_status;
}
- if (mlx4_ACCESS_MEM(dev, inbox->dma, slave,
- vhcr->in_param,
- MLX4_MAILBOX_SIZE, 1)) {
- mlx4_err(dev, "%s: Failed reading inbox (cmd:0x%x)\n",
- __func__, cmd->opcode);
+ ret = mlx4_ACCESS_MEM(dev, inbox->dma, slave,
+ vhcr->in_param,
+ MLX4_MAILBOX_SIZE, 1);
+ if (ret) {
+ if (!(dev->persist->state &
+ MLX4_DEVICE_STATE_INTERNAL_ERROR))
+ mlx4_err(dev, "%s: Failed reading inbox (cmd:0x%x)\n",
+ __func__, cmd->opcode);
vhcr_cmd->status = CMD_STAT_INTERNAL_ERR;
goto out_status;
}
@@ -1651,8 +1707,9 @@ static int mlx4_master_process_vhcr(struct mlx4_dev *dev, int slave,
}
if (err) {
- mlx4_warn(dev, "vhcr command:0x%x slave:%d failed with error:%d, status %d\n",
- vhcr->op, slave, vhcr->errno, err);
+ if (!(dev->persist->state & MLX4_DEVICE_STATE_INTERNAL_ERROR))
+ mlx4_warn(dev, "vhcr command:0x%x slave:%d failed with error:%d, status %d\n",
+ vhcr->op, slave, vhcr->errno, err);
vhcr_cmd->status = mlx4_errno_to_status(err);
goto out_status;
}
@@ -1667,7 +1724,9 @@ static int mlx4_master_process_vhcr(struct mlx4_dev *dev, int slave,
/* If we failed to write back the outbox after the
*command was successfully executed, we must fail this
* slave, as it is now in undefined state */
- mlx4_err(dev, "%s:Failed writing outbox\n", __func__);
+ if (!(dev->persist->state &
+ MLX4_DEVICE_STATE_INTERNAL_ERROR))
+ mlx4_err(dev, "%s:Failed writing outbox\n", __func__);
goto out;
}
}
diff --git a/drivers/net/ethernet/mellanox/mlx4/main.c b/drivers/net/ethernet/mellanox/mlx4/main.c
index 496d6e7..75f8133 100644
--- a/drivers/net/ethernet/mellanox/mlx4/main.c
+++ b/drivers/net/ethernet/mellanox/mlx4/main.c
@@ -1484,7 +1484,8 @@ static void mlx4_slave_exit(struct mlx4_dev *dev)
struct mlx4_priv *priv = mlx4_priv(dev);
mutex_lock(&priv->cmd.slave_cmd_mutex);
- if (mlx4_comm_cmd(dev, MLX4_COMM_CMD_RESET, 0, MLX4_COMM_TIME))
+ if (mlx4_comm_cmd(dev, MLX4_COMM_CMD_RESET, 0, MLX4_COMM_CMD_NA_OP,
+ MLX4_COMM_TIME))
mlx4_warn(dev, "Failed to close slave function\n");
mutex_unlock(&priv->cmd.slave_cmd_mutex);
}
@@ -1648,7 +1649,7 @@ static int mlx4_init_slave(struct mlx4_dev *dev)
mlx4_reset_vf_support(dev);
mlx4_warn(dev, "Sending reset\n");
ret_from_reset = mlx4_comm_cmd(dev, MLX4_COMM_CMD_RESET, 0,
- MLX4_COMM_TIME);
+ MLX4_COMM_CMD_NA_OP, MLX4_COMM_TIME);
/* if we are in the middle of flr the slave will try
* NUM_OF_RESET_RETRIES times before leaving.*/
if (ret_from_reset) {
@@ -1673,22 +1674,23 @@ static int mlx4_init_slave(struct mlx4_dev *dev)
mlx4_warn(dev, "Sending vhcr0\n");
if (mlx4_comm_cmd(dev, MLX4_COMM_CMD_VHCR0, dma >> 48,
- MLX4_COMM_TIME))
+ MLX4_COMM_CMD_NA_OP, MLX4_COMM_TIME))
goto err;
if (mlx4_comm_cmd(dev, MLX4_COMM_CMD_VHCR1, dma >> 32,
- MLX4_COMM_TIME))
+ MLX4_COMM_CMD_NA_OP, MLX4_COMM_TIME))
goto err;
if (mlx4_comm_cmd(dev, MLX4_COMM_CMD_VHCR2, dma >> 16,
- MLX4_COMM_TIME))
+ MLX4_COMM_CMD_NA_OP, MLX4_COMM_TIME))
goto err;
- if (mlx4_comm_cmd(dev, MLX4_COMM_CMD_VHCR_EN, dma, MLX4_COMM_TIME))
+ if (mlx4_comm_cmd(dev, MLX4_COMM_CMD_VHCR_EN, dma,
+ MLX4_COMM_CMD_NA_OP, MLX4_COMM_TIME))
goto err;
mutex_unlock(&priv->cmd.slave_cmd_mutex);
return 0;
err:
- mlx4_comm_cmd(dev, MLX4_COMM_CMD_RESET, 0, 0);
+ mlx4_comm_cmd(dev, MLX4_COMM_CMD_RESET, 0, MLX4_COMM_CMD_NA_OP, 0);
err_offline:
mutex_unlock(&priv->cmd.slave_cmd_mutex);
return -EIO;
diff --git a/drivers/net/ethernet/mellanox/mlx4/mcg.c b/drivers/net/ethernet/mellanox/mlx4/mcg.c
index d22d928..bd9ea0d 100644
--- a/drivers/net/ethernet/mellanox/mlx4/mcg.c
+++ b/drivers/net/ethernet/mellanox/mlx4/mcg.c
@@ -1350,6 +1350,9 @@ static int mlx4_QP_ATTACH(struct mlx4_dev *dev, struct mlx4_qp *qp,
MLX4_CMD_WRAPPED);
mlx4_free_cmd_mailbox(dev, mailbox);
+ if (err && !attach &&
+ dev->persist->state & MLX4_DEVICE_STATE_INTERNAL_ERROR)
+ err = 0;
return err;
}
diff --git a/drivers/net/ethernet/mellanox/mlx4/mlx4.h b/drivers/net/ethernet/mellanox/mlx4/mlx4.h
index 2a15b82..096a81c 100644
--- a/drivers/net/ethernet/mellanox/mlx4/mlx4.h
+++ b/drivers/net/ethernet/mellanox/mlx4/mlx4.h
@@ -123,6 +123,8 @@ enum mlx4_mpt_state {
#define MLX4_COMM_TIME 10000
#define MLX4_COMM_OFFLINE_TIME_OUT 30000
+#define MLX4_COMM_CMD_NA_OP 0x0
+
enum {
MLX4_COMM_CMD_RESET,
@@ -1173,7 +1175,7 @@ int mlx4_cmd_use_events(struct mlx4_dev *dev);
void mlx4_cmd_use_polling(struct mlx4_dev *dev);
int mlx4_comm_cmd(struct mlx4_dev *dev, u8 cmd, u16 param,
- unsigned long timeout);
+ u16 op, unsigned long timeout);
void mlx4_cq_tasklet_cb(unsigned long data);
void mlx4_cq_completion(struct mlx4_dev *dev, u32 cqn);
--
1.7.1
^ permalink raw reply related
* Re: [PATCH v5 2/5] can: kvaser_usb: Consolidate and unify state change handling
From: Andri Yngvason @ 2015-01-21 15:00 UTC (permalink / raw)
To: Ahmed S. Darwish, Wolfgang Grandegger
Cc: Olivier Sobrie, Oliver Hartkopp, Marc Kleine-Budde, Linux-CAN,
netdev, LKML
In-Reply-To: <20150121144323.GA17070@linux>
Quoting Ahmed S. Darwish (2015-01-21 14:43:23)
> Hi!
>
> On Wed, Jan 21, 2015 at 12:53:58PM +0100, Wolfgang Grandegger wrote:
> > On Wed, 21 Jan 2015 10:33:19 +0000, Andri Yngvason
> > <andri.yngvason@marel.com> wrote:
> > > Quoting Ahmed S. Darwish (2015-01-20 21:45:37)
> > >> From: Ahmed S. Darwish <ahmed.darwish@valeo.com>
> > >>
> > >> Replace most of the can interface's state and error counters
> > >> handling with the new can-dev can_change_state() mechanism.
> > >>
> > >> Suggested-by: Andri Yngvason <andri.yngvason@marel.com>
> > >> Signed-off-by: Ahmed S. Darwish <ahmed.darwish@valeo.com>
> > >> ---
> > >> drivers/net/can/usb/kvaser_usb.c | 114
> > >> +++++++++++++++++++--------------------
> > >> 1 file changed, 55 insertions(+), 59 deletions(-)
> > >>
> > >> diff --git a/drivers/net/can/usb/kvaser_usb.c
> > >> b/drivers/net/can/usb/kvaser_usb.c
> > >> index 971c5f9..0386d3f 100644
> > >> --- a/drivers/net/can/usb/kvaser_usb.c
> > >> +++ b/drivers/net/can/usb/kvaser_usb.c
> >
> > ...
> > >
> > > Looks good.
> >
> > Would be nice to see some "candump" traces as well.
>
> Sure. The USBCan-II device trace below is generated after applying
> all patches in the series, especially patch #3, which fixes some
> some invalid CAN state transitions logic in the original driver.
[...]
>
> Afterwards, candump on a PC, Kvaser USB device on the sending end:
>
> ...
> (000.008784) can0 60A [1] C1
> (000.011341) can0 2A8 [8] C2 0A 00 00 00 00 00 00
> (000.009873) can0 03D [7] C3 0A 00 00 00 00 00
> (000.010394) can0 55C [8] C4 0A 00 00 00 00 00 00
> (000.009979) can0 45A [8] C5 0A 00 00 00 00 00 00
> (000.010125) can0 6E8 [8] C6 0A 00 00 00 00 00 00
> (000.010149) can0 4EE [8] C7 0A 00 00 00 00 00 00
> (000.010102) can0 5D2 [8] C8 0A 00 00 00 00 00 00
> (000.010000) can0 61F [8] C9 0A 00 00 00 00 00 00
> (000.010271) can0 5F8 [8] CA 0A 00 00 00 00 00 00
>
> <-- Unplug the cable -->
>
> (000.009106) can0 20000080 [8] 00 00 00 00 00 00 08 00 ERRORFRAME
> bus-error
> error-counter-tx-rx{{8}{0}}
> (000.001872) can0 20000080 [8] 00 00 00 00 00 00 10 00 ERRORFRAME
> bus-error
> error-counter-tx-rx{{16}{0}}
[...]
> error-counter-tx-rx{{80}{0}}
> (000.001910) can0 20000080 [8] 00 00 00 00 00 00 58 00 ERRORFRAME
> bus-error
> error-counter-tx-rx{{88}{0}}
> (000.001753) can0 20000084 [8] 00 08 00 00 00 00 60 00 ERRORFRAME
> controller-problem{tx-error-warning}
Good.
> bus-error
> error-counter-tx-rx{{96}{0}}
> (000.001720) can0 20000080 [8] 00 00 00 00 00 00 68 00 ERRORFRAME
> bus-error
> error-counter-tx-rx{{104}{0}}
> (000.001876) can0 20000080 [8] 00 00 00 00 00 00 70 00 ERRORFRAME
> bus-error
> error-counter-tx-rx{{112}{0}}
> (000.001749) can0 20000080 [8] 00 00 00 00 00 00 78 00 ERRORFRAME
> bus-error
> error-counter-tx-rx{{120}{0}}
> (000.001771) can0 20000084 [8] 00 20 00 00 00 00 80 00 ERRORFRAME
> controller-problem{tx-error-passive}
Also good.
> bus-error
> error-counter-tx-rx{{128}{0}}
> (000.001868) can0 20000080 [8] 00 00 00 00 00 00 80 00 ERRORFRAME
> bus-error
> error-counter-tx-rx{{128}{0}}
> (000.001982) can0 20000080 [8] 00 00 00 00 00 00 80 00 ERRORFRAME
> bus-error
> error-counter-tx-rx{{128}{0}}
>
> (( Then a continous flood, exactly similar to the above packet, appears.
> Unfortunately this flooding is a firmware problem. ))
>
> <-- Replug the cable, after a good amount of time -->
>
Where are the reverse state transitions?
>
> (000.520485) can0 33D [4] CB 0A 00 00
> (000.002693) can0 42E [8] CC 0A 00 00 00 00 00 00
> (000.001795) can0 319 [4] CD 0A 00 00
> (000.002705) can0 3B1 [8] CE 0A 00 00 00 00 00 00
> (000.001295) can0 4CC [2] CF 0A
> (000.002205) can0 42B [6] D0 0A 00 00 00 00
> (000.001620) can0 5A2 [3] D1 0A 00
> (000.002636) can0 691 [8] D2 0A 00 00 00 00 00 00
> (000.002615) can0 36A [8] D3 0A 00 00 00 00 00 00
> (000.001729) can0 068 [4] D4 0A 00 00
> (000.001195) can0 4C8 [1] D5
> ...
>
> ##########################################################################
>
> Bus-off Testing:
>
> candump on a PC, Kvaser device on the sending end. An i.mx6 ARM
> board with flexcan is on the receiving end:
>
> (000.010319) can0 5CC [8] 90 02 00 00 00 00 00 00
> (000.008747) can0 679 [1] 91
> (000.011442) can0 011 [8] 92 02 00 00 00 00 00 00
> (000.008991) can0 631 [2] 93 02
> (000.011097) can0 532 [7] 94 02 00 00 00 00 00
> (000.009781) can0 0A9 [5] 95 02 00 00 00
> (000.010792) can0 1DD [8] 96 02 00 00 00 00 00 00
> (000.010026) can0 44E [8] 97 02 00 00 00 00 00 00
> (000.010181) can0 76A [8] 98 02 00 00 00 00 00 00
> (000.008867) can0 1A5 [1] 99
> (000.011322) can0 2B4 [8] 9A 02 00 00 00 00 00 00
>
> <-- Unplug the can low and high wires from the board -->
>
> (000.009688) can0 20000080 [8] 00 00 00 00 00 00 08 00 ERRORFRAME
> bus-error
> error-counter-tx-rx{{8}{0}}
> (000.002246) can0 20000080 [8] 00 00 00 00 00 00 10 00 ERRORFRAME
> bus-error
> error-counter-tx-rx{{16}{0}}
> (000.002124) can0 20000080 [8] 00 00 00 00 00 00 18 00 ERRORFRAME
> bus-error
> error-counter-tx-rx{{24}{0}}
> (000.002115) can0 20000080 [8] 00 00 00 00 00 00 20 00 ERRORFRAME
> bus-error
> error-counter-tx-rx{{32}{0}}
> (000.002132) can0 20000080 [8] 00 00 00 00 00 00 28 00 ERRORFRAME
> bus-error
> error-counter-tx-rx{{40}{0}}
> (000.002266) can0 20000080 [8] 00 00 00 00 00 00 30 00 ERRORFRAME
> bus-error
> error-counter-tx-rx{{48}{0}}
> (000.002187) can0 20000080 [8] 00 00 00 00 00 00 38 00 ERRORFRAME
> bus-error
> error-counter-tx-rx{{56}{0}}
> (000.002046) can0 20000080 [8] 00 00 00 00 00 00 40 00 ERRORFRAME
> bus-error
> error-counter-tx-rx{{64}{0}}
> (000.002076) can0 20000080 [8] 00 00 00 00 00 00 48 00 ERRORFRAME
> bus-error
> error-counter-tx-rx{{72}{0}}
> (000.002406) can0 20000080 [8] 00 00 00 00 00 00 50 00 ERRORFRAME
> bus-error
> error-counter-tx-rx{{80}{0}}
> (000.001969) can0 20000080 [8] 00 00 00 00 00 00 58 00 ERRORFRAME
> bus-error
> error-counter-tx-rx{{88}{0}}
> (000.002388) can0 20000084 [8] 00 08 00 00 00 00 60 00 ERRORFRAME
> controller-problem{tx-error-warning}
> bus-error
> error-counter-tx-rx{{96}{0}}
> (000.002021) can0 20000080 [8] 00 00 00 00 00 00 68 00 ERRORFRAME
> bus-error
> error-counter-tx-rx{{104}{0}}
> (000.002110) can0 20000080 [8] 00 00 00 00 00 00 70 00 ERRORFRAME
> bus-error
> error-counter-tx-rx{{112}{0}}
> (000.002155) can0 20000080 [8] 00 00 00 00 00 00 78 00 ERRORFRAME
> bus-error
> error-counter-tx-rx{{120}{0}}
> (000.002140) can0 20000084 [8] 00 20 00 00 00 00 80 00 ERRORFRAME
> controller-problem{tx-error-passive}
> bus-error
> error-counter-tx-rx{{128}{0}}
> (000.002324) can0 20000080 [8] 00 00 00 00 00 00 80 00 ERRORFRAME
> bus-error
> error-counter-tx-rx{{128}{0}}
> (000.002416) can0 20000080 [8] 00 00 00 00 00 00 80 00 ERRORFRAME
> bus-error
> error-counter-tx-rx{{128}{0}}
> (000.002237) can0 20000080 [8] 00 00 00 00 00 00 80 00 ERRORFRAME
> bus-error
> error-counter-tx-rx{{128}{0}}
>
> (( Then a continous flood, exactly similar to the above packet, appears ))
>
> <-- Short-circuit the can high and low wires -->
>
> (000.002364) can0 20000080 [8] 00 00 00 00 00 00 80 00 ERRORFRAME
> bus-error
> error-counter-tx-rx{{128}{0}}
> (000.002108) can0 20000080 [8] 00 00 00 00 00 00 88 00 ERRORFRAME
> bus-error
> error-counter-tx-rx{{136}{0}}
> (000.000494) can0 20000080 [8] 00 00 00 00 00 00 90 00 ERRORFRAME
> bus-error
> error-counter-tx-rx{{144}{0}}
> (000.000523) can0 20000080 [8] 00 00 00 00 00 00 98 00 ERRORFRAME
> bus-error
> error-counter-tx-rx{{152}{0}}
> (000.000661) can0 20000080 [8] 00 00 00 00 00 00 A0 00 ERRORFRAME
> bus-error
> error-counter-tx-rx{{160}{0}}
> (000.000464) can0 20000080 [8] 00 00 00 00 00 00 A8 00 ERRORFRAME
> bus-error
> error-counter-tx-rx{{168}{0}}
> (000.000534) can0 20000080 [8] 00 00 00 00 00 00 B0 00 ERRORFRAME
> bus-error
> error-counter-tx-rx{{176}{0}}
> (000.000499) can0 20000080 [8] 00 00 00 00 00 00 B8 00 ERRORFRAME
> bus-error
> error-counter-tx-rx{{184}{0}}
> (000.000626) can0 20000080 [8] 00 00 00 00 00 00 C0 00 ERRORFRAME
> bus-error
> error-counter-tx-rx{{192}{0}}
> (000.000373) can0 20000080 [8] 00 00 00 00 00 00 C8 00 ERRORFRAME
> bus-error
> error-counter-tx-rx{{200}{0}}
> (000.000627) can0 20000080 [8] 00 00 00 00 00 00 D0 00 ERRORFRAME
> bus-error
> error-counter-tx-rx{{208}{0}}
> (000.000507) can0 20000080 [8] 00 00 00 00 00 00 D8 00 ERRORFRAME
> bus-error
> error-counter-tx-rx{{216}{0}}
> (000.000501) can0 20000080 [8] 00 00 00 00 00 00 E0 00 ERRORFRAME
> bus-error
> error-counter-tx-rx{{224}{0}}
> (000.000459) can0 20000080 [8] 00 00 00 00 00 00 E8 00 ERRORFRAME
> bus-error
> error-counter-tx-rx{{232}{0}}
> (000.000606) can0 20000080 [8] 00 00 00 00 00 00 F0 00 ERRORFRAME
> bus-error
> error-counter-tx-rx{{240}{0}}
> (000.000454) can0 20000080 [8] 00 00 00 00 00 00 F8 00 ERRORFRAME
> bus-error
> error-counter-tx-rx{{248}{0}}
> (000.000664) can0 200000C0 [8] 00 00 00 00 00 00 00 00 ERRORFRAME
> bus-off
> bus-error
>
> (( No further bus activity ))
>
Bus-off seems OK. You could have just short circuited them without
disconnecting.
Reverse state transitions are missing from the logs. See comments above.
--
Andri
^ permalink raw reply
* Re: [PATCH net 0/2] net: marvell: Fix highmem support on non-TSO path
From: Russell King - ARM Linux @ 2015-01-21 15:01 UTC (permalink / raw)
To: Ezequiel Garcia; +Cc: netdev, David Miller, B38611, fabio.estevam
In-Reply-To: <1421844850-30886-1-git-send-email-ezequiel.garcia@free-electrons.com>
On Wed, Jan 21, 2015 at 09:54:08AM -0300, Ezequiel Garcia wrote:
> These two commits are fixes to the issue reported by Russell King on
> mv643xx_eth. Namely, the introduction of a regression by commit 69ad0dd7af22
> which removed the support for highmem skb fragments. The guilty commit
> introduced the assumption of fragment's payload being located in lowmem pages.
I do wonder whether 69ad0dd7af22 is the real culpret, or whether there is
some other change in the netdev layer that we're missing. That commit is
in 3.16, but from what I remember, 3.17 works fine, it's 3.18 which fails.
> A similar pattern can be found in the original mvneta driver (in fact, the
> regression was introduced by copy-pasting the mvneta code).
>
> These fixes are for the non-TSO egress path in mvneta and mv643xx_eth drivers.
> The TSO path needs a more intrusive change, as the TSO API needs to be fixed
> (e.g. to make it work in skb fragments, instead of pointers to data).
>
> Russell, as I'm still unable to reproduce this, do you think you can
> give it a spin over there?
Sure - I think the only one I can test is mv643xx_eth, I don't think I
have any device which supports mv_neta.
The test scenario is for a NFS mount (the Marvell device as the NFS
client) over IPv6.
Initial testing looks good, I'll let it run for a while with various
builds on the NFS share (which iirc was one of the triggering
workloads).
Thanks.
--
FTTC broadband for 0.8mile line: currently at 10.5Mbps down 400kbps up
according to speedtest.net.
^ permalink raw reply
* [PATCH net-next] cxgb4: Fixes cxgb4_inet6addr_notifier unregister call
From: Hariprasad Shenai @ 2015-01-21 15:27 UTC (permalink / raw)
To: netdev; +Cc: davem, leedom, anish, nirranjan, praveenm, Hariprasad Shenai
commit b5a02f503caa0837 ("cxgb4 : Update ipv6 address handling api") introduced
a regression where unregister cxgb4_inet6addr_notifier wasn't getting called
during module_exit.
Signed-off-by: Hariprasad Shenai <hariprasad@chelsio.com>
---
drivers/net/ethernet/chelsio/cxgb4/cxgb4_main.c | 2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/drivers/net/ethernet/chelsio/cxgb4/cxgb4_main.c b/drivers/net/ethernet/chelsio/cxgb4/cxgb4_main.c
index 1147e1e..12c1a3f 100644
--- a/drivers/net/ethernet/chelsio/cxgb4/cxgb4_main.c
+++ b/drivers/net/ethernet/chelsio/cxgb4/cxgb4_main.c
@@ -6290,7 +6290,7 @@ static int __init cxgb4_init_module(void)
static void __exit cxgb4_cleanup_module(void)
{
#if IS_ENABLED(CONFIG_IPV6)
- if (inet6addr_registered && list_empty(&adapter_list)) {
+ if (inet6addr_registered) {
unregister_inet6addr_notifier(&cxgb4_inet6addr_notifier);
inet6addr_registered = false;
}
--
1.7.1
^ permalink raw reply related
* Re: [PATCH v5 2/5] can: kvaser_usb: Consolidate and unify state change handling
From: Ahmed S. Darwish @ 2015-01-21 15:36 UTC (permalink / raw)
To: Andri Yngvason
Cc: Wolfgang Grandegger, Olivier Sobrie, Oliver Hartkopp,
Marc Kleine-Budde, Linux-CAN, netdev, LKML
In-Reply-To: <20150121150015.31351.6605@shannon>
On Wed, Jan 21, 2015 at 03:00:15PM +0000, Andri Yngvason wrote:
> Quoting Ahmed S. Darwish (2015-01-21 14:43:23)
> > Hi!
...
> > <-- Unplug the cable -->
> >
> > (000.009106) can0 20000080 [8] 00 00 00 00 00 00 08 00 ERRORFRAME
> > bus-error
> > error-counter-tx-rx{{8}{0}}
> > (000.001872) can0 20000080 [8] 00 00 00 00 00 00 10 00 ERRORFRAME
> > bus-error
> > error-counter-tx-rx{{16}{0}}
> [...]
> > error-counter-tx-rx{{80}{0}}
> > (000.001910) can0 20000080 [8] 00 00 00 00 00 00 58 00 ERRORFRAME
> > bus-error
> > error-counter-tx-rx{{88}{0}}
> > (000.001753) can0 20000084 [8] 00 08 00 00 00 00 60 00 ERRORFRAME
> > controller-problem{tx-error-warning}
> Good.
> > bus-error
> > error-counter-tx-rx{{96}{0}}
Nice.
> > (000.001720) can0 20000080 [8] 00 00 00 00 00 00 68 00 ERRORFRAME
> > bus-error
> > error-counter-tx-rx{{104}{0}}
> > (000.001876) can0 20000080 [8] 00 00 00 00 00 00 70 00 ERRORFRAME
> > bus-error
> > error-counter-tx-rx{{112}{0}}
> > (000.001749) can0 20000080 [8] 00 00 00 00 00 00 78 00 ERRORFRAME
> > bus-error
> > error-counter-tx-rx{{120}{0}}
> > (000.001771) can0 20000084 [8] 00 20 00 00 00 00 80 00 ERRORFRAME
> > controller-problem{tx-error-passive}
> Also good.
> > bus-error
> > error-counter-tx-rx{{128}{0}}
Also nice :-)
> > (000.001868) can0 20000080 [8] 00 00 00 00 00 00 80 00 ERRORFRAME
> > bus-error
> > error-counter-tx-rx{{128}{0}}
> > (000.001982) can0 20000080 [8] 00 00 00 00 00 00 80 00 ERRORFRAME
> > bus-error
> > error-counter-tx-rx{{128}{0}}
> >
> > (( Then a continous flood, exactly similar to the above packet, appears.
> > Unfortunately this flooding is a firmware problem. ))
> >
> > <-- Replug the cable, after a good amount of time -->
> >
> Where are the reverse state transitions?
> >
Hmmm...
[ ... ]
>
> Reverse state transitions are missing from the logs. See comments above.
>
When the device is on the _receiving_ end, and I unplug the CAN cable after
introducing some noise to the level of reaching WARNING or PASSIVE, I
receive a BUS_ERROR event with the rxerr count reset back to 0 or 1. In
that case, the driver correctly transitions back the state to ERROR_ACTIVE
and candump produces something similar to:
(000.000362) can0 2000008C [8] 00 40 40 00 00 00 00 01 ERRORFRAME
controller-problem{}
protocol-violation{{back-to-error-active}{}}
bus-error
error-counter-tx-rx{{0}{1}}
which is, AFAIK, the correct behaviour from the driver side.
Meanwhile, when the device is on the _sending_ end and I re-plug the CAN
cable again. Sometimes I receive events with txerr reset to 0 or 1, and
the driver correctly reverts back to ERROR_ACTIVE in that case. But on
another times like the quoted case above, I don't receive any events
resetting txerr back -- only data packets on the bus.
So, What can the driver do given the above?
Thanks,
Darwish
P.S. just in case, I'll also re-check now if the driver unintentionally
drops any important events resetting the txerr count back after a CAN
cable replug -- preventing the code from returning to ERROR_ACTIVE in
the process.
^ permalink raw reply
* unclear ipv6 redirect message (was Re: [PATCH v3 2/3] mfd: lubbock_io: add lubbock_io board)
From: Joe Perches @ 2015-01-21 16:05 UTC (permalink / raw)
To: Russell King - ARM Linux, netdev
Cc: Robert Jarzmik, Lee Jones, Mark Rutland, devicetree, Samuel Ortiz,
Pawel Moll, Ian Campbell, Dmitry Eremin-Solenikov, linux-kernel,
Haojian Zhuang, Rob Herring, Arnd Bergmann, linux-arm-kernel,
Kumar Gala, Daniel Mack
In-Reply-To: <20150121094453.GO26493@n2100.arm.linux.org.uk>
(adding netdev)
On Wed, 2015-01-21 at 09:44 +0000, Russell King - ARM Linux wrote:
> On Wed, Jan 21, 2015 at 08:46:29AM +0100, Robert Jarzmik wrote:
> > Russell King - ARM Linux <linux@arm.linux.org.uk> writes:
> >
> > > What I'd suggest (and always have done) is:
> > >
> > > dev_err(&pdev->dev, "couldn't request main irq%d: %d\n",
> > > irq, ret);
> > I like it, it's even more compact, I'll use it for next patch version.
>
> BTW, this is an example why I have the policy of always ensuring that
> the kernel messages print sufficient diagnostics. Right now, I have
> a problem - since I rebooted my firewall a few nights ago, I now get
> on one of my machines:
>
> rt6_redirect: source isn't a valid nexthop for redirect target
>
> and it spews that for a few minutes every 26 hours or so. No further
> information, and it leaves you wondering "well, what was the invalid
> next hop? What was the source?"
>
> Pretty much the only way to try and find out is to leave a tcpdump or
> wireshark running for 24 hours to try and get a dump - which is not
> that easy if you don't have lots of disk space. So, right now, I have
> no way to diagnose the above.
>
> If it printed that information, then I'd be able to see what the
> addresses were, and I'd probably be able to come up with a tcpdump
> filter which didn't involve logging all IPv6 traffic.
>
> Kernel messages need to be smart. If not, they might as well just be
> "The kernel encountered a problem. Abort, Retry or Fail?"
>
^ permalink raw reply
* Re: unclear ipv6 redirect message (was Re: [PATCH v3 2/3] mfd: lubbock_io: add lubbock_io board)
From: Russell King - ARM Linux @ 2015-01-21 16:11 UTC (permalink / raw)
To: Joe Perches
Cc: netdev, Robert Jarzmik, Lee Jones, Mark Rutland,
devicetree-u79uwXL29TY76Z2rM5mHXA, Samuel Ortiz, Pawel Moll,
Ian Campbell, Dmitry Eremin-Solenikov,
linux-kernel-u79uwXL29TY76Z2rM5mHXA, Haojian Zhuang, Rob Herring,
Arnd Bergmann, linux-arm-kernel-IAPFreCvJWM7uuMidbF8XUB+6BGkLq7r,
Kumar Gala, Daniel Mack
In-Reply-To: <1421856321.10574.13.camel-6d6DIl74uiNBDgjK7y7TUQ@public.gmane.org>
On Wed, Jan 21, 2015 at 08:05:21AM -0800, Joe Perches wrote:
> (adding netdev)
I wasn't actually reporting that as an issue; I was using it as an
example. It's from a very old kernel (2.6.27.21) which I run on one
of my old x86 machines.
--
FTTC broadband for 0.8mile line: currently at 10.5Mbps down 400kbps up
according to speedtest.net.
--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
^ permalink raw reply
* Re: [PATCH v5 2/5] can: kvaser_usb: Consolidate and unify state change handling
From: Wolfgang Grandegger @ 2015-01-21 16:13 UTC (permalink / raw)
To: Ahmed S. Darwish
Cc: Andri Yngvason, Olivier Sobrie, Oliver Hartkopp,
Marc Kleine-Budde, Linux-CAN, netdev, LKML
In-Reply-To: <20150121153647.GB17070@linux>
On Wed, 21 Jan 2015 10:36:47 -0500, "Ahmed S. Darwish"
<darwish.07@gmail.com> wrote:
> On Wed, Jan 21, 2015 at 03:00:15PM +0000, Andri Yngvason wrote:
>> Quoting Ahmed S. Darwish (2015-01-21 14:43:23)
>> > Hi!
>
> ...
>
>> > <-- Unplug the cable -->
>> >
>> > (000.009106) can0 20000080 [8] 00 00 00 00 00 00 08 00
>> > ERRORFRAME
>> > bus-error
>> > error-counter-tx-rx{{8}{0}}
>> > (000.001872) can0 20000080 [8] 00 00 00 00 00 00 10 00
For a bus-errors I would also expcect some more information in the
data[2..3] fields. But these are always zero.
>> > ERRORFRAME
>> > bus-error
>> > error-counter-tx-rx{{16}{0}}
>> [...]
>> > error-counter-tx-rx{{80}{0}}
>> > (000.001910) can0 20000080 [8] 00 00 00 00 00 00 58 00
>> > ERRORFRAME
>> > bus-error
>> > error-counter-tx-rx{{88}{0}}
>> > (000.001753) can0 20000084 [8] 00 08 00 00 00 00 60 00
>> > ERRORFRAME
>> > controller-problem{tx-error-warning}
>> Good.
>> > bus-error
>> > error-counter-tx-rx{{96}{0}}
>
> Nice.
>
>> > (000.001720) can0 20000080 [8] 00 00 00 00 00 00 68 00
>> > ERRORFRAME
>> > bus-error
>> > error-counter-tx-rx{{104}{0}}
>> > (000.001876) can0 20000080 [8] 00 00 00 00 00 00 70 00
>> > ERRORFRAME
>> > bus-error
>> > error-counter-tx-rx{{112}{0}}
>> > (000.001749) can0 20000080 [8] 00 00 00 00 00 00 78 00
>> > ERRORFRAME
>> > bus-error
>> > error-counter-tx-rx{{120}{0}}
>> > (000.001771) can0 20000084 [8] 00 20 00 00 00 00 80 00
>> > ERRORFRAME
>> > controller-problem{tx-error-passive}
>> Also good.
>> > bus-error
>> > error-counter-tx-rx{{128}{0}}
>
> Also nice :-)
>
>> > (000.001868) can0 20000080 [8] 00 00 00 00 00 00 80 00
>> > ERRORFRAME
>> > bus-error
>> > error-counter-tx-rx{{128}{0}}
>> > (000.001982) can0 20000080 [8] 00 00 00 00 00 00 80 00
>> > ERRORFRAME
>> > bus-error
>> > error-counter-tx-rx{{128}{0}}
>> >
>> > (( Then a continous flood, exactly similar to the above packet,
>> > appears.
>> > Unfortunately this flooding is a firmware problem. ))
>> >
>> > <-- Replug the cable, after a good amount of time -->
>> >
>> Where are the reverse state transitions?
>> >
>
> Hmmm...
>
> [ ... ]
>>
>> Reverse state transitions are missing from the logs. See comments
above.
>>
>
> When the device is on the _receiving_ end, and I unplug the CAN cable
after
> introducing some noise to the level of reaching WARNING or PASSIVE, I
> receive a BUS_ERROR event with the rxerr count reset back to 0 or 1. In
> that case, the driver correctly transitions back the state to
ERROR_ACTIVE
> and candump produces something similar to:
>
> (000.000362) can0 2000008C [8] 00 40 40 00 00 00 00 01
> ERRORFRAME
> controller-problem{}
> protocol-violation{{back-to-error-active}{}}
> bus-error
> error-counter-tx-rx{{0}{1}}
>
> which is, AFAIK, the correct behaviour from the driver side.
>
> Meanwhile, when the device is on the _sending_ end and I re-plug the CAN
> cable again. Sometimes I receive events with txerr reset to 0 or 1, and
> the driver correctly reverts back to ERROR_ACTIVE in that case. But on
> another times like the quoted case above, I don't receive any events
> resetting txerr back -- only data packets on the bus.
Well, the firmware seems to report *only* bus-errors via
CMD_CAN_ERROR_EVENT
messages, also carrying the new state, but no CMD_CHIP_STATE_EVENT just
for
the state changes.
> So, What can the driver do given the above?
Little if the notification does not come.
Wolfgang.
^ permalink raw reply
* Re: [PATCH v5 2/5] can: kvaser_usb: Consolidate and unify state change handling
From: Andri Yngvason @ 2015-01-21 16:20 UTC (permalink / raw)
To: Ahmed S. Darwish, Olivier Sobrie, Oliver Hartkopp,
Wolfgang Grandegger, Marc Kleine-Budde
Cc: Linux-CAN, netdev, LKML
In-Reply-To: <20150120214537.GB16828@linux>
Quoting Ahmed S. Darwish (2015-01-20 21:45:37)
> From: Ahmed S. Darwish <ahmed.darwish@valeo.com>
>
> Replace most of the can interface's state and error counters
> handling with the new can-dev can_change_state() mechanism.
>
> Suggested-by: Andri Yngvason <andri.yngvason@marel.com>
> Signed-off-by: Ahmed S. Darwish <ahmed.darwish@valeo.com>
> ---
> drivers/net/can/usb/kvaser_usb.c | 114 +++++++++++++++++++--------------------
> 1 file changed, 55 insertions(+), 59 deletions(-)
>
> diff --git a/drivers/net/can/usb/kvaser_usb.c b/drivers/net/can/usb/kvaser_usb.c
> index 971c5f9..0386d3f 100644
> --- a/drivers/net/can/usb/kvaser_usb.c
> +++ b/drivers/net/can/usb/kvaser_usb.c
> @@ -620,40 +620,43 @@ static void kvaser_usb_unlink_tx_urbs(struct kvaser_usb_net_priv *priv)
> }
>
> static void kvaser_usb_rx_error_update_can_state(struct kvaser_usb_net_priv *priv,
> - const struct kvaser_usb_error_summary *es)
> + const struct kvaser_usb_error_summary *es,
> + struct can_frame *cf)
> {
> struct net_device_stats *stats;
> - enum can_state new_state;
> -
> - stats = &priv->netdev->stats;
> - new_state = priv->can.state;
> + enum can_state cur_state, new_state, tx_state, rx_state;
>
> netdev_dbg(priv->netdev, "Error status: 0x%02x\n", es->status);
>
> - if (es->status & M16C_STATE_BUS_OFF) {
> - priv->can.can_stats.bus_off++;
> + stats = &priv->netdev->stats;
> + new_state = cur_state = priv->can.state;
> +
> + if (es->status & M16C_STATE_BUS_OFF)
> new_state = CAN_STATE_BUS_OFF;
> - } else if (es->status & M16C_STATE_BUS_PASSIVE) {
> - if (priv->can.state != CAN_STATE_ERROR_PASSIVE)
> - priv->can.can_stats.error_passive++;
> + else if (es->status & M16C_STATE_BUS_PASSIVE)
> new_state = CAN_STATE_ERROR_PASSIVE;
> - }
>
> if (es->status == M16C_STATE_BUS_ERROR) {
> - if ((priv->can.state < CAN_STATE_ERROR_WARNING) &&
> - ((es->txerr >= 96) || (es->rxerr >= 96))) {
> - priv->can.can_stats.error_warning++;
> + if ((cur_state < CAN_STATE_ERROR_WARNING) &&
> + ((es->txerr >= 96) || (es->rxerr >= 96)))
> new_state = CAN_STATE_ERROR_WARNING;
> - } else if (priv->can.state > CAN_STATE_ERROR_ACTIVE) {
> + else if (cur_state > CAN_STATE_ERROR_ACTIVE)
> new_state = CAN_STATE_ERROR_ACTIVE;
> - }
> }
>
> if (!es->status)
> new_state = CAN_STATE_ERROR_ACTIVE;
>
> + if (new_state != cur_state) {
> + tx_state = (es->txerr >= es->rxerr) ? new_state : 0;
> + rx_state = (es->txerr <= es->rxerr) ? new_state : 0;
> +
> + can_change_state(priv->netdev, cf, tx_state, rx_state);
This (below) is redundant. It doesn't harm but at this point can_change_state
has set priv->can.state to new_state.
> + new_state = priv->can.state;
> + }
> +
> if (priv->can.restart_ms &&
> - (priv->can.state >= CAN_STATE_BUS_OFF) &&
> + (cur_state >= CAN_STATE_BUS_OFF) &&
> (new_state < CAN_STATE_BUS_OFF)) {
> priv->can.can_stats.restarts++;
> }
> @@ -665,18 +668,17 @@ static void kvaser_usb_rx_error_update_can_state(struct kvaser_usb_net_priv *pri
>
> priv->bec.txerr = es->txerr;
> priv->bec.rxerr = es->rxerr;
> - priv->can.state = new_state;
> }
>
> static void kvaser_usb_rx_error(const struct kvaser_usb *dev,
> const struct kvaser_msg *msg)
> {
> - struct can_frame *cf;
> + struct can_frame *cf, tmp_cf = { .can_id = CAN_ERR_FLAG, .can_dlc = CAN_ERR_DLC };
> struct sk_buff *skb;
> struct net_device_stats *stats;
> struct kvaser_usb_net_priv *priv;
> struct kvaser_usb_error_summary es = { };
> - enum can_state old_state;
> + enum can_state old_state, new_state;
>
> switch (msg->id) {
> case CMD_CAN_ERROR_EVENT:
> @@ -721,60 +723,54 @@ static void kvaser_usb_rx_error(const struct kvaser_usb *dev,
> }
>
> /* Update all of the can interface's state and error counters before
> - * trying any skb allocation that can actually fail with -ENOMEM.
> + * trying any memory allocation that can actually fail with -ENOMEM.
> + *
> + * We send a temporary stack-allocated error can frame to
> + * can_change_state() for the very same reason.
> + *
> + * TODO: Split can_change_state() responsibility between updating the
> + * can interface's state and counters, and the setting up of can error
> + * frame ID and data to userspace. Remove stack allocation afterwards.
> */
> old_state = priv->can.state;
> - kvaser_usb_rx_error_update_can_state(priv, &es);
> + kvaser_usb_rx_error_update_can_state(priv, &es, &tmp_cf);
> + new_state = priv->can.state;
>
> skb = alloc_can_err_skb(priv->netdev, &cf);
> if (!skb) {
> stats->rx_dropped++;
> return;
> }
> + memcpy(cf, &tmp_cf, sizeof(*cf));
>
> - if (es.status & M16C_STATE_BUS_OFF) {
> - cf->can_id |= CAN_ERR_BUSOFF;
> -
> - if (!priv->can.restart_ms)
> - kvaser_usb_simple_msg_async(priv, CMD_STOP_CHIP);
> - netif_carrier_off(priv->netdev);
> - } else if (es.status & M16C_STATE_BUS_PASSIVE) {
> - if (old_state != CAN_STATE_ERROR_PASSIVE) {
> - cf->can_id |= CAN_ERR_CRTL;
> -
> - if (es.txerr || es.rxerr)
> - cf->data[1] = (es.txerr > es.rxerr)
> - ? CAN_ERR_CRTL_TX_PASSIVE
> - : CAN_ERR_CRTL_RX_PASSIVE;
> - else
> - cf->data[1] = CAN_ERR_CRTL_TX_PASSIVE |
> - CAN_ERR_CRTL_RX_PASSIVE;
> + if (new_state != old_state) {
> + if (es.status & M16C_STATE_BUS_OFF) {
> + if (!priv->can.restart_ms)
> + kvaser_usb_simple_msg_async(priv, CMD_STOP_CHIP);
> + netif_carrier_off(priv->netdev);
> + }
> +
This block is wrong. The usage of PROT_ACTIVE is based on a misunderstanding.
It's used in some drivers to signify back-to-error-active but its original
meaning is something completely different, AFAIK.
This is handled in can_change_state() using a new CTRL message; namely:
CAN_ERR_CTRL_ACTIVE. The newest version of can-utils is up to date with this.
> + if (es.status == M16C_STATE_BUS_ERROR) {
> + if ((old_state >= CAN_STATE_ERROR_WARNING) ||
> + (es.txerr < 96 && es.rxerr < 96)) {
> + if (old_state > CAN_STATE_ERROR_ACTIVE) {
> + cf->can_id |= CAN_ERR_PROT;
> + cf->data[2] = CAN_ERR_PROT_ACTIVE;
> + }
> + }
> }
> - }
>
> - if (es.status == M16C_STATE_BUS_ERROR) {
> - if ((old_state < CAN_STATE_ERROR_WARNING) &&
> - ((es.txerr >= 96) || (es.rxerr >= 96))) {
> - cf->can_id |= CAN_ERR_CRTL;
> - cf->data[1] = (es.txerr > es.rxerr)
> - ? CAN_ERR_CRTL_TX_WARNING
> - : CAN_ERR_CRTL_RX_WARNING;
> - } else if (old_state > CAN_STATE_ERROR_ACTIVE) {
This is also redundant, and wrong:
> + if (!es.status) {
> cf->can_id |= CAN_ERR_PROT;
> cf->data[2] = CAN_ERR_PROT_ACTIVE;
> }
> - }
>
> - if (!es.status) {
> - cf->can_id |= CAN_ERR_PROT;
> - cf->data[2] = CAN_ERR_PROT_ACTIVE;
> - }
> -
> - if (priv->can.restart_ms &&
> - (old_state >= CAN_STATE_BUS_OFF) &&
> - (priv->can.state < CAN_STATE_BUS_OFF)) {
> - cf->can_id |= CAN_ERR_RESTARTED;
> - netif_carrier_on(priv->netdev);
> + if (priv->can.restart_ms &&
> + (old_state >= CAN_STATE_BUS_OFF) &&
> + (new_state < CAN_STATE_BUS_OFF)) {
> + cf->can_id |= CAN_ERR_RESTARTED;
> + netif_carrier_on(priv->netdev);
> + }
> }
>
> if (es.error_factor) {
> --
> 1.9.1
Looking over the patch again, I've noticed that there are a few things that are
not quite right.
Marc, could you merge the "move bus_off++" patch before you merge this so that I
won't have to incorporate this patch-set into it?
--
Andri
^ permalink raw reply
* Re: [PATCH] net: dsa: set parent of hwmon device
From: Vivien Didelot @ 2015-01-21 16:26 UTC (permalink / raw)
To: Guenter Roeck
Cc: Florian Fainelli, netdev, David S . Miller, linux-kernel, kernel
In-Reply-To: <CAGVrzcawAF1=RPAJ+0O3B+PA7sDkstTjVjDQL_8C8a8kHL7CoA@mail.gmail.com>
Hi Guenter, Florian,
>>> - ds->hwmon_dev = hwmon_device_register_with_groups(NULL,
>>> + ds->hwmon_dev = hwmon_device_register_with_groups(parent,
>>> ds->hwmon_name, ds, dsa_hwmon_groups);
>>> if (IS_ERR(ds->hwmon_dev))
>>> ds->hwmon_dev = NULL;
>>
>> Looking into my old e-mail, turns out we did not add the parent
>> device because it affected the output of the "sensors" command, and
>> we wanted the device to be handled as 'virtual device' (which implies
>> no parent). That was an explicit part of the patch set (v2 of 'net:
>> dsa: Add support for reporting switch chip temperatures'), compared
>> to v1, which did set the parent device.
I don't know about the "sensors" output, but I found that convenient
from the sysfs side to have a hierarchy of sub-devices logically exposed
at the same place, i.e.:
# ls /sys/devices/platform/dsa.0
driver hwmon net ...
>> I would suggest to keep the code as is.
> Maybe follow-up with a comment adding that above the call to
> hwmon_device_register_with_groups()? I suspect the intent is clear if
> you are deep into hwmon devices, but not necessarily for the reader ;)
Thanks,
-v
^ permalink raw reply
* Re: [PATCH v5 2/5] can: kvaser_usb: Consolidate and unify state change handling
From: Andri Yngvason @ 2015-01-21 16:37 UTC (permalink / raw)
To: Ahmed S. Darwish
Cc: Wolfgang Grandegger, Olivier Sobrie, Oliver Hartkopp,
Marc Kleine-Budde, Linux-CAN, netdev, LKML
In-Reply-To: <20150121153647.GB17070@linux>
Quoting Ahmed S. Darwish (2015-01-21 15:36:47)
> On Wed, Jan 21, 2015 at 03:00:15PM +0000, Andri Yngvason wrote:
> > Quoting Ahmed S. Darwish (2015-01-21 14:43:23)
> > > Hi!
>
> ...
>
> > > <-- Unplug the cable -->
> > >
> > > (000.009106) can0 20000080 [8] 00 00 00 00 00 00 08 00 ERRORFRAME
> > > bus-error
> > > error-counter-tx-rx{{8}{0}}
> > > (000.001872) can0 20000080 [8] 00 00 00 00 00 00 10 00 ERRORFRAME
> > > bus-error
> > > error-counter-tx-rx{{16}{0}}
> > [...]
> > > error-counter-tx-rx{{80}{0}}
> > > (000.001910) can0 20000080 [8] 00 00 00 00 00 00 58 00 ERRORFRAME
> > > bus-error
> > > error-counter-tx-rx{{88}{0}}
> > > (000.001753) can0 20000084 [8] 00 08 00 00 00 00 60 00 ERRORFRAME
> > > controller-problem{tx-error-warning}
> > Good.
> > > bus-error
> > > error-counter-tx-rx{{96}{0}}
>
> Nice.
>
> > > (000.001720) can0 20000080 [8] 00 00 00 00 00 00 68 00 ERRORFRAME
> > > bus-error
> > > error-counter-tx-rx{{104}{0}}
> > > (000.001876) can0 20000080 [8] 00 00 00 00 00 00 70 00 ERRORFRAME
> > > bus-error
> > > error-counter-tx-rx{{112}{0}}
> > > (000.001749) can0 20000080 [8] 00 00 00 00 00 00 78 00 ERRORFRAME
> > > bus-error
> > > error-counter-tx-rx{{120}{0}}
> > > (000.001771) can0 20000084 [8] 00 20 00 00 00 00 80 00 ERRORFRAME
> > > controller-problem{tx-error-passive}
> > Also good.
> > > bus-error
> > > error-counter-tx-rx{{128}{0}}
>
> Also nice :-)
>
> > > (000.001868) can0 20000080 [8] 00 00 00 00 00 00 80 00 ERRORFRAME
> > > bus-error
> > > error-counter-tx-rx{{128}{0}}
> > > (000.001982) can0 20000080 [8] 00 00 00 00 00 00 80 00 ERRORFRAME
> > > bus-error
> > > error-counter-tx-rx{{128}{0}}
> > >
> > > (( Then a continous flood, exactly similar to the above packet, appears.
> > > Unfortunately this flooding is a firmware problem. ))
> > >
> > > <-- Replug the cable, after a good amount of time -->
> > >
> > Where are the reverse state transitions?
> > >
>
> Hmmm...
>
> [ ... ]
> >
> > Reverse state transitions are missing from the logs. See comments above.
> >
>
> When the device is on the _receiving_ end, and I unplug the CAN cable after
> introducing some noise to the level of reaching WARNING or PASSIVE, I
> receive a BUS_ERROR event with the rxerr count reset back to 0 or 1. In
> that case, the driver correctly transitions back the state to ERROR_ACTIVE
> and candump produces something similar to:
>
> (000.000362) can0 2000008C [8] 00 40 40 00 00 00 00 01 ERRORFRAME
> controller-problem{}
> protocol-violation{{back-to-error-active}{}}
> bus-error
> error-counter-tx-rx{{0}{1}}
>
> which is, AFAIK, the correct behaviour from the driver side.
>
> Meanwhile, when the device is on the _sending_ end and I re-plug the CAN
> cable again. Sometimes I receive events with txerr reset to 0 or 1, and
> the driver correctly reverts back to ERROR_ACTIVE in that case. But on
> another times like the quoted case above, I don't receive any events
> resetting txerr back -- only data packets on the bus.
>
> So, What can the driver do given the above?
>
So what you're telling us is that the state does not got back to error-active
unless there is something else transmitting on the bus?
If that's the case, it's almost definitely because state changes aren't
triggering interrupts.
An rx event will give you an interrupt which yields a napi poll which means that
the state will be polled, so in that case you don't need the interrupts.
Look for "Consolidate and unify state change handling" on gmane. There was a lot
of discussion about this kind of issues, especially regarding FlexCAN.
Best regards,
Andri
^ permalink raw reply
* Re: unclear ipv6 redirect message (was Re: [PATCH v3 2/3] mfd: lubbock_io: add lubbock_io board)
From: Joe Perches @ 2015-01-21 16:40 UTC (permalink / raw)
To: Russell King - ARM Linux
Cc: netdev, Robert Jarzmik, Lee Jones, Mark Rutland, devicetree,
Samuel Ortiz, Pawel Moll, Ian Campbell, Dmitry Eremin-Solenikov,
linux-kernel, Haojian Zhuang, Rob Herring, Arnd Bergmann,
linux-arm-kernel, Kumar Gala, Daniel Mack
In-Reply-To: <20150121161059.GT26493@n2100.arm.linux.org.uk>
On Wed, 2015-01-21 at 16:11 +0000, Russell King - ARM Linux wrote:
> On Wed, Jan 21, 2015 at 08:05:21AM -0800, Joe Perches wrote:
> > (adding netdev)
>
> I wasn't actually reporting that as an issue; I was using it as an
> example. It's from a very old kernel (2.6.27.21) which I run on one
> of my old x86 machines.
It's still the same code.
If the message can be improved, why not do it?
^ permalink raw reply
* [PATCH 3/3][v2] mdio-mux-gpio: use new gpiod_get_array and gpiod_put_array functions
From: Rojhalat Ibrahim @ 2015-01-21 16:46 UTC (permalink / raw)
To: linux-gpio@vger.kernel.org
Cc: Alexandre Courbot, Linus Walleij, David Miller, netdev
Use the new gpiod_get_array and gpiod_put_array functions for obtaining and
disposing of GPIO descriptors.
Cc: David Miller <davem@davemloft.net>
Signed-off-by: Rojhalat Ibrahim <imr@rtschenk.de>
---
This patch depends on my previous patch "gpiolib: add gpiod_get_array and
gpiod_put_array functions".
v2: use the new interface
Only compile-tested.
drivers/net/phy/mdio-mux-gpio.c | 61 +++++++++++-----------------------------
1 file changed, 18 insertions(+), 43 deletions(-)
diff --git a/drivers/net/phy/mdio-mux-gpio.c b/drivers/net/phy/mdio-mux-gpio.c
index 320eb15..3f6929c 100644
--- a/drivers/net/phy/mdio-mux-gpio.c
+++ b/drivers/net/phy/mdio-mux-gpio.c
@@ -12,33 +12,31 @@
#include <linux/module.h>
#include <linux/phy.h>
#include <linux/mdio-mux.h>
-#include <linux/of_gpio.h>
+#include <linux/gpio/consumer.h>
#define DRV_VERSION "1.1"
#define DRV_DESCRIPTION "GPIO controlled MDIO bus multiplexer driver"
-#define MDIO_MUX_GPIO_MAX_BITS 8
-
struct mdio_mux_gpio_state {
- struct gpio_desc *gpio[MDIO_MUX_GPIO_MAX_BITS];
- unsigned int num_gpios;
+ struct gpio_descs *gpios;
void *mux_handle;
};
static int mdio_mux_gpio_switch_fn(int current_child, int desired_child,
void *data)
{
- int values[MDIO_MUX_GPIO_MAX_BITS];
- unsigned int n;
struct mdio_mux_gpio_state *s = data;
+ int values[s->gpios->array_size];
+ unsigned int n;
if (current_child == desired_child)
return 0;
- for (n = 0; n < s->num_gpios; n++) {
+ for (n = 0; n < s->gpios->array_size; n++)
values[n] = (desired_child >> n) & 1;
- }
- gpiod_set_array_cansleep(s->num_gpios, s->gpio, values);
+
+ gpiod_set_array_cansleep(s->gpios->array_size, s->gpios->desc_array,
+ values);
return 0;
}
@@ -46,56 +44,33 @@ static int mdio_mux_gpio_switch_fn(int current_child, int desired_child,
static int mdio_mux_gpio_probe(struct platform_device *pdev)
{
struct mdio_mux_gpio_state *s;
- int num_gpios;
- unsigned int n;
int r;
- if (!pdev->dev.of_node)
- return -ENODEV;
-
- num_gpios = of_gpio_count(pdev->dev.of_node);
- if (num_gpios <= 0 || num_gpios > MDIO_MUX_GPIO_MAX_BITS)
- return -ENODEV;
-
s = devm_kzalloc(&pdev->dev, sizeof(*s), GFP_KERNEL);
if (!s)
return -ENOMEM;
- s->num_gpios = num_gpios;
-
- for (n = 0; n < num_gpios; ) {
- struct gpio_desc *gpio = gpiod_get_index(&pdev->dev, NULL, n,
- GPIOD_OUT_LOW);
- if (IS_ERR(gpio)) {
- r = PTR_ERR(gpio);
- goto err;
- }
- s->gpio[n] = gpio;
- n++;
- }
+ s->gpios = gpiod_get_array(&pdev->dev, NULL, GPIOD_OUT_LOW);
+ if (IS_ERR(s->gpios))
+ return PTR_ERR(s->gpios);
r = mdio_mux_init(&pdev->dev,
mdio_mux_gpio_switch_fn, &s->mux_handle, s);
- if (r == 0) {
- pdev->dev.platform_data = s;
- return 0;
- }
-err:
- while (n) {
- n--;
- gpiod_put(s->gpio[n]);
+ if (r != 0) {
+ gpiod_put_array(s->gpios);
+ return r;
}
- return r;
+
+ pdev->dev.platform_data = s;
+ return 0;
}
static int mdio_mux_gpio_remove(struct platform_device *pdev)
{
- unsigned int n;
struct mdio_mux_gpio_state *s = dev_get_platdata(&pdev->dev);
mdio_mux_uninit(s->mux_handle);
- for (n = 0; n < s->num_gpios; n++)
- gpiod_put(s->gpio[n]);
+ gpiod_put_array(s->gpios);
return 0;
}
--
2.0.5
^ permalink raw reply related
* Re: unclear ipv6 redirect message (was Re: [PATCH v3 2/3] mfd: lubbock_io: add lubbock_io board)
From: Russell King - ARM Linux @ 2015-01-21 16:46 UTC (permalink / raw)
To: Joe Perches
Cc: netdev, Robert Jarzmik, Lee Jones, Mark Rutland, devicetree,
Samuel Ortiz, Pawel Moll, Ian Campbell, Dmitry Eremin-Solenikov,
linux-kernel, Haojian Zhuang, Rob Herring, Arnd Bergmann,
linux-arm-kernel, Kumar Gala, Daniel Mack
In-Reply-To: <1421858444.10574.14.camel@perches.com>
On Wed, Jan 21, 2015 at 08:40:44AM -0800, Joe Perches wrote:
> On Wed, 2015-01-21 at 16:11 +0000, Russell King - ARM Linux wrote:
> > On Wed, Jan 21, 2015 at 08:05:21AM -0800, Joe Perches wrote:
> > > (adding netdev)
> >
> > I wasn't actually reporting that as an issue; I was using it as an
> > example. It's from a very old kernel (2.6.27.21) which I run on one
> > of my old x86 machines.
>
> It's still the same code.
> If the message can be improved, why not do it?
I assume you're taking the responsibility to test anything that comes
out of this then?
I'm not; I tried updating the kernel on the machine to 2.6.32 many
years ago and that was a no-go because of userspace (udev)
incompatibilities.
--
FTTC broadband for 0.8mile line: currently at 10.5Mbps down 400kbps up
according to speedtest.net.
^ permalink raw reply
* Re: [PATCH net-next 0/3] openvswitch: Add STT support.
From: Tom Herbert @ 2015-01-21 16:51 UTC (permalink / raw)
To: Pravin Shelar; +Cc: David Miller, Linux Netdev List
In-Reply-To: <CALnjE+rSXpPftiG4Vw3azkN2s2-Ey6ZcZr9Bv422pOdwrvU=tw@mail.gmail.com>
On Wed, Jan 21, 2015 at 1:08 AM, Pravin Shelar <pshelar@nicira.com> wrote:
> On Tue, Jan 20, 2015 at 3:06 PM, Tom Herbert <therbert@google.com> wrote:
>> On Tue, Jan 20, 2015 at 12:25 PM, Pravin B Shelar <pshelar@nicira.com> wrote:
>>> Following patch series adds support for Stateless Transport
>>> Tunneling protocol.
>>> STT uses TCP segmentation offload available in most of NIC. On
>>> packet xmit STT driver appends STT header along with TCP header
>>> to the packet. For GSO packet GSO parameters are set according
>>> to tunnel configuration and packet is handed over to networking
>>> stack. This allows use of segmentation offload available in NICs
>>>
>>> Netperf unidirectional test gives ~9.4 Gbits/s performance on 10Gbit
>>> NIC with 1500 byte MTU with two TCP streams.
>>>
>> Having packets marked TCP which really aren't TCP is a rather scary
>> prospect to deploy in a real data center (TCP is kind of an important
>> protocol ;-) ). Can you give some more motivation on this, more data
>> that shows what the benefits are and how this compares to equivalent
>> encapsulation protocols that implement GRO and GSO.
>>
> There are multi-year deployments of STT, So it is already in real data center.
> Biggest advantage is STT does not need new NIC with tunnel offload.
> Any NIC that supports TCP offload can be used to achieve better
> performance.
>
> Following are numbers you asked for.
> Setup: net-next branch on server and client.
> netperf: TCP unidirectional tests with 5 streams. Numbers are averaged
> over 3 runs of 50 sec.
>
Please provides more details on your configuration so that others
might be able to reproduce your results. Also, it would be quite
helpful if you could implement STT as a normal network interface like
VXLAN does so that we can isolate performance of the protocol. For
instance I have no problem getting line rate with VXLAN using 5
streams with or without RCO in my testing. I assume you tested with
OVS and maybe VMs which may have a significant impact beyond the
protocol changes.
Another thing to consider in your analysis is the performance with
flows using small packets. STT should demonstrate better performance
with bulk flows since LSO and LRO are better performing relative to
GSO and GRO. But for flows with small packets, I don't see how there
could be any performance advantage. We already have ways to leverage
simple UDP checksum offload with UDP encapsulations, seems like STT
might just represent unnecessary header overhead in those cases.
> VXLAN:
> CPU
> Client: 1.6
> Server: 14.2
> Throughput: 5.6 Gbit/s
>
> VXLAN with rcsum:
> CPU
> Client: 0.89
> Server: 12.4
> Throughput: 5.8 Gbit/s
>
> STT:
> CPU
> Client: 1.28
> Server: 4.0
> Throughput: 9.5 Gbit/s
>
9.5Gbps? Rounding error or is this 40Gbps or larger than 1500 byte MTU?
Thanks,
Tom
> Thanks,
> Pravin.
^ permalink raw reply
* Re: BW regression after "tcp: refine TSO autosizing"
From: Eric Dumazet @ 2015-01-21 17:01 UTC (permalink / raw)
To: David Laight
Cc: 'Rick Jones', Dave Taht, Eyal Perry, Yuchung Cheng,
Neal Cardwell, Eyal Perry, Or Gerlitz, Linux Netdev List,
Amir Vadai, Yevgeny Petrilin, Saeed Mahameed, Ido Shamay,
Amir Ancel
In-Reply-To: <063D6719AE5E284EB5DD2968C1650D6D1CAD027E@AcuExch.aculab.com>
On Wed, 2015-01-21 at 12:26 +0000, David Laight wrote:
> From: Of Rick Jones
> > >> Are you saying that at long last, delayed acks as we knew them are
> > >> dead, dead, dead?
> > >
> > > Sorry, I can not parse what you are saying.
> > >
> > > In case you missed it, it has nothing to do with delayed ACK but GRO on
> > > receiver.
> >
> > Dave - assuming I've interpreted Eric's comments correctly, I believe
> > the answer to your question is No. Your desire for a world brimming
> > with ack-every-other purity has not been fulfilled :)
> >
> > However, the engineers formerly at Mentat are probably pleased that a
> > functional near-equivalent to their ACK avoidance heuristic has ended-up
> > being implemented and tacitly accepted, albeit by the back door :)
>
> I must recheck something I discovered a while back with more recent kernels.
> There has been a bad interaction between 'slow start' and 'delayed acks'
> when nagle is disabled on 0 RTT local links with uni-directional traffic.
>
> 'Slow start' would refuse to send more than 4 messages until it received
> an ack (rather than 4 mss of data).
> The receiving system wouldn't send an ack until the timer expired
> (or several mss of data were received) by which time the sender could have
> a lot of data queued.
>
> Due to the 0 RTT and bursty nature of the data 'slow start' happened
> all the time.
Following packetdrill test suggests that current kernel send up to 10
messages without having to wait for any ACK
(IW10)
// Set up production and experiment configs
`../common/defaults.sh`
// Establish a connection.
0.000 socket(..., SOCK_STREAM, IPPROTO_TCP) = 3
0.000 setsockopt(3, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0
0.000 bind(3, ..., ...) = 0
0.000 listen(3, 1) = 0
0.100 < S 0:0(0) win 32792 <mss 1460,nop,wscale 7>
0.100 > S. 0:0(0) ack 1 <mss 1460,nop,wscale 6>
0.110 < . 1:1(0) ack 1 win 257
0.110 accept(3, ..., ...) = 4
0.200 %{ assert tcpi_snd_cwnd == 10 }%
+0 setsockopt(4, SOL_TCP, TCP_NODELAY, [1], 4) = 0
+0.01 write(4, ..., 100) = 100
+0 > P. 1:101(100) ack 1
+0.01 write(4, ..., 100) = 100
+0 > P. 101:201(100) ack 1
+0.01 write(4, ..., 100) = 100
+0 > P. 201:301(100) ack 1
+0.01 write(4, ..., 100) = 100
+0 > P. 301:401(100) ack 1
+0.01 write(4, ..., 100) = 100
+0 > P. 401:501(100) ack 1
+0.01 write(4, ..., 100) = 100
+0 > P. 501:601(100) ack 1
+0.01 write(4, ..., 100) = 100
+0 > P. 601:701(100) ack 1
+0.01 write(4, ..., 100) = 100
+0 > P. 701:801(100) ack 1
+0.01 write(4, ..., 100) = 100
+0 > P. 801:901(100) ack 1
+0.01 write(4, ..., 100) = 100
+0 > P. 901:1001(100) ack 1
^ permalink raw reply
* Questions about the sh_eth driver and hardware
From: Ben Hutchings @ 2015-01-21 17:24 UTC (permalink / raw)
To: Nobuhiro Iwamatsu, Mitsuhiro Kimura, Hisashi Nakamura,
Yoshihiro Kaneko
Cc: netdev, ct-linux-kernel
I found several more bugs in the sh_eth driver, but for some of them I'm
really unsure what to do. I only have a manual for the R8A7790 (R-Car
H2) and I don't know how all the other supported chips will behave.
Maybe you can answer some of these questions.
1. When freeing packet buffers, we currently try to stop the DMA engines
by clearing EDRRR and EDTRR but we *don't* wait after that. This seems
unsafe because in general register writes are not serialised with DMA.
The R8A7790 (R-Car H2) manual specifically says that the R bit of EDTRR
(aka CXR2) cannot be cleared by writing to it. I think that we could
stop TX DMA by clearing the active flags of all the descriptors and then
polling the R bit until it clears. What do you think?
As for RX DMA, I think we should wait some time after clearing the R bit
that we can be sure is long enough to transfer one packet. Do you know
how long that could be?
2. In case of a Receive Descriptor Empty error (RDE), we currently read
the RDFAR register to find the next descriptor the DMA engine will use.
But this register is not documented for the R8A7790 and the driver does
not define an offset for it on R-Car chips. The manual doesn't say how
to set the address of the next descriptor to use. Maybe we should
assume that R-Car chips will never skip descriptors after RDE?
Ben.
^ 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