Commit a1f240f1 authored by Yuval Avnery's avatar Yuval Avnery Committed by Saeed Mahameed

net/mlx5e: Adjust to max number of channles when re-attaching

When core driver enters deattach/attach flow after pci reset,
Number of logical CPUs may have changed.
As a result we need to update the cpu affiliated resource tables.
	1. indirect rqt list
	2. eq table

Reproduction (PowerPC):
	echo 1000 > /sys/kernel/debug/powerpc/eeh_max_freezes
	ppc64_cpu --smt=on
	# Restart driver
	modprobe -r ... ; modprobe ...
	# Link up
	ifconfig ...
	# Only physical CPUs
	ppc64_cpu --smt=off
	# Inject PCI errors so PCI will reset - calling the pci error handler
	echo 0x8000000000000000 > /sys/kernel/debug/powerpc/<PCI BUS>/err_injct_inboundA

Call trace when trying to add non-existing rqs to an indirect rqt:
	mlx5e_redirect_rqt+0x84/0x260 [mlx5_core] (unreliable)
	mlx5e_redirect_rqts+0x188/0x190 [mlx5_core]
	mlx5e_activate_priv_channels+0x488/0x570 [mlx5_core]
	mlx5e_open_locked+0xbc/0x140 [mlx5_core]
	mlx5e_open+0x50/0x130 [mlx5_core]
	mlx5e_nic_enable+0x174/0x1b0 [mlx5_core]
	mlx5e_attach_netdev+0x154/0x290 [mlx5_core]
	mlx5e_attach+0x88/0xd0 [mlx5_core]
	mlx5_attach_device+0x168/0x1e0 [mlx5_core]
	mlx5_load_one+0x1140/0x1210 [mlx5_core]
	mlx5_pci_resume+0x6c/0xf0 [mlx5_core]

Create cq will fail when trying to use non-existing EQ.

Fixes: 89d44f0a ("net/mlx5_core: Add pci error handlers to mlx5_core driver")
Signed-off-by: default avatarYuval Avnery <yuvalav@mellanox.com>
Signed-off-by: default avatarSaeed Mahameed <saeedm@mellanox.com>
parent 83621b7d
...@@ -1623,13 +1623,15 @@ static int mlx5e_alloc_cq_common(struct mlx5_core_dev *mdev, ...@@ -1623,13 +1623,15 @@ static int mlx5e_alloc_cq_common(struct mlx5_core_dev *mdev,
int err; int err;
u32 i; u32 i;
err = mlx5_vector2eqn(mdev, param->eq_ix, &eqn_not_used, &irqn);
if (err)
return err;
err = mlx5_cqwq_create(mdev, &param->wq, param->cqc, &cq->wq, err = mlx5_cqwq_create(mdev, &param->wq, param->cqc, &cq->wq,
&cq->wq_ctrl); &cq->wq_ctrl);
if (err) if (err)
return err; return err;
mlx5_vector2eqn(mdev, param->eq_ix, &eqn_not_used, &irqn);
mcq->cqe_sz = 64; mcq->cqe_sz = 64;
mcq->set_ci_db = cq->wq_ctrl.db.db; mcq->set_ci_db = cq->wq_ctrl.db.db;
mcq->arm_db = cq->wq_ctrl.db.db + 1; mcq->arm_db = cq->wq_ctrl.db.db + 1;
...@@ -1687,6 +1689,10 @@ static int mlx5e_create_cq(struct mlx5e_cq *cq, struct mlx5e_cq_param *param) ...@@ -1687,6 +1689,10 @@ static int mlx5e_create_cq(struct mlx5e_cq *cq, struct mlx5e_cq_param *param)
int eqn; int eqn;
int err; int err;
err = mlx5_vector2eqn(mdev, param->eq_ix, &eqn, &irqn_not_used);
if (err)
return err;
inlen = MLX5_ST_SZ_BYTES(create_cq_in) + inlen = MLX5_ST_SZ_BYTES(create_cq_in) +
sizeof(u64) * cq->wq_ctrl.buf.npages; sizeof(u64) * cq->wq_ctrl.buf.npages;
in = kvzalloc(inlen, GFP_KERNEL); in = kvzalloc(inlen, GFP_KERNEL);
...@@ -1700,8 +1706,6 @@ static int mlx5e_create_cq(struct mlx5e_cq *cq, struct mlx5e_cq_param *param) ...@@ -1700,8 +1706,6 @@ static int mlx5e_create_cq(struct mlx5e_cq *cq, struct mlx5e_cq_param *param)
mlx5_fill_page_frag_array(&cq->wq_ctrl.buf, mlx5_fill_page_frag_array(&cq->wq_ctrl.buf,
(__be64 *)MLX5_ADDR_OF(create_cq_in, in, pas)); (__be64 *)MLX5_ADDR_OF(create_cq_in, in, pas));
mlx5_vector2eqn(mdev, param->eq_ix, &eqn, &irqn_not_used);
MLX5_SET(cqc, cqc, cq_period_mode, param->cq_period_mode); MLX5_SET(cqc, cqc, cq_period_mode, param->cq_period_mode);
MLX5_SET(cqc, cqc, c_eqn, eqn); MLX5_SET(cqc, cqc, c_eqn, eqn);
MLX5_SET(cqc, cqc, uar_page, mdev->priv.uar->index); MLX5_SET(cqc, cqc, uar_page, mdev->priv.uar->index);
...@@ -1921,6 +1925,10 @@ static int mlx5e_open_channel(struct mlx5e_priv *priv, int ix, ...@@ -1921,6 +1925,10 @@ static int mlx5e_open_channel(struct mlx5e_priv *priv, int ix,
int err; int err;
int eqn; int eqn;
err = mlx5_vector2eqn(priv->mdev, ix, &eqn, &irq);
if (err)
return err;
c = kvzalloc_node(sizeof(*c), GFP_KERNEL, cpu_to_node(cpu)); c = kvzalloc_node(sizeof(*c), GFP_KERNEL, cpu_to_node(cpu));
if (!c) if (!c)
return -ENOMEM; return -ENOMEM;
...@@ -1937,7 +1945,6 @@ static int mlx5e_open_channel(struct mlx5e_priv *priv, int ix, ...@@ -1937,7 +1945,6 @@ static int mlx5e_open_channel(struct mlx5e_priv *priv, int ix,
c->xdp = !!params->xdp_prog; c->xdp = !!params->xdp_prog;
c->stats = &priv->channel_stats[ix].ch; c->stats = &priv->channel_stats[ix].ch;
mlx5_vector2eqn(priv->mdev, ix, &eqn, &irq);
c->irq_desc = irq_to_desc(irq); c->irq_desc = irq_to_desc(irq);
netif_napi_add(netdev, &c->napi, mlx5e_napi_poll, 64); netif_napi_add(netdev, &c->napi, mlx5e_napi_poll, 64);
...@@ -5010,11 +5017,21 @@ struct net_device *mlx5e_create_netdev(struct mlx5_core_dev *mdev, ...@@ -5010,11 +5017,21 @@ struct net_device *mlx5e_create_netdev(struct mlx5_core_dev *mdev,
int mlx5e_attach_netdev(struct mlx5e_priv *priv) int mlx5e_attach_netdev(struct mlx5e_priv *priv)
{ {
const struct mlx5e_profile *profile; const struct mlx5e_profile *profile;
int max_nch;
int err; int err;
profile = priv->profile; profile = priv->profile;
clear_bit(MLX5E_STATE_DESTROYING, &priv->state); clear_bit(MLX5E_STATE_DESTROYING, &priv->state);
/* max number of channels may have changed */
max_nch = mlx5e_get_max_num_channels(priv->mdev);
if (priv->channels.params.num_channels > max_nch) {
mlx5_core_warn(priv->mdev, "MLX5E: Reducing number of channels to %d\n", max_nch);
priv->channels.params.num_channels = max_nch;
mlx5e_build_default_indir_rqt(priv->channels.params.indirection_rqt,
MLX5E_INDIR_RQT_SIZE, max_nch);
}
err = profile->init_tx(priv); err = profile->init_tx(priv);
if (err) if (err)
goto out; goto out;
......
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment