[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20220405194845.c443x4gf522c2kgv@sx1>
Date: Tue, 5 Apr 2022 12:48:45 -0700
From: Saeed Mahameed <saeed@...nel.org>
To: Leon Romanovsky <leon@...nel.org>
Cc: Jason Gunthorpe <jgg@...dia.com>,
Patrisious Haddad <phaddad@...dia.com>,
Jakub Kicinski <kuba@...nel.org>, linux-rdma@...r.kernel.org,
netdev@...r.kernel.org, Paolo Abeni <pabeni@...hat.com>,
Saeed Mahameed <saeedm@...dia.com>,
Yishai Hadas <yishaih@...dia.com>
Subject: Re: [PATCH mlx5-next 1/3] net/mlx5: Nullify eq->dbg and qp->dbg
pointers post destruction
On 05 Apr 11:12, Leon Romanovsky wrote:
>From: Patrisious Haddad <phaddad@...dia.com>
>
>Prior to this patch in the case that destroy_unmap_eq()
>failed and was called again, it triggered an additional call of
Where is it being failed and called again ? this shouldn't even be an
option, we try to keep mlx5 symmetrical, constructors and destructors are
supposed to be called only once in their respective positions.
the callers must be fixed to avoid re-entry, or change destructors to clear
up all resources even on failures, no matter what do not invent a reentry
protocols to mlx5 destructors.
>mlx5_debug_eq_remove() which causes a kernel crash, since
>eq->dbg was not nullified in previous call.
>
[...]
> int mlx5_debug_cq_add(struct mlx5_core_dev *dev, struct mlx5_core_cq *cq)
>diff --git a/drivers/net/ethernet/mellanox/mlx5/core/eq.c b/drivers/net/ethernet/mellanox/mlx5/core/eq.c
>index 229728c80233..3c61f355cdac 100644
>--- a/drivers/net/ethernet/mellanox/mlx5/core/eq.c
>+++ b/drivers/net/ethernet/mellanox/mlx5/core/eq.c
>@@ -386,16 +386,20 @@ void mlx5_eq_disable(struct mlx5_core_dev *dev, struct mlx5_eq *eq,
> }
> EXPORT_SYMBOL(mlx5_eq_disable);
>
>-static int destroy_unmap_eq(struct mlx5_core_dev *dev, struct mlx5_eq *eq)
>+static int destroy_unmap_eq(struct mlx5_core_dev *dev, struct mlx5_eq *eq,
>+ bool reentry)
> {
> int err;
>
> mlx5_debug_eq_remove(dev, eq);
>
> err = mlx5_cmd_destroy_eq(dev, eq->eqn);
>- if (err)
>+ if (err) {
> mlx5_core_warn(dev, "failed to destroy a previously created eq: eqn %d\n",
> eq->eqn);
>+ if (reentry)
>+ return err;
>+ }
>
> mlx5_frag_buf_free(dev, &eq->frag_buf);
> return err;
>@@ -481,7 +485,7 @@ static int destroy_async_eq(struct mlx5_core_dev *dev, struct mlx5_eq *eq)
> int err;
>
> mutex_lock(&eq_table->lock);
>- err = destroy_unmap_eq(dev, eq);
>+ err = destroy_unmap_eq(dev, eq, false);
> mutex_unlock(&eq_table->lock);
> return err;
> }
>@@ -748,12 +752,15 @@ EXPORT_SYMBOL(mlx5_eq_create_generic);
>
> int mlx5_eq_destroy_generic(struct mlx5_core_dev *dev, struct mlx5_eq *eq)
> {
>+ struct mlx5_eq_table *eq_table = dev->priv.eq_table;
> int err;
>
> if (IS_ERR(eq))
> return -EINVAL;
>
>- err = destroy_async_eq(dev, eq);
>+ mutex_lock(&eq_table->lock);
Here you are inventing the re-entry.
Please drop this and fix properly. And avoid boolean parameters to mlx5 core
functions as much as possible, let's keep mlx5_core simple.
>+ err = destroy_unmap_eq(dev, eq, true);
>+ mutex_unlock(&eq_table->lock);
> if (err)
> goto out;
>
>@@ -851,7 +858,7 @@ static void destroy_comp_eqs(struct mlx5_core_dev *dev)
> list_for_each_entry_safe(eq, n, &table->comp_eqs_list, list) {
> list_del(&eq->list);
> mlx5_eq_disable(dev, &eq->core, &eq->irq_nb);
>- if (destroy_unmap_eq(dev, &eq->core))
>+ if (destroy_unmap_eq(dev, &eq->core, false))
> mlx5_core_warn(dev, "failed to destroy comp EQ 0x%x\n",
> eq->core.eqn);
> tasklet_disable(&eq->tasklet_ctx.task);
>@@ -915,7 +922,7 @@ static int create_comp_eqs(struct mlx5_core_dev *dev)
> goto clean_eq;
> err = mlx5_eq_enable(dev, &eq->core, &eq->irq_nb);
> if (err) {
>- destroy_unmap_eq(dev, &eq->core);
>+ destroy_unmap_eq(dev, &eq->core, false);
> goto clean_eq;
> }
>
>--
>2.35.1
>
Powered by blists - more mailing lists