]> git.proxmox.com Git - mirror_ubuntu-focal-kernel.git/commitdiff
IB: refcount race fixes
authorSean Hefty <mshefty@ichips.intel.com>
Fri, 12 May 2006 21:57:52 +0000 (14:57 -0700)
committerRoland Dreier <rolandd@cisco.com>
Fri, 12 May 2006 21:57:52 +0000 (14:57 -0700)
Fix race condition during destruction calls to avoid possibility of
accessing object after it has been freed.  Instead of waking up a wait
queue directly, which is susceptible to a race where the object is
freed between the reference count going to 0 and the wake_up(), use a
completion to wait in the function doing the freeing.

Signed-off-by: Sean Hefty <sean.hefty@intel.com>
Signed-off-by: Roland Dreier <rolandd@cisco.com>
drivers/infiniband/core/cm.c
drivers/infiniband/core/mad.c
drivers/infiniband/core/mad_priv.h
drivers/infiniband/core/mad_rmpp.c
drivers/infiniband/core/ucm.c

index 7cfedb8d9bcd963dc81a04ad11eac0322f30eee8..86fee43502cd2a977030d48fad69810fc7f0ac0f 100644 (file)
@@ -34,6 +34,8 @@
  *
  * $Id: cm.c 2821 2005-07-08 17:07:28Z sean.hefty $
  */
+
+#include <linux/completion.h>
 #include <linux/dma-mapping.h>
 #include <linux/err.h>
 #include <linux/idr.h>
@@ -122,7 +124,7 @@ struct cm_id_private {
        struct rb_node service_node;
        struct rb_node sidr_id_node;
        spinlock_t lock;        /* Do not acquire inside cm.lock */
-       wait_queue_head_t wait;
+       struct completion comp;
        atomic_t refcount;
 
        struct ib_mad_send_buf *msg;
@@ -159,7 +161,7 @@ static void cm_work_handler(void *data);
 static inline void cm_deref_id(struct cm_id_private *cm_id_priv)
 {
        if (atomic_dec_and_test(&cm_id_priv->refcount))
-               wake_up(&cm_id_priv->wait);
+               complete(&cm_id_priv->comp);
 }
 
 static int cm_alloc_msg(struct cm_id_private *cm_id_priv,
@@ -559,7 +561,7 @@ struct ib_cm_id *ib_create_cm_id(struct ib_device *device,
                goto error;
 
        spin_lock_init(&cm_id_priv->lock);
-       init_waitqueue_head(&cm_id_priv->wait);
+       init_completion(&cm_id_priv->comp);
        INIT_LIST_HEAD(&cm_id_priv->work_list);
        atomic_set(&cm_id_priv->work_count, -1);
        atomic_set(&cm_id_priv->refcount, 1);
@@ -724,8 +726,8 @@ retest:
        }
 
        cm_free_id(cm_id->local_id);
-       atomic_dec(&cm_id_priv->refcount);
-       wait_event(cm_id_priv->wait, !atomic_read(&cm_id_priv->refcount));
+       cm_deref_id(cm_id_priv);
+       wait_for_completion(&cm_id_priv->comp);
        while ((work = cm_dequeue_work(cm_id_priv)) != NULL)
                cm_free_work(work);
        if (cm_id_priv->private_data && cm_id_priv->private_data_len)
index 469b6923a2e2e0a2c59b388d8f93fedc2e961fac..5ad41a64314cd5b050f67dd4980c6ba0264419a4 100644 (file)
@@ -352,7 +352,7 @@ struct ib_mad_agent *ib_register_mad_agent(struct ib_device *device,
        INIT_WORK(&mad_agent_priv->local_work, local_completions,
                   mad_agent_priv);
        atomic_set(&mad_agent_priv->refcount, 1);
-       init_waitqueue_head(&mad_agent_priv->wait);
+       init_completion(&mad_agent_priv->comp);
 
        return &mad_agent_priv->agent;
 
@@ -467,7 +467,7 @@ struct ib_mad_agent *ib_register_mad_snoop(struct ib_device *device,
        mad_snoop_priv->agent.qp = port_priv->qp_info[qpn].qp;
        mad_snoop_priv->agent.port_num = port_num;
        mad_snoop_priv->mad_snoop_flags = mad_snoop_flags;
-       init_waitqueue_head(&mad_snoop_priv->wait);
+       init_completion(&mad_snoop_priv->comp);
        mad_snoop_priv->snoop_index = register_snoop_agent(
                                                &port_priv->qp_info[qpn],
                                                mad_snoop_priv);
@@ -486,6 +486,18 @@ error1:
 }
 EXPORT_SYMBOL(ib_register_mad_snoop);
 
+static inline void deref_mad_agent(struct ib_mad_agent_private *mad_agent_priv)
+{
+       if (atomic_dec_and_test(&mad_agent_priv->refcount))
+               complete(&mad_agent_priv->comp);
+}
+
+static inline void deref_snoop_agent(struct ib_mad_snoop_private *mad_snoop_priv)
+{
+       if (atomic_dec_and_test(&mad_snoop_priv->refcount))
+               complete(&mad_snoop_priv->comp);
+}
+
 static void unregister_mad_agent(struct ib_mad_agent_private *mad_agent_priv)
 {
        struct ib_mad_port_private *port_priv;
@@ -509,9 +521,8 @@ static void unregister_mad_agent(struct ib_mad_agent_private *mad_agent_priv)
        flush_workqueue(port_priv->wq);
        ib_cancel_rmpp_recvs(mad_agent_priv);
 
-       atomic_dec(&mad_agent_priv->refcount);
-       wait_event(mad_agent_priv->wait,
-                  !atomic_read(&mad_agent_priv->refcount));
+       deref_mad_agent(mad_agent_priv);
+       wait_for_completion(&mad_agent_priv->comp);
 
        kfree(mad_agent_priv->reg_req);
        ib_dereg_mr(mad_agent_priv->agent.mr);
@@ -529,9 +540,8 @@ static void unregister_mad_snoop(struct ib_mad_snoop_private *mad_snoop_priv)
        atomic_dec(&qp_info->snoop_count);
        spin_unlock_irqrestore(&qp_info->snoop_lock, flags);
 
-       atomic_dec(&mad_snoop_priv->refcount);
-       wait_event(mad_snoop_priv->wait,
-                  !atomic_read(&mad_snoop_priv->refcount));
+       deref_snoop_agent(mad_snoop_priv);
+       wait_for_completion(&mad_snoop_priv->comp);
 
        kfree(mad_snoop_priv);
 }
@@ -600,8 +610,7 @@ static void snoop_send(struct ib_mad_qp_info *qp_info,
                spin_unlock_irqrestore(&qp_info->snoop_lock, flags);
                mad_snoop_priv->agent.snoop_handler(&mad_snoop_priv->agent,
                                                    send_buf, mad_send_wc);
-               if (atomic_dec_and_test(&mad_snoop_priv->refcount))
-                       wake_up(&mad_snoop_priv->wait);
+               deref_snoop_agent(mad_snoop_priv);
                spin_lock_irqsave(&qp_info->snoop_lock, flags);
        }
        spin_unlock_irqrestore(&qp_info->snoop_lock, flags);
@@ -626,8 +635,7 @@ static void snoop_recv(struct ib_mad_qp_info *qp_info,
                spin_unlock_irqrestore(&qp_info->snoop_lock, flags);
                mad_snoop_priv->agent.recv_handler(&mad_snoop_priv->agent,
                                                   mad_recv_wc);
-               if (atomic_dec_and_test(&mad_snoop_priv->refcount))
-                       wake_up(&mad_snoop_priv->wait);
+               deref_snoop_agent(mad_snoop_priv);
                spin_lock_irqsave(&qp_info->snoop_lock, flags);
        }
        spin_unlock_irqrestore(&qp_info->snoop_lock, flags);
@@ -968,8 +976,7 @@ void ib_free_send_mad(struct ib_mad_send_buf *send_buf)
 
        free_send_rmpp_list(mad_send_wr);
        kfree(send_buf->mad);
-       if (atomic_dec_and_test(&mad_agent_priv->refcount))
-               wake_up(&mad_agent_priv->wait);
+       deref_mad_agent(mad_agent_priv);
 }
 EXPORT_SYMBOL(ib_free_send_mad);
 
@@ -1757,8 +1764,7 @@ static void ib_mad_complete_recv(struct ib_mad_agent_private *mad_agent_priv,
                mad_recv_wc = ib_process_rmpp_recv_wc(mad_agent_priv,
                                                      mad_recv_wc);
                if (!mad_recv_wc) {
-                       if (atomic_dec_and_test(&mad_agent_priv->refcount))
-                               wake_up(&mad_agent_priv->wait);
+                       deref_mad_agent(mad_agent_priv);
                        return;
                }
        }
@@ -1770,8 +1776,7 @@ static void ib_mad_complete_recv(struct ib_mad_agent_private *mad_agent_priv,
                if (!mad_send_wr) {
                        spin_unlock_irqrestore(&mad_agent_priv->lock, flags);
                        ib_free_recv_mad(mad_recv_wc);
-                       if (atomic_dec_and_test(&mad_agent_priv->refcount))
-                               wake_up(&mad_agent_priv->wait);
+                       deref_mad_agent(mad_agent_priv);
                        return;
                }
                ib_mark_mad_done(mad_send_wr);
@@ -1790,8 +1795,7 @@ static void ib_mad_complete_recv(struct ib_mad_agent_private *mad_agent_priv,
        } else {
                mad_agent_priv->agent.recv_handler(&mad_agent_priv->agent,
                                                   mad_recv_wc);
-               if (atomic_dec_and_test(&mad_agent_priv->refcount))
-                       wake_up(&mad_agent_priv->wait);
+               deref_mad_agent(mad_agent_priv);
        }
 }
 
@@ -2021,8 +2025,7 @@ void ib_mad_complete_send_wr(struct ib_mad_send_wr_private *mad_send_wr,
                                                   mad_send_wc);
 
        /* Release reference on agent taken when sending */
-       if (atomic_dec_and_test(&mad_agent_priv->refcount))
-               wake_up(&mad_agent_priv->wait);
+       deref_mad_agent(mad_agent_priv);
        return;
 done:
        spin_unlock_irqrestore(&mad_agent_priv->lock, flags);
index 6c9c133d71ef465ca1c7df621b1773cd3bdaa5cb..b4fa28d3160fc5abd2dc076e92bfb2f81928fb62 100644 (file)
@@ -37,6 +37,7 @@
 #ifndef __IB_MAD_PRIV_H__
 #define __IB_MAD_PRIV_H__
 
+#include <linux/completion.h>
 #include <linux/pci.h>
 #include <linux/kthread.h>
 #include <linux/workqueue.h>
@@ -108,7 +109,7 @@ struct ib_mad_agent_private {
        struct list_head rmpp_list;
 
        atomic_t refcount;
-       wait_queue_head_t wait;
+       struct completion comp;
 };
 
 struct ib_mad_snoop_private {
@@ -117,7 +118,7 @@ struct ib_mad_snoop_private {
        int snoop_index;
        int mad_snoop_flags;
        atomic_t refcount;
-       wait_queue_head_t wait;
+       struct completion comp;
 };
 
 struct ib_mad_send_wr_private {
index dfd4e588ce031ec8f5edb092f20fddfbc6d322bb..d4704e054e30746cdb5a377c1fa31ffddbdabafd 100644 (file)
@@ -49,7 +49,7 @@ struct mad_rmpp_recv {
        struct list_head list;
        struct work_struct timeout_work;
        struct work_struct cleanup_work;
-       wait_queue_head_t wait;
+       struct completion comp;
        enum rmpp_state state;
        spinlock_t lock;
        atomic_t refcount;
@@ -69,10 +69,16 @@ struct mad_rmpp_recv {
        u8 method;
 };
 
+static inline void deref_rmpp_recv(struct mad_rmpp_recv *rmpp_recv)
+{
+       if (atomic_dec_and_test(&rmpp_recv->refcount))
+               complete(&rmpp_recv->comp);
+}
+
 static void destroy_rmpp_recv(struct mad_rmpp_recv *rmpp_recv)
 {
-       atomic_dec(&rmpp_recv->refcount);
-       wait_event(rmpp_recv->wait, !atomic_read(&rmpp_recv->refcount));
+       deref_rmpp_recv(rmpp_recv);
+       wait_for_completion(&rmpp_recv->comp);
        ib_destroy_ah(rmpp_recv->ah);
        kfree(rmpp_recv);
 }
@@ -253,7 +259,7 @@ create_rmpp_recv(struct ib_mad_agent_private *agent,
                goto error;
 
        rmpp_recv->agent = agent;
-       init_waitqueue_head(&rmpp_recv->wait);
+       init_completion(&rmpp_recv->comp);
        INIT_WORK(&rmpp_recv->timeout_work, recv_timeout_handler, rmpp_recv);
        INIT_WORK(&rmpp_recv->cleanup_work, recv_cleanup_handler, rmpp_recv);
        spin_lock_init(&rmpp_recv->lock);
@@ -279,12 +285,6 @@ error:     kfree(rmpp_recv);
        return NULL;
 }
 
-static inline void deref_rmpp_recv(struct mad_rmpp_recv *rmpp_recv)
-{
-       if (atomic_dec_and_test(&rmpp_recv->refcount))
-               wake_up(&rmpp_recv->wait);
-}
-
 static struct mad_rmpp_recv *
 find_rmpp_recv(struct ib_mad_agent_private *agent,
               struct ib_mad_recv_wc *mad_recv_wc)
index f6a05965a4e8ab79c3529fa71884c792ea8e8c4e..9164a09b6ccd1d0aeea7153d29c8307e3c5903ca 100644 (file)
@@ -32,6 +32,8 @@
  *
  * $Id: ucm.c 2594 2005-06-13 19:46:02Z libor $
  */
+
+#include <linux/completion.h>
 #include <linux/init.h>
 #include <linux/fs.h>
 #include <linux/module.h>
@@ -72,7 +74,7 @@ struct ib_ucm_file {
 
 struct ib_ucm_context {
        int                 id;
-       wait_queue_head_t   wait;
+       struct completion   comp;
        atomic_t            ref;
        int                 events_reported;
 
@@ -138,7 +140,7 @@ static struct ib_ucm_context *ib_ucm_ctx_get(struct ib_ucm_file *file, int id)
 static void ib_ucm_ctx_put(struct ib_ucm_context *ctx)
 {
        if (atomic_dec_and_test(&ctx->ref))
-               wake_up(&ctx->wait);
+               complete(&ctx->comp);
 }
 
 static inline int ib_ucm_new_cm_id(int event)
@@ -178,7 +180,7 @@ static struct ib_ucm_context *ib_ucm_ctx_alloc(struct ib_ucm_file *file)
                return NULL;
 
        atomic_set(&ctx->ref, 1);
-       init_waitqueue_head(&ctx->wait);
+       init_completion(&ctx->comp);
        ctx->file = file;
        INIT_LIST_HEAD(&ctx->events);
 
@@ -586,8 +588,8 @@ static ssize_t ib_ucm_destroy_id(struct ib_ucm_file *file,
        if (IS_ERR(ctx))
                return PTR_ERR(ctx);
 
-       atomic_dec(&ctx->ref);
-       wait_event(ctx->wait, !atomic_read(&ctx->ref));
+       ib_ucm_ctx_put(ctx);
+       wait_for_completion(&ctx->comp);
 
        /* No new events will be generated after destroying the cm_id. */
        ib_destroy_cm_id(ctx->cm_id);