[Qemu-devel] [PATCH 1/5] Revert "iscsi: Fix NULL dereferences / races between task completion and abort"

[Paolo Bonzini <pbonzini@redhat.com>]

This reverts commit 64e69e80920d82df3fa679bc41b13770d2f99360.  The commit
returned immediately from iscsi_aio_cancel, risking corruption in case the
following happens:

    guest                  qemu                 target
  =========================================================================
    send write 1 -------->
                           send write 1 -------->
    cancel write 1 ------>
                           cancel write 1 ------>
       <------------------ cancellation processed
    send write 2 -------->
                           send write 2 -------->
                               <---------------- completed write 2
       <------------------ completed write 2
                               <---------------- completed write 1
                               <---------------- cancellation not done

Here, the guest would see write 2 superseding write 1, when in fact the
outcome could have been the opposite.  The right behavior is to return
only after the target says whether the cancellation was done or not, and
it will be implemented by the next three patches.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
---
 block/iscsi.c | 55 ++++++++++++++++++++++++++++++++-----------------------
 1 file modificato, 32 inserzioni(+), 23 rimozioni(-)

diff --git a/block/iscsi.c b/block/iscsi.c
index bb9cf82..219f927 100644
--- a/block/iscsi.c
+++ b/block/iscsi.c
@@ -76,10 +76,6 @@ static void
 iscsi_abort_task_cb(struct iscsi_context *iscsi, int status, void *command_data,
                     void *private_data)
 {
-    IscsiAIOCB *acb = (IscsiAIOCB *)private_data;
-
-    scsi_free_scsi_task(acb->task);
-    acb->task = NULL;
 }
 
 static void
@@ -88,15 +84,15 @@ iscsi_aio_cancel(BlockDriverAIOCB *blockacb)
     IscsiAIOCB *acb = (IscsiAIOCB *)blockacb;
     IscsiLun *iscsilun = acb->iscsilun;
 
-    acb->canceled = 1;
-
     acb->common.cb(acb->common.opaque, -ECANCELED);
+    acb->canceled = 1;
 
-    /* send a task mgmt call to the target to cancel the task on the target
-     * this also cancels the task in libiscsi
-     */
+    /* send a task mgmt call to the target to cancel the task on the target */
     iscsi_task_mgmt_abort_task_async(iscsilun->iscsi, acb->task,
-                                     iscsi_abort_task_cb, &acb);
+                                     iscsi_abort_task_cb, NULL);
+
+    /* then also cancel the task locally in libiscsi */
+    iscsi_scsi_task_cancel(iscsilun->iscsi, acb->task);
 }
 
 static AIOPool iscsi_aio_pool = {
@@ -183,18 +179,11 @@ iscsi_readv_writev_bh_cb(void *p)
 
     qemu_bh_delete(acb->bh);
 
-    if (!acb->canceled) {
+    if (acb->canceled == 0) {
         acb->common.cb(acb->common.opaque, acb->status);
     }
 
     qemu_aio_release(acb);
-
-    if (acb->canceled) {
-        return;
-    }
-
-    scsi_free_scsi_task(acb->task);
-    acb->task = NULL;
 }
 
 
@@ -208,8 +197,10 @@ iscsi_aio_write16_cb(struct iscsi_context *iscsi, int status,
 
     g_free(acb->buf);
 
-    if (acb->canceled) {
+    if (acb->canceled != 0) {
         qemu_aio_release(acb);
+        scsi_free_scsi_task(acb->task);
+        acb->task = NULL;
         return;
     }
 
@@ -221,6 +212,8 @@ iscsi_aio_write16_cb(struct iscsi_context *iscsi, int status,
     }
 
     iscsi_schedule_bh(iscsi_readv_writev_bh_cb, acb);
+    scsi_free_scsi_task(acb->task);
+    acb->task = NULL;
 }
 
 static int64_t sector_qemu2lun(int64_t sector, IscsiLun *iscsilun)
@@ -305,8 +298,10 @@ iscsi_aio_read16_cb(struct iscsi_context *iscsi, int status,
 
     trace_iscsi_aio_read16_cb(iscsi, status, acb, acb->canceled);
 
-    if (acb->canceled) {
+    if (acb->canceled != 0) {
         qemu_aio_release(acb);
+        scsi_free_scsi_task(acb->task);
+        acb->task = NULL;
         return;
     }
 
@@ -318,6 +313,8 @@ iscsi_aio_read16_cb(struct iscsi_context *iscsi, int status,
     }
 
     iscsi_schedule_bh(iscsi_readv_writev_bh_cb, acb);
+    scsi_free_scsi_task(acb->task);
+    acb->task = NULL;
 }
 
 static BlockDriverAIOCB *
@@ -417,8 +414,10 @@ iscsi_synccache10_cb(struct iscsi_context *iscsi, int status,
 {
     IscsiAIOCB *acb = opaque;
 
-    if (acb->canceled) {
+    if (acb->canceled != 0) {
         qemu_aio_release(acb);
+        scsi_free_scsi_task(acb->task);
+        acb->task = NULL;
         return;
     }
 
@@ -430,6 +429,8 @@ iscsi_synccache10_cb(struct iscsi_context *iscsi, int status,
     }
 
     iscsi_schedule_bh(iscsi_readv_writev_bh_cb, acb);
+    scsi_free_scsi_task(acb->task);
+    acb->task = NULL;
 }
 
 static BlockDriverAIOCB *
@@ -467,8 +468,10 @@ iscsi_unmap_cb(struct iscsi_context *iscsi, int status,
 {
     IscsiAIOCB *acb = opaque;
 
-    if (acb->canceled) {
+    if (acb->canceled != 0) {
         qemu_aio_release(acb);
+        scsi_free_scsi_task(acb->task);
+        acb->task = NULL;
         return;
     }
 
@@ -480,6 +483,8 @@ iscsi_unmap_cb(struct iscsi_context *iscsi, int status,
     }
 
     iscsi_schedule_bh(iscsi_readv_writev_bh_cb, acb);
+    scsi_free_scsi_task(acb->task);
+    acb->task = NULL;
 }
 
 static BlockDriverAIOCB *
@@ -523,8 +528,10 @@ iscsi_aio_ioctl_cb(struct iscsi_context *iscsi, int status,
 {
     IscsiAIOCB *acb = opaque;
 
-    if (acb->canceled) {
+    if (acb->canceled != 0) {
         qemu_aio_release(acb);
+        scsi_free_scsi_task(acb->task);
+        acb->task = NULL;
         return;
     }
 
@@ -553,6 +560,8 @@ iscsi_aio_ioctl_cb(struct iscsi_context *iscsi, int status,
     }
 
     iscsi_schedule_bh(iscsi_readv_writev_bh_cb, acb);
+    scsi_free_scsi_task(acb->task);
+    acb->task = NULL;
 }
 
 static BlockDriverAIOCB *iscsi_aio_ioctl(BlockDriverState *bs,
-- 
1.7.11.2