scsi: vmw_pscsi: Fix use-after-free in pvscsi_queue_lck()
Subsystems (3)
PATCH 1/1 scsi: vmw_pscsi: Fix use-after-free in pvscsi_queue_lck()
Once we unlock adapter->hw_lock in pvscsi_queue_lck() nothing prevents just queued scsi_cmnd from completing and freeing the request. Thus cmd->cmnd[0] dereference can dereference already freed request leading to kernel crashes or other issues (which one of our customers observed). Store cmd->cmnd[0] in a local variable before unlocking adapter->hw_lock to fix the issue.
--- drivers/scsi/vmw_pvscsi.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-)
drivers/scsi/vmw_pvscsi.c +4 -2
@@ -763,6 +763,7 @@ static int pvscsi_queue_lck(struct scsi_cmnd *cmd, void (*done)(struct scsi_cmndstruct pvscsi_adapter *adapter =shost_priv (host);struct pvscsi_ctx *ctx;unsigned long flags; +unsigned char op;spin_lock_irqsave (&adapter->hw_lock, flags);
@@ -775,13 +776,14 @@ static int pvscsi_queue_lck(struct scsi_cmnd *cmd, void (*done)(struct scsi_cmnd } cmd->scsi_done = done; + op = cmd->cmnd[0 ];dev_dbg (&cmd->device->sdev_gendev, -"queued cmd %p, ctx %p, op=%x\n" , cmd, ctx, cmd->cmnd[0 ]); +"queued cmd %p, ctx %p, op=%x\n" , cmd, ctx, op);spin_unlock_irqrestore (&adapter->hw_lock, flags); -pvscsi_kick_io (adapter, cmd->cmnd[0 ]); +pvscsi_kick_io (adapter, op);return 0 ; }
Signature · 1 line
2.16.4
Quoted 31 lines · Jan Kara
Once we unlock
adapter->hw_lockinpvscsi_queue_lck()nothing prevents just queuedscsi_cmndfrom completing and freeing the request. Thus cmd->cmnd[0] dereference can dereference already freed request leading to kernel crashes or other issues (which one of our customers observed). Store cmd->cmnd[0] in a local variable before unlockingadapter->hw_lockto fix the issue.Signed-off-by Jan Kara <[email protected]>--- drivers/scsi/vmw_pvscsi.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-)drivers/scsi/vmw_pvscsi.c +2
-0@@ -763,6 +763,7 @@ static int pvscsi_queue_lck(struct scsi_cmnd *cmd, void (*done)(struct scsi_cmndstruct pvscsi_adapter *adapter =shost_priv (host);struct pvscsi_ctx *ctx;unsigned long flags; +unsigned char op;spin_lock_irqsave (&adapter->hw_lock, flags);@@ -775,13 +776,14 @@ static int pvscsi_queue_lck(struct scsi_cmnd *cmd, void (*done)(struct scsi_cmnd } cmd->scsi_done = done; + op = cmd->cmnd[0 ]; dev_dbg(&cmd->device->sdev_gendev,
drivers/scsi/vmw_pvscsi.c static int pvscsi_queue_lck(struct scsi_cmnd *cmd, void (*done)(struct scsi_cmnd · line 789
Quoted 10 lines
drivers/scsi/vmw_pvscsi.c +2
-2- "queued cmd %p, ctx %p, op=%x\n", cmd, ctx, cmd->cmnd[0]); + "queued cmd %p, ctx %p, op=%x\n", cmd, ctx, op);spin_unlock_irqrestore (&adapter->hw_lock, flags); -pvscsi_kick_io (adapter, cmd->cmnd[0 ]); +pvscsi_kick_io (adapter, op);return 0 ; }
Martin K. Petersen
AppliedJan,
Quoted 6 lines
Once we unlock
adapter->hw_lockinpvscsi_queue_lck()nothing prevents just queuedscsi_cmndfrom completing and freeing the request. Thus cmd->cmnd[0] dereference can dereference already freed request leading to kernel crashes or other issues (which one of our customers observed). Store cmd->cmnd[0] in a local variable before unlockingadapter->hw_lockto fix the issue.
Applied to 5.2/scsi-fixes. Thanks!
Signature · 1 line
Martin K. Petersen Oracle Linux Engineering