[PATCH 1/2] TPM: Missing put_device()

[PATCH 1/2] TPM: Missing put_device()

[Rajiv Andrade]

In case one calling tpm_open() whilst other process is owning it,
the refcount wouldn't get 0.

This patch depends on [PATCH 3/4][resubmit] TPM: rcu locking, sent on 23 Sep 2008.

Signed-off-by: Mimi Zohar <zohar@linux.vnet.ibm.com>
Signed-off-by: Rajiv Andrade <srajiv@linux.vnet.ibm.com>
---
 drivers/char/tpm/tpm.c |    1 +
 1 files changed, 1 insertions(+), 0 deletions(-)

diff --git a/drivers/char/tpm/tpm.c b/drivers/char/tpm/tpm.c
index f6bbc1d..ab03b4d 100644
--- a/drivers/char/tpm/tpm.c
+++ b/drivers/char/tpm/tpm.c
@@ -979,6 +979,7 @@ int tpm_open(struct inode *inode, struct file *file)
 
 	if (test_and_set_bit(0, &chip->is_open)) {
 		dev_dbg(chip->dev, "Another process owns this TPM\n");
+		put_device(chip->dev);
 		return -EBUSY;
 	}
 
-- 
1.5.6.3

[PATCH 2/2] TPM: Fixed tpm release timing

[Rajiv Andrade]

As pointed by Jonathan Corbet, the timer must be deleted before flushing
the work queue in order to avoid a job being submitted after 
the chip had been released.

Signed-off-by: Rajiv Andrade <srajiv@linux.vnet.ibm.com>
Signed-off-by: Mimi Zohar <zohar@linux.vnet.ibm.com>
CC: Jonathan Corbet <corbet@lwn.net>
---
 drivers/char/tpm/tpm.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/drivers/char/tpm/tpm.c b/drivers/char/tpm/tpm.c
index ab03b4d..8d090d3 100644
--- a/drivers/char/tpm/tpm.c
+++ b/drivers/char/tpm/tpm.c
@@ -1004,9 +1004,9 @@ int tpm_release(struct inode *inode, struct file *file)
 {
 	struct tpm_chip *chip = file->private_data;
 
+	del_singleshot_timer_sync(&chip->user_read_timer);
 	flush_scheduled_work();
 	file->private_data = NULL;
-	del_singleshot_timer_sync(&chip->user_read_timer);
 	atomic_set(&chip->data_pending, 0);
 	kfree(chip->data_buffer);
 	clear_bit(0, &chip->is_open);
-- 
1.5.6.3