[PATCH 4.9 34/41] tpm: fix a kernel memory leak in tpm-sysfs.c

[Greg Kroah-Hartman]

4.9-stable review patch.  If anyone has any objections, please let me know.

------------------

From: Jarkko Sakkinen <jarkko.sakkinen@xxxxxxxxxxxxxxx>

commit 13b47cfcfc60495cde216eef4c01040d76174cbe upstream.

While cleaning up sysfs callback that prints EK we discovered a kernel
memory leak. This commit fixes the issue by zeroing the buffer used for
TPM command/response.

The leak happen when we use either tpm_vtpm_proxy, tpm_ibmvtpm or
xen-tpmfront.

Fixes: 0883743825e3 ("TPM: sysfs functions consolidation")
Reported-by: Jason Gunthorpe <jgunthorpe@xxxxxxxxxxxxxxxxxxxx>
Tested-by: Stefan Berger <stefanb@xxxxxxxxxxxxxxxxxx>
Signed-off-by: Jarkko Sakkinen <jarkko.sakkinen@xxxxxxxxxxxxxxx>
Signed-off-by: James Morris <james.l.morris@xxxxxxxxxx>
Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>

---
 drivers/char/tpm/tpm-sysfs.c |    3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

--- a/drivers/char/tpm/tpm-sysfs.c
+++ b/drivers/char/tpm/tpm-sysfs.c
@@ -35,9 +35,10 @@ static ssize_t pubek_show(struct device
 	ssize_t err;
 	int i, rc;
 	char *str = buf;
-
 	struct tpm_chip *chip = to_tpm_chip(dev);
 
+	memset(&tpm_cmd, 0, sizeof(tpm_cmd));
+
 	tpm_cmd.header.in = tpm_readpubek_header;
 	err = tpm_transmit_cmd(chip, &tpm_cmd, READ_PUBEK_RESULT_SIZE, 0,
 			       "attempting to read the PUBEK");