Re: [PATCH] platform/chrome: cros_ec_dev - Fix security issue

On Thu, Mar 18, 2021 at 1:20 AM Lee Jones lee.jones@linaro.org wrote:

On Thu, 18 Mar 2021, Greg KH wrote:

...

On Thu, Mar 18, 2021 at 07:54:43AM +0000, Lee Jones wrote:

...

On Wed, 17 Mar 2021, Gwendal Grignou wrote:

...

commit 5d749d0bbe811c10d9048cde6dfebc761713abfd upstream.

Prevent memory scribble by checking that ioctl buffer size parameters are sane. Without this check, on 32 bits system, if .insize = 0xffffffff - 20 and .outsize the amount to scribble, we would overflow, allocate a small amounts and be able to write outside of the malloc'ed area. Adding a hard limit allows argument checking of the ioctl. With the current EC, it is expected .insize and .outsize to be at around 512 bytes or less.

Signed-off-by: Olof Johansson olof@lixom.net Signed-off-by: Gwendal Grignou gwendal@chromium.org

---

drivers/platform/chrome/cros_ec_dev.c | 4 ++++ drivers/platform/chrome/cros_ec_proto.c | 4 ++-- include/linux/mfd/cros_ec.h | 6 ++++-- 3 files changed, 10 insertions(+), 4 deletions(-)

diff --git a/drivers/platform/chrome/cros_ec_dev.c b/drivers/platform/chrome/cros_ec_dev.c index 2b331d5b9e799..e16d82bb36a9d 100644 --- a/drivers/platform/chrome/cros_ec_dev.c +++ b/drivers/platform/chrome/cros_ec_dev.c @@ -137,6 +137,10 @@ static long ec_device_ioctl_xcmd(struct cros_ec_dev *ec, void __user *arg) if (copy_from_user(&u_cmd, arg, sizeof(u_cmd))) return -EFAULT;

• if ((u_cmd.outsize > EC_MAX_MSG_BYTES) ||

• (u_cmd.insize > EC_MAX_MSG_BYTES))
  

•     return -EINVAL;
  

• s_cmd = kmalloc(sizeof(*s_cmd) + max(u_cmd.outsize, u_cmd.insize), GFP_KERNEL); if (!s_cmd)

diff --git a/drivers/platform/chrome/cros_ec_proto.c b/drivers/platform/chrome/cros_ec_proto.c index 5c285f2b3a650..d20190c8f0c06 100644 --- a/drivers/platform/chrome/cros_ec_proto.c +++ b/drivers/platform/chrome/cros_ec_proto.c @@ -311,8 +311,8 @@ int cros_ec_query_all(struct cros_ec_device *ec_dev) ec_dev->max_response = EC_PROTO2_MAX_PARAM_SIZE; ec_dev->max_passthru = 0; ec_dev->pkt_xfer = NULL;

•             ec_dev->din_size = EC_MSG_BYTES;
  

•             ec_dev->dout_size = EC_MSG_BYTES;
  

•             ec_dev->din_size = EC_PROTO2_MSG_BYTES;
  

•             ec_dev->dout_size = EC_PROTO2_MSG_BYTES;
      } else {
              /*
               * It's possible for a test to occur too early when
  

diff --git a/include/linux/mfd/cros_ec.h b/include/linux/mfd/cros_ec.h index 3ab3cede28eac..93c14e9df6309 100644 --- a/include/linux/mfd/cros_ec.h +++ b/include/linux/mfd/cros_ec.h @@ -50,9 +50,11 @@ enum { EC_MSG_TX_TRAILER_BYTES, EC_MSG_RX_PROTO_BYTES = 3,

• /* Max length of messages */
• EC_MSG_BYTES = EC_PROTO2_MAX_PARAM_SIZE +

• /* Max length of messages for proto 2*/
• EC_PROTO2_MSG_BYTES = EC_PROTO2_MAX_PARAM_SIZE + EC_MSG_TX_PROTO_BYTES,

Nit: Better to not tab the '=' so far and place it all on one line.

Checkpatch now only complains about lines exceeding 100 chars.

Once fixed, feel free to apply my:

Acked-by: Lee Jones lee.jones@linaro.org

This commit is already in 4.7, and is from 2016, so I don't know why you are reviewing it :)

My mistake, I overlooked the branch information is removed when formatting a patch: it is for branch linux-4.4.y in linux stable git tree.

Heh! It looked like a standard patch at first glance.

Must have skipped over the "commit" line in the commit log.

I wonder why it doesn't have my Ack on it already then?

You were not in the loop on the original thread: https://lkml.org/lkml/2016/3/8/628 Sorry about that,

Gwendal.

-- Lee Jones [李琼斯] Senior Technical Lead - Developer Services Linaro.org │ Open source software for Arm SoCs Follow Linaro: Facebook | Twitter | Blog