All of lore.kernel.org
 help / color / mirror / Atom feed
* [bug report] rpmsg: glink: Use the local intents when receiving data
@ 2018-10-01 11:00 Dan Carpenter
  0 siblings, 0 replies; only message in thread
From: Dan Carpenter @ 2018-10-01 11:00 UTC (permalink / raw)
  To: sricharan; +Cc: linux-remoteproc

Hello Sricharan R,

The patch 64f95f87920d: "rpmsg: glink: Use the local intents when
receiving data" from Aug 24, 2017, leads to the following static
checker warning:

	drivers/rpmsg/qcom_glink_native.c:813 qcom_glink_rx_data()
	warn: integer overflows 'chunk_size + left_size'

drivers/rpmsg/qcom_glink_native.c
   781          if (avail < sizeof(hdr)) {
   782                  dev_dbg(glink->dev, "Not enough data in fifo\n");
   783                  return -EAGAIN;
   784          }
   785  
   786          qcom_glink_rx_peak(glink, &hdr, 0, sizeof(hdr));
   787          chunk_size = le32_to_cpu(hdr.chunk_size);
   788          left_size = le32_to_cpu(hdr.left_size);
                            ^^^^^^^^^^^^^^^^^^^^^^^^^^
My private build of Smatch marks le32_to_cpu() as untrusted.

   789  
   790          if (avail < sizeof(hdr) + chunk_size) {
                            ^^^^^^^^^^^^^^^^^^^^^^^^
This could have an integer overflow.

   791                  dev_dbg(glink->dev, "Payload not yet in fifo\n");
   792                  return -EAGAIN;
   793          }
   794  
   795          rcid = le16_to_cpu(hdr.msg.param1);
   796          spin_lock_irqsave(&glink->idr_lock, flags);
   797          channel = idr_find(&glink->rcids, rcid);
   798          spin_unlock_irqrestore(&glink->idr_lock, flags);
   799          if (!channel) {
   800                  dev_dbg(glink->dev, "Data on non-existing channel\n");
   801  
   802                  /* Drop the message */
   803                  goto advance_rx;
   804          }
   805  
   806          if (glink->intentless) {
   807                  /* Might have an ongoing, fragmented, message to append */
   808                  if (!channel->buf) {
   809                          intent = kzalloc(sizeof(*intent), GFP_ATOMIC);
   810                          if (!intent)
   811                                  return -ENOMEM;
   812  
   813                          intent->data = kmalloc(chunk_size + left_size,
                                                       ^^^^^^^^^^^^^^^^^^^^^^
We haven't check "left_size" at all so Smatch is very suspicious of
this.

   814                                                 GFP_ATOMIC);
   815                          if (!intent->data) {
   816                                  kfree(intent);
   817                                  return -ENOMEM;
   818                          }
   819  

See also:

drivers/rpmsg/qcom_smd.c:1227 qcom_channel_scan_worker() warn: potential integer overflow from user 'smem_items[tbl].info_base_id + cid'
drivers/rpmsg/qcom_smd.c:1228 qcom_channel_scan_worker() warn: potential integer overflow from user 'smem_items[tbl].fifo_base_id + cid'

regards,
dan carpenter

^ permalink raw reply	[flat|nested] only message in thread

only message in thread, other threads:[~2018-10-01 17:37 UTC | newest]

Thread overview: (only message) (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2018-10-01 11:00 [bug report] rpmsg: glink: Use the local intents when receiving data Dan Carpenter

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.