On 11/15/2017 9:42 PM, Bjorn Andersson wrote:
On Wed 15 Nov 12:10 PST 2017, Bjorn Andersson wrote:
diff --git a/drivers/soc/qcom/qmi_encdec.c b/drivers/soc/qcom/qmi_encdec.c
[..]
+void *qmi_encode_message(int type, unsigned int msg_id, size_t *len,
+ unsigned int txn_id, struct qmi_elem_info *ei,
+ const void *c_struct)
+{
+ struct qmi_header *hdr;
+ ssize_t msglen = 0;
+ void *msg;
+ int ret;
+
+ /* Check the possibility of a zero length QMI message */
+ if (!c_struct) {
+ ret = qmi_calc_min_msg_len(ei, 1);
+ if (ret) {
+ pr_err("%s: Calc. len %d != 0, but NULL c_struct\n",
+ __func__, ret);
+ return ERR_PTR(-EINVAL);
+ }
+ }
+
+ msg = kzalloc(sizeof(*hdr) + *len, GFP_KERNEL);
+ if (!msg)
+ return ERR_PTR(-ENOMEM);
+
+ msglen = qmi_encode(ei, msg + sizeof(*hdr), c_struct, *len, 1);
+ if (msglen < 0) {
+ kfree(msg);
+ return ERR_PTR(msglen);
+ }
Talked to Chris Lew about this earlier today;
The check above implies that it's valid to call this function with a
valid ei of minimum message length of 0 and c_struct being NULL. But the
call to qmi_encdec() will dereference c_struct in order to know that the
optional elements described in ei are unset.
So the call to qmi_encode() needs to only be done conditionally on
c_struct being non-NULL, logically interpreting c_struct being NULL as
all optional fields are unset.
Will post an update with this fixed.
I have tested this patch with QMI loopback servers, Qualcomm diag and
slimbus ngd.
Tested-By: Chris Lew <c...@codeaurora.org>
+
+ hdr = msg;
+ hdr->type = type;
+ hdr->txn_id = txn_id;
+ hdr->msg_id = msg_id;
+ hdr->msg_len = msglen;
+
+ *len = sizeof(*hdr) + msglen;
+
+ return msg;
+}
+EXPORT_SYMBOL(qmi_encode_message);
Regards,
Bjorn
--
Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum,
a Linux Foundation Collaborative Project
