commit 2ab4f4018cb6b8010ca5002c3bdc37783b5d28c2 upstream.
When mailboxes are used as a transport it is possible to setup the SCMI transport layer, depending on the underlying channels configuration, to use one or two mailboxes, associated, respectively, to one or two, distinct, shared memory areas: any other combination should be treated as invalid.
Add more strict checking of SCMI mailbox transport device node descriptors.
Fixes: fbc4d81ad285 ("firmware: arm_scmi: refactor in preparation to support per-protocol channels") Cc: stable@vger.kernel.org # 4.19 Signed-off-by: Cristian Marussi cristian.marussi@arm.com Link: https://lore.kernel.org/r/20230307162324.891866-1-cristian.marussi@arm.com Signed-off-by: Sudeep Holla sudeep.holla@arm.com [Cristian: backported to v4.19] Signed-off-by: Cristian Marussi cristian.marussi@arm.com --- Backporting was trivial but required since the patched function was moved around in a different file. --- drivers/firmware/arm_scmi/driver.c | 37 ++++++++++++++++++++++++++++++ 1 file changed, 37 insertions(+)
diff --git a/drivers/firmware/arm_scmi/driver.c b/drivers/firmware/arm_scmi/driver.c index e8cd66705ad7..5ccbbb3eb68e 100644 --- a/drivers/firmware/arm_scmi/driver.c +++ b/drivers/firmware/arm_scmi/driver.c @@ -705,6 +705,39 @@ static int scmi_remove(struct platform_device *pdev) return ret; }
+static int scmi_mailbox_chan_validate(struct device *cdev) +{ + int num_mb, num_sh, ret = 0; + struct device_node *np = cdev->of_node; + + num_mb = of_count_phandle_with_args(np, "mboxes", "#mbox-cells"); + num_sh = of_count_phandle_with_args(np, "shmem", NULL); + /* Bail out if mboxes and shmem descriptors are inconsistent */ + if (num_mb <= 0 || num_sh > 2 || num_mb != num_sh) { + dev_warn(cdev, "Invalid channel descriptor for '%s'\n", + of_node_full_name(np)); + return -EINVAL; + } + + if (num_sh > 1) { + struct device_node *np_tx, *np_rx; + + np_tx = of_parse_phandle(np, "shmem", 0); + np_rx = of_parse_phandle(np, "shmem", 1); + /* SCMI Tx and Rx shared mem areas have to be distinct */ + if (!np_tx || !np_rx || np_tx == np_rx) { + dev_warn(cdev, "Invalid shmem descriptor for '%s'\n", + of_node_full_name(np)); + ret = -EINVAL; + } + + of_node_put(np_tx); + of_node_put(np_rx); + } + + return ret; +} + static inline int scmi_mbox_chan_setup(struct scmi_info *info, struct device *dev, int prot_id) { @@ -720,6 +753,10 @@ scmi_mbox_chan_setup(struct scmi_info *info, struct device *dev, int prot_id) goto idr_alloc; }
+ ret = scmi_mailbox_chan_validate(dev); + if (ret) + return ret; + cinfo = devm_kzalloc(info->dev, sizeof(*cinfo), GFP_KERNEL); if (!cinfo) return -ENOMEM;
On Fri, Mar 31, 2023 at 11:49:55AM +0100, Cristian Marussi wrote:
commit 2ab4f4018cb6b8010ca5002c3bdc37783b5d28c2 upstream.
When mailboxes are used as a transport it is possible to setup the SCMI transport layer, depending on the underlying channels configuration, to use one or two mailboxes, associated, respectively, to one or two, distinct, shared memory areas: any other combination should be treated as invalid.
Add more strict checking of SCMI mailbox transport device node descriptors.
Fixes: fbc4d81ad285 ("firmware: arm_scmi: refactor in preparation to support per-protocol channels") Cc: stable@vger.kernel.org # 4.19 Signed-off-by: Cristian Marussi cristian.marussi@arm.com Link: https://lore.kernel.org/r/20230307162324.891866-1-cristian.marussi@arm.com Signed-off-by: Sudeep Holla sudeep.holla@arm.com [Cristian: backported to v4.19] Signed-off-by: Cristian Marussi cristian.marussi@arm.com
Backporting was trivial but required since the patched function was moved around in a different file.
Both backports now queued up, thanks.
greg k-h
linux-stable-mirror@lists.linaro.org