* [PATCH] Bad IOCTLs - time to fix them?
@ 2020-12-28 15:21 Hans Petter Selasky
2021-03-23 13:20 ` Mauro Carvalho Chehab
0 siblings, 1 reply; 2+ messages in thread
From: Hans Petter Selasky @ 2020-12-28 15:21 UTC (permalink / raw)
To: linux-media
Hi,
For some years now, FreeBSD's webcamd project at GitHub has carried
patches to fix up IOCTL's in the Linux Kernel and I'm reaching out once
again, that the bad read/write bits for certain IOCTLs may be seen as a
security issue and should be fixed. These are:
diff --git a/include/uapi/linux/dvb/ca.h b/include/uapi/linux/dvb/ca.h
index dffa59e95ebb..c4533a16d93e 100644
--- a/include/uapi/linux/dvb/ca.h
+++ b/include/uapi/linux/dvb/ca.h
@@ -134,9 +134,17 @@ struct ca_descr {
#define CA_RESET _IO('o', 128)
#define CA_GET_CAP _IOR('o', 129, struct ca_caps)
+/*
+ * At least CA_GET_SLOT_INFO and CA_GET_MSG need to be _IOWR not _IOR.
+ * This is wrong on Linux too but there the driver doesn't care.
+ *
#define CA_GET_SLOT_INFO _IOR('o', 130, struct ca_slot_info)
#define CA_GET_DESCR_INFO _IOR('o', 131, struct ca_descr_info)
#define CA_GET_MSG _IOR('o', 132, struct ca_msg)
+*/
+#define CA_GET_SLOT_INFO _IOWR('o', 130, struct ca_slot_info)
+#define CA_GET_DESCR_INFO _IOR('o', 131, struct ca_descr_info)
+#define CA_GET_MSG _IOWR('o', 132, struct ca_msg)
#define CA_SEND_MSG _IOW('o', 133, struct ca_msg)
#define CA_SET_DESCR _IOW('o', 134, struct ca_descr)
diff --git a/include/uapi/linux/dvb/frontend.h
b/include/uapi/linux/dvb/frontend.h
index 4f9b4551c534..a5e950404072 100644
--- a/include/uapi/linux/dvb/frontend.h
+++ b/include/uapi/linux/dvb/frontend.h
@@ -908,7 +908,8 @@ struct dtv_properties {
#define FE_DISHNETWORK_SEND_LEGACY_CMD _IO('o', 80) /* unsigned int */
#define FE_SET_PROPERTY _IOW('o', 82, struct dtv_properties)
-#define FE_GET_PROPERTY _IOR('o', 83, struct dtv_properties)
+#define FE_GET_PROPERTY _IOW('o', 83, struct dtv_properties)
+#define FE_GET_PROPERTY_OLD _IOR('o', 83, struct dtv_properties)
#if defined(__DVB_CORE__) || !defined(__KERNEL__)
diff --git a/include/uapi/linux/input.h b/include/uapi/linux/input.h
index 9a61c28ed3ae..94cc0a7cb0a5 100644
--- a/include/uapi/linux/input.h
+++ b/include/uapi/linux/input.h
@@ -131,8 +131,12 @@ struct input_mask {
#define EVIOCGREP _IOR('E', 0x03, unsigned int[2]) /* get repeat
settings */
#define EVIOCSREP _IOW('E', 0x03, unsigned int[2]) /* set repeat
settings */
-#define EVIOCGKEYCODE _IOR('E', 0x04, unsigned int[2]) /* get
keycode */
-#define EVIOCGKEYCODE_V2 _IOR('E', 0x04, struct input_keymap_entry)
+/*
+ * These two need to be _IOWR not _IOR (they're wrong on Linux too but
+ * there the driver doesn't care.)
+ */
+#define EVIOCGKEYCODE _IOWR('E', 0x04, unsigned int[2]) /* get
keycode */
+#define EVIOCGKEYCODE_V2 _IOWR('E', 0x04, struct input_keymap_entry)
#define EVIOCSKEYCODE _IOW('E', 0x04, unsigned int[2]) /* set
keycode */
#define EVIOCSKEYCODE_V2 _IOW('E', 0x04, struct input_keymap_entry)
@@ -165,7 +169,7 @@ struct input_mask {
*
* If the request code is not an ABS_MT value, -EINVAL is returned.
*/
-#define EVIOCGMTSLOTS(len) _IOC(_IOC_READ, 'E', 0x0a, len)
+#define EVIOCGMTSLOTS(len) _IOC(_IOC_READ|_IOC_WRITE, 'E', 0x0a, len)
#define EVIOCGKEY(len) _IOC(_IOC_READ, 'E', 0x18, len) /* get global
key state */
#define EVIOCGLED(len) _IOC(_IOC_READ, 'E', 0x19, len) /* get all
LEDs */
@@ -177,11 +181,11 @@ struct input_mask {
#define EVIOCSABS(abs) _IOW('E', 0xc0 + (abs), struct input_absinfo)
/* set abs value/limits */
#define EVIOCSFF _IOW('E', 0x80, struct ff_effect) /* send a force
effect to a force feedback device */
-#define EVIOCRMFF _IOW('E', 0x81, int) /* Erase a force effect */
+#define EVIOCRMFF _IOWINT('E', 0x81) /* Erase a force effect */
#define EVIOCGEFFECTS _IOR('E', 0x84, int) /* Report number of
effects playable at the same time */
-#define EVIOCGRAB _IOW('E', 0x90, int) /* Grab/Release device */
-#define EVIOCREVOKE _IOW('E', 0x91, int) /* Revoke device access */
+#define EVIOCGRAB _IOWINT('E', 0x90) /* Grab/Release device */
+#define EVIOCREVOKE _IOWINT('E', 0x91) /* Revoke device access */
/**
* EVIOCGMASK - Retrieve current event mask
@@ -212,7 +216,7 @@ struct input_mask {
* if the receive-buffer points to invalid memory, or EINVAL if the kernel
* does not implement the ioctl.
*/
-#define EVIOCGMASK _IOR('E', 0x92, struct input_mask) /* Get
event-masks */
+#define EVIOCGMASK _IOW('E', 0x92, struct input_mask) /* Get
event-masks */
/**
* EVIOCSMASK - Set event mask
--
2.29.2
^ permalink raw reply related [flat|nested] 2+ messages in thread* Re: [PATCH] Bad IOCTLs - time to fix them?
2020-12-28 15:21 [PATCH] Bad IOCTLs - time to fix them? Hans Petter Selasky
@ 2021-03-23 13:20 ` Mauro Carvalho Chehab
0 siblings, 0 replies; 2+ messages in thread
From: Mauro Carvalho Chehab @ 2021-03-23 13:20 UTC (permalink / raw)
To: Hans Petter Selasky; +Cc: linux-media
Em Mon, 28 Dec 2020 16:21:05 +0100
Hans Petter Selasky <hps@selasky.org> escreveu:
> Hi,
>
> For some years now, FreeBSD's webcamd project at GitHub has carried
> patches to fix up IOCTL's in the Linux Kernel and I'm reaching out once
> again, that the bad read/write bits for certain IOCTLs may be seen as a
> security issue and should be fixed. These are:
>
> diff --git a/include/uapi/linux/dvb/ca.h b/include/uapi/linux/dvb/ca.h
> index dffa59e95ebb..c4533a16d93e 100644
> --- a/include/uapi/linux/dvb/ca.h
> +++ b/include/uapi/linux/dvb/ca.h
> @@ -134,9 +134,17 @@ struct ca_descr {
>
> #define CA_RESET _IO('o', 128)
> #define CA_GET_CAP _IOR('o', 129, struct ca_caps)
> +/*
> + * At least CA_GET_SLOT_INFO and CA_GET_MSG need to be _IOWR not _IOR.
> + * This is wrong on Linux too but there the driver doesn't care.
> + *
> #define CA_GET_SLOT_INFO _IOR('o', 130, struct ca_slot_info)
> #define CA_GET_DESCR_INFO _IOR('o', 131, struct ca_descr_info)
> #define CA_GET_MSG _IOR('o', 132, struct ca_msg)
> +*/
> +#define CA_GET_SLOT_INFO _IOWR('o', 130, struct ca_slot_info)
> +#define CA_GET_DESCR_INFO _IOR('o', 131, struct ca_descr_info)
> +#define CA_GET_MSG _IOWR('o', 132, struct ca_msg)
> #define CA_SEND_MSG _IOW('o', 133, struct ca_msg)
> #define CA_SET_DESCR _IOW('o', 134, struct ca_descr)
>
> diff --git a/include/uapi/linux/dvb/frontend.h
> b/include/uapi/linux/dvb/frontend.h
> index 4f9b4551c534..a5e950404072 100644
> --- a/include/uapi/linux/dvb/frontend.h
> +++ b/include/uapi/linux/dvb/frontend.h
> @@ -908,7 +908,8 @@ struct dtv_properties {
> #define FE_DISHNETWORK_SEND_LEGACY_CMD _IO('o', 80) /* unsigned int */
>
> #define FE_SET_PROPERTY _IOW('o', 82, struct dtv_properties)
> -#define FE_GET_PROPERTY _IOR('o', 83, struct dtv_properties)
> +#define FE_GET_PROPERTY _IOW('o', 83, struct dtv_properties)
> +#define FE_GET_PROPERTY_OLD _IOR('o', 83, struct dtv_properties)
>
> #if defined(__DVB_CORE__) || !defined(__KERNEL__)
Commenting just the issues related to the media subsystem. You should
forward the evdev ones to linux-input@vger.kernel.org.
This has been wrongly defined for so long time that I'm not sure if
it makes sense to fix.
If this is touched, we need to have _OLD versions for all the ioctls,
and the code also need to be changed in order to accept both ways.
So, at the current stage, this patch seems incomplete.
Thanks,
Mauro
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2021-03-23 13:21 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2020-12-28 15:21 [PATCH] Bad IOCTLs - time to fix them? Hans Petter Selasky
2021-03-23 13:20 ` Mauro Carvalho Chehab
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.