From: Colin Ian King colin.king@canonical.com
Various usif_ioctl helper functions do not initialize the return variable ret and some of the error handling return paths just return garbage values that were on the stack (or in a register). I believe that in all the cases, the initial ret variable should be set to -EINVAL and subsequent paths through these helper functions set it appropriately otherwise.
Found via static analysis using cppcheck:
[drivers/gpu/drm/nouveau/nouveau_usif.c:138]: (error) Uninitialized variable: ret [drivers/gpu/drm/nouveau/nouveau_usif.c:179]: (error) Uninitialized variable: ret [drivers/gpu/drm/nouveau/nouveau_usif.c:202]: (error) Uninitialized variable: ret [drivers/gpu/drm/nouveau/nouveau_usif.c:241]: (error) Uninitialized variable: ret [drivers/gpu/drm/nouveau/nouveau_usif.c:157]: (error) Uninitialized variable: ret [drivers/gpu/drm/nouveau/nouveau_usif.c:288]: (error) Uninitialized variable: ret
Signed-off-by: Colin Ian King colin.king@canonical.com --- drivers/gpu/drm/nouveau/nouveau_usif.c | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-)
diff --git a/drivers/gpu/drm/nouveau/nouveau_usif.c b/drivers/gpu/drm/nouveau/nouveau_usif.c index cb1182d..01b50a2 100644 --- a/drivers/gpu/drm/nouveau/nouveau_usif.c +++ b/drivers/gpu/drm/nouveau/nouveau_usif.c @@ -129,7 +129,7 @@ usif_notify_new(struct drm_file *f, void *data, u32 size, void *argv, u32 argc) struct nvif_notify_req_v0 v0; } *req; struct usif_notify *ntfy; - int ret; + int ret = -EINVAL;
if (nvif_unpack(args->v0, 0, 0, true)) { if (usif_notify_find(f, args->v0.index)) @@ -170,7 +170,7 @@ usif_notify_del(struct drm_file *f, void *data, u32 size, void *argv, u32 argc) struct nvif_ioctl_ntfy_del_v0 v0; } *args = data; struct usif_notify *ntfy; - int ret; + int ret = -EINVAL;
if (nvif_unpack(args->v0, 0, 0, true)) { if (!(ntfy = usif_notify_find(f, args->v0.index))) @@ -193,7 +193,7 @@ usif_notify_get(struct drm_file *f, void *data, u32 size, void *argv, u32 argc) struct nvif_ioctl_ntfy_del_v0 v0; } *args = data; struct usif_notify *ntfy; - int ret; + int ret = -EINVAL;
if (nvif_unpack(args->v0, 0, 0, true)) { if (!(ntfy = usif_notify_find(f, args->v0.index))) @@ -232,7 +232,7 @@ usif_notify_put(struct drm_file *f, void *data, u32 size, void *argv, u32 argc) struct nvif_ioctl_ntfy_put_v0 v0; } *args = data; struct usif_notify *ntfy; - int ret; + int ret = -EINVAL;
if (nvif_unpack(args->v0, 0, 0, true)) { if (!(ntfy = usif_notify_find(f, args->v0.index))) @@ -269,7 +269,7 @@ usif_object_new(struct drm_file *f, void *data, u32 size, void *argv, u32 argc) struct nvif_ioctl_new_v0 v0; } *args = data; struct usif_object *object; - int ret; + int ret = -EINVAL;
if (!(object = kmalloc(sizeof(*object), GFP_KERNEL))) return -ENOMEM;
On Wed, Jul 1, 2015 at 12:51 PM, Colin King colin.king@canonical.com wrote:
From: Colin Ian King colin.king@canonical.com
Various usif_ioctl helper functions do not initialize the return variable ret and some of the error handling return paths just return garbage values that were on the stack (or in a register). I believe that in all the cases, the initial ret variable should be set to -EINVAL and subsequent paths through these helper functions set it appropriately otherwise.
Found via static analysis using cppcheck:
[drivers/gpu/drm/nouveau/nouveau_usif.c:138]: (error) Uninitialized variable: ret
It sure would seem that way, wouldn't it?
#define nvif_unpack(d,vl,vh,m) ({ \ if ((vl) == 0 || ret == -ENOSYS) { \ int _size = sizeof(d); \ if (_size <= size && (d).version >= (vl) && \ (d).version <= (vh)) { \ data = (u8 *)data + _size; \ size = size - _size; \ ret = ((m) || !size) ? 0 : -E2BIG; \ } else { \ ret = -ENOSYS; \ } \ } \ (ret == 0); \ })
So actually it does get initialized, and I guess cppcheck doesn't know about macros?
[drivers/gpu/drm/nouveau/nouveau_usif.c:179]: (error) Uninitialized variable: ret [drivers/gpu/drm/nouveau/nouveau_usif.c:202]: (error) Uninitialized variable: ret [drivers/gpu/drm/nouveau/nouveau_usif.c:241]: (error) Uninitialized variable: ret [drivers/gpu/drm/nouveau/nouveau_usif.c:157]: (error) Uninitialized variable: ret [drivers/gpu/drm/nouveau/nouveau_usif.c:288]: (error) Uninitialized variable: ret
Signed-off-by: Colin Ian King colin.king@canonical.com
drivers/gpu/drm/nouveau/nouveau_usif.c | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-)
diff --git a/drivers/gpu/drm/nouveau/nouveau_usif.c b/drivers/gpu/drm/nouveau/nouveau_usif.c index cb1182d..01b50a2 100644 --- a/drivers/gpu/drm/nouveau/nouveau_usif.c +++ b/drivers/gpu/drm/nouveau/nouveau_usif.c @@ -129,7 +129,7 @@ usif_notify_new(struct drm_file *f, void *data, u32 size, void *argv, u32 argc) struct nvif_notify_req_v0 v0; } *req; struct usif_notify *ntfy;
int ret;
int ret = -EINVAL; if (nvif_unpack(args->v0, 0, 0, true)) { if (usif_notify_find(f, args->v0.index))@@ -170,7 +170,7 @@ usif_notify_del(struct drm_file *f, void *data, u32 size, void *argv, u32 argc) struct nvif_ioctl_ntfy_del_v0 v0; } *args = data; struct usif_notify *ntfy;
int ret;
int ret = -EINVAL; if (nvif_unpack(args->v0, 0, 0, true)) { if (!(ntfy = usif_notify_find(f, args->v0.index)))@@ -193,7 +193,7 @@ usif_notify_get(struct drm_file *f, void *data, u32 size, void *argv, u32 argc) struct nvif_ioctl_ntfy_del_v0 v0; } *args = data; struct usif_notify *ntfy;
int ret;
int ret = -EINVAL; if (nvif_unpack(args->v0, 0, 0, true)) { if (!(ntfy = usif_notify_find(f, args->v0.index)))@@ -232,7 +232,7 @@ usif_notify_put(struct drm_file *f, void *data, u32 size, void *argv, u32 argc) struct nvif_ioctl_ntfy_put_v0 v0; } *args = data; struct usif_notify *ntfy;
int ret;
int ret = -EINVAL; if (nvif_unpack(args->v0, 0, 0, true)) { if (!(ntfy = usif_notify_find(f, args->v0.index)))@@ -269,7 +269,7 @@ usif_object_new(struct drm_file *f, void *data, u32 size, void *argv, u32 argc) struct nvif_ioctl_new_v0 v0; } *args = data; struct usif_object *object;
int ret;
int ret = -EINVAL; if (!(object = kmalloc(sizeof(*object), GFP_KERNEL))) return -ENOMEM;-- 2.1.4
-- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
On 1 July 2015 at 17:56, Ilia Mirkin imirkin@alum.mit.edu wrote:
On Wed, Jul 1, 2015 at 12:51 PM, Colin King colin.king@canonical.com wrote:
From: Colin Ian King colin.king@canonical.com
Various usif_ioctl helper functions do not initialize the return variable ret and some of the error handling return paths just return garbage values that were on the stack (or in a register). I believe that in all the cases, the initial ret variable should be set to -EINVAL and subsequent paths through these helper functions set it appropriately otherwise.
Found via static analysis using cppcheck:
[drivers/gpu/drm/nouveau/nouveau_usif.c:138]: (error) Uninitialized variable: ret
It sure would seem that way, wouldn't it?
#define nvif_unpack(d,vl,vh,m) ({ \ if ((vl) == 0 || ret == -ENOSYS) { \ int _size = sizeof(d); \ if (_size <= size && (d).version >= (vl) && \ (d).version <= (vh)) { \ data = (u8 *)data + _size; \ size = size - _size; \ ret = ((m) || !size) ? 0 : -E2BIG; \ } else { \ ret = -ENOSYS; \ } \ } \ (ret == 0); \ })
So actually it does get initialized, and I guess cppcheck doesn't know about macros?
I think I'm having deja-vu, but I do recall a similar mention to Ben. Although in my defence I've assumed that nvif_unpack was a function, as macros normally are normally all caps. Seems like the patch that capitalises nvif_unpack never made it upstream :'-(
Cheers, Emil
On 01/07/15 18:12, Emil Velikov wrote:
On 1 July 2015 at 17:56, Ilia Mirkin imirkin@alum.mit.edu wrote:
On Wed, Jul 1, 2015 at 12:51 PM, Colin King colin.king@canonical.com wrote:
From: Colin Ian King colin.king@canonical.com
Various usif_ioctl helper functions do not initialize the return variable ret and some of the error handling return paths just return garbage values that were on the stack (or in a register). I believe that in all the cases, the initial ret variable should be set to -EINVAL and subsequent paths through these helper functions set it appropriately otherwise.
Found via static analysis using cppcheck:
[drivers/gpu/drm/nouveau/nouveau_usif.c:138]: (error) Uninitialized variable: ret
It sure would seem that way, wouldn't it?
#define nvif_unpack(d,vl,vh,m) ({ \ if ((vl) == 0 || ret == -ENOSYS) { \ int _size = sizeof(d); \ if (_size <= size && (d).version >= (vl) && \ (d).version <= (vh)) { \ data = (u8 *)data + _size; \ size = size - _size; \ ret = ((m) || !size) ? 0 : -E2BIG; \ } else { \ ret = -ENOSYS; \ } \ } \ (ret == 0); \ })
So actually it does get initialized, and I guess cppcheck doesn't know about macros?
Hrm, what about the case when ((vl) == 0 || ret == -ENOSYS) is false, where is ret being set in that case?
I think I'm having deja-vu, but I do recall a similar mention to Ben. Although in my defence I've assumed that nvif_unpack was a function, as macros normally are normally all caps. Seems like the patch that capitalises nvif_unpack never made it upstream :'-(
Cheers, Emil
On Wed, Jul 1, 2015 at 1:18 PM, Colin Ian King colin.king@canonical.com wrote:
On 01/07/15 18:12, Emil Velikov wrote:
On 1 July 2015 at 17:56, Ilia Mirkin imirkin@alum.mit.edu wrote:
On Wed, Jul 1, 2015 at 12:51 PM, Colin King colin.king@canonical.com wrote:
From: Colin Ian King colin.king@canonical.com
Various usif_ioctl helper functions do not initialize the return variable ret and some of the error handling return paths just return garbage values that were on the stack (or in a register). I believe that in all the cases, the initial ret variable should be set to -EINVAL and subsequent paths through these helper functions set it appropriately otherwise.
Found via static analysis using cppcheck:
[drivers/gpu/drm/nouveau/nouveau_usif.c:138]: (error) Uninitialized variable: ret
It sure would seem that way, wouldn't it?
#define nvif_unpack(d,vl,vh,m) ({ \ if ((vl) == 0 || ret == -ENOSYS) { \ int _size = sizeof(d); \ if (_size <= size && (d).version >= (vl) && \ (d).version <= (vh)) { \ data = (u8 *)data + _size; \ size = size - _size; \ ret = ((m) || !size) ? 0 : -E2BIG; \ } else { \ ret = -ENOSYS; \ } \ } \ (ret == 0); \ })
So actually it does get initialized, and I guess cppcheck doesn't know about macros?
Hrm, what about the case when ((vl) == 0 || ret == -ENOSYS) is false, where is ret being set in that case?
Is that actually the case for any of the callsites? gcc would complain about that...
On 1 July 2015 at 18:37, Ilia Mirkin imirkin@alum.mit.edu wrote:
On Wed, Jul 1, 2015 at 1:18 PM, Colin Ian King colin.king@canonical.com wrote:
On 01/07/15 18:12, Emil Velikov wrote:
On 1 July 2015 at 17:56, Ilia Mirkin imirkin@alum.mit.edu wrote:
On Wed, Jul 1, 2015 at 12:51 PM, Colin King colin.king@canonical.com wrote:
From: Colin Ian King colin.king@canonical.com
Various usif_ioctl helper functions do not initialize the return variable ret and some of the error handling return paths just return garbage values that were on the stack (or in a register). I believe that in all the cases, the initial ret variable should be set to -EINVAL and subsequent paths through these helper functions set it appropriately otherwise.
Found via static analysis using cppcheck:
[drivers/gpu/drm/nouveau/nouveau_usif.c:138]: (error) Uninitialized variable: ret
It sure would seem that way, wouldn't it?
#define nvif_unpack(d,vl,vh,m) ({ \ if ((vl) == 0 || ret == -ENOSYS) { \ int _size = sizeof(d); \ if (_size <= size && (d).version >= (vl) && \ (d).version <= (vh)) { \ data = (u8 *)data + _size; \ size = size - _size; \ ret = ((m) || !size) ? 0 : -E2BIG; \ } else { \ ret = -ENOSYS; \ } \ } \ (ret == 0); \ })
So actually it does get initialized, and I guess cppcheck doesn't know about macros?
Hrm, what about the case when ((vl) == 0 || ret == -ENOSYS) is false, where is ret being set in that case?
Is that actually the case for any of the callsites? gcc would complain about that...
There is one:
drm/nouveau/nvkm/engine/disp/nv50.c: if (nvif_unpack(args->v1, 1, 1, true))
Seems like a recent addition though, I don't recall it with back when was introduced.
-Emil
On Wed, Jul 1, 2015 at 1:59 PM, Emil Velikov emil.l.velikov@gmail.com wrote:
On 1 July 2015 at 18:37, Ilia Mirkin imirkin@alum.mit.edu wrote:
On Wed, Jul 1, 2015 at 1:18 PM, Colin Ian King colin.king@canonical.com wrote:
On 01/07/15 18:12, Emil Velikov wrote:
On 1 July 2015 at 17:56, Ilia Mirkin imirkin@alum.mit.edu wrote:
On Wed, Jul 1, 2015 at 12:51 PM, Colin King colin.king@canonical.com wrote:
From: Colin Ian King colin.king@canonical.com
Various usif_ioctl helper functions do not initialize the return variable ret and some of the error handling return paths just return garbage values that were on the stack (or in a register). I believe that in all the cases, the initial ret variable should be set to -EINVAL and subsequent paths through these helper functions set it appropriately otherwise.
Found via static analysis using cppcheck:
[drivers/gpu/drm/nouveau/nouveau_usif.c:138]: (error) Uninitialized variable: ret
It sure would seem that way, wouldn't it?
#define nvif_unpack(d,vl,vh,m) ({ \ if ((vl) == 0 || ret == -ENOSYS) { \ int _size = sizeof(d); \ if (_size <= size && (d).version >= (vl) && \ (d).version <= (vh)) { \ data = (u8 *)data + _size; \ size = size - _size; \ ret = ((m) || !size) ? 0 : -E2BIG; \ } else { \ ret = -ENOSYS; \ } \ } \ (ret == 0); \ })
So actually it does get initialized, and I guess cppcheck doesn't know about macros?
Hrm, what about the case when ((vl) == 0 || ret == -ENOSYS) is false, where is ret being set in that case?
Is that actually the case for any of the callsites? gcc would complain about that...
There is one:
drm/nouveau/nvkm/engine/disp/nv50.c: if (nvif_unpack(args->v1, 1, 1, true))
Seems like a recent addition though, I don't recall it with back when was introduced.
It follows a call to nvif_unpack(0) though, which will initialize ret.
dri-devel@lists.freedesktop.org
-
Colin Ian King -
Colin King
-
Emil Velikov -
Ilia Mirkin