2021-11-30 15:14:01

by Zhou Qingyang

[permalink] [raw]
Subject: [PATCH] drm/radeon/radeon_kms: Fix a NULL pointer dereference in radeon_driver_open_kms()

In radeon_driver_open_kms(), radeon_vm_bo_add() is assigned to
vm->ib_bo_va and passes and used in radeon_vm_bo_set_addr(). In
radeon_vm_bo_set_addr(), there is a dereference of vm->ib_bo_va,
which could lead to a NULL pointer dereference on failure of
radeon_vm_bo_add().

Fix this bug by adding a check of vm->ib_bo_va.

This bug was found by a static analyzer. The analysis employs
differential checking to identify inconsistent security operations
(e.g., checks or kfrees) between two code paths and confirms that the
inconsistent operations are not recovered in the current function or
the callers, so they constitute bugs.

Note that, as a bug found by static analysis, it can be a false
positive or hard to trigger. Multiple researchers have cross-reviewed
the bug.

Builds with CONFIG_DRM_RADEON=m show no new warnings,
and our static analyzer no longer warns about this code.

Fixes: cc9e67e3d700 ("drm/radeon: fix VM IB handling")
Signed-off-by: Zhou Qingyang <[email protected]>
---
drivers/gpu/drm/radeon/radeon_kms.c | 7 +++++++
1 file changed, 7 insertions(+)

diff --git a/drivers/gpu/drm/radeon/radeon_kms.c b/drivers/gpu/drm/radeon/radeon_kms.c
index 482fb0ae6cb5..ead015c055fb 100644
--- a/drivers/gpu/drm/radeon/radeon_kms.c
+++ b/drivers/gpu/drm/radeon/radeon_kms.c
@@ -688,6 +688,13 @@ int radeon_driver_open_kms(struct drm_device *dev, struct drm_file *file_priv)
* virtual address space */
vm->ib_bo_va = radeon_vm_bo_add(rdev, vm,
rdev->ring_tmp_bo.bo);
+ if (!vm->ib_bo_va) {
+ r = -ENOMEM;
+ radeon_vm_fini(rdev, vm);
+ kfree(fpriv);
+ goto out_suspend;
+ }
+
r = radeon_vm_bo_set_addr(rdev, vm->ib_bo_va,
RADEON_VA_IB_OFFSET,
RADEON_VM_PAGE_READABLE |
--
2.25.1



2021-11-30 15:15:29

by Christian König

[permalink] [raw]
Subject: Re: [PATCH] drm/radeon/radeon_kms: Fix a NULL pointer dereference in radeon_driver_open_kms()

Am 30.11.21 um 16:04 schrieb Zhou Qingyang:
> In radeon_driver_open_kms(), radeon_vm_bo_add() is assigned to
> vm->ib_bo_va and passes and used in radeon_vm_bo_set_addr(). In
> radeon_vm_bo_set_addr(), there is a dereference of vm->ib_bo_va,
> which could lead to a NULL pointer dereference on failure of
> radeon_vm_bo_add().
>
> Fix this bug by adding a check of vm->ib_bo_va.
>
> This bug was found by a static analyzer. The analysis employs
> differential checking to identify inconsistent security operations
> (e.g., checks or kfrees) between two code paths and confirms that the
> inconsistent operations are not recovered in the current function or
> the callers, so they constitute bugs.
>
> Note that, as a bug found by static analysis, it can be a false
> positive or hard to trigger. Multiple researchers have cross-reviewed
> the bug.
>
> Builds with CONFIG_DRM_RADEON=m show no new warnings,
> and our static analyzer no longer warns about this code.
>
> Fixes: cc9e67e3d700 ("drm/radeon: fix VM IB handling")
> Signed-off-by: Zhou Qingyang <[email protected]>
> ---
> drivers/gpu/drm/radeon/radeon_kms.c | 7 +++++++
> 1 file changed, 7 insertions(+)
>
> diff --git a/drivers/gpu/drm/radeon/radeon_kms.c b/drivers/gpu/drm/radeon/radeon_kms.c
> index 482fb0ae6cb5..ead015c055fb 100644
> --- a/drivers/gpu/drm/radeon/radeon_kms.c
> +++ b/drivers/gpu/drm/radeon/radeon_kms.c
> @@ -688,6 +688,13 @@ int radeon_driver_open_kms(struct drm_device *dev, struct drm_file *file_priv)
> * virtual address space */
> vm->ib_bo_va = radeon_vm_bo_add(rdev, vm,
> rdev->ring_tmp_bo.bo);
> + if (!vm->ib_bo_va) {
> + r = -ENOMEM;
> + radeon_vm_fini(rdev, vm);
> + kfree(fpriv);
> + goto out_suspend;
> + }
> +

Impressive catch for an automated checker.

Please improve the error handling into goto style since we now add the
fourth instance of the same error handling code. Apart from that looks
good to me.

Thanks,
Christian.

> r = radeon_vm_bo_set_addr(rdev, vm->ib_bo_va,
> RADEON_VA_IB_OFFSET,
> RADEON_VM_PAGE_READABLE |


2021-11-30 15:35:20

by Zhou Qingyang

[permalink] [raw]
Subject: [PATCH v2] drm/radeon/radeon_kms: Fix a NULL pointer dereference in radeon_driver_open_kms()

In radeon_driver_open_kms(), radeon_vm_bo_add() is assigned to
vm->ib_bo_va and passes and used in radeon_vm_bo_set_addr(). In
radeon_vm_bo_set_addr(), there is a dereference of vm->ib_bo_va,
which could lead to a NULL pointer dereference on failure of
radeon_vm_bo_add().

Fix this bug by adding a check of vm->ib_bo_va.

This bug was found by a static analyzer. The analysis employs
differential checking to identify inconsistent security operations
(e.g., checks or kfrees) between two code paths and confirms that the
inconsistent operations are not recovered in the current function or
the callers, so they constitute bugs.

Note that, as a bug found by static analysis, it can be a false
positive or hard to trigger. Multiple researchers have cross-reviewed
the bug.

Builds with CONFIG_DRM_RADEON=m show no new warnings,
and our static analyzer no longer warns about this code.

Fixes: cc9e67e3d700 ("drm/radeon: fix VM IB handling")
Signed-off-by: Zhou Qingyang <[email protected]>
---
Changes in v2:
- Improve the error handling into goto style

drivers/gpu/drm/radeon/radeon_kms.c | 24 ++++++++++++++----------
1 file changed, 14 insertions(+), 10 deletions(-)

diff --git a/drivers/gpu/drm/radeon/radeon_kms.c b/drivers/gpu/drm/radeon/radeon_kms.c
index 482fb0ae6cb5..e49a9d160e52 100644
--- a/drivers/gpu/drm/radeon/radeon_kms.c
+++ b/drivers/gpu/drm/radeon/radeon_kms.c
@@ -649,6 +649,8 @@ int radeon_driver_open_kms(struct drm_device *dev, struct drm_file *file_priv)
{
struct radeon_device *rdev = dev->dev_private;
int r;
+ struct radeon_fpriv *fpriv;
+ struct radeon_vm *vm;

file_priv->driver_priv = NULL;

@@ -660,8 +662,6 @@ int radeon_driver_open_kms(struct drm_device *dev, struct drm_file *file_priv)

/* new gpu have virtual address space support */
if (rdev->family >= CHIP_CAYMAN) {
- struct radeon_fpriv *fpriv;
- struct radeon_vm *vm;

fpriv = kzalloc(sizeof(*fpriv), GFP_KERNEL);
if (unlikely(!fpriv)) {
@@ -673,34 +673,38 @@ int radeon_driver_open_kms(struct drm_device *dev, struct drm_file *file_priv)
vm = &fpriv->vm;
r = radeon_vm_init(rdev, vm);
if (r) {
- kfree(fpriv);
- goto out_suspend;
+ goto out_fpriv;
}

r = radeon_bo_reserve(rdev->ring_tmp_bo.bo, false);
if (r) {
- radeon_vm_fini(rdev, vm);
- kfree(fpriv);
- goto out_suspend;
+ goto out_vm_fini;
}

/* map the ib pool buffer read only into
* virtual address space */
vm->ib_bo_va = radeon_vm_bo_add(rdev, vm,
rdev->ring_tmp_bo.bo);
+ if (!vm->ib_bo_va) {
+ r = -ENOMEM;
+ goto out_vm_fini;
+ }
+
r = radeon_vm_bo_set_addr(rdev, vm->ib_bo_va,
RADEON_VA_IB_OFFSET,
RADEON_VM_PAGE_READABLE |
RADEON_VM_PAGE_SNOOPED);
if (r) {
- radeon_vm_fini(rdev, vm);
- kfree(fpriv);
- goto out_suspend;
+ goto out_vm_fini;
}
}
file_priv->driver_priv = fpriv;
}

+out_vm_fini:
+ radeon_vm_fini(rdev, vm);
+out_fpriv:
+ kfree(fpriv);
out_suspend:
pm_runtime_mark_last_busy(dev->dev);
pm_runtime_put_autosuspend(dev->dev);
--
2.25.1


2021-11-30 15:37:52

by Christian König

[permalink] [raw]
Subject: Re: [PATCH v2] drm/radeon/radeon_kms: Fix a NULL pointer dereference in radeon_driver_open_kms()



Am 30.11.21 um 16:33 schrieb Zhou Qingyang:
> In radeon_driver_open_kms(), radeon_vm_bo_add() is assigned to
> vm->ib_bo_va and passes and used in radeon_vm_bo_set_addr(). In
> radeon_vm_bo_set_addr(), there is a dereference of vm->ib_bo_va,
> which could lead to a NULL pointer dereference on failure of
> radeon_vm_bo_add().
>
> Fix this bug by adding a check of vm->ib_bo_va.
>
> This bug was found by a static analyzer. The analysis employs
> differential checking to identify inconsistent security operations
> (e.g., checks or kfrees) between two code paths and confirms that the
> inconsistent operations are not recovered in the current function or
> the callers, so they constitute bugs.
>
> Note that, as a bug found by static analysis, it can be a false
> positive or hard to trigger. Multiple researchers have cross-reviewed
> the bug.
>
> Builds with CONFIG_DRM_RADEON=m show no new warnings,
> and our static analyzer no longer warns about this code.
>
> Fixes: cc9e67e3d700 ("drm/radeon: fix VM IB handling")
> Signed-off-by: Zhou Qingyang <[email protected]>
> ---
> Changes in v2:
> - Improve the error handling into goto style
>
> drivers/gpu/drm/radeon/radeon_kms.c | 24 ++++++++++++++----------
> 1 file changed, 14 insertions(+), 10 deletions(-)
>
> diff --git a/drivers/gpu/drm/radeon/radeon_kms.c b/drivers/gpu/drm/radeon/radeon_kms.c
> index 482fb0ae6cb5..e49a9d160e52 100644
> --- a/drivers/gpu/drm/radeon/radeon_kms.c
> +++ b/drivers/gpu/drm/radeon/radeon_kms.c
> @@ -649,6 +649,8 @@ int radeon_driver_open_kms(struct drm_device *dev, struct drm_file *file_priv)
> {
> struct radeon_device *rdev = dev->dev_private;
> int r;
> + struct radeon_fpriv *fpriv;
> + struct radeon_vm *vm;

Please keep variables like "i" or "r" declared last.

>
> file_priv->driver_priv = NULL;
>
> @@ -660,8 +662,6 @@ int radeon_driver_open_kms(struct drm_device *dev, struct drm_file *file_priv)
>
> /* new gpu have virtual address space support */
> if (rdev->family >= CHIP_CAYMAN) {
> - struct radeon_fpriv *fpriv;
> - struct radeon_vm *vm;
>
> fpriv = kzalloc(sizeof(*fpriv), GFP_KERNEL);
> if (unlikely(!fpriv)) {
> @@ -673,34 +673,38 @@ int radeon_driver_open_kms(struct drm_device *dev, struct drm_file *file_priv)
> vm = &fpriv->vm;
> r = radeon_vm_init(rdev, vm);
> if (r) {
> - kfree(fpriv);
> - goto out_suspend;
> + goto out_fpriv;
> }
>
> r = radeon_bo_reserve(rdev->ring_tmp_bo.bo, false);
> if (r) {
> - radeon_vm_fini(rdev, vm);
> - kfree(fpriv);
> - goto out_suspend;
> + goto out_vm_fini;
> }
>
> /* map the ib pool buffer read only into
> * virtual address space */
> vm->ib_bo_va = radeon_vm_bo_add(rdev, vm,
> rdev->ring_tmp_bo.bo);
> + if (!vm->ib_bo_va) {
> + r = -ENOMEM;
> + goto out_vm_fini;
> + }
> +
> r = radeon_vm_bo_set_addr(rdev, vm->ib_bo_va,
> RADEON_VA_IB_OFFSET,
> RADEON_VM_PAGE_READABLE |
> RADEON_VM_PAGE_SNOOPED);
> if (r) {
> - radeon_vm_fini(rdev, vm);
> - kfree(fpriv);
> - goto out_suspend;
> + goto out_vm_fini;
> }
> }
> file_priv->driver_priv = fpriv;
> }
>

That here won't work.

> +out_vm_fini:
> + radeon_vm_fini(rdev, vm);
> +out_fpriv:
> + kfree(fpriv);

You are finishing the VM and freeing up the memory in the good case now
as well.

Christian.

> out_suspend:
> pm_runtime_mark_last_busy(dev->dev);
> pm_runtime_put_autosuspend(dev->dev);


2021-11-30 15:58:05

by Zhou Qingyang

[permalink] [raw]
Subject: [PATCH v3] drm/radeon/radeon_kms: Fix a NULL pointer dereference in radeon_driver_open_kms()

In radeon_driver_open_kms(), radeon_vm_bo_add() is assigned to
vm->ib_bo_va and passes and used in radeon_vm_bo_set_addr(). In
radeon_vm_bo_set_addr(), there is a dereference of vm->ib_bo_va,
which could lead to a NULL pointer dereference on failure of
radeon_vm_bo_add().

Fix this bug by adding a check of vm->ib_bo_va.

This bug was found by a static analyzer. The analysis employs
differential checking to identify inconsistent security operations
(e.g., checks or kfrees) between two code paths and confirms that the
inconsistent operations are not recovered in the current function or
the callers, so they constitute bugs.

Note that, as a bug found by static analysis, it can be a false
positive or hard to trigger. Multiple researchers have cross-reviewed
the bug.

Builds with CONFIG_DRM_RADEON=m show no new warnings,
and our static analyzer no longer warns about this code.

Fixes: cc9e67e3d700 ("drm/radeon: fix VM IB handling")
Signed-off-by: Zhou Qingyang <[email protected]>
---
Changes in v3:
- Fix the bug that good case will also be freed
- Improve code style

Changes in v2:
- Improve the error handling into goto style

drivers/gpu/drm/radeon/radeon_kms.c | 35 ++++++++++++++++-------------
1 file changed, 19 insertions(+), 16 deletions(-)

diff --git a/drivers/gpu/drm/radeon/radeon_kms.c b/drivers/gpu/drm/radeon/radeon_kms.c
index 482fb0ae6cb5..439f4d1fdd65 100644
--- a/drivers/gpu/drm/radeon/radeon_kms.c
+++ b/drivers/gpu/drm/radeon/radeon_kms.c
@@ -648,6 +648,8 @@ void radeon_driver_lastclose_kms(struct drm_device *dev)
int radeon_driver_open_kms(struct drm_device *dev, struct drm_file *file_priv)
{
struct radeon_device *rdev = dev->dev_private;
+ struct radeon_fpriv *fpriv;
+ struct radeon_vm *vm;
int r;

file_priv->driver_priv = NULL;
@@ -660,8 +662,6 @@ int radeon_driver_open_kms(struct drm_device *dev, struct drm_file *file_priv)

/* new gpu have virtual address space support */
if (rdev->family >= CHIP_CAYMAN) {
- struct radeon_fpriv *fpriv;
- struct radeon_vm *vm;

fpriv = kzalloc(sizeof(*fpriv), GFP_KERNEL);
if (unlikely(!fpriv)) {
@@ -672,35 +672,38 @@ int radeon_driver_open_kms(struct drm_device *dev, struct drm_file *file_priv)
if (rdev->accel_working) {
vm = &fpriv->vm;
r = radeon_vm_init(rdev, vm);
- if (r) {
- kfree(fpriv);
- goto out_suspend;
- }
+ if (r)
+ goto out_fpriv;

r = radeon_bo_reserve(rdev->ring_tmp_bo.bo, false);
- if (r) {
- radeon_vm_fini(rdev, vm);
- kfree(fpriv);
- goto out_suspend;
- }
+ if (r)
+ goto out_vm_fini;

/* map the ib pool buffer read only into
* virtual address space */
vm->ib_bo_va = radeon_vm_bo_add(rdev, vm,
rdev->ring_tmp_bo.bo);
+ if (!vm->ib_bo_va) {
+ r = -ENOMEM;
+ goto out_vm_fini;
+ }
+
r = radeon_vm_bo_set_addr(rdev, vm->ib_bo_va,
RADEON_VA_IB_OFFSET,
RADEON_VM_PAGE_READABLE |
RADEON_VM_PAGE_SNOOPED);
- if (r) {
- radeon_vm_fini(rdev, vm);
- kfree(fpriv);
- goto out_suspend;
- }
+ if (r)
+ goto out_vm_fini;
}
file_priv->driver_priv = fpriv;
}

+out_vm_fini:
+ if (r)
+ radeon_vm_fini(rdev, vm);
+out_fpriv:
+ if (r)
+ kfree(fpriv);
out_suspend:
pm_runtime_mark_last_busy(dev->dev);
pm_runtime_put_autosuspend(dev->dev);
--
2.25.1


2021-12-01 03:22:25

by Zhou Qingyang

[permalink] [raw]
Subject: [PATCH v4] drm/radeon/radeon_kms: Fix a NULL pointer dereference in radeon_driver_open_kms()

In radeon_driver_open_kms(), radeon_vm_bo_add() is assigned to
vm->ib_bo_va and passes and used in radeon_vm_bo_set_addr(). In
radeon_vm_bo_set_addr(), there is a dereference of vm->ib_bo_va,
which could lead to a NULL pointer dereference on failure of
radeon_vm_bo_add().

Fix this bug by adding a check of vm->ib_bo_va.

This bug was found by a static analyzer. The analysis employs
differential checking to identify inconsistent security operations
(e.g., checks or kfrees) between two code paths and confirms that the
inconsistent operations are not recovered in the current function or
the callers, so they constitute bugs.

Note that, as a bug found by static analysis, it can be a false
positive or hard to trigger. Multiple researchers have cross-reviewed
the bug.

Builds with CONFIG_DRM_RADEON=m show no new warnings,
and our static analyzer no longer warns about this code.

Fixes: cc9e67e3d700 ("drm/radeon: fix VM IB handling")
Reported-by: kernel test robot <[email protected]>
Signed-off-by: Zhou Qingyang <[email protected]>
---
Changes in v2:
- Initialize the variables to silence warning

Changes in v3:
- Fix the bug that good case will also be freed
- Improve code style

Changes in v2:
- Improve the error handling into goto style

drivers/gpu/drm/radeon/radeon_kms.c | 37 ++++++++++++++++-------------
1 file changed, 20 insertions(+), 17 deletions(-)

diff --git a/drivers/gpu/drm/radeon/radeon_kms.c b/drivers/gpu/drm/radeon/radeon_kms.c
index 482fb0ae6cb5..9d0f840286a1 100644
--- a/drivers/gpu/drm/radeon/radeon_kms.c
+++ b/drivers/gpu/drm/radeon/radeon_kms.c
@@ -648,7 +648,9 @@ void radeon_driver_lastclose_kms(struct drm_device *dev)
int radeon_driver_open_kms(struct drm_device *dev, struct drm_file *file_priv)
{
struct radeon_device *rdev = dev->dev_private;
- int r;
+ struct radeon_fpriv *fpriv = NULL;
+ struct radeon_vm *vm = NULL;
+ int r = 0;

file_priv->driver_priv = NULL;

@@ -660,8 +662,6 @@ int radeon_driver_open_kms(struct drm_device *dev, struct drm_file *file_priv)

/* new gpu have virtual address space support */
if (rdev->family >= CHIP_CAYMAN) {
- struct radeon_fpriv *fpriv;
- struct radeon_vm *vm;

fpriv = kzalloc(sizeof(*fpriv), GFP_KERNEL);
if (unlikely(!fpriv)) {
@@ -672,35 +672,38 @@ int radeon_driver_open_kms(struct drm_device *dev, struct drm_file *file_priv)
if (rdev->accel_working) {
vm = &fpriv->vm;
r = radeon_vm_init(rdev, vm);
- if (r) {
- kfree(fpriv);
- goto out_suspend;
- }
+ if (r)
+ goto out_fpriv;

r = radeon_bo_reserve(rdev->ring_tmp_bo.bo, false);
- if (r) {
- radeon_vm_fini(rdev, vm);
- kfree(fpriv);
- goto out_suspend;
- }
+ if (r)
+ goto out_vm_fini;

/* map the ib pool buffer read only into
* virtual address space */
vm->ib_bo_va = radeon_vm_bo_add(rdev, vm,
rdev->ring_tmp_bo.bo);
+ if (!vm->ib_bo_va) {
+ r = -ENOMEM;
+ goto out_vm_fini;
+ }
+
r = radeon_vm_bo_set_addr(rdev, vm->ib_bo_va,
RADEON_VA_IB_OFFSET,
RADEON_VM_PAGE_READABLE |
RADEON_VM_PAGE_SNOOPED);
- if (r) {
- radeon_vm_fini(rdev, vm);
- kfree(fpriv);
- goto out_suspend;
- }
+ if (r)
+ goto out_vm_fini;
}
file_priv->driver_priv = fpriv;
}

+out_vm_fini:
+ if (r)
+ radeon_vm_fini(rdev, vm);
+out_fpriv:
+ if (r)
+ kfree(fpriv);
out_suspend:
pm_runtime_mark_last_busy(dev->dev);
pm_runtime_put_autosuspend(dev->dev);
--
2.25.1


2021-12-01 06:57:34

by Christian König

[permalink] [raw]
Subject: Re: [PATCH v3] drm/radeon/radeon_kms: Fix a NULL pointer dereference in radeon_driver_open_kms()

Am 30.11.21 um 16:57 schrieb Zhou Qingyang:
> In radeon_driver_open_kms(), radeon_vm_bo_add() is assigned to
> vm->ib_bo_va and passes and used in radeon_vm_bo_set_addr(). In
> radeon_vm_bo_set_addr(), there is a dereference of vm->ib_bo_va,
> which could lead to a NULL pointer dereference on failure of
> radeon_vm_bo_add().
>
> Fix this bug by adding a check of vm->ib_bo_va.
>
> This bug was found by a static analyzer. The analysis employs
> differential checking to identify inconsistent security operations
> (e.g., checks or kfrees) between two code paths and confirms that the
> inconsistent operations are not recovered in the current function or
> the callers, so they constitute bugs.
>
> Note that, as a bug found by static analysis, it can be a false
> positive or hard to trigger. Multiple researchers have cross-reviewed
> the bug.
>
> Builds with CONFIG_DRM_RADEON=m show no new warnings,
> and our static analyzer no longer warns about this code.
>
> Fixes: cc9e67e3d700 ("drm/radeon: fix VM IB handling")
> Signed-off-by: Zhou Qingyang <[email protected]>

Reviewed-by: Christian König <[email protected]>

> ---
> Changes in v3:
> - Fix the bug that good case will also be freed
> - Improve code style
>
> Changes in v2:
> - Improve the error handling into goto style
>
> drivers/gpu/drm/radeon/radeon_kms.c | 35 ++++++++++++++++-------------
> 1 file changed, 19 insertions(+), 16 deletions(-)
>
> diff --git a/drivers/gpu/drm/radeon/radeon_kms.c b/drivers/gpu/drm/radeon/radeon_kms.c
> index 482fb0ae6cb5..439f4d1fdd65 100644
> --- a/drivers/gpu/drm/radeon/radeon_kms.c
> +++ b/drivers/gpu/drm/radeon/radeon_kms.c
> @@ -648,6 +648,8 @@ void radeon_driver_lastclose_kms(struct drm_device *dev)
> int radeon_driver_open_kms(struct drm_device *dev, struct drm_file *file_priv)
> {
> struct radeon_device *rdev = dev->dev_private;
> + struct radeon_fpriv *fpriv;
> + struct radeon_vm *vm;
> int r;
>
> file_priv->driver_priv = NULL;
> @@ -660,8 +662,6 @@ int radeon_driver_open_kms(struct drm_device *dev, struct drm_file *file_priv)
>
> /* new gpu have virtual address space support */
> if (rdev->family >= CHIP_CAYMAN) {
> - struct radeon_fpriv *fpriv;
> - struct radeon_vm *vm;
>
> fpriv = kzalloc(sizeof(*fpriv), GFP_KERNEL);
> if (unlikely(!fpriv)) {
> @@ -672,35 +672,38 @@ int radeon_driver_open_kms(struct drm_device *dev, struct drm_file *file_priv)
> if (rdev->accel_working) {
> vm = &fpriv->vm;
> r = radeon_vm_init(rdev, vm);
> - if (r) {
> - kfree(fpriv);
> - goto out_suspend;
> - }
> + if (r)
> + goto out_fpriv;
>
> r = radeon_bo_reserve(rdev->ring_tmp_bo.bo, false);
> - if (r) {
> - radeon_vm_fini(rdev, vm);
> - kfree(fpriv);
> - goto out_suspend;
> - }
> + if (r)
> + goto out_vm_fini;
>
> /* map the ib pool buffer read only into
> * virtual address space */
> vm->ib_bo_va = radeon_vm_bo_add(rdev, vm,
> rdev->ring_tmp_bo.bo);
> + if (!vm->ib_bo_va) {
> + r = -ENOMEM;
> + goto out_vm_fini;
> + }
> +
> r = radeon_vm_bo_set_addr(rdev, vm->ib_bo_va,
> RADEON_VA_IB_OFFSET,
> RADEON_VM_PAGE_READABLE |
> RADEON_VM_PAGE_SNOOPED);
> - if (r) {
> - radeon_vm_fini(rdev, vm);
> - kfree(fpriv);
> - goto out_suspend;
> - }
> + if (r)
> + goto out_vm_fini;
> }
> file_priv->driver_priv = fpriv;
> }
>
> +out_vm_fini:
> + if (r)
> + radeon_vm_fini(rdev, vm);
> +out_fpriv:
> + if (r)
> + kfree(fpriv);
> out_suspend:
> pm_runtime_mark_last_busy(dev->dev);
> pm_runtime_put_autosuspend(dev->dev);


2021-12-01 07:20:42

by Christian König

[permalink] [raw]
Subject: Re: [PATCH v4] drm/radeon/radeon_kms: Fix a NULL pointer dereference in radeon_driver_open_kms()

Am 01.12.21 um 04:22 schrieb Zhou Qingyang:
> In radeon_driver_open_kms(), radeon_vm_bo_add() is assigned to
> vm->ib_bo_va and passes and used in radeon_vm_bo_set_addr(). In
> radeon_vm_bo_set_addr(), there is a dereference of vm->ib_bo_va,
> which could lead to a NULL pointer dereference on failure of
> radeon_vm_bo_add().
>
> Fix this bug by adding a check of vm->ib_bo_va.
>
> This bug was found by a static analyzer. The analysis employs
> differential checking to identify inconsistent security operations
> (e.g., checks or kfrees) between two code paths and confirms that the
> inconsistent operations are not recovered in the current function or
> the callers, so they constitute bugs.
>
> Note that, as a bug found by static analysis, it can be a false
> positive or hard to trigger. Multiple researchers have cross-reviewed
> the bug.
>
> Builds with CONFIG_DRM_RADEON=m show no new warnings,
> and our static analyzer no longer warns about this code.
>
> Fixes: cc9e67e3d700 ("drm/radeon: fix VM IB handling")
> Reported-by: kernel test robot <[email protected]>
> Signed-off-by: Zhou Qingyang <[email protected]>
> ---
> Changes in v2:
> - Initialize the variables to silence warning

What warning do you get? Double checking the code that shouldn't be
necessary and is usually rather frowned upon.

Thanks,
Christian.

>
> Changes in v3:
> - Fix the bug that good case will also be freed
> - Improve code style
>
> Changes in v2:
> - Improve the error handling into goto style
>
> drivers/gpu/drm/radeon/radeon_kms.c | 37 ++++++++++++++++-------------
> 1 file changed, 20 insertions(+), 17 deletions(-)
>
> diff --git a/drivers/gpu/drm/radeon/radeon_kms.c b/drivers/gpu/drm/radeon/radeon_kms.c
> index 482fb0ae6cb5..9d0f840286a1 100644
> --- a/drivers/gpu/drm/radeon/radeon_kms.c
> +++ b/drivers/gpu/drm/radeon/radeon_kms.c
> @@ -648,7 +648,9 @@ void radeon_driver_lastclose_kms(struct drm_device *dev)
> int radeon_driver_open_kms(struct drm_device *dev, struct drm_file *file_priv)
> {
> struct radeon_device *rdev = dev->dev_private;
> - int r;
> + struct radeon_fpriv *fpriv = NULL;
> + struct radeon_vm *vm = NULL;
> + int r = 0;
>
> file_priv->driver_priv = NULL;
>
> @@ -660,8 +662,6 @@ int radeon_driver_open_kms(struct drm_device *dev, struct drm_file *file_priv)
>
> /* new gpu have virtual address space support */
> if (rdev->family >= CHIP_CAYMAN) {
> - struct radeon_fpriv *fpriv;
> - struct radeon_vm *vm;
>
> fpriv = kzalloc(sizeof(*fpriv), GFP_KERNEL);
> if (unlikely(!fpriv)) {
> @@ -672,35 +672,38 @@ int radeon_driver_open_kms(struct drm_device *dev, struct drm_file *file_priv)
> if (rdev->accel_working) {
> vm = &fpriv->vm;
> r = radeon_vm_init(rdev, vm);
> - if (r) {
> - kfree(fpriv);
> - goto out_suspend;
> - }
> + if (r)
> + goto out_fpriv;
>
> r = radeon_bo_reserve(rdev->ring_tmp_bo.bo, false);
> - if (r) {
> - radeon_vm_fini(rdev, vm);
> - kfree(fpriv);
> - goto out_suspend;
> - }
> + if (r)
> + goto out_vm_fini;
>
> /* map the ib pool buffer read only into
> * virtual address space */
> vm->ib_bo_va = radeon_vm_bo_add(rdev, vm,
> rdev->ring_tmp_bo.bo);
> + if (!vm->ib_bo_va) {
> + r = -ENOMEM;
> + goto out_vm_fini;
> + }
> +
> r = radeon_vm_bo_set_addr(rdev, vm->ib_bo_va,
> RADEON_VA_IB_OFFSET,
> RADEON_VM_PAGE_READABLE |
> RADEON_VM_PAGE_SNOOPED);
> - if (r) {
> - radeon_vm_fini(rdev, vm);
> - kfree(fpriv);
> - goto out_suspend;
> - }
> + if (r)
> + goto out_vm_fini;
> }
> file_priv->driver_priv = fpriv;
> }
>
> +out_vm_fini:
> + if (r)
> + radeon_vm_fini(rdev, vm);
> +out_fpriv:
> + if (r)
> + kfree(fpriv);
> out_suspend:
> pm_runtime_mark_last_busy(dev->dev);
> pm_runtime_put_autosuspend(dev->dev);