2020-01-23 15:26:04

by Colin King

Subject: [PATCH][next] drm/i915/gem: fix null pointer dereference on vm

From: Colin Ian King <[email protected]>

Currently if the call to function context_get_vm_rcu returns
a null pointer for vm then the error exit path via label err_put
will call i915_vm_put on the null vm, causing a null pointer
dereference. Fix this by adding a null check on vm and returning
without calling the i915_vm_put.

Fixes: 5dbd2b7be61e ("drm/i915/gem: Convert vm idr to xarray")
Signed-off-by: Colin Ian King <[email protected]>
---
drivers/gpu/drm/i915/gem/i915_gem_context.c | 9 ++++++---
1 file changed, 6 insertions(+), 3 deletions(-)

diff --git a/drivers/gpu/drm/i915/gem/i915_gem_context.c b/drivers/gpu/drm/i915/gem/i915_gem_context.c
index 5d4157e1ccf7..3e6e34ec9fa8 100644
--- a/drivers/gpu/drm/i915/gem/i915_gem_context.c
+++ b/drivers/gpu/drm/i915/gem/i915_gem_context.c
@@ -1005,9 +1005,12 @@ static int get_ppgtt(struct drm_i915_file_private *file_priv,
err = -ENODEV;
rcu_read_lock();
vm = context_get_vm_rcu(ctx);
- if (vm)
- err = xa_alloc(&file_priv->vm_xa, &id, vm,
- xa_limit_32b, GFP_KERNEL);
+ if (!vm) {
+ rcu_read_unlock();
+ return err;
+ }
+ err = xa_alloc(&file_priv->vm_xa, &id, vm,
+ xa_limit_32b, GFP_KERNEL);
rcu_read_unlock();
if (err)
goto err_put;
--
2.24.0

2020-01-23 15:26:23

by Chris Wilson

Subject: Re: [PATCH][next] drm/i915/gem: fix null pointer dereference on vm

Quoting Colin King (2020-01-23 15:14:06)
> From: Colin Ian King <[email protected]>
>
> Currently if the call to function context_get_vm_rcu returns
> a null pointer for vm then the error exit path via label err_put
> will call i915_vm_put on the null vm, causing a null pointer
> dereference. Fix this by adding a null check on vm and returning
> without calling the i915_vm_put.
>
> Fixes: 5dbd2b7be61e ("drm/i915/gem: Convert vm idr to xarray")
> Signed-off-by: Colin Ian King <[email protected]>

Hmm. Actually, we can drop the rcu_read_lock as soon as we've acquire
the local ref to ctx->vm. So something like,

if (!rcu_access_pointer(ctx->vm))
return -ENODEV;

- err = -ENODEV;
rcu_read_lock();
vm = context_get_vm_rcu(ctx);
- if (vm)
- err = xa_alloc(&file_priv->vm_xa, &id, vm,
- xa_limit_32b, GFP_KERNEL);
rcu_read_unlock();
+ if (!vm)
+ return -ENODEV;
+
+ err = xa_alloc(&file_priv->vm_xa, &id, vm,
+ xa_limit_32b, GFP_KERNEL);
if (err)
goto err_put;

would work.
-Chris