[Intel-xe] [PATCH] drm/xe: Fix lockdep warning from xe_vm_madvise

[Matthew Brost]

On Wed, Jul 12, 2023 at 07:01:05PM -0700, Brian Welty wrote:
> We need to hold vm->lock before the xe_vm_is_closed_or_banned().
> 
> Else we get this splat:
> [  802.555227] ------------[ cut here ]------------
> [  802.555234] WARNING: CPU: 33 PID: 3122 at drivers/gpu/drm/xe/xe_vm.h:60 xe_vm_madvise_ioctl+0x2de/0x440 [xe]
> [  802.555515] CPU: 33 PID: 3122 Comm: xe_exec_fault_m Tainted: G        W          6.4.0-xe+ #16
> ...
> [  802.555709] Call Trace:
> [  802.555714]  <TASK>
> [  802.555720]  ? __warn+0x81/0x170
> [  802.555737]  ? xe_vm_madvise_ioctl+0x2de/0x440 [xe]
> 
> Fixes: 627d78b9d038 ("drm/xe: Ban a VM if rebind worker hits an error")
> Signed-off-by: Brian Welty <brian.welty at intel.com>

Thanks for the fix. LGTM.

Reviewed-by: Matthew Brost <matthew.brost at intel.com>

> ---
>  drivers/gpu/drm/xe/xe_vm_madvise.c | 10 +++++-----
>  1 file changed, 5 insertions(+), 5 deletions(-)
> 
> diff --git a/drivers/gpu/drm/xe/xe_vm_madvise.c b/drivers/gpu/drm/xe/xe_vm_madvise.c
> index d6e3b76bf590..7532a60a24f0 100644
> --- a/drivers/gpu/drm/xe/xe_vm_madvise.c
> +++ b/drivers/gpu/drm/xe/xe_vm_madvise.c
> @@ -267,11 +267,6 @@ int xe_vm_madvise_ioctl(struct drm_device *dev, void *data,
>  	if (XE_IOCTL_ERR(xe, !vm))
>  		return -EINVAL;
>  
> -	if (XE_IOCTL_ERR(xe, xe_vm_is_closed_or_banned(vm))) {
> -		err = -ENOENT;
> -		goto put_vm;
> -	}
> -
>  	if (XE_IOCTL_ERR(xe, !xe_vm_in_fault_mode(vm))) {
>  		err = -EINVAL;
>  		goto put_vm;
> @@ -279,6 +274,11 @@ int xe_vm_madvise_ioctl(struct drm_device *dev, void *data,
>  
>  	down_read(&vm->lock);
>  
> +	if (XE_IOCTL_ERR(xe, xe_vm_is_closed_or_banned(vm))) {
> +		err = -ENOENT;
> +		goto unlock_vm;
> +	}
> +
>  	vmas = get_vmas(vm, &num_vmas, args->addr, args->range);
>  	if (XE_IOCTL_ERR(xe, err))
>  		goto unlock_vm;
> -- 
> 2.38.0
>