Re: [PATCH v6 08/12] drm/xe/bo: Block mmap of DONTNEED/purged BOs

["Yadav, Arvind" <arvind.yadav@intel.com>]

On 10-03-2026 15:47, Thomas Hellström wrote:
> On Tue, 2026-03-03 at 20:50 +0530, Arvind Yadav wrote:
>> Don't allow new CPU mmaps to BOs marked DONTNEED or PURGED.
>> DONTNEED BOs can have their contents discarded at any time, making
>> CPU access undefined behavior. PURGED BOs have no backing store and
>> are permanently invalid.
>>
>> Return -EBUSY for DONTNEED BOs (temporary purgeable state) and
>> -EINVAL for purged BOs (permanent, no backing store).
>>
>> The mmap offset ioctl now checks the BO's purgeable state before
>> allowing userspace to establish a new CPU mapping. This prevents
>> the race where userspace gets a valid offset but the BO is purged
>> before actual faulting begins.
>>
>> Existing mmaps (established before DONTNEED) may still work until
>> pages are purged, at which point CPU faults fail with SIGBUS.
>>
>> v6:
>> - Split DONTNEED → -EBUSY and PURGED → -EINVAL for consistency
>>    with the rest of the series (Thomas, Matt)
>>
>> Cc: Matthew Brost <matthew.brost@intel.com>
>> Cc: Thomas Hellström <thomas.hellstrom@linux.intel.com>
>> Cc: Himal Prasad Ghimiray <himal.prasad.ghimiray@intel.com>
>> Signed-off-by: Arvind Yadav <arvind.yadav@intel.com>
>> ---
>>   drivers/gpu/drm/xe/xe_bo.c | 28 +++++++++++++++++++++++++++-
>>   1 file changed, 27 insertions(+), 1 deletion(-)
>>
>> diff --git a/drivers/gpu/drm/xe/xe_bo.c b/drivers/gpu/drm/xe/xe_bo.c
>> index d05a73756905..3a4965bdadf2 100644
>> --- a/drivers/gpu/drm/xe/xe_bo.c
>> +++ b/drivers/gpu/drm/xe/xe_bo.c
>> @@ -3396,6 +3396,8 @@ int xe_gem_mmap_offset_ioctl(struct drm_device
> This needs to be done in the mmap() callback. It should be OK to wrap
> the existing callback there.


I will move this to mmap() callback.

>
>> *dev, void *data,
>>   	struct xe_device *xe = to_xe_device(dev);
>>   	struct drm_xe_gem_mmap_offset *args = data;
>>   	struct drm_gem_object *gem_obj;
>> +	struct xe_bo *bo;
>> +	int err;
>>   
>>   	if (XE_IOCTL_DBG(xe, args->extensions) ||
>>   	    XE_IOCTL_DBG(xe, args->reserved[0] || args-
>>> reserved[1]))
>> @@ -3425,11 +3427,35 @@ int xe_gem_mmap_offset_ioctl(struct
>> drm_device *dev, void *data,
>>   	if (XE_IOCTL_DBG(xe, !gem_obj))
>>   		return -ENOENT;
>>   
>> +	bo = gem_to_xe_bo(gem_obj);
>> +
>> +	/*
>> +	 * Reject new mmap to purgeable BOs. DONTNEED BOs can be
>> purged
>> +	 * at any time, making CPU access undefined behavior. Purged
>> BOs have
>> +	 * no backing store and are permanently invalid.
>> +	 */
>> +	xe_bo_lock(bo, false);
> sleeping locks need to be interruptible whenever possible.


Noted,

Thanks,
Arvind

>
> Thanks,
> Thomas
>
>
>
>> +	if (xe_bo_madv_is_dontneed(bo)) {
>> +		err = -EBUSY;
>> +		goto out_unlock;
>> +	}
>> +
>> +	if (xe_bo_is_purged(bo)) {
>> +		err = -EINVAL;
>> +		goto out_unlock;
>> +	}
>> +	xe_bo_unlock(bo);
>> +
>>   	/* The mmap offset was set up at BO allocation time. */
>>   	args->offset = drm_vma_node_offset_addr(&gem_obj->vma_node);
>> +	xe_bo_put(bo);
>>   
>> -	xe_bo_put(gem_to_xe_bo(gem_obj));
>>   	return 0;
>> +
>> +out_unlock:
>> +	xe_bo_unlock(bo);
>> +	xe_bo_put(bo);
>> +	return err;
>>   }
>>   
>>   /**