Re: [PATCH 1/1] iommu/vtd: Fix NULL pointer dereference in prq_event_thread()

[Lu Baolu <baolu.lu-VuQAYsv1563Yd54FQh9/CA@public.gmane.org>]

Hi Yi,

On 11/5/18 1:21 PM, Liu, Yi L wrote:
> Hi Baolu,
> 
>> From: iommu-bounces-cunTk1MwBs9QetFLy7KEm3xJsTq8ys+cHZ5vskTnxNA@public.gmane.org [mailto:iommu-
>> bounces-cunTk1MwBs9QetFLy7KEm3xJsTq8ys+cHZ5vskTnxNA@public.gmane.org] On Behalf Of Lu Baolu
>> Sent: Monday, November 5, 2018 10:19 AM
>> To: Joerg Roedel <joro-zLv9SwRftAIdnm+yROfE0A@public.gmane.org>; David Woodhouse <dwmw2-wEGCiKHe2LqWVfeAwA7xHQ@public.gmane.org>
>> Cc: Raj, Ashok <ashok.raj-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>; linux-kernel-u79uwXL29TY76Z2rM5mHXA@public.gmane.org;
>> iommu-cunTk1MwBs9QetFLy7KEm3xJsTq8ys+cHZ5vskTnxNA@public.gmane.org
>> Subject: [PATCH 1/1] iommu/vtd: Fix NULL pointer dereference in prq_event_thread()
>>
>> When handling page request without pasid event, go to "no_pasid"
>> branch instead of "bad_req". Otherwise, a NULL pointer deference will happen there.
>>
>> Cc: Ashok Raj <ashok.raj-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
>> Cc: Jacob Pan <jacob.jun.pan-VuQAYsv1563Yd54FQh9/CA@public.gmane.org>
>> Cc: Sohil Mehta <sohil.mehta-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
>> Signed-off-by: Lu Baolu <baolu.lu-VuQAYsv1563Yd54FQh9/CA@public.gmane.org>
>> ---
>>   drivers/iommu/intel-svm.c | 2 +-
>>   1 file changed, 1 insertion(+), 1 deletion(-)
>>
>> diff --git a/drivers/iommu/intel-svm.c b/drivers/iommu/intel-svm.c index
>> db301efe126d..887150907526 100644
>> --- a/drivers/iommu/intel-svm.c
>> +++ b/drivers/iommu/intel-svm.c
>> @@ -595,7 +595,7 @@ static irqreturn_t prq_event_thread(int irq, void *d)
>>   			pr_err("%s: Page request without PASID: %08llx %08llx\n",
>>   			       iommu->name, ((unsigned long long *)req)[0],
>>   			       ((unsigned long long *)req)[1]);
>> -			goto bad_req;
>> +			goto no_pasid;
>>   		}
>>
>>   		if (!svm || svm->pasid != req->pasid) {
>> --
> 
> I'm afraid it is still necessary to goto "bad_req". The following code behind
> "bad_req" will trigger fault_cb registered by in-kernel drivers. It is reasonable
> that PRQ without PASID can be handled by such callbacks. So I would suggest
> to keep the existing logic.

A page fault without a pasid is triggered by a DMA transfer without
PASID. It doesn't relate to the SVM functionality hence there's no
@svm or @sdev related to it. It's unnecessary to report it to the
drivers as far as I can see.

Best regards,
Lu Baolu

> 
>                  if (sdev && sdev->ops && sdev->ops->fault_cb) {
>                          int rwxp = (req->rd_req << 3) | (req->wr_req << 2) |
>                                   (req->exe_req << 1) | (req->priv_req);
>                          sdev->ops->fault_cb(sdev->dev, req->pasid, req->addr, req->private, rwxp, result);
>                   }
> 
> Thanks,
> Yi Liu
>