On 2015/3/26 15:58, Qiu, Michael wrote: > On 3/26/2015 3:52 PM, Xie, Huawei wrote: >> On 3/26/2015 3:05 PM, Qiu, Michael wrote: >>> Function gpa_to_vva() could return zero, while this will lead >>> a Segmentation fault. >>> >>> This patch is to fix this issue. >>> >>> Signed-off-by: Michael Qiu <michael.qiu at intel.com> >>> --- >>> lib/librte_vhost/vhost_rxtx.c | 3 +++ >>> 1 file changed, 3 insertions(+) >>> >>> diff --git a/lib/librte_vhost/vhost_rxtx.c b/lib/librte_vhost/vhost_rxtx.c >>> index 535c7a1..23c8acb 100644 >>> --- a/lib/librte_vhost/vhost_rxtx.c >>> +++ b/lib/librte_vhost/vhost_rxtx.c >>> @@ -587,6 +587,9 @@ rte_vhost_dequeue_burst(struct virtio_net *dev, >>> uint16_t queue_id, >>> >>> /* Buffer address translation. */ >>> vb_addr = gpa_to_vva(dev, desc->addr); >>> + if (!vb_addr) >>> + return entry_success; >>> + >> Firstly we should add check for all gpa_to_vva translation, and do >> reporting and cleanup on error. We should avoid the case that some buggy >> or malicious guest virtio driver gives us an invalid GPA(for example, >> GPA for some MMIO space) and crash our vhost process. > > Yes, agree, I will do this for next version. > >> As we discuss, you meet segfault here, but our virtio PMD shouldn't give >> us the GPA that has no translation, so we should root cause first and > > Yes, root cause is very important, but it will spend lots time, and I > think we could be possible to apply this first(All check version). >
How to deal with invalid address but not NULL? > Thanks, > Michael >> fix the problem, and then submit the patch checking all gpa_to_vva >> translation. >> >> -Huawei >>> /* Prefetch buffer address. */ >>> rte_prefetch0((void *)(uintptr_t)vb_addr); >>> >> > > >
