From: Eric Auger <eric.auger@redhat.com>
To: "Duan, Zhenzhong" <zhenzhong.duan@intel.com>,
"eric.auger.pro@gmail.com" <eric.auger.pro@gmail.com>,
"qemu-devel@nongnu.org" <qemu-devel@nongnu.org>,
"mst@redhat.com" <mst@redhat.com>,
"jasowang@redhat.com" <jasowang@redhat.com>,
"sgarzare@redhat.com" <sgarzare@redhat.com>,
"lvivier@redhat.com" <lvivier@redhat.com>
Subject: Re: [PATCH] hw/virtio/vhost: Disable IOTLB callbacks when IOMMU gets disabled
Date: Tue, 21 Jan 2025 11:34:44 +0100 [thread overview]
Message-ID: <0a9acf3d-184d-4947-be72-63e7187ef82f@redhat.com> (raw)
In-Reply-To: <SJ0PR11MB67445918D03892472C78C2DD92E62@SJ0PR11MB6744.namprd11.prod.outlook.com>
Hi Zhenzhong,
On 1/21/25 10:18 AM, Duan, Zhenzhong wrote:
> Hi Eric,
>
>> -----Original Message-----
>> From: Eric Auger <eric.auger@redhat.com>
>> Subject: [PATCH] hw/virtio/vhost: Disable IOTLB callbacks when IOMMU gets
>> disabled
>>
>> When a guest exposed with a vhost device and protected by an
>> intel IOMMU gets rebooted, we sometimes observe a spurious warning:
>>
>> Fail to lookup the translated address ffffe000
> Do you see this print once during one time reboot?
Actually this happens rarely on reboot. The reproducibility is of the
order of magnitude of 1/10 for me. I use a vm with vhost net device +
virtual intel iommu featuring a crontab job.
@reboot /usr/sbin/reboot
>
>> We observe that the IOMMU gets disabled through a write to the global
>> command register (CMAR_GCMD.TE) before the vhost device gets stopped.
>> When this warning happens it can be observed an inflight IOTLB
>> miss occurs after the IOMMU disable and before the vhost stop. In
>> that case a flat translation occurs and the check in
>> vhost_memory_region_lookup() fails.
>>
>> Let's disable the IOTLB callbacks when all IOMMU MRs have been
>> unregistered.
> Try to understand the sequence, is it like below?
>
> vhost vcpu
>
> call into vtd_iommu_translate();
No that's a kernel vhost translate request that normally tries to find
out the translated address on kernel side in the IOTLB but since the
data is not there it ends up asking for the translation to user space ...
>
> set s->dmar_enabled = false;
> switch off iommu address space;
> disable vhost IOTLB callbacks;
vtd_handle_gcmd_write/vtd_handle_gcmd_te/vtd_handle_gcmd_te which
eventually calls vhost_iommu_region_del
>
> check if !s->dmar_enabled;
> return flat translation and trigger warning
vhost inflight translation reaches user space through
vhost_device_iotlb_miss()
>
> Thanks
> Zhenzhong
>
>> Signed-off-by: Eric Auger <eric.auger@redhat.com>
>> ---
>> hw/virtio/vhost.c | 4 ++++
>> 1 file changed, 4 insertions(+)
>>
>> diff --git a/hw/virtio/vhost.c b/hw/virtio/vhost.c
>> index 6aa72fd434..128c2ab094 100644
>> --- a/hw/virtio/vhost.c
>> +++ b/hw/virtio/vhost.c
>> @@ -931,6 +931,10 @@ static void vhost_iommu_region_del(MemoryListener
>> *listener,
>> break;
>> }
>> }
>> + if (QLIST_EMPTY(&dev->iommu_list) &&
>> + dev->vhost_ops->vhost_set_iotlb_callback) {
>> + dev->vhost_ops->vhost_set_iotlb_callback(dev, false);
>> + }
>> }
>>
>> void vhost_toggle_device_iotlb(VirtIODevice *vdev)
>> --
>> 2.47.1
next prev parent reply other threads:[~2025-01-21 10:36 UTC|newest]
Thread overview: 34+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-01-20 17:33 [PATCH] hw/virtio/vhost: Disable IOTLB callbacks when IOMMU gets disabled Eric Auger
2025-01-21 3:27 ` Jason Wang
2025-01-21 7:15 ` Eric Auger
2025-01-21 16:25 ` Eric Auger
2025-01-22 7:17 ` Jason Wang
2025-01-22 7:55 ` Eric Auger
2025-01-23 1:34 ` Jason Wang
2025-01-23 8:31 ` Eric Auger
2025-01-24 1:48 ` Jason Wang
2025-01-24 2:44 ` Duan, Zhenzhong
2025-01-24 3:30 ` Jason Wang
2025-01-24 3:41 ` Jason Wang
2025-01-24 4:00 ` Duan, Zhenzhong
2025-01-24 9:20 ` Jason Wang
2025-01-24 9:50 ` Duan, Zhenzhong
2025-01-24 17:56 ` Eric Auger
2025-01-24 15:18 ` Peter Xu
2025-01-26 7:56 ` Jason Wang
2025-01-27 0:44 ` Jason Wang
2025-01-30 17:35 ` Peter Xu
2025-01-24 17:47 ` Eric Auger
2025-01-26 7:09 ` Duan, Zhenzhong
2025-01-31 9:55 ` Eric Auger
2025-02-20 15:25 ` Michael S. Tsirkin
2025-02-20 15:57 ` Eric Auger
2025-02-20 23:27 ` Michael S. Tsirkin
2025-01-21 8:31 ` Laurent Vivier
2025-01-21 8:45 ` Stefano Garzarella
2025-01-21 8:49 ` Eric Auger
2025-01-21 8:48 ` Eric Auger
2025-01-21 16:32 ` Eric Auger
2025-01-21 9:18 ` Duan, Zhenzhong
2025-01-21 10:34 ` Eric Auger [this message]
2025-01-21 10:43 ` Duan, Zhenzhong
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=0a9acf3d-184d-4947-be72-63e7187ef82f@redhat.com \
--to=eric.auger@redhat.com \
--cc=eric.auger.pro@gmail.com \
--cc=jasowang@redhat.com \
--cc=lvivier@redhat.com \
--cc=mst@redhat.com \
--cc=qemu-devel@nongnu.org \
--cc=sgarzare@redhat.com \
--cc=zhenzhong.duan@intel.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).