From: Eric Auger <eric.auger@redhat.com>
To: "Liu, Yi L" <yi.l.liu@intel.com>,
"alex.williamson@redhat.com" <alex.williamson@redhat.com>,
"jgg@nvidia.com" <jgg@nvidia.com>,
"Tian, Kevin" <kevin.tian@intel.com>
Cc: "linux-s390@vger.kernel.org" <linux-s390@vger.kernel.org>,
"yi.y.sun@linux.intel.com" <yi.y.sun@linux.intel.com>,
"kvm@vger.kernel.org" <kvm@vger.kernel.org>,
"mjrosato@linux.ibm.com" <mjrosato@linux.ibm.com>,
"intel-gvt-dev@lists.freedesktop.org"
<intel-gvt-dev@lists.freedesktop.org>,
"joro@8bytes.org" <joro@8bytes.org>,
"cohuck@redhat.com" <cohuck@redhat.com>,
"Hao, Xudong" <xudong.hao@intel.com>,
"peterx@redhat.com" <peterx@redhat.com>,
"Zhao, Yan Y" <yan.y.zhao@intel.com>,
"Xu, Terrence" <terrence.xu@intel.com>,
"nicolinc@nvidia.com" <nicolinc@nvidia.com>,
"shameerali.kolothum.thodi@huawei.com"
<shameerali.kolothum.thodi@huawei.com>,
"suravee.suthikulpanit@amd.com" <suravee.suthikulpanit@amd.com>,
"intel-gfx@lists.freedesktop.org"
<intel-gfx@lists.freedesktop.org>,
"chao.p.peng@linux.intel.com" <chao.p.peng@linux.intel.com>,
"lulu@redhat.com" <lulu@redhat.com>,
"robin.murphy@arm.com" <robin.murphy@arm.com>,
"jasowang@redhat.com" <jasowang@redhat.com>,
"Jiang, Yanting" <yanting.jiang@intel.com>
Subject: Re: [Intel-gfx] [PATCH v3 02/12] vfio/pci: Only check ownership of opened devices in hot reset
Date: Tue, 4 Apr 2023 17:59:01 +0200 [thread overview]
Message-ID: <5781064c-8742-d37d-57dc-7a7238e948d5@redhat.com> (raw)
In-Reply-To: <DS0PR11MB7529441450FE32DC9578C6B8C3939@DS0PR11MB7529.namprd11.prod.outlook.com>
On 4/4/23 17:29, Liu, Yi L wrote:
>> From: Eric Auger <eric.auger@redhat.com>
>> Sent: Tuesday, April 4, 2023 11:19 PM
>>
>> Hi Yi,
>>
>> On 4/4/23 16:37, Liu, Yi L wrote:
>>> Hi Eric,
>>>
>>>> From: Eric Auger <eric.auger@redhat.com>
>>>> Sent: Tuesday, April 4, 2023 10:00 PM
>>>>
>>>> Hi YI,
>>>>
>>>> On 4/1/23 16:44, Yi Liu wrote:
>>>>> If the affected device is not opened by any user, it's safe to reset it
>>>>> given it's not in use.
>>>>>
>>>>> Reviewed-by: Kevin Tian <kevin.tian@intel.com>
>>>>> Reviewed-by: Jason Gunthorpe <jgg@nvidia.com>
>>>>> Tested-by: Yanting Jiang <yanting.jiang@intel.com>
>>>>> Signed-off-by: Yi Liu <yi.l.liu@intel.com>
>>>>> ---
>>>>> drivers/vfio/pci/vfio_pci_core.c | 14 +++++++++++---
>>>>> include/uapi/linux/vfio.h | 8 ++++++++
>>>>> 2 files changed, 19 insertions(+), 3 deletions(-)
>>>>>
>>>>> diff --git a/drivers/vfio/pci/vfio_pci_core.c b/drivers/vfio/pci/vfio_pci_core.c
>>>>> index 65bbef562268..5d745c9abf05 100644
>>>>> --- a/drivers/vfio/pci/vfio_pci_core.c
>>>>> +++ b/drivers/vfio/pci/vfio_pci_core.c
>>>>> @@ -2429,10 +2429,18 @@ static int vfio_pci_dev_set_hot_reset(struct
>>>> vfio_device_set *dev_set,
>>>>> list_for_each_entry(cur_vma, &dev_set->device_list, vdev.dev_set_list) {
>>>>> /*
>>>>> - * Test whether all the affected devices are contained by the
>>>>> - * set of groups provided by the user.
>>>>> + * Test whether all the affected devices can be reset by the
>>>>> + * user.
>>>>> + *
>>>>> + * Resetting an unused device (not opened) is safe, because
>>>>> + * dev_set->lock is held in hot reset path so this device
>>>>> + * cannot race being opened by another user simultaneously.
>>>>> + *
>>>>> + * Otherwise all opened devices in the dev_set must be
>>>>> + * contained by the set of groups provided by the user.
>>>>> */
>>>>> - if (!vfio_dev_in_groups(cur_vma, groups)) {
>>>>> + if (cur_vma->vdev.open_count &&
>>>>> + !vfio_dev_in_groups(cur_vma, groups)) {
>>>>> ret = -EINVAL;
>>>>> goto err_undo;
>>>>> }
>>>>> diff --git a/include/uapi/linux/vfio.h b/include/uapi/linux/vfio.h
>>>>> index 0552e8dcf0cb..f96e5689cffc 100644
>>>>> --- a/include/uapi/linux/vfio.h
>>>>> +++ b/include/uapi/linux/vfio.h
>>>>> @@ -673,6 +673,14 @@ struct vfio_pci_hot_reset_info {
>>>>> * VFIO_DEVICE_PCI_HOT_RESET - _IOW(VFIO_TYPE, VFIO_BASE + 13,
>>>>> * struct vfio_pci_hot_reset)
>>>>> *
>>>>> + * Userspace requests hot reset for the devices it uses. Due to the
>>>>> + * underlying topology, multiple devices can be affected in the reset
>>>> by the reset
>>>>> + * while some might be opened by another user. To avoid interference
>>>> s/interference/hot reset failure?
>>> I don’t think user can really avoid hot reset failure since there may
>>> be new devices plugged into the affected slot. Even user has opened
>> I don't know the legacy wrt that issue but this sounds a serious issue,
>> meaning the reset of an assigned device could impact another device
>> belonging to another group not not owned by the user?
> but the hot reset shall fail as the group is not owned by the user.
sure it shall but I fail to understand if the reset fails or the device
plug is somehow delayed until the reset completes.
>
>>> all the groups/devices reported by VFIO_DEVICE_GET_PCI_HOT_RESET_INFO,
>>> the hot reset can fail if new device is plugged in and has not been
>>> bound to vfio or opened by another user during the window of
>>> _INFO and HOT_RESET.
>> with respect to the latter isn't the dev_set lock held during the hot
>> reset and sufficient to prevent any new opening to occur?
> yes. new open needs to acquire the dev_set lock. So when hot reset
> acquires the dev_set lock, then no new open can occur.
>
> Regards,
> Yi Liu
>
>>> maybe the whole statement should be as below:
>>>
>>> To avoid interference, the hot reset can only be conducted when all
>>> the affected devices are either opened by the calling user or not
>>> opened yet at the moment of the hot reset attempt.
>> OK
>>
>> Eric
>>>>> + * the calling user must ensure all affected devices, if opened, are
>>>>> + * owned by itself.
>>>>> + *
>>>>> + * The ownership is proved by an array of group fds.
>>>>> + *
>>>>> * Return: 0 on success, -errno on failure.
>>>>> */
>>>>> struct vfio_pci_hot_reset {
>>> Regards,
>>> Yi Liu
next prev parent reply other threads:[~2023-04-04 15:59 UTC|newest]
Thread overview: 145+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-04-01 14:44 [Intel-gfx] [PATCH v3 00/12] Introduce new methods for verifying ownership in vfio PCI hot reset Yi Liu
2023-04-01 14:44 ` [Intel-gfx] [PATCH v3 01/12] vfio/pci: Update comment around group_fd get in vfio_pci_ioctl_pci_hot_reset() Yi Liu
2023-04-04 13:59 ` Eric Auger
2023-04-04 14:37 ` Liu, Yi L
2023-04-01 14:44 ` [Intel-gfx] [PATCH v3 02/12] vfio/pci: Only check ownership of opened devices in hot reset Yi Liu
2023-04-04 13:59 ` Eric Auger
2023-04-04 14:37 ` Liu, Yi L
2023-04-04 15:18 ` Eric Auger
2023-04-04 15:29 ` Liu, Yi L
2023-04-04 15:59 ` Eric Auger [this message]
2023-04-05 11:41 ` Jason Gunthorpe
2023-04-05 15:14 ` Eric Auger
2023-04-01 14:44 ` [Intel-gfx] [PATCH v3 03/12] vfio/pci: Move the existing hot reset logic to be a helper Yi Liu
2023-04-04 13:59 ` Eric Auger
2023-04-04 14:24 ` Liu, Yi L
2023-04-01 14:44 ` [Intel-gfx] [PATCH v3 04/12] vfio-iommufd: Add helper to retrieve iommufd_ctx and devid for vfio_device Yi Liu
2023-04-04 15:28 ` Eric Auger
2023-04-04 21:48 ` Alex Williamson
2023-04-21 7:11 ` Liu, Yi L
2023-04-01 14:44 ` [Intel-gfx] [PATCH v3 05/12] vfio/pci: Allow passing zero-length fd array in VFIO_DEVICE_PCI_HOT_RESET Yi Liu
2023-04-04 16:54 ` Eric Auger
2023-04-04 20:18 ` Alex Williamson
2023-04-05 7:55 ` Liu, Yi L
2023-04-05 8:01 ` Liu, Yi L
2023-04-05 15:36 ` Alex Williamson
2023-04-05 16:46 ` Jason Gunthorpe
2023-04-05 8:02 ` Eric Auger
2023-04-05 8:09 ` Liu, Yi L
2023-04-01 14:44 ` [Intel-gfx] [PATCH v3 06/12] vfio: Refine vfio file kAPIs for vfio PCI hot reset Yi Liu
2023-04-05 8:27 ` Eric Auger
2023-04-05 9:23 ` Liu, Yi L
2023-04-01 14:44 ` [Intel-gfx] [PATCH v3 07/12] vfio: Accpet device file from vfio PCI hot reset path Yi Liu
2023-04-04 20:31 ` Alex Williamson
2023-04-05 8:07 ` Eric Auger
2023-04-05 8:10 ` Liu, Yi L
2023-04-01 14:44 ` [Intel-gfx] [PATCH v3 08/12] vfio/pci: Renaming for accepting device fd in " Yi Liu
2023-04-04 21:23 ` Alex Williamson
2023-04-05 9:32 ` Eric Auger
2023-04-01 14:44 ` [Intel-gfx] [PATCH v3 09/12] vfio/pci: Accept device fd in VFIO_DEVICE_PCI_HOT_RESET ioctl Yi Liu
2023-04-05 9:36 ` Eric Auger
2023-04-01 14:44 ` [Intel-gfx] [PATCH v3 10/12] vfio: Mark cdev usage in vfio_device Yi Liu
2023-04-05 11:48 ` Eric Auger
2023-04-21 7:06 ` Liu, Yi L
2023-04-01 14:44 ` [Intel-gfx] [PATCH v3 11/12] iommufd: Define IOMMUFD_INVALID_ID in uapi Yi Liu
2023-04-04 21:00 ` Alex Williamson
2023-04-05 9:31 ` Liu, Yi L
2023-04-05 15:13 ` Alex Williamson
2023-04-05 15:17 ` Liu, Yi L
2023-04-05 11:46 ` Eric Auger
2023-04-01 14:44 ` [Intel-gfx] [PATCH v3 12/12] vfio/pci: Report dev_id in VFIO_DEVICE_GET_PCI_HOT_RESET_INFO Yi Liu
2023-04-03 9:25 ` Liu, Yi L
2023-04-03 15:01 ` Alex Williamson
2023-04-03 15:22 ` Liu, Yi L
2023-04-03 15:32 ` Alex Williamson
2023-04-03 16:12 ` Jason Gunthorpe
2023-04-07 10:09 ` Liu, Yi L
2023-04-07 12:03 ` Alex Williamson
2023-04-07 13:24 ` Liu, Yi L
2023-04-07 13:51 ` Alex Williamson
2023-04-07 14:04 ` Liu, Yi L
2023-04-07 15:14 ` Alex Williamson
2023-04-07 15:47 ` Liu, Yi L
2023-04-07 21:07 ` Alex Williamson
2023-04-08 5:07 ` Liu, Yi L
2023-04-08 14:20 ` Alex Williamson
2023-04-09 11:58 ` Yi Liu
2023-04-09 13:29 ` Alex Williamson
2023-04-10 8:48 ` Liu, Yi L
2023-04-10 14:41 ` Alex Williamson
2023-04-10 15:18 ` Liu, Yi L
2023-04-10 15:23 ` Alex Williamson
2023-04-11 13:34 ` Jason Gunthorpe
2023-04-11 13:33 ` Jason Gunthorpe
2023-04-11 6:16 ` Liu, Yi L
2023-04-04 22:20 ` Alex Williamson
2023-04-05 12:19 ` Eric Auger
2023-04-05 14:04 ` Liu, Yi L
2023-04-05 16:25 ` Alex Williamson
2023-04-05 16:37 ` Jason Gunthorpe
2023-04-05 16:52 ` Alex Williamson
2023-04-05 17:23 ` Jason Gunthorpe
2023-04-05 18:56 ` Alex Williamson
2023-04-05 19:18 ` Alex Williamson
2023-04-05 19:21 ` Jason Gunthorpe
2023-04-05 19:49 ` Alex Williamson
2023-04-05 23:22 ` Jason Gunthorpe
2023-04-06 10:02 ` Liu, Yi L
2023-04-06 17:53 ` Alex Williamson
2023-04-07 10:09 ` Liu, Yi L
2023-04-11 13:24 ` Jason Gunthorpe
2023-04-11 15:54 ` Alex Williamson
2023-04-11 17:11 ` Alex Williamson
2023-04-11 18:40 ` Jason Gunthorpe
2023-04-11 21:58 ` Alex Williamson
2023-04-12 0:01 ` Jason Gunthorpe
2023-04-12 7:27 ` Tian, Kevin
2023-04-12 15:05 ` Jason Gunthorpe
2023-04-12 17:01 ` Alex Williamson
2023-04-13 2:57 ` Tian, Kevin
2023-04-12 10:09 ` Liu, Yi L
2023-04-12 16:54 ` Alex Williamson
2023-04-12 16:50 ` Alex Williamson
2023-04-12 20:06 ` Jason Gunthorpe
2023-04-13 8:25 ` Tian, Kevin
2023-04-13 11:50 ` Jason Gunthorpe
2023-04-13 14:35 ` Liu, Yi L
2023-04-13 14:41 ` Jason Gunthorpe
2023-04-13 18:07 ` Alex Williamson
2023-04-14 9:11 ` Tian, Kevin
2023-04-14 11:38 ` Liu, Yi L
2023-04-14 17:10 ` Alex Williamson
2023-04-17 4:20 ` Liu, Yi L
2023-04-17 19:01 ` Alex Williamson
2023-04-17 19:31 ` Jason Gunthorpe
2023-04-17 20:06 ` Alex Williamson
2023-04-18 3:24 ` Tian, Kevin
2023-04-18 4:10 ` Alex Williamson
2023-04-18 5:02 ` Tian, Kevin
2023-04-18 12:59 ` Jason Gunthorpe
2023-04-18 16:44 ` Alex Williamson
2023-04-18 10:34 ` Liu, Yi L
2023-04-18 16:49 ` Alex Williamson
2023-04-18 12:57 ` Jason Gunthorpe
2023-04-18 18:39 ` Alex Williamson
2023-04-20 12:10 ` Liu, Yi L
2023-04-20 14:08 ` Alex Williamson
2023-04-21 22:35 ` Jason Gunthorpe
2023-04-23 14:46 ` Liu, Yi L
2023-04-26 7:22 ` Liu, Yi L
2023-04-26 13:20 ` Alex Williamson
2023-04-26 15:08 ` Liu, Yi L
2023-04-14 16:34 ` Alex Williamson
2023-04-17 13:39 ` Jason Gunthorpe
2023-04-18 1:28 ` Tian, Kevin
2023-04-18 10:23 ` Liu, Yi L
2023-04-18 13:02 ` Jason Gunthorpe
2023-04-23 10:28 ` Liu, Yi L
2023-04-24 17:38 ` Jason Gunthorpe
2023-04-17 14:05 ` Jason Gunthorpe
2023-04-12 7:14 ` Tian, Kevin
2023-04-06 6:34 ` Liu, Yi L
2023-04-06 17:07 ` Alex Williamson
2023-04-05 17:58 ` Eric Auger
2023-04-06 5:31 ` Liu, Yi L
2023-04-01 14:47 ` [Intel-gfx] ✗ Fi.CI.BUILD: failure for Introduce new methods for verifying ownership in vfio PCI hot reset (rev4) Patchwork
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=5781064c-8742-d37d-57dc-7a7238e948d5@redhat.com \
--to=eric.auger@redhat.com \
--cc=alex.williamson@redhat.com \
--cc=chao.p.peng@linux.intel.com \
--cc=cohuck@redhat.com \
--cc=intel-gfx@lists.freedesktop.org \
--cc=intel-gvt-dev@lists.freedesktop.org \
--cc=jasowang@redhat.com \
--cc=jgg@nvidia.com \
--cc=joro@8bytes.org \
--cc=kevin.tian@intel.com \
--cc=kvm@vger.kernel.org \
--cc=linux-s390@vger.kernel.org \
--cc=lulu@redhat.com \
--cc=mjrosato@linux.ibm.com \
--cc=nicolinc@nvidia.com \
--cc=peterx@redhat.com \
--cc=robin.murphy@arm.com \
--cc=shameerali.kolothum.thodi@huawei.com \
--cc=suravee.suthikulpanit@amd.com \
--cc=terrence.xu@intel.com \
--cc=xudong.hao@intel.com \
--cc=yan.y.zhao@intel.com \
--cc=yanting.jiang@intel.com \
--cc=yi.l.liu@intel.com \
--cc=yi.y.sun@linux.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).