From: "Cédric Le Goater" <clg@redhat.com>
To: Joao Martins <joao.m.martins@oracle.com>,
eric.auger@redhat.com, qemu-devel@nongnu.org
Cc: Yi Liu <yi.l.liu@intel.com>,
Zhenzhong Duan <zhenzhong.duan@intel.com>,
Alex Williamson <alex.williamson@redhat.com>,
Jason Gunthorpe <jgg@nvidia.com>,
Avihai Horon <avihaih@nvidia.com>
Subject: Re: [PATCH v6 4/9] vfio/{iommufd,container}: Invoke HostIOMMUDevice::realize() during attach_device()
Date: Tue, 23 Jul 2024 10:08:10 +0200 [thread overview]
Message-ID: <cc253045-be6b-4c45-afeb-6aea45b9cbc4@redhat.com> (raw)
In-Reply-To: <e0ad88d2-c78d-4d28-9655-08fd5a92fe2b@oracle.com>
On 7/23/24 10:05, Joao Martins wrote:
> On 23/07/2024 08:55, Eric Auger wrote:
>>
>>
>> On 7/23/24 09:44, Cédric Le Goater wrote:
>>> On 7/23/24 09:38, Eric Auger wrote:
>>>> Hi Joao,
>>>>
>>>> On 7/22/24 23:13, Joao Martins wrote:
>>>>> Move the HostIOMMUDevice::realize() to be invoked during the attach
>>>>> of the device
>>>>> before we allocate IOMMUFD hardware pagetable objects (HWPT). This
>>>>> allows the use
>>>>> of the hw_caps obtained by IOMMU_GET_HW_INFO that essentially tell
>>>>> if the IOMMU
>>>>> behind the device supports dirty tracking.
>>>>>
>>>>> Note: The HostIOMMUDevice data from legacy backend is static and
>>>>> doesn't
>>>>> need any information from the (type1-iommu) backend to be initialized.
>>>>> In contrast however, the IOMMUFD HostIOMMUDevice data requires the
>>>>> iommufd FD to be connected and having a devid to be able to
>>>>> successfully
>>>> Nit: maybe this comment shall be also added in iommufd.c before the call
>>>> to vfio_device_hiod_realize() to avoid someone else to move that call
>>>> earlier at some point
>>>>> GET_HW_INFO. This means vfio_device_hiod_realize() is called in
>>>>> different places within the backend .attach_device() implementation.
>>>>>
>>>>> Suggested-by: Cédric Le Goater <clg@redhat.cm>
>>>>> Signed-off-by: Joao Martins <joao.m.martins@oracle.com>
>>>>> Reviewed-by: Zhenzhong Duan <zhenzhong.duan@intel.com>
>>>>> ---
>>>>> include/hw/vfio/vfio-common.h | 1 +
>>>>> hw/vfio/common.c | 16 ++++++----------
>>>>> hw/vfio/container.c | 4 ++++
>>>>> hw/vfio/helpers.c | 11 +++++++++++
>>>>> hw/vfio/iommufd.c | 4 ++++
>>>>> 5 files changed, 26 insertions(+), 10 deletions(-)
>>>>>
>>>>> diff --git a/include/hw/vfio/vfio-common.h
>>>>> b/include/hw/vfio/vfio-common.h
>>>>> index 1a96678f8c38..4e44b26d3c45 100644
>>>>> --- a/include/hw/vfio/vfio-common.h
>>>>> +++ b/include/hw/vfio/vfio-common.h
>>>>> @@ -242,6 +242,7 @@ void vfio_region_finalize(VFIORegion *region);
>>>>> void vfio_reset_handler(void *opaque);
>>>>> struct vfio_device_info *vfio_get_device_info(int fd);
>>>>> bool vfio_device_is_mdev(VFIODevice *vbasedev);
>>>>> +bool vfio_device_hiod_realize(VFIODevice *vbasedev, Error **errp);
>>>>> bool vfio_attach_device(char *name, VFIODevice *vbasedev,
>>>>> AddressSpace *as, Error **errp);
>>>>> void vfio_detach_device(VFIODevice *vbasedev);
>>>>> diff --git a/hw/vfio/common.c b/hw/vfio/common.c
>>>>> index 784e266e6aab..da12cbd56408 100644
>>>>> --- a/hw/vfio/common.c
>>>>> +++ b/hw/vfio/common.c
>>>>> @@ -1537,7 +1537,7 @@ bool vfio_attach_device(char *name, VFIODevice
>>>>> *vbasedev,
>>>>> {
>>>>> const VFIOIOMMUClass *ops =
>>>>>
>>>>> VFIO_IOMMU_CLASS(object_class_by_name(TYPE_VFIO_IOMMU_LEGACY));
>>>>> - HostIOMMUDevice *hiod;
>>>>> + HostIOMMUDevice *hiod = NULL;
>>>>> if (vbasedev->iommufd) {
>>>>> ops =
>>>>> VFIO_IOMMU_CLASS(object_class_by_name(TYPE_VFIO_IOMMU_IOMMUFD));
>>>>> @@ -1545,21 +1545,17 @@ bool vfio_attach_device(char *name,
>>>>> VFIODevice *vbasedev,
>>>>> assert(ops);
>>>>> - if (!ops->attach_device(name, vbasedev, as, errp)) {
>>>>> - return false;
>>>>> - }
>>>>> - if (vbasedev->mdev) {
>>>>> - return true;
>>>>> + if (!vbasedev->mdev) {
>>>>> + hiod = HOST_IOMMU_DEVICE(object_new(ops->hiod_typename));
>>>>> + vbasedev->hiod = hiod;
>>>>> }
>>>>> - hiod = HOST_IOMMU_DEVICE(object_new(ops->hiod_typename));
>>>>> - if (!HOST_IOMMU_DEVICE_GET_CLASS(hiod)->realize(hiod, vbasedev,
>>>>> errp)) {
>>>>> + if (!ops->attach_device(name, vbasedev, as, errp)) {
>>>>> object_unref(hiod);
>>>>> - ops->detach_device(vbasedev);
>>>>> + vbasedev->hiod = NULL;
>>>>> return false;
>>>>> }
>>>>> - vbasedev->hiod = hiod;
>>>>> return true;
>>>>> }
>>>>> diff --git a/hw/vfio/container.c b/hw/vfio/container.c
>>>>> index 10cb4b4320ac..9ccdb639ac84 100644
>>>>> --- a/hw/vfio/container.c
>>>>> +++ b/hw/vfio/container.c
>>>>> @@ -914,6 +914,10 @@ static bool vfio_legacy_attach_device(const
>>>>> char *name, VFIODevice *vbasedev,
>>>>> trace_vfio_attach_device(vbasedev->name, groupid);
>>>>> + if (!vfio_device_hiod_realize(vbasedev, errp)) {
>>>>> + return false;
>>>> don't you want to go to err_alloc_ioas instead?
>>>
>>> hmm, the err_alloc_ioas label is in a different function
>>> iommufd_cdev_attach().
>>>
>>> may be you meant the comment for routine iommufd_cdev_attach() and
>>> label err_connect_bind ?
>>>
>>>
>>> Thanks,
>>>
>>> C.
>>>
>>>
>>>>> + }
>>>>> +
>>>>> group = vfio_get_group(groupid, as, errp);
>>>>> if (!group) {
>>>>> return false;
>>>>> diff --git a/hw/vfio/helpers.c b/hw/vfio/helpers.c
>>>>> index 7e23e9080c9d..ea15c79db0a3 100644
>>>>> --- a/hw/vfio/helpers.c
>>>>> +++ b/hw/vfio/helpers.c
>>>>> @@ -689,3 +689,14 @@ bool vfio_device_is_mdev(VFIODevice *vbasedev)
>>>>> subsys = realpath(tmp, NULL);
>>>>> return subsys && (strcmp(subsys, "/sys/bus/mdev") == 0);
>>>>> }
>>>>> +
>>>>> +bool vfio_device_hiod_realize(VFIODevice *vbasedev, Error **errp)
>>>>> +{
>>>>> + HostIOMMUDevice *hiod = vbasedev->hiod;
>>>>> +
>>>>> + if (!hiod) {
>>>>> + return true;
>>>>> + }
>>>>> +
>>>>> + return HOST_IOMMU_DEVICE_GET_CLASS(hiod)->realize(hiod,
>>>>> vbasedev, errp);
>>>>> +}
>>>>> diff --git a/hw/vfio/iommufd.c b/hw/vfio/iommufd.c
>>>>> index 5e2fc1ce089d..2324bf892c56 100644
>>>>> --- a/hw/vfio/iommufd.c
>>>>> +++ b/hw/vfio/iommufd.c
>>>>> @@ -403,6 +403,10 @@ static bool iommufd_cdev_attach(const char
>>>>> *name, VFIODevice *vbasedev,
>>>>> space = vfio_get_address_space(as);
>>>>> + if (!vfio_device_hiod_realize(vbasedev, errp)) {
>>>>> + return false;
>> Hum sorry my previous comment was targetting that place. I think
>> unrolling is needed up to put_address_space
>>
>> so effectively this does not match err_alloc_ioas but I guess we would
>> need another label
>>
>
> You're right. We haven't yet attached rthe device and that's what err_alloc_ioas
> would do. Adding another label not sure would make things cleaner given the
> ordering requirement. So maybe this instead?
>
> @@ -482,7 +483,8 @@ static bool iommufd_cdev_attach(const char *name, VFIODevice
> *vbasedev,
> space = vfio_get_address_space(as);
>
> if (!vfio_device_hiod_realize(vbasedev, errp)) {
> - return false;
> + vfio_put_address_space(space);
> + goto err_connect_bind;
> }
>
> /* try to attach to an existing container in this space */
>
LGTM.
Thanks,
C.
next prev parent reply other threads:[~2024-07-23 8:08 UTC|newest]
Thread overview: 51+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-07-22 21:13 [PATCH v6 0/9] hw/iommufd: IOMMUFD Dirty Tracking Joao Martins
2024-07-22 21:13 ` [PATCH v6 1/9] vfio/iommufd: Introduce auto domain creation Joao Martins
2024-07-23 4:38 ` Duan, Zhenzhong
2024-07-23 6:57 ` Cédric Le Goater
2024-07-23 7:18 ` Eric Auger
2024-07-22 21:13 ` [PATCH v6 2/9] vfio/{iommufd,container}: Remove caps::aw_bits Joao Martins
2024-07-23 7:21 ` Eric Auger
2024-07-22 21:13 ` [PATCH v6 3/9] vfio/iommufd: Add hw_caps field to HostIOMMUDeviceCaps Joao Martins
2024-07-23 5:11 ` Duan, Zhenzhong
2024-07-23 7:26 ` Eric Auger
2024-07-22 21:13 ` [PATCH v6 4/9] vfio/{iommufd, container}: Invoke HostIOMMUDevice::realize() during attach_device() Joao Martins via
2024-07-23 7:38 ` [PATCH v6 4/9] vfio/{iommufd,container}: " Eric Auger
2024-07-23 7:44 ` Cédric Le Goater
2024-07-23 7:55 ` Eric Auger
2024-07-23 8:05 ` Joao Martins
2024-07-23 8:08 ` Cédric Le Goater [this message]
2024-07-23 8:10 ` Eric Auger
2024-07-23 8:20 ` Duan, Zhenzhong
2024-07-23 8:24 ` Eric Auger
2024-07-23 8:26 ` Joao Martins
2024-07-23 7:53 ` Joao Martins
2024-07-23 8:00 ` Cédric Le Goater
2024-07-22 21:13 ` [PATCH v6 5/9] vfio/iommufd: Probe and request hwpt dirty tracking capability Joao Martins
2024-07-23 5:11 ` Duan, Zhenzhong
2024-07-23 6:13 ` Joao Martins
2024-07-23 6:57 ` Cédric Le Goater
2024-07-23 7:02 ` Duan, Zhenzhong
2024-07-23 7:50 ` Eric Auger
2024-07-23 8:00 ` Joao Martins
2024-07-23 8:09 ` Eric Auger
2024-07-23 8:17 ` Joao Martins
2024-07-23 11:59 ` Jason Gunthorpe
2024-07-22 21:13 ` [PATCH v6 6/9] vfio/iommufd: Implement VFIOIOMMUClass::set_dirty_tracking support Joao Martins
2024-07-23 8:03 ` Eric Auger
2024-07-23 8:14 ` Joao Martins
2024-07-23 8:17 ` Eric Auger
2024-07-22 21:13 ` [PATCH v6 7/9] vfio/iommufd: Implement VFIOIOMMUClass::query_dirty_bitmap support Joao Martins
2024-07-22 21:13 ` [PATCH v6 8/9] vfio/migration: Don't block migration device dirty tracking is unsupported Joao Martins
2024-07-23 4:45 ` Duan, Zhenzhong
2024-07-23 8:22 ` Eric Auger
2024-07-22 21:13 ` [PATCH v6 9/9] vfio/common: Allow disabling device dirty page tracking Joao Martins
2024-07-23 5:05 ` Duan, Zhenzhong
2024-07-23 8:31 ` Eric Auger
2024-07-23 8:42 ` Joao Martins
2024-07-23 10:11 ` Eric Auger
2024-07-23 8:35 ` [PATCH v6 0/9] hw/iommufd: IOMMUFD Dirty Tracking Cédric Le Goater
2024-07-23 8:56 ` Joao Martins
2024-07-23 9:08 ` Cédric Le Goater
2024-07-23 14:23 ` Yi Liu
2024-07-23 14:21 ` Joao Martins
2024-07-23 14:24 ` Cédric Le Goater
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=cc253045-be6b-4c45-afeb-6aea45b9cbc4@redhat.com \
--to=clg@redhat.com \
--cc=alex.williamson@redhat.com \
--cc=avihaih@nvidia.com \
--cc=eric.auger@redhat.com \
--cc=jgg@nvidia.com \
--cc=joao.m.martins@oracle.com \
--cc=qemu-devel@nongnu.org \
--cc=yi.l.liu@intel.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).