From: Nicolin Chen <nicolinc@nvidia.com>
To: <kevin.tian@intel.com>, Jason Gunthorpe <jgg@nvidia.com>
Cc: <corbet@lwn.net>, <will@kernel.org>, <bagasdotme@gmail.com>,
<robin.murphy@arm.com>, <joro@8bytes.org>,
<thierry.reding@gmail.com>, <vdumpa@nvidia.com>,
<jonathanh@nvidia.com>, <shuah@kernel.org>, <jsnitsel@redhat.com>,
<nathan@kernel.org>, <peterz@infradead.org>, <yi.l.liu@intel.com>,
<mshavit@google.com>, <praan@google.com>,
<zhangzekun11@huawei.com>, <iommu@lists.linux.dev>,
<linux-doc@vger.kernel.org>, <linux-kernel@vger.kernel.org>,
<linux-arm-kernel@lists.infradead.org>,
<linux-tegra@vger.kernel.org>, <linux-kselftest@vger.kernel.org>,
<patches@lists.linux.dev>, <mochs@nvidia.com>,
<alok.a.tiwari@oracle.com>, <vasant.hegde@amd.com>
Subject: Re: [PATCH v4 05/23] iommufd/driver: Let iommufd_viommu_alloc helper save ictx to viommu->ictx
Date: Thu, 15 May 2025 19:05:45 -0700 [thread overview]
Message-ID: <aCadeeP+Z4s6WzOi@Asurada-Nvidia> (raw)
In-Reply-To: <20250514170637.GE382960@nvidia.com>
On Wed, May 14, 2025 at 02:06:37PM -0300, Jason Gunthorpe wrote:
> On Thu, May 08, 2025 at 08:02:26PM -0700, Nicolin Chen wrote:
> > When an IOMMU driver calls iommufd_viommu_alloc(), it must pass in an ictx
> > pointer as the underlying _iommufd_object_alloc() helper function requires
> > that to allocate a new object. However, neither the iommufd_viommu_alloc()
> > nor its underlying _iommufd_object_alloc() saves the ictx in the allocated
> > viommu object, although viommu could hold an ictx pointer.
> >
> > When the IOMMU driver wants to use another iommufd function passing in the
> > allocated viommu, it could have avoided passing in the ictx pointer again,
> > if viommu->ictx is valid.
> >
> > Save ictx to viommu->ictx in the iommufd_viommu_alloc(), in order to ease
> > a new vIOMMU-based helper that would then get the ictx from viommu->ictx.
> >
> > Signed-off-by: Nicolin Chen <nicolinc@nvidia.com>
> > ---
> > include/linux/iommufd.h | 4 +++-
> > 1 file changed, 3 insertions(+), 1 deletion(-)
>
> It is OK, but please think carefully if the ictx is actually
> needed. The reason most objects don't have an ictx in them is because
> their contexts are always inside an ioctl so they get the ictx from
> there.
>
> We don't have a lot of viommu allocations so it isn't such a big deal,
> but just generally that is how it was built that the ictx comes from
> the ioctl not the object.
It's more of providing cleaner for-driver APIs.
Especially with the new ucmd changes, it can be just:
// in my_viommu_alloc()
my_viommu = iommufd_viommu_alloc(ucmd, ...);
iommufd_viommu_alloc_mmap(my_viommu, ...);
iommufd_viommu_destroy_mmap(my_viommu, ...);
// in my_viommu_destory()
iommufd_viommu_destroy_mmap(my_viommu, ...);
Otherwise, they would be something like:
// in my_viommu_alloc()
my_viommu = iommufd_viommu_alloc(ucmd, ...);
iommufd_viommu_alloc_mmap(ucmd->ictx, my_viommu, ...);
iommufd_viommu_destroy_mmap(ucmd->ictx, my_viommu, ...);
// in my_viommu_destory()
iommufd_viommu_destroy_mmap(my_viommu->ictx, my_viommu ...);
That being said, I ended up doing something more in this patch
for the ucmd rework and Kevin's comments:
diff --git a/drivers/iommu/iommufd/iommufd_private.h b/drivers/iommu/iommufd/iommufd_private.h
index 98fe5a49c9606..ddd144031af5d 100644
--- a/drivers/iommu/iommufd/iommufd_private.h
+++ b/drivers/iommu/iommufd/iommufd_private.h
@@ -136,14 +136,6 @@ int iopt_pin_pages(struct io_pagetable *iopt, unsigned long iova,
void iopt_unpin_pages(struct io_pagetable *iopt, unsigned long iova,
unsigned long length);
-struct iommufd_ucmd {
- struct iommufd_ctx *ictx;
- void __user *ubuffer;
- u32 user_size;
- void *cmd;
- struct iommufd_object *alloced_obj;
-};
-
int iommufd_vfio_ioctl(struct iommufd_ctx *ictx, unsigned int cmd,
unsigned long arg);
diff --git a/drivers/iommu/iommufd/viommu.c b/drivers/iommu/iommufd/viommu.c
index 416cd281eb2b5..6c0a2a0885a32 100644
--- a/drivers/iommu/iommufd/viommu.c
+++ b/drivers/iommu/iommufd/viommu.c
@@ -60,9 +60,14 @@ int iommufd_viommu_alloc_ioctl(struct iommufd_ucmd *ucmd)
goto out_put_hwpt;
}
+ /* The iommufd_viommu_alloc helper saves ucmd->ictx in viommu->ictx */
+ if (WARN_ON_ONCE(viommu->ictx != ucmd->ictx)) {
+ rc = -EINVAL;
+ goto out_put_hwpt;
+ }
+
xa_init(&viommu->vdevs);
viommu->type = cmd->type;
- viommu->ictx = ucmd->ictx;
viommu->hwpt = hwpt_paging;
refcount_inc(&viommu->hwpt->common.obj.users);
INIT_LIST_HEAD(&viommu->veventqs);
diff --git a/include/linux/iommufd.h b/include/linux/iommufd.h
index a24b924cf6872..74d11c6779739 100644
--- a/include/linux/iommufd.h
+++ b/include/linux/iommufd.h
@@ -52,6 +52,14 @@ struct iommufd_object {
unsigned int id;
};
+struct iommufd_ucmd {
+ struct iommufd_ctx *ictx;
+ void __user *ubuffer;
+ u32 user_size;
+ void *cmd;
+ struct iommufd_object *alloced_obj;
+};
+
struct iommufd_device *iommufd_device_bind(struct iommufd_ctx *ictx,
struct device *dev, u32 *id);
void iommufd_device_unbind(struct iommufd_device *idev);
@@ -252,8 +260,10 @@ static inline int iommufd_viommu_report_event(struct iommufd_viommu *viommu,
static_assert(offsetof(drv_struct, member.obj) == 0); \
ret = (drv_struct *)_iommufd_object_alloc_ucmd( \
ucmd, sizeof(drv_struct), IOMMUFD_OBJ_VIOMMU); \
- if (!IS_ERR(ret)) \
+ if (!IS_ERR(ret)) { \
ret->member.ops = viommu_ops; \
+ ret->member.ictx = ucmd->ictx; \
+ } \
ret; \
})
#endif
Thanks
Nicolin
next prev parent reply other threads:[~2025-05-16 2:06 UTC|newest]
Thread overview: 106+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-05-09 3:02 [PATCH v4 00/23] iommufd: Add vIOMMU infrastructure (Part-4 HW QUEUE) Nicolin Chen
2025-05-09 3:02 ` [PATCH v4 01/23] iommufd/viommu: Add driver-allocated vDEVICE support Nicolin Chen
2025-05-15 5:42 ` Tian, Kevin
2025-05-15 16:55 ` Nicolin Chen
2025-05-09 3:02 ` [PATCH v4 02/23] iommu: Pass in a driver-level user data structure to viommu_alloc op Nicolin Chen
2025-05-15 5:44 ` Tian, Kevin
2025-05-09 3:02 ` [PATCH v4 03/23] iommufd/viommu: Allow driver-specific user data for a vIOMMU object Nicolin Chen
2025-05-15 5:45 ` Tian, Kevin
2025-05-15 16:56 ` Nicolin Chen
2025-05-09 3:02 ` [PATCH v4 04/23] iommu: Add iommu_copy_struct_to_user helper Nicolin Chen
2025-05-15 5:46 ` Tian, Kevin
2025-05-09 3:02 ` [PATCH v4 05/23] iommufd/driver: Let iommufd_viommu_alloc helper save ictx to viommu->ictx Nicolin Chen
2025-05-14 17:06 ` Jason Gunthorpe
2025-05-16 2:05 ` Nicolin Chen [this message]
2025-05-16 13:28 ` Jason Gunthorpe
2025-05-16 20:56 ` Nicolin Chen
2025-05-26 13:30 ` Jason Gunthorpe
2025-05-27 18:41 ` Nicolin Chen
2025-05-30 18:27 ` Jason Gunthorpe
2025-05-30 18:34 ` Nicolin Chen
2025-05-15 5:48 ` Tian, Kevin
2025-05-09 3:02 ` [PATCH v4 06/23] iommufd/driver: Add iommufd_struct_destroy to revert iommufd_viommu_alloc Nicolin Chen
2025-05-14 18:26 ` Jason Gunthorpe
2025-05-14 19:21 ` Nicolin Chen
2025-05-15 12:49 ` Jason Gunthorpe
2025-05-15 16:55 ` Nicolin Chen
2025-05-09 3:02 ` [PATCH v4 07/23] iommufd/selftest: Support user_data in mock_viommu_alloc Nicolin Chen
2025-05-15 5:49 ` Tian, Kevin
2025-05-09 3:02 ` [PATCH v4 08/23] iommufd/selftest: Add covearge for viommu data Nicolin Chen
2025-05-15 5:50 ` Tian, Kevin
2025-05-09 3:02 ` [PATCH v4 09/23] iommufd: Abstract iopt_pin_pages and iopt_unpin_pages helpers Nicolin Chen
2025-05-14 18:45 ` Jason Gunthorpe
2025-05-15 5:54 ` Tian, Kevin
2025-05-09 3:02 ` [PATCH v4 10/23] iommufd/viommu: Introduce IOMMUFD_OBJ_HW_QUEUE and its related struct Nicolin Chen
2025-05-15 5:58 ` Tian, Kevin
2025-05-15 17:14 ` Nicolin Chen
2025-05-16 2:30 ` Nicolin Chen
2025-05-16 2:59 ` Tian, Kevin
2025-05-19 17:05 ` Vasant Hegde
2025-05-15 15:39 ` Jason Gunthorpe
2025-05-15 17:17 ` Nicolin Chen
2025-05-09 3:02 ` [PATCH v4 11/23] iommufd/viommu: Add IOMMUFD_CMD_HW_QUEUE_ALLOC ioctl Nicolin Chen
2025-05-15 6:30 ` Tian, Kevin
2025-05-15 18:44 ` Nicolin Chen
2025-05-16 2:49 ` Tian, Kevin
2025-05-16 3:16 ` Nicolin Chen
2025-05-16 3:52 ` Tian, Kevin
2025-05-16 4:05 ` Nicolin Chen
2025-05-18 15:19 ` Nicolin Chen
2025-05-15 16:06 ` Jason Gunthorpe
2025-05-15 18:16 ` Nicolin Chen
2025-05-15 18:59 ` Jason Gunthorpe
2025-05-15 20:32 ` Nicolin Chen
2025-05-16 13:26 ` Jason Gunthorpe
2025-05-16 2:42 ` Tian, Kevin
2025-05-16 13:25 ` Jason Gunthorpe
2025-05-19 17:29 ` Vasant Hegde
2025-05-19 18:14 ` Nicolin Chen
2025-05-20 8:38 ` Vasant Hegde
2025-05-23 1:51 ` Tian, Kevin
2025-05-26 13:29 ` Jason Gunthorpe
2025-05-09 3:02 ` [PATCH v4 12/23] iommufd/driver: Add iommufd_hw_queue_depend/undepend() helpers Nicolin Chen
2025-05-15 16:12 ` Jason Gunthorpe
2025-05-16 4:51 ` Nicolin Chen
2025-05-09 3:02 ` [PATCH v4 13/23] iommufd/selftest: Add coverage for IOMMUFD_CMD_HW_QUEUE_ALLOC Nicolin Chen
2025-05-09 3:02 ` [PATCH v4 14/23] iommufd: Add mmap interface Nicolin Chen
2025-05-09 14:13 ` kernel test robot
2025-05-09 19:30 ` Nicolin Chen
2025-05-15 6:41 ` Tian, Kevin
2025-05-15 16:47 ` Jason Gunthorpe
2025-05-16 4:08 ` Tian, Kevin
2025-05-16 13:29 ` Jason Gunthorpe
2025-05-16 17:42 ` Nicolin Chen
2025-05-09 3:02 ` [PATCH v4 15/23] iommufd/selftest: Add coverage for the new " Nicolin Chen
2025-05-09 3:02 ` [PATCH v4 16/23] Documentation: userspace-api: iommufd: Update HW QUEUE Nicolin Chen
2025-05-15 6:42 ` Tian, Kevin
2025-05-15 16:58 ` Jason Gunthorpe
2025-05-09 3:02 ` [PATCH v4 17/23] iommu/arm-smmu-v3-iommufd: Add vsmmu_alloc impl op Nicolin Chen
2025-05-15 7:52 ` Tian, Kevin
2025-05-15 17:19 ` Jason Gunthorpe
2025-05-15 17:32 ` Nicolin Chen
2025-05-09 3:02 ` [PATCH v4 18/23] iommu/arm-smmu-v3-iommufd: Support implementation-defined hw_info Nicolin Chen
2025-05-15 7:54 ` Tian, Kevin
2025-05-15 17:17 ` Jason Gunthorpe
2025-05-15 18:52 ` Nicolin Chen
2025-05-15 18:56 ` Jason Gunthorpe
2025-05-15 19:21 ` Nicolin Chen
2025-05-15 19:23 ` Jason Gunthorpe
2025-05-15 20:17 ` Nicolin Chen
2025-05-16 13:22 ` Jason Gunthorpe
2025-05-16 17:34 ` Nicolin Chen
2025-05-09 3:02 ` [PATCH v4 19/23] iommu/tegra241-cmdqv: Use request_threaded_irq Nicolin Chen
2025-05-15 7:57 ` Tian, Kevin
2025-05-15 17:21 ` Jason Gunthorpe
2025-05-09 3:02 ` [PATCH v4 20/23] iommu/tegra241-cmdqv: Simplify deinit flow in tegra241_cmdqv_remove_vintf() Nicolin Chen
2025-05-15 8:00 ` Tian, Kevin
2025-05-15 17:27 ` Jason Gunthorpe
2025-05-09 3:02 ` [PATCH v4 21/23] iommu/tegra241-cmdqv: Do not statically map LVCMDQs Nicolin Chen
2025-05-15 8:20 ` Tian, Kevin
2025-05-15 17:03 ` Nicolin Chen
2025-05-09 3:02 ` [PATCH v4 22/23] iommu/tegra241-cmdqv: Add user-space use support Nicolin Chen
2025-05-15 8:27 ` Tian, Kevin
2025-05-15 17:13 ` Nicolin Chen
2025-05-16 4:00 ` Tian, Kevin
2025-05-16 4:10 ` Nicolin Chen
2025-05-09 3:02 ` [PATCH v4 23/23] iommu/tegra241-cmdqv: Add IOMMU_VEVENTQ_TYPE_TEGRA241_CMDQV support Nicolin Chen
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=aCadeeP+Z4s6WzOi@Asurada-Nvidia \
--to=nicolinc@nvidia.com \
--cc=alok.a.tiwari@oracle.com \
--cc=bagasdotme@gmail.com \
--cc=corbet@lwn.net \
--cc=iommu@lists.linux.dev \
--cc=jgg@nvidia.com \
--cc=jonathanh@nvidia.com \
--cc=joro@8bytes.org \
--cc=jsnitsel@redhat.com \
--cc=kevin.tian@intel.com \
--cc=linux-arm-kernel@lists.infradead.org \
--cc=linux-doc@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-kselftest@vger.kernel.org \
--cc=linux-tegra@vger.kernel.org \
--cc=mochs@nvidia.com \
--cc=mshavit@google.com \
--cc=nathan@kernel.org \
--cc=patches@lists.linux.dev \
--cc=peterz@infradead.org \
--cc=praan@google.com \
--cc=robin.murphy@arm.com \
--cc=shuah@kernel.org \
--cc=thierry.reding@gmail.com \
--cc=vasant.hegde@amd.com \
--cc=vdumpa@nvidia.com \
--cc=will@kernel.org \
--cc=yi.l.liu@intel.com \
--cc=zhangzekun11@huawei.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).