Linux IOMMU Development
 help / color / mirror / Atom feed
From: Nicolin Chen <nicolinc@nvidia.com>
To: Jason Gunthorpe <jgg@nvidia.com>
Cc: <kevin.tian@intel.com>, <will@kernel.org>, <robin.murphy@arm.com>,
	<joro@8bytes.org>, <ddutile@redhat.com>, <yi.l.liu@intel.com>,
	<peterz@infradead.org>, <jsnitsel@redhat.com>, <praan@google.com>,
	<linux-arm-kernel@lists.infradead.org>, <iommu@lists.linux.dev>,
	<linux-kernel@vger.kernel.org>, <patches@lists.linux.dev>,
	<baolu.lu@linux.intel.com>
Subject: Re: [PATCH v1 06/12] iommufd/selftest: Implement mock_get_viommu_size and mock_viommu_init
Date: Fri, 13 Jun 2025 16:37:05 -0700	[thread overview]
Message-ID: <aEy2IYL5XTbR54Z3@nvidia.com> (raw)
In-Reply-To: <20250613232318.GP1174925@nvidia.com>

On Fri, Jun 13, 2025 at 08:23:18PM -0300, Jason Gunthorpe wrote:
> On Fri, Jun 13, 2025 at 01:19:33PM -0700, Nicolin Chen wrote:
> > > The patches will read better if you add the call logic for init along
> > > side alloc based on init or alloc ops being non-NULL in the prior
> > > patch and then have these driver patches replace alloc with init.
> > > 
> > > Duplicating alloc into init and leaving both makes the patch harder to
> > > check.
> > 
> > I see. That will add an additional patch tentatively supporting
> > both ops.
> 
> An additional patch is not needed, just add calls to the ops in the
> same patch where you add the ops.

I have a long writing in that iommu ops patch, as Don requested
per an offline email. So separating them would be a bit cleaner.

Yet, I am merging two cleanup patches (viommu.c + iommu.h):

iommu: Introduce get_viommu_size and viommu_init ops
iommufd/viommu: Support get_viommu_size and viommu_init ops
iommufd/selftest: Drop parent domain from mock_iommu_domain_nested
iommufd/selftest: Replace mock_viommu_alloc with mock_viommu_init
iommu/arm-smmu-v3: Replace arm_vsmmu_alloc with arm_vsmmu_init
iommu: Deprecate viommu_alloc op

Net-net will be the same. And the driver patch has viommu_alloc
and viommu_init parts side by side.

> Then the patch that removes the old ops removes the extra logic
> deciding which set to call. This is the preferred way to do driver
> conversions.

Yes. I made a very small diff in the last patch:

@@ -33,8 +33,6 @@ int iommufd_viommu_alloc_ioctl(struct iommufd_ucmd *ucmd)

        ops = dev_iommu_ops(idev->dev);
        if (!ops->get_viommu_size || !ops->viommu_init) {
-               if (ops->viommu_alloc)
-                       goto get_hwpt_paging;
                rc = -EOPNOTSUPP;
                goto out_put_idev;
        }
@@ -54,7 +52,6 @@ int iommufd_viommu_alloc_ioctl(struct iommufd_ucmd *ucmd)
                goto out_put_idev;
        }

-get_hwpt_paging:
        hwpt_paging = iommufd_get_hwpt_paging(ucmd, cmd->hwpt_id);
        if (IS_ERR(hwpt_paging)) {
                rc = PTR_ERR(hwpt_paging);
@@ -66,13 +63,8 @@ int iommufd_viommu_alloc_ioctl(struct iommufd_ucmd *ucmd)
                goto out_put_hwpt;
        }

-       if (!ops->viommu_init)
-               viommu = ops->viommu_alloc(idev->dev,
-                                          hwpt_paging->common.domain,
-                                          ucmd->ictx, cmd->type);
-       else
-               viommu = (struct iommufd_viommu *)_iommufd_object_alloc(
-                       ucmd->ictx, viommu_size, IOMMUFD_OBJ_VIOMMU);
+       viommu = (struct iommufd_viommu *)_iommufd_object_alloc(
+               ucmd->ictx, viommu_size, IOMMUFD_OBJ_VIOMMU);
        if (IS_ERR(viommu)) {
                rc = PTR_ERR(viommu);
                goto out_put_hwpt;

Thanks!
Nicolin

  reply	other threads:[~2025-06-13 23:37 UTC|newest]

Thread overview: 46+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-06-09 17:13 [PATCH v1 00/12] iommufd: Repare for IOMMUFD_OBJ_HW_QUEUE Nicolin Chen
2025-06-09 17:13 ` [PATCH v1 01/12] iommufd: Apply obvious cosmetic fixes Nicolin Chen
2025-06-09 17:13 ` [PATCH v1 02/12] iommufd: Drop unused ictx in struct iommufd_vdevice Nicolin Chen
2025-06-12  8:06   ` Tian, Kevin
2025-06-13 13:36   ` Jason Gunthorpe
2025-06-09 17:13 ` [PATCH v1 03/12] iommufd: Use enum iommu_viommu_type for type in struct iommufd_viommu Nicolin Chen
2025-06-12  8:06   ` Tian, Kevin
2025-06-13 13:36   ` Jason Gunthorpe
2025-06-09 17:13 ` [PATCH v1 04/12] iommufd: Use enum iommu_veventq_type for type in struct iommufd_veventq Nicolin Chen
2025-06-12  8:07   ` Tian, Kevin
2025-06-13 13:36   ` Jason Gunthorpe
2025-06-09 17:13 ` [PATCH v1 05/12] iommu: Introduce get_viommu_size and viommu_init ops Nicolin Chen
2025-06-12  8:12   ` Tian, Kevin
2025-06-12 17:06     ` Nicolin Chen
2025-06-13  7:33       ` Tian, Kevin
2025-06-13 13:41   ` Jason Gunthorpe
2025-06-13 20:45     ` Nicolin Chen
2025-06-09 17:13 ` [PATCH v1 06/12] iommufd/selftest: Implement mock_get_viommu_size and mock_viommu_init Nicolin Chen
2025-06-12  8:17   ` Tian, Kevin
2025-06-12 17:12     ` Nicolin Chen
2025-06-13  7:34       ` Tian, Kevin
2025-06-13 13:45   ` Jason Gunthorpe
2025-06-13 20:19     ` Nicolin Chen
2025-06-13 23:23       ` Jason Gunthorpe
2025-06-13 23:37         ` Nicolin Chen [this message]
2025-06-09 17:13 ` [PATCH v1 07/12] iommu/arm-smmu-v3: Implement arm_smmu_get_viommu_size and arm_vsmmu_init Nicolin Chen
2025-06-12  8:20   ` Tian, Kevin
2025-06-12 17:18     ` Nicolin Chen
2025-06-13  7:36       ` Tian, Kevin
2025-06-09 17:13 ` [PATCH v1 08/12] iommufd/viommu: Replace ops->viommu_alloc with ops->viommu_init Nicolin Chen
2025-06-10  5:55   ` Baolu Lu
2025-06-10  6:19     ` Nicolin Chen
2025-06-12  8:22       ` Tian, Kevin
2025-06-12 17:35         ` Nicolin Chen
2025-06-12  8:27   ` Tian, Kevin
2025-06-12 17:24     ` Nicolin Chen
2025-06-09 17:13 ` [PATCH v1 09/12] iommu: Deprecate viommu_alloc op Nicolin Chen
2025-06-12  8:23   ` Tian, Kevin
2025-06-09 17:13 ` [PATCH v1 10/12] iommufd: Move _iommufd_object_alloc out of driver.c Nicolin Chen
2025-06-12  8:23   ` Tian, Kevin
2025-06-13 13:46   ` Jason Gunthorpe
2025-06-09 17:13 ` [PATCH v1 11/12] iommufd: Introduce iommufd_object_alloc_ucmd helper Nicolin Chen
2025-06-13 13:58   ` Jason Gunthorpe
2025-06-13 23:17     ` Nicolin Chen
2025-06-09 17:13 ` [PATCH v1 12/12] iommufd: Apply the new " Nicolin Chen
2025-06-10 18:32 ` [PATCH v1 00/12] iommufd: Repare for IOMMUFD_OBJ_HW_QUEUE 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=aEy2IYL5XTbR54Z3@nvidia.com \
    --to=nicolinc@nvidia.com \
    --cc=baolu.lu@linux.intel.com \
    --cc=ddutile@redhat.com \
    --cc=iommu@lists.linux.dev \
    --cc=jgg@nvidia.com \
    --cc=joro@8bytes.org \
    --cc=jsnitsel@redhat.com \
    --cc=kevin.tian@intel.com \
    --cc=linux-arm-kernel@lists.infradead.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=patches@lists.linux.dev \
    --cc=peterz@infradead.org \
    --cc=praan@google.com \
    --cc=robin.murphy@arm.com \
    --cc=will@kernel.org \
    --cc=yi.l.liu@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