From: Alex Williamson <alex.williamson@redhat.com>
To: Pierre Morel <pmorel@linux.ibm.com>
Cc: sebott@linux.vnet.ibm.com, gerald.schaefer@de.ibm.com,
pasic@linux.vnet.ibm.com, borntraeger@de.ibm.com,
walling@linux.ibm.com, linux-s390@vger.kernel.org,
iommu@lists.linux-foundation.org, joro@8bytes.org,
linux-kernel@vger.kernel.org, kvm@vger.kernel.org,
schwidefsky@de.ibm.com, heiko.carstens@de.ibm.com
Subject: Re: [PATCH 4/4] vfio: vfio_iommu_type1: implement VFIO_IOMMU_INFO_CAPABILITIES
Date: Thu, 16 May 2019 12:40:26 -0600 [thread overview]
Message-ID: <20190516124026.415bf671@x1.home> (raw)
In-Reply-To: <1557476555-20256-5-git-send-email-pmorel@linux.ibm.com>
On Fri, 10 May 2019 10:22:35 +0200
Pierre Morel <pmorel@linux.ibm.com> wrote:
> We implement a capability intercafe for VFIO_IOMMU_GET_INFO and add the
> first capability: VFIO_IOMMU_INFO_CAPABILITIES.
>
> When calling the ioctl, the user must specify
> VFIO_IOMMU_INFO_CAPABILITIES to retrieve the capabilities and must check
> in the answer if capabilities are supported.
> Older kernel will not check nor set the VFIO_IOMMU_INFO_CAPABILITIES in
> the flags of vfio_iommu_type1_info.
>
> The iommu get_attr callback will be called to retrieve the specific
> attributes and fill the capabilities, VFIO_IOMMU_INFO_CAP_QFN for the
> PCI query function attributes and VFIO_IOMMU_INFO_CAP_QGRP for the
> PCI query function group.
>
> Signed-off-by: Pierre Morel <pmorel@linux.ibm.com>
> ---
> drivers/vfio/vfio_iommu_type1.c | 95 ++++++++++++++++++++++++++++++++++++++++-
> 1 file changed, 94 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/vfio/vfio_iommu_type1.c b/drivers/vfio/vfio_iommu_type1.c
> index d0f731c..f7f8120 100644
> --- a/drivers/vfio/vfio_iommu_type1.c
> +++ b/drivers/vfio/vfio_iommu_type1.c
> @@ -1658,6 +1658,70 @@ static int vfio_domains_have_iommu_cache(struct vfio_iommu *iommu)
> return ret;
> }
>
> +int vfio_iommu_type1_caps(struct vfio_iommu *iommu, struct vfio_info_cap *caps,
> + size_t size)
> +{
> + struct vfio_domain *d;
> + struct vfio_iommu_type1_info_block *info_fn;
> + struct vfio_iommu_type1_info_block *info_grp;
> + unsigned long total_size, fn_size, grp_size;
> + int ret;
> +
> + d = list_first_entry(&iommu->domain_list, struct vfio_domain, next);
> + if (!d)
> + return -ENODEV;
> + /* The size of these capabilities are device dependent */
> + fn_size = iommu_domain_get_attr(d->domain,
> + DOMAIN_ATTR_ZPCI_FN_SIZE, NULL);
> + if (fn_size < 0)
> + return fn_size;
What if non-Z archs want to use this? The function is architected
specifically for this one use case, fail if any component is not there
which means it requires a re-write to add further support. If
ZPCI_FN_SIZE isn't support, move on to the next thing.
> + fn_size += sizeof(struct vfio_info_cap_header);
> + total_size = fn_size;
Here too, total_size should be initialized to zero and each section +=
the size they'd like to add.
> +
> + grp_size = iommu_domain_get_attr(d->domain,
> + DOMAIN_ATTR_ZPCI_GRP_SIZE, NULL);
> + if (grp_size < 0)
> + return grp_size;
> + grp_size += sizeof(struct vfio_info_cap_header);
> + total_size += grp_size;
> +
> + /* Tell caller to call us with a greater buffer */
> + if (total_size > size) {
> + caps->size = total_size;
> + return 0;
> + }
> +
> + info_fn = kzalloc(fn_size, GFP_KERNEL);
> + if (!info_fn)
> + return -ENOMEM;
Maybe fn_size was zero because we're not on Z.
> + ret = iommu_domain_get_attr(d->domain,
> + DOMAIN_ATTR_ZPCI_FN, &info_fn->data);
Kernel internal structures != user api. Thanks,
Alex
> + if (ret < 0)
> + return ret;
> +
> + info_fn->header.id = VFIO_IOMMU_INFO_CAP_QFN;
> +
> + ret = vfio_info_add_capability(caps, &info_fn->header, fn_size);
> + if (ret)
> + goto err_fn;
> +
> + info_grp = kzalloc(grp_size, GFP_KERNEL);
> + if (!info_grp)
> + goto err_fn;
> + ret = iommu_domain_get_attr(d->domain,
> + DOMAIN_ATTR_ZPCI_GRP, &info_grp->data);
> + if (ret < 0)
> + goto err_grp;
> + info_grp->header.id = VFIO_IOMMU_INFO_CAP_QGRP;
> + ret = vfio_info_add_capability(caps, &info_grp->header, grp_size);
> +
> +err_grp:
> + kfree(info_grp);
> +err_fn:
> + kfree(info_fn);
> + return ret;
> +}
> +
> static long vfio_iommu_type1_ioctl(void *iommu_data,
> unsigned int cmd, unsigned long arg)
> {
> @@ -1679,6 +1743,8 @@ static long vfio_iommu_type1_ioctl(void *iommu_data,
> }
> } else if (cmd == VFIO_IOMMU_GET_INFO) {
> struct vfio_iommu_type1_info info;
> + struct vfio_info_cap caps = { .buf = NULL, .size = 0 };
> + int ret;
>
> minsz = offsetofend(struct vfio_iommu_type1_info, iova_pgsizes);
>
> @@ -1688,7 +1754,34 @@ static long vfio_iommu_type1_ioctl(void *iommu_data,
> if (info.argsz < minsz)
> return -EINVAL;
>
> - info.flags = VFIO_IOMMU_INFO_PGSIZES;
> + if (info.flags & VFIO_IOMMU_INFO_CAPABILITIES) {
> + minsz = offsetofend(struct vfio_iommu_type1_info,
> + cap_offset);
> + if (info.argsz < minsz)
> + return -EINVAL;
> + ret = vfio_iommu_type1_caps(iommu, &caps,
> + info.argsz - sizeof(info));
> + if (ret)
> + return ret;
> + }
> + if (caps.size) {
> + if (info.argsz < sizeof(info) + caps.size) {
> + info.argsz = sizeof(info) + caps.size;
> + info.cap_offset = 0;
> + } else {
> + if (copy_to_user((void __user *)arg +
> + sizeof(info), caps.buf,
> + caps.size)) {
> + kfree(caps.buf);
> + return -EFAULT;
> + }
> +
> + info.cap_offset = sizeof(info);
> + }
> + kfree(caps.buf);
> + }
> +
> + info.flags |= VFIO_IOMMU_INFO_PGSIZES;
>
> info.iova_pgsizes = vfio_pgsize_bitmap(iommu);
>
next prev parent reply other threads:[~2019-05-16 18:40 UTC|newest]
Thread overview: 13+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-05-10 8:22 [PATCH 0/4] Retrieving zPCI specific info with VFIO Pierre Morel
2019-05-10 8:22 ` [PATCH 1/4] s390: pci: Exporting access to CLP PCI function and PCI group Pierre Morel
2019-05-10 10:21 ` Robin Murphy
2019-05-10 14:45 ` Pierre Morel
2019-05-10 8:22 ` [PATCH 2/4] vfio: vfio_iommu_type1: Define VFIO_IOMMU_INFO_CAPABILITIES Pierre Morel
2019-05-16 14:57 ` Christian Borntraeger
2019-05-16 18:54 ` Alex Williamson
2019-05-16 18:31 ` Alex Williamson
2019-05-17 8:18 ` Pierre Morel
2019-05-10 8:22 ` [PATCH 3/4] s390: iommu: Adding get attributes for s390_iommu Pierre Morel
2019-05-10 8:22 ` [PATCH 4/4] vfio: vfio_iommu_type1: implement VFIO_IOMMU_INFO_CAPABILITIES Pierre Morel
2019-05-16 18:40 ` Alex Williamson [this message]
2019-05-17 8:17 ` Pierre Morel
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=20190516124026.415bf671@x1.home \
--to=alex.williamson@redhat.com \
--cc=borntraeger@de.ibm.com \
--cc=gerald.schaefer@de.ibm.com \
--cc=heiko.carstens@de.ibm.com \
--cc=iommu@lists.linux-foundation.org \
--cc=joro@8bytes.org \
--cc=kvm@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-s390@vger.kernel.org \
--cc=pasic@linux.vnet.ibm.com \
--cc=pmorel@linux.ibm.com \
--cc=schwidefsky@de.ibm.com \
--cc=sebott@linux.vnet.ibm.com \
--cc=walling@linux.ibm.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).