From: Alex Williamson <alex.williamson@redhat.com>
To: Alexey Kardashevskiy <aik@ozlabs.ru>
Cc: linuxppc-dev@lists.ozlabs.org, kvm@vger.kernel.org,
Paul Mackerras <paulus@samba.org>,
linux-kernel@vger.kernel.org
Subject: Re: [PATCH kernel v6 26/29] vfio: powerpc/spapr: Define v2 IOMMU
Date: Mon, 16 Mar 2015 13:45:37 -0600 [thread overview]
Message-ID: <1426535137.3643.270.camel@redhat.com> (raw)
In-Reply-To: <1426234057-16165-27-git-send-email-aik@ozlabs.ru>
On Fri, 2015-03-13 at 19:07 +1100, Alexey Kardashevskiy wrote:
> The existing IOMMU requires VFIO_IOMMU_ENABLE call to enable actual use
> of the container (i.e. call DMA map/unmap) and this is where we check
> the rlimit for locked pages. It assumes that only as much memory
> as a default DMA window can be mapped. Every DMA map/unmap request will
> do pinning/unpinning of physical pages.
>
> New IOMMU will split physical pages pinning and TCE table update.
> It will require guest pages to be registered first and consequent
> map/unmap requests to work only with pre-registered memory.
> For the default single window case this means that the entire guest
> (instead of 2GB) needs to be pinned before using VFIO.
> However when a huge DMA window is added, no additional pinning will be
> required, otherwise it would be guest RAM + 2GB.
>
> This advertises v2 SPAPR TCE IOMMU and restricts what the userspace
> can do with v1 or v2 IOMMUs.
>
> Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru>
> ---
> Changes:
> v6:
> * enforced limitations imposed by the SPAPR TCE IOMMU version
> ---
> drivers/vfio/vfio_iommu_spapr_tce.c | 18 +++++++++++++++++-
> include/uapi/linux/vfio.h | 2 ++
> 2 files changed, 19 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/vfio/vfio_iommu_spapr_tce.c b/drivers/vfio/vfio_iommu_spapr_tce.c
> index 9d240b4..e191438 100644
> --- a/drivers/vfio/vfio_iommu_spapr_tce.c
> +++ b/drivers/vfio/vfio_iommu_spapr_tce.c
> @@ -95,6 +95,7 @@ struct tce_container {
> bool enabled;
> unsigned long locked_pages;
> struct list_head mem_list;
> + bool v2;
> };
>
> struct tce_memory {
> @@ -398,7 +399,7 @@ static void *tce_iommu_open(unsigned long arg)
> {
> struct tce_container *container;
>
> - if (arg != VFIO_SPAPR_TCE_IOMMU) {
> + if ((arg != VFIO_SPAPR_TCE_IOMMU) && (arg != VFIO_SPAPR_TCE_v2_IOMMU)) {
> pr_err("tce_vfio: Wrong IOMMU type\n");
> return ERR_PTR(-EINVAL);
> }
> @@ -410,6 +411,8 @@ static void *tce_iommu_open(unsigned long arg)
> mutex_init(&container->lock);
> INIT_LIST_HEAD_RCU(&container->mem_list);
>
> + container->v2 = arg == VFIO_SPAPR_TCE_v2_IOMMU;
> +
> return container;
> }
>
> @@ -580,6 +583,7 @@ static long tce_iommu_ioctl(void *iommu_data,
> case VFIO_CHECK_EXTENSION:
> switch (arg) {
> case VFIO_SPAPR_TCE_IOMMU:
> + case VFIO_SPAPR_TCE_v2_IOMMU:
> ret = 1;
> break;
> default:
> @@ -719,6 +723,9 @@ static long tce_iommu_ioctl(void *iommu_data,
> case VFIO_IOMMU_SPAPR_REGISTER_MEMORY: {
> struct vfio_iommu_spapr_register_memory param;
>
> + if (!container->v2)
> + return -EPERM;
> +
> minsz = offsetofend(struct vfio_iommu_spapr_register_memory,
> size);
>
> @@ -741,6 +748,9 @@ static long tce_iommu_ioctl(void *iommu_data,
> case VFIO_IOMMU_SPAPR_UNREGISTER_MEMORY: {
> struct vfio_iommu_spapr_register_memory param;
>
> + if (!container->v2)
> + return -EPERM;
> +
> minsz = offsetofend(struct vfio_iommu_spapr_register_memory,
> size);
>
> @@ -761,6 +771,9 @@ static long tce_iommu_ioctl(void *iommu_data,
> return 0;
> }
> case VFIO_IOMMU_ENABLE:
> + if (container->v2)
> + return -EPERM;
> +
> mutex_lock(&container->lock);
> ret = tce_iommu_enable(container);
> mutex_unlock(&container->lock);
> @@ -768,6 +781,9 @@ static long tce_iommu_ioctl(void *iommu_data,
>
>
> case VFIO_IOMMU_DISABLE:
> + if (container->v2)
> + return -EPERM;
> +
> mutex_lock(&container->lock);
> tce_iommu_disable(container);
> mutex_unlock(&container->lock);
I wouldn't have guessed; nothing in the documentation suggests these
ioctls are deprecated in v2 (ie. please document). If the ioctl doesn't
exist for the IOMMU type, why not simply break and let it fall out at
-ENOTTY? Same for the above, v1 would have previously returned -ENOTTY
for those ioctls, why change to -EPERM?
> diff --git a/include/uapi/linux/vfio.h b/include/uapi/linux/vfio.h
> index b17e120..fbc5286 100644
> --- a/include/uapi/linux/vfio.h
> +++ b/include/uapi/linux/vfio.h
> @@ -36,6 +36,8 @@
> /* Two-stage IOMMU */
> #define VFIO_TYPE1_NESTING_IOMMU 6 /* Implies v2 */
>
> +#define VFIO_SPAPR_TCE_v2_IOMMU 7
> +
> /*
> * The IOCTL interface is designed for extensibility by embedding the
> * structure length (argsz) and flags into structures passed between
next prev parent reply other threads:[~2015-03-16 19:45 UTC|newest]
Thread overview: 35+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-03-13 8:07 [PATCH kernel v6 00/29] powerpc/iommu/vfio: Enable Dynamic DMA windows Alexey Kardashevskiy
2015-03-13 8:07 ` [PATCH kernel v6 01/29] vfio: powerpc/spapr: Move page pinning from arch code to VFIO IOMMU driver Alexey Kardashevskiy
2015-03-13 8:07 ` [PATCH kernel v6 02/29] vfio: powerpc/spapr: Do cleanup when releasing the group Alexey Kardashevskiy
2015-03-13 8:07 ` [PATCH kernel v6 03/29] vfio: powerpc/spapr: Check that TCE page size is equal to it_page_size Alexey Kardashevskiy
2015-03-13 8:07 ` [PATCH kernel v6 04/29] vfio: powerpc/spapr: Use it_page_size Alexey Kardashevskiy
2015-03-13 8:07 ` [PATCH kernel v6 05/29] vfio: powerpc/spapr: Move locked_vm accounting to helpers Alexey Kardashevskiy
2015-03-13 8:07 ` [PATCH kernel v6 06/29] vfio: powerpc/spapr: Disable DMA mappings on disabled container Alexey Kardashevskiy
2015-03-13 8:07 ` [PATCH kernel v6 07/29] vfio: powerpc/spapr: Moving pinning/unpinning to helpers Alexey Kardashevskiy
2015-03-13 8:07 ` [PATCH kernel v6 08/29] vfio: powerpc/spapr: Register memory Alexey Kardashevskiy
2015-03-13 8:07 ` [PATCH kernel v6 09/29] vfio: powerpc/spapr: Rework attach/detach Alexey Kardashevskiy
2015-03-13 8:07 ` [PATCH kernel v6 10/29] powerpc/powernv: Do not set "read" flag if direction==DMA_NONE Alexey Kardashevskiy
2015-03-13 8:07 ` [PATCH kernel v6 11/29] powerpc/iommu: Move tce_xxx callbacks from ppc_md to iommu_table Alexey Kardashevskiy
2015-03-13 8:07 ` [PATCH kernel v6 12/29] powerpc/iommu: Introduce iommu_table_alloc() helper Alexey Kardashevskiy
2015-03-13 8:07 ` [PATCH kernel v6 13/29] powerpc/spapr: vfio: Switch from iommu_table to new iommu_table_group Alexey Kardashevskiy
2015-03-13 8:07 ` [PATCH kernel v6 14/29] vfio: powerpc/spapr: powerpc/iommu: Rework IOMMU ownership control Alexey Kardashevskiy
2015-03-13 8:07 ` [PATCH kernel v6 15/29] vfio: powerpc/spapr: powerpc/powernv/ioda2: " Alexey Kardashevskiy
2015-03-13 8:07 ` [PATCH kernel v6 16/29] powerpc/iommu: Fix IOMMU ownership control functions Alexey Kardashevskiy
2015-03-13 8:07 ` [PATCH kernel v6 17/29] powerpc/powernv/ioda/ioda2: Rework tce_build()/tce_free() Alexey Kardashevskiy
2015-03-13 8:07 ` [PATCH kernel v6 18/29] powerpc/iommu/powernv: Release replaced TCE Alexey Kardashevskiy
2015-03-13 8:07 ` [PATCH kernel v6 19/29] powerpc/powernv/ioda2: Rework iommu_table creation Alexey Kardashevskiy
2015-03-13 8:07 ` [PATCH kernel v6 20/29] powerpc/powernv/ioda2: Introduce pnv_pci_ioda2_create_table/pnc_pci_free_table Alexey Kardashevskiy
2015-03-13 8:07 ` [PATCH kernel v6 21/29] powerpc/powernv/ioda2: Introduce pnv_pci_ioda2_set_window Alexey Kardashevskiy
2015-03-13 8:07 ` [PATCH kernel v6 22/29] powerpc/iommu: Split iommu_free_table into 2 helpers Alexey Kardashevskiy
2015-03-13 8:07 ` [PATCH kernel v6 23/29] powerpc/powernv: Implement multilevel TCE tables Alexey Kardashevskiy
2015-03-13 8:07 ` [PATCH kernel v6 24/29] powerpc/powernv: Change prototypes to receive iommu Alexey Kardashevskiy
2015-03-13 8:07 ` [PATCH kernel v6 25/29] powerpc/powernv/ioda: Define and implement DMA table/window management callbacks Alexey Kardashevskiy
2015-03-13 8:07 ` [PATCH kernel v6 26/29] vfio: powerpc/spapr: Define v2 IOMMU Alexey Kardashevskiy
2015-03-16 19:45 ` Alex Williamson [this message]
2015-03-17 2:59 ` Alexey Kardashevskiy
2015-03-13 8:07 ` [PATCH kernel v6 27/29] vfio: powerpc/spapr: powerpc/powernv/ioda2: Rework ownership Alexey Kardashevskiy
2015-03-13 8:07 ` [PATCH kernel v6 28/29] vfio: powerpc/spapr: Support multiple groups in one container if possible Alexey Kardashevskiy
2015-03-13 8:07 ` [PATCH kernel v6 29/29] vfio: powerpc/spapr: Support Dynamic DMA windows Alexey Kardashevskiy
2015-03-16 19:38 ` Alex Williamson
2015-03-17 1:02 ` Alexey Kardashevskiy
2015-03-17 2:49 ` Alex Williamson
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=1426535137.3643.270.camel@redhat.com \
--to=alex.williamson@redhat.com \
--cc=aik@ozlabs.ru \
--cc=kvm@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linuxppc-dev@lists.ozlabs.org \
--cc=paulus@samba.org \
/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).