qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Steve Sistare <steven.sistare@oracle.com>
To: qemu-devel@nongnu.org
Cc: Alex Williamson <alex.williamson@redhat.com>,
	Cedric Le Goater <clg@redhat.com>, Yi Liu <yi.l.liu@intel.com>,
	Eric Auger <eric.auger@redhat.com>,
	Zhenzhong Duan <zhenzhong.duan@intel.com>,
	"Michael S. Tsirkin" <mst@redhat.com>,
	Marcel Apfelbaum <marcel.apfelbaum@gmail.com>,
	Peter Xu <peterx@redhat.com>, Fabiano Rosas <farosas@suse.de>,
	Steve Sistare <steven.sistare@oracle.com>
Subject: [PATCH V2 44/45] iommufd: preserve DMA mappings
Date: Fri, 14 Feb 2025 06:14:26 -0800	[thread overview]
Message-ID: <1739542467-226739-45-git-send-email-steven.sistare@oracle.com> (raw)
In-Reply-To: <1739542467-226739-1-git-send-email-steven.sistare@oracle.com>

During cpr-transfer load in new QEMU, the vfio_memory_listener causes
spurious calls to map and unmap DMA regions, as devices are created and
the address space is built.  This memory was already already mapped by the
device in old QEMU, so suppress the map and unmap callbacks during CPR --
eg, if the reused flag is set.  The reused flag is cleared in the post_load
handler.

Signed-off-by: Steve Sistare <steven.sistare@oracle.com>
---
 backends/iommufd.c    | 8 ++++++++
 hw/vfio/cpr-iommufd.c | 1 +
 2 files changed, 9 insertions(+)

diff --git a/backends/iommufd.c b/backends/iommufd.c
index da90b21..dfcfd6b 100644
--- a/backends/iommufd.c
+++ b/backends/iommufd.c
@@ -209,6 +209,10 @@ int iommufd_backend_map_file_dma(IOMMUFDBackend *be, uint32_t ioas_id,
         .length = size,
     };
 
+    if (be->cpr_reused) {
+        return 0;
+    }
+
     if (!readonly) {
         map.flags |= IOMMU_IOAS_MAP_WRITEABLE;
     }
@@ -240,6 +244,10 @@ int iommufd_backend_unmap_dma(IOMMUFDBackend *be, uint32_t ioas_id,
         .length = size,
     };
 
+    if (be->cpr_reused) {
+        return 0;
+    }
+
     ret = ioctl(fd, IOMMU_IOAS_UNMAP, &unmap);
     /*
      * IOMMUFD takes mapping as some kind of object, unmapping
diff --git a/hw/vfio/cpr-iommufd.c b/hw/vfio/cpr-iommufd.c
index 92b101d..286597a 100644
--- a/hw/vfio/cpr-iommufd.c
+++ b/hw/vfio/cpr-iommufd.c
@@ -66,6 +66,7 @@ static const VMStateDescription vfio_container_vmstate = {
     .name = "vfio-iommufd-container",
     .version_id = 0,
     .minimum_version_id = 0,
+    .priority = MIG_PRI_LOW,   /* Must happen after devices and groups */
     .pre_save = vfio_container_pre_save,
     .post_load = vfio_container_post_load,
     .needed = cpr_needed_for_reuse,
-- 
1.8.3.1



  parent reply	other threads:[~2025-02-14 14:21 UTC|newest]

Thread overview: 72+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-02-14 14:13 [PATCH V2 00/45] Live update: vfio and iommufd Steve Sistare
2025-02-14 14:13 ` [PATCH V2 01/45] MAINTAINERS: Add reviewer for CPR Steve Sistare
2025-02-14 14:53   ` Peter Xu
2025-02-14 20:14     ` Steven Sistare
2025-02-14 14:13 ` [PATCH V2 02/45] migration: cpr helpers Steve Sistare
2025-02-14 16:37   ` Peter Xu
2025-02-14 20:31     ` Steven Sistare
2025-02-18 16:26       ` Peter Xu
2025-02-24 16:51         ` Steven Sistare
2025-02-14 14:13 ` [PATCH V2 03/45] migration: lower handler priority Steve Sistare
2025-02-14 15:58   ` Peter Xu
2025-02-14 14:13 ` [PATCH V2 04/45] vfio: vfio_find_ram_discard_listener Steve Sistare
2025-02-14 14:13 ` [PATCH V2 05/45] vfio/container: ram discard disable helper Steve Sistare
2025-02-17 17:58   ` Cédric Le Goater
2025-02-14 14:13 ` [PATCH V2 06/45] vfio/container: reform vfio_connect_container cleanup Steve Sistare
2025-02-17 18:01   ` Cédric Le Goater
2025-02-14 14:13 ` [PATCH V2 07/45] vfio/container: vfio_container_group_add Steve Sistare
2025-02-17 18:02   ` Cédric Le Goater
2025-02-14 14:13 ` [PATCH V2 08/45] vfio/container: register container for cpr Steve Sistare
2025-02-14 14:13 ` [PATCH V2 09/45] vfio/container: preserve descriptors Steve Sistare
2025-02-14 14:13 ` [PATCH V2 10/45] vfio/container: export vfio_legacy_dma_map Steve Sistare
2025-02-14 14:13 ` [PATCH V2 11/45] vfio/container: discard old DMA vaddr Steve Sistare
2025-02-14 14:13 ` [PATCH V2 12/45] vfio/container: restore " Steve Sistare
2025-02-14 14:13 ` [PATCH V2 13/45] vfio/container: mdev cpr blocker Steve Sistare
2025-02-14 14:13 ` [PATCH V2 14/45] vfio/container: recover from unmap-all-vaddr failure Steve Sistare
2025-02-14 14:13 ` [PATCH V2 15/45] pci: export msix_is_pending Steve Sistare
2025-02-14 14:45   ` Steven Sistare
2025-02-14 14:46     ` Steven Sistare
2025-02-14 14:13 ` [PATCH V2 16/45] pci: skip reset during cpr Steve Sistare
2025-02-14 14:13 ` [PATCH V2 17/45] vfio-pci: " Steve Sistare
2025-02-14 14:14 ` [PATCH V2 18/45] vfio/pci: vfio_vector_init Steve Sistare
2025-02-14 14:14 ` [PATCH V2 19/45] vfio/pci: vfio_notifier_init Steve Sistare
2025-02-14 14:14 ` [PATCH V2 20/45] vfio/pci: pass vector to virq functions Steve Sistare
2025-02-14 14:14 ` [PATCH V2 21/45] vfio/pci: vfio_notifier_init cpr parameters Steve Sistare
2025-02-14 14:14 ` [PATCH V2 22/45] vfio/pci: vfio_notifier_cleanup Steve Sistare
2025-02-14 14:14 ` [PATCH V2 23/45] vfio/pci: export MSI functions Steve Sistare
2025-02-14 14:14 ` [PATCH V2 24/45] vfio-pci: preserve MSI Steve Sistare
2025-02-14 14:14 ` [PATCH V2 25/45] vfio-pci: preserve INTx Steve Sistare
2025-02-14 14:14 ` [PATCH V2 26/45] migration: close kvm after cpr Steve Sistare
2025-02-14 15:51   ` Steven Sistare
2025-02-14 14:14 ` [PATCH V2 27/45] migration: cpr_get_fd_param helper Steve Sistare
2025-02-14 14:14 ` [PATCH V2 28/45] vfio: return mr from vfio_get_xlat_addr Steve Sistare
2025-02-14 14:38   ` Steven Sistare
2025-02-14 16:48   ` Peter Xu
2025-02-14 20:40     ` Steven Sistare
2025-02-14 22:42       ` Peter Xu
2025-02-24 16:50         ` Steven Sistare
2025-02-24 19:20           ` Peter Xu
2025-02-24 19:35             ` Steven Sistare
2025-02-14 14:14 ` [PATCH V2 29/45] vfio: pass ramblock to vfio_container_dma_map Steve Sistare
2025-02-14 14:14 ` [PATCH V2 30/45] backends/iommufd: iommufd_backend_map_file_dma Steve Sistare
2025-02-14 14:14 ` [PATCH V2 31/45] backends/iommufd: change process ioctl Steve Sistare
2025-02-14 14:14 ` [PATCH V2 32/45] physmem: qemu_ram_get_fd_offset Steve Sistare
2025-02-14 14:39   ` Steven Sistare
2025-02-14 16:49   ` Peter Xu
2025-02-14 14:14 ` [PATCH V2 33/45] vfio/iommufd: use IOMMU_IOAS_MAP_FILE Steve Sistare
2025-02-14 14:14 ` [PATCH V2 34/45] vfio/iommufd: export iommufd_cdev_get_info_iova_range Steve Sistare
2025-02-14 14:14 ` [PATCH V2 35/45] vfio/iommufd: define hwpt constructors Steve Sistare
2025-02-14 14:14 ` [PATCH V2 36/45] vfio/iommufd: invariant device name Steve Sistare
2025-02-14 14:14 ` [PATCH V2 37/45] vfio/iommufd: fix cpr register Steve Sistare
2025-02-14 14:14 ` [PATCH V2 38/45] vfio/iommufd: register container for cpr Steve Sistare
2025-02-14 14:14 ` [PATCH V2 39/45] vfio/iommufd: preserve descriptors Steve Sistare
2025-02-14 14:14 ` [PATCH V2 40/45] vfio/iommufd: reconstruct device Steve Sistare
2025-02-14 14:14 ` [PATCH V2 41/45] vfio/iommufd: reconstruct hw_caps Steve Sistare
2025-02-14 14:14 ` [PATCH V2 42/45] vfio/iommufd: reconstruct hwpt Steve Sistare
2025-02-14 14:14 ` [PATCH V2 43/45] vfio/iommufd: change process Steve Sistare
2025-02-14 14:14 ` Steve Sistare [this message]
2025-02-14 14:14 ` [PATCH V2 45/45] vfio/container: delete old cpr register Steve Sistare
2025-02-14 15:56 ` [PATCH V2 00/45] Live update: vfio and iommufd Steven Sistare
2025-02-14 16:06   ` Peter Xu
2025-02-14 16:20     ` Steven Sistare
2025-02-14 16:48       ` 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=1739542467-226739-45-git-send-email-steven.sistare@oracle.com \
    --to=steven.sistare@oracle.com \
    --cc=alex.williamson@redhat.com \
    --cc=clg@redhat.com \
    --cc=eric.auger@redhat.com \
    --cc=farosas@suse.de \
    --cc=marcel.apfelbaum@gmail.com \
    --cc=mst@redhat.com \
    --cc=peterx@redhat.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).