From: "Michael S. Tsirkin" <mst@redhat.com>
To: qemu-devel@nongnu.org
Cc: Peter Maydell <peter.maydell@linaro.org>,
Alejandro Jimenez <alejandro.j.jimenez@oracle.com>,
Marcel Apfelbaum <marcel.apfelbaum@gmail.com>,
Paolo Bonzini <pbonzini@redhat.com>,
Richard Henderson <richard.henderson@linaro.org>,
Eduardo Habkost <eduardo@habkost.net>
Subject: [PULL 67/75] amd_iommu: Toggle address translation mode on devtab entry invalidation
Date: Sun, 5 Oct 2025 15:18:30 -0400 [thread overview]
Message-ID: <5670069df966ac146cc404011af23c257380368d.1759691708.git.mst@redhat.com> (raw)
In-Reply-To: <cover.1759691708.git.mst@redhat.com>
From: Alejandro Jimenez <alejandro.j.jimenez@oracle.com>
A guest must issue an INVALIDATE_DEVTAB_ENTRY command after changing a
Device Table entry (DTE) e.g. after attaching a device and setting up its
DTE. When intercepting this event, determine if the DTE has been configured
for paging or not, and toggle the appropriate memory regions to allow DMA
address translation for the address space if needed. Requires dma-remap=on.
Signed-off-by: Alejandro Jimenez <alejandro.j.jimenez@oracle.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Message-ID: <20250919213515.917111-19-alejandro.j.jimenez@oracle.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
---
hw/i386/amd_iommu.c | 122 +++++++++++++++++++++++++++++++++++++++++++-
1 file changed, 120 insertions(+), 2 deletions(-)
diff --git a/hw/i386/amd_iommu.c b/hw/i386/amd_iommu.c
index ce5d4c3662..e916dcb2be 100644
--- a/hw/i386/amd_iommu.c
+++ b/hw/i386/amd_iommu.c
@@ -1032,18 +1032,136 @@ static void amdvi_reset_address_translation_all(AMDVIState *s)
}
}
+static void enable_dma_mode(AMDVIAddressSpace *as, bool inval_current)
+{
+ /*
+ * When enabling DMA mode for the purpose of isolating guest devices on
+ * a failure to retrieve or invalid DTE, all existing mappings must be
+ * dropped.
+ */
+ if (inval_current) {
+ IOMMUNotifier *n;
+ IOMMU_NOTIFIER_FOREACH(n, &as->iommu) {
+ amdvi_address_space_unmap(as, n);
+ }
+ }
+
+ if (as->addr_translation) {
+ return;
+ }
+
+ /* Installing DTE enabling translation, activate region */
+ as->addr_translation = true;
+ amdvi_switch_address_space(as);
+ /* Sync shadow page tables */
+ amdvi_address_space_sync(as);
+}
+
+/*
+ * If paging was previously in use in the address space
+ * - invalidate all existing mappings
+ * - switch to no_dma memory region
+ */
+static void enable_nodma_mode(AMDVIAddressSpace *as)
+{
+ IOMMUNotifier *n;
+
+ if (!as->addr_translation) {
+ /* passthrough is already active, nothing to do */
+ return;
+ }
+
+ as->addr_translation = false;
+ IOMMU_NOTIFIER_FOREACH(n, &as->iommu) {
+ /* Drop all mappings for the address space */
+ amdvi_address_space_unmap(as, n);
+ }
+ amdvi_switch_address_space(as);
+}
+
+/*
+ * A guest driver must issue the INVALIDATE_DEVTAB_ENTRY command to the IOMMU
+ * after changing a Device Table entry. We can use this fact to detect when a
+ * Device Table entry is created for a device attached to a paging domain and
+ * enable the corresponding IOMMU memory region to allow for DMA translation if
+ * appropriate.
+ */
+static void amdvi_update_addr_translation_mode(AMDVIState *s, uint16_t devid)
+{
+ uint8_t bus_num, devfn, dte_mode;
+ AMDVIAddressSpace *as;
+ uint64_t dte[4] = { 0 };
+ int ret;
+
+ /*
+ * Convert the devid encoded in the command to a bus and devfn in
+ * order to retrieve the corresponding address space.
+ */
+ bus_num = PCI_BUS_NUM(devid);
+ devfn = devid & 0xff;
+
+ /*
+ * The main buffer of size (AMDVIAddressSpace *) * (PCI_BUS_MAX) has already
+ * been allocated within AMDVIState, but must be careful to not access
+ * unallocated devfn.
+ */
+ if (!s->address_spaces[bus_num] || !s->address_spaces[bus_num][devfn]) {
+ return;
+ }
+ as = s->address_spaces[bus_num][devfn];
+
+ ret = amdvi_as_to_dte(as, dte);
+
+ if (!ret) {
+ dte_mode = (dte[0] >> AMDVI_DEV_MODE_RSHIFT) & AMDVI_DEV_MODE_MASK;
+ }
+
+ switch (ret) {
+ case 0:
+ /* DTE was successfully retrieved */
+ if (!dte_mode) {
+ enable_nodma_mode(as); /* DTE[V]=1 && DTE[Mode]=0 => passthrough */
+ } else {
+ enable_dma_mode(as, false); /* Enable DMA translation */
+ }
+ break;
+ case -AMDVI_FR_DTE_V:
+ /* DTE[V]=0, address is passed untranslated */
+ enable_nodma_mode(as);
+ break;
+ case -AMDVI_FR_DTE_RTR_ERR:
+ case -AMDVI_FR_DTE_TV:
+ /*
+ * Enforce isolation by using DMA in rare scenarios where the DTE cannot
+ * be retrieved or DTE[TV]=0. Existing mappings are dropped.
+ */
+ enable_dma_mode(as, true);
+ break;
+ }
+}
+
/* log error without aborting since linux seems to be using reserved bits */
static void amdvi_inval_devtab_entry(AMDVIState *s, uint64_t *cmd)
{
uint16_t devid = cpu_to_le16((uint16_t)extract64(cmd[0], 0, 16));
+ trace_amdvi_devtab_inval(PCI_BUS_NUM(devid), PCI_SLOT(devid),
+ PCI_FUNC(devid));
+
/* This command should invalidate internal caches of which there isn't */
if (extract64(cmd[0], 16, 44) || cmd[1]) {
amdvi_log_illegalcom_error(s, extract64(cmd[0], 60, 4),
s->cmdbuf + s->cmdbuf_head);
+ return;
+ }
+
+ /*
+ * When DMA remapping capability is enabled, check if updated DTE is setup
+ * for paging or not, and configure the corresponding memory regions.
+ */
+ if (s->dma_remap) {
+ amdvi_update_addr_translation_mode(s, devid);
}
- trace_amdvi_devtab_inval(PCI_BUS_NUM(devid), PCI_SLOT(devid),
- PCI_FUNC(devid));
}
static void amdvi_complete_ppr(AMDVIState *s, uint64_t *cmd)
--
MST
next prev parent reply other threads:[~2025-10-05 19:22 UTC|newest]
Thread overview: 88+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-10-05 19:16 [PULL 00/75] virtio,pci,pc: features, fixes Michael S. Tsirkin
2025-10-05 19:16 ` [PULL 01/75] net: bundle all offloads in a single struct Michael S. Tsirkin
2025-10-05 19:16 ` [PULL 02/75] linux-headers: deal with counted_by annotation Michael S. Tsirkin
2025-10-05 19:16 ` [PULL 03/75] linux-headers: Update to Linux v6.17-rc1 Michael S. Tsirkin
2025-10-05 19:16 ` [PULL 04/75] virtio: introduce extended features type Michael S. Tsirkin
2025-10-05 19:16 ` [PULL 05/75] virtio: serialize extended features state Michael S. Tsirkin
2025-10-05 19:16 ` [PULL 06/75] virtio: add support for negotiating extended features Michael S. Tsirkin
2025-10-05 19:16 ` [PULL 07/75] virtio-pci: implement support for " Michael S. Tsirkin
2025-10-05 19:16 ` [PULL 08/75] vhost: add support for negotiating " Michael S. Tsirkin
2025-10-05 19:16 ` [PULL 09/75] qmp: update virtio features map to support " Michael S. Tsirkin
2025-10-05 19:16 ` [PULL 10/75] vhost-backend: implement extended features support Michael S. Tsirkin
2025-10-05 19:16 ` [PULL 11/75] vhost-net: " Michael S. Tsirkin
2025-10-05 19:16 ` [PULL 12/75] virtio-net: " Michael S. Tsirkin
2025-10-05 19:16 ` [PULL 13/75] net: implement tunnel probing Michael S. Tsirkin
2025-10-05 19:16 ` [PULL 14/75] net: implement UDP tunnel features offloading Michael S. Tsirkin
2025-10-05 19:16 ` [PULL 15/75] Revert "hw/acpi/ghes: Make ghes_record_cper_errors() static" Michael S. Tsirkin
2025-10-05 19:16 ` [PULL 16/75] acpi/ghes: Cleanup the code which gets ghes ged state Michael S. Tsirkin
2025-10-05 19:16 ` [PULL 17/75] acpi/ghes: prepare to change the way HEST offsets are calculated Michael S. Tsirkin
2025-10-05 19:16 ` [PULL 18/75] acpi/ghes: add a firmware file with HEST address Michael S. Tsirkin
2025-10-05 19:16 ` [PULL 19/75] acpi/ghes: Use HEST table offsets when preparing GHES records Michael S. Tsirkin
2025-10-05 19:16 ` [PULL 20/75] acpi/ghes: don't hard-code the number of sources for HEST table Michael S. Tsirkin
2025-10-05 19:16 ` [PULL 21/75] acpi/ghes: add a notifier to notify when error data is ready Michael S. Tsirkin
2025-10-05 19:16 ` [PULL 22/75] acpi/generic_event_device: Update GHES migration to cover hest addr Michael S. Tsirkin
2025-10-05 19:16 ` [PULL 23/75] acpi/generic_event_device: add logic to detect if HEST addr is available Michael S. Tsirkin
2025-10-05 19:16 ` [PULL 24/75] acpi/generic_event_device: add an APEI error device Michael S. Tsirkin
2025-10-05 19:16 ` [PULL 25/75] tests/acpi: virt: allow acpi table changes at DSDT and HEST tables Michael S. Tsirkin
2025-10-05 19:16 ` [PULL 26/75] arm/virt: Wire up a GED error device for ACPI / GHES Michael S. Tsirkin
2025-10-05 19:17 ` [PULL 27/75] qapi/acpi-hest: add an interface to do generic CPER error injection Michael S. Tsirkin
2025-10-05 19:17 ` [PULL 28/75] acpi/generic_event_device.c: enable use_hest_addr for QEMU 10.x Michael S. Tsirkin
2025-10-05 19:17 ` [PULL 29/75] tests/acpi: virt: update HEST and DSDT tables Michael S. Tsirkin
2025-10-05 19:17 ` [PULL 30/75] docs: hest: add new "etc/acpi_table_hest_addr" and update workflow Michael S. Tsirkin
2025-10-05 19:17 ` [PULL 31/75] scripts/ghes_inject: add a script to generate GHES error inject Michael S. Tsirkin
2025-10-05 19:17 ` [PULL 32/75] hw/smbios: allow clearing the VM bit in SMBIOS table 0 Michael S. Tsirkin
2025-10-05 19:17 ` [PULL 33/75] hw/i386/pc: Avoid overlap between CXL window and PCI 64bit BARs in QEMU Michael S. Tsirkin
2025-10-06 17:08 ` Michael Tokarev
2025-10-28 19:26 ` Michael Tokarev
2025-10-29 11:01 ` Jonathan Cameron via
2025-10-29 14:10 ` Michael Tokarev
2025-10-31 10:13 ` Jonathan Cameron via
2025-10-05 19:17 ` [PULL 34/75] pcie_sriov: Fix broken MMIO accesses from SR-IOV VFs Michael S. Tsirkin
2025-10-05 19:17 ` [PULL 35/75] hw/virtio: rename vhost-user-device and make user creatable Michael S. Tsirkin
2025-10-05 19:17 ` [PULL 36/75] smbios: cap DIMM size to 2Tb as workaround for broken Windows Michael S. Tsirkin
2025-10-05 19:17 ` [PULL 37/75] pcie: Add a way to get the outstanding page request allocation (pri) from the config space Michael S. Tsirkin
2025-10-05 19:17 ` [PULL 38/75] intel_iommu: Bypass barrier wait descriptor Michael S. Tsirkin
2025-10-05 19:17 ` [PULL 39/75] intel_iommu: Declare PRI constants and structures Michael S. Tsirkin
2025-10-05 19:17 ` [PULL 40/75] intel_iommu: Declare registers for PRI Michael S. Tsirkin
2025-10-05 19:17 ` [PULL 41/75] intel_iommu: Add PRI operations support Michael S. Tsirkin
2025-10-05 19:17 ` [PULL 42/75] x86: ich9: fix default value of 'No Reboot' bit in GCS Michael S. Tsirkin
2025-10-05 19:17 ` [PULL 43/75] vhost: use virtio_config_get_guest_notifier() Michael S. Tsirkin
2025-10-05 19:17 ` [PULL 44/75] virtio: unify virtio_notify_irqfd() and virtio_notify() Michael S. Tsirkin
2025-10-05 19:17 ` [PULL 45/75] virtio: support irqfd in virtio_notify_config() Michael S. Tsirkin
2025-10-05 19:17 ` [PULL 46/75] tests/libqos: extract qvirtqueue_set_avail_idx() Michael S. Tsirkin
2025-10-05 19:17 ` [PULL 47/75] tests/virtio-scsi: add a virtio_error() IOThread test Michael S. Tsirkin
2025-10-05 19:17 ` [PULL 48/75] pcie_sriov: make pcie_sriov_pf_exit() safe on non-SR-IOV devices Michael S. Tsirkin
2025-10-05 19:17 ` [PULL 49/75] virtio: Add function name to error messages Michael S. Tsirkin
2025-10-05 20:13 ` Alessandro Ratti
2025-10-05 20:24 ` Michael S. Tsirkin
2025-10-08 10:01 ` Michael S. Tsirkin
2025-10-08 16:53 ` Alessandro Ratti
2025-10-05 20:19 ` [PULL v2 75/75] virtio: improve virtqueue mapping " Michael S. Tsirkin
2025-10-05 19:17 ` [PULL 50/75] memory: Adjust event ranges to fit within notifier boundaries Michael S. Tsirkin
2025-10-05 19:17 ` [PULL 51/75] amd_iommu: Document '-device amd-iommu' common options Michael S. Tsirkin
2025-10-05 19:17 ` [PULL 52/75] amd_iommu: Reorder device and page table helpers Michael S. Tsirkin
2025-10-05 19:17 ` [PULL 53/75] amd_iommu: Helper to decode size of page invalidation command Michael S. Tsirkin
2025-10-05 19:18 ` [PULL 54/75] amd_iommu: Add helper function to extract the DTE Michael S. Tsirkin
2025-10-05 19:18 ` [PULL 55/75] amd_iommu: Return an error when unable to read PTE from guest memory Michael S. Tsirkin
2025-10-05 19:18 ` [PULL 56/75] amd_iommu: Add helpers to walk AMD v1 Page Table format Michael S. Tsirkin
2025-10-05 19:18 ` [PULL 57/75] amd_iommu: Add a page walker to sync shadow page tables on invalidation Michael S. Tsirkin
2025-10-05 19:18 ` [PULL 58/75] amd_iommu: Add basic structure to support IOMMU notifier updates Michael S. Tsirkin
2025-10-05 19:18 ` [PULL 59/75] amd_iommu: Sync shadow page tables on page invalidation Michael S. Tsirkin
2025-10-05 19:18 ` [PULL 60/75] amd_iommu: Use iova_tree records to determine large page size on UNMAP Michael S. Tsirkin
2025-10-05 19:18 ` [PULL 61/75] amd_iommu: Unmap all address spaces under the AMD IOMMU on reset Michael S. Tsirkin
2025-10-05 19:18 ` [PULL 62/75] amd_iommu: Add replay callback Michael S. Tsirkin
2025-10-05 19:18 ` [PULL 63/75] amd_iommu: Invalidate address translations on INVALIDATE_IOMMU_ALL Michael S. Tsirkin
2025-10-05 19:18 ` [PULL 64/75] amd_iommu: Toggle memory regions based on address translation mode Michael S. Tsirkin
2025-10-05 19:18 ` [PULL 65/75] amd_iommu: Set all address spaces to use passthrough mode on reset Michael S. Tsirkin
2025-10-05 19:18 ` [PULL 66/75] amd_iommu: Add dma-remap property to AMD vIOMMU device Michael S. Tsirkin
2025-10-05 19:18 ` Michael S. Tsirkin [this message]
2025-10-05 19:18 ` [PULL 68/75] amd_iommu: Do not assume passthrough translation when DTE[TV]=0 Michael S. Tsirkin
2025-10-05 19:18 ` [PULL 69/75] amd_iommu: Refactor amdvi_page_walk() to use common code for page walk Michael S. Tsirkin
2025-10-05 19:18 ` [PULL 70/75] intel-iommu: Move dma_translation to x86-iommu Michael S. Tsirkin
2025-10-05 19:18 ` [PULL 71/75] amd_iommu: HATDis/HATS=11 support Michael S. Tsirkin
2025-10-05 19:18 ` [PULL 72/75] vdpa-dev: add get_vhost() callback for vhost-vdpa device Michael S. Tsirkin
2025-10-05 19:18 ` [PULL 73/75] intel_iommu: Enable Enhanced Set Root Table Pointer Support (ESRTPS) Michael S. Tsirkin
2025-10-05 19:18 ` [PULL 74/75] intel_iommu: Simplify caching mode check with VFIO device Michael S. Tsirkin
2025-10-05 19:18 ` [PULL 75/75] pci: Fix wrong parameter passing to pci_device_get_iommu_bus_devfn() Michael S. Tsirkin
2025-10-05 20:20 ` [PULL 00/75] virtio,pci,pc: features, fixes Michael S. Tsirkin
2025-10-06 21:59 ` Richard Henderson
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=5670069df966ac146cc404011af23c257380368d.1759691708.git.mst@redhat.com \
--to=mst@redhat.com \
--cc=alejandro.j.jimenez@oracle.com \
--cc=eduardo@habkost.net \
--cc=marcel.apfelbaum@gmail.com \
--cc=pbonzini@redhat.com \
--cc=peter.maydell@linaro.org \
--cc=qemu-devel@nongnu.org \
--cc=richard.henderson@linaro.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).