48e88e8e19
- kvm-hw-virtio-virtio-iommu-Migrate-to-3-phase-reset.patch [RHEL-7188] - kvm-hw-i386-intel-iommu-Migrate-to-3-phase-reset.patch [RHEL-7188] - kvm-hw-arm-smmuv3-Move-reset-to-exit-phase.patch [RHEL-7188] - kvm-hw-vfio-common-Add-a-trace-point-in-vfio_reset_handl.patch [RHEL-7188] - kvm-docs-devel-reset-Document-reset-expectations-for-DMA.patch [RHEL-7188] - kvm-qga-implement-a-guest-get-load-command.patch [RHEL-69622] - kvm-migration-Fix-UAF-for-incoming-migration-on-Migratio.patch [RHEL-69775] - kvm-scripts-improve-error-from-qemu-trace-stap-on-missin.patch [RHEL-47340] - kvm-Recommend-systemtap-client-from-qemu-tools.patch [RHEL-47340] - Resolves: RHEL-7188 ([intel iommu][PF] DMAR: DRHD: handling fault status reg) - Resolves: RHEL-69622 ([qemu-guest-agent][RFE] Report CPU load average) - Resolves: RHEL-69775 (Guest crashed on the target host when the migration was canceled) - Resolves: RHEL-47340 ([Qemu RHEL-9] qemu-trace-stap should handle lack of stap more gracefully)
54 lines
2.0 KiB
Diff
54 lines
2.0 KiB
Diff
From 389c3c6b4215c9be3fd784c73af0e9795e796380 Mon Sep 17 00:00:00 2001
|
|
From: Eric Auger <eric.auger@redhat.com>
|
|
Date: Tue, 18 Feb 2025 19:25:35 +0100
|
|
Subject: [PATCH 5/9] docs/devel/reset: Document reset expectations for DMA and
|
|
IOMMU
|
|
MIME-Version: 1.0
|
|
Content-Type: text/plain; charset=UTF-8
|
|
Content-Transfer-Encoding: 8bit
|
|
|
|
RH-Author: Eric Auger <eric.auger@redhat.com>
|
|
RH-MergeRequest: 341: Fix vIOMMU reset order
|
|
RH-Jira: RHEL-7188
|
|
RH-Acked-by: Peter Xu <peterx@redhat.com>
|
|
RH-Acked-by: Donald Dutile <None>
|
|
RH-Acked-by: Cédric Le Goater <clg@redhat.com>
|
|
RH-Commit: [5/5] be8b9d9e34a2b301430dfa229c6785ab17d3fb16 (eauger1/centos-qemu-kvm)
|
|
|
|
To avoid any translation faults, the IOMMUs are expected to be
|
|
reset after the devices they protect. Document that we expect
|
|
DMA requests to be stopped during the 'enter' or 'hold' phase
|
|
while IOMMUs should be reset during the 'exit' phase.
|
|
|
|
Signed-off-by: Eric Auger <eric.auger@redhat.com>
|
|
Reviewed-by: Zhenzhong Duan <zhenzhong.duan@intel.com>
|
|
Message-Id: <20250218182737.76722-6-eric.auger@redhat.com>
|
|
Reviewed-by: Peter Xu <peterx@redhat.com>
|
|
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
|
|
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
|
|
(cherry picked from commit dd6d545e8f2d9a0e8a8c287ec16469f03ef5c198)
|
|
Signed-off-by: Eric Auger <eric.auger@redhat.com>
|
|
---
|
|
docs/devel/reset.rst | 5 +++++
|
|
1 file changed, 5 insertions(+)
|
|
|
|
diff --git a/docs/devel/reset.rst b/docs/devel/reset.rst
|
|
index 9746a4e8a0..24ab630465 100644
|
|
--- a/docs/devel/reset.rst
|
|
+++ b/docs/devel/reset.rst
|
|
@@ -123,6 +123,11 @@ The *exit* phase is executed only when the last reset operation ends. Therefore
|
|
the object does not need to care how many of reset controllers it has and how
|
|
many of them have started a reset.
|
|
|
|
+DMA capable devices are expected to cancel all outstanding DMA operations
|
|
+during either 'enter' or 'hold' phases. IOMMUs are expected to reset during
|
|
+the 'exit' phase and this sequencing makes sure no outstanding DMA request
|
|
+will fault.
|
|
+
|
|
|
|
Handling reset in a resettable object
|
|
-------------------------------------
|
|
--
|
|
2.48.1
|
|
|