app/virtio-ha: add PF reset before DMA table clean up #105
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
When vfe-vhostd crashes, it could happen that some adminQ command is in-flight, for example, some adminQ command is sent just before the crash. Before this commit, DMA mapping of global container will be cleaned up upon vfe-vhostd quit. So for PFs, it could happen that adminQ command response comes after DMA mapping clean-up, resulting in IO_PAGE_FAULT in kernel. This commit fixes this issue by doing a PF reset before DMA clean-up. RM: 3957706 Signed-off-by: Chenbo Xia <chenbox@nvidia.com>
When doing hot-upgrade, we need to know that vfe-vhostd and vfe-vhostd-ha init finish or not. This commit adds the related log and corresponding HA IPC message so that vfe-vhostd could notify vfe-vhostd-ha that init finishes. Signed-off-by: Chenbo Xia <chenbox@nvidia.com>
Before this commit, we use HPA for checking if old and new memory region is the same or not. This is for a corner case that when vhostd restart, qemu also restart, then qemu could send different memory region with same info (QEMU_VA, GPA, SIZE). Previously we use HPA to handle this case, but the side effect is we need to use MAP_POPULATE flag for mmap call, which results in more time used in mmap. In real environment, the time could be several seconds when mmap hundreds of GB memory. This commit removes the usage of HPA and MAP_POPULATE flag, but use QEMU process id to handle the corner case. Signed-off-by: Chenbo Xia <chenbox@nvidia.com>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
When vfe-vhostd crashes, it could happen that some adminQ command is in-flight, for example, some adminQ command is sent just before the crash. Before this commit, DMA mapping of global container will be cleaned up upon vfe-vhostd quit. So for PFs, it could happen that adminQ command response comes after DMA mapping clean-up, resulting in IO_PAGE_FAULT in kernel.
This commit fixes this issue by doing a PF reset before DMA clean-up.
RM: 3957706