* [PATCH V3] memory: RAM_NAMED_FILE flag
@ 2023-06-07 15:18 Steve Sistare
2023-06-13 9:33 ` Philippe Mathieu-Daudé
0 siblings, 1 reply; 2+ messages in thread
From: Steve Sistare @ 2023-06-07 15:18 UTC (permalink / raw)
To: qemu-devel
Cc: Paolo Bonzini, Juan Quintela, Philippe Mathieu-Daudé,
David Hildenbrand, Igor Mammedov, Peter Xu, Steve Sistare
migrate_ignore_shared() is an optimization that avoids copying memory
that is visible and can be mapped on the target. However, a
memory-backend-ram or a memory-backend-memfd block with the RAM_SHARED
flag set is not migrated when migrate_ignore_shared() is true. This is
wrong, because the block has no named backing store, and its contents will
be lost. To fix, ignore shared memory iff it is a named file. Define a
new flag RAM_NAMED_FILE to distinguish this case.
Signed-off-by: Steve Sistare <steven.sistare@oracle.com>
Reviewed-by: Peter Xu <peterx@redhat.com>
---
backends/hostmem-file.c | 1 +
include/exec/cpu-common.h | 1 +
include/exec/memory.h | 3 +++
migration/ram.c | 3 ++-
qapi/migration.json | 4 ++--
softmmu/physmem.c | 7 ++++++-
6 files changed, 15 insertions(+), 4 deletions(-)
diff --git a/backends/hostmem-file.c b/backends/hostmem-file.c
index 38ea65b..b4335a8 100644
--- a/backends/hostmem-file.c
+++ b/backends/hostmem-file.c
@@ -57,6 +57,7 @@ file_backend_memory_alloc(HostMemoryBackend *backend, Error **errp)
ram_flags = backend->share ? RAM_SHARED : 0;
ram_flags |= backend->reserve ? 0 : RAM_NORESERVE;
ram_flags |= fb->is_pmem ? RAM_PMEM : 0;
+ ram_flags |= RAM_NAMED_FILE;
memory_region_init_ram_from_file(&backend->mr, OBJECT(backend), name,
backend->size, fb->align, ram_flags,
fb->mem_path, fb->offset, fb->readonly,
diff --git a/include/exec/cpu-common.h b/include/exec/cpu-common.h
index e5a55ed..87dc9a7 100644
--- a/include/exec/cpu-common.h
+++ b/include/exec/cpu-common.h
@@ -93,6 +93,7 @@ void qemu_ram_set_uf_zeroable(RAMBlock *rb);
bool qemu_ram_is_migratable(RAMBlock *rb);
void qemu_ram_set_migratable(RAMBlock *rb);
void qemu_ram_unset_migratable(RAMBlock *rb);
+bool qemu_ram_is_named_file(RAMBlock *rb);
int qemu_ram_get_fd(RAMBlock *rb);
size_t qemu_ram_pagesize(RAMBlock *block);
diff --git a/include/exec/memory.h b/include/exec/memory.h
index c3661b2..47c2e02 100644
--- a/include/exec/memory.h
+++ b/include/exec/memory.h
@@ -232,6 +232,9 @@ typedef struct IOMMUTLBEvent {
/* RAM that isn't accessible through normal means. */
#define RAM_PROTECTED (1 << 8)
+/* RAM is an mmap-ed named file */
+#define RAM_NAMED_FILE (1 << 9)
+
static inline void iommu_notifier_init(IOMMUNotifier *n, IOMMUNotify fn,
IOMMUNotifierFlag flags,
hwaddr start, hwaddr end,
diff --git a/migration/ram.c b/migration/ram.c
index 88a6c82..cc98da3 100644
--- a/migration/ram.c
+++ b/migration/ram.c
@@ -197,7 +197,8 @@ static bool postcopy_preempt_active(void)
bool ramblock_is_ignored(RAMBlock *block)
{
return !qemu_ram_is_migratable(block) ||
- (migrate_ignore_shared() && qemu_ram_is_shared(block));
+ (migrate_ignore_shared() && qemu_ram_is_shared(block) &&
+ qemu_ram_is_named_file(block));
}
#undef RAMBLOCK_FOREACH
diff --git a/qapi/migration.json b/qapi/migration.json
index 179af0c..5bb5ab8 100644
--- a/qapi/migration.json
+++ b/qapi/migration.json
@@ -465,8 +465,8 @@
# block devices (and thus take locks) immediately at the end of
# migration. (since 3.0)
#
-# @x-ignore-shared: If enabled, QEMU will not migrate shared memory
-# (since 4.0)
+# @x-ignore-shared: If enabled, QEMU will not migrate shared memory that is
+# accessible on the destination machine. (since 4.0)
#
# @validate-uuid: Send the UUID of the source to allow the destination
# to ensure it is the same. (since 4.2)
diff --git a/softmmu/physmem.c b/softmmu/physmem.c
index 9d7e172..a8f742a 100644
--- a/softmmu/physmem.c
+++ b/softmmu/physmem.c
@@ -1570,6 +1570,11 @@ void qemu_ram_unset_migratable(RAMBlock *rb)
rb->flags &= ~RAM_MIGRATABLE;
}
+bool qemu_ram_is_named_file(RAMBlock *rb)
+{
+ return rb->flags & RAM_NAMED_FILE;
+}
+
int qemu_ram_get_fd(RAMBlock *rb)
{
return rb->fd;
@@ -1880,7 +1885,7 @@ RAMBlock *qemu_ram_alloc_from_fd(ram_addr_t size, MemoryRegion *mr,
/* Just support these ram flags by now. */
assert((ram_flags & ~(RAM_SHARED | RAM_PMEM | RAM_NORESERVE |
- RAM_PROTECTED)) == 0);
+ RAM_PROTECTED | RAM_NAMED_FILE)) == 0);
if (xen_enabled()) {
error_setg(errp, "-mem-path not supported with Xen");
--
1.8.3.1
^ permalink raw reply related [flat|nested] 2+ messages in thread
* Re: [PATCH V3] memory: RAM_NAMED_FILE flag
2023-06-07 15:18 [PATCH V3] memory: RAM_NAMED_FILE flag Steve Sistare
@ 2023-06-13 9:33 ` Philippe Mathieu-Daudé
0 siblings, 0 replies; 2+ messages in thread
From: Philippe Mathieu-Daudé @ 2023-06-13 9:33 UTC (permalink / raw)
To: Steve Sistare, qemu-devel
Cc: Paolo Bonzini, Juan Quintela, David Hildenbrand, Igor Mammedov,
Peter Xu
On 7/6/23 17:18, Steve Sistare wrote:
> migrate_ignore_shared() is an optimization that avoids copying memory
> that is visible and can be mapped on the target. However, a
> memory-backend-ram or a memory-backend-memfd block with the RAM_SHARED
> flag set is not migrated when migrate_ignore_shared() is true. This is
> wrong, because the block has no named backing store, and its contents will
> be lost. To fix, ignore shared memory iff it is a named file. Define a
> new flag RAM_NAMED_FILE to distinguish this case.
>
> Signed-off-by: Steve Sistare <steven.sistare@oracle.com>
> Reviewed-by: Peter Xu <peterx@redhat.com>
> ---
> backends/hostmem-file.c | 1 +
> include/exec/cpu-common.h | 1 +
> include/exec/memory.h | 3 +++
> migration/ram.c | 3 ++-
> qapi/migration.json | 4 ++--
> softmmu/physmem.c | 7 ++++++-
> 6 files changed, 15 insertions(+), 4 deletions(-)
I'm taking this in my next PR, thanks.
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2023-06-13 9:36 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2023-06-07 15:18 [PATCH V3] memory: RAM_NAMED_FILE flag Steve Sistare
2023-06-13 9:33 ` Philippe Mathieu-Daudé
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).