[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[PULL 16/45] softmmu/physmem: Fix ram_block_discard_range() to handle sh
From: |
Paolo Bonzini |
Subject: |
[PULL 16/45] softmmu/physmem: Fix ram_block_discard_range() to handle shared anonymous memory |
Date: |
Thu, 17 Jun 2021 11:31:05 +0200 |
From: David Hildenbrand <david@redhat.com>
We can create shared anonymous memory via
"-object memory-backend-ram,share=on,..."
which is, for example, required by PVRDMA for mremap() to work.
Shared anonymous memory is weird, though. Instead of MADV_DONTNEED, we
have to use MADV_REMOVE: MADV_DONTNEED will only remove / zap all
relevant page table entries of the current process, the backend storage
will not get removed, resulting in no reduced memory consumption and
a repopulation of previous content on next access.
Shared anonymous memory is internally really just shmem, but without a
fd exposed. As we cannot use fallocate() without the fd to discard the
backing storage, MADV_REMOVE gets the same job done without a fd as
documented in "man 2 madvise". Removing backing storage implicitly
invalidates all page table entries with relevant mappings - an additional
MADV_DONTNEED is not required.
Fixes: 06329ccecfa0 ("mem: add share parameter to memory-backend-ram")
Reviewed-by: Peter Xu <peterx@redhat.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Signed-off-by: David Hildenbrand <david@redhat.com>
Message-Id: <20210406080126.24010-3-david@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
---
include/qemu/osdep.h | 4 ++--
softmmu/physmem.c | 7 ++++++-
2 files changed, 8 insertions(+), 3 deletions(-)
diff --git a/include/qemu/osdep.h b/include/qemu/osdep.h
index 236a045671..18a9e3fb4c 100644
--- a/include/qemu/osdep.h
+++ b/include/qemu/osdep.h
@@ -410,7 +410,7 @@ void qemu_anon_ram_free(void *ptr, size_t size);
#ifdef MADV_REMOVE
#define QEMU_MADV_REMOVE MADV_REMOVE
#else
-#define QEMU_MADV_REMOVE QEMU_MADV_INVALID
+#define QEMU_MADV_REMOVE QEMU_MADV_DONTNEED
#endif
#elif defined(CONFIG_POSIX_MADVISE)
@@ -424,7 +424,7 @@ void qemu_anon_ram_free(void *ptr, size_t size);
#define QEMU_MADV_DONTDUMP QEMU_MADV_INVALID
#define QEMU_MADV_HUGEPAGE QEMU_MADV_INVALID
#define QEMU_MADV_NOHUGEPAGE QEMU_MADV_INVALID
-#define QEMU_MADV_REMOVE QEMU_MADV_INVALID
+#define QEMU_MADV_REMOVE QEMU_MADV_DONTNEED
#else /* no-op */
diff --git a/softmmu/physmem.c b/softmmu/physmem.c
index b78b30e7ba..c0a3c47167 100644
--- a/softmmu/physmem.c
+++ b/softmmu/physmem.c
@@ -3527,6 +3527,7 @@ int ram_block_discard_range(RAMBlock *rb, uint64_t start,
size_t length)
/* The logic here is messy;
* madvise DONTNEED fails for hugepages
* fallocate works on hugepages and shmem
+ * shared anonymous memory requires madvise REMOVE
*/
need_madvise = (rb->page_size == qemu_host_page_size);
need_fallocate = rb->fd != -1;
@@ -3560,7 +3561,11 @@ int ram_block_discard_range(RAMBlock *rb, uint64_t
start, size_t length)
* fallocate'd away).
*/
#if defined(CONFIG_MADVISE)
- ret = madvise(host_startaddr, length, MADV_DONTNEED);
+ if (qemu_ram_is_shared(rb) && rb->fd < 0) {
+ ret = madvise(host_startaddr, length, QEMU_MADV_REMOVE);
+ } else {
+ ret = madvise(host_startaddr, length, QEMU_MADV_DONTNEED);
+ }
if (ret) {
ret = -errno;
error_report("ram_block_discard_range: Failed to discard range
"
--
2.31.1
- [PULL 02/45] vl: Fix an assert failure in error path, (continued)
- [PULL 02/45] vl: Fix an assert failure in error path, Paolo Bonzini, 2021/06/17
- [PULL 04/45] block/scsi: correctly emulate the VPD block limits page, Paolo Bonzini, 2021/06/17
- [PULL 03/45] qemu-config: use qemu_opts_from_qdict, Paolo Bonzini, 2021/06/17
- [PULL 08/45] esp: allow non-DMA callback in esp_transfer_data() initial transfer, Paolo Bonzini, 2021/06/17
- [PULL 05/45] runstate: Initialize Error * to NULL, Paolo Bonzini, 2021/06/17
- [PULL 06/45] esp: only assert INTR_DC interrupt flag if selection fails, Paolo Bonzini, 2021/06/17
- [PULL 07/45] esp: only set ESP_RSEQ at the start of the select sequence, Paolo Bonzini, 2021/06/17
- [PULL 11/45] esp: revert 75ef849696 "esp: correctly fill bus id with requested lun", Paolo Bonzini, 2021/06/17
- [PULL 14/45] esp: store lun coming from the MESSAGE OUT phase, Paolo Bonzini, 2021/06/17
- [PULL 15/45] softmmu/physmem: Mark shared anonymous memory RAM_SHARED, Paolo Bonzini, 2021/06/17
- [PULL 16/45] softmmu/physmem: Fix ram_block_discard_range() to handle shared anonymous memory,
Paolo Bonzini <=
- [PULL 18/45] util/mmap-alloc: Factor out calculation of the pagesize for the guard page, Paolo Bonzini, 2021/06/17
- [PULL 09/45] esp: handle non-DMA transfers from the target one byte at a time, Paolo Bonzini, 2021/06/17
- [PULL 10/45] esp: ensure PDMA write transfers are flushed from the FIFO to the target immediately, Paolo Bonzini, 2021/06/17
- [PULL 13/45] esp: fix migration version check in esp_is_version_5(), Paolo Bonzini, 2021/06/17
- [PULL 12/45] esp: correctly accumulate extended messages for PDMA, Paolo Bonzini, 2021/06/17
- [PULL 17/45] softmmu/physmem: Fix qemu_ram_remap() to handle shared anonymous memory, Paolo Bonzini, 2021/06/17
- [PULL 19/45] util/mmap-alloc: Factor out reserving of a memory region to mmap_reserve(), Paolo Bonzini, 2021/06/17
- [PULL 20/45] util/mmap-alloc: Factor out activating of memory to mmap_activate(), Paolo Bonzini, 2021/06/17
- [PULL 22/45] softmmu/memory: Pass ram_flags to memory_region_init_ram_shared_nomigrate(), Paolo Bonzini, 2021/06/17
- [PULL 26/45] util/mmap-alloc: Support RAM_NORESERVE via MAP_NORESERVE under Linux, Paolo Bonzini, 2021/06/17