[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[PULL 22/68] libvhost-user: Use most of mmap_offset as fd_offset
From: |
Michael S. Tsirkin |
Subject: |
[PULL 22/68] libvhost-user: Use most of mmap_offset as fd_offset |
Date: |
Tue, 12 Mar 2024 18:26:37 -0400 |
From: David Hildenbrand <david@redhat.com>
In the past, QEMU would create memory regions that could partially cover
hugetlb pages, making mmap() fail if we would use the mmap_offset as an
fd_offset. For that reason, we never used the mmap_offset as an offset into
the fd and instead always mapped the fd from the very start.
However, that can easily result in us mmap'ing a lot of unnecessary
parts of an fd, possibly repeatedly.
QEMU nowadays does not create memory regions that partially cover huge
pages -- it never really worked with postcopy. QEMU handles merging of
regions that partially cover huge pages (due to holes in boot memory) since
2018 in c1ece84e7c93 ("vhost: Huge page align and merge").
Let's be a bit careful and not unconditionally convert the
mmap_offset into an fd_offset. Instead, let's simply detect the hugetlb
size and pass as much as we can as fd_offset, making sure that we call
mmap() with a properly aligned offset.
With QEMU and a virtio-mem device that is fully plugged (50GiB using 50
memslots) the qemu-storage daemon process consumes in the VA space
1281GiB before this change and 58GiB after this change.
================ Vhost user message ================
Request: VHOST_USER_ADD_MEM_REG (37)
Flags: 0x9
Size: 40
Fds: 59
Adding region 4
guest_phys_addr: 0x0000000200000000
memory_size: 0x0000000040000000
userspace_addr: 0x00007fb73bffe000
old mmap_offset: 0x0000000080000000
fd_offset: 0x0000000080000000
new mmap_offset: 0x0000000000000000
mmap_addr: 0x00007f02f1bdc000
Successfully added new region
================ Vhost user message ================
Request: VHOST_USER_ADD_MEM_REG (37)
Flags: 0x9
Size: 40
Fds: 59
Adding region 5
guest_phys_addr: 0x0000000240000000
memory_size: 0x0000000040000000
userspace_addr: 0x00007fb77bffe000
old mmap_offset: 0x00000000c0000000
fd_offset: 0x00000000c0000000
new mmap_offset: 0x0000000000000000
mmap_addr: 0x00007f0284000000
Successfully added new region
Reviewed-by: Raphael Norwitz <raphael@enfabrica.net>
Acked-by: Stefano Garzarella <sgarzare@redhat.com>
Signed-off-by: David Hildenbrand <david@redhat.com>
Message-Id: <20240214151701.29906-12-david@redhat.com>
Tested-by: Mario Casquero <mcasquer@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
---
subprojects/libvhost-user/libvhost-user.c | 54 ++++++++++++++++++++---
1 file changed, 48 insertions(+), 6 deletions(-)
diff --git a/subprojects/libvhost-user/libvhost-user.c
b/subprojects/libvhost-user/libvhost-user.c
index ef6353d847..55aef5fcc6 100644
--- a/subprojects/libvhost-user/libvhost-user.c
+++ b/subprojects/libvhost-user/libvhost-user.c
@@ -43,6 +43,8 @@
#include <fcntl.h>
#include <sys/ioctl.h>
#include <linux/vhost.h>
+#include <sys/vfs.h>
+#include <linux/magic.h>
#ifdef __NR_userfaultfd
#include <linux/userfaultfd.h>
@@ -281,12 +283,32 @@ vu_remove_all_mem_regs(VuDev *dev)
dev->nregions = 0;
}
+static size_t
+get_fd_hugepagesize(int fd)
+{
+#if defined(__linux__)
+ struct statfs fs;
+ int ret;
+
+ do {
+ ret = fstatfs(fd, &fs);
+ } while (ret != 0 && errno == EINTR);
+
+ if (!ret && (unsigned int)fs.f_type == HUGETLBFS_MAGIC) {
+ return fs.f_bsize;
+ }
+#endif
+ return 0;
+}
+
static void
_vu_add_mem_reg(VuDev *dev, VhostUserMemoryRegion *msg_region, int fd)
{
const uint64_t start_gpa = msg_region->guest_phys_addr;
const uint64_t end_gpa = start_gpa + msg_region->memory_size;
int prot = PROT_READ | PROT_WRITE;
+ uint64_t mmap_offset, fd_offset;
+ size_t hugepagesize;
VuDevRegion *r;
void *mmap_addr;
int low = 0;
@@ -300,7 +322,7 @@ _vu_add_mem_reg(VuDev *dev, VhostUserMemoryRegion
*msg_region, int fd)
msg_region->memory_size);
DPRINT(" userspace_addr: 0x%016"PRIx64"\n",
msg_region->userspace_addr);
- DPRINT(" mmap_offset: 0x%016"PRIx64"\n",
+ DPRINT(" old mmap_offset: 0x%016"PRIx64"\n",
msg_region->mmap_offset);
if (dev->postcopy_listening) {
@@ -335,11 +357,31 @@ _vu_add_mem_reg(VuDev *dev, VhostUserMemoryRegion
*msg_region, int fd)
idx = low;
/*
- * We don't use offset argument of mmap() since the mapped address has
- * to be page aligned, and we use huge pages.
+ * Convert most of msg_region->mmap_offset to fd_offset. In almost all
+ * cases, this will leave us with mmap_offset == 0, mmap()'ing only
+ * what we really need. Only if a memory region would partially cover
+ * hugetlb pages, we'd get mmap_offset != 0, which usually doesn't happen
+ * anymore (i.e., modern QEMU).
+ *
+ * Note that mmap() with hugetlb would fail if the offset into the file
+ * is not aligned to the huge page size.
*/
- mmap_addr = mmap(0, msg_region->memory_size + msg_region->mmap_offset,
- prot, MAP_SHARED | MAP_NORESERVE, fd, 0);
+ hugepagesize = get_fd_hugepagesize(fd);
+ if (hugepagesize) {
+ fd_offset = ALIGN_DOWN(msg_region->mmap_offset, hugepagesize);
+ mmap_offset = msg_region->mmap_offset - fd_offset;
+ } else {
+ fd_offset = msg_region->mmap_offset;
+ mmap_offset = 0;
+ }
+
+ DPRINT(" fd_offset: 0x%016"PRIx64"\n",
+ fd_offset);
+ DPRINT(" new mmap_offset: 0x%016"PRIx64"\n",
+ mmap_offset);
+
+ mmap_addr = mmap(0, msg_region->memory_size + mmap_offset,
+ prot, MAP_SHARED | MAP_NORESERVE, fd, fd_offset);
if (mmap_addr == MAP_FAILED) {
vu_panic(dev, "region mmap error: %s", strerror(errno));
return;
@@ -354,7 +396,7 @@ _vu_add_mem_reg(VuDev *dev, VhostUserMemoryRegion
*msg_region, int fd)
r->size = msg_region->memory_size;
r->qva = msg_region->userspace_addr;
r->mmap_addr = (uint64_t)(uintptr_t)mmap_addr;
- r->mmap_offset = msg_region->mmap_offset;
+ r->mmap_offset = mmap_offset;
dev->nregions++;
if (dev->postcopy_listening) {
--
MST
- [PULL 13/68] libvhost-user: Bump up VHOST_USER_MAX_RAM_SLOTS to 509, (continued)
- [PULL 13/68] libvhost-user: Bump up VHOST_USER_MAX_RAM_SLOTS to 509, Michael S. Tsirkin, 2024/03/12
- [PULL 14/68] libvhost-user: Factor out removing all mem regions, Michael S. Tsirkin, 2024/03/12
- [PULL 15/68] libvhost-user: Merge vu_set_mem_table_exec_postcopy() into vu_set_mem_table_exec(), Michael S. Tsirkin, 2024/03/12
- [PULL 16/68] libvhost-user: Factor out adding a memory region, Michael S. Tsirkin, 2024/03/12
- [PULL 17/68] libvhost-user: No need to check for NULL when unmapping, Michael S. Tsirkin, 2024/03/12
- [PULL 18/68] libvhost-user: Don't zero out memory for memory regions, Michael S. Tsirkin, 2024/03/12
- [PULL 19/68] libvhost-user: Don't search for duplicates when removing memory regions, Michael S. Tsirkin, 2024/03/12
- [PULL 23/68] libvhost-user: Factor out vq usability check, Michael S. Tsirkin, 2024/03/12
- [PULL 26/68] pcie: Support PCIe Gen5/Gen6 link speeds, Michael S. Tsirkin, 2024/03/12
- [PULL 21/68] libvhost-user: Speedup gpa_to_mem_region() and vu_gpa_to_va(), Michael S. Tsirkin, 2024/03/12
- [PULL 22/68] libvhost-user: Use most of mmap_offset as fd_offset,
Michael S. Tsirkin <=
- [PULL 27/68] vdpa: stash memory region properties in vars, Michael S. Tsirkin, 2024/03/12
- [PULL 20/68] libvhost-user: Factor out search for memory region by GPA and simplify, Michael S. Tsirkin, 2024/03/12
- [PULL 24/68] libvhost-user: Dynamically remap rings after (temporarily?) removing memory regions, Michael S. Tsirkin, 2024/03/12
- [PULL 28/68] vdpa: trace skipped memory sections, Michael S. Tsirkin, 2024/03/12
- [PULL 29/68] hw/pci-bridge/pxb-cxl: Drop RAS capability from host bridge., Michael S. Tsirkin, 2024/03/12
- [PULL 30/68] hw/audio/virtio-sound: return correct command response size, Michael S. Tsirkin, 2024/03/12
- [PULL 31/68] hw/virtio: check owner for removing objects, Michael S. Tsirkin, 2024/03/12
- [PULL 33/68] hw/cxl/cxl-host: Fix missing ERRP_GUARD() in cxl_fixed_memory_window_config(), Michael S. Tsirkin, 2024/03/12
- [PULL 25/68] libvhost-user: Mark mmap'ed region memory as MADV_DONTDUMP, Michael S. Tsirkin, 2024/03/12