qemu-kvm/SOURCES/kvm-vfio-container-Move-pgsizes-and-dma_max_mappings-to-.patch

243 lines
10 KiB
Diff

From d798939fbbe6c27200c165edd6f3771413821b34 Mon Sep 17 00:00:00 2001
From: Eric Auger <eric.auger@redhat.com>
Date: Thu, 2 Nov 2023 15:12:36 +0800
Subject: [PATCH 011/101] vfio/container: Move pgsizes and dma_max_mappings to
base container
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: 211: IOMMUFD backend backport
RH-Jira: RHEL-19302 RHEL-21057
RH-Acked-by: Cédric Le Goater <clg@redhat.com>
RH-Acked-by: Sebastian Ott <sebott@redhat.com>
RH-Commit: [10/67] e80696175aba159a17ce9a869535db66682deb08 (eauger1/centos-qemu-kvm)
No functional change intended.
Signed-off-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Yi Liu <yi.l.liu@intel.com>
Signed-off-by: Yi Sun <yi.y.sun@linux.intel.com>
Signed-off-by: Zhenzhong Duan <zhenzhong.duan@intel.com>
Reviewed-by: Cédric Le Goater <clg@redhat.com>
Signed-off-by: Cédric Le Goater <clg@redhat.com>
(cherry picked from commit 7ab1cb74ffdbf92ef237243b41bde5c7067d5298)
Signed-off-by: Eric Auger <eric.auger@redhat.com>
---
hw/vfio/common.c | 17 +++++++++--------
hw/vfio/container-base.c | 1 +
hw/vfio/container.c | 11 +++++------
hw/vfio/spapr.c | 10 ++++++----
include/hw/vfio/vfio-common.h | 2 --
include/hw/vfio/vfio-container-base.h | 2 ++
6 files changed, 23 insertions(+), 20 deletions(-)
diff --git a/hw/vfio/common.c b/hw/vfio/common.c
index cf6618f6ed..1cb53d369e 100644
--- a/hw/vfio/common.c
+++ b/hw/vfio/common.c
@@ -401,6 +401,7 @@ static int vfio_ram_discard_notify_populate(RamDiscardListener *rdl,
static void vfio_register_ram_discard_listener(VFIOContainer *container,
MemoryRegionSection *section)
{
+ VFIOContainerBase *bcontainer = &container->bcontainer;
RamDiscardManager *rdm = memory_region_get_ram_discard_manager(section->mr);
VFIORamDiscardListener *vrdl;
@@ -419,8 +420,8 @@ static void vfio_register_ram_discard_listener(VFIOContainer *container,
section->mr);
g_assert(vrdl->granularity && is_power_of_2(vrdl->granularity));
- g_assert(container->pgsizes &&
- vrdl->granularity >= 1ULL << ctz64(container->pgsizes));
+ g_assert(bcontainer->pgsizes &&
+ vrdl->granularity >= 1ULL << ctz64(bcontainer->pgsizes));
ram_discard_listener_init(&vrdl->listener,
vfio_ram_discard_notify_populate,
@@ -441,7 +442,7 @@ static void vfio_register_ram_discard_listener(VFIOContainer *container,
* number of sections in the address space we could have over time,
* also consuming DMA mappings.
*/
- if (container->dma_max_mappings) {
+ if (bcontainer->dma_max_mappings) {
unsigned int vrdl_count = 0, vrdl_mappings = 0, max_memslots = 512;
#ifdef CONFIG_KVM
@@ -462,11 +463,11 @@ static void vfio_register_ram_discard_listener(VFIOContainer *container,
}
if (vrdl_mappings + max_memslots - vrdl_count >
- container->dma_max_mappings) {
+ bcontainer->dma_max_mappings) {
warn_report("%s: possibly running out of DMA mappings. E.g., try"
" increasing the 'block-size' of virtio-mem devies."
" Maximum possible DMA mappings: %d, Maximum possible"
- " memslots: %d", __func__, container->dma_max_mappings,
+ " memslots: %d", __func__, bcontainer->dma_max_mappings,
max_memslots);
}
}
@@ -626,7 +627,7 @@ static void vfio_listener_region_add(MemoryListener *listener,
iommu_idx);
ret = memory_region_iommu_set_page_size_mask(giommu->iommu_mr,
- container->pgsizes,
+ bcontainer->pgsizes,
&err);
if (ret) {
g_free(giommu);
@@ -675,7 +676,7 @@ static void vfio_listener_region_add(MemoryListener *listener,
llsize = int128_sub(llend, int128_make64(iova));
if (memory_region_is_ram_device(section->mr)) {
- hwaddr pgmask = (1ULL << ctz64(container->pgsizes)) - 1;
+ hwaddr pgmask = (1ULL << ctz64(bcontainer->pgsizes)) - 1;
if ((iova & pgmask) || (int128_get64(llsize) & pgmask)) {
trace_vfio_listener_region_add_no_dma_map(
@@ -777,7 +778,7 @@ static void vfio_listener_region_del(MemoryListener *listener,
if (memory_region_is_ram_device(section->mr)) {
hwaddr pgmask;
- pgmask = (1ULL << ctz64(container->pgsizes)) - 1;
+ pgmask = (1ULL << ctz64(bcontainer->pgsizes)) - 1;
try_unmap = !((iova & pgmask) || (int128_get64(llsize) & pgmask));
} else if (memory_region_has_ram_discard_manager(section->mr)) {
vfio_unregister_ram_discard_listener(container, section);
diff --git a/hw/vfio/container-base.c b/hw/vfio/container-base.c
index 5d654ae172..dcce111349 100644
--- a/hw/vfio/container-base.c
+++ b/hw/vfio/container-base.c
@@ -52,6 +52,7 @@ void vfio_container_init(VFIOContainerBase *bcontainer, VFIOAddressSpace *space,
bcontainer->ops = ops;
bcontainer->space = space;
bcontainer->dirty_pages_supported = false;
+ bcontainer->dma_max_mappings = 0;
QLIST_INIT(&bcontainer->giommu_list);
}
diff --git a/hw/vfio/container.c b/hw/vfio/container.c
index 7bd81eab09..c5a6262882 100644
--- a/hw/vfio/container.c
+++ b/hw/vfio/container.c
@@ -154,7 +154,7 @@ static int vfio_legacy_dma_unmap(VFIOContainerBase *bcontainer, hwaddr iova,
if (errno == EINVAL && unmap.size && !(unmap.iova + unmap.size) &&
container->iommu_type == VFIO_TYPE1v2_IOMMU) {
trace_vfio_legacy_dma_unmap_overflow_workaround();
- unmap.size -= 1ULL << ctz64(container->pgsizes);
+ unmap.size -= 1ULL << ctz64(bcontainer->pgsizes);
continue;
}
error_report("VFIO_UNMAP_DMA failed: %s", strerror(errno));
@@ -559,7 +559,6 @@ static int vfio_connect_container(VFIOGroup *group, AddressSpace *as,
container = g_malloc0(sizeof(*container));
container->fd = fd;
container->error = NULL;
- container->dma_max_mappings = 0;
container->iova_ranges = NULL;
QLIST_INIT(&container->vrdl_list);
bcontainer = &container->bcontainer;
@@ -589,13 +588,13 @@ static int vfio_connect_container(VFIOGroup *group, AddressSpace *as,
}
if (info->flags & VFIO_IOMMU_INFO_PGSIZES) {
- container->pgsizes = info->iova_pgsizes;
+ bcontainer->pgsizes = info->iova_pgsizes;
} else {
- container->pgsizes = qemu_real_host_page_size();
+ bcontainer->pgsizes = qemu_real_host_page_size();
}
- if (!vfio_get_info_dma_avail(info, &container->dma_max_mappings)) {
- container->dma_max_mappings = 65535;
+ if (!vfio_get_info_dma_avail(info, &bcontainer->dma_max_mappings)) {
+ bcontainer->dma_max_mappings = 65535;
}
vfio_get_info_iova_range(info, container);
diff --git a/hw/vfio/spapr.c b/hw/vfio/spapr.c
index 83da2f7ec2..4f76bdd3ca 100644
--- a/hw/vfio/spapr.c
+++ b/hw/vfio/spapr.c
@@ -226,6 +226,7 @@ static int vfio_spapr_create_window(VFIOContainer *container,
hwaddr *pgsize)
{
int ret = 0;
+ VFIOContainerBase *bcontainer = &container->bcontainer;
IOMMUMemoryRegion *iommu_mr = IOMMU_MEMORY_REGION(section->mr);
uint64_t pagesize = memory_region_iommu_get_min_page_size(iommu_mr), pgmask;
unsigned entries, bits_total, bits_per_level, max_levels;
@@ -239,13 +240,13 @@ static int vfio_spapr_create_window(VFIOContainer *container,
if (pagesize > rampagesize) {
pagesize = rampagesize;
}
- pgmask = container->pgsizes & (pagesize | (pagesize - 1));
+ pgmask = bcontainer->pgsizes & (pagesize | (pagesize - 1));
pagesize = pgmask ? (1ULL << (63 - clz64(pgmask))) : 0;
if (!pagesize) {
error_report("Host doesn't support page size 0x%"PRIx64
", the supported mask is 0x%lx",
memory_region_iommu_get_min_page_size(iommu_mr),
- container->pgsizes);
+ bcontainer->pgsizes);
return -EINVAL;
}
@@ -421,6 +422,7 @@ void vfio_container_del_section_window(VFIOContainer *container,
int vfio_spapr_container_init(VFIOContainer *container, Error **errp)
{
+ VFIOContainerBase *bcontainer = &container->bcontainer;
struct vfio_iommu_spapr_tce_info info;
bool v2 = container->iommu_type == VFIO_SPAPR_TCE_v2_IOMMU;
int ret, fd = container->fd;
@@ -461,7 +463,7 @@ int vfio_spapr_container_init(VFIOContainer *container, Error **errp)
}
if (v2) {
- container->pgsizes = info.ddw.pgsizes;
+ bcontainer->pgsizes = info.ddw.pgsizes;
/*
* There is a default window in just created container.
* To make region_add/del simpler, we better remove this
@@ -476,7 +478,7 @@ int vfio_spapr_container_init(VFIOContainer *container, Error **errp)
}
} else {
/* The default table uses 4K pages */
- container->pgsizes = 0x1000;
+ bcontainer->pgsizes = 0x1000;
vfio_host_win_add(container, info.dma32_window_start,
info.dma32_window_start +
info.dma32_window_size - 1,
diff --git a/include/hw/vfio/vfio-common.h b/include/hw/vfio/vfio-common.h
index bc67e1316c..d3dc2f9dcb 100644
--- a/include/hw/vfio/vfio-common.h
+++ b/include/hw/vfio/vfio-common.h
@@ -85,8 +85,6 @@ typedef struct VFIOContainer {
bool initialized;
uint64_t dirty_pgsizes;
uint64_t max_dirty_bitmap_size;
- unsigned long pgsizes;
- unsigned int dma_max_mappings;
QLIST_HEAD(, VFIOHostDMAWindow) hostwin_list;
QLIST_HEAD(, VFIOGroup) group_list;
QLIST_HEAD(, VFIORamDiscardListener) vrdl_list;
diff --git a/include/hw/vfio/vfio-container-base.h b/include/hw/vfio/vfio-container-base.h
index 7090962496..85ec7e1a56 100644
--- a/include/hw/vfio/vfio-container-base.h
+++ b/include/hw/vfio/vfio-container-base.h
@@ -36,6 +36,8 @@ typedef struct VFIOAddressSpace {
typedef struct VFIOContainerBase {
const VFIOIOMMUOps *ops;
VFIOAddressSpace *space;
+ unsigned long pgsizes;
+ unsigned int dma_max_mappings;
bool dirty_pages_supported;
QLIST_HEAD(, VFIOGuestIOMMU) giommu_list;
QLIST_ENTRY(VFIOContainerBase) next;
--
2.39.3