Adjust PCI infrastructure to reserve device memory through the
new device memory API. Any hotplug event will reserve memory, any
hot-unplug event will release memory back to the system.
This allows for more reliable PCI mappings in secondary processes,
and will be crucial to support multiprocess hotplug.
Signed-off-by: Anatoly Burakov <anatoly.bura...@intel.com>
---
drivers/bus/pci/linux/pci_init.h | 1 -
drivers/bus/pci/linux/pci_uio.c | 11 +----------
drivers/bus/pci/linux/pci_vfio.c | 27 ++++++++++++---------------
lib/librte_pci/Makefile | 1 +
lib/librte_pci/rte_pci.c | 20 +++++++++++++++++++-
5 files changed, 33 insertions(+), 27 deletions(-)
diff --git a/drivers/bus/pci/linux/pci_init.h b/drivers/bus/pci/linux/pci_init.h
index c2e603a37..bc9279c66 100644
--- a/drivers/bus/pci/linux/pci_init.h
+++ b/drivers/bus/pci/linux/pci_init.h
@@ -14,7 +14,6 @@
/*
* Helper function to map PCI resources right after hugepages in virtual memory
*/
-extern void *pci_map_addr;
void *pci_find_max_end_va(void);
/* parse one line of the "resource" sysfs file (note that the 'line'
diff --git a/drivers/bus/pci/linux/pci_uio.c b/drivers/bus/pci/linux/pci_uio.c
index d423e4bb0..dbf108b6f 100644
--- a/drivers/bus/pci/linux/pci_uio.c
+++ b/drivers/bus/pci/linux/pci_uio.c
@@ -26,8 +26,6 @@
#include "eal_filesystem.h"
#include "pci_init.h"
-void *pci_map_addr = NULL;
-
#define OFF_MAX ((uint64_t)(off_t)-1)
int
@@ -316,19 +314,12 @@ pci_uio_map_resource_by_index(struct rte_pci_device *dev,
int res_idx,
goto error;
}
- /* try mapping somewhere close to the end of hugepages */
- if (pci_map_addr == NULL)
- pci_map_addr = pci_find_max_end_va();
-
- mapaddr = pci_map_resource(pci_map_addr, fd, 0,
+ mapaddr = pci_map_resource(NULL, fd, 0,
(size_t)dev->mem_resource[res_idx].len, 0);
close(fd);
if (mapaddr == MAP_FAILED)
goto error;
- pci_map_addr = RTE_PTR_ADD(mapaddr,
- (size_t)dev->mem_resource[res_idx].len);
-
maps[map_idx].phaddr = dev->mem_resource[res_idx].phys_addr;
maps[map_idx].size = dev->mem_resource[res_idx].len;
maps[map_idx].addr = mapaddr;
diff --git a/drivers/bus/pci/linux/pci_vfio.c b/drivers/bus/pci/linux/pci_vfio.c
index aeeaa9ed8..f390ea37a 100644
--- a/drivers/bus/pci/linux/pci_vfio.c
+++ b/drivers/bus/pci/linux/pci_vfio.c
@@ -324,7 +324,7 @@ pci_rte_vfio_setup_device(struct rte_pci_device *dev, int
vfio_dev_fd)
static int
pci_vfio_mmap_bar(int vfio_dev_fd, struct mapped_pci_resource *vfio_res,
- int bar_index, int additional_flags)
+ int bar_index)
{
struct memreg {
unsigned long offset, size;
@@ -371,9 +371,14 @@ pci_vfio_mmap_bar(int vfio_dev_fd, struct
mapped_pci_resource *vfio_res,
memreg[0].size = bar->size;
}
- /* reserve the address using an inaccessible mapping */
- bar_addr = mmap(bar->addr, bar->size, 0, MAP_PRIVATE |
- MAP_ANONYMOUS | additional_flags, -1, 0);
+ if (bar->addr == NULL) {
+ bar_addr = rte_mem_dev_memory_alloc(bar->size, 0);
+ if (bar_addr == NULL) {
+ RTE_LOG(ERR, EAL, "%s(): cannot reserve space for
device\n",
+ __func__);
+ return -1;
+ }
+ }
if (bar_addr != MAP_FAILED) {
void *map_addr = NULL;
if (memreg[0].size) {
@@ -469,7 +474,6 @@ pci_vfio_map_resource_primary(struct rte_pci_device *dev)
for (i = 0; i < (int) vfio_res->nb_maps; i++) {
struct vfio_region_info reg = { .argsz = sizeof(reg) };
- void *bar_addr;
reg.index = i;
@@ -494,19 +498,12 @@ pci_vfio_map_resource_primary(struct rte_pci_device *dev)
if ((reg.flags & VFIO_REGION_INFO_FLAG_MMAP) == 0)
continue;
- /* try mapping somewhere close to the end of hugepages */
- if (pci_map_addr == NULL)
- pci_map_addr = pci_find_max_end_va();
-
- bar_addr = pci_map_addr;
- pci_map_addr = RTE_PTR_ADD(bar_addr, (size_t) reg.size);
-
- maps[i].addr = bar_addr;
+ maps[i].addr = NULL;
maps[i].offset = reg.offset;
maps[i].size = reg.size;
maps[i].path = NULL; /* vfio doesn't have per-resource paths */
- ret = pci_vfio_mmap_bar(vfio_dev_fd, vfio_res, i, 0);
+ ret = pci_vfio_mmap_bar(vfio_dev_fd, vfio_res, i);
if (ret < 0) {
RTE_LOG(ERR, EAL, " %s mapping BAR%i failed: %s\n",
pci_addr, i, strerror(errno));
@@ -574,7 +571,7 @@ pci_vfio_map_resource_secondary(struct rte_pci_device *dev)
maps = vfio_res->maps;
for (i = 0; i < (int) vfio_res->nb_maps; i++) {
- ret = pci_vfio_mmap_bar(vfio_dev_fd, vfio_res, i, MAP_FIXED);
+ ret = pci_vfio_mmap_bar(vfio_dev_fd, vfio_res, i);
if (ret < 0) {
RTE_LOG(ERR, EAL, " %s mapping BAR%i failed: %s\n",
pci_addr, i, strerror(errno));
diff --git a/lib/librte_pci/Makefile b/lib/librte_pci/Makefile
index 94a632670..f996fe33c 100644
--- a/lib/librte_pci/Makefile
+++ b/lib/librte_pci/Makefile
@@ -8,6 +8,7 @@ LIB = librte_pci.a
CFLAGS := -I$(SRCDIR) $(CFLAGS)
CFLAGS += $(WERROR_FLAGS) -O3
+CFLAGS += -DALLOW_EXPERIMENTAL_API
LDLIBS += -lrte_eal
EXPORT_MAP := rte_pci_version.map
diff --git a/lib/librte_pci/rte_pci.c b/lib/librte_pci/rte_pci.c
index 530738dbd..c425a624e 100644
--- a/lib/librte_pci/rte_pci.c
+++ b/lib/librte_pci/rte_pci.c
@@ -151,6 +151,16 @@ pci_map_resource(void *requested_addr, int fd, off_t
offset, size_t size,
{
void *mapaddr;
+ if (requested_addr == NULL) {
+ requested_addr = rte_mem_dev_memory_alloc(size, 0);
+ if (requested_addr == NULL) {
+ RTE_LOG(ERR, EAL, "%s(): cannot reserve space for
device\n",
+ __func__);
+ return MAP_FAILED;
+ }
+ }
+ additional_flags |= MAP_FIXED;
+
/* Map the PCI memory resource of device */
mapaddr = mmap(requested_addr, size, PROT_READ | PROT_WRITE,
MAP_SHARED | additional_flags, fd, offset);
@@ -170,15 +180,23 @@ pci_map_resource(void *requested_addr, int fd, off_t
offset, size_t size,
void
pci_unmap_resource(void *requested_addr, size_t size)
{
+ void *mapped;
if (requested_addr == NULL)
return;
+ mapped = mmap(requested_addr, size, PROT_READ,
+ MAP_PRIVATE | MAP_ANONYMOUS, -1, 0);
+
/* Unmap the PCI memory resource of device */
- if (munmap(requested_addr, size)) {
+ if (mapped == MAP_FAILED) {
RTE_LOG(ERR, EAL, "%s(): cannot munmap(%p, %#zx): %s\n",
__func__, requested_addr, size,
strerror(errno));
} else
RTE_LOG(DEBUG, EAL, " PCI memory unmapped at %p\n",
requested_addr);
+ if (rte_mem_dev_memory_free(requested_addr, size))
+ RTE_LOG(ERR, EAL, "%s(): cannot mark %p-%p as free\n",
+ __func__, requested_addr,
+ RTE_PTR_ADD(requested_addr, size));
}
--
2.17.0
