/* this is really a sanity check */
static int
test_macros(int __rte_unused unused_parm)
{
#define SMALLER 0x1000U
#define BIGGER 0x2000U
#define PTR_DIFF BIGGER - SMALLER
#define FAIL_MACRO(x)\
{printf(#x "() test failed!\n");\
return -1;}
uintptr_t unused = 0;
RTE_SET_USED(unused);
if ((uintptr_t)RTE_PTR_ADD(SMALLER, PTR_DIFF) != BIGGER)
FAIL_MACRO(RTE_PTR_ADD);
if ((uintptr_t)RTE_PTR_SUB(BIGGER, PTR_DIFF) != SMALLER)
FAIL_MACRO(RTE_PTR_SUB);
if (RTE_PTR_DIFF(BIGGER, SMALLER) != PTR_DIFF)
FAIL_MACRO(RTE_PTR_DIFF);
if (RTE_MAX(SMALLER, BIGGER) != BIGGER)
FAIL_MACRO(RTE_MAX);
if (RTE_MIN(SMALLER, BIGGER) != SMALLER)
FAIL_MACRO(RTE_MIN);
if (strncmp(RTE_STR(test), "test", sizeof("test")))
FAIL_MACRO(RTE_STR);
return 0;
}
static struct rte_memseg *
virt2memseg(const void *addr, const struct rte_memseg_list *msl)
{
const struct rte_fbarray *arr;
void *start, *end;
int ms_idx;
if (msl == NULL)
return NULL;
/* a memseg list was specified, check if it's the right one */
start = msl->base_va;
end = RTE_PTR_ADD(start, (size_t)msl->page_sz * msl->memseg_arr.len);
if (addr < start || addr >= end)
return NULL;
/* now, calculate index */
arr = &msl->memseg_arr;
ms_idx = RTE_PTR_DIFF(addr, msl->base_va) / msl->page_sz;
return rte_fbarray_get(arr, ms_idx);
}
/* fills hugepage cache entry for a given start virt_addr */
static int
get_hugefile_by_virt_addr(uint64_t virt_addr, struct memseg_cache_entry * e)
{
uint64_t start_addr, end_addr;
char *start,*path_end;
char buf[PATH_MAX*2];
FILE *f;
start = NULL;
path_end = NULL;
start_addr = 0;
memset(e->filepath, 0, sizeof(e->filepath));
/* open /proc/self/maps */
f = fopen("/proc/self/maps", "r");
if (f == NULL) {
RTE_LOG(ERR, EAL, "cannot open /proc/self/maps!\n");
return -1;
}
/* parse maps */
while (fgets(buf, sizeof(buf), f) != NULL) {
/* get endptr to end of start addr */
start = buf;
GET_PAGEMAP_ADDR(start,start_addr,'-',
"Cannot find start address in maps!\n");
/* if start address is bigger than our address, skip */
if (start_addr > virt_addr)
continue;
GET_PAGEMAP_ADDR(start,end_addr,' ',
"Cannot find end address in maps!\n");
/* if end address is less than our address, skip */
if (end_addr <= virt_addr)
continue;
/* find where the path starts */
start = strstr(start, "/");
if (start == NULL)
continue;
/* at this point, we know that this is our map.
* now let's find the file */
path_end = strstr(start, "\n");
break;
}
if (path_end == NULL) {
RTE_LOG(ERR, EAL, "Hugefile path not found!\n");
goto error;
}
/* calculate offset and copy the file path */
snprintf(e->filepath, RTE_PTR_DIFF(path_end, start) + 1, "%s", start);
e->offset = virt_addr - start_addr;
fclose(f);
return 0;
error:
fclose(f);
return -1;
}
static int
test_memzone_reserve_max_aligned(void)
{
const struct rte_memzone *mz;
const struct rte_config *config;
const struct rte_memseg *ms;
int memseg_idx = 0;
int memzone_idx = 0;
uintptr_t addr_offset;
size_t len = 0;
void* last_addr;
size_t maxlen = 0;
/* random alignment */
rte_srand((unsigned)rte_rdtsc());
const unsigned align = 1 << ((rte_rand() % 8) + 5); /* from 128 up to 4k alignment */
/* get pointer to global configuration */
config = rte_eal_get_configuration();
ms = rte_eal_get_physmem_layout();
addr_offset = 0;
for (memseg_idx = 0; memseg_idx < RTE_MAX_MEMSEG; memseg_idx++){
/* ignore smaller memsegs as they can only get smaller */
if (ms[memseg_idx].len < maxlen)
continue;
/* align everything */
last_addr = RTE_PTR_ALIGN_CEIL(ms[memseg_idx].addr, RTE_CACHE_LINE_SIZE);
len = ms[memseg_idx].len - RTE_PTR_DIFF(last_addr, ms[memseg_idx].addr);
len &= ~((size_t) RTE_CACHE_LINE_MASK);
/* cycle through all memzones */
for (memzone_idx = 0; memzone_idx < RTE_MAX_MEMZONE; memzone_idx++) {
/* stop when reaching last allocated memzone */
if (config->mem_config->memzone[memzone_idx].addr == NULL)
break;
/* check if the memzone is in our memseg and subtract length */
if ((config->mem_config->memzone[memzone_idx].addr >=
ms[memseg_idx].addr) &&
(config->mem_config->memzone[memzone_idx].addr <
(RTE_PTR_ADD(ms[memseg_idx].addr, ms[memseg_idx].len)))) {
/* since the zones can now be aligned and occasionally skip
* some space, we should calculate the length based on
* reported length and start addresses difference.
*/
len -= (uintptr_t) RTE_PTR_SUB(
config->mem_config->memzone[memzone_idx].addr,
(uintptr_t) last_addr);
len -= config->mem_config->memzone[memzone_idx].len;
last_addr =
RTE_PTR_ADD(config->mem_config->memzone[memzone_idx].addr,
(size_t) config->mem_config->memzone[memzone_idx].len);
}
}
/* make sure we get the alignment offset */
if (len > maxlen) {
addr_offset = RTE_PTR_ALIGN_CEIL((uintptr_t) last_addr, align) - (uintptr_t) last_addr;
maxlen = len;
}
}
if (maxlen == 0 || maxlen == addr_offset) {
printf("There is no space left for biggest %u-aligned memzone!\n", align);
return 0;
}
maxlen -= addr_offset;
mz = rte_memzone_reserve_aligned("max_zone_aligned", 0,
SOCKET_ID_ANY, 0, align);
if (mz == NULL){
printf("Failed to reserve a big chunk of memory\n");
rte_dump_physmem_layout(stdout);
rte_memzone_dump(stdout);
return -1;
}
if (mz->len != maxlen) {
printf("Memzone reserve with 0 size and alignment %u did not return"
" bigest block\n", align);
printf("Expected size = %zu, actual size = %zu\n",
maxlen, mz->len);
rte_dump_physmem_layout(stdout);
rte_memzone_dump(stdout);
return -1;
}
return 0;
}
static int
test_memzone_reserve_max(void)
{
const struct rte_memzone *mz;
const struct rte_config *config;
const struct rte_memseg *ms;
int memseg_idx = 0;
int memzone_idx = 0;
size_t len = 0;
void* last_addr;
size_t maxlen = 0;
/* get pointer to global configuration */
config = rte_eal_get_configuration();
ms = rte_eal_get_physmem_layout();
for (memseg_idx = 0; memseg_idx < RTE_MAX_MEMSEG; memseg_idx++){
/* ignore smaller memsegs as they can only get smaller */
if (ms[memseg_idx].len < maxlen)
continue;
/* align everything */
last_addr = RTE_PTR_ALIGN_CEIL(ms[memseg_idx].addr, RTE_CACHE_LINE_SIZE);
len = ms[memseg_idx].len - RTE_PTR_DIFF(last_addr, ms[memseg_idx].addr);
len &= ~((size_t) RTE_CACHE_LINE_MASK);
/* cycle through all memzones */
for (memzone_idx = 0; memzone_idx < RTE_MAX_MEMZONE; memzone_idx++) {
/* stop when reaching last allocated memzone */
if (config->mem_config->memzone[memzone_idx].addr == NULL)
break;
/* check if the memzone is in our memseg and subtract length */
if ((config->mem_config->memzone[memzone_idx].addr >=
ms[memseg_idx].addr) &&
(config->mem_config->memzone[memzone_idx].addr <
(RTE_PTR_ADD(ms[memseg_idx].addr, ms[memseg_idx].len)))) {
/* since the zones can now be aligned and occasionally skip
* some space, we should calculate the length based on
* reported length and start addresses difference. Addresses
* are allocated sequentially so we don't need to worry about
* them being in the right order.
*/
len -= RTE_PTR_DIFF(
config->mem_config->memzone[memzone_idx].addr,
last_addr);
len -= config->mem_config->memzone[memzone_idx].len;
last_addr = RTE_PTR_ADD(config->mem_config->memzone[memzone_idx].addr,
(size_t) config->mem_config->memzone[memzone_idx].len);
}
}
/* we don't need to calculate offset here since length
* is always cache-aligned */
if (len > maxlen)
maxlen = len;
}
if (maxlen == 0) {
printf("There is no space left!\n");
return 0;
}
mz = rte_memzone_reserve("max_zone", 0, SOCKET_ID_ANY, 0);
if (mz == NULL){
printf("Failed to reserve a big chunk of memory\n");
rte_dump_physmem_layout(stdout);
rte_memzone_dump(stdout);
return -1;
}
if (mz->len != maxlen) {
printf("Memzone reserve with 0 size did not return bigest block\n");
printf("Expected size = %zu, actual size = %zu\n",
maxlen, mz->len);
rte_dump_physmem_layout(stdout);
rte_memzone_dump(stdout);
return -1;
}
return 0;
}
void *
eal_get_virtual_area(void *requested_addr, size_t *size,
size_t page_sz, int flags, int mmap_flags)
{
bool addr_is_hint, allow_shrink, unmap, no_align;
uint64_t map_sz;
void *mapped_addr, *aligned_addr;
if (system_page_sz == 0)
system_page_sz = sysconf(_SC_PAGESIZE);
mmap_flags |= MAP_PRIVATE | MAP_ANONYMOUS;
RTE_LOG(DEBUG, EAL, "Ask a virtual area of 0x%zx bytes\n", *size);
addr_is_hint = (flags & EAL_VIRTUAL_AREA_ADDR_IS_HINT) > 0;
allow_shrink = (flags & EAL_VIRTUAL_AREA_ALLOW_SHRINK) > 0;
unmap = (flags & EAL_VIRTUAL_AREA_UNMAP) > 0;
if (next_baseaddr == NULL && internal_config.base_virtaddr != 0 &&
rte_eal_process_type() == RTE_PROC_PRIMARY)
next_baseaddr = (void *) internal_config.base_virtaddr;
if (requested_addr == NULL && next_baseaddr != NULL) {
requested_addr = next_baseaddr;
requested_addr = RTE_PTR_ALIGN(requested_addr, page_sz);
addr_is_hint = true;
}
/* we don't need alignment of resulting pointer in the following cases:
*
* 1. page size is equal to system size
* 2. we have a requested address, and it is page-aligned, and we will
* be discarding the address if we get a different one.
*
* for all other cases, alignment is potentially necessary.
*/
no_align = (requested_addr != NULL &&
requested_addr == RTE_PTR_ALIGN(requested_addr, page_sz) &&
!addr_is_hint) ||
page_sz == system_page_sz;
do {
map_sz = no_align ? *size : *size + page_sz;
if (map_sz > SIZE_MAX) {
RTE_LOG(ERR, EAL, "Map size too big\n");
rte_errno = E2BIG;
return NULL;
}
mapped_addr = mmap(requested_addr, (size_t)map_sz, PROT_READ,
mmap_flags, -1, 0);
if (mapped_addr == MAP_FAILED && allow_shrink)
*size -= page_sz;
} while (allow_shrink && mapped_addr == MAP_FAILED && *size > 0);
/* align resulting address - if map failed, we will ignore the value
* anyway, so no need to add additional checks.
*/
aligned_addr = no_align ? mapped_addr :
RTE_PTR_ALIGN(mapped_addr, page_sz);
if (*size == 0) {
RTE_LOG(ERR, EAL, "Cannot get a virtual area of any size: %s\n",
strerror(errno));
rte_errno = errno;
return NULL;
} else if (mapped_addr == MAP_FAILED) {
RTE_LOG(ERR, EAL, "Cannot get a virtual area: %s\n",
strerror(errno));
/* pass errno up the call chain */
rte_errno = errno;
return NULL;
} else if (requested_addr != NULL && !addr_is_hint &&
aligned_addr != requested_addr) {
RTE_LOG(ERR, EAL, "Cannot get a virtual area at requested address: %p (got %p)\n",
requested_addr, aligned_addr);
munmap(mapped_addr, map_sz);
rte_errno = EADDRNOTAVAIL;
return NULL;
} else if (requested_addr != NULL && addr_is_hint &&
aligned_addr != requested_addr) {
RTE_LOG(WARNING, EAL, "WARNING! Base virtual address hint (%p != %p) not respected!\n",
requested_addr, aligned_addr);
RTE_LOG(WARNING, EAL, " This may cause issues with mapping memory into secondary processes\n");
} else if (next_baseaddr != NULL) {
next_baseaddr = RTE_PTR_ADD(aligned_addr, *size);
}
RTE_LOG(DEBUG, EAL, "Virtual area found at %p (size = 0x%zx)\n",
aligned_addr, *size);
if (unmap) {
munmap(mapped_addr, map_sz);
} else if (!no_align) {
void *map_end, *aligned_end;
size_t before_len, after_len;
/* when we reserve space with alignment, we add alignment to
* mapping size. On 32-bit, if 1GB alignment was requested, this
* would waste 1GB of address space, which is a luxury we cannot
* afford. so, if alignment was performed, check if any unneeded
* address space can be unmapped back.
*/
map_end = RTE_PTR_ADD(mapped_addr, (size_t)map_sz);
aligned_end = RTE_PTR_ADD(aligned_addr, *size);
/* unmap space before aligned mmap address */
before_len = RTE_PTR_DIFF(aligned_addr, mapped_addr);
if (before_len > 0)
munmap(mapped_addr, before_len);
/* unmap space after aligned end mmap address */
after_len = RTE_PTR_DIFF(map_end, aligned_end);
if (after_len > 0)
munmap(aligned_end, after_len);
}
return aligned_addr;
}
/*
* Get physical address of any mapped virtual address in the current process.
*/
phys_addr_t
rte_mem_virt2phy(const void *virtaddr)
{
int fd, retval;
uint64_t page, physaddr;
unsigned long virt_pfn;
int page_size;
off_t offset;
/* when using dom0, /proc/self/pagemap always returns 0, check in
* dpdk memory by browsing the memsegs */
if (rte_xen_dom0_supported()) {
struct rte_mem_config *mcfg;
struct rte_memseg *memseg;
unsigned i;
mcfg = rte_eal_get_configuration()->mem_config;
for (i = 0; i < RTE_MAX_MEMSEG; i++) {
memseg = &mcfg->memseg[i];
if (memseg->addr == NULL)
break;
if (virtaddr > memseg->addr &&
virtaddr < RTE_PTR_ADD(memseg->addr,
memseg->len)) {
return memseg->phys_addr +
RTE_PTR_DIFF(virtaddr, memseg->addr);
}
}
return RTE_BAD_PHYS_ADDR;
}
/* Cannot parse /proc/self/pagemap, no need to log errors everywhere */
if (!proc_pagemap_readable)
return RTE_BAD_PHYS_ADDR;
/* standard page size */
page_size = getpagesize();
fd = open("/proc/self/pagemap", O_RDONLY);
if (fd < 0) {
RTE_LOG(ERR, EAL, "%s(): cannot open /proc/self/pagemap: %s\n",
__func__, strerror(errno));
return RTE_BAD_PHYS_ADDR;
}
virt_pfn = (unsigned long)virtaddr / page_size;
offset = sizeof(uint64_t) * virt_pfn;
if (lseek(fd, offset, SEEK_SET) == (off_t) -1) {
RTE_LOG(ERR, EAL, "%s(): seek error in /proc/self/pagemap: %s\n",
__func__, strerror(errno));
close(fd);
return RTE_BAD_PHYS_ADDR;
}
retval = read(fd, &page, PFN_MASK_SIZE);
close(fd);
if (retval < 0) {
RTE_LOG(ERR, EAL, "%s(): cannot read /proc/self/pagemap: %s\n",
__func__, strerror(errno));
return RTE_BAD_PHYS_ADDR;
} else if (retval != PFN_MASK_SIZE) {
RTE_LOG(ERR, EAL, "%s(): read %d bytes from /proc/self/pagemap "
"but expected %d:\n",
__func__, retval, PFN_MASK_SIZE);
return RTE_BAD_PHYS_ADDR;
}
/*
* the pfn (page frame number) are bits 0-54 (see
* pagemap.txt in linux Documentation)
*/
physaddr = ((page & 0x7fffffffffffffULL) * page_size)
+ ((unsigned long)virtaddr % page_size);
return physaddr;
}
