2
0
mirror of https://github.com/openvswitch/ovs synced 2025-08-22 18:07:40 +00:00
ovs/lib/util.c

2530 lines
62 KiB
C
Raw Normal View History

/*
* Copyright (c) 2008, 2009, 2010, 2011, 2012, 2013, 2014, 2015, 2016 Nicira, Inc.
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at:
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
#include <config.h>
#include "util.h"
#include <ctype.h>
#include <errno.h>
#include <limits.h>
#include <pthread.h>
#include <stdarg.h>
#include <stdint.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#ifdef __linux__
#include <sys/prctl.h>
#include <sys/utsname.h>
#endif
#include <sys/stat.h>
#include <unistd.h>
#include "bitmap.h"
#include "byte-order.h"
#include "coverage.h"
#include "ovs-rcu.h"
#include "ovs-thread.h"
#include "socket-util.h"
#include "timeval.h"
#include "openvswitch/vlog.h"
#ifdef HAVE_PTHREAD_SET_NAME_NP
#include <pthread_np.h>
#endif
#ifdef _WIN32
#include <shlwapi.h>
#endif
VLOG_DEFINE_THIS_MODULE(util);
#ifdef __linux__
#define LINUX 1
#include <asm/param.h>
#else
#define LINUX 0
#endif
COVERAGE_DEFINE(util_xalloc);
/* argv[0] without directory names. */
char *program_name;
/* Name for the currently running thread or process, for log messages, process
* listings, and debuggers. */
DEFINE_PER_THREAD_MALLOCED_DATA(char *, subprogram_name);
/* --version option output. */
static char *program_version;
vswitchd: Only lock pages that are faulted in. The main purpose of locking the memory is to ensure that OVS can keep doing what it did before in case of increased memory pressure, e.g., during VM ingest / migration. Fulfilling this requirement can be achieved without locking all the allocated memory, but only the pages already accessed in the past (faulted in). Processing of the new traffic involves new memory allocations. Latency on these operations can't be guaranteed by the locking. The main difference would be the pre-faulting of the stack memory. However, in order to revalidate or process upcalls on the same traffic, the same amount of stack is likely needed, so all the necessary memory will already be faulted in. Switch 'mlockall' to MCL_ONFAULT to avoid consuming unnecessarily large amounts of RAM on systems with high core counts. For example, in a densely populated OVN cluster this saves about 650 MB of RAM per node on a system with 64 cores. This equates to 320 GB of allocated but unused RAM in a 500 node cluster. This also makes OVS better suited by default for small systems with limited amount of memory. The MCL_ONFAULT flag was introduced in Linux kernel 4.4 and wasn't available at the time of '--mlockall' introduction, but we can use it now. Falling back to an old way of locking in case we're running on an older kernel just in case. Only locking the faulted in pages also makes locking compatible with vhost post-copy live migration by default, because we'll no longer pre-fault all the guest's memory. Post-copy relies on userfaultfd to work on shared huge pages, which is only available in 4.11+ kernels. So, technically, it should not be possible for MCL_ONFAULT to fail and the call without it to succeed. But keeping the check just in case for now. Acked-by: Simon Horman <horms@ovn.org> Acked-by: Eelco Chaudron <echaudro@redhat.com> Signed-off-by: Ilya Maximets <i.maximets@ovn.org>
2024-06-14 14:22:47 +02:00
/* 'true' if mlockall() succeeded, but doesn't support ONFAULT. */
static bool is_all_memory_locked = false;
/* Buffer used by ovs_strerror() and ovs_format_message(). */
DEFINE_STATIC_PER_THREAD_DATA(struct { char s[128]; },
strerror_buffer,
{ "" });
static char *xreadlink(const char *filename);
void
ovs_assert_failure(const char *where, const char *function,
const char *condition)
{
/* Prevent an infinite loop (or stack overflow) in case VLOG_ABORT happens
* to trigger an assertion failure of its own. */
static int reentry = 0;
switch (reentry++) {
case 0:
VLOG_ABORT("%s: assertion %s failed in %s()",
where, condition, function);
OVS_NOT_REACHED();
case 1:
fprintf(stderr, "%s: assertion %s failed in %s()",
where, condition, function);
abort();
default:
abort();
}
}
void
vswitchd: Only lock pages that are faulted in. The main purpose of locking the memory is to ensure that OVS can keep doing what it did before in case of increased memory pressure, e.g., during VM ingest / migration. Fulfilling this requirement can be achieved without locking all the allocated memory, but only the pages already accessed in the past (faulted in). Processing of the new traffic involves new memory allocations. Latency on these operations can't be guaranteed by the locking. The main difference would be the pre-faulting of the stack memory. However, in order to revalidate or process upcalls on the same traffic, the same amount of stack is likely needed, so all the necessary memory will already be faulted in. Switch 'mlockall' to MCL_ONFAULT to avoid consuming unnecessarily large amounts of RAM on systems with high core counts. For example, in a densely populated OVN cluster this saves about 650 MB of RAM per node on a system with 64 cores. This equates to 320 GB of allocated but unused RAM in a 500 node cluster. This also makes OVS better suited by default for small systems with limited amount of memory. The MCL_ONFAULT flag was introduced in Linux kernel 4.4 and wasn't available at the time of '--mlockall' introduction, but we can use it now. Falling back to an old way of locking in case we're running on an older kernel just in case. Only locking the faulted in pages also makes locking compatible with vhost post-copy live migration by default, because we'll no longer pre-fault all the guest's memory. Post-copy relies on userfaultfd to work on shared huge pages, which is only available in 4.11+ kernels. So, technically, it should not be possible for MCL_ONFAULT to fail and the call without it to succeed. But keeping the check just in case for now. Acked-by: Simon Horman <horms@ovn.org> Acked-by: Eelco Chaudron <echaudro@redhat.com> Signed-off-by: Ilya Maximets <i.maximets@ovn.org>
2024-06-14 14:22:47 +02:00
set_all_memory_locked(void)
{
vswitchd: Only lock pages that are faulted in. The main purpose of locking the memory is to ensure that OVS can keep doing what it did before in case of increased memory pressure, e.g., during VM ingest / migration. Fulfilling this requirement can be achieved without locking all the allocated memory, but only the pages already accessed in the past (faulted in). Processing of the new traffic involves new memory allocations. Latency on these operations can't be guaranteed by the locking. The main difference would be the pre-faulting of the stack memory. However, in order to revalidate or process upcalls on the same traffic, the same amount of stack is likely needed, so all the necessary memory will already be faulted in. Switch 'mlockall' to MCL_ONFAULT to avoid consuming unnecessarily large amounts of RAM on systems with high core counts. For example, in a densely populated OVN cluster this saves about 650 MB of RAM per node on a system with 64 cores. This equates to 320 GB of allocated but unused RAM in a 500 node cluster. This also makes OVS better suited by default for small systems with limited amount of memory. The MCL_ONFAULT flag was introduced in Linux kernel 4.4 and wasn't available at the time of '--mlockall' introduction, but we can use it now. Falling back to an old way of locking in case we're running on an older kernel just in case. Only locking the faulted in pages also makes locking compatible with vhost post-copy live migration by default, because we'll no longer pre-fault all the guest's memory. Post-copy relies on userfaultfd to work on shared huge pages, which is only available in 4.11+ kernels. So, technically, it should not be possible for MCL_ONFAULT to fail and the call without it to succeed. But keeping the check just in case for now. Acked-by: Simon Horman <horms@ovn.org> Acked-by: Eelco Chaudron <echaudro@redhat.com> Signed-off-by: Ilya Maximets <i.maximets@ovn.org>
2024-06-14 14:22:47 +02:00
is_all_memory_locked = true;
}
bool
vswitchd: Only lock pages that are faulted in. The main purpose of locking the memory is to ensure that OVS can keep doing what it did before in case of increased memory pressure, e.g., during VM ingest / migration. Fulfilling this requirement can be achieved without locking all the allocated memory, but only the pages already accessed in the past (faulted in). Processing of the new traffic involves new memory allocations. Latency on these operations can't be guaranteed by the locking. The main difference would be the pre-faulting of the stack memory. However, in order to revalidate or process upcalls on the same traffic, the same amount of stack is likely needed, so all the necessary memory will already be faulted in. Switch 'mlockall' to MCL_ONFAULT to avoid consuming unnecessarily large amounts of RAM on systems with high core counts. For example, in a densely populated OVN cluster this saves about 650 MB of RAM per node on a system with 64 cores. This equates to 320 GB of allocated but unused RAM in a 500 node cluster. This also makes OVS better suited by default for small systems with limited amount of memory. The MCL_ONFAULT flag was introduced in Linux kernel 4.4 and wasn't available at the time of '--mlockall' introduction, but we can use it now. Falling back to an old way of locking in case we're running on an older kernel just in case. Only locking the faulted in pages also makes locking compatible with vhost post-copy live migration by default, because we'll no longer pre-fault all the guest's memory. Post-copy relies on userfaultfd to work on shared huge pages, which is only available in 4.11+ kernels. So, technically, it should not be possible for MCL_ONFAULT to fail and the call without it to succeed. But keeping the check just in case for now. Acked-by: Simon Horman <horms@ovn.org> Acked-by: Eelco Chaudron <echaudro@redhat.com> Signed-off-by: Ilya Maximets <i.maximets@ovn.org>
2024-06-14 14:22:47 +02:00
memory_all_locked(void)
{
vswitchd: Only lock pages that are faulted in. The main purpose of locking the memory is to ensure that OVS can keep doing what it did before in case of increased memory pressure, e.g., during VM ingest / migration. Fulfilling this requirement can be achieved without locking all the allocated memory, but only the pages already accessed in the past (faulted in). Processing of the new traffic involves new memory allocations. Latency on these operations can't be guaranteed by the locking. The main difference would be the pre-faulting of the stack memory. However, in order to revalidate or process upcalls on the same traffic, the same amount of stack is likely needed, so all the necessary memory will already be faulted in. Switch 'mlockall' to MCL_ONFAULT to avoid consuming unnecessarily large amounts of RAM on systems with high core counts. For example, in a densely populated OVN cluster this saves about 650 MB of RAM per node on a system with 64 cores. This equates to 320 GB of allocated but unused RAM in a 500 node cluster. This also makes OVS better suited by default for small systems with limited amount of memory. The MCL_ONFAULT flag was introduced in Linux kernel 4.4 and wasn't available at the time of '--mlockall' introduction, but we can use it now. Falling back to an old way of locking in case we're running on an older kernel just in case. Only locking the faulted in pages also makes locking compatible with vhost post-copy live migration by default, because we'll no longer pre-fault all the guest's memory. Post-copy relies on userfaultfd to work on shared huge pages, which is only available in 4.11+ kernels. So, technically, it should not be possible for MCL_ONFAULT to fail and the call without it to succeed. But keeping the check just in case for now. Acked-by: Simon Horman <horms@ovn.org> Acked-by: Eelco Chaudron <echaudro@redhat.com> Signed-off-by: Ilya Maximets <i.maximets@ovn.org>
2024-06-14 14:22:47 +02:00
return is_all_memory_locked;
}
void
out_of_memory(void)
{
ovs_abort(0, "virtual memory exhausted");
}
void *
xcalloc__(size_t count, size_t size)
{
void *p = count && size ? calloc(count, size) : malloc(1);
if (p == NULL) {
out_of_memory();
}
return p;
}
void *
xzalloc__(size_t size)
{
return xcalloc__(1, size);
}
void *
xmalloc__(size_t size)
{
void *p = malloc(size ? size : 1);
if (p == NULL) {
out_of_memory();
}
return p;
}
void *
xrealloc__(void *p, size_t size)
{
p = realloc(p, size ? size : 1);
if (p == NULL) {
out_of_memory();
}
return p;
}
void *
xcalloc(size_t count, size_t size)
{
COVERAGE_INC(util_xalloc);
return xcalloc__(count, size);
}
void *
xzalloc(size_t size)
{
COVERAGE_INC(util_xalloc);
return xzalloc__(size);
}
void *
xmalloc(size_t size)
{
COVERAGE_INC(util_xalloc);
return xmalloc__(size);
}
void *
xrealloc(void *p, size_t size)
{
COVERAGE_INC(util_xalloc);
return xrealloc__(p, size);
}
void *
xmemdup(const void *p_, size_t size)
{
void *p = xmalloc(size);
nullable_memcpy(p, p_, size);
return p;
}
char *
xmemdup0(const char *p_, size_t length)
{
char *p = xmalloc(length + 1);
memcpy(p, p_, length);
p[length] = '\0';
return p;
}
char *
xstrdup(const char *s)
{
return xmemdup0(s, strlen(s));
}
char * MALLOC_LIKE
nullable_xstrdup(const char *s)
{
return s ? xstrdup(s) : NULL;
}
bool
nullable_string_is_equal(const char *a, const char *b)
{
return a ? b && !strcmp(a, b) : !b;
}
char *
xvasprintf(const char *format, va_list args)
{
va_list args2;
size_t needed;
char *s;
va_copy(args2, args);
needed = vsnprintf(NULL, 0, format, args);
s = xmalloc(needed + 1);
vsnprintf(s, needed + 1, format, args2);
va_end(args2);
return s;
}
void *
x2nrealloc(void *p, size_t *n, size_t s)
{
*n = *n == 0 ? 1 : 2 * *n;
return xrealloc(p, *n * s);
}
/* Allocates and returns 'size' bytes of memory aligned to 'alignment' bytes.
* 'alignment' must be a power of two and a multiple of sizeof(void *).
*
* Use free_size_align() to free the returned memory block. */
void *
xmalloc_size_align(size_t size, size_t alignment)
{
#ifdef HAVE_POSIX_MEMALIGN
void *p;
int error;
COVERAGE_INC(util_xalloc);
error = posix_memalign(&p, alignment, size ? size : 1);
if (error != 0) {
out_of_memory();
}
return p;
#else
/* Allocate room for:
*
* - Header padding: Up to alignment - 1 bytes, to allow the
* pointer 'q' to be aligned exactly sizeof(void *) bytes before the
* beginning of the alignment.
*
* - Pointer: A pointer to the start of the header padding, to allow us
* to free() the block later.
*
* - User data: 'size' bytes.
*
* - Trailer padding: Enough to bring the user data up to a alignment
* multiple.
*
* +---------------+---------+------------------------+---------+
* | header | pointer | user data | trailer |
* +---------------+---------+------------------------+---------+
* ^ ^ ^
* | | |
* p q r
*
*/
void *p, *r, **q;
bool runt;
if (!IS_POW2(alignment) || (alignment % sizeof(void *) != 0)) {
ovs_abort(0, "Invalid alignment");
}
p = xmalloc((alignment - 1)
+ sizeof(void *)
+ ROUND_UP(size, alignment));
runt = PAD_SIZE((uintptr_t) p, alignment) < sizeof(void *);
/* When the padding size < sizeof(void*), we don't have enough room for
* pointer 'q'. As a reuslt, need to move 'r' to the next alignment.
* So ROUND_UP when xmalloc above, and ROUND_UP again when calculate 'r'
* below.
*/
r = (void *) ROUND_UP((uintptr_t) p + (runt ? alignment : 0), alignment);
q = (void **) r - 1;
*q = p;
return r;
#endif
}
void
free_size_align(void *p)
{
#ifdef HAVE_POSIX_MEMALIGN
free(p);
#else
if (p) {
void **q = (void **) p - 1;
free(*q);
}
#endif
}
/* Allocates and returns 'size' bytes of memory aligned to a cache line and in
* dedicated cache lines. That is, the memory block returned will not share a
* cache line with other data, avoiding "false sharing".
*
* Use free_cacheline() to free the returned memory block. */
void *
xmalloc_cacheline(size_t size)
{
return xmalloc_size_align(size, CACHE_LINE_SIZE);
}
/* Like xmalloc_cacheline() but clears the allocated memory to all zero
* bytes. */
void *
xzalloc_cacheline(size_t size)
{
void *p = xmalloc_cacheline(size);
memset(p, 0, size);
return p;
}
/* Frees a memory block allocated with xmalloc_cacheline() or
* xzalloc_cacheline(). */
void
free_cacheline(void *p)
{
free_size_align(p);
}
void *
xmalloc_pagealign(size_t size)
{
return xmalloc_size_align(size, get_page_size());
}
void
free_pagealign(void *p)
{
free_size_align(p);
}
char *
xasprintf(const char *format, ...)
{
va_list args;
char *s;
va_start(args, format);
s = xvasprintf(format, args);
va_end(args);
return s;
}
/* Similar to strlcpy() from OpenBSD, but it never reads more than 'size - 1'
* bytes from 'src' and doesn't return anything. */
void
ovs_strlcpy(char *dst, const char *src, size_t size)
{
if (size > 0) {
size_t len = strnlen(src, size - 1);
memcpy(dst, src, len);
dst[len] = '\0';
}
}
/* Copies 'src' to 'dst'. Reads no more than 'size - 1' bytes from 'src'.
* Always null-terminates 'dst' (if 'size' is nonzero), and writes a zero byte
* to every otherwise unused byte in 'dst'.
*
* Except for performance, the following call:
* ovs_strzcpy(dst, src, size);
* is equivalent to these two calls:
* memset(dst, '\0', size);
* ovs_strlcpy(dst, src, size);
*
* (Thus, ovs_strzcpy() is similar to strncpy() without some of the pitfalls.)
*/
void
ovs_strzcpy(char *dst, const char *src, size_t size)
{
if (size > 0) {
size_t len = strnlen(src, size - 1);
memcpy(dst, src, len);
memset(dst + len, '\0', size - len);
}
}
/*
* Returns true if 'str' ends with given 'suffix'.
*/
int
string_ends_with(const char *str, const char *suffix)
{
int str_len = strlen(str);
int suffix_len = strlen(suffix);
return (str_len >= suffix_len) &&
(0 == strcmp(str + (str_len - suffix_len), suffix));
}
/* Prints 'format' on stderr, formatting it like printf() does. If 'err_no' is
* nonzero, then it is formatted with ovs_retval_to_string() and appended to
* the message inside parentheses. Then, terminates with abort().
*
* This function is preferred to ovs_fatal() in a situation where it would make
* sense for a monitoring process to restart the daemon.
*
* 'format' should not end with a new-line, because this function will add one
* itself. */
void
ovs_abort(int err_no, const char *format, ...)
{
va_list args;
va_start(args, format);
ovs_abort_valist(err_no, format, args);
}
/* Same as ovs_abort() except that the arguments are supplied as a va_list. */
void
ovs_abort_valist(int err_no, const char *format, va_list args)
{
ovs_error_valist(err_no, format, args);
abort();
}
/* Prints 'format' on stderr, formatting it like printf() does. If 'err_no' is
* nonzero, then it is formatted with ovs_retval_to_string() and appended to
* the message inside parentheses. Then, terminates with EXIT_FAILURE.
*
* 'format' should not end with a new-line, because this function will add one
* itself. */
void
ovs_fatal(int err_no, const char *format, ...)
{
va_list args;
va_start(args, format);
ovs_fatal_valist(err_no, format, args);
}
/* Same as ovs_fatal() except that the arguments are supplied as a va_list. */
void
ovs_fatal_valist(int err_no, const char *format, va_list args)
{
ovs_error_valist(err_no, format, args);
exit(EXIT_FAILURE);
}
/* Prints 'format' on stderr, formatting it like printf() does. If 'err_no' is
* nonzero, then it is formatted with ovs_retval_to_string() and appended to
* the message inside parentheses.
*
* 'format' should not end with a new-line, because this function will add one
* itself. */
void
ovs_error(int err_no, const char *format, ...)
{
va_list args;
va_start(args, format);
ovs_error_valist(err_no, format, args);
va_end(args);
}
/* Same as ovs_error() except that the arguments are supplied as a va_list. */
void
ovs_error_valist(int err_no, const char *format, va_list args)
{
const char *subprogram_name = get_subprogram_name();
int save_errno = errno;
if (subprogram_name[0]) {
fprintf(stderr, "%s(%s): ", program_name, subprogram_name);
} else {
fprintf(stderr, "%s: ", program_name);
}
vfprintf(stderr, format, args);
if (err_no != 0) {
fprintf(stderr, " (%s)", ovs_retval_to_string(err_no));
}
putc('\n', stderr);
errno = save_errno;
}
/* Many OVS functions return an int which is one of:
* - 0: no error yet
* - >0: errno value
* - EOF: end of file (not necessarily an error; depends on the function called)
*
* Returns the appropriate human-readable string. The caller must copy the
* string if it wants to hold onto it, as the storage may be overwritten on
* subsequent function calls.
*/
const char *
ovs_retval_to_string(int retval)
{
return (!retval ? ""
: retval == EOF ? "End of file"
: ovs_strerror(retval));
}
/* This function returns the string describing the error number in 'error'
* for POSIX platforms. For Windows, this function can be used for C library
* calls. For socket calls that are also used in Windows, use sock_strerror()
* instead. For WINAPI calls, look at ovs_lasterror_to_string(). */
const char *
ovs_strerror(int error)
{
enum { BUFSIZE = sizeof strerror_buffer_get()->s };
int save_errno;
char *buffer;
char *s;
if (error == 0) {
/*
* strerror(0) varies among platforms:
*
* Success
* No error
* Undefined error: 0
*
* We want to provide a consistent result here because
* our testsuite has test cases which strictly matches
* log messages containing this string.
*/
return "Success";
}
save_errno = errno;
buffer = strerror_buffer_get()->s;
#if STRERROR_R_CHAR_P
/* GNU style strerror_r() might return an immutable static string, or it
* might write and return 'buffer', but in either case we can pass the
* returned string directly to the caller. */
s = strerror_r(error, buffer, BUFSIZE);
#else /* strerror_r() returns an int. */
s = buffer;
if (strerror_r(error, buffer, BUFSIZE)) {
/* strerror_r() is only allowed to fail on ERANGE (because the buffer
* is too short). We don't check the actual failure reason because
* POSIX requires strerror_r() to return the error but old glibc
* (before 2.13) returns -1 and sets errno. */
snprintf(buffer, BUFSIZE, "Unknown error %d", error);
}
#endif
errno = save_errno;
return s;
}
/* Sets global "program_name" and "program_version" variables. Should
* be called at the beginning of main() with "argv[0]" as the argument
* to 'argv0'.
*
* 'version' should contain the version of the caller's program. If 'version'
* is the same as the VERSION #define, the caller is assumed to be part of Open
* vSwitch. Otherwise, it is assumed to be an external program linking against
* the Open vSwitch libraries.
*
*/
void
ovs_set_program_name(const char *argv0, const char *version)
{
char *basename;
#ifdef _WIN32
size_t max_len = strlen(argv0) + 1;
SetErrorMode(GetErrorMode() | SEM_NOGPFAULTERRORBOX);
#if _MSC_VER < 1900
/* This function is deprecated from 1900 (Visual Studio 2015) */
_set_output_format(_TWO_DIGIT_EXPONENT);
#endif
basename = xmalloc(max_len);
_splitpath_s(argv0, NULL, 0, NULL, 0, basename, max_len, NULL, 0);
#else
const char *slash = strrchr(argv0, '/');
basename = xstrdup(slash ? slash + 1 : argv0);
#endif
assert_single_threaded();
free(program_name);
/* Remove libtool prefix, if it is there */
if (strncmp(basename, "lt-", 3) == 0) {
char *tmp_name = basename;
basename = xstrdup(basename + 3);
free(tmp_name);
}
program_name = basename;
free(program_version);
if (!strcmp(version, VERSION)) {
program_version = xasprintf("%s (Open vSwitch) "VERSION"\n",
program_name);
} else {
program_version = xasprintf("%s %s\n"
"Open vSwitch Library "VERSION"\n",
program_name, version);
}
}
/* Returns the name of the currently running thread or process. */
const char *
get_subprogram_name(void)
{
const char *name = subprogram_name_get();
return name ? name : "";
}
/* Sets 'subprogram_name' as the name of the currently running thread or
* process. (This appears in log messages and may also be visible in system
* process listings and debuggers.) */
void
set_subprogram_name(const char *subprogram_name)
{
char *pname = xstrdup(subprogram_name ? subprogram_name : program_name);
free(subprogram_name_set(pname));
#if HAVE_GLIBC_PTHREAD_SETNAME_NP
/* The maximum supported thread name including '\0' is 16.
* Add '>' at 0th position to highlight that the name was truncated. */
if (strlen(pname) > 15) {
memmove(pname, &pname[strlen(pname) - 15], 15 + 1);
pname[0] = '>';
}
pthread_setname_np(pthread_self(), pname);
#elif HAVE_NETBSD_PTHREAD_SETNAME_NP
pthread_setname_np(pthread_self(), "%s", pname);
#elif HAVE_PTHREAD_SET_NAME_NP
pthread_set_name_np(pthread_self(), pname);
#endif
}
unsigned int
get_page_size(void)
{
static unsigned int cached;
if (!cached) {
#ifndef _WIN32
long int value = sysconf(_SC_PAGESIZE);
#else
long int value;
SYSTEM_INFO sysinfo;
GetSystemInfo(&sysinfo);
value = sysinfo.dwPageSize;
#endif
if (value >= 0) {
cached = value;
}
}
return cached;
}
/* Returns the time at which the system booted, as the number of milliseconds
* since the epoch, or 0 if the time of boot cannot be determined. */
long long int
get_boot_time(void)
{
static long long int cache_expiration = LLONG_MIN;
static long long int boot_time;
ovs_assert(LINUX);
if (time_msec() >= cache_expiration) {
static const char stat_file[] = "/proc/stat";
char line[128];
FILE *stream;
cache_expiration = time_msec() + 5 * 1000;
stream = fopen(stat_file, "r");
if (!stream) {
VLOG_ERR_ONCE("%s: open failed (%s)",
stat_file, ovs_strerror(errno));
return boot_time;
}
while (fgets(line, sizeof line, stream)) {
long long int btime;
if (ovs_scan(line, "btime %lld", &btime)) {
boot_time = btime * 1000;
goto done;
}
}
VLOG_ERR_ONCE("%s: btime not found", stat_file);
done:
fclose(stream);
}
return boot_time;
}
/* This is a wrapper for setting timeout in control utils.
* The value of OVS_CTL_TIMEOUT environment variable will be used by
* default if 'secs' is not specified. */
void
ctl_timeout_setup(unsigned int secs)
{
if (!secs) {
char *env = getenv("OVS_CTL_TIMEOUT");
if (env && env[0]) {
str_to_uint(env, 10, &secs);
}
}
if (secs) {
time_alarm(secs);
}
}
/* Returns a pointer to a string describing the program version. The
* caller must not modify or free the returned string.
2011-08-04 16:50:25 -07:00
*/
const char *
ovs_get_program_version(void)
{
return program_version;
}
/* Returns a pointer to a string describing the program name. The
* caller must not modify or free the returned string.
*/
const char *
ovs_get_program_name(void)
{
return program_name;
}
/* Print the version information for the program. */
void
ovs_print_version(uint8_t min_ofp, uint8_t max_ofp)
{
printf("%s", program_version);
if (min_ofp || max_ofp) {
printf("OpenFlow versions %#x:%#x\n", min_ofp, max_ofp);
}
}
/* Writes the 'size' bytes in 'buf' to 'stream' as hex bytes arranged 16 per
* line. Numeric offsets are also included, starting at 'ofs' for the first
* byte in 'buf'. If 'ascii' is true then the corresponding ASCII characters
* are also rendered alongside. */
void
ovs_hex_dump(FILE *stream, const void *buf_, size_t size,
uintptr_t ofs, bool ascii)
{
const uint8_t *buf = buf_;
const size_t per_line = 16; /* Maximum bytes per line. */
while (size > 0) {
size_t i;
/* Number of bytes on this line. */
size_t start = ofs % per_line;
size_t end = per_line;
if (end - start > size) {
end = start + size;
}
size_t n = end - start;
/* Print line. */
fprintf(stream, "%08"PRIxMAX" ",
(uintmax_t) ROUND_DOWN(ofs, per_line));
for (i = 0; i < start; i++) {
fprintf(stream, " ");
}
for (; i < end; i++) {
fprintf(stream, "%c%02x",
i == per_line / 2 ? '-' : ' ', buf[i - start]);
}
if (ascii) {
fprintf(stream, " ");
for (; i < per_line; i++) {
fprintf(stream, " ");
}
fprintf(stream, "|");
for (i = 0; i < start; i++) {
fprintf(stream, " ");
}
for (; i < end; i++) {
int c = buf[i - start];
putc(c >= 32 && c < 127 ? c : '.', stream);
}
for (; i < per_line; i++) {
fprintf(stream, " ");
}
fprintf(stream, "|");
}
fprintf(stream, "\n");
ofs += n;
buf += n;
size -= n;
}
}
bool
str_to_int(const char *s, int base, int *i)
{
long long ll;
bool ok = str_to_llong(s, base, &ll);
if (!ok || ll < INT_MIN || ll > INT_MAX) {
*i = 0;
return false;
}
*i = ll;
return true;
}
bool
str_to_long(const char *s, int base, long *li)
{
long long ll;
bool ok = str_to_llong(s, base, &ll);
if (!ok || ll < LONG_MIN || ll > LONG_MAX) {
*li = 0;
return false;
}
*li = ll;
return true;
}
bool
str_to_llong(const char *s, int base, long long *x)
{
char *tail;
bool ok = str_to_llong_with_tail(s, &tail, base, x);
if (*tail != '\0') {
*x = 0;
return false;
}
return ok;
}
bool
str_to_llong_with_tail(const char *s, char **tail, int base, long long *x)
{
int save_errno = errno;
errno = 0;
*x = strtoll(s, tail, base);
if (errno == EINVAL || errno == ERANGE || *tail == s) {
errno = save_errno;
*x = 0;
return false;
} else {
errno = save_errno;
return true;
}
}
bool
str_to_uint(const char *s, int base, unsigned int *u)
{
long long ll;
bool ok = str_to_llong(s, base, &ll);
if (!ok || ll < 0 || ll > UINT_MAX) {
*u = 0;
return false;
} else {
*u = ll;
return true;
}
}
bool
str_to_ullong(const char *s, int base, unsigned long long *x)
{
int save_errno = errno;
char *tail;
errno = 0;
*x = strtoull(s, &tail, base);
if (errno == EINVAL || errno == ERANGE || tail == s || *tail != '\0') {
errno = save_errno;
*x = 0;
return false;
} else {
errno = save_errno;
return true;
}
}
bool
str_to_llong_range(const char *s, int base, long long *begin,
long long *end)
{
char *tail;
if (str_to_llong_with_tail(s, &tail, base, begin)
&& *tail == '-'
&& str_to_llong(tail + 1, base, end)) {
return true;
}
*begin = 0;
*end = 0;
return false;
}
/* Converts floating-point string 's' into a double. If successful, stores
* the double in '*d' and returns true; on failure, stores 0 in '*d' and
* returns false.
*
* Underflow (e.g. "1e-9999") is not considered an error, but overflow
* (e.g. "1e9999)" is. */
bool
str_to_double(const char *s, double *d)
{
int save_errno = errno;
char *tail;
errno = 0;
*d = strtod(s, &tail);
if (errno == EINVAL || (errno == ERANGE && *d != 0)
|| tail == s || *tail != '\0') {
errno = save_errno;
*d = 0;
return false;
} else {
errno = save_errno;
return true;
}
}
/* Returns the value of 'c' as a hexadecimal digit. */
int
hexit_value(unsigned char c)
{
static const signed char tbl[UCHAR_MAX + 1] = {
#define TBL(x) \
( x >= '0' && x <= '9' ? x - '0' \
: x >= 'a' && x <= 'f' ? x - 'a' + 0xa \
: x >= 'A' && x <= 'F' ? x - 'A' + 0xa \
: -1)
#define TBL0(x) TBL(x), TBL((x) + 1), TBL((x) + 2), TBL((x) + 3)
#define TBL1(x) TBL0(x), TBL0((x) + 4), TBL0((x) + 8), TBL0((x) + 12)
#define TBL2(x) TBL1(x), TBL1((x) + 16), TBL1((x) + 32), TBL1((x) + 48)
TBL2(0), TBL2(64), TBL2(128), TBL2(192)
};
return tbl[c];
}
2009-10-19 14:04:14 -07:00
/* Returns the integer value of the 'n' hexadecimal digits starting at 's', or
* UINTMAX_MAX if one of those "digits" is not really a hex digit. Sets '*ok'
* to true if the conversion succeeds or to false if a non-hex digit is
* detected. */
uintmax_t
hexits_value(const char *s, size_t n, bool *ok)
{
uintmax_t value;
size_t i;
value = 0;
for (i = 0; i < n; i++) {
int hexit = hexit_value(s[i]);
if (hexit < 0) {
*ok = false;
return UINTMAX_MAX;
}
value = (value << 4) + hexit;
}
*ok = true;
return value;
}
/* Parses the string in 's' as an integer in either hex or decimal format and
* puts the result right justified in the array 'valuep' that is 'field_width'
* big. If the string is in hex format, the value may be arbitrarily large;
* integers are limited to 64-bit values. (The rationale is that decimal is
* likely to represent a number and 64 bits is a reasonable maximum whereas
* hex could either be a number or a byte string.)
*
* On return 'tail' points to the first character in the string that was
* not parsed as part of the value. ERANGE is returned if the value is too
* large to fit in the given field. */
int
parse_int_string(const char *s, uint8_t *valuep, int field_width, char **tail)
{
unsigned long long int integer;
int i;
if (!strncmp(s, "0x", 2) || !strncmp(s, "0X", 2)) {
uint8_t *hexit_str;
int len = 0;
int val_idx;
int err = 0;
s += 2;
hexit_str = xmalloc(field_width * 2);
for (;;) {
uint8_t hexit;
bool ok;
s += strspn(s, " \t\r\n");
hexit = hexits_value(s, 1, &ok);
if (!ok) {
*tail = CONST_CAST(char *, s);
break;
}
if (hexit != 0 || len) {
if (DIV_ROUND_UP(len + 1, 2) > field_width) {
err = ERANGE;
goto free;
}
hexit_str[len] = hexit;
len++;
}
s++;
}
val_idx = field_width;
for (i = len - 1; i >= 0; i -= 2) {
val_idx--;
valuep[val_idx] = hexit_str[i];
if (i > 0) {
valuep[val_idx] += hexit_str[i - 1] << 4;
}
}
memset(valuep, 0, val_idx);
free:
free(hexit_str);
return err;
}
errno = 0;
integer = strtoull(s, tail, 0);
if (errno || s == *tail) {
return errno ? errno : EINVAL;
}
for (i = field_width - 1; i >= 0; i--) {
valuep[i] = integer;
integer >>= 8;
}
if (integer) {
return ERANGE;
}
return 0;
}
/* Returns the current working directory as a malloc()'d string, or a null
* pointer if the current working directory cannot be determined. */
char *
get_cwd(void)
{
long int path_max;
size_t size;
/* Get maximum path length or at least a reasonable estimate. */
#ifndef _WIN32
path_max = pathconf(".", _PC_PATH_MAX);
#else
path_max = MAX_PATH;
#endif
size = (path_max < 0 ? 1024
: path_max > 10240 ? 10240
: path_max);
/* Get current working directory. */
for (;;) {
char *buf = xmalloc(size);
if (getcwd(buf, size)) {
return xrealloc(buf, strlen(buf) + 1);
} else {
int error = errno;
free(buf);
if (error != ERANGE) {
VLOG_WARN("getcwd failed (%s)", ovs_strerror(error));
return NULL;
}
size *= 2;
}
}
}
2010-11-09 14:38:28 -08:00
static char *
all_slashes_name(const char *s)
{
return xstrdup(s[0] == '/' && s[1] == '/' && s[2] != '/' ? "//"
: s[0] == '/' ? "/"
: ".");
}
#ifndef _WIN32
2009-10-19 14:04:14 -07:00
/* Returns the directory name portion of 'file_name' as a malloc()'d string,
* similar to the POSIX dirname() function but thread-safe. */
char *
dir_name(const char *file_name)
{
size_t len = strlen(file_name);
while (len > 0 && file_name[len - 1] == '/') {
len--;
}
while (len > 0 && file_name[len - 1] != '/') {
len--;
}
while (len > 0 && file_name[len - 1] == '/') {
len--;
}
2010-11-09 14:38:28 -08:00
return len ? xmemdup0(file_name, len) : all_slashes_name(file_name);
}
/* Returns the file name portion of 'file_name' as a malloc()'d string,
* similar to the POSIX basename() function but thread-safe. */
char *
base_name(const char *file_name)
{
size_t end, start;
end = strlen(file_name);
while (end > 0 && file_name[end - 1] == '/') {
end--;
}
if (!end) {
return all_slashes_name(file_name);
2009-10-19 14:04:14 -07:00
}
2010-11-09 14:38:28 -08:00
start = end;
while (start > 0 && file_name[start - 1] != '/') {
start--;
}
return xmemdup0(file_name + start, end - start);
2009-10-19 14:04:14 -07:00
}
#endif /* _WIN32 */
bool
is_file_name_absolute(const char *fn)
{
#ifdef _WIN32
/* Use platform specific API */
return !PathIsRelative(fn);
#else
/* An absolute path begins with /. */
return fn[0] == '/';
#endif
}
/* If 'file_name' is absolute, returns a copy of 'file_name'. Otherwise,
* returns an absolute path to 'file_name' considering it relative to 'dir',
* which itself must be absolute. 'dir' may be null or the empty string, in
* which case the current working directory is used.
*
* Returns a null pointer if 'dir' is null and getcwd() fails. */
char *
abs_file_name(const char *dir, const char *file_name)
{
/* If it's already absolute, return a copy. */
if (is_file_name_absolute(file_name)) {
return xstrdup(file_name);
}
/* If a base dir was supplied, use it. We assume, without checking, that
* the base dir is absolute.*/
if (dir && dir[0]) {
char *separator = dir[strlen(dir) - 1] == '/' ? "" : "/";
return xasprintf("%s%s%s", dir, separator, file_name);
}
#if _WIN32
/* It's a little complicated to make an absolute path on Windows because a
* relative path might still specify a drive letter. The OS has a function
* to do the job for us, so use it. */
char abs_path[MAX_PATH];
DWORD n = GetFullPathName(file_name, sizeof abs_path, abs_path, NULL);
return n > 0 && n <= sizeof abs_path ? xmemdup0(abs_path, n) : NULL;
#else
/* Outside Windows, do the job ourselves. */
char *cwd = get_cwd();
if (!cwd) {
return NULL;
}
char *abs_name = xasprintf("%s/%s", cwd, file_name);
free(cwd);
return abs_name;
#endif
}
/* Like readlink(), but returns the link name as a null-terminated string in
* allocated memory that the caller must eventually free (with free()).
* Returns NULL on error, in which case errno is set appropriately. */
static char *
xreadlink(const char *filename)
{
#ifdef _WIN32
errno = ENOENT;
return NULL;
#else
size_t size;
for (size = 64; ; size *= 2) {
char *buf = xmalloc(size);
ssize_t retval = readlink(filename, buf, size);
int error = errno;
if (retval >= 0 && retval < size) {
buf[retval] = '\0';
return buf;
}
free(buf);
if (retval < 0) {
errno = error;
return NULL;
}
}
#endif
}
/* Returns a version of 'filename' with symlinks in the final component
* dereferenced. This differs from realpath() in that:
*
* - 'filename' need not exist.
*
* - If 'filename' does exist as a symlink, its referent need not exist.
*
* - Only symlinks in the final component of 'filename' are dereferenced.
*
* For Windows platform, this function returns a string that has the same
* value as the passed string.
*
* The caller must eventually free the returned string (with free()). */
char *
follow_symlinks(const char *filename)
{
#ifndef _WIN32
struct stat s;
char *fn;
int i;
fn = xstrdup(filename);
for (i = 0; i < 10; i++) {
char *linkname;
char *next_fn;
if (lstat(fn, &s) != 0 || !S_ISLNK(s.st_mode)) {
return fn;
}
linkname = xreadlink(fn);
if (!linkname) {
VLOG_WARN("%s: readlink failed (%s)",
filename, ovs_strerror(errno));
return fn;
}
if (linkname[0] == '/') {
/* Target of symlink is absolute so use it raw. */
next_fn = linkname;
} else {
/* Target of symlink is relative so add to 'fn''s directory. */
char *dir = dir_name(fn);
if (!strcmp(dir, ".")) {
next_fn = linkname;
} else {
char *separator = dir[strlen(dir) - 1] == '/' ? "" : "/";
next_fn = xasprintf("%s%s%s", dir, separator, linkname);
free(linkname);
}
free(dir);
}
free(fn);
fn = next_fn;
}
VLOG_WARN("%s: too many levels of symlinks", filename);
free(fn);
#endif
return xstrdup(filename);
}
/* Pass a value to this function if it is marked with
* __attribute__((warn_unused_result)) and you genuinely want to ignore
* its return value. (Note that every scalar type can be implicitly
* converted to bool.) */
void ignore(bool x OVS_UNUSED) { }
/* Returns an appropriate delimiter for inserting just before the 0-based item
* 'index' in a list that has 'total' items in it. */
const char *
english_list_delimiter(size_t index, size_t total)
{
return (index == 0 ? ""
: index < total - 1 ? ", "
: total > 2 ? ", and "
: " and ");
}
/* Returns the number of trailing 0-bits in 'n'. Undefined if 'n' == 0. */
#if __GNUC__ >= 4 || _MSC_VER
/* Defined inline in util.h. */
#else
/* Returns the number of trailing 0-bits in 'n'. Undefined if 'n' == 0. */
int
raw_ctz(uint64_t n)
{
uint64_t k;
int count = 63;
#define CTZ_STEP(X) \
k = n << (X); \
if (k) { \
count -= X; \
n = k; \
}
CTZ_STEP(32);
CTZ_STEP(16);
CTZ_STEP(8);
CTZ_STEP(4);
CTZ_STEP(2);
CTZ_STEP(1);
#undef CTZ_STEP
return count;
}
/* Returns the number of leading 0-bits in 'n'. Undefined if 'n' == 0. */
int
raw_clz64(uint64_t n)
{
uint64_t k;
int count = 63;
#define CLZ_STEP(X) \
k = n >> (X); \
if (k) { \
count -= X; \
n = k; \
}
CLZ_STEP(32);
CLZ_STEP(16);
CLZ_STEP(8);
CLZ_STEP(4);
CLZ_STEP(2);
CLZ_STEP(1);
#undef CLZ_STEP
return count;
}
#endif
#if NEED_COUNT_1BITS_8
util: New function popcount(). This is the fastest portable implementation among the ones below, as measured with GCC 4.4 on a Xeon X3430. The measeured times were, in seconds: popcount1 25.6 popcount2 6.9 (but is not portable) popcount3 31.4 popcount4 25.6 popcount5 61.6 (and is buggy) popcount6 64.6 popcount7 32.3 popcount8 11.2 int popcount1(unsigned int x) { return __builtin_popcount(x); } int popcount2(unsigned int x) { unsigned int y; asm("popcnt %1, %0" : "=r" (y) : "g" (x)); return y; } int popcount3(unsigned int x) { unsigned int n; n = (x >> 1) & 033333333333; x -= n; n = (n >> 1) & 033333333333; x -= n; x = (x + (x >> 3)) & 030707070707; return x % 63; } int popcount4(unsigned int x) { x -= (x >> 1) & 0x55555555; x = (x & 0x33333333) + ((x >> 2) & 0x33333333); x = (x + (x >> 4)) & 0x0f0f0f0f; x += x >> 8; x += x >> 16; return x & 0x3f; } int popcount5(unsigned int x) { int n; n = 0; while (x) { if (x & 0xf) { n += ((0xe9949440 >> (x & 0xf)) & 3) + 1; } x >>= 4; } return n; } int popcount6(unsigned int x) { int n; n = 0; while (x) { n += (0xe994 >> (x & 7)) & 3; x >>= 3; } return n; } int popcount7(unsigned int x) { static const int table[16] = { 0, 1, 1, 2, 1, 2, 2, 3, 1, 2, 2, 3, 2, 3, 3, 4 }; return (table[x & 0xf] + table[(x >> 4) & 0xf] + table[(x >> 8) & 0xf] + table[(x >> 12) & 0xf] + table[(x >> 16) & 0xf] + table[(x >> 20) & 0xf] + table[(x >> 24) & 0xf] + table[x >> 28]); } static int popcount8(unsigned int x) { ((((X) & (1 << 0)) != 0) + \ (((X) & (1 << 1)) != 0) + \ (((X) & (1 << 2)) != 0) + \ (((X) & (1 << 3)) != 0) + \ (((X) & (1 << 4)) != 0) + \ (((X) & (1 << 5)) != 0) + \ (((X) & (1 << 6)) != 0) + \ (((X) & (1 << 7)) != 0)) static const uint8_t popcount8[256] = { INIT64(0), INIT64(64), INIT64(128), INIT64(192) }; return (popcount8[x & 0xff] + popcount8[(x >> 8) & 0xff] + popcount8[(x >> 16) & 0xff] + popcount8[x >> 24]); } int main(void) { unsigned long long int x; int n; n = 0; for (x = 0; x <= UINT32_MAX; x++) { n += popcount8(x); } printf("%d\n", n); return 0; } Signed-off-by: Ben Pfaff <blp@nicira.com>
2012-07-20 12:38:59 -07:00
#define INIT1(X) \
((((X) & (1 << 0)) != 0) + \
(((X) & (1 << 1)) != 0) + \
(((X) & (1 << 2)) != 0) + \
(((X) & (1 << 3)) != 0) + \
(((X) & (1 << 4)) != 0) + \
(((X) & (1 << 5)) != 0) + \
(((X) & (1 << 6)) != 0) + \
(((X) & (1 << 7)) != 0))
#define INIT2(X) INIT1(X), INIT1((X) + 1)
#define INIT4(X) INIT2(X), INIT2((X) + 2)
#define INIT8(X) INIT4(X), INIT4((X) + 4)
#define INIT16(X) INIT8(X), INIT8((X) + 8)
#define INIT32(X) INIT16(X), INIT16((X) + 16)
#define INIT64(X) INIT32(X), INIT32((X) + 32)
const uint8_t count_1bits_8[256] = {
INIT64(0), INIT64(64), INIT64(128), INIT64(192)
};
#endif
/* Returns true if the 'n' bytes starting at 'p' are 'byte'. */
bool
is_all_byte(const void *p_, size_t n, uint8_t byte)
{
const uint8_t *p = p_;
size_t i;
for (i = 0; i < n; i++) {
if (p[i] != byte) {
return false;
}
}
return true;
}
/* Returns true if the 'n' bytes starting at 'p' are zeros. */
bool
is_all_zeros(const void *p, size_t n)
{
return is_all_byte(p, n, 0);
}
/* Returns true if the 'n' bytes starting at 'p' are 0xff. */
bool
is_all_ones(const void *p, size_t n)
{
return is_all_byte(p, n, 0xff);
}
odp-util: Fix clearing match mask if set action is partially unnecessary. While committing set() actions, commit() could wildcard all the fields that are same in match key and in the set action. This leads to situation where mask after commit could actually contain less bits than it was before. And if set action was partially committed, all the fields that were the same will be cleared out from the matching key resulting in the incorrect (too wide) flow. For example, for the flow that matches on both src and dst mac addresses, if the dst mac is the same and only src should be changed by the set() action, destination address will be wildcarded in the match key and will never be matched, i.e. flows with any destination mac will match, which is not correct. Setting OF rule: in_port=1,dl_src=50:54:00:00:00:09 actions=mod_dl_dst(50:54:00:00:00:0a),output(2) Sending following packets on port 1: 1. eth(src=50:54:00:00:00:09,dst=50:54:00:00:00:0a),eth_type(0x0800) 2. eth(src=50:54:00:00:00:09,dst=50:54:00:00:00:0c),eth_type(0x0800) 3. eth(src=50:54:00:00:00:0b,dst=50:54:00:00:00:0c),eth_type(0x0800) Resulted datapath flows: eth(dst=50:54:00:00:00:0c),<...>, actions:set(eth(dst=50:54:00:00:00:0a)),2 eth(src=50:54:00:00:00:09,dst=50:54:00:00:00:0a),<...>, actions:2 The first flow doesn't have any match on source MAC address and the third packet successfully matched on it while it must be dropped. Fix that by updating the match mask with only the new bits set by commit(), but keeping those that were cleared (OR operation). With fix applied, resulted correct flows are: eth(src=50:54:00:00:00:09,dst=50:54:00:00:00:0a),<...>, actions:2 eth(src=50:54:00:00:00:09,dst=50:54:00:00:00:0c),<...>, actions:set(eth(dst=50:54:00:00:00:0a)),2 eth(src=50:54:00:00:00:0b),<...>, actions:drop The code before commit dbf4a92800d0 was not able to reduce the mask, it was only possible to expand it to exact match, so it was OK to update original matching mask with the new value in all cases. Fixes: dbf4a92800d0 ("odp-util: Do not rewrite fields with the same values as matched") Reported-at: https://bugzilla.redhat.com/show_bug.cgi?id=1854376 Acked-by: Eli Britstein <elibr@mellanox.com> Tested-by: Adrián Moreno <amorenoz@redhat.com> Signed-off-by: Ilya Maximets <i.maximets@ovn.org>
2020-07-27 17:41:35 +02:00
/* *dst |= *src for 'n' bytes. */
void
or_bytes(void *dst_, const void *src_, size_t n)
{
const uint8_t *src = src_;
uint8_t *dst = dst_;
size_t i;
for (i = 0; i < n; i++) {
*dst++ |= *src++;
}
}
/* Copies 'n_bits' bits starting from bit 'src_ofs' in 'src' to the 'n_bits'
* starting from bit 'dst_ofs' in 'dst'. 'src' is 'src_len' bytes long and
* 'dst' is 'dst_len' bytes long.
*
* If you consider all of 'src' to be a single unsigned integer in network byte
* order, then bit N is the bit with value 2**N. That is, bit 0 is the bit
* with value 1 in src[src_len - 1], bit 1 is the bit with value 2, bit 2 is
* the bit with value 4, ..., bit 8 is the bit with value 1 in src[src_len -
* 2], and so on. Similarly for 'dst'.
*
* Required invariants:
* src_ofs + n_bits <= src_len * 8
* dst_ofs + n_bits <= dst_len * 8
* 'src' and 'dst' must not overlap.
*/
void
bitwise_copy(const void *src_, unsigned int src_len, unsigned int src_ofs,
void *dst_, unsigned int dst_len, unsigned int dst_ofs,
unsigned int n_bits)
{
const uint8_t *src = src_;
uint8_t *dst = dst_;
src += src_len - (src_ofs / 8 + 1);
src_ofs %= 8;
dst += dst_len - (dst_ofs / 8 + 1);
dst_ofs %= 8;
if (src_ofs == 0 && dst_ofs == 0) {
unsigned int n_bytes = n_bits / 8;
if (n_bytes) {
dst -= n_bytes - 1;
src -= n_bytes - 1;
memcpy(dst, src, n_bytes);
n_bits %= 8;
src--;
dst--;
}
if (n_bits) {
uint8_t mask = (1 << n_bits) - 1;
*dst = (*dst & ~mask) | (*src & mask);
}
} else {
while (n_bits > 0) {
unsigned int max_copy = 8 - MAX(src_ofs, dst_ofs);
unsigned int chunk = MIN(n_bits, max_copy);
uint8_t mask = ((1 << chunk) - 1) << dst_ofs;
*dst &= ~mask;
*dst |= ((*src >> src_ofs) << dst_ofs) & mask;
src_ofs += chunk;
if (src_ofs == 8) {
src--;
src_ofs = 0;
}
dst_ofs += chunk;
if (dst_ofs == 8) {
dst--;
dst_ofs = 0;
}
n_bits -= chunk;
}
}
}
/* Zeros the 'n_bits' bits starting from bit 'dst_ofs' in 'dst'. 'dst' is
* 'dst_len' bytes long.
*
* If you consider all of 'dst' to be a single unsigned integer in network byte
* order, then bit N is the bit with value 2**N. That is, bit 0 is the bit
* with value 1 in dst[dst_len - 1], bit 1 is the bit with value 2, bit 2 is
* the bit with value 4, ..., bit 8 is the bit with value 1 in dst[dst_len -
* 2], and so on.
*
* Required invariant:
* dst_ofs + n_bits <= dst_len * 8
*/
void
bitwise_zero(void *dst_, unsigned int dst_len, unsigned dst_ofs,
unsigned int n_bits)
{
uint8_t *dst = dst_;
if (!n_bits) {
return;
}
dst += dst_len - (dst_ofs / 8 + 1);
dst_ofs %= 8;
if (dst_ofs) {
unsigned int chunk = MIN(n_bits, 8 - dst_ofs);
*dst &= ~(((1 << chunk) - 1) << dst_ofs);
n_bits -= chunk;
if (!n_bits) {
return;
}
dst--;
}
while (n_bits >= 8) {
*dst-- = 0;
n_bits -= 8;
}
if (n_bits) {
*dst &= ~((1 << n_bits) - 1);
}
}
/* Sets to 1 all of the 'n_bits' bits starting from bit 'dst_ofs' in 'dst'.
* 'dst' is 'dst_len' bytes long.
*
* If you consider all of 'dst' to be a single unsigned integer in network byte
* order, then bit N is the bit with value 2**N. That is, bit 0 is the bit
* with value 1 in dst[dst_len - 1], bit 1 is the bit with value 2, bit 2 is
* the bit with value 4, ..., bit 8 is the bit with value 1 in dst[dst_len -
* 2], and so on.
*
* Required invariant:
* dst_ofs + n_bits <= dst_len * 8
*/
void
bitwise_one(void *dst_, unsigned int dst_len, unsigned dst_ofs,
unsigned int n_bits)
{
uint8_t *dst = dst_;
if (!n_bits) {
return;
}
dst += dst_len - (dst_ofs / 8 + 1);
dst_ofs %= 8;
if (dst_ofs) {
unsigned int chunk = MIN(n_bits, 8 - dst_ofs);
*dst |= ((1 << chunk) - 1) << dst_ofs;
n_bits -= chunk;
if (!n_bits) {
return;
}
dst--;
}
while (n_bits >= 8) {
*dst-- = 0xff;
n_bits -= 8;
}
if (n_bits) {
*dst |= (1 << n_bits) - 1;
}
}
/* Scans the 'n_bits' bits starting from bit 'dst_ofs' in 'dst' for 1-bits.
* Returns false if any 1-bits are found, otherwise true. 'dst' is 'dst_len'
* bytes long.
*
* If you consider all of 'dst' to be a single unsigned integer in network byte
* order, then bit N is the bit with value 2**N. That is, bit 0 is the bit
* with value 1 in dst[dst_len - 1], bit 1 is the bit with value 2, bit 2 is
* the bit with value 4, ..., bit 8 is the bit with value 1 in dst[dst_len -
* 2], and so on.
*
* Required invariant:
* dst_ofs + n_bits <= dst_len * 8
*/
bool
bitwise_is_all_zeros(const void *p_, unsigned int len, unsigned int ofs,
unsigned int n_bits)
{
const uint8_t *p = p_;
if (!n_bits) {
return true;
}
p += len - (ofs / 8 + 1);
ofs %= 8;
if (ofs) {
unsigned int chunk = MIN(n_bits, 8 - ofs);
if (*p & (((1 << chunk) - 1) << ofs)) {
return false;
}
n_bits -= chunk;
if (!n_bits) {
return true;
}
p--;
}
while (n_bits >= 8) {
if (*p) {
return false;
}
n_bits -= 8;
p--;
}
if (n_bits && *p & ((1 << n_bits) - 1)) {
return false;
}
return true;
}
/* Scans the bits in 'p' that have bit offsets 'start' (inclusive) through
* 'end' (exclusive) for the first bit with value 'target'. If one is found,
* returns its offset, otherwise 'end'. 'p' is 'len' bytes long.
*
* If you consider all of 'p' to be a single unsigned integer in network byte
* order, then bit N is the bit with value 2**N. That is, bit 0 is the bit
* with value 1 in p[len - 1], bit 1 is the bit with value 2, bit 2 is the bit
* with value 4, ..., bit 8 is the bit with value 1 in p[len - 2], and so on.
*
* Required invariant:
* start <= end
*/
unsigned int
bitwise_scan(const void *p, unsigned int len, bool target, unsigned int start,
unsigned int end)
{
unsigned int ofs;
for (ofs = start; ofs < end; ofs++) {
if (bitwise_get_bit(p, len, ofs) == target) {
break;
}
}
return ofs;
}
/* Scans the bits in 'p' that have bit offsets 'start' (inclusive) through
* 'end' (exclusive) for the first bit with value 'target', in reverse order.
* If one is found, returns its offset, otherwise 'end'. 'p' is 'len' bytes
* long.
*
* If you consider all of 'p' to be a single unsigned integer in network byte
* order, then bit N is the bit with value 2**N. That is, bit 0 is the bit
* with value 1 in p[len - 1], bit 1 is the bit with value 2, bit 2 is the bit
* with value 4, ..., bit 8 is the bit with value 1 in p[len - 2], and so on.
*
* To scan an entire bit array in reverse order, specify start == len * 8 - 1
* and end == -1, in which case the return value is nonnegative if successful
* and -1 if no 'target' match is found.
*
* Required invariant:
* start >= end
*/
int
bitwise_rscan(const void *p, unsigned int len, bool target, int start, int end)
{
const uint8_t *s = p;
int start_byte = len - (start / 8 + 1);
int end_byte = len - (end / 8 + 1);
int ofs_byte;
int ofs;
uint8_t the_byte;
/* Find the target in the start_byte from starting offset */
ofs_byte = start_byte;
the_byte = s[ofs_byte];
for (ofs = start % 8; ofs >= 0; ofs--) {
if (((the_byte & (1u << ofs)) != 0) == target) {
break;
}
}
if (ofs < 0) {
/* Target not found in start byte, continue searching byte by byte */
for (ofs_byte = start_byte + 1; ofs_byte <= end_byte; ofs_byte++) {
if ((target && s[ofs_byte])
|| (!target && (s[ofs_byte] != 0xff))) {
break;
}
}
if (ofs_byte > end_byte) {
return end;
}
the_byte = s[ofs_byte];
/* Target is in the_byte, find it bit by bit */
for (ofs = 7; ofs >= 0; ofs--) {
if (((the_byte & (1u << ofs)) != 0) == target) {
break;
}
}
}
int ret = (len - ofs_byte) * 8 - (8 - ofs);
if (ret < end) {
return end;
}
return ret;
}
/* Copies the 'n_bits' low-order bits of 'value' into the 'n_bits' bits
* starting at bit 'dst_ofs' in 'dst', which is 'dst_len' bytes long.
*
* If you consider all of 'dst' to be a single unsigned integer in network byte
* order, then bit N is the bit with value 2**N. That is, bit 0 is the bit
* with value 1 in dst[dst_len - 1], bit 1 is the bit with value 2, bit 2 is
* the bit with value 4, ..., bit 8 is the bit with value 1 in dst[dst_len -
* 2], and so on.
*
* Required invariants:
* dst_ofs + n_bits <= dst_len * 8
* n_bits <= 64
*/
void
bitwise_put(uint64_t value,
void *dst, unsigned int dst_len, unsigned int dst_ofs,
unsigned int n_bits)
{
ovs_be64 n_value = htonll(value);
bitwise_copy(&n_value, sizeof n_value, 0,
dst, dst_len, dst_ofs,
n_bits);
}
/* Returns the value of the 'n_bits' bits starting at bit 'src_ofs' in 'src',
* which is 'src_len' bytes long.
*
* If you consider all of 'src' to be a single unsigned integer in network byte
* order, then bit N is the bit with value 2**N. That is, bit 0 is the bit
* with value 1 in src[src_len - 1], bit 1 is the bit with value 2, bit 2 is
* the bit with value 4, ..., bit 8 is the bit with value 1 in src[src_len -
* 2], and so on.
*
* Required invariants:
* src_ofs + n_bits <= src_len * 8
* n_bits <= 64
*/
uint64_t
bitwise_get(const void *src, unsigned int src_len,
unsigned int src_ofs, unsigned int n_bits)
{
ovs_be64 value = htonll(0);
bitwise_copy(src, src_len, src_ofs,
&value, sizeof value, 0,
n_bits);
return ntohll(value);
}
/* Returns the value of the bit with offset 'ofs' in 'src', which is 'len'
* bytes long.
*
* If you consider all of 'src' to be a single unsigned integer in network byte
* order, then bit N is the bit with value 2**N. That is, bit 0 is the bit
* with value 1 in src[len - 1], bit 1 is the bit with value 2, bit 2 is the
* bit with value 4, ..., bit 8 is the bit with value 1 in src[len - 2], and so
* on.
*
* Required invariants:
* ofs < len * 8
*/
bool
bitwise_get_bit(const void *src_, unsigned int len, unsigned int ofs)
{
const uint8_t *src = src_;
return (src[len - (ofs / 8 + 1)] & (1u << (ofs % 8))) != 0;
}
/* Sets the bit with offset 'ofs' in 'dst', which is 'len' bytes long, to 0.
*
* If you consider all of 'dst' to be a single unsigned integer in network byte
* order, then bit N is the bit with value 2**N. That is, bit 0 is the bit
* with value 1 in dst[len - 1], bit 1 is the bit with value 2, bit 2 is the
* bit with value 4, ..., bit 8 is the bit with value 1 in dst[len - 2], and so
* on.
*
* Required invariants:
* ofs < len * 8
*/
void
bitwise_put0(void *dst_, unsigned int len, unsigned int ofs)
{
uint8_t *dst = dst_;
dst[len - (ofs / 8 + 1)] &= ~(1u << (ofs % 8));
}
/* Sets the bit with offset 'ofs' in 'dst', which is 'len' bytes long, to 1.
*
* If you consider all of 'dst' to be a single unsigned integer in network byte
* order, then bit N is the bit with value 2**N. That is, bit 0 is the bit
* with value 1 in dst[len - 1], bit 1 is the bit with value 2, bit 2 is the
* bit with value 4, ..., bit 8 is the bit with value 1 in dst[len - 2], and so
* on.
*
* Required invariants:
* ofs < len * 8
*/
void
bitwise_put1(void *dst_, unsigned int len, unsigned int ofs)
{
uint8_t *dst = dst_;
dst[len - (ofs / 8 + 1)] |= 1u << (ofs % 8);
}
/* Sets the bit with offset 'ofs' in 'dst', which is 'len' bytes long, to 'b'.
*
* If you consider all of 'dst' to be a single unsigned integer in network byte
* order, then bit N is the bit with value 2**N. That is, bit 0 is the bit
* with value 1 in dst[len - 1], bit 1 is the bit with value 2, bit 2 is the
* bit with value 4, ..., bit 8 is the bit with value 1 in dst[len - 2], and so
* on.
*
* Required invariants:
* ofs < len * 8
*/
void
bitwise_put_bit(void *dst, unsigned int len, unsigned int ofs, bool b)
{
if (b) {
bitwise_put1(dst, len, ofs);
} else {
bitwise_put0(dst, len, ofs);
}
}
/* Flips the bit with offset 'ofs' in 'dst', which is 'len' bytes long.
*
* If you consider all of 'dst' to be a single unsigned integer in network byte
* order, then bit N is the bit with value 2**N. That is, bit 0 is the bit
* with value 1 in dst[len - 1], bit 1 is the bit with value 2, bit 2 is the
* bit with value 4, ..., bit 8 is the bit with value 1 in dst[len - 2], and so
* on.
*
* Required invariants:
* ofs < len * 8
*/
void
bitwise_toggle_bit(void *dst_, unsigned int len, unsigned int ofs)
{
uint8_t *dst = dst_;
dst[len - (ofs / 8 + 1)] ^= 1u << (ofs % 8);
}
/* ovs_scan */
struct scan_spec {
unsigned int width;
enum {
SCAN_DISCARD,
SCAN_CHAR,
SCAN_SHORT,
SCAN_INT,
SCAN_LONG,
SCAN_LLONG,
SCAN_INTMAX_T,
SCAN_PTRDIFF_T,
SCAN_SIZE_T
} type;
};
static const char *
skip_spaces(const char *s)
{
while (isspace((unsigned char) *s)) {
s++;
}
return s;
}
static const char *
scan_int(const char *s, const struct scan_spec *spec, int base, va_list *args)
{
const char *start = s;
uintmax_t value;
bool negative;
int n_digits;
negative = *s == '-';
s += *s == '-' || *s == '+';
if ((!base || base == 16) && *s == '0' && (s[1] == 'x' || s[1] == 'X')) {
base = 16;
s += 2;
} else if (!base) {
base = *s == '0' ? 8 : 10;
}
if (s - start >= spec->width) {
return NULL;
}
value = 0;
n_digits = 0;
while (s - start < spec->width) {
int digit = hexit_value(*s);
if (digit < 0 || digit >= base) {
break;
}
value = value * base + digit;
n_digits++;
s++;
}
if (!n_digits) {
return NULL;
}
if (negative) {
value = -value;
}
switch (spec->type) {
case SCAN_DISCARD:
break;
case SCAN_CHAR:
*va_arg(*args, char *) = value;
break;
case SCAN_SHORT:
*va_arg(*args, short int *) = value;
break;
case SCAN_INT:
*va_arg(*args, int *) = value;
break;
case SCAN_LONG:
*va_arg(*args, long int *) = value;
break;
case SCAN_LLONG:
*va_arg(*args, long long int *) = value;
break;
case SCAN_INTMAX_T:
*va_arg(*args, intmax_t *) = value;
break;
case SCAN_PTRDIFF_T:
*va_arg(*args, ptrdiff_t *) = value;
break;
case SCAN_SIZE_T:
*va_arg(*args, size_t *) = value;
break;
}
return s;
}
static const char *
skip_digits(const char *s)
{
while (*s >= '0' && *s <= '9') {
s++;
}
return s;
}
static const char *
scan_float(const char *s, const struct scan_spec *spec, va_list *args)
{
const char *start = s;
long double value;
char *tail;
char *copy;
bool ok;
s += *s == '+' || *s == '-';
s = skip_digits(s);
if (*s == '.') {
s = skip_digits(s + 1);
}
if (*s == 'e' || *s == 'E') {
s++;
s += *s == '+' || *s == '-';
s = skip_digits(s);
}
if (s - start > spec->width) {
s = start + spec->width;
}
copy = xmemdup0(start, s - start);
value = strtold(copy, &tail);
ok = *tail == '\0';
free(copy);
if (!ok) {
return NULL;
}
switch (spec->type) {
case SCAN_DISCARD:
break;
case SCAN_INT:
*va_arg(*args, float *) = value;
break;
case SCAN_LONG:
*va_arg(*args, double *) = value;
break;
case SCAN_LLONG:
*va_arg(*args, long double *) = value;
break;
case SCAN_CHAR:
case SCAN_SHORT:
case SCAN_INTMAX_T:
case SCAN_PTRDIFF_T:
case SCAN_SIZE_T:
OVS_NOT_REACHED();
}
return s;
}
static void
scan_output_string(const struct scan_spec *spec,
const char *s, size_t n,
va_list *args)
{
if (spec->type != SCAN_DISCARD) {
char *out = va_arg(*args, char *);
memcpy(out, s, n);
out[n] = '\0';
}
}
static const char *
scan_string(const char *s, const struct scan_spec *spec, va_list *args)
{
size_t n;
for (n = 0; n < spec->width; n++) {
if (!s[n] || isspace((unsigned char) s[n])) {
break;
}
}
if (!n) {
return NULL;
}
scan_output_string(spec, s, n, args);
return s + n;
}
static const char *
parse_scanset(const char *p_, unsigned long *set, bool *complemented)
{
const uint8_t *p = (const uint8_t *) p_;
*complemented = *p == '^';
p += *complemented;
if (*p == ']') {
bitmap_set1(set, ']');
p++;
}
while (*p && *p != ']') {
if (p[1] == '-' && p[2] != ']' && p[2] > *p) {
bitmap_set_multiple(set, *p, p[2] - *p + 1, true);
p += 3;
} else {
bitmap_set1(set, *p++);
}
}
if (*p == ']') {
p++;
}
return (const char *) p;
}
static const char *
scan_set(const char *s, const struct scan_spec *spec, const char **pp,
va_list *args)
{
unsigned long set[BITMAP_N_LONGS(UCHAR_MAX + 1)];
bool complemented;
unsigned int n;
/* Parse the scan set. */
memset(set, 0, sizeof set);
*pp = parse_scanset(*pp, set, &complemented);
/* Parse the data. */
n = 0;
while (s[n]
&& bitmap_is_set(set, (unsigned char) s[n]) == !complemented
&& n < spec->width) {
n++;
}
if (!n) {
return NULL;
}
scan_output_string(spec, s, n, args);
return s + n;
}
static const char *
scan_chars(const char *s, const struct scan_spec *spec, va_list *args)
{
unsigned int n = spec->width == UINT_MAX ? 1 : spec->width;
if (strlen(s) < n) {
return NULL;
}
if (spec->type != SCAN_DISCARD) {
memcpy(va_arg(*args, char *), s, n);
}
return s + n;
}
static bool
ovs_scan__(const char *s, int *n, const char *format, va_list *args)
{
const char *const start = s;
bool ok = false;
const char *p;
p = format;
while (*p != '\0') {
struct scan_spec spec;
unsigned char c = *p++;
bool discard;
if (isspace(c)) {
s = skip_spaces(s);
continue;
} else if (c != '%') {
if (*s != c) {
goto exit;
}
s++;
continue;
} else if (*p == '%') {
if (*s++ != '%') {
goto exit;
}
p++;
continue;
}
/* Parse '*' flag. */
discard = *p == '*';
p += discard;
/* Parse field width. */
spec.width = 0;
while (*p >= '0' && *p <= '9') {
spec.width = spec.width * 10 + (*p++ - '0');
}
if (spec.width == 0) {
spec.width = UINT_MAX;
}
/* Parse type modifier. */
switch (*p) {
case 'h':
if (p[1] == 'h') {
spec.type = SCAN_CHAR;
p += 2;
} else {
spec.type = SCAN_SHORT;
p++;
}
break;
case 'j':
spec.type = SCAN_INTMAX_T;
p++;
break;
case 'l':
if (p[1] == 'l') {
spec.type = SCAN_LLONG;
p += 2;
} else {
spec.type = SCAN_LONG;
p++;
}
break;
case 'L':
case 'q':
spec.type = SCAN_LLONG;
p++;
break;
case 't':
spec.type = SCAN_PTRDIFF_T;
p++;
break;
case 'z':
spec.type = SCAN_SIZE_T;
p++;
break;
default:
spec.type = SCAN_INT;
break;
}
if (discard) {
spec.type = SCAN_DISCARD;
}
c = *p++;
if (c != 'c' && c != 'n' && c != '[') {
s = skip_spaces(s);
}
switch (c) {
case 'd':
s = scan_int(s, &spec, 10, args);
break;
case 'i':
s = scan_int(s, &spec, 0, args);
break;
case 'o':
s = scan_int(s, &spec, 8, args);
break;
case 'u':
s = scan_int(s, &spec, 10, args);
break;
case 'x':
case 'X':
s = scan_int(s, &spec, 16, args);
break;
case 'e':
case 'f':
case 'g':
case 'E':
case 'G':
s = scan_float(s, &spec, args);
break;
case 's':
s = scan_string(s, &spec, args);
break;
case '[':
s = scan_set(s, &spec, &p, args);
break;
case 'c':
s = scan_chars(s, &spec, args);
break;
case 'n':
if (spec.type != SCAN_DISCARD) {
*va_arg(*args, int *) = s - start;
}
break;
}
if (!s) {
goto exit;
}
}
if (n) {
*n = s - start;
}
ok = true;
exit:
return ok;
}
/* This is an implementation of the standard sscanf() function, with the
* following exceptions:
*
* - It returns true if the entire format was successfully scanned and
* converted, false if any conversion failed.
*
* - The standard doesn't define sscanf() behavior when an out-of-range value
* is scanned, e.g. if a "%"PRIi8 conversion scans "-1" or "0x1ff". Some
* implementations consider this an error and stop scanning. This
* implementation never considers an out-of-range value an error; instead,
* it stores the least-significant bits of the converted value in the
* destination, e.g. the value 255 for both examples earlier.
*
* - Only single-byte characters are supported, that is, the 'l' modifier
* on %s, %[, and %c is not supported. The GNU extension 'a' modifier is
* also not supported.
*
* - %p is not supported.
*/
bool
ovs_scan(const char *s, const char *format, ...)
{
va_list args;
bool res;
va_start(args, format);
res = ovs_scan__(s, NULL, format, &args);
va_end(args);
return res;
}
/*
* This function is similar to ovs_scan(), with an extra parameter `n` added to
* return the number of scanned characters.
*/
bool
ovs_scan_len(const char *s, int *n, const char *format, ...)
{
va_list args;
bool success;
int n1;
va_start(args, format);
success = ovs_scan__(s + *n, &n1, format, &args);
va_end(args);
if (success) {
*n = *n + n1;
}
return success;
}
void
xsleep(unsigned int seconds)
{
ovsrcu_quiesce_start();
#ifdef _WIN32
Sleep(seconds * 1000);
#else
sleep(seconds);
#endif
ovsrcu_quiesce_end();
}
static void
xnanosleep__(uint64_t nanoseconds)
{
#ifndef _WIN32
int retval;
struct timespec ts_sleep;
nsec_to_timespec(nanoseconds, &ts_sleep);
int error = 0;
do {
retval = nanosleep(&ts_sleep, NULL);
error = retval < 0 ? errno : 0;
} while (error == EINTR);
#else
HANDLE timer = CreateWaitableTimer(NULL, FALSE, NULL);
if (timer) {
LARGE_INTEGER duetime;
duetime.QuadPart = -nanoseconds;
if (SetWaitableTimer(timer, &duetime, 0, NULL, NULL, FALSE)) {
WaitForSingleObject(timer, INFINITE);
} else {
VLOG_ERR_ONCE("SetWaitableTimer Failed (%s)",
ovs_lasterror_to_string());
}
CloseHandle(timer);
} else {
VLOG_ERR_ONCE("CreateWaitableTimer Failed (%s)",
ovs_lasterror_to_string());
}
#endif
}
/* High resolution sleep with thread quiesce. */
void
xnanosleep(uint64_t nanoseconds)
{
ovsrcu_quiesce_start();
xnanosleep__(nanoseconds);
ovsrcu_quiesce_end();
}
/* High resolution sleep without thread quiesce. */
void
xnanosleep_no_quiesce(uint64_t nanoseconds)
{
xnanosleep__(nanoseconds);
}
#if __linux__
void
set_timer_resolution(unsigned long nanoseconds)
{
prctl(PR_SET_TIMERSLACK, nanoseconds);
}
#else
void
set_timer_resolution(unsigned long nanoseconds OVS_UNUSED)
{
}
#endif
/* Determine whether standard output is a tty or not. This is useful to decide
* whether to use color output or not when --color option for utilities is set
* to `auto`.
*/
bool
is_stdout_a_tty(void)
{
char const *t = getenv("TERM");
return (isatty(STDOUT_FILENO) && t && strcmp(t, "dumb") != 0);
}
#ifdef _WIN32
char *
ovs_format_message(int error)
{
enum { BUFSIZE = sizeof strerror_buffer_get()->s };
char *buffer = strerror_buffer_get()->s;
if (error == 0) {
/* See ovs_strerror */
return "Success";
}
FormatMessage(FORMAT_MESSAGE_FROM_SYSTEM | FORMAT_MESSAGE_IGNORE_INSERTS,
NULL, error, 0, buffer, BUFSIZE, NULL);
return buffer;
}
/* Returns a null-terminated string that explains the last error.
* Use this function to get the error string for WINAPI calls. */
char *
ovs_lasterror_to_string(void)
{
return ovs_format_message(GetLastError());
}
int
ftruncate(int fd, off_t length)
{
int error;
error = _chsize_s(fd, length);
if (error) {
return -1;
}
return 0;
}
OVS_CONSTRUCTOR(winsock_start) {
WSADATA wsaData;
int error;
error = WSAStartup(MAKEWORD(2, 2), &wsaData);
if (error != 0) {
VLOG_FATAL("WSAStartup failed: %s", sock_strerror(sock_errno()));
}
}
#endif
#ifdef __linux__
bool
ovs_kernel_is_version_or_newer(int target_major, int target_minor)
{
static struct ovsthread_once once = OVSTHREAD_ONCE_INITIALIZER;
static int current_major, current_minor = -1;
if (ovsthread_once_start(&once)) {
struct utsname utsname;
if (uname(&utsname) == -1) {
VLOG_WARN("uname failed (%s)", ovs_strerror(errno));
} else if (!ovs_scan(utsname.release, "%d.%d",
&current_major, &current_minor)) {
VLOG_WARN("uname reported bad OS release (%s)", utsname.release);
}
ovsthread_once_done(&once);
}
if (current_major == -1 || current_minor == -1) {
return false;
}
return current_major > target_major || (
current_major == target_major && current_minor >= target_minor);
}
#endif