- html - 出于某种原因,IE8 对我的 Sass 文件中继承的 html5 CSS 不友好?
- JMeter 在响应断言中使用 span 标签的问题
- html - 在 :hover and :active? 上具有不同效果的 CSS 动画
- html - 相对于居中的 html 内容固定的 CSS 重复背景?
我试图了解内存分配如何在 fork 上工作,即使在静态或动态分配上也是如此。我很难理解一些泄漏,如下所示。
有了这个程序:
#include <unistd.h>
#include <sys/wait.h>
int main(int argc, char **argv)
{
pid_t pid;
int status;
pid = fork();
if (pid == 0)
return (0);
else
{
waitpid(pid, &status, 0);
return (0);
}
}
我正在使用 valgrind 获取此日志文件:
==81268== Memcheck, a memory error detector
==81268== Copyright (C) 2002-2017, and GNU GPL'd, by Julian Seward et al.
==81268== Using Valgrind-3.13.0 and LibVEX; rerun with -h for copyright info
==81268== Command: ./a.out
==81268== Parent PID: 73556
==81268==
==81268== Syscall param msg->desc.port.name points to uninitialised byte(s)
==81268== at 0x1003A834A: mach_msg_trap (in /usr/lib/system/libsystem_kernel.dylib)
==81268== by 0x1003A7796: mach_msg (in /usr/lib/system/libsystem_kernel.dylib)
==81268== by 0x1003A1485: task_set_special_port (in /usr/lib/system/libsystem_kernel.dylib)
==81268== by 0x10053D10E: _os_trace_create_debug_control_port (in /usr/lib/system/libsystem_trace.dylib)
==81268== by 0x10053D458: _libtrace_init (in /usr/lib/system/libsystem_trace.dylib)
==81268== by 0x1000A69DF: libSystem_initializer (in /usr/lib/libSystem.B.dylib)
==81268== by 0x100018A1A: ImageLoaderMachO::doModInitFunctions(ImageLoader::LinkContext const&) (in /usr/lib/dyld)
==81268== by 0x100018C1D: ImageLoaderMachO::doInitialization(ImageLoader::LinkContext const&) (in /usr/lib/dyld)
==81268== by 0x1000144A9: ImageLoader::recursiveInitialization(ImageLoader::LinkContext const&, unsigned int, char const*, ImageLoader::InitializerTimingList&, ImageLoader::UninitedUpwards&) (in /usr/lib/dyld)
==81268== by 0x100014440: ImageLoader::recursiveInitialization(ImageLoader::LinkContext const&, unsigned int, char const*, ImageLoader::InitializerTimingList&, ImageLoader::UninitedUpwards&) (in /usr/lib/dyld)
==81268== by 0x100013523: ImageLoader::processInitializers(ImageLoader::LinkContext const&, unsigned int, ImageLoader::InitializerTimingList&, ImageLoader::UninitedUpwards&) (in /usr/lib/dyld)
==81268== by 0x1000135B8: ImageLoader::runInitializers(ImageLoader::LinkContext const&, ImageLoader::InitializerTimingList&) (in /usr/lib/dyld)
==81268== Address 0x10488bc6c is on thread 1's stack
==81268== in frame #2, created by task_set_special_port (???:)
==81268==
==81269==
==81269== HEAP SUMMARY:
==81269== in use at exit: 28,557 bytes in 172 blocks
==81269== total heap usage: 188 allocs, 16 frees, 34,701 bytes allocated
==81269==
==81269== 72 bytes in 3 blocks are possibly lost in loss record 31 of 55
==81269== at 0x100098BF2: calloc (in /Users/simrossi/.brew/Cellar/valgrind/3.13.0/lib/valgrind/vgpreload_memcheck-amd64-darwin.so)
==81269== by 0x1005B1846: map_images_nolock (in /usr/lib/libobjc.A.dylib)
==81269== by 0x1005C4FE8: objc_object::sidetable_retainCount() (in /usr/lib/libobjc.A.dylib)
==81269== by 0x10000803B: dyld::notifyBatchPartial(dyld_image_states, bool, char const* (*)(dyld_image_states, unsigned int, dyld_image_info const*), bool, bool) (in /usr/lib/dyld)
==81269== by 0x100008255: dyld::registerObjCNotifiers(void (*)(unsigned int, char const* const*, mach_header const* const*), void (*)(char const*, mach_header const*), void (*)(char const*, mach_header const*)) (in /usr/lib/dyld)
==81269== by 0x1001FF00A: _dyld_objc_notify_register (in /usr/lib/system/libdyld.dylib)
==81269== by 0x1005B1074: _objc_init (in /usr/lib/libobjc.A.dylib)
==81269== by 0x10019268D: _os_object_init (in /usr/lib/system/libdispatch.dylib)
==81269== by 0x10019263A: libdispatch_init (in /usr/lib/system/libdispatch.dylib)
==81269== by 0x1000A69D5: libSystem_initializer (in /usr/lib/libSystem.B.dylib)
==81269== by 0x100018A1A: ImageLoaderMachO::doModInitFunctions(ImageLoader::LinkContext const&) (in /usr/lib/dyld)
==81269== by 0x100018C1D: ImageLoaderMachO::doInitialization(ImageLoader::LinkContext const&) (in /usr/lib/dyld)
==81269==
==81269== 2,064 (16 direct, 2,048 indirect) bytes in 1 blocks are definitely lost in loss record 51 of 55
==81269== at 0x100098586: malloc (in /Users/simrossi/.brew/Cellar/valgrind/3.13.0/lib/valgrind/vgpreload_memcheck-amd64-darwin.so)
==81269== by 0x1004E9103: _nc_table_new (in /usr/lib/system/libsystem_notify.dylib)
==81269== by 0x1004E4FB5: _notify_init_globals (in /usr/lib/system/libsystem_notify.dylib)
==81269== by 0x1004F2AC0: _os_once (in /usr/lib/system/libsystem_platform.dylib)
==81269== by 0x1004F2A91: _os_alloc_once (in /usr/lib/system/libsystem_platform.dylib)
==81269== by 0x1004E50AD: _notify_fork_child (in /usr/lib/system/libsystem_notify.dylib)
==81269== by 0x1000A6B20: libSystem_atfork_child (in /usr/lib/libSystem.B.dylib)
==81269== by 0x10026F436: fork (in /usr/lib/system/libsystem_c.dylib)
==81269== by 0x100000F4A: main (in ./a.out)
==81269==
==81269== 2,064 (16 direct, 2,048 indirect) bytes in 1 blocks are definitely lost in loss record 52 of 55
==81269== at 0x100098586: malloc (in /Users/simrossi/.brew/Cellar/valgrind/3.13.0/lib/valgrind/vgpreload_memcheck-amd64-darwin.so)
==81269== by 0x1004E9103: _nc_table_new (in /usr/lib/system/libsystem_notify.dylib)
==81269== by 0x1004E4FC6: _notify_init_globals (in /usr/lib/system/libsystem_notify.dylib)
==81269== by 0x1004F2AC0: _os_once (in /usr/lib/system/libsystem_platform.dylib)
==81269== by 0x1004F2A91: _os_alloc_once (in /usr/lib/system/libsystem_platform.dylib)
==81269== by 0x1004E50AD: _notify_fork_child (in /usr/lib/system/libsystem_notify.dylib)
==81269== by 0x1000A6B20: libSystem_atfork_child (in /usr/lib/libSystem.B.dylib)
==81269== by 0x10026F436: fork (in /usr/lib/system/libsystem_c.dylib)
==81269== by 0x100000F4A: main (in ./a.out)
==81269==
==81269== LEAK SUMMARY:
==81269== definitely lost: 32 bytes in 2 blocks
==81269== indirectly lost: 4,096 bytes in 2 blocks
==81269== possibly lost: 72 bytes in 3 blocks
==81269== still reachable: 200 bytes in 6 blocks
==81269== suppressed: 24,157 bytes in 159 blocks
==81269== Reachable blocks (those to which a pointer was found) are not shown.
==81269== To see them, rerun with: --leak-check=full --show-leak-kinds=all
==81269==
==81269== For counts of detected and suppressed errors, rerun with: -v
==81269== Use --track-origins=yes to see where uninitialised values come from
==81269== ERROR SUMMARY: 5 errors from 4 contexts (suppressed: 13 from 13)
==81268==
==81268== HEAP SUMMARY:
==81268== in use at exit: 18,127 bytes in 159 blocks
==81268== total heap usage: 175 allocs, 16 frees, 24,271 bytes allocated
==81268==
==81268== 72 bytes in 3 blocks are possibly lost in loss record 26 of 42
==81268== at 0x100098BF2: calloc (in /Users/simrossi/.brew/Cellar/valgrind/3.13.0/lib/valgrind/vgpreload_memcheck-amd64-darwin.so)
==81268== by 0x1005B1846: map_images_nolock (in /usr/lib/libobjc.A.dylib)
==81268== by 0x1005C4FE8: objc_object::sidetable_retainCount() (in /usr/lib/libobjc.A.dylib)
==81268== by 0x10000803B: dyld::notifyBatchPartial(dyld_image_states, bool, char const* (*)(dyld_image_states, unsigned int, dyld_image_info const*), bool, bool) (in /usr/lib/dyld)
==81268== by 0x100008255: dyld::registerObjCNotifiers(void (*)(unsigned int, char const* const*, mach_header const* const*), void (*)(char const*, mach_header const*), void (*)(char const*, mach_header const*)) (in /usr/lib/dyld)
==81268== by 0x1001FF00A: _dyld_objc_notify_register (in /usr/lib/system/libdyld.dylib)
==81268== by 0x1005B1074: _objc_init (in /usr/lib/libobjc.A.dylib)
==81268== by 0x10019268D: _os_object_init (in /usr/lib/system/libdispatch.dylib)
==81268== by 0x10019263A: libdispatch_init (in /usr/lib/system/libdispatch.dylib)
==81268== by 0x1000A69D5: libSystem_initializer (in /usr/lib/libSystem.B.dylib)
==81268== by 0x100018A1A: ImageLoaderMachO::doModInitFunctions(ImageLoader::LinkContext const&) (in /usr/lib/dyld)
==81268== by 0x100018C1D: ImageLoaderMachO::doInitialization(ImageLoader::LinkContext const&) (in /usr/lib/dyld)
==81268==
==81268== LEAK SUMMARY:
==81268== definitely lost: 0 bytes in 0 blocks
==81268== indirectly lost: 0 bytes in 0 blocks
==81268== possibly lost: 72 bytes in 3 blocks
==81268== still reachable: 200 bytes in 6 blocks
==81268== suppressed: 17,855 bytes in 150 blocks
==81268== Reachable blocks (those to which a pointer was found) are not shown.
==81268== To see them, rerun with: --leak-check=full --show-leak-kinds=all
==81268==
==81268== For counts of detected and suppressed errors, rerun with: -v
==81268== Use --track-origins=yes to see where uninitialised values come from
==81268== ERROR SUMMARY: 2 errors from 2 contexts (suppressed: 12 from 12)
我不明白泄漏的那些 32 字节和 4096 字节是从哪里来的?它来自标准库吗,我在 stackoverflow 上看到了类似的东西。我在 Mac OS Sierra x86_64 上。
编辑:如果我不 fork ,我不会有任何泄漏,它似乎与函数调用有关。
最佳答案
macOS 上程序的启动代码分配了大量从未释放的内存。对于正在分配的内存,您无能为力。但是,您可以将泄漏告诉 Valgrind 并说服它忽略它们。我相信 fork()
是无关紧要的 — 即使没有它,您也可能会看到泄漏。
您需要运行这样的程序,并带有生成抑制的选项,在文件中捕获抑制,然后(将来)运行指定您创建的抑制文件。
您可以在 Valgrind shows memory leak with empty main without including headers 看到一个有效的例子.
关于C,没有 malloc 的 fork 泄漏,我们在Stack Overflow上找到一个类似的问题: https://stackoverflow.com/questions/54671093/
我是 C 的新手,在 Linux 中使用带有开关 gcc -g -std=c89 -Wall ... 的 gcc4.4.6 进行编程,我在许多函数深处遇到了这个错误我的程序名为 compute: **
今天阅读Rust subreddit时,我发现以下评论: jemalloc针对(多线程)速度而不是内存使用进行了优化 经过更多研究后,我发现还有更多选择(例如calloc)。 我想了解不同内存分配器的
相关代码: write(-1, "test", sizeof("test")); void * p = malloc(1024); void * p2 = malloc(510); w
我正在比较不同的 malloc 实现,我想比较它们的运行时间和内存使用情况。 特别是,我对运行时和最大常驻内存感兴趣。重要的是最大常驻内存将是真实的(没有代码段等)。 我不能使用像 valgrind
我承认这三个都有不同的含义。但是,我不明白这些具体情况适用于哪些特定情况。任何人都可以分享每个例子吗?谢谢。 malloc(sizeof(int)) malloc(size
GLib 文档推荐使用 GLib Slice Allocator 而不是 malloc: "For newly written code it is recommended to use the ne
我正在分配一个字符串 int main(){ int buf = 1024; char *input = malloc(sizeof(char*) * buf); //CODE
Here有一个关于 malloc 包的环境变量列表: MallocStackLogging MallocStackLoggingNoCompact MallocPreScribble MallocSc
总体问题:当您将通过malloc分配的返回值分配给一个指针时,您是否需要malloc该指针以及,还是您可以简单地声明并分配它? 例如,假设我有一个函数 foo,它在执行过程中使用 malloc 创建了
这个问题在这里已经有了答案: String assignment in C (4 个答案) 关闭 7 年前。 这是有问题的片段。 int main() { char** RESERV = (
任务是将一个二进制文件解析到内存中。但是,我事先不知道需要分配的内存量。 哪种方法更可取:在解析例程中进行多个小 malloc,或者首先遍历文件以确定所需的内存量,然后再次解析? 感谢任何提示。 最佳
我最近一直在尝试理解严格别名的一个特定方面,我认为我已经制作了尽可能最小的有趣代码。 (对我来说很有趣,就是这样!) 更新:根据到目前为止的答案,很明显我需要澄清这个问题。从某个角度来看,这里的第一个
我一直在为我创建的一个简单程序创建测试。我总是使用类似这样的方法检查使用 malloc 分配内存是否失败 int* ptr = malloc(sizeof(int) * x); if(!ptr){
我是 malloc 和对齐 malloc 的新手。我知道如何使用它们。但是,我不知道在什么情况下我们应该使用对齐的 malloc 而不是标准的 malloc。你能给我解释一下吗? 最佳答案 glibc
这样分配内存是不好的做法吗?: FOO *foo; while (!(foo = malloc(sizeof(FOO)))) ; 最佳答案 我不知道有什么不好的做法,但这种情况并不常见。 malloc
有人可以向我解释使用和不使用 malloc 创建结构之间的区别吗?什么时候应该使用 malloc,什么时候应该使用常规初始化? 例如: struct person { char* name;
假设我有一个类型 node_t typedef struct node{ char* value; struct node *next; }node_t; 当我想创建一个名为 n1 的
关闭。这个问题需要更多 focused .它目前不接受答案。 想改进这个问题?更新问题,使其仅关注一个问题 editing this post . 3年前关闭。 Improve this questi
我对指针感到困惑。这是交换两个名称的代码。请看代码。考虑输入:hellohai(对于 d)和 asd(对于 e)。我得到的输出:asd 1ellohai 1ellohai #include #incl
我已经编写了这个函数(如下)。它应该逐行读取文件。编辑该行并将某些单词/字符放入各种功能中。然后将这些函数放入“entrant”结构的数组(malloc)中。 问题是,当我退出循环并尝试打印数组时,放
我是一名优秀的程序员,十分优秀!