楊志璿
    • Create new note
    • Create a note from template
      • Sharing URL Link copied
      • /edit
      • View mode
        • Edit mode
        • View mode
        • Book mode
        • Slide mode
        Edit mode View mode Book mode Slide mode
      • Customize slides
      • Note Permission
      • Read
        • Only me
        • Signed-in users
        • Everyone
        Only me Signed-in users Everyone
      • Write
        • Only me
        • Signed-in users
        • Everyone
        Only me Signed-in users Everyone
      • Engagement control Commenting, Suggest edit, Emoji Reply
    • Invite by email
      Invitee

      This note has no invitees

    • Publish Note

      Share your work with the world Congratulations! 🎉 Your note is out in the world Publish Note

      Your note will be visible on your profile and discoverable by anyone.
      Your note is now live.
      This note is visible on your profile and discoverable online.
      Everyone on the web can find and read all notes of this public team.
      See published notes
      Unpublish note
      Please check the box to agree to the Community Guidelines.
      View profile
    • Commenting
      Permission
      Disabled Forbidden Owners Signed-in users Everyone
    • Enable
    • Permission
      • Forbidden
      • Owners
      • Signed-in users
      • Everyone
    • Suggest edit
      Permission
      Disabled Forbidden Owners Signed-in users Everyone
    • Enable
    • Permission
      • Forbidden
      • Owners
      • Signed-in users
    • Emoji Reply
    • Enable
    • Versions and GitHub Sync
    • Note settings
    • Note Insights New
    • Engagement control
    • Transfer ownership
    • Delete this note
    • Save as template
    • Insert from template
    • Import from
      • Dropbox
      • Google Drive
      • Gist
      • Clipboard
    • Export to
      • Dropbox
      • Google Drive
      • Gist
    • Download
      • Markdown
      • HTML
      • Raw HTML
Menu Note settings Note Insights Versions and GitHub Sync Sharing URL Create Help
Create Create new note Create a note from template
Menu
Options
Engagement control Transfer ownership Delete this note
Import from
Dropbox Google Drive Gist Clipboard
Export to
Dropbox Google Drive Gist
Download
Markdown HTML Raw HTML
Back
Sharing URL Link copied
/edit
View mode
  • Edit mode
  • View mode
  • Book mode
  • Slide mode
Edit mode View mode Book mode Slide mode
Customize slides
Note Permission
Read
Only me
  • Only me
  • Signed-in users
  • Everyone
Only me Signed-in users Everyone
Write
Only me
  • Only me
  • Signed-in users
  • Everyone
Only me Signed-in users Everyone
Engagement control Commenting, Suggest edit, Emoji Reply
  • Invite by email
    Invitee

    This note has no invitees

  • Publish Note

    Share your work with the world Congratulations! 🎉 Your note is out in the world Publish Note

    Your note will be visible on your profile and discoverable by anyone.
    Your note is now live.
    This note is visible on your profile and discoverable online.
    Everyone on the web can find and read all notes of this public team.
    See published notes
    Unpublish note
    Please check the box to agree to the Community Guidelines.
    View profile
    Engagement control
    Commenting
    Permission
    Disabled Forbidden Owners Signed-in users Everyone
    Enable
    Permission
    • Forbidden
    • Owners
    • Signed-in users
    • Everyone
    Suggest edit
    Permission
    Disabled Forbidden Owners Signed-in users Everyone
    Enable
    Permission
    • Forbidden
    • Owners
    • Signed-in users
    Emoji Reply
    Enable
    Import from Dropbox Google Drive Gist Clipboard
       Owned this note    Owned this note      
    Published Linked with GitHub
    • Any changes
      Be notified of any changes
    • Mention me
      Be notified of mention me
    • Unsubscribe
    # 2020q1 Homework1 (lab0) contributed by < [25077667](https://github.com/25077667) > :::danger > [name=25077667]編輯中... > 人在做,Google/GitHub 在看,不用特別強調你的閒置,快去動手! > :notes: jserv ::: :::info 作業要求:[【連結】](https://hackmd.io/JSFDnHn0Rpe0V7d-AqLAXA?view#-%E4%BD%9C%E6%A5%AD%E8%A6%81%E6%B1%82) ::: # 開發環境 `$ neofetch` ![img](https://i.imgur.com/LH6fsQJ.png) # 實作思路 ### q_new 很明顯,就是要 `new` 一個 queue,像這樣 ```cpp= queue_t *q_new() { queue_t *q = malloc(sizeof(queue_t)); if (q == NULL) return NULL; // bad alloc memset(q, 0, sizeof(queue_t)); return q; } ``` 可是呢~ 我原本其實是想用 calloc,像下面這樣,但是會: > ERROR: Attempted to free unallocated block. Address = 0x7ffff4e93890 > ```cpp= queue_t *q_new() { return calloc(1, sizeof(queue_t)); } ``` ::: warning 接下來這邊要先看過[追蹤記憶體配置和釋放的狀況](https://hackmd.io/JSFDnHn0Rpe0V7d-AqLAXA?view#%E8%BF%BD%E8%B9%A4%E8%A8%98%E6%86%B6%E9%AB%94%E9%85%8D%E7%BD%AE%E5%92%8C%E9%87%8B%E6%94%BE%E7%9A%84%E7%8B%80%E6%B3%81)才能知道我說 function hooking 在 `harness.c` 的部分 ::: 原本以為是 function hooking 沒改,所以會錯,但是進去看了之後似乎不是。因為報錯是 "Attempted to free unallocated block...",而這是在 `harness.c` 的 `find_header()` 的這幾行描述的: ```cpp= if (cautious_mode) { /* Make sure this is really an allocated block */ block_ele_t *ab = allocated; bool found = false; while (ab && !found) { found = ab == b; ab = ab->next; } if (!found) { report_event(MSG_ERROR, "Attempted to free unallocated block. Address = %p", p); error_occurred = true; } } ``` 很明顯,這是線性走訪所有 allocated heap 去找有沒有申請過這塊記憶體。 可是當我們回顧到 `harness.c` 的 `test_calloc()` 的 line 7 ```cpp= void *test_calloc(size_t nelem, size_t elsize) { /* Reference: Malloc tutorial * https://danluu.com/malloc-tutorial/ */ size_t size = nelem * elsize; // TODO: check for overflow void *ptr = test_malloc(size); if(!ptr) return NULL; memset(ptr, 0, size); return ptr; } ``` 同樣是走 `test_malloc` 就是不知道為啥沒有把 `test_calloc` 的 address 放到 `allocated` 裡面去。不然怎麼會找不到呢? >[name=25077667]所以就先 malloc + memset 湊合著用,先解決其他問題在深入去追 code ### q_free 因為太直觀,所以就不放 code 了。 就是一個一個 free,不然還有其他辦法嗎? ### q_insert_head 跟下面 `q_insert_tail` 很像。 ```cpp= bool q_insert_head(queue_t *q, char *s) { if (!q) return false; list_ele_t *newh = malloc(sizeof(list_ele_t)); if (!newh) return false; memset(newh, 0, sizeof(list_ele_t)); if (!(newh->value = strdup(s))) { free(newh); return false; } newh->next = q->head; q->head = newh; if (!q->tail) q->tail = newh; q->size++; return true; } ``` ### q_insert_tail 總之就是要 $O(1)$ 完成插入。 目前程式碼如下 ```cpp= bool q_insert_tail(queue_t *q, char *s) { if (!q) return false; list_ele_t *newTail = malloc(sizeof(list_ele_t)); if (!newTail) return false; memset(newTail, 0, sizeof(list_ele_t)); if (!(newTail->value = strdup(s))) { free(newTail); return false; } if (q->tail != NULL) q->tail->next = newTail; q->tail = newTail; if (q->head == NULL) q->head = newTail; q->size++; return true; } ``` 第 6 行,一開始是使用 calloc,因為 `newTail->next` 要設成 `NULL`,NULL本質上也是 0,所以就打算 calloc = malloc + memset(0)。 但是問題同上 `q_new` 出現: > ERROR: Attempted to free unallocated or corrupted block. Address = 0x7fffbae30820 > 這東西。 再來這邊還可以優化的地方,大概是把他跟 q_insert_head 寫再一起(當 `q == NULL` 或加入一個新節點時: line 6-13 )。 ### q_remove_head ```cpp= bool q_remove_head(queue_t *q, char *sp, size_t bufsize) { if (q == NULL || !q->head || sp == NULL) return false; list_ele_t *toBeDeleted = q->head; memset(sp, 0, bufsize); strncpy(sp, toBeDeleted->value, bufsize - 1); q->head = q->head->next; q->size--; free(toBeDeleted->value); free(toBeDeleted); return true; } ``` 原本我 line 8 是用 memcpy,可是跑 Valgrind 的時候,被告知越界讀取,我還不是很懂為什麼,我猜是 memcpy 會往後多看一格(這應該是舉燭吧) :::warning 工程人員用字遣詞應當精準! :notes: jserv >[name=25077667] fixed ::: 原本寫法在這 [commit](https://github.com/25077667/lab0-c/commit/9da5ccc1ed96ef65d9266e648500408f569c4a1e#diff-a5618994268beaa9fba3b5fc78562a8b),現在看一看覺得我 commit 名子寫錯,這才不是 buffer overflow,這是 Out-of-Bounds Reading。 ### q_size 為了要 $O(1)$,就多放一個元素在 `queue_t` 叫做 size 存起來。 ### q_reverse 就是動一動指標最快,所以就直接寫了。 寫完後覺得好像還可以改善,我想了一下下,就發現 `q->tail->next` 可以當作 cache 用,所以變成下面這樣子。恩這也很顯而易見。 ```cpp= void q_reverse(queue_t *q) { if (q == NULL || !q->size) return; list_ele_t *current = q->head->next; q->tail = q->head; while (current) { /* * Use q->tail->next to be the cache. * Because q->tail->next is not using, and it can reduce declare a * variable. */ q->tail->next = current->next; current->next = q->head; q->head = current; current = q->tail->next; } } ``` ### q_sort 這 `q_sort` 不是 `quick sort` 而是對 `queue sort`。 遇到的問題是 Stack Overflow,因為我原本 merge 的寫法是: >list1->next = merge(list1->next, list2); > 而這肯定是要遍歷完全部節點在 stack 上。 每一次 recursive function call 都要 push RBP,所以,會有很大的需求在 stack size 上。 (這是我在 Facebook 群組的[討論串](https://www.facebook.com/groups/system.software2020/permalink/325709451696594/),感謝大家花時間教我><) 改善方式:==把 `merge` function 換成 iterative 的方式== 目前寫法還有改善空間(不優雅、不利他)。 ```cpp= list_ele_t *merge(list_ele_t *list1, list_ele_t *list2) { list_ele_t *result, *current; if (!list1) return list2; if (!list2) return list1; // init result and current if (strcmp(list1->value, list2->value) < 0) { result = list1; list1 = list1->next; } else { result = list2; list2 = list2->next; } current = result; // merge value of list nodes until one's end while (list1 && list2) { if (strcmp(list1->value, list2->value) < 0) { current->next = list1; list1 = list1->next; } else { current->next = list2; list2 = list2->next; } current = current->next; } if (list1) current->next = list1; if (list2) current->next = list2; return result; } list_ele_t *mergeSort(list_ele_t *head) { if (!head || !head->next) return head; // spilt into 2 lists, slow will be the half element of list list_ele_t *fast = head->next; list_ele_t *slow = head; while (fast && fast->next) { slow = slow->next; fast = fast->next->next; } fast = slow->next; slow->next = NULL; list_ele_t *list1 = mergeSort(head); list_ele_t *list2 = mergeSort(fast); return merge(list1, list2); } void q_sort(queue_t *q) { if (!q || q->size < 2) return; q->head = mergeSort(q->head); list_ele_t *last = q->head; while (last->next) { last = last->next; } q->tail = last; } ``` :::warning 剛剛看覺得糟糕,我 sort 啥東西都沒動,竟然有次會出現: > +++ TESTING trace trace-15-perf: > \# Test performance of insert_tail, size, reverse, and sort > ERROR: Time limit exceeded. Either you are in an infinite loop, or your code is too inefficient > --- trace-15-perf 0/6 > 難道真落在那 5% 拒絕區?還是我的表現只是近似 const time,但不是 const? 這我還要在找找原因(也可能是我 merge 太沒效率啦) ::: # natural sort <!-- Will add this at tomorrow --> 1. 下載他 ``` $ wget https://raw.githubusercontent.com/sourcefrog/natsort/master/strnatcmp.c $ wget https://raw.githubusercontent.com/sourcefrog/natsort/master/strnatcmp.h ``` 2. Makefile 在 Makefile 的 OBJS 中加入 strnatcmp.o 讓他生成 strnatcmp 的 object file 3. replace `strcmp()` to `strnatcmp()` in queue.c and qtest.c: 原本以為這樣就 ok 了,但是在 git pre-commit 的 hook 擋下來: ``` strnatcmp.c:112:14: style: The scope of the variable 'ca' can be reduced. [variableScope] nat_char ca, cb; ^ strnatcmp.c:112:18: style: The scope of the variable 'cb' can be reduced. [variableScope] nat_char ca, cb; ^ strnatcmp.c:163:0: style: The function 'strnatcmp' is never used. [unusedFunction] ^ Fail to pass static analysis. ``` 於是我把 `nat_char ca, cb;` 移進 `while(1)` 中,把未使用的 `strnatcmp` 註解起來。 >[name=25077667] CPPCHECK 好嚴格QQ > 然後就可以成功 commit 了。 ## Time limit exceeded 在這邊,每次 `make test` 的第 15 個測試都會超時。 ``` +++ TESTING trace trace-15-perf: # Test performance of insert_tail, size, reverse, and sort ERROR: Time limit exceeded. Either you are in an infinite loop, or your code is too inefficient --- trace-15-perf 0/6 ``` 跟上面的 "偶爾" 超時,有很大差異。 但是因為修改的部分只有 compare function 因此考慮的是 code 本身的效率問題與 function call 的成本。 因此我們拿 perf 去看他 ```shell= $ sudo perf top -p ${pid} ``` 使用 natsort 超時的關鍵發現: > Overhead Shared Objec Symbol > ==28.64% qtest [.] strnatcasecmp== > ==20.87% qtest [.] mergeSort== > ==12.11% qtest [.] show_queue== > 6.43% libc-2.27.so [.] _int_malloc > 3.21% qtest [.] test_malloc > 3.05% qtest [.] q_reverse > 2.80% qtest [.] test_strdup > 2.78% qtest [.] merge > 2.39% qtest [.] __ctype_toupper_loc@plt > 2.24% libc-2.27.so [.] __ctype_toupper_loc > 2.19% qtest [.] q_sort > 1.47% libc-2.27.so [.] __random > 1.32% libc-2.27.so [.] malloc > 1.06% libc-2.27.so [.] __strcasecmp_l_avx > 0.96% [kernel] [k] prepare_exit_to_usermode [結果檔案](https://github.com/25077667/lab0-c/blob/master/trace-15-perf-test) 我們應該先解決高 cpu 週期占比的。 所以是先考慮 ==標註起來== 的這三個 但是 `strnatcasecmp` 跟 `show_queue` 不可控。因此我就再確認如果不用 `strnatcasecmp` 的 cpu 使用週期其況。 > Overhead Shared Objec Symbol > 27.15% qtest [.] mergeSort > 17.55% qtest [.] show_queue > 6.96% libc-2.27.so [.] _int_malloc > 6.58% libc-2.27.so [.] __strcmp_sse2_unaligned > 4.31% qtest [.] q_reverse > 3.97% qtest [.] merge > 果然消耗最大的是 `mergeSort`。但是我不知道從何改善起,讓我想想這要怎麼優化。 :::danger TODO: 優化 mergeSort() ::: # 記憶體分析 `$ make valgrind` >[name=25077667]這在 WSL 上有可能會[錯誤](https://stackoverflow.com/questions/57206233/valgrind-showing-error-calling-pr-set-ptracer-vgdb-might-block)。 ![Imgur](https://imgur.com/GWJsDua.png) 基本上就可能 RAM 吃太飽 malloc 不到東西。其他沒啥問題: ![Imgur](https://imgur.com/jgBbhLf.png) `$ htop` ![Imgur](https://imgur.com/7fmjhbR.png) ## Massif `$ valgrind --tool=massif ./qtest -f traces/trace-18-natsort.cmd` 這是跑 nat_sort 的 valgrind 結果 ![Imgur](https://imgur.com/7BlCdfJ.png) [Output File](https://github.com/25077667/lab0-c/blob/master/massif.out.7408) # 論文 Dude, is my code constant time? 論文[連結](https://eprint.iacr.org/2016/1123.pdf) 論文說他這方法是時序攻擊的一種,然後是基於統計分析的,不會因為硬體環境差異而有所錯誤。 他也提到一個 [ctgrind](https://github.com/agl/ctgrind),挺有趣的,也是檢測一你是否為 const time。 至於他的統計分析,就是拿兩個不同的 data 丟進去,統計上透過 Student-T 的 model 去做用 Welch’s 去算 p-value. 至於 Welch’s 可以參考我去年統計學寫的[筆記](https://hackmd.io/@25077667/statistic#Case-II-%E6%AF%8D%E9%AB%94%E5%B8%B8%E6%85%8B%EF%BC%8C%E8%AE%8A%E7%95%B0%E6%95%B8%E7%9A%86%E6%9C%AA%E7%9F%A5),這是關於母體常態,變異數皆未知的情況。 這篇方法主要分三個步驟 1. Measure execution time * Classes definition: * 分兩類 input data * Cycle counters: * 計算 clock 數 * 在不同環境用不同方法算:systick、high-resolution timers、external equipment 或是有些處理器自己就有提供 * Environmental conditions: * 最小化環境差異造成影響 2. Apply post-processing * Cropping: * 去掉離群值 * 可能因為 interrupt 或是 extraneous activities 導致 * Higher-order preprocessing: * 這我看不懂 :::warning 和 higher-order function 無關,後者是 programming language 的範疇。這裡著重 "preprocessing",注意看論文的訴求。另外,不懂就標註出來,不要說「看不太懂」這種含糊的話 —— 誠實面對自己! :notes: jserv ::: 3. Apply statistical test * t-test: * T檢定 * 裡面一段提到說前面的 cropping 因為是 non-linear 所以會造成 T 分配的不適用。 * 這邊我有點不懂的是:去掉離群值不是很正常嗎,即便是 non-linear 的操作資料,不也只是讓 sample without bias 而已嗎? * 然後這邊他的 Null hypothesis 是兩個樣本同分布,但是如果只是要檢測是否比標準複雜度低,只需要單尾檢定即可。 * Non-parametric test: * 原文說也可以使用 nonparametric test,因為這依賴比較少的假設,但是缺點是收斂慢(需要大樣本)。 ## Dudect codes 這作法跟我之前的寫統計學作業的經驗不太一樣,過去我們統計學要求出 p-value,所以多個麻煩的積分。(那時候不能用 for loop 因為 IEEE-754 的誤差會影響到,然後每次的黎曼求法的 dx 又很小,所以浮點數的影響就會對後續的數據造成有點嚴重的 $\pm 1$ 誤差) https://jonisalonen.com/2013/deriving-welfords-method-for-computing-variance/ ==File: **ttest.[ch]**== ### t_ctx * mean: $H_0$ 與 $H_1$ 的平均數 * m2: $H_0$ 與 $H_1$ 的 $(N-1)*Var()$ * n: $H_0$ 與 $H_1$ 的 smaple size ### t_compute 計算檢定統計數,在 $T$ 分配底下。 這邊挺直覺得,就是 $t = \frac{\bar{X_0} - \bar{X_1}}{\sqrt{\frac{Var_0}{N_0} + \frac{Var_1}{N_1}}}$ ### t_push 這其實是提供 mean、m2 給 `fixture.c` 的 `update_statistics`。 * `dalta`: * `double delta = x - ctx->mean[class];` * 這個 `x` 就是他的執行時間 * `ctx->n[class]++` * 因為之前是 $N-1$,接下來要計算第 $N$ 個 * `mean` * 前一個的平均加上這一個的 $\Delta \over N$ ,相當合理 * 最難的是 `m2` * 推導如下: > 根據平均值公式:$\Delta \bar x = \bar x_1 - \bar x_2$ > $\rightarrow {x_1 - (N-1) \bar x_2 \over N}$ > 然後根據變異數的平移對稱性([之前筆記](https://hackmd.io/@25077667/statistic#%E8%AE%8A%E7%95%B0%E6%95%B8)),把 `m2` 展開 $M_{2,i} = M_{2,{i-1}} + (x_i - \bar{x_{i-1}}) \cdot (x_i - \bar{x_{i}})$ > 就是這行了 > ```cpp= > ctx->m2[class] = ctx->m2[class] + delta * (x - ctx->mean[class]); > ``` ==File: **fixture.[ch]**== ### do_it * `mode`: 傳入`constant.c` 的 `measure` 用於控制要 `test_insert_tail` 還是 `test_insert_tail`。 * 有疑問的點: * 這邊有使用到 5 個空間都是使用 `calloc`,其中 `input_data` 就是測試資料,竟然沒有出現跟我上面相同的錯誤: > Attempted to free unallocated block. Address = 0x7ffff4e93890 * `prepare_inputs`: 產生 input ,而這邊比較有趣的是 rand 出來的 class[i] 有二分之一的機率,會把下面兩個 byte 也寫成 0。控制下面 `measure` 函式的插入次數。 > *(uint16_t *) (input_data + i * chunk_size) = 0x00; * 其中的 `randombytes` 會使用到 `/dev/urandom/`,去填 `input_data`。但是似乎沒有 close(fd)。不確定是什麼原因不需要,或是沒有寫 close(fd)。 * 果然我跑 valgrind 看,他會說我沒有關檔。![Imgur](https://imgur.com/ZRvPQNg.png) * `measure`: 對頭尾插入 `*(uint16_t *) (input_data + i * chunk_size)` 次。分別儲存於 `before_ticks` 與 `after_ticks`。 * `differentiate`: 將前述 `before_ticks` 與 `after_ticks` 的相差,儲存於 `exec_times`。 * `update_statistics`: 根據 `differentiate` 算完,放在`exec_times` 的時間差,去做 T 檢定。 * `cpucycles`: 這是在 `measure` 中測量執行時間的關鍵,他定義在 `cpucycles.h` 中。 ```cpp= #include <stdint.h> // http://www.intel.com/content/www/us/en/embedded/training/ia-32-ia-64-benchmark-code-execution-paper.html static inline int64_t cpucycles(void) { #if defined(__i386__) || defined(__x86_64__) unsigned int hi, lo; __asm__ volatile("rdtsc\n\t" : "=a"(lo), "=d"(hi)); return ((int64_t) lo) | (((int64_t) hi) << 32); #else #error Unsupported Architecture #endif } ``` 這邊覺得有趣的是 `bitwise` 運算,為何這會是 `cpucycles` 呢? cpu_high, cpu_low 都是使用 RDTSC 這個 [x86_Read Time-Stamp Counter](https://c9x.me/x86/html/file_module_x86_id_278.html) > `hi`: 就是 cpu_high , `lo`: 就是 cpu_low > 我們可見在 `((int64_t) lo) | (((int64_t) hi) << 32);` 就是把 `cpu_high` 放在高位元(32),把 `cpu_low` 放在低位元(32)。一起組成 64 位元。傳回 `before_ticks` 跟 `after_ticks`。 然後這邊有關於 volatile 的[參考資料](https://bluelove1968.pixnet.net/blog/post/222282820) # select <!-- 正在看 --> 覺得看完[重點提示](https://hackmd.io/@sysprog/H1TtmVTTz),還是只有舉燭。 所以以下是我的舉燭。 * `select`: * 是一個 sys_call。 * 宣告:`int select(int nfds, fd_set *readfds, fd_set *writefds, fd_set *exceptfds, struct timeval *timeout);` * `nfds`: An integer one more than the maximum of any file descriptor in any of the sets. * `readfds`: Holding the file descriptors to be checked for being ready to **read**, and on output indicates which file descriptors are ready to **read**. * `writefds`: Holding the file descriptors to be checked for being ready to **write**, and on output indicates which file descriptors are ready to **write**. * `errorfds`: Holding the file descriptors to be checked for **error conditions pending**, and on output indicates which file descriptors have **error conditions pending**. * `timeout`: Specifies a **maximum interval to wait** for the selection to complete. * 其中 `readfds`, `writefds`, `errorfds` 可以為 `NULL` * 但是如果「全部」是 NULL,意義是什麼?我不知道。 * man 的描述: * `select()` and `pselect()` allow a program to monitor multiple file descriptors, waiting until one or more of the file descriptors become "ready" for some class of I/O operation (e.g., input possible). A file descriptor is considered ready if it is possible to perform a corresponding I/O operation (e.g., `read(2)`, or a sufficiently small `write(2)`) without blocking. * I/O Model: * [參考資料](https://medium.com/@clu1022/%E6%B7%BA%E8%AB%87i-o-model-32da09c619e6) * 這有張很有趣的圖: * ![](https://miro.medium.com/max/614/1*RcsvMcPF2wC_3tV5ckU8vg.jpeg) * 從上面這張圖,描述以下這四種 I/O 方式 * Blocking I/O * Non-blocking I/O * Synchronous I/O * Asynchronous I/O * `select()` 作用: * `select()` * +`read()` 是一種 Blocking I/O * +`recvfrom()` 是一種 Non-blocking I/O * 當我們給 `select` 多個 `fd` 時,可以同時觀察多個 I/O 是否 "is ready"。 然後我們可以從 `select()` 的 return 值,取得第幾個 `fd` 使用 `read()` 進行 I/O。 # 小結 寫不完 la # 參考資料 [perf tool](https://hackmd.io/@sysprog/B11109rdg)

    Import from clipboard

    Paste your markdown or webpage here...

    Advanced permission required

    Your current role can only read. Ask the system administrator to acquire write and comment permission.

    This team is disabled

    Sorry, this team is disabled. You can't edit this note.

    This note is locked

    Sorry, only owner can edit this note.

    Reach the limit

    Sorry, you've reached the max length this note can be.
    Please reduce the content or divide it to more notes, thank you!

    Import from Gist

    Import from Snippet

    or

    Export to Snippet

    Are you sure?

    Do you really want to delete this note?
    All users will lose their connection.

    Create a note from template

    Create a note from template

    Oops...
    This template has been removed or transferred.
    Upgrade
    All
    • All
    • Team
    No template.

    Create a template

    Upgrade

    Delete template

    Do you really want to delete this template?
    Turn this template into a regular note and keep its content, versions, and comments.

    This page need refresh

    You have an incompatible client version.
    Refresh to update.
    New version available!
    See releases notes here
    Refresh to enjoy new features.
    Your user state has changed.
    Refresh to load new user state.

    Sign in

    Forgot password

    or

    By clicking below, you agree to our terms of service.

    Sign in via Facebook Sign in via Twitter Sign in via GitHub Sign in via Dropbox Sign in with Wallet
    Wallet ( )
    Connect another wallet

    New to HackMD? Sign up

    Help

    • English
    • 中文
    • Français
    • Deutsch
    • 日本語
    • Español
    • Català
    • Ελληνικά
    • Português
    • italiano
    • Türkçe
    • Русский
    • Nederlands
    • hrvatski jezik
    • język polski
    • Українська
    • हिन्दी
    • svenska
    • Esperanto
    • dansk

    Documents

    Help & Tutorial

    How to use Book mode

    Slide Example

    API Docs

    Edit in VSCode

    Install browser extension

    Contacts

    Feedback

    Discord

    Send us email

    Resources

    Releases

    Pricing

    Blog

    Policy

    Terms

    Privacy

    Cheatsheet

    Syntax Example Reference
    # Header Header 基本排版
    - Unordered List
    • Unordered List
    1. Ordered List
    1. Ordered List
    - [ ] Todo List
    • Todo List
    > Blockquote
    Blockquote
    **Bold font** Bold font
    *Italics font* Italics font
    ~~Strikethrough~~ Strikethrough
    19^th^ 19th
    H~2~O H2O
    ++Inserted text++ Inserted text
    ==Marked text== Marked text
    [link text](https:// "title") Link
    ![image alt](https:// "title") Image
    `Code` Code 在筆記中貼入程式碼
    ```javascript
    var i = 0;
    ```
    var i = 0;
    :smile: :smile: Emoji list
    {%youtube youtube_id %} Externals
    $L^aT_eX$ LaTeX
    :::info
    This is a alert area.
    :::

    This is a alert area.

    Versions and GitHub Sync
    Get Full History Access

    • Edit version name
    • Delete

    revision author avatar     named on  

    More Less

    Note content is identical to the latest version.
    Compare
      Choose a version
      No search result
      Version not found
    Sign in to link this note to GitHub
    Learn more
    This note is not linked with GitHub
     

    Feedback

    Submission failed, please try again

    Thanks for your support.

    On a scale of 0-10, how likely is it that you would recommend HackMD to your friends, family or business associates?

    Please give us some advice and help us improve HackMD.

     

    Thanks for your feedback

    Remove version name

    Do you want to remove this version name and description?

    Transfer ownership

    Transfer to
      Warning: is a public team. If you transfer note to this team, everyone on the web can find and read this note.

        Link with GitHub

        Please authorize HackMD on GitHub
        • Please sign in to GitHub and install the HackMD app on your GitHub repo.
        • HackMD links with GitHub through a GitHub App. You can choose which repo to install our App.
        Learn more  Sign in to GitHub

        Push the note to GitHub Push to GitHub Pull a file from GitHub

          Authorize again
         

        Choose which file to push to

        Select repo
        Refresh Authorize more repos
        Select branch
        Select file
        Select branch
        Choose version(s) to push
        • Save a new version and push
        • Choose from existing versions
        Include title and tags
        Available push count

        Pull from GitHub

         
        File from GitHub
        File from HackMD

        GitHub Link Settings

        File linked

        Linked by
        File path
        Last synced branch
        Available push count

        Danger Zone

        Unlink
        You will no longer receive notification when GitHub file changes after unlink.

        Syncing

        Push failed

        Push successfully