---
tags: 2022 linux kernel
---
# 2022q1 Homework1 (lab0)
contributed by < [`Risheng1128`](https://github.com/Risheng1128) >
> [作業說明](https://hackmd.io/@sysprog/linux2022-lab0)
## 實驗環境
```shell
$ gcc --version
gcc (Ubuntu 9.3.0-17ubuntu1~20.04) 9.3.0
Architecture: x86_64
CPU op-mode(s): 32-bit, 64-bit
Byte Order: Little Endian
Address sizes: 39 bits physical, 48 bits virtual
CPU(s): 4
On-line CPU(s) list: 0-3
Thread(s) per core: 2
Core(s) per socket: 2
Socket(s): 1
NUMA node(s): 1
Vendor ID: GenuineIntel
CPU family: 6
Model: 142
Model name: Intel(R) Core(TM) i5-7200U CPU @ 2.50GHz
Stepping: 9
CPU MHz: 2700.000
CPU max MHz: 3100.0000
CPU min MHz: 400.0000
BogoMIPS: 5399.81
Virtualization: VT-x
L1d cache: 64 KiB
L1i cache: 64 KiB
L2 cache: 512 KiB
L3 cache: 3 MiB
NUMA node0 CPU(s): 0-3
```
## 開發紀錄
:::info
作業要求
- [x] q_new: 建立新的「空」佇列
- [x] q_free: 釋放佇列所佔用的記憶體
- [x] q_insert_head: 在佇列開頭 (head) 插入 (insert) 給定的新節點 (以 LIFO 準則)
- [x] q_insert_tail: 在佇列尾端 (tail) 插入 (insert) 給定的新節點 (以 FIFO 準則)
- [x] q_remove_head: 在佇列開頭 (head) 移去 (remove) 給定的節點
- [x] q_release_element: 釋放特定節點的記憶體空間
- [x] q_size: 計算目前佇列的節點總量
- [x] q_delete_mid: 移走佇列中間的節點
- [x] q_delete_dup: 在已經排序的狀況,移走佇列中具備重複內容的節點
- [x] q_swap: 交換佇列中鄰近的節點
- [x] q_reverse: 以反向順序重新排列鏈結串列,該函式不該配置或釋放任何鏈結串列節點,換言之,它只能重排已經存在的節點
- [x] q_sort: 以遞增順序來排序鏈結串列的節點
:::
### 二個重要的結構體
```c
struct list_head {
struct list_head *prev;
struct list_head *next;
};
```
```c
/* Linked list element */
typedef struct {
char *value;
struct list_head list;
} element_t;
```
### q_new
根據 `list.h` 參考函式 `INIT_LIST_HEAD` ,其功能為將 linked list 的 `next` 及 `prev` 指向本身,即初始化
```c
static inline void INIT_LIST_HEAD(struct list_head *head)
{
head->next = head;
head->prev = head;
}
```
- 實作流程
1. 使用 `malloc` 分配記憶體空間,並由指標 `new` 指向
2. 如果空間分配失敗 `malloc` 會回傳 `NULL` ,此時回傳 `NULL`
3. 使用 `INIT_LIST_HEAD` 初始化
:::spoiler 實際程式碼
```c
struct list_head *q_new()
{
struct list_head *new = malloc(sizeof(struct list_head));
if (!new)
return NULL;
INIT_LIST_HEAD(new);
return new;
}
```
:::
### q_free
這邊利用到了一個特殊的巨集 `list_entry` ,可以從 `list.h` 查到該巨集
```c
#define list_entry(node, type, member) container_of(node, type, member)
```
可以看到 `list_entry` 會使用到 `container_of` ,參考 [Linux 核心原始程式碼巨集: container_of](https://hackmd.io/@sysprog/linux-macro-containerof)可以得知 `container_of` 可以用來得到包含 `ptr` 之 `object` 的起始地址
```c
#ifndef container_of
#ifdef __LIST_HAVE_TYPEOF
#define container_of(ptr, type, member) \
__extension__({ \
const __typeof__(((type *) 0)->member) *__pmember = (ptr); \
(type *) ((char *) __pmember - offsetof(type, member)); \
})
#else
#define container_of(ptr, type, member) \
((type *) ((char *) (ptr) -offsetof(type, member)))
#endif
#endif
```
- 實作流程
1. 如果傳入的 `head` 為 `NULL` → `return`
2. 走訪整個 `linked list` ,每經過一個節點,就對其使用 `list_entry()` 並釋放其空間
3. 釋放分配給 `head` 的記憶體空間
:::spoiler 實際程式碼,套用 API 可以寫出更精簡的寫法
```diff
void q_free(struct list_head *l)
{
if (!l)
return;
struct list_head *node = l->next;
- while (node != l) {
- struct list_head *del = node;
- node = node->next;
- q_release_element(list_entry(del, element_t, list));
- }
+ element_t *node, *next;
+ list_for_each_entry_safe (node, next, l, list) {
+ list_del(&node->list);
+ q_release_element(node);
+ }
free(l);
}
```
:::
### q_insert_head
根據 `list.h` 參考函式 `list_add` ,查看原始碼可以得知其功能為把 `node` 節點加在 `head` 之後
```c
static inline void list_add(struct list_head *node, struct list_head *head)
{
struct list_head *next = head->next;
next->prev = node;
node->next = next;
node->prev = head;
head->next = node;
}
```
- 實作流程
1. 使用 `malloc()` 分配 `element_t` 大小的記憶體空間,若分配失敗則回傳 `false`
```c
element_t *new = malloc(sizeof(element_t));
if (!new)
return false;
```
2. 使用 `strlen()` 取得字串長度,並分配要複製字串的空間
:::warning
分配的空間大小要比字串長度多一個位元組,記得加入 `'\0'`
:::
```c
int str_size = strlen(s);
new->value = malloc((str_size + 1) * sizeof(char));
if (!new->value) {
free(new);
return false;
}
strncpy(new->value, s, str_size);
*(new->value + str_size) = '\0';
```
3. 使用 `list_add()` 將節點加在 `head` 的下一個
```c
list_add(&new->list, head);
```
:::spoiler 實際程式碼
```c=
bool q_insert_head(struct list_head *head, char *s)
{
if (!head)
return false;
element_t *new = malloc(sizeof(element_t));
if (!new)
return false;
INIT_LIST_HEAD(&new->list);
int str_size = strlen(s);
new->value = malloc((str_size + 1) * sizeof(char));
if (!new->value) {
free(new);
return false;
}
strncpy(new->value, s, str_size);
*(new->value + str_size) = '\0';
list_add(&new->list, head);
return true;
}
```
:::
- 踩雷小紀錄: `q_insert_head()` 第 15 行
在執行測試檔 `trace-12-malloc.cmd` 時,一開始都會有 [memory leak](https://en.wikipedia.org/wiki/Memory_leak) 的問題,原本以為是 `q_free()` 有問題,結果最後發現問題是出在 `q_insert_head()` 和 `q_insert_tail()`
原因: 缺乏考慮當 `new` 成功被分配空間而 `new->value` 分配空間失敗的情況
解法: 在 `new->value` 分配失敗時記得要將 `new` 的空間也釋放掉
### q_insert_tail
`q_insert_tail()` 的思考幾乎都和 `q_insert_head()` 相同,兩者只差在插入的位置不同,也就使用不同函式
- `q_insert_head()` 使用 `list_add()`
- `q_insert_tail()` 使用 `list_add_tail()`
:::spoiler 實際程式碼
```c
bool q_insert_tail(struct list_head *head, char *s)
{
if (!head)
return false;
element_t *new = malloc(sizeof(element_t));
if (!new)
return false;
INIT_LIST_HEAD(&new->list);
int str_size = strlen(s);
new->value = malloc((str_size + 1) * sizeof(char));
if (!new->value) {
free(new);
return false;
}
strncpy(new->value, s, str_size);
*(new->value + str_size) = '\0';
list_add_tail(&new->list, head);
return true;
}
```
:::
### q_remove_head
根據 `list.h` 參考函式 `list_del` ,其功能為將節點 `node` 從 `linked list` 上 **remove**,成為單獨的節點
```c
static inline void list_del(struct list_head *node)
{
struct list_head *next = node->next;
struct list_head *prev = node->prev;
next->prev = prev;
prev->next = next;
#ifdef LIST_POISONING
node->prev = (struct list_head *) (0x00100100);
node->next = (struct list_head *) (0x00200200);
#endif
}
```
- 思考方法
1. 如果 `head` 為 `NULL` 或是 `empty` ,則回傳 `NULL`
2. 將佇列的第一個 element 取出(利用 `list_entry` )並且 `remove` 該 element
3. 將 element->value 根據 `bufsize` 大小做複製
:::spoiler 實際程式碼
```c
element_t *q_remove_head(struct list_head *head, char *sp, size_t bufsize)
{
if (!head || list_empty(head))
return NULL;
element_t *target = list_entry(head->next, element_t, list);
list_del(&target->list);
if (bufsize) {
strncpy(sp, target->value, bufsize - 1);
*(sp + bufsize - 1) = '\0';
}
return target;
}
```
:::
### q_remove_tail
`q_remove_tail()` 的作法和 `q_remove_head()` 大致相同,兩者只差在移除的位置不同
- `q_remove_tail()`: 被移除的節點位置為 `head->prev`
- `q_remove_head()`: 被移除的節點位置為 `head->next`
:::spoiler 實際程式碼
```c
element_t *q_remove_tail(struct list_head *head, char *sp, size_t bufsize)
{
if (!head || list_empty(head))
return NULL;
element_t *target = list_entry(head->prev, element_t, list);
list_del(&target->list);
if (bufsize) {
strncpy(sp, target->value, bufsize - 1);
*(sp + bufsize - 1) = '\0';
}
return target;
}
```
:::
### q_size
根據 `list.h` 參考巨集函式 `list_for_each()`,其功能為走訪整個 linked list
```c
#define list_for_each(node, head) \
for (node = (head)->next; node != (head); node = node->next)
```
利用 `list_for_each()` 可以完成 `q_size()`
:::spoiler 實際程式碼
```c
int q_size(struct list_head *head)
{
if (!head)
return 0;
int size = 0;
struct list_head *node;
list_for_each (node, head)
size++;
return size;
}
```
:::
### q_delete_mid
- 思考方法
1. 使用「龜兔賽跑」[Floyd’s Cycle detection](https://en.wikipedia.org/wiki/Cycle_detection)來尋找中間的節點
- `rabbit` 每一次迭代會走訪兩個節點
- `turtle` 每一次迭代會走訪一個節點
2. 當 `rabbit` 走到 `head` 或是 `rabbit->next` 為 `head` 表示已經抵達「終點」,此時 `turtle` 指到的位址剛好會是佇列正中間的節點
3. 透過 `list_del` 可以將正中間的節點移出 linked list
4. 透過 `list_entry` 可以得到正中間 element 的地址,即可以進行記憶體空間釋放
:::spoiler 實際程式碼
```c
bool q_delete_mid(struct list_head *head)
{
// https://leetcode.com/problems/delete-the-middle-node-of-a-linked-list/
if (!head || !head->next)
return false;
struct list_head *rabbit = head->next, *turtle = head->next;
while (rabbit != head && rabbit->next != head) {
rabbit = rabbit->next->next;
turtle = turtle->next;
}
list_del(turtle);
q_release_element(list_entry(turtle, element_t, list));
return true;
}
```
:::
### q_delete_dup
- 思考方法
1. 如果 `head` 為 `NULL` 或 empty 或只有一個 element 時,則會回傳 `false`
→ 表示一定不會有重複的資料
```c
if (!head || list_empty(head) || list_is_singular(head))
return false;
```
2. 變數解釋
- `curr` 為指向當前節點的指標
- `next` 為指向 `curr` 下一個節點的指標
- 宣告布林型態變數 `key` 表示是否發生資料相同的情況
`key` 為 `false` 表示當前的 `curr` 和 `next` 沒有發生資料相同的情況
`key` 為 `true` 表示當前的 `curr` 和 `next` 有發生資料相同的情況
3. 走訪整個 linked list 並取得 `curr` 和 `next` 的 element 地址
```c
element_t *curr_entry = list_entry(curr, element_t, list);
element_t *next_entry = list_entry(next, element_t, list);
```
4. 當資料相同時,會對 `next` 指向的節點進行移除並釋放空間,同時 `key` 被設置為 `true`
這邊資料的比較使用 [strcmp()](http://tw.gitbook.net/c_standard_library/c_function_strcmp.html) 函式,當資料相同時回傳 `0`
```c
while (next != head && !strcmp(curr_entry->value, next_entry->value)) {
list_del(next);
q_release_element(next_entry);
next = curr->next;
next_entry = list_entry(next, element_t, list);
key = true;
}
```
5. 當 `key` 為 `true` 時,表示發生過資料相同的情況,但 `curr` 指到的節點尚未釋放,因此要記得釋放該空間
```c
if (key) {
list_del(curr);
q_release_element(curr_entry);
key = false;
}
```
:::spoiler 實際程式碼
```c
bool q_delete_dup(struct list_head *head)
{
// https://leetcode.com/problems/remove-duplicates-from-sorted-list-ii/
if (!head || list_empty(head) || list_is_singular(head))
return false;
struct list_head *curr = head->next, *next = curr->next;
bool key = false;
while (curr != head && next != head) {
element_t *curr_entry = list_entry(curr, element_t, list);
element_t *next_entry = list_entry(next, element_t, list);
while (next != head && !strcmp(curr_entry->value, next_entry->value)) {
list_del(next);
q_release_element(next_entry);
next = curr->next;
next_entry = list_entry(next, element_t, list);
key = true;
}
if (key) {
list_del(curr);
q_release_element(curr_entry);
key = false;
}
curr = next;
next = next->next;
}
return true;
}
```
:::
### q_swap
- 思考方法
1. 建立兩個指標 `first` 及 `second`
2. 利用 `list_del_init` 將 `first` 指向的節點從 linked list 取出
3. 利用 `list_add(first, second);` 將 `first` 指向節點加到 `second` 指向節點的下一個位置,即兩節點交換位置
4. 當 `first` 或 `first->next` 為 `head` 時,表示已經到 linked list 的盡頭
:::spoiler 實際程式碼
```c
void q_swap(struct list_head *head)
{
// https://leetcode.com/problems/swap-nodes-in-pairs/
if (!head)
return;
struct list_head *first = head->next;
while (first != head && first->next != head) {
struct list_head *second = first->next;
list_del_init(first);
list_add(first, second);
first = first->next;
}
}
```
:::
### q_reverse
1. 建立三個指標 `prev`, `curr` 及 `next`,並依照 `prev` → `curr` → `next` 的順序指向不同節點
```c
struct list_head *prev = head->prev, *curr = head, *next = NULL;
```
```graphviz
digraph ele_list {
rankdir=LR;
node[shape=record];
head [label="Head"]
e1 [label="bear"]
e2 [label="dolphin"]
e3 [label="gerbil"]
e4 [label="..."]
e5 [label="meerkat"]
prev [shape=plaintext;label="prev"]
curr [shape=plaintext;label="curr"]
next [shape=plaintext;label="next"]
// next 方向
head -> e1
e1 -> e2
e2 -> e3
e3 -> e4
e4 -> e5
e5 -> head
// prev 方向
head -> e5
e5 -> e4
e4 -> e3
e3 -> e2
e2 -> e1
e1 -> head
// pointer 初始化
prev -> e5 [color=red]
curr -> head [color=red]
}
```
2. `next = curr->next;` (`next` 指向 `curr` 的下一個節點 )
```graphviz
digraph ele_list {
rankdir=LR;
node[shape=record];
head [label="Head"]
e1 [label="bear"]
e2 [label="dolphin"]
e3 [label="gerbil"]
e4 [label="..."]
e5 [label="meerkat"]
prev [shape=plaintext;label="prev"]
curr [shape=plaintext;label="curr"]
next [shape=plaintext;label="next"]
// next 方向
head -> e1
e1 -> e2
e2 -> e3
e3 -> e4
e4 -> e5
e5 -> head
// prev 方向
head -> e5
e5 -> e4
e4 -> e3
e3 -> e2
e2 -> e1
e1 -> head
// pointer 初始化
prev -> e5
curr -> head
next -> e1 [color=red]
}
```
3. `curr->next = prev;` ( `curr` 的 `next` 指向原本的上一個節點 `prev` )
```graphviz
digraph ele_list {
rankdir=LR;
node[shape=record];
head [label="Head"]
e1 [label="bear"]
e2 [label="dolphin"]
e3 [label="gerbil"]
e4 [label="..."]
e5 [label="meerkat"]
prev [shape=plaintext;label="prev"]
curr [shape=plaintext;label="curr"]
next [shape=plaintext;label="next"]
// next 方向
head -> e5 [label="next";color=red]
e1 -> e2
e2 -> e3
e3 -> e4
e4 -> e5
e5 -> head
// prev 方向
head -> e5
e5 -> e4
e4 -> e3
e3 -> e2
e2 -> e1
e1 -> head
// pointer 初始化
prev -> e5
curr -> head
next -> e1
}
```
4. `curr->prev = next;` ( `curr` 的 `prev` 指向原本的下一個節點 `next` )
```graphviz
digraph ele_list {
rankdir=LR;
node[shape=record];
head [label="Head"]
e1 [label="bear"]
e2 [label="dolphin"]
e3 [label="gerbil"]
e4 [label="..."]
e5 [label="meerkat"]
prev [shape=plaintext;label="prev"]
curr [shape=plaintext;label="curr"]
next [shape=plaintext;label="next"]
// next 方向
head -> e5 [label="next"]
e1 -> e2
e2 -> e3
e3 -> e4
e4 -> e5
e5 -> head
// prev 方向
head -> e1 [label="prev";color=red]
e5 -> e4
e4 -> e3
e3 -> e2
e2 -> e1
e1 -> head
// pointer 初始化
prev -> e5
curr -> head
next -> e1
}
```
5. `prev = curr;` ( `prev` 往下移動一個節點 )
```graphviz
digraph ele_list {
rankdir=LR;
node[shape=record];
head [label="Head"]
e1 [label="bear"]
e2 [label="dolphin"]
e3 [label="gerbil"]
e4 [label="..."]
e5 [label="meerkat"]
prev [shape=plaintext;label="prev"]
curr [shape=plaintext;label="curr"]
next [shape=plaintext;label="next"]
// next 方向
head -> e5 [label="next"]
e1 -> e2
e2 -> e3
e3 -> e4
e4 -> e5
e5 -> head
// prev 方向
head -> e1 [label="prev"]
e5 -> e4
e4 -> e3
e3 -> e2
e2 -> e1
e1 -> head
// pointer 初始化
prev -> head [color=red]
curr -> head
next -> e1
}
```
6. `curr = next;`
```graphviz
digraph ele_list {
rankdir=LR;
node[shape=record];
head [label="Head"]
e1 [label="bear"]
e2 [label="dolphin"]
e3 [label="gerbil"]
e4 [label="..."]
e5 [label="meerkat"]
prev [shape=plaintext;label="prev"]
curr [shape=plaintext;label="curr"]
next [shape=plaintext;label="next"]
// next 方向
head -> e5 [label="next"]
e1 -> e2
e2 -> e3
e3 -> e4
e4 -> e5
e5 -> head
// prev 方向
head -> e1 [label="prev"]
e5 -> e4
e4 -> e3
e3 -> e2
e2 -> e1
e1 -> head
// pointer 初始化
prev -> head
curr -> e1 [color=red]
next -> e1
}
```
經過這六個步驟,可以發現 `head` 的 `prev` 和 `next` 的變化
- `head` 的 `next` 從指向 `bear` (第一個 element)改成指向 `meekat` (最後一個 element)
- `head` 的 `prev` 從指向 `meerkat` (最後一個 element)改成指向 `bear` (第一個 element)
:::spoiler 實際程式碼
```c
void q_reverse(struct list_head *head)
{
if (!head)
return;
struct list_head *prev = head->prev, *curr = head, *next = NULL;
while (next != head) {
next = curr->next;
curr->next = prev;
curr->prev = next;
prev = curr;
curr = next;
}
}
```
:::
### q_sort
#### 實作 quick sort 遞迴版本
參考 [quick-sort.c](https://github.com/sysprog21/linux-list/blob/master/examples/quick-sort.c) 使用 Linux Kernel API 實做 quick sort
參數介紹
- `less`: 存放資料比 `pivot` 小的 linked list
- `greater`: 存放資料比 `pivot` 大或相同的 linked list
首先將 `pivot` 設定為 linked list 的第一個節點,並且從 linked list上移除
```c
pivot = list_first_entry(head, element_t, list);
list_del(&pivot->list);
```
接著走訪整個節點,並讓每個節點和 `pivot` 的資料互相比較(使用 `strcmp()` 並根據其結果決定將節點加在 `less` 或是 `greater`)
```c
list_for_each_entry_safe (item, is, head, list) {
if (strcmp(item->value, pivot->value) < 0)
list_move_tail(&item->list, &less);
else
list_move(&item->list, &greater);
}
```
使用遞迴分別將 `less` 及 `greater` 進行比較和分割
```c
INIT_LIST_HEAD(&less);
INIT_LIST_HEAD(&greater);
```
將每次分割好的節點依照 `less` → `head` → `greater` 重組起來
```c
list_add(&pivot->list, head);
list_splice(&less, head);
list_splice_tail(&greater, head);
```
:::spoiler 實際程式碼
```c
void q_sort(struct list_head *head)
{
struct list_head less, greater;
element_t *pivot, *item, *is = NULL;
if (!head || list_empty(head) || list_is_singular(head))
return;
INIT_LIST_HEAD(&less);
INIT_LIST_HEAD(&greater);
pivot = list_first_entry(head, element_t, list);
list_del(&pivot->list);
list_for_each_entry_safe (item, is, head, list) {
if (strcmp(item->value, pivot->value) < 0)
list_move_tail(&item->list, &less);
else
list_move(&item->list, &greater);
}
q_sort(&less);
q_sort(&greater);
list_add(&pivot->list, head);
list_splice(&less, head);
list_splice_tail(&greater, head);
}
```
:::
結果 `trace-14-perf.cmd` 和 `trace-15-perf.cmd` 沒有通過QQ
:::warning
注意排序操作期待是 stable sorting
:notes: jserv
:::
:::info
好的,那學生試試看使用 merge sort 實作,不過想請問老師為什麼要使用 stable sorting ,是因為 unstable sorting 的效能比較不好,還是有其他的原因,謝謝老師!
> 「效能」只是評估的一個因素,stable sorting 和 unstable sorting 的應用場景不同,請回去讀書。
> :notes: jserv
了解老師的意思了,謝謝老師
:::
決定改成用 merge sort 試試看
#### 實作 merge sort 遞迴版本
merge sort 的實作被拆為三個函式 `q_sort()` `mergesort()` 及 `mergelist()`
- `q_sort()`
1. 首先將指向 `head` 的指標改成 `NULL`,目的在於把 doubly linked list 的終點從 `head` 改為 `NULL`,變成單向 linked list
```c
head->prev->next = NULL;
```
2. 接著呼叫 `mergesort()` 開始進行 merge sort
```c
head->next = mergesort(head->next);
```
3. 排序完後每個節點的 `prev` 會亂掉,因此需要走訪各個節點把所有 `prev` 接回來
```c
struct list_head *curr = head, *next = head->next;
while (next) {
next->prev = curr;
curr = next;
next = next->next;
}
```
- `mergesort()`
參考[你所不知道的 C 語言: linked list 和非連續記憶體](https://hackmd.io/@sysprog/c-linked-list)裡頭, merge sort 的實作方式並做修改
1. 使用「龜兔賽跑」[Floyd’s Cycle detection](https://en.wikipedia.org/wiki/Cycle_detection)演算法找出 linked list 正中間的節點,並將 linked list 切成 `left` 及 `right` 兩個 linked list
```c
struct list_head *rabbit = head, *turtle = head;
while (rabbit && rabbit->next) {
rabbit = rabbit->next->next;
turtle = turtle->next;
}
struct list_head *mid = turtle;
turtle->prev->next = NULL;
struct list_head *left = mergesort(head);
struct list_head *right = mergesort(mid);
```
2. 呼叫 `mergelist()` 合併 `left` 及 `right`
```c
return mergelist(left, right);
```
- `mergelist()`
參考[21. Merge Two Sorted Lists](https://leetcode.com/problems/merge-two-sorted-lists/),並實作合併兩個 linked list
1. 利用 pointer to pointer `indirect` 指向 pointer `res` ,並藉由修改指標完成 linked list 合併的動作
```c
struct list_head *res = NULL;
struct list_head **indirect = &res;
```
2. 使用 `strcmp` 判斷資料的大小
```c
node = strcmp(list1_entry->value, list2_entry->value) < 0 ? &list1
```
:::spoiler 實際程式碼
```c
/*
* Merge the two lists in a one sorted list.
*/
static struct list_head *mergelist(struct list_head *list1,
struct list_head *list2)
{
struct list_head *res = NULL;
struct list_head **indirect = &res;
for (struct list_head **node = NULL; list1 && list2;
*node = (*node)->next) {
element_t *list1_entry = list_entry(list1, element_t, list);
element_t *list2_entry = list_entry(list2, element_t, list);
node = strcmp(list1_entry->value, list2_entry->value) < 0 ? &list1
: &list2;
*indirect = *node;
indirect = &(*indirect)->next;
}
*indirect = (struct list_head *) ((u_int64_t) list1 | (u_int64_t) list2);
return res;
}
/*
* Divide the list into several nodes and merge to sorted list.
* No effect if q is NULL or empty.
*/
static struct list_head *mergesort(struct list_head *head)
{
if (!head || !head->next)
return head;
struct list_head *rabbit = head, *turtle = head;
while (rabbit && rabbit->next) {
rabbit = rabbit->next->next;
turtle = turtle->next;
}
struct list_head *mid = turtle;
turtle->prev->next = NULL;
struct list_head *left = mergesort(head);
struct list_head *right = mergesort(mid);
return mergelist(left, right);
}
/*
* Sort elements of queue in ascending order
* No effect if q is NULL or empty. In addition, if q has only one
* element, do nothing.
*/
void q_sort(struct list_head *head)
{
if (!head || list_empty(head))
return;
head->prev->next = NULL;
head->next = mergesort(head->next);
struct list_head *curr = head, *next = head->next;
while (next) {
next->prev = curr;
curr = next;
next = next->next;
}
curr->next = head;
head->prev = curr;
}
```
:::
### 研讀 Linux 核心原始程式碼 [list_sort.c](https://github.com/torvalds/linux/blob/master/lib/list_sort.c)
:::info
作業要求
- [x] 研讀 Linux 核心原始程式碼的 [lib/list_sort.c](https://github.com/torvalds/linux/blob/master/lib/list_sort.c)
- [x] 嘗試引入 Linux 核心原始程式碼到 lab0-c 專案
- [x] 比較你自己實作的 merge sort 和 Linux 核心程式碼之間效能落差
:::
為了讓篇幅小一點,這邊再開了一個新的筆記用來紀錄分析過程: [研讀 Linux 核心原始程式碼 list_sort](https://hackmd.io/@Risheng/list_sort)
### 引入 Linux 核心原始碼到 lab0-c
為了引入 Linux 核心原始碼,必須要先了解 Linux 核心的 merge sort 整個的運作流程,已將整個分析放在[額外的筆記本](https://hackmd.io/@Risheng/list_sort)上,根據分析結果,以下是我規劃的實作流程
:::info
實作流程
1. 將檔案 [lib/list_sort.c](https://github.com/torvalds/linux/blob/master/lib/list_sort.c) 加進 lab0-c
2. 將位於其他檔案的巨集函式加進 `list_sort.h`
3. 在 `qtest.c` 新增函式 `cmp()`
4. 修改 `qtest.c` 的 `do_sort()` 函式
:::
建立 `list_sort.c` 及 `list_sort.h` ,前者包含 Linux 核心原始碼,後者包含必要的定義
首先需要將 lab0-c 所沒有的定義加進 `list_sort.h` ,如下所示
> list_cmp_func_t (位於 [include/linux/list_sort.h](https://github.com/torvalds/linux/blob/master/include/linux/list_sort.h))
> list_sort (位於 [include/linux/list_sort.h](https://github.com/torvalds/linux/blob/master/include/linux/list_sort.h))
> likely() (位於 [include/linux/compiler.h](https://github.com/torvalds/linux/blob/master/include/linux/compiler.h))
> unlikely() (位於 [include/linux/compiler.h](https://github.com/torvalds/linux/blob/master/include/linux/compiler.h))
以下為 `list_sort.h` 的程式碼
```c
#ifndef _LIST_SORT_H
#define _LIST_SORT_H
#include <stdint.h>
#include "list.h"
#define USE_LINUX_SORT 1
#define likely(x) __builtin_expect(!!(x), 1)
#define unlikely(x) __builtin_expect(!!(x), 0)
typedef int __attribute__((nonnull(2,3))) (*list_cmp_func_t)(void *, const struct list_head *, const struct list_head *);
void list_sort(void *priv, struct list_head *head, list_cmp_func_t cmp);
#endif /* End of _LIST_SORT_H_ */
```
接著在 makefile 的 `OBJS` 新增 `list_sort.o`
```shell
OBJS := qtest.o report.o console.o harness.o queue.o list_sort.o \
random.o dudect/constant.o dudect/fixture.o dudect/ttest.o \
linenoise.o
```
在 `qtest.c` 上建立函式 `cmp()`
```c
__attribute__((nonnull(2,3)))
int cmp(void *priv, const struct list_head *list1, const struct list_head *list2)
{
element_t *list1_entry = list_entry(list1, element_t, list);
element_t *list2_entry = list_entry(list2, element_t, list);
return strcmp(list1_entry->value, list2_entry->value) < 0 ? 0 : 1;
}
```
採雷小紀錄: 修改 `list_sort()` 的 prototype ,移除 `__attribute__((nonnull(2,3)))`
原因: 在測資 `trace-10-robust.cmd` , 會輸入 `NULL` 給 `head` ,原本在 `list_sort` 裡新增以下程式,但編譯器會編不過,錯誤訊息如下
```c
if(!head)
return;
```
```shell
CC list_sort.o
list_sort.c: In function ‘list_sort’:
list_sort.c:182:7: error: nonnull argument ‘head’ compared to NULL [-Werror=nonnull-compare]
182 | if(!head)
| ^
cc1: all warnings being treated as errors
make: *** [Makefile:50: list_sort.o] Error 1
```
解法: 移除 `__attribute__((nonnull(2,3)))` 即可
最後為了可以切換 `list_sort()` 及 `q_sort()` ,在 `list_sort.h` 新增新的巨集 `USE_LINUX_SORT`
→ 如果 `USE_LINUX_SORT` 為 `1` ,則使用 `list_sort()` ,反之則使用 `q_sort()`
修改 `qtest.c` 裡的函式 `do_sort()` ,如下所示
```c
if (exception_setup(true)) {
#if (USE_LINUX_SORT == 1)
list_sort(NULL, l_meta.l, cmp);
#else
q_sort(l_meta.l);
#endif
}
```
最後結果: `q_sort()` 及 `list_sort()` 都可以正常執行
### 比較自己實作的 merge sort 和 Linux 核心程式碼
首先這邊新增了一個測資 `trace-sort.cmd` ,如以下所示,使用 `qtest` 提供的命令 `time` 可以算出排序使用的時間
```
option fail 0
option malloc 0
new
ih RAND 100000
time
sort
time
```
輸入命令 `make check` 測試的示意如下 (這邊的排序是自己寫的 `q_sort()`)
```
./qtest -v 3 -f traces/trace-sort.cmd
cmd> option fail 0
cmd> option malloc 0
cmd> new
l = []
cmd> ih RAND 100000
l = [lhajuiamm nptum qdmcmfl nxyasup hemjdlj tetxmj pznuvn buyczxacm wbplbcipt vnbsjzf chtbzfy hpxnv upzkmvv bpvlvjz boocfn ucnxwj kemoxa azrxtyor duuks tebrmfg vfrxshpxu evpfw nmwfcp qrdld rahnv xqqroncbl qfppumm kwcrpvz puxlhhhcu ertih ... ]
cmd> time
Elapsed time = 0.087, Delta time = 0.087
cmd> sort
l = [aaaze aaazhb aabfh aabhbdo aabim aabpzxwgi aabqblx aabubl aacdjotav aacnapcm aacqj aacvkdmfq aacwcqd aacwg aacxso aadaocyzf aadldsjdq aadtapna aaeaciodj aaeakdh aaejadsff aaelwk aaeqgzt aaevmbu aafcfy aafeary aafjlj aafklh aafosjbi aafqy ... ]
cmd> time
Elapsed time = 0.216, Delta time = 0.129
Freeing queue
```
接著可以統計出 `q_sort()` 和 `list_sort()` 的排序的時間,這邊每一組測試三次
| 資料總數 | `q_sort()` 第一次 (s) | `q_sort()` 第二次 (s) | `q_sort()` 第三次 (s) | `list_sort()` 第一次 (s) | `list_sort()` 第二次 (s) | `list_sort()` 第三次 (s) |
| ------- | ----- | ----- | ----- | ----- | ----- | ----- |
| 100000 | 0.130 | 0.128 | 0.129 | 0.085 | 0.084 | 0.084 |
| 300000 | 0.504 | 0.496 | 0.492 | 0.305 | 0.306 | 0.306 |
| 500000 | 0.905 | 0.907 | 0.900 | 0.546 | 0.543 | 0.548 |
最後統計效能差異
| 資料總數 | `q_sort()` 平均 (s) | `list_sort()` 平均 (s) | `list_sort()` 和 `q_sort()` 效能比較 (%) |
| ------- | ----- | ----- | ------ |
| 100000 | 0.129 | 0.084 | 34.88% |
| 300000 | 0.497 | 0.306 | 38.43% |
| 500000 | 0.904 | 0.546 | 39.60% |
接著探索 `list_sort()` 比較快的原因,這邊參考[Linux 效能分析工具: Perf](http://wiki.csie.ncku.edu.tw/embedded/perf-tutorial#%E7%B0%A1%E4%BB%8B)
這邊採用的資料點為 500000 筆
輸入命令 `perf stat --repeat 5 -e cache-misses,cache-references,instructions,cycles ./qtest -f traces/trace-sort.cmd`
→ 程式會執行 5 次,並且自動平均,以下為 `q_sort()` 及 `list_sort()` 執行的結果
首先是 `q_sort()` 的結果
```shell
Performance counter stats for './qtest -f traces/trace-sort.cmd' (5 runs):
7430,5695 cache-misses # 69.606 % of all cache refs ( +- 0.92% )
1,0675,1677 cache-references ( +- 0.24% )
14,6569,3251 instructions # 0.46 insn per cycle ( +- 0.12% )
32,1192,4749 cycles ( +- 0.90% )
1.2041 +- 0.0109 seconds time elapsed ( +- 0.90% )
```
接著是 `list_sort()` 的結果
```shell
Performance counter stats for './qtest -f traces/trace-sort.cmd' (5 runs):
5698,1650 cache-misses # 69.677 % of all cache refs ( +- 0.84% )
8178,0080 cache-references ( +- 0.43% )
14,6192,8852 instructions # 0.65 insn per cycle ( +- 0.08% )
22,3567,4206 cycles ( +- 1.22% )
0.84768 +- 0.00768 seconds time elapsed ( +- 0.91% )
```
將輸出的結果做成表格,如下表所示
| | `q_sort()` | `list_sort()` |
| ---------------- | ------------- | -------------- |
| cycles | 3,211,924,749 | 2,235,674,206 |
| instructions | 1,465,693,251 | 1,461,928,852 |
| cache-references | 106,751,677 | 81,780,080 |
| cache-misses | 74,305,695 | 56,981,650 |
| insn per cycle | 0.46 | 0.65 |
這邊可以發現很有趣的事
1. 兩者的 instructions 執行的次數沒有差很多,但 cycles 卻差了非常多
2. `list_sort()` 的 cache miss 比 `q_sort()` 來的少很多
## 運用 Valgrind 排除 qtest 實作的記憶體錯誤
:::info
作業目標
- [x] 運用 Valgrind 排除 qtest 實作的記憶體錯誤,並透過 Massif 視覺化 “simulation” 過程中的記憶體使用量,需要設計對應的實驗
:::
由於不知從何下手,因此起頭參考[laneser](https://hackmd.io/@laneser/linux2022_lab0#2022q1-Homework1-lab0)的開發紀錄
:::danger
command 是「命令」,instruction 是「指令」,二者語意不同。
:notes: jserv
已修正!
:::
使用命令 `valgrind -q --leak-check=full ./qtest -f traces/trace-01-ops.cmd` 對檔案 `trace-01-ops.cmd` 進行測試
:::spoiler 執行結果
```shell
cmd> # Test of insert_head and remove_head
cmd> option fail 0
cmd> option malloc 0
cmd> new
l = []
cmd> ih gerbil
l = [gerbil]
cmd> ih bear
l = [bear gerbil]
cmd> ih dolphin
l = [dolphin bear gerbil]
cmd> rh dolphin
Removed dolphin from queue
l = [bear gerbil]
cmd> rh bear
Removed bear from queue
l = [gerbil]
cmd> rh gerbil
Removed gerbil from queue
l = []
cmd>
Freeing queue
==41370== 4 bytes in 1 blocks are still reachable in loss record 1 of 3
==41370== at 0x483B7F3: malloc (in /usr/lib/x86_64-linux-gnu/valgrind/vgpreload_memcheck-amd64-linux.so)
==41370== by 0x4A4E50E: strdup (strdup.c:42)
==41370== by 0x1108F9: linenoiseHistoryAdd (linenoise.c:1236)
==41370== by 0x11148C: linenoiseHistoryLoad (linenoise.c:1325)
==41370== by 0x10C6B0: main (qtest.c:951)
==41370==
==41370== 84 bytes in 19 blocks are still reachable in loss record 2 of 3
==41370== at 0x483B7F3: malloc (in /usr/lib/x86_64-linux-gnu/valgrind/vgpreload_memcheck-amd64-linux.so)
==41370== by 0x4A4E50E: strdup (strdup.c:42)
==41370== by 0x11086D: linenoiseHistoryAdd (linenoise.c:1236)
==41370== by 0x11148C: linenoiseHistoryLoad (linenoise.c:1325)
==41370== by 0x10C6B0: main (qtest.c:951)
==41370==
==41370== 160 bytes in 1 blocks are still reachable in loss record 3 of 3
==41370== at 0x483B7F3: malloc (in /usr/lib/x86_64-linux-gnu/valgrind/vgpreload_memcheck-amd64-linux.so)
==41370== by 0x1108B9: linenoiseHistoryAdd (linenoise.c:1224)
==41370== by 0x11148C: linenoiseHistoryLoad (linenoise.c:1325)
==41370== by 0x10C6B0: main (qtest.c:951)
==41370==
```
:::
可以看到有出現 still reachable 類型的 memory lost
> still reachable: 程式結束時有未釋放的記憶體,不過卻還有指標指著,通常會發生在 global 變數
跟著錯誤訊息的追隨路徑(`main` → `linenoiseHistoryLoad` → `linenoiseHistoryAdd`)可以找到沒有被釋放的記憶體空間
```c=
int linenoiseHistoryAdd(const char *line)
{
char *linecopy;
if (history_max_len == 0)
return 0;
/* Initialization on first call. */
if (history == NULL) {
history = malloc(sizeof(char *) * history_max_len);
if (history == NULL)
return 0;
memset(history, 0, (sizeof(char *) * history_max_len));
}
/* Don't add duplicated lines. */
if (history_len && !strcmp(history[history_len - 1], line))
return 0;
/* Add an heap allocated copy of the line in the history.
* If we reached the max length, remove the older line. */
linecopy = strdup(line);
if (!linecopy)
return 0;
if (history_len == history_max_len) {
free(history[0]);
memmove(history, history + 1, sizeof(char *) * (history_max_len - 1));
history_len--;
}
history[history_len] = linecopy;
history_len++;
return 1;
}
```
:::warning
錯誤訊息一共發生兩種
1. 第一種位於 `linenoiseHistoryAdd` 第 10 行的位置,即 `history = malloc(sizeof(char *) * history_max_len);`
2. 第二種位於 `linenoiseHistoryAdd` 第 22 行的位置,即 `linecopy = strdup(line);`
:::
在終端機輸入命令 `man strdup` 可以得到其訊息,其回傳的空間會經過 `malloc()`
> The strdup() function returns a pointer to a new string which is a duplicate of the string s. Memory for the new string is obtained with malloc(3), and can be freed with free(3).
查看了 `linenoiseHistoryAdd` 的原始碼後,發現 `histroy` 分配的每個指標指到的空間為 `linecopy` 分配的空間,因此只要將整個 `history` 釋放即可
```c
history[history_len] = linecopy;
```
接著在 `lineoise.c` 中發現 API `linenoiseAtExit()` 包含 `freeHistory()` ,可以用來釋放 `history`
```c
static void linenoiseAtExit(void)
{
disableRawMode(STDIN_FILENO);
freeHistory();
}
static void freeHistory(void)
{
if (history) {
int j;
for (j = 0; j < history_len; j++)
free(history[j]);
free(history);
}
}
```
:::info
實作步驟
1. 將 `static void linenoiseAtExit()` 改成 `void linenoiseAtExit()` ,目的在於讓 `console.c` 裡的 `finish_cmd()` 可以呼叫該 API ,最後可以把 `history` 釋放
2. 將 `linenoiseAtExit()` 的宣告從 `linenoise.c` 移到 `linenoise.h`
3. 在 `console.h` 裡 `include linenoise.h`
4. 在 `finish_cmd()` 裡呼叫 `linenoiseAtExit()`
:::
重新測試 `valgrind -q --leak-check=full ./qtest -f traces/trace-01-ops.cmd`
:::spoiler 測試結果
```shell
cmd> # Test of insert_head and remove_head
cmd> option fail 0
cmd> option malloc 0
cmd> new
l = []
cmd> ih gerbil
l = [gerbil]
cmd> ih bear
l = [bear gerbil]
cmd> ih dolphin
l = [dolphin bear gerbil]
cmd> rh dolphin
Removed dolphin from queue
l = [bear gerbil]
cmd> rh bear
Removed bear from queue
l = [gerbil]
cmd> rh gerbil
Removed gerbil from queue
l = []
cmd>
Freeing queue
```
:::
可以看到原本的錯誤訊息已經消失,表示 [memory leak](https://en.wikipedia.org/wiki/Memory_leak) 已經解決
接著從第一筆測資測試到最後一筆測資
在測試到 `trace-14-perf.cmd` 時發生其他的問題,這邊先用最後出現問題的地方將不同的錯誤訊息分類
:::spoiler 錯誤訊息分類
```shell
# 第一種
==339801== 8 bytes in 1 blocks are still reachable in loss record 2 of 37
==339801== at 0x483DD99: calloc (in /usr/lib/x86_64-linux-gnu/valgrind/vgpreload_memcheck-amd64-linux.so)
==339801== by 0x10CDD8: calloc_or_fail (report.c:208)
==339801== by 0x10D6A2: parse_args (console.c:154)
==339801== by 0x10D6A2: interpret_cmd (console.c:208)
==339801== by 0x10DF5D: cmd_select (console.c:584)
==339801== by 0x10E22A: run_console (console.c:657)
==339801== by 0x10C6E0: main (qtest.c:962)
# 第二種
==339801== 5 bytes in 1 blocks are still reachable in loss record 1 of 37
==339801== at 0x483B7F3: malloc (in /usr/lib/x86_64-linux-gnu/valgrind/vgpreload_memcheck-amd64-linux.so)
==339801== by 0x10CE7F: strsave_or_fail (report.c:230)
==339801== by 0x10D6D9: parse_args (console.c:157)
==339801== by 0x10D6D9: interpret_cmd (console.c:208)
==339801== by 0x10DF5D: cmd_select (console.c:584)
==339801== by 0x10E22A: run_console (console.c:657)
==339801== by 0x10C6E0: main (qtest.c:962)
# 第三種
==339801== 32 bytes in 1 blocks are still reachable in loss record 23 of 37
==339801== at 0x483B7F3: malloc (in /usr/lib/x86_64-linux-gnu/valgrind/vgpreload_memcheck-amd64-linux.so)
==339801== by 0x10CD55: malloc_or_fail (report.c:189)
==339801== by 0x10D7DF: add_cmd (console.c:90)
==339801== by 0x10C631: console_init (qtest.c:796)
==339801== by 0x10C631: main (qtest.c:945)
# 第四種
==339801== 48 bytes in 1 blocks are definitely lost in loss record 31 of 37
==339801== at 0x483B7F3: malloc (in /usr/lib/x86_64-linux-gnu/valgrind/vgpreload_memcheck-amd64-linux.so)
==339801== by 0x10E29C: test_malloc (harness.c:138)
==339801== by 0x10E800: q_insert_head (queue.c:62)
==339801== by 0x10C017: do_ih (qtest.c:201)
==339801== by 0x10D1CB: interpret_cmda (console.c:186)
==339801== by 0x10D717: interpret_cmd (console.c:209)
==339801== by 0x10DF5D: cmd_select (console.c:584)
==339801== by 0x10E22A: run_console (console.c:657)
==339801== by 0x10C6E0: main (qtest.c:962)
# 第五種
==339801== Invalid read of size 8
==339801== at 0x10AC42: do_sort (qtest.c:631)
==339801== by 0x10D1CB: interpret_cmda (console.c:186)
==339801== by 0x10D717: interpret_cmd (console.c:209)
==339801== by 0x10DF5D: cmd_select (console.c:584)
==339801== by 0x10E22A: run_console (console.c:657)
==339801== by 0x10C6E0: main (qtest.c:962)
==339801== Address 0xfffffffffffffff8 is not stack'd, malloc'd or (recently) free'd
```
:::
:::warning
錯誤訊息一共有五種
1. 第一種位於 `report.c` 的函式 `calloc_or_fail()` ,即 `void *p = calloc(cnt, bytes);` ,類型為 still reachable
2. 第二種位於 `report.c` 的函式 `strsave_or_fail()` ,即 `char *ss = malloc(len + 1);` ,類型為 still reachable
3. 第三種位於 `report.c` 的函式 `malloc_or_fail()` ,即 `void *p = malloc(bytes);` ,類型為 still reachable
4. 第四種位於 `harness.c` 的函式 `test_malloc()` ,即 `block_ele_t *new_block = malloc(size + sizeof(block_ele_t) + sizeof(size_t));` ,類型為 definitely lost
> definitely lost: 真的 memory leak
5. 第五種位於 `qtest.c` 的函式 `do_sort` ,即 `if (strcasecmp(item->value, next_item->value) > 0) {` ,類型為 Invalid read
> Invalid read: 表示被讀取的地址為 process 不可以讀取的地址, size 8 表示 process 想要讀取 8 bytes
:::
最後發現,以上全部的訊息是因為超出測試的時間,程式進中斷後結束時沒有把記憶體釋放而產生
首先查了一段時間後發現在 `harness.c` 裡有一個變數 `time_limit` ,將其數值增加(例如改成 `5`) ,以上的問題都會解決
```c
static int time_limit = 1;
```
以 `time_limit = 5` 重新輸入命令 `valgrind -q --leak-check=full ./qtest -f traces/trace-14-perf.cmd` 測試後,可以發現通過測試,且之後測資經過測試後也都通過
```c
cmd> # Test performance of insert_tail, reverse, and sort
cmd> option fail 0
cmd> option malloc 0
cmd> new
l = []
cmd> ih dolphin 1000000
l = [dolphin dolphin dolphin dolphin dolphin dolphin dolphin dolphin dolphin dolphin dolphin dolphin dolphin dolphin dolphin dolphin dolphin dolphin dolphin dolphin dolphin dolphin dolphin dolphin dolphin dolphin dolphin dolphin dolphin dolphin ... ]
cmd> it gerbil 1000000
l = [dolphin dolphin dolphin dolphin dolphin dolphin dolphin dolphin dolphin dolphin dolphin dolphin dolphin dolphin dolphin dolphin dolphin dolphin dolphin dolphin dolphin dolphin dolphin dolphin dolphin dolphin dolphin dolphin dolphin dolphin ... ]
cmd> reverse
l = [gerbil gerbil gerbil gerbil gerbil gerbil gerbil gerbil gerbil gerbil gerbil gerbil gerbil gerbil gerbil gerbil gerbil gerbil gerbil gerbil gerbil gerbil gerbil gerbil gerbil gerbil gerbil gerbil gerbil gerbil ... ]
cmd> sort
l = [dolphin dolphin dolphin dolphin dolphin dolphin dolphin dolphin dolphin dolphin dolphin dolphin dolphin dolphin dolphin dolphin dolphin dolphin dolphin dolphin dolphin dolphin dolphin dolphin dolphin dolphin dolphin dolphin dolphin dolphin ... ]
cmd>
Freeing queue
```
最後透過 massif 查看記憶體的狀況,輸入命令 `valgrind -v --tool=massif ./qtest -f traces/trace-01-ops.cmd`
→ 產生了名為 `massif.out.81335` 的檔案
接著使用 massif-visualizer 打開檔案,輸入命令 `massif-visualizer ./massif.out.81335` ,以下為最終結果
![](https://i.imgur.com/Mz0a6fN.png)
可以看到建立的動態記憶體最後歸為 `0` ,表示所有記憶體已被釋放
## 在 qtest 提供新的命令 shuffle
:::info
作業目標
- [x] 在 `qtest` 提供新的命令 `shuffle` ,允許藉由 [Fisher–Yates shuffle](https://en.wikipedia.org/wiki/Fisher%E2%80%93Yates_shuffle) 演算法,對佇列中所有節點進行洗牌 (shuffle) 操作
- [ ] 解釋 [select](https://man7.org/linux/man-pages/man2/select.2.html) 系統呼叫在本程式的使用方式,並分析 console.c 的實作,說明其中運用 CS:APP [RIO](http://csapp.cs.cmu.edu/2e/ch10-preview.pdf) 套件 的原理和考量點。可對照參考 CS:APP [第 10 章重點提示](https://hackmd.io/@sysprog/H1TtmVTTz)
:::
為了了解怎麼新增命令,這邊先分析整個 qtest 是如何從得到命令一直到使用 `console.c` 來添加以及執行命令的運作流程,並做紀錄
首先,從 `main` 可以得知 qtest 是如何接收使用者給出的命令,其中使用到了函式 `getopt()`,輸入命令 `man getopt` 可以得到其描述
> getopt is used to break up (parse) options in command lines for easy parsing by shell procedures, and to check for valid options. It uses the GNU getopt(3) routines to do this.
:::spoiler 原始碼
```c
while ((c = getopt(argc, argv, "hv:f:l:")) != -1) {
switch (c) {
case 'h':
usage(argv[0]);
break;
case 'f':
strncpy(buf, optarg, BUFSIZE);
buf[BUFSIZE - 1] = '\0';
infile_name = buf;
break;
case 'v': {
char *endptr;
errno = 0;
level = strtol(optarg, &endptr, 10);
if (errno != 0 || endptr == optarg) {
fprintf(stderr, "Invalid verbosity level\n");
exit(EXIT_FAILURE);
}
break;
}
case 'l':
strncpy(lbuf, optarg, BUFSIZE);
buf[BUFSIZE - 1] = '\0';
logfile_name = lbuf;
break;
default:
printf("Unknown option '%c'\n", c);
usage(argv[0]);
break;
}
}
```
:::
得知 qtest 如何得到命令後,接著繼續了解 `console.c` 如何添加命令
找到函式 `console_init()` 可以發現裡頭加入了所有佇列會用到的所有命令,因此可以在這裡定義 `shuffle` 命令
:::spoiler console_init()
```c
static void console_init()
{
ADD_COMMAND(new, " | Create new queue");
ADD_COMMAND(free, " | Delete queue");
ADD_COMMAND(
ih,
" str [n] | Insert string str at head of queue n times. "
"Generate random string(s) if str equals RAND. (default: n == 1)");
ADD_COMMAND(
it,
" str [n] | Insert string str at tail of queue n times. "
"Generate random string(s) if str equals RAND. (default: n == 1)");
ADD_COMMAND(
rh,
" [str] | Remove from head of queue. Optionally compare "
"to expected value str");
ADD_COMMAND(
rt,
" [str] | Remove from tail of queue. Optionally compare "
"to expected value str");
ADD_COMMAND(
rhq,
" | Remove from head of queue without reporting value.");
ADD_COMMAND(reverse, " | Reverse queue");
ADD_COMMAND(sort, " | Sort queue in ascending order");
ADD_COMMAND(
size, " [n] | Compute queue size n times (default: n == 1)");
ADD_COMMAND(show, " | Show queue contents");
ADD_COMMAND(dm, " | Delete middle node in queue");
ADD_COMMAND(
dedup, " | Delete all nodes that have duplicate string");
ADD_COMMAND(swap,
" | Swap every two adjacent nodes in queue");
add_param("length", &string_length, "Maximum length of displayed string",
NULL);
add_param("malloc", &fail_probability, "Malloc failure probability percent",
NULL);
add_param("fail", &fail_limit,
"Number of times allow queue operations to return false", NULL);
}
```
:::
,開始分析巨集函式 `ADD_COMMAND` ,可以發現 `ADD_COMMAND` 被定義為, `add_cmd(#cmd, do_##cmd, msg)`
```c
#define ADD_COMMAND(cmd, msg) add_cmd(#cmd, do_##cmd, msg)
```
查看函式 `add_cmd()` ,可以看到一個有趣的結構 `CELE` (位於 `add_cmd()` 第 10 行, `cmd_ele` 為結構 `CLEL` 的變數)
發現該函式做的事情就只是把輸入的命令存到一個名為 `CELE` 的結構 (從 11 行到 14 行)
:::spoiler add_cmd()
```c=
void add_cmd(char *name, cmd_function operation, char *documentation)
{
cmd_ptr next_cmd = cmd_list;
cmd_ptr *last_loc = &cmd_list;
while (next_cmd && strcmp(name, next_cmd->name) > 0) {
last_loc = &next_cmd->next;
next_cmd = next_cmd->next;
}
cmd_ptr ele = (cmd_ptr) malloc_or_fail(sizeof(cmd_ele), "add_cmd");
ele->name = name;
ele->operation = operation;
ele->documentation = documentation;
ele->next = next_cmd;
*last_loc = ele;
}
```
:::
分析結構 `CELE` 後,可以發現 `CELE` 為 linked list 的結構形式, 其中比較重要的變數為 `operation` ,被定義為函式指標,存放不同命令的函式地址。
由此可知,在一開始初始化的時候,所有命令會被存到一組 linked list 裡
```c
/* Each command defined in terms of a function */
typedef bool (*cmd_function)(int argc, char *argv[]);
typedef struct CELE cmd_ele, *cmd_ptr;
struct CELE {
char *name;
cmd_function operation;
char *documentation;
cmd_ptr next;
};
```
最後開始分析 `console.c` 是如何執行命令
在 `main` 找到函式 `run_console()` ,仔細觀察原始碼後可以得知,執行命令一共分為兩種模式
- 一種為讓使用者一個命令接著一個命令輸入並執行 (`run_console()` 的第 8 行到第 15 行)
- 另一種則為直接讀取一個檔案執行 ( `run_console()` 的第 16 行到第 19 行 )
:::spoiler run_console()
```c=
bool run_console(char *infile_name)
{
if (!push_file(infile_name)) {
report(1, "ERROR: Could not open source file '%s'", infile_name);
return false;
}
if (!has_infile) {
char *cmdline;
while ((cmdline = linenoise(prompt)) != NULL) {
interpret_cmd(cmdline);
linenoiseHistoryAdd(cmdline); /* Add to the history. */
linenoiseHistorySave(HISTORY_FILE); /* Save the history on disk. */
linenoiseFree(cmdline);
}
} else {
while (!cmd_done())
cmd_select(0, NULL, NULL, NULL, NULL);
}
return err_cnt == 0;
}
```
:::
首先先討論第一種模式,在 `run_console()` 第 11 行可以找到函式 `interpret_cmd()` ,用途在於解析使用者輸入的命令並執行
- 解析命令使用函式 `parse_args()`
- 執行命令使用函式 `interpret_cmda()`
:::spoiler interpret_cmd()
```c
/* Execute a command from a command line */
static bool interpret_cmd(char *cmdline)
{
if (quit_flag)
return false;
#if RPT >= 6
report(6, "Interpreting command '%s'\n", cmdline);
#endif
int argc;
char **argv = parse_args(cmdline, &argc);
bool ok = interpret_cmda(argc, argv);
for (int i = 0; i < argc; i++)
free_string(argv[i]);
free_array(argv, argc, sizeof(char *));
return ok;
}
```
:::
接著探討很重要的用來執行命令的函式 `interpret_cmda()` ,可以歸納以下幾點
1. 首先利用輸入命令的名字開始尋找是否有該命令存在,走訪 linked list 得知 (`interpret_cmd()` 的第 9 行和第 10 行)
2. 如果有該命令的話,則利用函式指標開始呼叫該命令要執行的函式 (`interpret_cmd()` 的第 12 行)
:::spoiler interpret_cmda
```c=
/* Execute a command that has already been split into arguments */
static bool interpret_cmda(int argc, char *argv[])
{
if (argc == 0)
return true;
/* Try to find matching command */
cmd_ptr next_cmd = cmd_list;
bool ok = true;
while (next_cmd && strcmp(argv[0], next_cmd->name) != 0)
next_cmd = next_cmd->next;
if (next_cmd) {
ok = next_cmd->operation(argc, argv);
if (!ok)
record_error();
} else {
report(1, "Unknown command '%s'", argv[0]);
record_error();
ok = false;
}
return ok;
}
```
:::
接下來換成討論第二種執行情況,也就是直接讀取檔案並執行
從之前提到的函式 `run_console()` ,查看函式 `cmd_select()` ,可以歸類以下幾點
1. 在第 33 行,這邊用到一個函式庫 `select.h` 的函式 `select()` ==詳細資料之後補上==
2. 在第 46 行的位置,開始執行命令
:::spoiler cmd_select()
```c=
int cmd_select(int nfds,
fd_set *readfds,
fd_set *writefds,
fd_set *exceptfds,
struct timeval *timeout)
{
int infd;
fd_set local_readset;
if (cmd_done())
return 0;
if (!block_flag) {
/* Process any commands in input buffer */
if (!readfds)
readfds = &local_readset;
/* Add input fd to readset for select */
infd = buf_stack->fd;
FD_SET(infd, readfds);
if (infd == STDIN_FILENO && prompt_flag) {
printf("%s", prompt);
fflush(stdout);
prompt_flag = true;
}
if (infd >= nfds)
nfds = infd + 1;
}
if (nfds == 0)
return 0;
int result = select(nfds, readfds, writefds, exceptfds, timeout);
if (result <= 0)
return result;
infd = buf_stack->fd;
if (readfds && FD_ISSET(infd, readfds)) {
/* Commandline input available */
FD_CLR(infd, readfds);
result--;
if (has_infile) {
char *cmdline;
cmdline = readline();
if (cmdline)
interpret_cmd(cmdline);
}
}
return result;
}
```
:::
:::success
結論
1. 一開始使用函式 `getopt()` 取得使用者對 qtest 的輸入
2. 接著在函式 `console_init()` 裡一路追蹤,得知了所有的命令一開始都會被儲存在一個 linked list 上
3. 執行命令一共有兩種模式,等待使用者輸入以及讀取檔案
5. 利用函式指標執行每種不同的命令
:::
從上述的分析,可以得知 qtest 從執行、初始化命令到執行命令的過程,並想出了增加命令的實作方法,如下所示
:::info
實作步驟
1. 模仿其他的命令新增函式 `do_shuffle()` ,作為命令 `shuffle` 的執行函式
2. 實作函式 `q_shuffle()` ,用來實現 [Fisher–Yates shuffle 演算法](https://en.wikipedia.org/wiki/Fisher%E2%80%93Yates_shuffle)
3. 在 `console_init()` 裡新增 `shuffle` 命令
:::
在 `qtest.c` 新增函式 `do_shuffle()`
```c
static bool do_shuffle(int argc, char *argv[])
{
if (argc != 1) {
report(1, "%s takes no arguments", argv[0]);
return false;
}
if (!l_meta.l)
report(3, "Warning: Try to access null queue");
error_check();
bool ok = true;
if (exception_setup(true))
ok = q_shuffle(l_meta.l);
exception_cancel();
show_queue(3);
return ok && !error_check();
}
```
由於 `queue.h` 無法更改,因此將 `q_shuffle()` 宣告在 `do_shuffle()` 的上面
這邊參考 [Fisher–Yates shuffle](https://en.wikipedia.org/wiki/Fisher%E2%80%93Yates_shuffle#Fisher_and_Yates'_original_method) 裡提到的 Pencil-and-paper method (這邊使用 Fisher and Yates' original method) ,流程如下表所示
| Range | Roll | Scratch | Result |
| ----- | ---- | ------- | ------- |
| | | A B C D | |
| 1~4 | 2 | A C D | B |
| 1~3 | 3 | A C | B D |
| 1~2 | 1 | C | B D A |
| 1 | 1 | | B D A C |
- 思考方式
1. 變數 `number` 決定目前要取第幾個 element
2. 取得節點後,將該節點 remove 並加到 linked list 的最後一個
3. 重複這些動作後,最一開始移動的節點最後會回到第一個
:::spoiler q_shuffle()
```c
static bool q_shuffle(struct list_head *head)
{
if (!head)
return false;
int size = q_size(head);
while (size) {
struct list_head *tmp = head->next;
int number = rand() % size--;
for(int i = 0; i < number; i++)
tmp = tmp->next;
list_del(tmp);
list_add_tail(tmp, head);
}
return true;
}
```
:::
在 `console_init()` 裡新增 `shuffle` 命令
```c
ADD_COMMAND(
shuffle, " | Shuffle the queue by using Fisher–Yates shuffle");
```
接著輸入命令 `./qtest` 以及 `help` ,可以看到所有可使用的命令,可以發現 `shuffle` 已被加入
```diff
./qtest
cmd> help
Commands:
# ... | Display comment
dedup | Delete all nodes that have duplicate string
dm | Delete middle node in queue
free | Delete queue
help | Show documentation
ih str [n] | Insert string str at head of queue n times. Generate random string(s) if str equals RAND. (default: n == 1)
it str [n] | Insert string str at tail of queue n times. Generate random string(s) if str equals RAND. (default: n == 1)
log file | Copy output to file
new | Create new queue
option [name val] | Display or set options
quit | Exit program
reverse | Reverse queue
rh [str] | Remove from head of queue. Optionally compare to expected value str
rhq | Remove from head of queue without reporting value.
rt [str] | Remove from tail of queue. Optionally compare to expected value str
show | Show queue contents
+ shuffle | Shuffle the queue by using Fisher–Yates shuffle
size [n] | Compute queue size n times (default: n == 1)
sort | Sort queue in ascending order
source file | Read commands from source file
swap | Swap every two adjacent nodes in queue
time cmd arg ... | Time command execution
Options:
echo 1 Do/don't echo commands
error 5 Number of errors until exit
fail 30 Number of times allow queue operations to return false
length 1024 Maximum length of displayed string
malloc 0 Malloc failure probability percent
simulation 0 Start/Stop simulation mode
verbose 4 Verbosity level
cmd>
```
:::spoiler 測試程式結果
```shell
cmd> it 9
l = [1 2 3 4 5 6 7 8 9]
cmd> shuffle
l = [6 2 7 3 1 4 9 5 8]
cmd> shuffle
l = [4 8 1 5 2 7 6 9 3]
cmd> sort
l = [1 2 3 4 5 6 7 8 9]
cmd> shuffle
l = [1 4 8 3 2 5 6 7 9
cmd> sort
l = [1 2 3 4 5 6 7 8 9]
cmd> shuffle
l = [2 6 3 7 8 5 9 1 4]
```
:::
## 在 qtest 提供新的命令 web
:::info
作業要求
- [x] 在 qtest 提供新的命令 web,提供 web 伺服器功能,注意: web 伺服器運作過程中,qtest 仍可接受其他命令
- 依據前述說明,嘗試整合 [tiny-web-server](https://github.com/7890/tiny-web-server)
:::
首先查看了 tiny-web-server 後,發現有些步驟會用到 `tiny-web-server` 的函式 ,因此把 `tiny.c` 額外分出兩個檔案 `tiny_function.c` 及 `tiny_function.h` ,以下為 `tiny_function.h` 程式碼
:::spoiler `tiny_function.h`
```c
#ifndef _TINY_FUNCTION_H
#define _TINY_FUNCTION_H
#include <arpa/inet.h> /* inet_ntoa */
#include <dirent.h>
#include <errno.h>
#include <fcntl.h>
#include <netinet/in.h>
#include <netinet/tcp.h>
#include <signal.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <sys/sendfile.h>
#include <sys/socket.h>
#include <sys/stat.h>
#include <sys/types.h>
#include <time.h>
#include <unistd.h>
#define LISTENQ 1024 /* second argument to listen() */
#define MAXLINE 1024 /* max length of a line */
#define RIO_BUFSIZE 8192
#ifndef DEFAULT_PORT
#define DEFAULT_PORT 9999 /* use this port if none given as arg to main() */
#endif
#ifndef FORK_COUNT
#define FORK_COUNT 4
#endif
#ifndef NO_LOG_ACCESS
#define LOG_ACCESS
#endif
typedef struct {
int rio_fd; /* descriptor for this buf */
int rio_cnt; /* unread byte in this buf */
char *rio_bufptr; /* next unread byte in this buf */
char rio_buf[RIO_BUFSIZE]; /* internal buffer */
} rio_t;
/* Simplifies calls to bind(), connect(), and accept() */
typedef struct sockaddr SA;
typedef struct {
char filename[512];
off_t offset; /* for support Range */
size_t end;
} http_request;
typedef struct {
const char *extension;
const char *mime_type;
} mime_map;
// https://developer.mozilla.org/en-US/docs/Web/HTTP/Basics_of_HTTP/MIME_types/Common_types
void rio_readinitb(rio_t *rp, int fd);
ssize_t writen(int fd, void *usrbuf, size_t n);
ssize_t rio_readlineb(rio_t *rp, void *usrbuf, size_t maxlen);
void format_size(char *buf, struct stat *stat);
void handle_directory_request(int out_fd, int dir_fd, char *filename);
int open_listenfd(int port);
void url_decode(char *src, char *dest, int max);
void parse_request(int fd, http_request *req);
void log_access(int status, struct sockaddr_in *c_addr, http_request *req);
void client_error(int fd, int status, char *msg, char *longmsg);
void serve_static(int out_fd, int in_fd, http_request *req, size_t total_size);
void process(int fd, struct sockaddr_in *clientaddr);
void print_help();
#endif /* _TINY_FUNCTION_H */
```
:::
修改 Makefile ,讓 lab0-c 可以使用 tiny-web-server 的函式,另外這邊新增命令 `make tiny` ,保留原本 tiny-web-server 的功能
```diff
OBJS := qtest.o report.o console.o harness.o queue.o list_sort.o \
random.o dudect/constant.o dudect/fixture.o dudect/ttest.o \
- linenoise.o
+ linenoise.o tiny_function.o
+ tiny: tiny.c tiny_function.c
+ $(CC) $(CFLAGS) -o $@ $^
+ ./$@
```
開始新增命令 `web` ,在 `init_cmd()` 新增以下程式碼
```c
ADD_COMMAND(web, " | Use web server");
```
另外,由於會使用到的函式分散在不同檔案,這邊宣告一些全域變數(位於 `linenoise.h`)
```c
struct sockaddr_in clientaddr;
socklen_t clientlen;
int listenfd;
bool noise;
```
實作函式 `do_web()` ,這邊參考 `tiny.c` 的函式 `main()` 的作法,使用預設的 port `9999`
```c
if(!listenfd) {
listenfd = open_listenfd(DEFAULT_PORT);
noise = false;
printf("listen on port %d, fd is %d\n", DEFAULT_PORT, listenfd);
} else
printf("Server has been launched\n");
return true;
```
接著的步驟幾乎都是參考[作業說明](https://hackmd.io/@sysprog/linux2022-lab0),跟著整合 [tiny-web-server](https://github.com/7890/tiny-web-server) 的步驟慢慢做
修改 `run_console()` ,使其可以選擇使用 `linenoise()` 或是 `cmd_select()`
```c
if (!has_infile) {
char *cmdline;
while (noise && (cmdline = linenoise(prompt)) != NULL) {
interpret_cmd(cmdline);
linenoiseHistoryAdd(cmdline); /* Add to the history. */
linenoiseHistorySave(
HISTORY_FILE); /* Save the history on disk. */
linenoiseFree(cmdline);
}
if (!noise) {
while (!cmd_done())
cmd_select(0, NULL, NULL, NULL, NULL);
}
} else {
while (!cmd_done())
cmd_select(0, NULL, NULL, NULL, NULL);
}
```
修改 `cmd_select()`
```diff
int cmd_select(int nfds,
fd_set *readfds,
fd_set *writefds,
fd_set *exceptfds,
struct timeval *timeout)
{
int infd;
fd_set local_readset;
if (cmd_done())
return 0;
if (!block_flag) {
/* Process any commands in input buffer */
if (!readfds)
readfds = &local_readset;
/* Add input fd to readset for select */
infd = buf_stack->fd;
+ FD_ZERO(readfds);
FD_SET(infd, readfds);
+ /* If web not ready listen */
+ if (listenfd != -1)
+ FD_SET(listenfd, readfds);
if (infd == STDIN_FILENO && prompt_flag) {
printf("%s", prompt);
fflush(stdout);
prompt_flag = true;
}
if (infd >= nfds)
nfds = infd + 1;
+ if (listenfd >= nfds)
+ nfds = listenfd + 1;
}
if (nfds == 0)
return 0;
int result = select(nfds, readfds, writefds, exceptfds, timeout);
if (result <= 0)
return result;
infd = buf_stack->fd;
if (readfds && FD_ISSET(infd, readfds)) {
/* Commandline input available */
FD_CLR(infd, readfds);
result--;
char *cmdline;
cmdline = readline();
if (cmdline)
interpret_cmd(cmdline);
}
+ else if (readfds && FD_ISSET(listenfd, readfds)) {
+ FD_CLR(listenfd, readfds);
+ result--;
+ int connfd = accept(listenfd, (SA *) &clientaddr, &clientlen);
+ char *p = url_process(connfd, &clientaddr);
+ if (p)
+ interpret_cmd(p);
+ free(p);
+ close(connfd);
+ }
return result;
}
```
繼續照著作業說明實作,在 `linenoiseEdit` 新增程式碼
```c
if(listenfd) {
fd_set set;
FD_ZERO(&set);
FD_SET(listenfd, &set);
FD_SET(stdin_fd, &set);
clientlen = sizeof(clientaddr);
int rv = select(listenfd + 1, &set, NULL, NULL, NULL);
int connfd;
switch (rv) {
case -1:
perror("select"); /* an error occurred */
continue;
case 0:
printf("timeout occurred\n"); /* a timeout occurred */
continue;
default:
if (FD_ISSET(listenfd, &set)) {
connfd = accept(listenfd, (SA *) &clientaddr, &clientlen);
char *p = url_process(connfd, &clientaddr);
strncpy(buf, p, strlen(p) + 1);
close(connfd);
free(p);
return strlen(p);
} else if (FD_ISSET(stdin_fd, &set)) {
nread = read(l.ifd, &c, 1);
if (nread <= 0)
return l.len;
}
break;
}
} else {
nread = read(l.ifd, &c, 1);
if (nread <= 0)
return l.len;
}
```
最後新增函式 `url_process()` ,透過 HTTP 的 GET method 傳送要執行的命令,這邊參考原本 tiny-web-server 的實作,不同在於 tiny-web-server 會讀取所在目錄的檔案,而 `url_process()` 指單純開啟空白頁面,實作如下
```c
char *url_process(int fd, struct sockaddr_in *clientaddr)
{
#ifdef LOG_ACCESS
printf("accept request, fd is %d, pid is %d\n", fd, getpid());
#endif
http_request req;
parse_request(fd, &req);
int status = 200;
char buf[MAXLINE];
snprintf(buf, MAXLINE, "HTTP/1.1 %d\r\n",status);
writen(fd, buf, strlen(buf));
char *p = req.filename;
/* Change '/' to ' ' */
while (*p) {
++p;
if (*p == '/') {
*p = ' ';
}
}
#ifdef LOG_ACCESS
log_access(status, clientaddr, &req);
#endif
char *ret = malloc(strlen(req.filename) + 1);
strncpy(ret, req.filename, strlen(req.filename) + 1);
return ret;
}
```
最後結果如以下所示,可以成功開啟網頁,以及接收使用者命令
```shell
cmd> web
listen on port 9999, fd is 3
cmd> accept request, fd is 4, pid is 84485
0.0.0.0:0 200 - '.' (text/plain)
cmd> accept request, fd is 4, pid is 84485
127.0.0.1:33774 200 - 'new' (text/plain)
l = []
cmd> accept request, fd is 4, pid is 84485
127.0.0.1:33776 200 - 'ih 4' (text/plain)
l = [4]
cmd> accept request, fd is 4, pid is 84485
127.0.0.1:33778 200 - 'ih 5' (text/plain)
l = [5 4]
cmd> accept request, fd is 4, pid is 84485
```
不過還有一個小問題,開啟 server 之後,會經過很久的時間才能從終端機輸入命令,目前還沒有解決
## 進度紀錄
:::warning
這些都在該 git log 描述,不需要在此提及,有 git commit 簡述即可。
:notes: jserv
已修正!
:::
## 參考資料
- [K01: lab0](https://hackmd.io/@sysprog/linux2022-lab0)
- [你所不知道的 C 語言: linked list 和非連續記憶體](https://hackmd.io/@sysprog/c-linked-list)
- [Linux 核心原始程式碼巨集: container_of](https://hackmd.io/@sysprog/linux-macro-containerof)
- [Floyd’s Cycle detection](https://en.wikipedia.org/wiki/Cycle_detection)
- [quick-sort.c](https://github.com/sysprog21/linux-list/blob/master/examples/quick-sort.c)
- [edotor](https://edotor.net/)