If you have been compiled kernel from source code,and you will have linux-headers-x.x.x (Personality i compile kernel with make deb-pkg command).
Next we will use add ours custom defined system call into kernel.
The aricle based on kernel-v5.0.0-rc6
The first: please know what bits belongs to your system.
uname -m #uname -m => i686
So, my machine system is 32 bits
Yes, you can git clone kernel or download kernel tar file.
You extact the source code and then cd root directory and new create a dir:
mkdir eudyptula && cd eudyptula
Open the file arch/x86/kernel/syscall_table_32.S and add the following line. (64 bits should open arch/x86/kernel/syscall_table_64.S)
387 i386 eudyptula sys_eudyptula __ia32_sys_eudyptula
This is the last one entry in the file above.
Source: Documentation/process/adding-syscalls.rst
As an open source code management tool, Git’s excellent collaboration features quickly capture the market for code management and become a qualified programmer. Standard. You can learn this skill from 0. This article is a brief introduction to one of the features: Combine multiple commits into one Commit.
Git作为一个开源的代码管理工具，其优异的协作特性迅速占领代码管理的市场，并且也成为一个合格程序员的 标配。你可以从0开始学习这个技能，本篇文章就是简单介绍其中的一个特性：将多个commits合并成一个 commit。
The first thing to do is to invoke git to start an interactive rebase session: 首先要做的是调用git来启动交互式rebase会话：
git rebase --interactive HEAD^[N]
Or, shorter: 或者，简写为：
git rebase -i HEAD^[N]
Where N is the number of commits you want to join, starting from the most recent
one. For example, i will show your my git log to show how to use it.
这里的N是你想要合并的commit的数量。接下来我将会使用我自己的git log进行展示。
commit 53130bcdf0355382a6edbf7b3f42cfd9bb0ea7f2
Author: Bo Yu <[email protected]>
Date: Wed Feb 20 16:16:42 2019 +0800
x86: fix warnings from checkpatch
Signed-off-by: Bo Yu <[email protected]>
commit 30e75b0d0d04b2eacbdbd7042f993c7e0522eb89
Author: Bo Yu <[email protected]>
Date: Wed Feb 20 16:12:12 2019 +0800
x86: Add a syscall for task 15
Please consider the patch into v5.0.0-rc7
Signed-off-by: Bo Yu <[email protected]>
commit 40e196a906d969fd10d885c692d2674b3d657006
Merge: b5372fe5dc84 1765f5dcd009
Author: Linus Torvalds <[email protected]>
Date: Tue Feb 19 16:13:19 2019 -0800
Merge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net
Pull networking fixes from David Miller:
1) Fix suspend and resume in mt76x0u USB driver, from Stanislaw
Gruszka.
2) Missing memory barriers in xsk, from Magnus Karlsson.
3) rhashtable fixes in mac80211 from Herbert Xu.
Now I want to combine 53130b and 30e75b into one commit. 现在我想将 53130b 和 30e75b 合并为一个commit.
In fact, there is one commit only to squash, simple.
git rebase -i 40e196a906d969fd10d885c69
小心，这个操作危害性很大，如果有问题，请看最后的提示。
pick 30e75b0d0d04 x86: Add a syscall for task 15
pick 53130bcdf035 x86: fix warnings from checkpatch
# 变基 40e196a906d9..53130bcdf035 到 40e196a906d9（2 个提交）
#
# 命令:
# p, pick = 使用提交
# r, reword = 使用提交，但修改提交说明
# e, edit = 使用提交，但停止以便进行提交修补
# s, squash = 使用提交，但和前一个版本融合
# f, fixup = 类似于 "squash"，但丢弃提交说明日志
# x, exec = 使用 shell 运行命令（此行剩余部分）
# d, drop = 删除提交
#
# 这些行可以被重新排序；它们会被从上至下地执行。
#
#
# 如果您在这里删除一行，对应的提交将会丢失。
#
# 然而，如果您删除全部内容，变基操作将会终止。
#
# 注意空提交已被注释掉
注意，将上面第二行日志的pick 53130bcd的pick改为s,然后保存退出。
然后就进行变基了。进行完变基过程，接着又进入了一个vim 日志的系统：
# 这是一个 2 个提交的组合。
# 这是第一个提交说明：
x86: Add a syscall for task 15
Please consider the patch into v5.0.0-rc7
Signed-off-by: Bo Yu <[email protected]>
# 这是提交说明 #2：
x86: fix warnings from checkpatch
Signed-off-by: Bo Yu <[email protected]>
# 请为您的变更输入提交说明。以 '#' 开始的行将被忽略，而一个空的提交
# 说明将会终止提交。
#
# 日期： Wed Feb 20 16:12:12 2019 +0800
#
# 交互式变基操作正在进行中；至 40e196a906d9
# 最后的命令已完成（2 条命令被执行）：
# pick 30e75b0d0d04 x86: Add a syscall for task 15
# s 53130bcdf035 x86: fix warnings from checkpatch
# 未剩下任何命令。
# 您在执行变基操作时编辑提交。
#
# 要提交的变更：
# 修改： arch/x86/entry/syscalls/syscall_32.tbl
# 修改： include/linux/syscalls.h
# 修改： kernel/sys.c
这个时候，就可以把这两个commit合并成一个，使用’#’注释掉相关的语句。 然后保存退出，你在屏幕上就会看到：
git rebase -i 40e196a906d969fd10d885c69
[分离头指针 7947f9d65c6c] x86: Add a syscall for task 15
Date: Wed Feb 20 16:12:12 2019 +0800
3 files changed, 14 insertions(+)
成功变基并更新 detached HEAD。
No， not at all.You just put git commit-hash , which efore the first one you want
to rewrite from.
不，一点也不用，你仅仅把想要合并的那些commits中的第一个commit的hash值写上就行。
c b a old 1---->2----->3----->4-----5---->6
For instance,a, b,c is the commits you want to squash,so,you just
git rebase --interactive old
Remenber, please put a replace with hash-value 例如，上面中的a,b,c是你想要合并的commits，你仅仅把上面命令中的a替换成commit value就行
When you type command above,your editor will show up : 当你键入上面的命令，你的编辑器将会展示：
pick 9d814b2a0e78 fs/proc: fix minor error
#变基 acb685c130ea..9d814b2a0e78 到 acb685c130ea(1个提交)
...
Save the file and exit 保存文件并且退出
快捷键 定义 操作
c 撰写 撰写新邮件
作为一个c的使用者，void恐怕是你比较喜欢的一个单词，为什么呢？因为简单，记得在 编译原理 这门课中，任课老师特意讲了下 空操作 这个概念，也算学过这门课了，哎！
今天这练习一个内核oops的检测时，遇到了一个有意思的事情：
error: function declaration isn’t a prototype
这个函数是这样定义的：
static void create_oops() {
*(innt *)0 = 0;
}
static int __init my_oops_init(void)
{
...
create_oops();
...
}
注意在下面的使用中，这里给它传递了一个空的参数，以前没有注意到编译器会报这个错，这次编译器就抛出了这个错误。其实，按照目前(有待考证)gcc的特性，如果不想给一个函数传递函数，那么就要在定义的使用使用void(在形参位置)填充，这样就明确的告诉编译器，该函数没有参数。
This is a note i refer to the article
So, i have to record it simply.
In fact, if you have interesting to learn kernel programming, the oops is your better choice or tool.There is plenty of resources on Internet.And if you have some issues on that, i am grate happy to receive your email.
in oops.c
#include <linux/kernel.h>
#include <linux/module.h>
#include <linux/init.h>
static void create_oops(void) {
*(int *)0 = 0;
}
static int __init my_oops_init(void)
{
pr_info("oops from the module\n");
create_oops();
return 0;
}
static void __exit my_oops_exit(void) {
pr_info("Goodbye\n");
}
module_init(my_oops_init);
module_exit(my_oops_exit);
And you have to editor the Makefile file, like below:
obj-m += oops.o
all:
make -C /lib/modules/$(shell uname -r)/build M=$(PWD) modules
CFLAGS_task12-2.o := -DDEBUG
clean:
make -C /lib/modules/$(shell uname -r)/build M=$(PWD) clean
In the file, there is an entry may be different from other tutorial.Yes, CFLAGS, it is a flag how to compile the module.Why does we use the option? The answer is enable DEBUG for kernel module. In oops.c file, we can print some message with pr_info, but if we use pr_debug, it may be not display the information without -DDEBUG option.
So, then we:
make && sudo insmod oops.ko
And the result:
yubo@debian:~/lit$ make
make -C /lib/modules/5.0.0-rc4/build M=/home/yubo/lit modules
make[1]: Entering directory '/usr/src/linux-headers-5.0.0-rc4'
CC [M] /home/yubo/lit/oops.o
Building modules, stage 2.
MODPOST 1 modules
WARNING: modpost: missing MODULE_LICENSE() in /home/yubo/lit/oops.o
see include/linux/module.h for more information
CC /home/yubo/lit/oops.mod.o
LD [M] /home/yubo/lit/oops.ko
make[1]: Leaving directory '/usr/src/linux-headers-5.0.0-rc4'
Haha, it will kill the prograss after you insmod the module.
Let’s cat dmesg:
[ 3341.170623] oops: loading out-of-tree module taints kernel.
[ 3341.170628] oops: module license 'unspecified' taints kernel.
[ 3341.170629] Disabling lock debugging due to kernel taint
[ 3341.170907] oops from the module
[ 3341.170913] BUG: unable to handle kernel NULL pointer dereference at 00000000
[ 3341.170917] #PF error: [WRITE]
[ 3341.170919] *pdpt = 0000000029111001 *pde = 0000000000000000
[ 3341.170924] Oops: 0002 [#1] SMP PTI
[ 3341.170928] CPU: 3 PID: 4066 Comm: insmod Tainted: P O 5.0.0-rc4 #3
[ 3341.170932] Hardware name: Acer Aspire E1-471G/EA40_HC, BIOS V1.24 10/11/2012
[ 3341.170937] EIP: my_oops_init+0x12/0x1000 [oops]
[ 3341.170942] Code: Bad RIP value.
[ 3341.170945] EAX: 00000014 EBX: f86d1000 ECX: f6f755ac EDX: 00000007
[ 3341.170947] ESI: 00000001 EDI: f86d4000 EBP: e7ce7dec ESP: e7ce7de8
[ 3341.170950] DS: 007b ES: 007b FS: 00d8 GS: 00e0 SS: 0068 EFLAGS: 00010246
[ 3341.170953] CR0: 80050033 CR2: f86d3fe8 CR3: 27c32000 CR4: 000406f0
[ 3341.170955] Call Trace:
[ 3341.170962] do_one_initcall+0x42/0x1ae
[ 3341.170966] ? _cond_resched+0x18/0x40
[ 3341.170972] ? do_init_module+0x21/0x1dd
[ 3341.170975] do_init_module+0x50/0x1dd
[ 3341.170978] load_module+0x1d06/0x2540
[ 3341.170982] ? kernel_read_file+0x222/0x250
[ 3341.170986] sys_finit_module+0xa7/0xe0
[ 3341.170990] do_fast_syscall_32+0x7f/0x1d0
[ 3341.170993] entry_SYSENTER_32+0x6b/0xbe
[ 3341.170995] EIP: 0xb7f81851
[ 3341.170997] Code: 8b 98 58 cd ff ff 85 d2 89 c8 74 02 89 0a 5b 5d c3 8b 04 24 c3 8b 14 24 c3 8b 1c 24 c3 8b 3c 24 c3 51 52 55 89 e5 0f 34 cd 80 <5d> 5a 59 c3 90 90 90 90 8d 76 00 58 b8 77 00 00 00 cd 80 90 8d 76
[ 3341.170999] EAX: ffffffda EBX: 00000003 ECX: 004dc970 EDX: 00000000
[ 3341.171001] ESI: 01d19d90 EDI: 01d1a7a8 EBP: 004dae9c ESP: bfe2a18c
[ 3341.171003] DS: 007b ES: 007b FS: 0000 GS: 0033 SS: 007b EFLAGS: 00000292
Let’s have a closer look at the above dump, to understand some of the important bits of information.
[ 3341.170623] oops: loading out-of-tree module taints kernel.
[ 3341.170628] oops: module license 'unspecified' taints kernel.
[ 3341.170629] Disabling lock debugging due to kernel taint
[ 3341.170907] oops from the module
This is basic information to show the module.
[ 3341.170913] BUG: unable to handle kernel NULL pointer dereference at 00000000
It is obvious that to show the type of oops.
[ 3341.170919] *pdpt = 0000000029111001 *pde = 0000000000000000
May be this is relative to page table entry.If you have definited understanding about this, please tell me, thanks :)
[ 3341.170924] Oops: 0002 [#1] SMP PTI
This is the error code value in hex. Each bit has a significance of its own:
bit 0 == 0 means no page found, 1 means a protection fault bit 1 == 0 means read, 1 means write bit 2 == 0 means kernel, 1 means user-mode [#1] — this value is the number of times the Oops occurred. Multiple Oops can be triggered as a cascading effect of the first one.
Refer to above the site.
[ 3341.170928] CPU: 3 PID: 4066 Comm: insmod Tainted: P O 5.0.0-rc4 #3
This is your choice to have a look at kernel/panic.c:
/*
* TAINT_FORCED_RMMOD could be a per-module flag but the module
* is being removed anyway.
*/
const struct taint_flag taint_flags[TAINT_FLAGS_COUNT] = {
[ TAINT_PROPRIETARY_MODULE ] = { 'P', 'G', true },
[ TAINT_FORCED_MODULE ] = { 'F', ' ', true },
[ TAINT_CPU_OUT_OF_SPEC ] = { 'S', ' ', false },
[ TAINT_FORCED_RMMOD ] = { 'R', ' ', false },
[ TAINT_MACHINE_CHECK ] = { 'M', ' ', false },
[ TAINT_BAD_PAGE ] = { 'B', ' ', false },
[ TAINT_USER ] = { 'U', ' ', false },
[ TAINT_DIE ] = { 'D', ' ', false },
[ TAINT_OVERRIDDEN_ACPI_TABLE ] = { 'A', ' ', false },
[ TAINT_WARN ] = { 'W', ' ', false },
[ TAINT_CRAP ] = { 'C', ' ', true },
[ TAINT_FIRMWARE_WORKAROUND ] = { 'I', ' ', false },
[ TAINT_OOT_MODULE ] = { 'O', ' ', true },
[ TAINT_UNSIGNED_MODULE ] = { 'E', ' ', true },
[ TAINT_SOFTLOCKUP ] = { 'L', ' ', false },
[ TAINT_LIVEPATCH ] = { 'K', ' ', true },
[ TAINT_AUX ] = { 'X', ' ', true },
[ TAINT_RANDSTRUCT ] = { 'T', ' ', true },
};
[ 3341.170937] EIP: my_oops_init+0x12/0x1000 [oops]
There is oop happened on the program. my_oops_init+0x12/0x21 is the
[ 3341.170942] Code: Bad RIP value.
Oh, Is it a bug? Should it be Bad EIP value? Now， it will show the information about original code.RIP is the CPU register containing the address of the instruction that is getting executed.
[ 3341.170945] EAX: 00000014 EBX: f86d1000 ECX: f6f755ac EDX: 00000007
[ 3341.170947] ESI: 00000001 EDI: f86d4000 EBP: e7ce7dec ESP: e7ce7de8
[ 3341.170950] DS: 007b ES: 007b FS: 00d8 GS: 00e0 SS: 0068 EFLAGS: 00010246
[ 3341.170953] CR0: 80050033 CR2: f86d3fe8 CR3: 27c32000 CR4: 000406f0
I do not know it.
[ 3341.170955] Call Trace:
[ 3341.170962] do_one_initcall+0x42/0x1ae
[ 3341.170966] ? _cond_resched+0x18/0x40
[ 3341.170972] ? do_init_module+0x21/0x1dd
[ 3341.170975] do_init_module+0x50/0x1dd
[ 3341.170978] load_module+0x1d06/0x2540
[ 3341.170982] ? kernel_read_file+0x222/0x250
[ 3341.170986] sys_finit_module+0xa7/0xe0
[ 3341.170990] do_fast_syscall_32+0x7f/0x1d0
[ 3341.170993] entry_SYSENTER_32+0x6b/0xbe
Interesting…
[ 3341.170995] EIP: 0xb7f81851
[ 3341.170997] Code: 8b 98 58 cd ff ff 85 d2 89 c8 74 02 89 0a 5b 5d c3 8b 04 24 c3 8b 14 24 c3 8b 1c 24 c3 8b 3c 24 c3 51 52 55 89 e5 0f 34 cd 80 <5d> 5a 59 c3 90 90 90 90 8d 76 00 58 b8 77 00 00 00 cd 80 90 8d 76
[ 3341.170999] EAX: ffffffda EBX: 00000003 ECX: 004dc970 EDX: 00000000
[ 3341.171001] ESI: 01d19d90 EDI: 01d1a7a8 EBP: 004dae9c ESP: bfe2a18c
[ 3341.171003] DS: 007b ES: 007b FS: 0000 GS: 0033 SS: 007b EFLAGS: 00000292
Ok
[ 3341.171005] Modules linked in: oops(PO+) ctr ...
Used module in linked list maybe.
[ 3341.171049] CR2: 0000000000000000
[ 3341.171052] ---[ end trace dd8bceded905e5e1 ]---
[ 3341.171054] EIP: my_oops_init+0x12/0x1000 [oops]
[ 3341.171057] Code: Bad RIP value.
[ 3341.171058] EAX: 00000014 EBX: f86d1000 ECX: f6f755ac EDX: 00000007
[ 3341.171060] ESI: 00000001 EDI: f86d4000 EBP: e7ce7dec ESP: c5a27ddc
[ 3341.171062] DS: 007b ES: 007b FS: 00d8 GS: 00e0 SS: 0068 EFLAGS: 00010246
[ 3341.171064] CR0: 80050033 CR2: f86d3fe8 CR3: 27c32000 CR4: 000406f0
The first step is to load the offending module into the GDB debugger, as follows:
yubo@debian:~/lit$ gdb oops.ko
GNU gdb (Debian 7.12-6) 7.12.0.20161007-git
Copyright (C) 2016 Free Software Foundation, Inc.
License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html>
This is free software: you are free to change and redistribute it.
There is NO WARRANTY, to the extent permitted by law. Type "show copying"
and "show warranty" for details.
This GDB was configured as "i686-linux-gnu".
Type "show configuration" for configuration details.
For bug reporting instructions, please see:
<http://www.gnu.org/software/gdb/bugs/>.
Find the GDB manual and other documentation resources online at:
<http://www.gnu.org/software/gdb/documentation/>.
For help, type "help".
Type "apropos word" to search for commands related to "word"...
Reading symbols from oops.ko...done.
(gdb)
(gdb) add-symbol-file oops.o 0xf86d4000
add symbol table from file "oops.o" at
.text_addr = 0xf86d4000
(y or n) y
Reading symbols from oops.o...done.
Next, add the symbol file to the debugger. The add-symbol-file command’s first argument is oops.o and the second argument is the address of the text section of the module. You can obtain this address from /sys/module/oops/sections/.init.text (where oops is the module name):
From the RIP instruction line, we can get the name of the offending function, and disassemble it.
(gdb) disassemble my_oops_init
Dump of assembler code for function my_oops_init:
0x00000024 <+0>: call 0x25 <my_oops_init+1>
0x00000029 <+5>: push %ebp
0x0000002a <+6>: mov %esp,%ebp
0x0000002c <+8>: push $0x0
0x00000031 <+13>: call 0x32 <my_oops_init+14>
0x00000036 <+18>: movl $0x0,0x0
0x00000040 <+28>: xor %eax,%eax
0x00000042 <+30>: leave
0x00000043 <+31>: ret
End of assembler dump.
(gdb)
Now, to pin point the actual line of offending code, we add the starting address and the offset. The offset is available in the same RIP instruction line. In our case, we are adding 0x00000024 + 0x12(which is came from my_oops_init+0x12/0x1000 ) = 0x00000036.It points the 0x00000036 the instruction. It is a movl instruction.
(gdb) list *0x00000036
0x36 is in my_oops_init (/home/yubo/lit/oops.c:12).
7 #include <linux/kernel.h>
8 #include <linux/module.h>
9 #include <linux/init.h>
10
11 static void create_oops(void) {
12 *(int *)0 = 0;
13 }
14
15 static int __init my_oops_init(void)
16 {
(gdb)
So, we have found it.
Here is some usefully link: Shi shougang
这样写就是太乱了，但是为了安排到kernel目录下，暂时也就这样吧。 文章的题目什么意思呢？ 在kernel的主目录下，我们会看到一个tools/include/目录，这个目录是做什么的呢？他就是给内核某个特性的提供一个库，这个库你可以编译成动态库也可以编译成静态库，这取决于你的选择，通过这个库，你就可以像其他库文件一样使用了。
cd linux/tools/lib/bpf/
make && sudo make install
sudo make install_headers
sudo ldconfig