转自：https://linux.cn/article-3587-1.html

‘dmesg’命令显示linux内核的环形缓冲区信息，我们可以从中获得诸如系统架构、cpu、挂载的硬件，RAM等多个运行级别的大量的系统信息。当计算机启动时，系统内核（操作系统的核心部分）将会被加载到内存中。在加载的过程中会显示很多的信息，在这些信息中我们可以看到内核检测硬件设备。

当驱动装不上去的时候，可以使用dmesg来查看安装驱动时打印出来的调试信息，缺少哪些函数。

dmesg 命令的使用范例

‘dmesg’命令设备故障的诊断是非常重要的。在‘dmesg’命令的帮助下进行硬件的连接或断开连接操作时，我们可以看到硬件的检测或者断开连接的信息。‘dmesg’命令在多数基于Linux和Unix的操作系统中都可以使用。

下面我们展示一些最负盛名的‘dmesg’命令工具以及其实际使用举例。‘dmesg’命令的使用语法如下。

1. # dmesg [options...]

1. 列出加载到内核中的所有驱动

我们可以使用如‘more’。 ‘tail’, ‘less ’或者‘grep’文字处理工具来处理‘dmesg’命令的输出。由于dmesg日志的输出不适合在一页中完全显示，因此我们使用管道（pipe）将其输出送到more或者less命令单页显示。

1. [root@tecmint.com ~]# dmesg | more
2. [root@tecmint.com ~]# dmesg | less

输出

1. [ 0.000000] Initializing cgroup subsys cpuset
2. [ 0.000000] Initializing cgroup subsys cpu
3. [ 0.000000] Initializing cgroup subsys cpuacct
4. [ 0.000000] Linux version 3.11.0-13-generic (buildd@aatxe) (gcc version 4.8.1 (Ubuntu/Linaro 4.8.1-10ubuntu8) ) #20-Ubuntu SMP Wed Oct 23 17:26:33 UTC 2013
5. (Ubuntu 3.11.0-13.20-generic 3.11.6)
6. [ 0.000000] KERNEL supported cpus:
7. [ 0.000000] Intel GenuineIntel
8. [ 0.000000] AMD AuthenticAMD
9. [ 0.000000] NSC Geode by NSC
10. [ 0.000000] Cyrix CyrixInstead
11. [ 0.000000] Centaur CentaurHauls
12. [ 0.000000] Transmeta GenuineTMx86
13. [ 0.000000] Transmeta TransmetaCPU
14. [ 0.000000] UMC UMC UMC UMC
15. [ 0.000000] e820: BIOS-provided physical RAM map:
16. [ 0.000000] BIOS-e820: [mem 0x0000000000000000-0x000000000009fbff] usable
17. [ 0.000000] BIOS-e820: [mem 0x00000000000f0000-0x00000000000fffff] reserved
18. [ 0.000000] BIOS-e820: [mem 0x0000000000100000-0x000000007dc08bff] usable
19. [ 0.000000] BIOS-e820: [mem 0x000000007dc08c00-0x000000007dc5cbff] ACPI NVS
20. [ 0.000000] BIOS-e820: [mem 0x000000007dc5cc00-0x000000007dc5ebff] ACPI data
21. [ 0.000000] BIOS-e820: [mem 0x000000007dc5ec00-0x000000007fffffff] reserved
22. [ 0.000000] BIOS-e820: [mem 0x00000000e0000000-0x00000000efffffff] reserved
23. [ 0.000000] BIOS-e820: [mem 0x00000000fec00000-0x00000000fed003ff] reserved
24. [ 0.000000] BIOS-e820: [mem 0x00000000fed20000-0x00000000fed9ffff] reserved
25. [ 0.000000] BIOS-e820: [mem 0x00000000fee00000-0x00000000feefffff] reserved
26. [ 0.000000] BIOS-e820: [mem 0x00000000ffb00000-0x00000000ffffffff] reserved
27. [ 0.000000] NX (Execute Disable) protection: active
28. .....

2. 列出所有被检测到的硬件

要显示所有被内核检测到的硬盘设备，你可以使用‘grep’命令搜索‘sda’关键词，如下：

1. [root@tecmint.com ~]# dmesg | grep sda
2.  
3. [ 1.280971] sd 2:0:0:0: [sda] 488281250 512-byte logical blocks: (250 GB/232 GiB)
4. [ 1.281014] sd 2:0:0:0: [sda] Write Protect is off
5. [ 1.281016] sd 2:0:0:0: [sda] Mode Sense: 00 3a 00 00
6. [ 1.281039] sd 2:0:0:0: [sda] Write cache: enabled, read cache: enabled, doesn't support DPO or FUA
7. [ 1.359585] sda: sda1 sda2 < sda5 sda6 sda7 sda8 >
8. [ 1.360052] sd 2:0:0:0: [sda] Attached SCSI disk
9. [ 2.347887] EXT4-fs (sda1): mounted filesystem with ordered data mode. Opts: (null)
10. [ 22.928440] Adding 3905532k swap on /dev/sda6. Priority:-1 extents:1 across:3905532k FS
11. [ 23.950543] EXT4-fs (sda1): re-mounted. Opts: errors=remount-ro
12. [ 24.134016] EXT4-fs (sda5): mounted filesystem with ordered data mode. Opts: (null)
13. [ 24.330762] EXT4-fs (sda7): mounted filesystem with ordered data mode. Opts: (null)
14. [ 24.561015] EXT4-fs (sda8): mounted filesystem with ordered data mode. Opts: (null)

注解 ‘sda’表示第一块 SATA硬盘，‘sdb’表示第二块SATA硬盘。若想查看IDE硬盘搜索‘hda’或‘hdb’关键词。

3. 只输出dmesg命令的前20行日志

在‘dmesg’命令后跟随‘head’命令来显示开始几行，‘dmesg | head -20′命令将显示开始的前20行。

1. [root@tecmint.com ~]# dmesg | head -20
2.  
3. [ 0.000000] Initializing cgroup subsys cpuset
4. [ 0.000000] Initializing cgroup subsys cpu
5. [ 0.000000] Initializing cgroup subsys cpuacct
6. [ 0.000000] Linux version 3.11.0-13-generic (buildd@aatxe) (gcc version 4.8.1 (Ubuntu/Linaro 4.8.1-10ubuntu8) ) #20-Ubuntu SMP Wed Oct 23 17:26:33 UTC 2013 (Ubuntu 3.11.0-13.20-generic 3.11.6)
7. [ 0.000000] KERNEL supported cpus:
8. [ 0.000000] Intel GenuineIntel
9. [ 0.000000] AMD AuthenticAMD
10. [ 0.000000] NSC Geode by NSC
11. [ 0.000000] Cyrix CyrixInstead
12. [ 0.000000] Centaur CentaurHauls
13. [ 0.000000] Transmeta GenuineTMx86
14. [ 0.000000] Transmeta TransmetaCPU
15. [ 0.000000] UMC UMC UMC UMC
16. [ 0.000000] e820: BIOS-provided physical RAM map:
17. [ 0.000000] BIOS-e820: [mem 0x0000000000000000-0x000000000009fbff] usable
18. [ 0.000000] BIOS-e820: [mem 0x00000000000f0000-0x00000000000fffff] reserved
19. [ 0.000000] BIOS-e820: [mem 0x0000000000100000-0x000000007dc08bff] usable
20. [ 0.000000] BIOS-e820: [mem 0x000000007dc08c00-0x000000007dc5cbff] ACPI NVS
21. [ 0.000000] BIOS-e820: [mem 0x000000007dc5cc00-0x000000007dc5ebff] ACPI data
22. [ 0.000000] BIOS-e820: [mem 0x000000007dc5ec00-0x000000007fffffff] reserved

4. 只输出dmesg命令最后20行日志

在‘dmesg’命令后跟随‘tail’命令（‘ dmesg | tail -20’）来输出‘dmesg’命令的最后20行日志，当你插入可移动设备时它是非常有用的。

1. [root@tecmint.com ~]# dmesg | tail -20
2.  
3. parport0: PC-style at 0x378, irq 7 [PCSPP,TRISTATE]
4. ppdev: user-space parallel port driver
5. EXT4-fs (sda1): mounted filesystem with ordered data mode
6. Adding 2097144k swap on /dev/sda2. Priority:-1 extents:1 across:2097144k
7. readahead-disable-service: delaying service auditd
8. ip_tables: (C) 2000-2006 Netfilter Core Team
9. nf_conntrack version 0.5.0 (16384 buckets, 65536 max)
10. NET: Registered protocol family 10
11. lo: Disabled Privacy Extensions
12. e1000: eth0 NIC Link is Up 1000 Mbps Full Duplex, Flow Control: None
13. Slow work thread pool: Starting up
14. Slow work thread pool: Ready
15. FS-Cache: Loaded
16. CacheFiles: Loaded
17. CacheFiles: Security denies permission to nominate security context: error -95
18. eth0: no IPv6 routers present
19. type=1305 audit(1398268784.593:18630): audit_enabled=0 old=1 auid=4294967295 ses=4294967295 res=1
20. readahead-collector: starting delayed service auditd
21. readahead-collector: sorting
22. readahead-collector: finished

5. 搜索包含特定字符串的被检测到的硬件

由于‘dmesg’命令的输出实在太长了，在其中搜索某个特定的字符串是非常困难的。因此，有必要过滤出一些包含‘usb’ ‘dma’ ‘tty’ ‘memory’等字符串的日志行。 的‘-i’选项表示忽略大小写。

1. [root@tecmint.com log]# dmesg | grep -i usb
2. [root@tecmint.com log]# dmesg | grep -i dma
3. [root@tecmint.com log]# dmesg | grep -i tty
4. [root@tecmint.com log]# dmesg | grep -i memory

输出

1. [ 0.000000] Scanning 1 areas for low memory corruption
2. [ 0.000000] initial memory mapped: [mem 0x00000000-0x01ffffff]
3. [ 0.000000] Base memory trampoline at [c009b000] 9b000 size 16384
4. [ 0.000000] init_memory_mapping: [mem 0x00000000-0x000fffff]
5. [ 0.000000] init_memory_mapping: [mem 0x37800000-0x379fffff]
6. [ 0.000000] init_memory_mapping: [mem 0x34000000-0x377fffff]
7. [ 0.000000] init_memory_mapping: [mem 0x00100000-0x33ffffff]
8. [ 0.000000] init_memory_mapping: [mem 0x37a00000-0x37bfdfff]
9. [ 0.000000] Early memory node ranges
10. [ 0.000000] PM: Registered nosave memory: [mem 0x0009f000-0x000effff]
11. [ 0.000000] PM: Registered nosave memory: [mem 0x000f0000-0x000fffff]
12. [ 0.000000] please try 'cgroup_disable=memory' option if you don't want memory cgroups
13. [ 0.000000] Memory: 2003288K/2059928K available (6352K kernel code, 607K rwdata, 2640K rodata, 880K init, 908K bss, 56640K reserved, 1146920K highmem)
14. [ 0.000000] virtual kernel memory layout:
15. [ 0.004291] Initializing cgroup subsys memory
16. [ 0.004609] Freeing SMP alternatives memory: 28K (c1a3e000 - c1a45000)
17. [ 0.899622] Freeing initrd memory: 23616K (f51d0000 - f68e0000)
18. [ 0.899813] Scanning for low memory corruption every 60 seconds
19. [ 0.946323] agpgart-intel 0000:00:00.0: detected 32768K stolen memory
20. [ 1.360318] Freeing unused kernel memory: 880K (c1962000 - c1a3e000)
21. [ 1.429066] [drm] Memory usable by graphics device = 2048M

6. 清空dmesg缓冲区日志

我们可以使用如下命令来清空dmesg的日志。该命令会清空dmesg环形缓冲区中的日志。但是你依然可以查看存储在‘/var/log/dmesg’文件中的日志。你连接任何的设备都会产生dmesg日志输出。

1. [root@tecmint.com log]# dmesg -c

7. 实时监控dmesg日志输出

在某些发行版中可以使用命令‘tail -f /var/log/dmesg’来实时监控dmesg的日志输出。

1. [root@tecmint.com log]# watch "dmesg | tail -20"

结论：dmesg命令在系统dmesg记录实时更改或产生的情况下是非常有用的。你可以使用man dmesg来获取更多关于dmesg的信息。