Return to how_to_enable_and_run_dtrace.mdwn CVS log

Up to [NetBSD Developer Wiki] / wikisrc / tutorials

DTrace on armv4 confirmed using hpcarm port on a Jornada 720.

    1: DTrace is a Dynamic Tracing framework developed by Sun and ported to NetBSD. It enables extensive instrumentation of the kernel and user space. See the [DTrace Community Page](http://dtrace.org) for more information.
    2: Also see [DTrace Introduction](http://dtrace.org/guide/preface.html), Brendan Gregg's [DTrace one liners](http://www.brendangregg.com/DTrace/dtrace_oneliners.txt) and his notes for [DTrace on FreeBSD](https://wiki.freebsd.org/DTrace/).
    3: 
    4: # Current status
    5: 
    6: ## Supported platforms
    7: 
    8: DTrace is a work-in-progress effort and it is for x86 systems and some arm boards.
    9: 
   10: * i386 and amd64
   11: * earm* (evbarm and armv4 based ports)
   12: 
   13: ## Supported providers
   14: 
   15: * SDT: Statically Defined Tracing
   16: * FBT: Function Boundary Tracing
   17: * Lockstat: Kernel Lock Statistics
   18: * Profile: Time based interrupt event source for Profiling
   19: * Syscall: System Calls
   20: * Syscall Linux (32bit & 64 bit): System calls via the Linux binary emulation layer
   21: 
   22: ## TODO for netbsd-7
   23: 
   24: * Measure effect of `options KDTRACE_HOOKS` on system performance.
   25: * Determine whether the profile module works and list it here.
   26: * Integrate [[riz|users/riz]]'s syscall provider patch.
   27: 
   28: ## TODO for netbsd-6
   29: 
   30: Need to identify changes to pull up to netbsd-6 and pull them up.
   31: Candidates:
   32: 
   33: * Profile provider.
   34: 
   35: # How to use
   36: 
   37: ##  Building DTrace 
   38: 
   39: You need the following options in your kernel: 
   40:     
   41:     options         KDTRACE_HOOKS   # kernel DTrace hooks
   42:     options         MODULAR
   43: 
   44: Optionally:
   45: 
   46:     options         INSECURE   # permit modules to loaded from user space once system has gone multiuser and securelevel has been raised.
   47: 
   48: A Distribution needs to be built with the options `MKDTRACE=yes` and `MKCTF=yes`, this is taken care of automatically and doesn't need to be specified manually. The list of platforms it is applied to automatically is set in `src/share/mk/bsd.own.mk`
   49: 
   50: Set the system to load the solaris and dtrace related modules in `/etc/modules.conf`, for a list of available modules, see `/stand/$MACHINE/$VERSION/modules/`
   51: 
   52: For example, add the following to `/etc/modules.conf` (the file may not exist already on a system):
   53:     
   54:     solaris
   55:     dtrace
   56:     dtrace_sdt
   57:     dtrace_fbt
   58:     dtrace_lockstat
   59:     dtrace_profile
   60:     dtrace_syscall
   61:     
   62: A `dtrace` device node is created automatically in `/dev/dtrace` when the modules are loaded into place.
   63:     
   64: List the dtrace probes 
   65:     
   66:     dtrace -l
   67:     
   68:        ID   PROVIDER            MODULE                          FUNCTION NAME
   69:         1     dtrace                                                     BEGIN
   70:         2     dtrace                                                     END
   71:         3     dtrace                                                     ERROR
   72:         4        fbt            netbsd             AcpiAcquireGlobalLock entry
   73:         5        fbt            netbsd             AcpiAcquireGlobalLock return
   74:         6        fbt            netbsd             AcpiAllocateRootTable entry
   75:         7        fbt            netbsd                    AcpiAttachData entry
   76:         . 
   77:         .
   78:     29129        fbt           solaris                   zfs_vop_getattr entry 
   79:     29130        fbt           solaris                   zfs_vop_getattr return
   80:     29131       proc                                                     create
   81:     29132       proc                                                     exec 
   82:         .
   83:         .
   84:     29140       proc                                                     lwp_start
   85:     29141       proc                                                     lwp_exit
   86: 
   87: 
   88: ##  Running hello world 
   89: 
   90: Put the following into the file hello.d:
   91:     
   92:     BEGIN
   93:     {
   94:         trace("Hello world");
   95:         exit(0);
   96:     }
   97:     
   98: 
   99: Run the hello world script: 
  100:     
  101:     dtrace -s hello.d
  102:     
  103:     dtrace: script './hello.d' matched 1 probe
  104:     CPU     ID                    FUNCTION:NAME
  105:       0      1                           :BEGIN   Hello world
  106:     
  107: 
  108: A more complex example that traces the execution of a sleep operation
  109: in the kernel. Put it in sleep.d:
  110:     
  111:     #pragma D option flowindent
  112:     
  113:     fbt::syscall:entry
  114:     /execname == "sleep" && guard++ == 0/
  115:     {
  116:             self->traceme = 1;
  117:             printf("fd: %d", arg0);
  118:     }
  119:     
  120:     fbt::syscall:entry /self->traceme/ {}
  121:     
  122:     fbt::syscall:return
  123:     /self->traceme/
  124:     {
  125:             self->traceme = 0;
  126:             exit(0);
  127:     }
  128:     
  129: 
  130: Start the script running (dtrace -s sleep.d) and then execute a "sleep 2" in another shell.