Hi Henk,
This one is a slight nuisance on systems with large CPU counts.
When using dtrace(1M), the DTrace subsystem ends up allocating
a pair of 4MB buffers for its principal buffers and the same again
if you are using aggregations. This becomes a ton of memory to
setup and tear down on a system
Just for the record, I had an offline exchange with Ryan on this.
The startup time when using dtrace(1M) on systems with many
CPU's such as the T5440 can be large owing to the default
principal and aggregation buffer sizes (4MB). To reduce the
startup latency, tune these down (if possible) by sett
Hi Joe,
Thanks for the good test case.
Yes, this seems to be a bug. It appears that we end up generating
malformed DOF when unreferenced providers are included as in your
example. This will mean that we reject the DOF when it's loaded
into the kernel and you therefore don't get any probes. I've
Hi Daniel,
One script has 10k probes, but they are rarely triggered. Another
script has only one probe, but the probe triggered 10k per second.
Which one will impact the OS performance to a larger extent? In
another word: performance impact depends on the number of probes or
times any probe t
Hi Drew,
I'm wondering if this is expected behavior, or a bug.
Expected but not all that obvious really. See below.
Given this D:
# cat smb_logon.d
pid$target::smb_logon:entry {
add=&((struct netr_client *)arg0)->username;
printf("%p",add);
exit();
}
and this defintion:
typedef struct
Hi Joel,
Is there a setting I can use to make sure probefunc is not cut off at 127
characters?
As James said, this is a hardcoded limit so there's not an awful
lot you can do about it. Sorry.
Sometime we should really revisit the current static allocations.
FYI, there is an open bug which co
Hi Max,
I am using the hotuser script to trace user functions in the output I
get some lines that show hex address instead of function:
libc.so.1`0xfe54c080 2375 2.4%
libc.so.1`0xfe54b884 3254 3.2%
libc.so.1`0xfe54c364
Hi,
You're probably running into the following issue which is
an OpenSolaris upgrade related problem:
http://defect.opensolaris.org/bz/show_bug.cgi?id=11609
Unfortunately, this breaks the type information that DTrace uses.
I'm not really aware of anything that you can do to get DTrace
working c
Hi Brian,
I did find this:
http://hub.opensolaris.org/bin/view/Community+Group+dtrace/ChangeLog
But unfortunately, it doesn't map DTrace API Version numbers to the SNV build
numbers. This would be useful, for example, for knowing which of these DTrace
features, if any, are available in a Sol
Hi Robert,
I am trying to get documentation how how this lazyloading happens
when you are dealing with dtrace probes being added to user
applications. In particular if it somehow avoids the overhead of
registering dtrace probes when they aren't needed, how does it know if
they are needed
Hi Joseph,
What is the status of DTrace substr function? It was integrated into Neveda build 15 and it is available on Solaris 10 systems (U4 - u8 x64) I am working on. I cannot find any documentation on it, and there appears to be a bug in it that I cannot reproduced on OpenSolaris.
I thin
Hi Thomas,
I want to trace the thread_reaper and want to find out, how often it runs and
how many threads it really reaps.
I use this one-liner in the first step:
[r...@itotcsol104 bin]# dtrace -n 'fbt:genunix:thread_reaper: { @num[probefunc]
= count(); }'
dtrace: description 'fbt:genunix:thre
Hi Max,
From taking a quick look at plockstat.c, the nsec column is a sum
of the number of nanoseconds that a mutex is held. Count is a count
of the number of times the mutex is acquired. According to this output,
a malloc lock is aquired and released
approximately 4 million times, but the t
Hi Krish,
Oops. Yes, there seems to be a bit of an issue here so thanks
for flagging it. I'll follow up offline with you.
Jon.
Hi all,
I ran the following one-liner, dtrace -n 'pid3524:::*read* {}' and dtrace
dumped core. Of course the intention was to place '*read*' as probefunc.
Shouldn't dt
Hi Joel,
Is there a way to truncate multiple aggregations collectively?
No. The trunc() function just operates on a single aggregation.
Jon.
For Example, to display 20 or less records for the below:
...
@a[pid,self->addr,self->name,self->id] = sum(itema);
@b[pid,self->addr,self->name,self-
Does the same apply for the pid provider?
I have seen pid providers with dtrace -l.
At the moment the command does not list any pid providers.
Yes, pid provider probes won't appear in the ouput of `dtrace -l`
until they have been explicitly created (by enabling them):
# dtrace -l -P pid*
I
Hi Hans,
I would like to know why dtrace -l |grep plockstat show plockstat28099
lines.
Process with pid 28099 is a java process.
Why can I not examine other processes with the plockstat provider?
Is it a java startup option?
No. The probes that plockstat uses only become visible when you
ac
Hi Joel,
I'm running dtrace with multiple scripts, e.g. -s foo.d -s bar.d -s baz.d.
It looks like variables in one script overwrite variables in another. I
was expecting, perhaps wrongfully, that variables would be per-script
and not per dtrace run.
I guess it may not be too obvious from th
OK. Latest version I can see is a year old,
http://dlc.sun.com/pdf/817-6223/817-6223.pdf from
http://docs.sun.com/app/docs/doc/817-6223/6mlkidlgj?l=en&a=view
That's the latest version.
With the help of the most excellent Alta I have exported
the wiki guide to pdf and it doesn't look too bad
Hi Andrew,
Does the Wiki dtrace manual get converted into the PDF version anywhere?
The one on docs.sun.com seems to be significantly older (although I may
be looking in the wrong place). I want to print off and bind a couple of
copies to leave with a customer after a session there. Can't see
Hi Daniel,
If use truss on an oracle process. we could see something like the
following, so semtimedop should be a syscall
22459/1: semtimedop(9, 0xFD7FFFDFE648, 1, 0xFD7FFFDFE488) EAGAIN
22459/1: semnum=15semop=-1semflg=0
22459/1: timeout: 2.06000 sec
b
Hi Rayson,
Is it possible to provide an interface for other providers to read
data from the hardware performance counters?
The current CPC provider gives us hardware "events", but not the
actual value of the counters. Say if I am interested in the number of
cache misses caused a function, I can
Sorry, one more question... is it safe to fire up multiple dtrace_hdl_t
within a process?
Sure, a consumer can do multiple opens.
Chad's advice is good (as always). Especially the bit about
offering to send you a stripped down consumer :-) .
Jon.
Ryan
Jon Haslam wrote:
In parti
In particular, this shouldn't be an issue unless you are enabling foo12345:::
probes (that is, any providers with PIDs in them)
Ah, but I am... that was the way for my app to tell dtrace it wants the
latest stats... the question was how to get the answer back since I
can't copyout() the ag
Can you provide any specifics regarding these bad things?
Well, libdtrace is rather underdocumented, but according to the folks
behind TclDTrace, "Note that *dtrace_close* is needed to prevent an epic
fail. It does a major cleanup, that involves restoring the state of all
grabbed processes, w
rs may well
chime in with suggestions. Yes, with libdtrace you would have to do
some leg work yourself with coding but the joy and pleasure it
will give you will far outweigh any pain you may have encountered
along the way :-) .
Have a go. You know you want to.
Jon.
Did I miss something?
Ryan
J
Hi Ryan,
You need to use libdtrace(3LIB) to get this done in a clean manner.
There are a bunch of tools in Solaris that you can look at for examples
of how to use the interfaces. These include plockstat(1M), instrat(1M),
lockstat(1M) and powertop(1M) to name a few.
Jon.
Hi all, I'm trying to ga
Is this because /usr/lib/dtrace/iscsit.d references the translator iscsiinfo_t
that appears to be defined in /usr/lib/dtrace/iscsi.d ? Would the
#pragma D depends_on lines help pkg find *.d file dependencies?
Yes, it is because of that. As you say, D Library dependencies are
expressed by usi
I too had the same intuition that DTrace works completely in software,
but after knowing that there is Performance Monitoring Unit(PMU)
present in intel's chip, so I got confused if DTRace uses it, also
same type of hardware is present in AMD's chip. I might be wrong.
The intel/amd micro
/onnv/onnv-gate/usr/src/lib/libdtrace/i386/dt_isadep.c#78
Jon.
Thanks
Zhihui
2009/4/23 Jon Haslam <mailto:jonathan.has...@sun.com>>
Hi Zhihi,
This could be owing to the fact that there we've found jump
tables in the instructions. If DTrace finds jump tables when
a
Hi Zhihi,
This could be owing to the fact that there we've found jump
tables in the instructions. If DTrace finds jump tables when
attempting to create probes it goes into an ultra conservative
mode of operation.
To see if this is the problem, set the DTRACE_DEBUG environment
variable before inv
Hi Angelo,
One of my ISVs have been looking for the atomic operation for global
variables as well. While we wait for this feature, do you have any
ideas for a reasonably good work around, now?
There isn't an RFE for this; if there is a need (or a perceived need)
then please log an RFE on beha
Hi,
The process being traced is written in C++ and the symbols are mangled
and we don't do any auto-demangling in Solaris. The demangled version of
the function looks like:
$ dem _Z12write_recordP3THDP8st_tableP12st_copy_info
_Z12write_recordP3THDP8st_tableP12st_copy_info == write_record(THD*,
Hi Pavan,
And given below is the output that I see:
unix`mutex_enter 802562
unix`ip_ocsum 2644713
unix`0xfb842691253
^^
Usually when I see unresolved functions o
Hi Paul,
Is there a list anywhere of the DTrace FBT provider calls for each version of
Solaris since the Nevada build that hosed the DTrace Toolkit. I have no issue
with rewriting what I need, but I am hoping to stave off having to look at all
the source code to find everything. I know ther
Hi,
A fix for this was put into snv_94. Upgrade to the latest bits if you can
and you should be OK.
Jon.
On SNV91 we are getting the following kernel panic using DTrace.
Core dump are available.
Mar 9 18:14:55 SVR24 unix: [ID 10 kern.notice]
Mar 9 18:14:55 SVR24 unix: [ID 839527 kern.
Hi,
do we have any possibility of including a data structure in dtrace??
it would be advantageous and can act like a buffer
Sorry but I don't understand what you want to do here. Can you give
an example of the kind of thing you are referring to?
Jon.
___
Hi Jin,
# ./ex1.d
dtrace: failed to compile script ./ex1.d: line 3: probe description
cpc:::inst_retired-1 does not match any probes
Does snv_109 support cpc provider?
Yes, the provider is in snv_109 and, as Dan says, you need to specify a
mode.
You may know this anyway as you're
Hi Rafael,
Works fine for me :-) .
I'll pick up 6805610 and work this offline with you.
Jon.
Here's a simpler example. I'm running OpenSolaris snv_107 on x86.
r...@catfish:/sandbox/tmp$ pfexec dtrace -n
'cpu-change-speed{printf("%d\n",
((cpudrv_devstate_t *)arg0)->cpu_id)}'
dtrace: invalid
Hi,
Finally, the long promised CPU performance counter provider has
arrived into Nevada build 109! For those that need an injection of
excitement at the start of the week, you can check out some usage
examples on my blog:
http://blogs.sun.com/jonh/entry/finally_dtrace_meets_the_cpu
Also, you ca
file
> with function's prototypes and manually create probe for each
> function. To have
> better idea there is an application which I have tried which does
> similar thing on
> windows platform: www.autodebug.com.
>
> Remek
>
> On Wed, Jan 21, 2009 at 1:01
> probe arbitrary statements in code symbolically (function entry, exit,
> interior, source code co-ordinates):
> yes (using debugging information)
>
> If that is true than they are a bit ahead in this area. I am really
> missing it in dtrace.
> Having this it would make dtrace user-land instrumen
Hi Toby,
> However when I run D code using this probe I don't see any output
> unless I add the dtrace_kernel privilege.
>
> Here the D code I am using (this is in a script that I run as dtrace
> -s tick-1.d):
>
> /*
> * Count off and report the number of seconds elapsed
> */
> dtrace:::BEGIN
Hi Deba,
> Now I want to know which script is updating this file. I would like to the
> see the output of "ptree 18786". I am not sure how to this structure of ptree
> or the name of the script.
The simple approach may just be to stop the target process,
ptree it and then continue it:
#!/usr/s
I'm following this up with Shiv directly and I'll post to
the alias if anything interesting falls out of it.
Jon.
> printa() in the below script results in segmentation fault. If %A in
> dtrace:::END is changed to %a, there is not segmentation fault but the
> output is in hex form (Ex: OUT: 8 0xf
> This is what exactly what I have in mind. Did you do it the way I am
> thinking? The 'jzgraph' project is well but dead. So, probably I shall
> have to generate a 'dot' file and then externally feed it to 'dotty'.
Firstly, the approach you're taking is OK but you'll probably
run into some issue
Hi Amit,
> Basically, I would like to have a graphical display showing me the
> call flow between the various functions- starting and ending at the
> starting function. I have thought about it a bit. I have written a
> regex to identify the various columns. After some intermediate data
> storage u
Hi Mike,
If nobody has picked this up, I'll sponsor you. Contact me offline
and we'll take it from there.
Jon.
> On Sat, Sep 20, 2008 at 7:57 AM, Mike Gerdts <[EMAIL PROTECTED]> wrote:
>
>> On Fri, Sep 19, 2008 at 9:31 PM, Mike Gerdts <[EMAIL PROTECTED]> wrote:
>>
>>> I still think that
Hi Shiv,
> DTraceToolkit has a dtrace script (Proc/pidpersec.d) that refers to a
> variable mpid as below
>
> profile:::tick-1sec
> {
> printf("%-22Y %8d %6d\n", walltimestamp, `mpid, pids);
> pids = 0;
> }
>
> Where is `mpid getting picked from (it isn't declared anywhere else)?
The ba
Hi,
For those that are interested, the DTrace change log has been
updated (up to and including build 101):
http://opensolaris.org/os/community/dtrace/ChangeLog
Jon.
___
dtrace-discuss mailing list
dtrace-discuss@opensolaris.org
Hi Manfred,
> Unfortunately this doesn't appear to be in the docs - sorry about
> that. You can feel free to add it to the wiki yourself or I'll try
> and do it shortly.
I just saw your wiki update for the above - many thanks!
Much appreciated.
Jon.
__
Hi Manfred,
> DTrace features the built-in variable caller (from chapter 3 of the Solaris
> Dynamic Tracing Guide: uintptr_t caller - The program counter location of the
> current thread just before entering the current probe).
>
> I tried to use caller (trace(caller)) in a script of mine and go
>> The DTrace documentation on docs.sun.com should be following the online wiki
>> documentation. The wiki is the primary source of documentation:
>>
>> http://wikis.sun.com/display/DTrace/Documentation
>
> Is there any convenient way to convert wiki source into some printable
> format (pdf?) ?
>> From a features point of view they should be pretty much the same as U4.
> Is this to be read as "there will only be bugfixes in U6"?
Yes. Pretty much only escalated bugs.
> Is there a Sun policy on what features (not) to propagate from OpenSolaris
> Dtrace to Solaris? And are these backpor
Hi Manfred,
> Full Ack. It's quite difficult, though, to predict what DTrace features one
> will get when installing Solaris 10 (whatever release). There is nothing
> about the fine details of DTrace in the Solaris 10 What's New documents.
>
> "printa() with multiple aggregations" is a nice fea
Hi Thomas,
Apologies for not following up on the list to your reply.
I'm glad that you can now move forward with your DTrace based
endeavors and hope that this wasn't too much of a pain to get
sorted.
> Maybe this historical "restriction" and DTrace-Zone-NFS restriction find
> their way into the
Hi Thomas,
Thanks for the debug data - I think I know what's going
on now (for those who are wondering - I contacted
Thomas offline to get some debug and ask some questions).
> I get this strange behavior when (as non-root with dtrace-proc and
> dtrace-user privs) I try to trace a process in a n
For those that are interested there is an open bug on this:
6266202: DTrace io provider doesn't work with ZFS
Jon.
> On Aug 22, 2008, at 7:31 AM, Ben Rockwood wrote:
>
>
>> John Dzilvelis wrote:
>>
>>> Hello,
>>> I have recently started using iosnoop from the dtrace toolkit to
>>> moni
ther mmu context/ustack() related issue on 4v
and I'll verify the source of this then.
Jon.
>
> Jon Haslam wrote:
>
>>> One more thing. On the T1000, I run this script:
>>>
>>> #!/usr/sbin/dtrace -s
>>> profile-997
>>> /arg1 &a
> One more thing. On the T1000, I run this script:
>
> #!/usr/sbin/dtrace -s
> profile-997
> /arg1 && execname == "ssh"/
> {
> @[execname,pid,ustack()]=count()
> }
> profile-997
> /arg1 && execname == "sshd"/
> {
> @[execname,pid,ustack()]=count()
> }
> profile-50s
> {
>
d just be using
> dtrace as a quick way of enabling and disabling the specific counters
> we want to track.)
The user guide chapter will have more and different examples. I'll
have a play around with that idea but I'm not sure how useful it is
to correlate values that we've
Tracing Fans,
I know it's been a long time in coming but the CPU Performance
Counter (CPC) provider is almost here! The code is currently in
for review and a proposed architecture document is attached here
for review.
Many thanks to all those that gave me feedback on this proposal.
A revised ve
Jon.
> On Jul 14, 2008, at 3:42 AM, Jon Haslam wrote:
>
>
>> Tracing Fans,
>>
>> I know it's been a long time in coming but the CPU Performance
>> Counter (CPC) provider is almost here! The code is currently in
>> for review and a proposed ar
Hi Brian,
> I'd like to suggest adding something to the error to indicate that the
> problem is that another probe is already enabled that conflicts with
> this one. The current message "Failed to enable probe" is pretty much
> the same as "it didn't work" and will be doubly confusing because t
Hi James,
> You may need a "per type" of overflow limit. A 3 - 4GHz cpu will
> generate
> 300k - 400k events per second when tracking cycles.
>
> Some of the more exotic cache behavior counters might fire only 1-2k
> events
> per second, though. When you combine that with the need
Hi Chad,
>>
>> Only one of cpustat and DTrace may use the counter hardware at any one time.
>> Ownership of the counters is given on a first-come, first-served basis.
>
> I'm curious, how does DTrace interact with DTrace in this situation?
> Specifically, if two DTrace invocations (either separate
Hi Phil,
> Looks good!
Thanks!
> I guess the thing which most people are going to pick up on is the
> sampling granularity. I just wonder whether you should always have to
> specify this, or whether you should pick up some sensible (safe)
> default; I guess there are pros and cons, the downside
Tracing Fans,
I know it's been a long time in coming but the CPU Performance
Counter (CPC) provider is almost here! The code is currently in
for review and a proposed architecture document is attached here
for review.
Any and all feedback/questions on the proposed implementation
is welcome.
Tha
Hi Chris,
Scratch space is only valid for the duration of a clause so
the scratch space you had the alias pointers in is most likely
getting trashed by the copyinstr(). You need to stash the pointers
away in something more permanent if you want to access them
at a later time.
Jon.
> Dtracing get
Hi Chris,
Looks like this is being caused by the change to direct bindings
that happened in snv_83. Basically, when direct binding is specified
we get an implicit -z lazyload on the link so libraries that ON
tools link against are now are marked for lazy loading:
#elfdump -d /bin/getent
Dynamic
Hi Chris,
Yeah, something looks to be broken. Use the '-Z' option
for now as a workaround and I'll take a look at it.
Jon.
> On nevada build 93 if I specify anything other than a wild card for the pid
> providers probefunc or probemod dtrace can't find the probes.
>
> So with wild cards I get:
a vestige of a tracing system from
bygone days.
Jon.
>
> Thanks.
>
>
> On Fri, May 30, 2008 at 7:28 PM, Jon Haslam <[EMAIL PROTECTED]> wrote:
>>> The comments at the top of sys/vtrace.h explains the background, but
>>> doesn't state in detail how to use it
> The comments at the top of sys/vtrace.h explains the background, but
> doesn't state in detail how to use it. In particular, what the type
> and the format of the arguments are.
>
> For example, to monitor this (in uts/common/inet/ip/ip.c)
> TRACE_1(TR_FAC_IP, TR_IP_RPUT_START, "ip_input
Hi Tuma,
> Anyone knows about the vtrace provider? When I was reading the code, I
> find some traces like TRACE_1, TRACE_2, etc. The comments says they're
> now replaced by Dtrace's vtrace provider, but I couldn't find any
> documents about this vtrace provider. I googled and searched on
> docs.su
Hi Mikael,
> So with this limitation and the limitation that dtrace -G -s only
> applies to object
> files and not to libraries the only workaround I find is that one
> needs to build all
> object files in them same directory as the binary or can anyone come
> up with
> a different workaroun
Hi,
Question for the Apple team I guess... Do Apple have any
plans to ship the DTrace Java API anytime? Just to be clear,
I'm referring to the code under:
http://src.opensolaris.org/source/xref/onnv/onnv-gate/usr/src/lib/libdtrace_jni
Thanks.
Jon.
___
Hi Jenny,
By the looks of it this fix went into the Solaris 10 KU patch - at
least I can see it in 120011-14 which was released in Sep 2007. No
idea what the current KU patch is though.
Jon.
> Hi, Adam,
>
> Is there any patch for sol10 for the bug#6370454(support USDT probes in
> static functio
Hi Mitsu,
> I have a problem to sort out a memory overwritten of our Application
> on Solaris 10.
> I wonder if it is possible to detect memory overwritten by using
> dtrace. I have worked for power pc platform before and I have usually
> used the DABR register to detect the memory overwritte
> However, I must make a meta-comment. This wiki (Confluence)
> has some real drawbacks. Why isn't Sun using a time-tested,
> Open Source, Enterprise wiki such as TWiki?
We just have to use the infrastructure that's provided so there's
not a lot that we can do for now. Your points are duly note
Hi Rich,
Apologies for taking a while but I've put a page on the
DTrace wiki space where one and all can record their
deepest desires (purely for DTrace enhancements though!).
http://wikis.sun.com/display/DTrace/The+Wish+List
Jon.
> [ I posted this a week ago, but haven't seen any response.
>
>
>> Using time based probes to iterate over data structures is
>> problematic as the data structures may well change beneath you.
>> I offered this up as an example of how to iterate over data structures
>> using a tick probe as it gets referenced quite a bit but there aren't that
>> many examples
>> self->pidp = `pidhash[$1 & (`pid_hashsz - 1)];
>
> what is backtick doing here in front of pidhash? Is it a way
> of accessing an arbitrary kernel variable?
Yes. The backquote is a scoping operator for kernel variables.
See the External Variables section in the docs:
http://wikis.sun.
Hi Roman,
> Here's the problem I'm facing: I need to sample a particular
> attribute of all the LWPs in a given process. Now, sampling
> itself is no problem -- I can use profile/tick provider and
> all is good. The problem is: I can't seem to think of a
> DTrace-friendly way to iterate over all
Hi Rich,
In case you weren't aware, you can view and log RFE's
(Requests for Enhancements) via bugs.opensolaris.org. However,
this interface is a bit limited and I do understand it's probably
not the greatest way to interact with the bug tracking system
that we use.
I'm quite open to there being
> That said, I've often thought a quick reference card might be handy; I'm
> pretty sure Jon H showed me a Sun-only version of such a thing back in
> the early days, although that could be my memory playing tricks. But it
> seems so old-fashioned; perhaps some DTrace quick reference wallpaper,
> w
Sorry for the bad form of replying to myself but I didn't finish
what I had to say before I sent it. Sigh.
Without the unit specifier we should be cleaning at the highest
rate possible - 5000hz. I just wanted to make sure you were
cleaning at this rate and that you really were getting the rate you
Hi Paul,
> Trying to get a script to run for an hour. Up till now every attempt failed
> before the hour was over. Latest command was:
>
> dtrace -q -x dynvarsize=16g -x cleanrate=400 -s /DTrace/newstr2.
Try specifying a unit for your cleanrate frequency i.e. 400hz. Without
a unit I'd have exp
Hi Bill,
At the minute we don't have an in-built mechanism for dumping the
data in a packet. There is an RFE which may be of interest:
6314638 Dtrace needs a packet dumper
For now you'd have to roll something yourself to get this done.
Jon.
> I sent this to Brendan but never got a response. M
notes there.
>>>
>>> It may also be useful to add a chapter that discusses platform
>>> differences at a
>>> high level.
>>>
>>> Just some thoughts...if the platform differences are not vast enough to
>>> warrant
>>> adding format
Hi James,
> I just got word, the OS X dtrace source has been posted:
>
> http://www.opensource.apple.com/darwinsource/10.5/dtrace-48/
>
> In the interest of full disclosure, the sources as they sit will not
> build correctly. This isn't a plot, plan, or other scheming, its simply
Hi Jan,
I think that putting a probe on fbt::closef:entry should give
you what you want:
# cat rel.d
#!/usr/sbin/dtrace -s
fbt::closef:entry
/execname == "foo && args[0]->f_vnode->v_path != NULL/
{
@[stringof(args[0]->f_vnode->v_path)] = count();
}
# ./rel.d
dtrace: script './rel.d' mat
> There's an easy way to dump all the elements of a structure like the
> sdump scat command or the print mdb command
No, such a thing doesn't exist at the minute. There is an open
RFE for it though:
4968681: would like ::printf-like functionality in DTrace
Jon.
___
entioned though, using fbt may be
easier for you. You always have options with DTrace!
Jon.
> Jon Haslam wrote:
>
>>> What determines what struct dtrace knows about by default?
>>>
>>> I am trying to use dtrace to trace the input and output to
>>> the
> What determines what struct dtrace knows about by default?
>
> I am trying to use dtrace to trace the input and output to
> the ntp_adjtime call. The ntp_adjtime call takes a single
> argument, namely a pointer to a timex struct. Even though
> that struct is the argument to a syscall, dtrace doe
Tracers,
Good news! The Solaris Dynamic Tracing Guide has entered
the participation age at last and has been converted into
wiki format. The new and shiny guide can be found at:
http://wikis.sun.com/display/DTrace/Documentation
The guide has fallen behind with developments in DTrace
since it was
> Typically, when I assign a variable in dtrace, I cast it first, like
> this:
>
> fbt::foo:entry
> {
> self->vpp = (vnode_t **)arg0;
> }
There's no need to cast in this situation though as, with
the fbt provider, you can just access typed arguments through
the args[] array. Just use args[2
>>(Where I'll be using the "sqrt(avg(x^2) - avg(x)^2)" approximation to
>>standard deviation.)
>>
>>While this solution is fast, the problem is that data[2] faces
>>overflow much sooner than data[1], especially if you're looking at
>>data like time in nanoseconds.
>>
>
>H...yes. Your two choi
Hi Neelam,
>I am also looking something very similar. I want to use hardware performance
>counters inside dtrace. Could you provide me some pointers if this could be
>done.
>
>
The CPC provider I'm been working on is taking shape, albeit slowly.
I hope to have some archives ready for around t
Hi Jeevan,
>how can i get the value of a variable in some application code
>(hastorageplus) without using usdt probes ?
>
>can we get it directly using dtrace ?
>
You can but the level of effort required is dependent upon
the scope of the variable in question (i.e. if it's global
or local).
Thi
Hi Roland,
> Is there anywhere a list of DTrace probes/provides which are supported
> by libc ?
As with any other user-land library, DTrace can dynamically instrument
libc down to individual instructions via the 'pid' provider.
For example, to see a list of which function entry points can be
ins
1 - 100 of 112 matches
Mail list logo
