Re: Article on Sun's DTrace
But if you *do* happen to know 60 good programmers who are willing to work on FreeBSD full time for very little money, let me know and I'll see what I can do about that baby thing. the great idea: get a bank account, ask users if they throw an 1$ (1eur) on it and let some indian (or something else) work on it. i think the idea is not as bad. bye, Ingo ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-hackers To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Article on Sun's DTrace
Steven Smith [EMAIL PROTECTED] writes: I don't know enough about Sparcs to even speculate how it's done there. The UltraSparc is a RISC processor, which amongst other things implies constant instruction size. Memory barriers take care of any cache coherence issues that may arise. DES -- Dag-Erling Smørgrav - [EMAIL PROTECTED] ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-hackers To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Article on Sun's DTrace
On Mon, Jul 12, 2004 at 07:09:33PM +0100, Steven Smith wrote: On their DTrace support forum there is the article about the problem with different byte patterns of movl %esp, %ebp produced by different assemblers. Do you have an URL for that? I can't seem to find it. http://forum.sun.com/thread.jsp?forum=211thread=19877 Also modifying functions on-the-fly require some sort of synchronization: noone should run function which currently is being modified (fbt provider). I suspect that the actual probe trigger is an int3 instruction, rather than a call, since that's a single byte and can therefore be atomically copied in over the start of any instruction. Having read that bug report I began to think that they change several continues bytes in a function, probably they just search for well known commands sequence and atomically change one of them. I think it is possible to change almost any instruction on x86, just because changed instruction should be emulated after return from DTrace probe (this very actual for probes in userspace). Yes, you are right, using classic debugging technique for activating DTrace trampoline should work. One can find description of probe's activating for x86 in the 4.1 paragraph of the DTrace Usenix report. They talked about IDT and interrupt handler. I know that you know this, but... If an interrupt call for activating probe is used on x86, then this explains how it is possible to get offset of ret command (cs:eip from trap frame) and how probes work in the userspace (control goes to kernel, where it works with script's variables). Again, if every ret instructions or instructions before ret instructions are changed in a function (because an offset of ret instruction is available in :return probe), then to speed up instruction changing, it is possible to save offsets of probes entries in some well known sections of object files (during compilation phase for example) and if there isn't such section, then try to find probes entries on-the-fly by disassembling binary code. Wildcard probes can require changing at least two instructions in every of tens of thousands functions. ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-hackers To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Article on Sun's DTrace
Hi Andrey, (B (BAndrey Simonenko wrote: (B (B (B Having read that bug report I began to think that they change several (B continues bytes in a function, probably they just search for well known (B commands sequence and atomically change one of them. I think it is (B possible (B to change almost any instruction on x86, just because changed instruction (B should be emulated after return from DTrace probe (this very actual for (B probes in userspace). Yes, you are right, using classic debugging (B technique for activating DTrace trampoline should work. (B (B One can find description of probe's activating for x86 in the 4.1 (B paragraph of the DTrace Usenix report. They talked about IDT and (B interrupt handler. (B (B I know that you know this, but... (B (B If an interrupt call for activating probe is used on x86, then this (B explains how it is possible to get offset of "ret" command (cs:eip from (B trap frame) and how probes work in the userspace (control goes to kernel, (B where it works with script's variables). (B (B Again, if every "ret" instructions or instructions before "ret" (B instructions (B are changed in a function (because an offset of "ret" instruction is (B available in :return probe), then to speed up instruction changing, it (B is possible to save offsets of probes entries in some well known sections (B of object files (during compilation phase for example) and if there isn't (B such section, then try to find probes entries on-the-fly by disassembling (B binary code. Wildcard probes can require changing at least two (B instructions (B in every of tens of thousands functions. (B ___ (B (B (BYou seem to know well about DTrace. I was taught a lot by you (Bon the off line too. By the way, Are you plan to port DTrace like (Btool to FreeBSD? or are you Sun or DTrace developer? Sorry, (BI'm not sure who and what you are, and I'm not old-timer on this list. (B (BEitarou (B (B-- (B (B*** (BEitarou Kamo (B (BTel. +81 75 7035997 (BFax +81 75 7035997 (BVoIP 050 10585997(domestic only) (Be$B!>(Bmail [EMAIL PROTECTED] (B (BFor business: (BFeel free to mail me(above), please. (B (BDonation http://www.PayPal.Com (B (BGPG FingerPrint: (B032D FDF9 D27B 23F7 9A81 BF4C 626C FBAA BC3A 9895 (B (B (B (B (B (B (B___ (B[EMAIL PROTECTED] mailing list (Bhttp://lists.freebsd.org/mailman/listinfo/freebsd-hackers (BTo unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: Article on Sun's DTrace
On Sat, 10 Jul 2004 20:45:14 +0100 in lucky.freebsd.hackers, Steven Smith wrote: It's also possible to put probes on the return instruction of the function. I'm not sure how they're actually finding that, though. I think the return probe is done by adding a call probe that changes the return address. Yeah, I thought that when I first saw it, but the probe is passed the address of the return instruction when it fires, and I can't see how you could get that if it was just invoked by modifying the return address on the call stack. Don't you think that they disassemble functions on-the-fly to find out prolog and return sequence of a function? On their DTrace support forum there is the article about the problem with different byte patterns of movl %esp, %ebp produced by different assemblers. (As an optimization fbt:::entry and fbt:::return probes' entry points can be found before and be placed in well known section.) At least if the control goes from the target function to some DTrace probe dispatch function, then this DTrace function should know which commands to emulate before returning to the target function, and it is impossible to run original function's commands without knowing their size, I mean that DTrace function should copy exactly complete sequence of commands from target function, not some bytes. I haven't opportunity to test DTrace, but there is another interesting question. fbt:::entry probe (or similar so called wild card probe) can create tens of thousands entry points on-the-fly. How does this creation affect on whole system? Also modifying functions on-the-fly require some sort of synchronization: noone should run function which currently is being modified (fbt provider). ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-hackers To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Article on Sun's DTrace
It's also possible to put probes on the return instruction of the function. I'm not sure how they're actually finding that, though. I think the return probe is done by adding a call probe that changes the return address. Yeah, I thought that when I first saw it, but the probe is passed the address of the return instruction when it fires, and I can't see how you could get that if it was just invoked by modifying the return address on the call stack. Don't you think that they disassemble functions on-the-fly to find out prolog and return sequence of a function? That is entirely plausible. On their DTrace support forum there is the article about the problem with different byte patterns of movl %esp, %ebp produced by different assemblers. Do you have an URL for that? I can't seem to find it. Also modifying functions on-the-fly require some sort of synchronization: noone should run function which currently is being modified (fbt provider). I suspect that the actual probe trigger is an int3 instruction, rather than a call, since that's a single byte and can therefore be atomically copied in over the start of any instruction. Any other processor either sees the value before the probe was activated (which is fine; it's just equivalent to the probe activating a split second later) ot afterwards (which is also fine). The x86 memory model is (I think; someone with more knowledge may want to correct me) strong enough to make that perfectly safe. (So the push %ebp part of the prolog becomes an int3 instruction in a single atomic operation, rather than just the first byte of a call instruction). I don't know enough about Sparcs to even speculate how it's done there. Steven. pgpp87t1IesDg.pgp Description: PGP signature
Re: Article on Sun's DTrace
It's also possible to put probes on the return instruction of the function. I'm not sure how they're actually finding that, though. I think the return probe is done by adding a call probe that changes the return address. Yeah, I thought that when I first saw it, but the probe is passed the address of the return instruction when it fires, and I can't see how you could get that if it was just invoked by modifying the return address on the call stack. Steven. pgpicSF3SLP7i.pgp Description: PGP signature
Re: Article on Sun's DTrace
On Wed, 7 Jul 2004, David Schultz wrote: The page referenced earlier in this thread pointed out that 6 staff-years went into DTrace. That's accurate, and we're not talking about part-time employees or people who don't know what they're doing. The D compiler aside, this is not a small matter of programming that can just be ported to a new OS or machine architecture in a few months. I don't doubt that DTrace took a long time to do. However, in most projects the design phase consumes a lot of time, and it is often the case that unforeseen problems or changes in the feature set cost the developers a lot of time. So while it might have taken six years to write DTrace the first time, I suspect it would take a fraction of that time to re-implement. (It certainly might be longer than a few months and I'm not going to quibble. We won't know the precise number until someone does the port.) -Dan ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-hackers To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Article on Sun's DTrace
On Thu, 8 Jul 2004, Steven Smith wrote: It's also possible to put probes on the return instruction of the function. I'm not sure how they're actually finding that, though. I think the return probe is done by adding a call probe that changes the return address. ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-hackers To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Article on Sun's DTrace
On Thu, Jul 08, 2004 at 03:46:24AM -0400, Daniel Ellard wrote: I don't doubt that DTrace took a long time to do. However, in most projects the design phase consumes a lot of time, and it is often the case that unforeseen problems or changes in the feature set cost the developers a lot of time. So while it might have taken six years to write DTrace the first time, I suspect it would take a fraction of that time to re-implement. (It certainly might be longer than a few months and I'm not going to quibble. We won't know the precise number until someone does the port.) They said 6 staff-years. This means if they have 6 people working on it full time, it took 1 year to complete. If they had 60 people full time, it took just over 5 weeks (technically, i doubt that would work practically). From speaking to a friend at sun, I do know it took a long time and a lot of effort, and was *not* a simple thing to implement. ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-hackers To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Article on Sun's DTrace
Hi Julian, Julian Elischer wrote: On Thu, 8 Jul 2004, Steven Smith wrote: It's also possible to put probes on the return instruction of the function. I'm not sure how they're actually finding that, though. I think the return probe is done by adding a call probe that changes the return address. The pointer to function seems to be the key word. return_p is the pointer to function. In case of normal mode return _p points the dummy address or Null pointer. In case of DTrace mode return_p points the address of probe. This may be magic of the DTrace. What all the DTrace do are to know the address of probe. If DTrace get the pointer of probes return_ps via structure or else, lang D would get the return value of the probe from memory. -- *** Eitarou Kamo Tel. +81 75 7035997 Fax +81 75 7035997 VoIP 050 10585997(domestic only) e-mail [EMAIL PROTECTED] For business: Feel free to mail me(above), please. Donation http://www.PayPal.Com GPG FingerPrint: 032D FDF9 D27B 23F7 9A81 BF4C 626C FBAA BC3A 9895 ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-hackers To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Article on Sun's DTrace
Hi, Avleen Vig wrote: They said 6 staff-years. This means if they have 6 people working on it full time, it took 1 year to complete. If they had 60 people full time, it took just over 5 weeks (technically, i doubt that would work practically). From speaking to a friend at sun, I do know it took a long time and a lot of effort, and was *not* a simple thing to implement. They may say so to add value-added. But I have never run it. So I don't know what it is. They ought to finish Solaris10 itself before DTrace... Say hello to Sun guys. Eitarou -- *** Eitarou Kamo Tel. +81 75 7035997 Fax +81 75 7035997 VoIP 050 10585997(domestic only) e-mail [EMAIL PROTECTED] For business: Feel free to mail me(above), please. Donation http://www.PayPal.Com GPG FingerPrint: 032D FDF9 D27B 23F7 9A81 BF4C 626C FBAA BC3A 9895 ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-hackers To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Article on Sun's DTrace
Avleen Vig [EMAIL PROTECTED] writes: They said 6 staff-years. No, they said three engineers working full-time for two years. DES -- Dag-Erling Smørgrav - [EMAIL PROTECTED] ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-hackers To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Article on Sun's DTrace
On Thu, Jul 08, 2004, Wilko Bulte wrote: On Thu, Jul 08, 2004 at 01:23:04AM -0700, Avleen Vig wrote: On Thu, Jul 08, 2004 at 03:46:24AM -0400, Daniel Ellard wrote: I don't doubt that DTrace took a long time to do. However, in most projects the design phase consumes a lot of time, and it is often the case that unforeseen problems or changes in the feature set cost the developers a lot of time. So while it might have taken six years to write DTrace the first time, I suspect it would take a fraction of that time to re-implement. (It certainly might be longer than a few months and I'm not going to quibble. We won't know the precise number until someone does the port.) They said 6 staff-years. This means if they have 6 people working on it full time, it took 1 year to complete. If they had 60 people full time, it took just over 5 weeks (technically, i doubt that would work practically). It works about as well as having 3 women deliver a baby in 3 months. ;-) See also Fred Brooks' book _The_Mythical_Man-Month_. But if you *do* happen to know 60 good programmers who are willing to work on FreeBSD full time for very little money, let me know and I'll see what I can do about that baby thing. ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-hackers To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Article on Sun's DTrace
Hi Dan, Daniel Ellard wrote: In a nutshell, here is what DTrace is about: - It has no impact on the system when it is not used. So you can leave it in all the time, instead of having a debug kernel and a production kernel. [I don't know how they achieve the no impact but they claim that they really mean no, not just negligible.] snip /snip Thanks good comment. This is the one everyone, of course me too, wants to know. I wanted to get brief instruction. So, you could think of it as a million debugging printf's magically inserted into the kernel for you along with a tool to analyze the output, but it's really much more sophisticated than that. It looks very nice. I wish I'd had it during my forays into the FreeBSD kernel. Is it hopelessly solaris-specific? Well, I was at the presentation that Bryan Cantrill gave at USENIX, where he was asked about the possibility of porting DTrace to linux. His response was something like well, we're really trying to encourage people to use the *best* possible operating system, so no. (Of course, one might argue that this means that a FreeBSD port is imminent, but I don't think that's what he meant.) -Dan FreeBSD has good features such as jail, chroot e.t.c. which can controll process or resources in parallel. So you need not port DTrace entirely. You can implement DTrace like one from scratch. Using legacy system sometimes makes new system feature. I would rather expect new one than porting. DTrace is one of example, I think. You may be able to fork new debug process in parallel in the future. If I dare name it, It's B(SD)Trace? But it's up to your effort. DTrace is a pioneer work. And for the people like me who bothers to put the debug lines in kernel this must be powerful tool. Eitarou -- *** Eitarou Kamo Tel. +81 75 7035997 Fax +81 75 7035997 VoIP 050 10585997(domestic only) e-mail [EMAIL PROTECTED] For business: Feel free to mail me(above), please. Donation http://www.PayPal.Com GPG FingerPrint: 032D FDF9 D27B 23F7 9A81 BF4C 626C FBAA BC3A 9895 ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-hackers To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Article on Sun's DTrace
FreeBSD has good features such as jail, chroot e.t.c. which can controll process or resources in parallel. So you need not port DTrace entirely. I think the control of ressources in a jail is very limited right now. I don't know if some work in that direction is in progress but it would be a great project for the future. You can implement DTrace like one from scratch. Using legacy system sometimes makes new system feature. I would rather expect new one than porting. DTrace is one of example, I think. You may be able to fork new debug process in parallel in the future. If I dare name it, It's B(SD)Trace? But it's up to your effort. DTrace is a pioneer work. And for the people like me who bothers to put the debug lines in kernel this must be powerful tool. Dtrace is (seems, at least) to be a very powerful tool. Eventual coders could/should get their inspiration out of the work of Sun engineers. But remember, the volunteers of the FreeBSD project aren't paid to do what they do. 2 years and 3 full-time engineers were needed to accomplish Dtrace so I think seeing a similar utility in the near future for FreeBSD is very hopeless. -- Nicolas Bérard Nault ([EMAIL PROTECTED]) http://www.quebecbsd.org http://www.xeatech.net http://staff.xeatech.net/nicobn Je ne sais pas avec quelles armes se combattra la troisième guerre mondiale mais je peux vous assurer que la quatrième se combattra avec des pierres et des bâtons. -- Albert Einstein. ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-hackers To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Article on Sun's DTrace
Hi Nicpolas, Nicolas Bérard Nault wrote: I think the control of ressources in a jail is very limited right now. I don't know if some work in that direction is in progress but it would be a great project for the future. If you attend the prj, You can tweak as you like, I think. Dtrace is (seems, at least) to be a very powerful tool. Eventual coders could/should get their inspiration out of the work of Sun engineers. But remember, the volunteers of the FreeBSD project aren't paid to do what they do. 2 years and 3 full-time engineers were needed to accomplish Dtrace so I think seeing a similar utility in the near future for FreeBSD is very hopeless. So what? -- *** Eitarou Kamo Tel. +81 75 7035997 Fax +81 75 7035997 VoIP 050 10585997(domestic only) e-mail [EMAIL PROTECTED] For business: Feel free to mail me(above), please. Donation http://www.PayPal.Com GPG FingerPrint: 032D FDF9 D27B 23F7 9A81 BF4C 626C FBAA BC3A 9895 ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-hackers To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Article on Sun's DTrace
Dtrace is (seems, at least) to be a very powerful tool. Eventual coders could/should get their inspiration out of the work of Sun engineers. But remember, the volunteers of the FreeBSD project aren't paid to do what they do. 2 years and 3 full-time engineers were needed to accomplish Dtrace so I think seeing a similar utility in the near future for FreeBSD is very hopeless. So what? Don't you think this is a candidate for funding from the FreeBSD foundation ? -- Nicolas Bérard Nault ([EMAIL PROTECTED]) http://staff.xeatech.net/nicobn PGP public key: 0x64159509 ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-hackers To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Article on Sun's DTrace
Hi Nicolas, (B (BNicolas Be'rard Nault wrote: (B (B Dtrace is (seems, at least) to be a very powerful tool. Eventual coders (B could/should get their inspiration out of the work of Sun engineers. (B But (B remember, the volunteers of the FreeBSD project aren't paid to do what (B they do. 2 years and 3 full-time engineers were needed to accomplish (B Dtrace so I think seeing a similar utility in the near future for (B FreeBSD (B is very hopeless. (B (B (B So what? (B (B (B Don't you think this is a candidate for funding from the FreeBSD (B foundation ? (B (B (B (BDo you think it is worth for that? (B (BBTW, don't you know my previous post which was not received to this list (Bbut you quoted. (B (B-- (B (B (B*** (BEitarou Kamo (B (BTel. +81 75 7035997 (BFax +81 75 7035997 (BVoIP 050 10585997(domestic only) (Be$B!>(Bmail [EMAIL PROTECTED] (B (BFor business: (BFeel free to mail me(above), please. (B (BDonation http://www.PayPal.Com (B (BGPG FingerPrint: (B032D FDF9 D27B 23F7 9A81 BF4C 626C FBAA BC3A 9895 (B (B (B (B (B (B (B___ (B[EMAIL PROTECTED] mailing list (Bhttp://lists.freebsd.org/mailman/listinfo/freebsd-hackers (BTo unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: Article on Sun's DTrace
On Wed, Jul 07, 2004, Eitarou Kamo wrote: FreeBSD has good features such as jail, chroot e.t.c. which can controll Solaris 10 has these features, too, but I'm not sure what that has to do with DTrace. process or resources in parallel. So you need not port DTrace entirely. You can implement DTrace like one from scratch. Using legacy system sometimes makes new system feature. I would rather expect new one than porting. DTrace is one of example, I think. You may be able to fork new debug process in parallel in the future. If I dare name it, It's B(SD)Trace? But it's up to your effort. DTrace is a pioneer work. And for the people like me who bothers to put the debug lines in kernel this must be powerful tool. The page referenced earlier in this thread pointed out that 6 staff-years went into DTrace. That's accurate, and we're not talking about part-time employees or people who don't know what they're doing. The D compiler aside, this is not a small matter of programming that can just be ported to a new OS or machine architecture in a few months. That said, there is prior work in this area, such as: http://oss.software.ibm.com/developer/opensource/linux/projects/dprobes/ But these other efforts don't come close to DTrace. ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-hackers To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Article on Sun's DTrace
Hi David, David Schultz wrote: On Wed, Jul 07, 2004, Eitarou Kamo wrote: FreeBSD has good features such as jail, chroot e.t.c. which can controll Solaris 10 has these features, too, but I'm not sure what that has to do with DTrace. If You use the feature with DTrace like tool, You can operate it in parallel. This is just a idea. I know Sol10 has chroot or zfs which behave like jail. What I'd like to say is that you need not insist on just DTrace you can develop original one. The page referenced earlier in this thread pointed out that 6 staff-years went into DTrace. That's accurate, and we're not talking about part-time employees or people who don't know what they're doing. The D compiler aside, this is not a small matter of programming that can just be ported to a new OS or machine architecture in a few months. I think so. If you develop the DTrace like tool, you need the resources at least equivalent with DTrace team, I guess. If you do, your develop team will make any action to get a resources. That said, there is prior work in this area, such as: http://oss.software.ibm.com/developer/opensource/linux/projects/dprobes/ But these other efforts don't come close to DTrace. -- *** Eitarou Kamo Tel. +81 75 7035997 Fax +81 75 7035997 VoIP 050 10585997(domestic only) e-mail [EMAIL PROTECTED] For business: Feel free to mail me(above), please. Donation http://www.PayPal.Com GPG FingerPrint: 032D FDF9 D27B 23F7 9A81 BF4C 626C FBAA BC3A 9895 ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-hackers To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Article on Sun's DTrace
In the last episode (Jul 07), David Schultz said: The page referenced earlier in this thread pointed out that 6 staff-years went into DTrace. That's accurate, and we're not talking about part-time employees or people who don't know what they're doing. The D compiler aside, this is not a small matter of programming that can just be ported to a new OS or machine architecture in a few months. Pawel Jakub Dawidek has already written a C-like language for his Cerber project that looks like it could be used for a FreeBSD DTrace. It doesn't support associative arrays for stat collecting like D does, but it's got just about everything else. If you just wanted to track syscalls, you could almost use Cerber as-is. http://cerber.sourceforge.net/ -- Dan Nelson [EMAIL PROTECTED] ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-hackers To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Article on Sun's DTrace
Hi Dan N, Dan Nelson wrote: In the last episode (Jul 07), David Schultz said: Pawel Jakub Dawidek has already written a C-like language for his Cerber project that looks like it could be used for a FreeBSD DTrace. It doesn't support associative arrays for stat collecting like D does, but it's got just about everything else. If you just wanted to track syscalls, you could almost use Cerber as-is. http://cerber.sourceforge.net/ Indeed, he have already made a action. Eitarou -- *** Eitarou Kamo Tel. +81 75 7035997 Fax +81 75 7035997 VoIP 050 10585997(domestic only) e-mail [EMAIL PROTECTED] For business: Feel free to mail me(above), please. Donation http://www.PayPal.Com GPG FingerPrint: 032D FDF9 D27B 23F7 9A81 BF4C 626C FBAA BC3A 9895 ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-hackers To unsubscribe, send any mail to [EMAIL PROTECTED]
closefrom (was Re: Article on Sun's DTrace)
Hi; The article suggests we should learn from Solaris and implement closefrom(3C): http://www.freebsd.org/cgi/man.cgi?query=closefromapropos=0sektion=0manpath=SunOS+5.9format=html And make sure we use it in sendmail (and Javac). cheers, Pedro. Yahoo! Companion - Scarica gratis la toolbar di Ricerca di Yahoo! http://companion.yahoo.it ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-hackers To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Article on Sun's DTrace
In a nutshell, here is what DTrace is about: - It has no impact on the system when it is not used. So you can leave it in all the time, instead of having a debug kernel and a production kernel. [I don't know how they achieve the no impact but they claim that they really mean no, not just negligible.] - It allows you to analyze pretty much any aspect of the kernel that you like, and it has hooks into userland as well. So if you have strange behaviors happening due to a badly-written process, you can track down what that process is doing. They demonstrate some nice examples of this. - The D language is what you use to specify how the trace info is filtered/processed/presented to you. D is not a complete programming language. It is highly constrained (i.e. no loops, no recursion, etc) in order to make sure that every path through a D program completes in finite time (otherwise, a bug in your D program might effectively hang the kernel, which is a no-no). So, you could think of it as a million debugging printf's magically inserted into the kernel for you along with a tool to analyze the output, but it's really much more sophisticated than that. It looks very nice. I wish I'd had it during my forays into the FreeBSD kernel. Is it hopelessly solaris-specific? Well, I was at the presentation that Bryan Cantrill gave at USENIX, where he was asked about the possibility of porting DTrace to linux. His response was something like well, we're really trying to encourage people to use the *best* possible operating system, so no. (Of course, one might argue that this means that a FreeBSD port is imminent, but I don't think that's what he meant.) -Dan ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-hackers To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Article on Sun's DTrace
On Wed, Jun 30, 2004, Bruce M Simpson wrote: This recently caught my eye: http://www.samag.com/documents/s=9171/sam0406h/0406h.htm There are a number of good sounding suggestions in there. DTrace is pure magic. It would be well worth your time to install Solaris 10 just to try out DTrace for a day. Keep in mind, however, that it took them several man-years to develop just for sparc64 and x86, so we're not talking about a port of it any time soon... ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-hackers To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Article on Sun's DTrace
HI Schultz and all, (B (BDavid Schultz wrote: (B (B On Wed, Jun 30, 2004, Bruce M Simpson wrote: (B (B (B This recently caught my eye: (B http://www.samag.com/documents/s=9171/sam0406h/0406h.htm (B (B There are a number of good sounding suggestions in there. (B (B (B DTrace is pure magic. It would be well worth your time to install (B Solaris 10 just to try out DTrace for a day. Keep in mind, (B however, that it took them several man-years to develop just for (B sparc64 and x86, so we're not talking about a port of it any time (B soon... (B (B (BIs DTrace open source? I know presence of DTrace. But (BSun's products usually aren't open source especially OS (Bitself. Recently JS(Sun's COO and president) announced (BSolaris might be open source in the near future. But he often (Btells a lie, sorry this may be unsuitable, tells a something uncertain. (B (BEitarou (B (B (B (B-- (B (B (B*** (BEitarou Kamo (B (BTel. +81 75 7035997 (BFax +81 75 7035997 (BVoIP 050 10585997(domestic only) (Be$B!>(Bmail [EMAIL PROTECTED] (B (BFor business: (BFeel free to mail me(above), please. (B (BDonation http://www.PayPal.Com (B (BGPG FingerPrint: (B032D FDF9 D27B 23F7 9A81 BF4C 626C FBAA BC3A 9895 (B (B (B (B (B (B (B___ (B[EMAIL PROTECTED] mailing list (Bhttp://lists.freebsd.org/mailman/listinfo/freebsd-hackers (BTo unsubscribe, send any mail to "[EMAIL PROTECTED]"
Re: Article on Sun's DTrace
Hi, David Schultz wrote: On Wed, Jun 30, 2004, Bruce M Simpson wrote: This recently caught my eye: http://www.samag.com/documents/s=9171/sam0406h/0406h.htm There are a number of good sounding suggestions in there. DTrace is pure magic. It would be well worth your time to install Solaris 10 just to try out DTrace for a day. Keep in mind, however, that it took them several man-years to develop just for sparc64 and x86, so we're not talking about a port of it any time soon... See also, http://www.sun.com/bigadmin/content/dtrace/ I haven't seen above well yet. But A article says that DTrace sounds like 30,000 lines of debug print. I have written about 50 lines of debug print in my 4.10-R kernel to chase my umass issue. If the output of it exists on a trace log file, It may be one of DTrace feature. I saw the some sort of debug print lines in the kernel source of 4.10-R. If you enable them and gather the information to a file, it may be nearly equal DTrace. But I can't warrant or guarantee you the performance. DTrace may be a kinda elegant debug mode kernel, I guess. Eitarou -- *** Eitarou Kamo Tel. +81 75 7035997 Fax +81 75 7035997 VoIP 050 10585997(domestic only) e-mail [EMAIL PROTECTED] For business: Feel free to mail me(above), please. Donation http://www.PayPal.Com GPG FingerPrint: 032D FDF9 D27B 23F7 9A81 BF4C 626C FBAA BC3A 9895 ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-hackers To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Article on Sun's DTrace
In local.freebsd-hackers, you wrote: I haven't seen above well yet. But A article says that DTrace sounds like 30,000 lines of debug print. No, already the first article tells you that they use a VM with byte-code for the C-like language D. And it's not compiled into the kernel but hooked in and removed on-the-fly. Volker -- http://www-i2.informatik.rwth-aachen.de/stolz/ *** PGP *** S/MIME Neu! Ändern Sie den Anfangstag Ihrer Woche ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-hackers To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Article on Sun's DTrace
Hi Volker and all, Volker Stolz wrote: In local.freebsd-hackers, you wrote: I haven't seen above well yet. But A article says that DTrace sounds like 30,000 lines of debug print. No, already the first article tells you that they use a VM with byte-code for the C-like language D. And it's not compiled into the kernel but hooked in and removed on-the-fly. Volker I don't know langage D well. But my guess is that they trim the info valuable from the debug print outputs. Sorry I haven't read the articles compleately yet. Eitarou -- *** Eitarou Kamo Tel. +81 75 7035997 Fax +81 75 7035997 VoIP 050 10585997(domestic only) e-mail [EMAIL PROTECTED] For business: Feel free to mail me(above), please. Donation http://www.PayPal.Com GPG FingerPrint: 032D FDF9 D27B 23F7 9A81 BF4C 626C FBAA BC3A 9895 ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-hackers To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Article on Sun's DTrace
Hi, Eitarou Kamo wrote: No, already the first article tells you that they use a VM with byte-code for the C-like language D. And it's not compiled into the kernel but hooked in and removed on-the-fly. I don't know langage D well. But my guess is that they trim the info valuable from the debug print outputs. Sorry I haven't read the articles compleately yet. Eitarou Again DTrace seems to observe the resources via system analyze tool and command. But language D seems to be great enough to call C/C++ function or even assembler. -- *** Eitarou Kamo Tel. +81 75 7035997 Fax +81 75 7035997 VoIP 050 10585997(domestic only) e-mail [EMAIL PROTECTED] For business: Feel free to mail me(above), please. Donation http://www.PayPal.Com GPG FingerPrint: 032D FDF9 D27B 23F7 9A81 BF4C 626C FBAA BC3A 9895 ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-hackers To unsubscribe, send any mail to [EMAIL PROTECTED]
Re: Article on Sun's DTrace
On Tuesday 29 June 2004 09:16 pm, Bruce M Simpson wrote: This recently caught my eye: http://www.samag.com/documents/s=9171/sam0406h/0406h.htm There are a number of good sounding suggestions in there. They gave a paper on it at USENIX ATC as well. -- John Baldwin [EMAIL PROTECTED]http://www.FreeBSD.org/~jhb/ Power Users Use the Power to Serve = http://www.FreeBSD.org ___ [EMAIL PROTECTED] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-hackers To unsubscribe, send any mail to [EMAIL PROTECTED]