Re: linux-next: add utrace tree
* Ananth N Mavinakayanahalli ana...@in.ibm.com wrote: On Fri, Jan 29, 2010 at 08:39:07AM +0100, Ingo Molnar wrote: ... When we merged kprobes ~10 years ago we made the (rather bad) mistake of merging a raw, opaque facility and leaving 'the rest' up to some other entity. IBM kprobes hackers vanished the day the original kprobes code went upstream and the high level entity never truly materialized in-kernel, for nearly a decade! I don't know what you are referring to here... Kprobes was merged in 2.6.9 (~August 2004 -- less than 6 years ago). [...] Ok, 6 years then :-) [...] Since then, we did work on ports to powerpc and s390. We implemented kretprobes. We made it much scalable using RCU; we did the powerpc booster to skip single-step when possible, not to mention various bug fixes over the years. Except it had no real in-kernel user. Yes, we did not do the perf integration, but perf did not exist then, either. Its simply wrong to say people 'vanished'. It has certainly was a bit stale for years - and with no real users that's certainly not a surprise. That has changed recently so i'm not complaining. We just dont want to repeat the same mistake with uprobes. Ingo
Re: linux-next: add utrace tree
On Fri, Jan 29, 2010 at 10:11:16AM +0100, Ingo Molnar wrote: * Ananth N Mavinakayanahalli ana...@in.ibm.com wrote: On Fri, Jan 29, 2010 at 08:39:07AM +0100, Ingo Molnar wrote: ... When we merged kprobes ~10 years ago we made the (rather bad) mistake of merging a raw, opaque facility and leaving 'the rest' up to some other entity. IBM kprobes hackers vanished the day the original kprobes code went upstream and the high level entity never truly materialized in-kernel, for nearly a decade! I don't know what you are referring to here... Kprobes was merged in 2.6.9 (~August 2004 -- less than 6 years ago). [...] Ok, 6 years then :-) [...] Since then, we did work on ports to powerpc and s390. We implemented kretprobes. We made it much scalable using RCU; we did the powerpc booster to skip single-step when possible, not to mention various bug fixes over the years. Except it had no real in-kernel user. Not that I want to rebut you Ingo, but there were in-kernel users since 2006 (net/ipv4/tcp_probe.c) :-) Aside, I am also glad that we have more flexibility with the perf integration. Ananth
Re: linux-next: add utrace tree
* Ananth N Mavinakayanahalli ana...@in.ibm.com wrote: On Fri, Jan 29, 2010 at 10:11:16AM +0100, Ingo Molnar wrote: * Ananth N Mavinakayanahalli ana...@in.ibm.com wrote: On Fri, Jan 29, 2010 at 08:39:07AM +0100, Ingo Molnar wrote: ... When we merged kprobes ~10 years ago we made the (rather bad) mistake of merging a raw, opaque facility and leaving 'the rest' up to some other entity. IBM kprobes hackers vanished the day the original kprobes code went upstream and the high level entity never truly materialized in-kernel, for nearly a decade! I don't know what you are referring to here... Kprobes was merged in 2.6.9 (~August 2004 -- less than 6 years ago). [...] Ok, 6 years then :-) [...] Since then, we did work on ports to powerpc and s390. We implemented kretprobes. We made it much scalable using RCU; we did the powerpc booster to skip single-step when possible, not to mention various bug fixes over the years. Except it had no real in-kernel user. Not that I want to rebut you Ingo, but there were in-kernel users since 2006 (net/ipv4/tcp_probe.c) :-) i said 'real' users. That usage in tcp_probe.c was (and is) really minimal and never expanded really. Aside, I am also glad that we have more flexibility with the perf integration. ok, good :) Ingo
Final do Festival Termómetro - O Convite
Se não visualizares esta página correctamente, clica aqui http://www.divulgacaodigital.com/festival_termometro/newsletters/1/email.html Adiciona-nos à tua safe-list, para garantir que recebes sempre a info dos nossos eventos. Olá, No próximo Sábado, dia 30 de Janeiro do profícuo ano de 2010, acontece a final do Festival Termómetro no Lx Factory em Lisboa e gostaríamos de contar com a presença de todos, sobretudo da pessoa que está a ler o presente parágrafo, justamente esta, e que pelos vistos - quase apostamos - irá ler o segundo. Não tínhamos dito? Pois adivinhamos e já que estão no segundo parágrafo, saibam pois que esta será uma final diferente de todas as outras . E porquê? Porque é uma edição comemorativa dos 15 anos do festival, porque pela primeira vez terá em competição bandas de outros países , do estrangeiro - sem ser da América por Deus! - e, como convidados especiais: Manuel Cruz, Samuel Úria e B Fachada. Três dos mais geniais intérpretes dos dias de hoje, 3 vozes, que pela primeira vez se decidiram juntar, para tocar temas uns dos outros, emprestando a sua voz e talento a cada um deles, e tornando um dia absolutamente normal – ninguém dava nada pelo 30 de Janeiro até hoje - em absolutamente inédito, que muito nos honra de ter provocado. E assim, para que possam ver isto, para que se inscrevam também na história e não se sintam “historio-excluídos”, devem desde logo fazer a vossa reserva no site oficial do festival: www.termometro-online.com http://www.termometro-online.com/ . Um site muito bonito, com cores garridas e onde podem saber mais sobre cada uma das bandas presentes, as fotos de cada uma das eliminatórias e um sem número de outras coisas que agora não nos lembramos. O que lembramos sim, é que é este Sábado, é a partir das 23 horas, é no lx factory e tem transmissão em directo na Speaky.Tv http://speaky.tv/ . É a final do Festival Termómetro e este é o convite especial, que é extensível a todos os que no vosso entender vão gostar de uma coisa destas. E mesmo os que não vão gostar caramba. Divulguem pois esta mensagem por todos os vossos contactos. O nosso sentido e antecipado agradecimento por isso. Já agora podem igualmente fazer gratuitamente o download do novo número da revista 365 aqui: www.revista365.com http://www.revista365.com/ . Do sempre vosso Fernando Alvim (Director do Festival Termómetro) - é favor divulgar - http://www.divulgacaodigital.com Para RE-ENVIAR / To FORWARD - http://www.divulgacaodigital.com/festival_termometro/?p=forwarduid=8796d6f78d5efbb8958965a0e70ab9c8mid=13 Para REMOVER / To REMOVE - http://www.divulgacaodigital.com/festival_termometro/?p=unsubscribeuid=8796d6f78d5efbb8958965a0e70ab9c8 Para MODIFICAR / To MODIFY - http://www.divulgacaodigital.com/festival_termometro/?p=preferencesuid=8796d6f78d5efbb8958965a0e70ab9c8 -- Powered by PHPlist, www.phplist.com -- inline: powerphplist.png
Re: linux-next: add utrace tree
Ingo Molnar mi...@elte.hu writes: [...] So, to sum it up: utrace XOL, which is rather complex already, needs even more complexity (which is not yet implemented) than the much simpler common-case emulator approach i outlined, just to break even with the performance of the much simpler approach. [...] Is it an uncontroversial claim that emulation of CISC instructions should perform better than their native execution, followed by an int3 (as in the simplest working scheme) or boosting (as done by kprobes)? From my experience with simulators, simple software emulation of cpus can be hundreds of times slower or worse than native execution. - FChE