[lttng-dev] 回复：回复： 回复：回复： 回复： 回复： 回复： shm leak in traced application?

and only need some littile fixes,just like this time ,I need a single patch to ustcomm_recv_fds_unix_sock(). Again I am very very sorry for you take so much time to think our cases. Lttng is the best trace toolsets in the world. Thanks zhenyu.ren

[lttng-dev] 回复： 回复：回复： 回复： 回复： 回复： shm leak in traced application?

Oh, I see. I have an old ust(2.7). So I have no FD_CLOEXEC in ustcomm_recv_fds_unix_sock(). Thanks very much!!! zhenyu.ren -- 发件人：zhenyu.ren via lttng-dev 发送时间：2022年3月10日(星期四) 11:24 收件人：Mathieu Desnoyers 抄　送：lttng-dev 主　题：[lttng

[lttng-dev] 回复：回复： 回复： 回复： 回复： shm leak in traced application?

FD_CLOEXEC by fcntl() but not shm_fd. Why your patch do with wait_fd but not shm_fd? As far as I know, wait_fd is just a pipe and it seems not related to shm resource. ------ 发件人：Mathieu Desnoyers 发送时间：2022年3月10日(星期四) 00:46 收件人：zhenyu

[lttng-dev] 回复： 回复： 回复： 回复： shm leak in traced application?

Thanks a lot for reply. I do not reply it in bug tracker since I have not gotten a reliable way to reproduce the leak case. -- 发件人：Mathieu Desnoyers 发送时间：2022年3月8日(星期二) 23:26 收件人：zhenyu.ren 抄　送：Jonathan Rajotte ; lttng-dev 主　题

[lttng-dev] 回复： 回复： 回复： shm leak in traced application?

Hi, In shm_object_table_append_shm()/alloc_shm()， why not calling FD_CLOEXEC fcntl() to shmfds? I guess this omission leads to shm fds leak. Thanks zhenyu.ren -- 发件人：Jonathan Rajotte-Julien 发送时间：2022年2月25日(星期五) 22:31 收件人

[lttng-dev] 回复： 回复： shm leak in traced application?

without killling ust applications(I tried to dig into kernel shm_open and /dev/shm, but not found any ideas)? Thanks in advance zhenyu.ren ------ 发件人：zhenyu.ren via lttng-dev 发送时间：2022年2月23日(星期三) 23:09 收件人：lttng-dev 主　题：[lttng-dev] 回复

[lttng-dev] 回复： shm leak in traced application?

>"I found these items also exist in a traced application which is a long-time >running daemon" Even if lttng-sessiond has been killed!! Thanks zhenyu.ren ------ 发件人：zhenyu.ren via lttng-dev 发送时间：2022年2月23日(星期三) 22:44

[lttng-dev] shm leak in traced application?

Hi, There are many items such as "/dev/shm/ust-shm-consumer-81132 (deleted)" exist in lttng-sessiond fd spaces. I know it is the result of shm_open() and shm_unlnik() in create_posix_shm(). However, today, I found these items also exist in a traced application which is a long-time

[lttng-dev] 回复：回复： lttng-consumerd can NOT get notification to consume ring buffer data

f missing all the spans in that sub-buffer. Thanks zhenyu.ren -- 发件人：Mathieu Desnoyers 发送时间：2021年12月10日(星期五) 04:27 收件人：zhenyu.ren 抄　送：lttng-dev 主　题：Re: 回复：[lttng-dev] lttng-consumerd can NOT get notification to consume ring

[lttng-dev] 回复： lttng-consumerd can NOT get notification to consume ring buffer data

all span producers since they are too many... So, Can I make lttng-consumer just skip out this subbuffer and try to consume the next subbuffer? Thanks zhenyu.ren -- 发件人：Mathieu Desnoyers 发送时间：2021年12月8日(星期三) 04:17 收件人

[lttng-dev] lttng-consumerd can NOT get notification to consume ring buffer data

Hi, Lttng-dev: I found a strange problem related to lttng-consumed and ctf files recently. The ctf files belongs to some CPUs have been stopped rotating but other ctf files(belong to other CPUs) keeped working as usual. I am very sure all CPUs were producing spans all the time. #date; ls

[lttng-dev] what will happen during tracing if getcpu() doesn't return correct cpu number occasionally?

Hi, I know lttng-ust tracepoint process uses per cpu lockless ringbuffer algorithm for high performance so that it relies on getcpu() to return the cpu number on which the app is running. Unfortunately, I am working on arm that linux kernel does not support vdso getcpu() implemention and

[lttng-dev] 回复：回复： 回复： some tracepoints not exist in metadata?

发件人：Mathieu Desnoyers 发送时间：2020年7月1日(星期三) 20:07 收件人：zhenyu.ren 抄　送：lttng-dev 主　题：Re: 回复：[lttng-dev] 回复： some tracepoints not exist in metadata? Hi, Make sure you use single quotes in your shell around wildcards, else the shell will try to expand them to files in the current directory, e

[lttng-dev] 回复： 回复： some tracepoints not exist in metadata?

. -- 发件人：zhenyu.ren via lttng-dev 发送时间：2020年7月1日(星期三) 11:44 收件人：lttng-dev ; mathieu.desnoyers 主　题：[lttng-dev] 回复： some tracepoints not exist in metadata? I'd like to add that both "lttng list -u" and "lttng list $session_name" can get thos

[lttng-dev] 回复： some tracepoints not exist in metadata?

I'd like to add that both "lttng list -u" and "lttng list $session_name" can get those tracepoints' names... ------ 发件人：zhenyu.ren via lttng-dev 发送时间：2020年7月1日(星期三) 11:05 收件人：lttng-dev ; mathieu.desnoyers

[lttng-dev] some tracepoints not exist in metadata?

Hi, I find sometimes there are some tracepoints not exist in metadata file(but they are in the list of lttng list -u) and the end result is they can NOT be logged. I can not always reproduce it. It bothered me for 2 days. The lttng version is 2.7, I know it is not the latest version but

[lttng-dev] A question about staring lttng-sessiond

-consumerdexec ./lttng-sessiond --sig-parent -g  admin[admin@rs1k04068 /home/admin/install]sudo sh ./start  [zhenyu.ren@rs1k04068.et2sqa /home/zhenyu.ren]Note that user changed from admin to zhenyu.ren(zhenyu.ren sudo su admin previously). Where can I have a code fix to avoid logging out

[lttng-dev] 答复： 答复： 2.7.0 not support enable-event -k -a

--发件人：zhenyu.ren <zhenyu@aliyun.com>发送时间：2015年11月11日(星期三) 09:32收件人：Jérémie Galarneau <jeremie.galarn...@efficios.com>抄　送：lttng-dev <lttng-dev@lists.lttng.org>主　 题：[lttng-dev] 答复： 2.7.0 not support enable-event -k -a   In fact, I made some changes in lttng toolcha

[lttng-dev] 答复： 2.7.0 not support enable-event -k -a

zhenyu.ren--发件人：Jérémie Galarneau <jeremie.galarn...@efficios.com>发送时间：2015年11月10日(星期二) 23:09收件人：zhenyu.ren <zhenyu@aliyun.com>抄　送：lttng-dev <lttng-dev@lists.lttng.org>，Jonathan Rajotte Julien <jonathan.rajotte-jul...@efficios.com>主　题

[lttng-dev] lttng list -u return Error: Command error

];/* Can't go bigger than /proc/65535/cmdline */ I think path[20] is not big enough .Do you have any better idea? Thanks zhenyu.ren ___ lttng-dev mailing list lttng-dev@lists.lttng.org http://lists.lttng.org/cgi-bin/mailman/listinfo/lttng-dev

[lttng-dev] 答复： UST performance

-bef.tv_sec)*10+(aft.tv_nsec-bef.tv_nsec); printf(using %ld ns\n, diff); return 0; } Thankszhenyu.ren--发件人：Mathieu Desnoyers mathieu.desnoy...@efficios.com发送时间：2015年4月28日(星期二) 23:47收件人：zhenyu.ren zhenyu@aliyun.com抄　送

[lttng-dev] UST performance

Hi,dev    About a month ago,I asked a question about how to do time recording for ust event when vdso_clock_gettime not availble on my platform（Glibc too old）.Recently I implemented a lib to analyze vdso in memory and lookuped the symbol vdso_clock_gettime successfully.With this

[lttng-dev] 答复： clock_gettime vdso issue

to call    __vdso_clock_gettime  ? Thankszhenyu.ren --发件人：Mathieu Desnoyers mathieu.desnoy...@efficios.com发送时间：2015年3月27日(星期五) 00:33收件人：Jesper Derehag jdere...@hotmail.com抄　送：zhenyu.ren zhenyu@aliyun.com，lttng-dev lttng-dev

[lttng-dev] 答复：答复： Can lttng support kernel function tracer

Oh,What I want is something like lttng enable-event -a -k --function to enable ALL function trace. Thanks zhenyu.ren -- 发件人：zhenyu.ren zhenyu@aliyun.com 发送时间：2015年2月18日(星期三) 14:29 收件人：Francis Deslauriers francis.deslauri

[lttng-dev] Can lttng support kernel function tracer

Hi, It's well known that lttng can make use of kernel tracepoints to do block tracing ,system call tracing etc.Is it possible that lttng can produce anything like kernel function tracer does? Thanks zhenyu.ren ___ lttng-dev mailing list lttng

[lttng-dev] How to analysis a huge trace-collected data efficiently

Hi,   It's well known that tracing can produce huge data easily.So if we keep tracing on a production system for a while(hour/day/week)  ,we may get  a huge CTF file. It's very inefficient to analysis a huge CTF file ,especially the interested events are in some narrow time window.Is there any

[lttng-dev] trace time out of order

-2430 0 2.20GHz(24HT) ， 96G mem urcu 0.8 lttng 2.3 Thanks zhenyu.ren ___ lttng-dev mailing list lttng-dev@lists.lttng.org http://lists.lttng.org/cgi-bin/mailman/listinfo/lttng-dev

[lttng-dev] tracepoint sys_enter and sys_exit

, But where are these two tracepoints (hits) in kernel source code? Thanks zhenyu.ren ___ lttng-dev mailing list lttng-dev@lists.lttng.org http://lists.lttng.org/cgi-bin/mailman/listinfo/lttng-dev

Re: [lttng-dev] DEFINE_EVENT_MAP_NOARGS missing DEFINE_TRACE

Hi,dev Neither does TRACE_EVENT_NOARGS define in probe/define_trace.h Thanks zhenyu.ren -- 发件人：zhenyu.ren zhenyu@aliyun.com 发送时间：2014年3月28日(星期五) 12:07 收件人：lttng-dev lttng-dev@lists.lttng.org 主　题：[lttng-dev

Re: [lttng-dev] epoll_wait reboot kernel?

mov%rbx,%rdx | 49c: 4c 89 f2mov %r14,%rdx here rdx is the 3rd parameter to inline_memcpy(). Thanks zhenyu.ren Hi,lttng-dev I am working on rhel6(2.6.32.220) and ust is working very well.But

Re: [lttng-dev] epoll_wait reboot kernel?

that is the 1st parameter of inline_memcpy(dst,src,len). So you can image what will happen if len is too large. I don't know why GCC produce the wrong code.Does anyone has any clue? Thanks zhenyu.ren - Original Message - From: zhenyu.ren zhenyu@aliyun.com To: lttng-dev lttng-dev

[lttng-dev] epoll_wait reboot kernel?

. It seems lttng-tools in ubuntu 12.04 is using poll,isnt it? Does anyone have any suggestions to me ? PS:.lttng(ust/module/tool) version is 2.3.0 and urcu version is 0.8 Thanks in advance zhenyu.ren ___ lttng-dev mailing list lttng-dev@lists.lttng.org

[lttng-dev] no trace data are collected for UST

Hi, lttng-dev     All seems well but no trace data are produced for UST. (1) link static lib(.a) into target program   g++ -o target ... provider.o xx.o ... -ldl -llttng-ust -lurcu-cds -lurcu-bp -llttng-ust-tracepoint(2) run target and lttng-sessiond(rootuid)/consumerd

[lttng-dev] The problem with compiling multiple tracepoint sources with app directly

Hi,dev-list    It seems work well if there is only one tracepoint file(sample_tracepoint.h) including in app .i.esample_tracepoint.h#undef TRACEPOINT_PROVIDER #define TRACEPOINT_PROVIDER sample #if !defined(SAMPLE_TRACEPOINT_H) || defined(TRACEPOINT_HEADER_MULTI_READ) #define SAMPLE_TRACEPOINT_H

[lttng-dev] 答复: 答复: 答复: lttng list -u coredump

Can you point out what in the C99 specification makes you believe that the pointer is actually dereferenced in the expression (pos)-member ?AFAIK, this only evaluates to the address of field member within pos. The compiler should not dereference the pointer pos at all. Yes,you are right.