[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-Id: <20160927114026.eafeb5a2ea9a75ae52229bbe@kernel.org>
Date: Tue, 27 Sep 2016 11:40:26 -0700
From: Masami Hiramatsu <mhiramat@...nel.org>
To: Arnaldo Carvalho de Melo <acme@...nel.org>
Cc: wang@...radead.org, Jiri Olsa <jolsa@...nel.org>,
David Ahern <dsahern@...il.com>,
Namhyung Kim <namhyung@...nel.org>,
Linux Kernel Mailing List <linux-kernel@...r.kernel.org>
Subject: Re: using SDT events
On Tue, 27 Sep 2016 10:55:38 -0300
Arnaldo Carvalho de Melo <acme@...nel.org> wrote:
> Masami,
>
> Testing something else (event aliases/JSON) I saw the many sdt
> events available via 'perf list', decided to use the usual workflow,
> failed, what am I doing wrong?
No, that is not supported yet. So I dropped perf-list support from
the latest series...
https://lkml.org/lkml/2016/7/1/133
The unresolved problems are
- how we treat sdt events after recorded? remove or leave it as tracepoint?
- what should be done if we found the tracepoint which has same name of the
SDT. (we can not reserve tracepoint name beforehand)
- Sometimes SDT name can be conflict, even on same binary.
>
>
> [root@...et ~]# perf list sdt_qemu:pci*
>
> List of pre-defined events (to be used in -e):
>
> sdt_qemu:pci_cfg_read [SDT event]
> sdt_qemu:pci_cfg_write [SDT event]
> sdt_qemu:pci_update_mappings_add [SDT event]
> sdt_qemu:pci_update_mappings_del [SDT event]
> [root@...et ~]# perf record -e sdt_qemu:pci* -a
> event syntax error: 'sdt_qemu:pci*'
> \___ unknown tracepoint
>
> Error: File /sys/kernel/debug/tracing/events/sdt_qemu/pci* not found.
> Hint: Perhaps this kernel misses some CONFIG_ setting to enable this
> feature?.
>
> Run 'perf list' for a list of valid events
>
> Usage: perf record [<options>] [<command>]
> or: perf record [<options>] -- <command> [<options>]
>
> -e, --event <event> event selector. use 'perf list' to list
> available events
> [root@...et ~]#
--
Masami Hiramatsu <mhiramat@...nel.org>
Powered by blists - more mailing lists