[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20241126120237.1598854d@gandalf.local.home>
Date: Tue, 26 Nov 2024 12:02:37 -0500
From: Steven Rostedt <rostedt@...dmis.org>
To: Shiju Jose <shiju.jose@...wei.com>
Cc: "dave.jiang@...el.com" <dave.jiang@...el.com>,
"dan.j.williams@...el.com" <dan.j.williams@...el.com>, Jonathan Cameron
<jonathan.cameron@...wei.com>, "alison.schofield@...el.com"
<alison.schofield@...el.com>, "nifan.cxl@...il.com" <nifan.cxl@...il.com>,
"vishal.l.verma@...el.com" <vishal.l.verma@...el.com>,
"ira.weiny@...el.com" <ira.weiny@...el.com>, "dave@...olabs.net"
<dave@...olabs.net>, "linux-cxl@...r.kernel.org"
<linux-cxl@...r.kernel.org>, "linux-kernel@...r.kernel.org"
<linux-kernel@...r.kernel.org>, Linuxarm <linuxarm@...wei.com>, tanxiaofei
<tanxiaofei@...wei.com>, "Zengtao (B)" <prime.zeng@...ilicon.com>
Subject: Re: [PATCH v4 3/6] cxl/events: Update General Media Event Record to
CXL spec rev 3.1
On Tue, 26 Nov 2024 11:51:23 +0000
Shiju Jose <shiju.jose@...wei.com> wrote:
> We are encountering a parsing error ("FAILED TO PARSE") from libtraceevent when it
> tries to parse some of the CXL trace events for the user-space tool rasdaemon.
> This issue appeared after new fields were added to the trace events.
> It was found that the issue does not occur when all or some of the decoded strings
> for the event's data and flags are removed from the TP_printk() function in the kernel,
> and only the values are printed instead.
> https://elixir.bootlin.com/linux/v6.12/source/drivers/cxl/core/trace.h
> https://lore.kernel.org/lkml/20241120093745.1847-1-shiju.jose@huawei.com/
>
> Below is the information from the debugging in libtraceevent:
> The failure occurs in the following functions and locations within libtraceevent:
> File: src/event-parse.c
> Function: event_read_format()
> ret = event_read_fields(event->tep, event, &event->format.fields); if (ret < 0)
> return ret;
>
> Function: event_read_fields()
> if (test_type_token(type, token, TEP_EVENT_ITEM, "field"))
> goto fail;
>
> Can you recognize if there are any limitations or issues that would prevent
> libtraceevent from parsing the trace event in the condition described above?
Can you show me the output of the format files for the affected trace events:
# cat /sys/kernel/tracing/cxl/<affected_event>/format
You can attach it too if your email does whitespace mangling.
Thanks,
-- Steve
Powered by blists - more mailing lists