linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Steven Rostedt <rostedt@goodmis.org>
To: Prasad Sodagudi <psodagud@codeaurora.org>
Cc: mingo@redhat.com, keescook@chromium.org,
	saiprakash.ranjan@codeaurora.org,
	linux-arm-kernel@lists.infradead.org,
	linux-kernel@vger.kernel.org, linux-arm-msm@vger.kernel.org,
	gregkh@linuxfoundation.org, anton@enomsg.org, arnd@arndb.de,
	catalin.marinas@arm.com, ccross@android.com, jbaron@akamai.com,
	jim.cromie@gmail.com, joe@perches.com, joel@joelfernandes.org
Subject: Re: [PATCH] tracing: Add register read and write tracing support
Date: Mon, 28 Sep 2020 10:55:01 -0400	[thread overview]
Message-ID: <20200928105501.7e29df65@oasis.local.home> (raw)
In-Reply-To: <1601253290-400618-2-git-send-email-psodagud@codeaurora.org>

On Sun, 27 Sep 2020 17:34:50 -0700
Prasad Sodagudi <psodagud@codeaurora.org> wrote:

> Add register read/write operations tracing support.
> ftrace events helps trace register read and write
> location details of memory mapped IO registers. Also
> add _no_log variants the writel_relaxed/readl_relaed
> APIs to avoid excessive logging for certain register
> operations.

As mentioned elsewhere, I don't see a reason for "nolog" variants if it
is just to avoid logging too much. You can easily filter on the
recording side.


> --- /dev/null
> +++ b/include/linux/iorw.h
> @@ -0,0 +1,20 @@
> +/* SPDX-License-Identifier: GPL-2.0 */
> +/*
> + *
> + */
> +#ifndef __LOG_IORW_H__
> +#define __LOG_IORW_H__
> +
> +#include <linux/types.h>
> +
> +#if IS_ENABLED(CONFIG_TRACE_RW)
> +void log_write_io(volatile void __iomem *addr);
> +void log_read_io(const volatile void __iomem *addr);

So basically, this is always doing a function call, even when tracing
is not enabled. You may want to turn this into a macro, and use the new
interface I'm about to push:

 See https://lore.kernel.org/r/20200925211206.423598568@goodmis.org

Although I'm about to push a v3 (found a config that breaks msr.h)

#if IS_ENABLED(CONFIG_TRACE_RW)
#include <linux/atomic.h>
#include <linux/tracepoint-defs.h>

DECLARE_TRACEPOINT(rwio_write);
DECLARE_TRACEPOINT(rwio_read);

void __log_write_io(volatile void __iomem *addr);
void __log_read_io(const volatile void __iomem *addr);

#define log_write_io(addr) \
	if (tracepoint_enabled(rwio_write)
		__log_write_io(addr)

#define log_read_io(addr) \
	if (tracepoint_enabled(rwio_read)
		__log_read_io(addr)


> +#else
> +static inline void log_write_io(volatile void __iomem *addr)
> +{ }
> +static inline void log_read_io(const volatile void __iomem *addr)
> +{ }
> +#endif /* CONFIG_TRACE_RW */
> +
> +#endif /* __LOG_IORW_H__  */
> diff --git a/include/trace/events/rwio.h b/include/trace/events/rwio.h
> new file mode 100644
> index 0000000..b829629
> --- /dev/null
> +++ b/include/trace/events/rwio.h
> @@ -0,0 +1,51 @@
> +/* SPDX-License-Identifier: GPL-2.0 */
> +
> +#undef TRACE_SYSTEM
> +#define TRACE_SYSTEM rwio
> +
> +#if !defined(_TRACE_RWIO_H) || defined(TRACE_HEADER_MULTI_READ)
> +#define _TRACE_RWIO_H
> +
> +#include <linux/tracepoint.h>
> +
> +TRACE_EVENT(raw_write,

"raw" is too generic. Call this rwio_write.

> +
> +	TP_PROTO(unsigned long fn, volatile void __iomem *addr),
> +
> +	TP_ARGS(fn, addr),
> +
> +	TP_STRUCT__entry(
> +		__field(u64, fn)
> +		__field(u64, addr)
> +	),
> +
> +	TP_fast_assign(
> +		__entry->fn = fn;
> +		__entry->addr = (u64)addr;
> +	),
> +
> +	TP_printk("%pS write addr=%p\n", __entry->fn, __entry->addr)
> +);
> +
> +TRACE_EVENT(raw_read,

And this "rwio_read"

-- Steve

> +
> +	TP_PROTO(unsigned long fn, const volatile void __iomem *addr),
> +
> +	TP_ARGS(fn, addr),
> +
> +	TP_STRUCT__entry(
> +		__field(u64, fn)
> +		__field(u64, addr)
> +	),
> +
> +	TP_fast_assign(
> +		__entry->fn = fn;
> +		__entry->addr = (u64)addr;
> +	),
> +
> +	TP_printk("%pS read addr=%p\n", __entry->fn, __entry->addr)
> +);
> +
> +#endif /* _TRACE_PREEMPTIRQ_H */
> +

  parent reply	other threads:[~2020-09-28 14:55 UTC|newest]

Thread overview: 6+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-09-28  0:34 [PATCH] Register read and writes tracing Prasad Sodagudi
2020-09-28  0:34 ` [PATCH] tracing: Add register read and write tracing support Prasad Sodagudi
2020-09-28  5:17   ` Greg KH
2020-09-28  5:20   ` Greg KH
2020-09-28 14:55   ` Steven Rostedt [this message]
2020-09-28  5:45 ` [PATCH] Register read and writes tracing Sai Prakash Ranjan

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20200928105501.7e29df65@oasis.local.home \
    --to=rostedt@goodmis.org \
    --cc=anton@enomsg.org \
    --cc=arnd@arndb.de \
    --cc=catalin.marinas@arm.com \
    --cc=ccross@android.com \
    --cc=gregkh@linuxfoundation.org \
    --cc=jbaron@akamai.com \
    --cc=jim.cromie@gmail.com \
    --cc=joe@perches.com \
    --cc=joel@joelfernandes.org \
    --cc=keescook@chromium.org \
    --cc=linux-arm-kernel@lists.infradead.org \
    --cc=linux-arm-msm@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mingo@redhat.com \
    --cc=psodagud@codeaurora.org \
    --cc=saiprakash.ranjan@codeaurora.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).