linux-kernel - Re: [PATCH v9] dma-buf: add some tracepoints to debug.

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <20251223114424.1c539f7a@gandalf.local.home>
Date: Tue, 23 Dec 2025 11:44:24 -0500
From: Steven Rostedt <rostedt@...dmis.org>
To: Xiang Gao <gxxa03070307@...il.com>
Cc: sumit.semwal@...aro.org, christian.koenig@....com, mhiramat@...nel.org,
 linux-media@...r.kernel.org, dri-devel@...ts.freedesktop.org,
 linux-kernel@...r.kernel.org, mathieu.desnoyers@...icios.com,
 dhowells@...hat.com, kuba@...nel.org, brauner@...nel.org,
 akpm@...ux-foundation.org, linux-trace-kernel@...r.kernel.org, gaoxiang17
 <gaoxiang17@...omi.com>
Subject: Re: [PATCH v9] dma-buf: add some tracepoints to debug.

On Tue, 23 Dec 2025 11:27:49 +0800
Xiang Gao <gxxa03070307@...il.com> wrote:

>  
> +#define CREATE_TRACE_POINTS
> +#include <trace/events/dma_buf.h>
> +
> +/*
> + * dmabuf->name must be accessed with holding dmabuf->name_lock.
> + * we need to take the lock around the tracepoint call itself where
> + * it is called in the code.
> + *
> + * Note: FUNC##_enabled() is a static branch that will only
> + *       be set when the trace event is enabled.
> + */
> +#define DMA_BUF_TRACE(FUNC, ...)					\
> +	do {											\
> +		if (FUNC##_enabled()) {						\
> +			guard(spinlock)(&dmabuf->name_lock);	\
> +			FUNC(__VA_ARGS__);						\
> +		} else if (IS_ENABLED(CONFIG_LOCKDEP)) {	\
> +			/* Expose this lock when lockdep is enabled */	\
> +			guard(spinlock)(&dmabuf->name_lock);	\
> +		}											\
> +	} while (0)
> +

I hate to make another comment here, but I was just thinking that this can
be made to look a little nicer. Basically, we want to make sure that when
LOCKDEP is active, we always take the lock. But we also need to take the
lock when tracing is enabled. The tracepoint itself is a static branch,
which means it is a nop when not active, so there's no real problem with
calling it. Thus, this could look better as:

#define DMA_BUF_TRACE(FUNC, ...)					\
	do {								\
		/* Always expose lock if lockdep is enabled */		\
		if (IS_ENABLED(CONFIG_LOCKDEP) || FUNC##_enabled()) {	\
			guard(spinlock)(&dmabuf->name_lock);		\
			FUNC(__VA_ARGS__);				\
		}							\
	} while (0)

-- Steve