linux-kernel - Re: [PATCH 2/3] dma-mapping: Use trace_dma_alloc for dma_alloc* instead of using trace_dma

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

Message-ID: <20241018052738.GA20045@lst.de>
Date: Fri, 18 Oct 2024 07:27:38 +0200
From: Christoph Hellwig <hch@....de>
To: Sean Anderson <sean.anderson@...ux.dev>
Cc: Christoph Hellwig <hch@....de>, iommu@...ts.linux.dev,
	linux-trace-kernel@...r.kernel.org,
	Masami Hiramatsu <mhiramat@...nel.org>,
	linux-kernel@...r.kernel.org, Robin Murphy <robin.murphy@....com>,
	Mathieu Desnoyers <mathieu.desnoyers@...icios.com>,
	Steven Rostedt <rostedt@...dmis.org>,
	Marek Szyprowski <m.szyprowski@...sung.com>
Subject: Re: [PATCH 2/3] dma-mapping: Use trace_dma_alloc for dma_alloc*
 instead of using trace_dma_map

On Thu, Oct 17, 2024 at 02:13:53PM -0400, Sean Anderson wrote:
> +DECLARE_EVENT_CLASS(_dma_alloc,
>  	TP_PROTO(struct device *dev, void *virt_addr, dma_addr_t dma_addr,
>  		 size_t size, enum dma_data_direction dir, gfp_t flags,
>  		 unsigned long attrs),
> @@ -149,7 +149,60 @@ TRACE_EVENT(dma_alloc,
>  		decode_dma_attrs(__entry->attrs))
>  );
>  
> -TRACE_EVENT(dma_free,
> +DEFINE_EVENT(_dma_alloc, dma_alloc,
> +	TP_PROTO(struct device *dev, void *virt_addr, dma_addr_t dma_addr,
> +		 size_t size, enum dma_data_direction dir, gfp_t flags,
> +		 unsigned long attrs),
> +	TP_ARGS(dev, virt_addr, dma_addr, size, dir, flags, attrs));
> +
> +DEFINE_EVENT(_dma_alloc, dma_alloc_pages,

The scheme we used in XFS (fs/xfs/xfs_trace.h) for the event classes is
to give the class a _class postdix, and use macros to avoid the repeated
DEFINE_EVENT boilerplate.  Any chance you could rewrite this to use
a similar scheme?