From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from [213.170.72.194] (helo=shelob.oktetlabs.ru) by canuck.infradead.org with esmtp (Exim 4.42 #1 (Red Hat Linux)) id 1CSwX4-00061H-09 for linux-mtd@lists.infradead.org; Sat, 13 Nov 2004 06:57:31 -0500 Message-ID: <4195F689.9070705@yandex.ru> Date: Sat, 13 Nov 2004 14:56:57 +0300 From: "Artem B. Bityuckiy" MIME-Version: 1.0 To: David Woodhouse References: <4195EF83.8050104@yandex.ru> <1100345900.8015.17.camel@localhost.localdomain> In-Reply-To: <1100345900.8015.17.camel@localhost.localdomain> Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Cc: linux-mtd@lists.infradead.org Subject: Re: JFFS2 prints List-Id: Linux MTD discussion mailing list List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , David Woodhouse wrote: > On Sat, 2004-11-13 at 14:26 +0300, Artem B. Bityuckiy wrote: > >>But the main thing why I think it would be nice to introduce macros like >>these is that the messages will be a bit standard. I mean that all >>errors will have the same prefix. It is useful when you develop JFFS2 >>with debugging output on. In this case you will just s/JFFS2 Error and >>see all the error messages. But currently, it is hard to find all the >>error messages in JFFS2 because some of them use prefix "Error", some >>just contain the word "failed", etc. > > > I'm not sure that just a standard prefix will really help with > debugging. The trick is still to know what's important, and what's > noise. And running it with full debugging is slow enough already over a > 115200 baud serial line, without adding more stuff to each line :) I'm debugging JFFS2 on the P4 host with the Flash simulator, so I do not restricted by the serial line speed. May be somebody use USB which is also faster. The situation is that I have *lots* of debugging output and want to see all warnings/errors. May be not only I :-) Yes, some of the output is unneeded and too noisy, but many is very useful. Moreover, we will not need to write the name of function in the debugging printk. I mean D1(printk(KERN_DEBUG "some_long_function_name(): some important message")); We will use shorter calls like JFFS2DMSG(1, "some important message"); > > I'd much rather see an effort to weed out the unnecessary prints at > debug level 1, moving them to level 2. I don't consider level 2 to be > particularly useful in general; it's _too_ verbose. > Yes, I've mentioned this and I switch off the lists content output when developing JFFS2. I think these messages should be moved to the level 2. So, do you think we should not introduce such output macros? -- Best Regards, Artem B. Bityuckiy, St.-Petersburg, Russia.