From mboxrd@z Thu Jan 1 00:00:00 1970 From: "Rafael J. Wysocki" Subject: Re: [PATCH v7 4/9] acpi/arm64: Add GTDT table parse driver Date: Sat, 16 Jul 2016 14:35:16 +0200 Message-ID: <16716612.zdLA2RMyjn@vostro.rjw.lan> References: <1468432402-4872-1-git-send-email-fu.wei@linaro.org> <1580087.dlcVOb6fx1@vostro.rjw.lan> Mime-Version: 1.0 Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: 7Bit Return-path: Received: from cloudserver094114.home.net.pl ([79.96.170.134]:62447 "HELO cloudserver094114.home.net.pl" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with SMTP id S1751471AbcGPMa2 (ORCPT ); Sat, 16 Jul 2016 08:30:28 -0400 In-Reply-To: Sender: linux-acpi-owner@vger.kernel.org List-Id: linux-acpi@vger.kernel.org To: Fu Wei Cc: "Rafael J. Wysocki" , Len Brown , Daniel Lezcano , Thomas Gleixner , Marc Zyngier , Lorenzo Pieralisi , Sudeep Holla , Hanjun Guo , "linux-arm-kernel@lists.infradead.org" , "linaro-acpi@lists.linaro.org" , Linux Kernel Mailing List , ACPI Devel Maling List , rruigrok@codeaurora.org, harba@codeaurora.org, Christopher Covington , Timur Tabi , G Gregory , Al Stone , Jon Masters , wei@redhat.com, Arnd Bergmann , Wim Van Sebroeck , Catal On Saturday, July 16, 2016 10:24:35 AM Fu Wei wrote: > Hi Rafeal, > > On 16 July 2016 at 05:22, Rafael J. Wysocki wrote: > > On Saturday, July 16, 2016 12:32:14 AM Fu Wei wrote: > >> Hi Rafael, > >> > >> On 15 July 2016 at 21:07, Rafael J. Wysocki wrote: > >> > On Friday, July 15, 2016 02:15:27 PM Rafael J. Wysocki wrote: > >> >> On Friday, July 15, 2016 03:32:35 PM Fu Wei wrote: > >> >> > Hi Rafael, > >> >> > > >> > > >> > [cut] > >> > > >> >> > > > >> >> > >> + return 0; > >> >> > >> + } > >> >> > >> + > >> >> > >> + if (!gtdt->platform_timer_count) { > >> >> > >> + pr_info("No Platform Timer.\n"); > >> >> > >> + return 0; > >> >> > >> + } > >> >> > >> + > >> >> > >> + acpi_gtdt_desc.platform_timer_start = (void *)gtdt + > >> >> > >> + gtdt->platform_timer_offset; > >> >> > >> + if (acpi_gtdt_desc.platform_timer_start < > >> >> > >> + (void *)table + sizeof(struct acpi_table_gtdt)) { > >> >> > >> + pr_err(FW_BUG "Platform Timer pointer error.\n"); > >> >> > > > >> >> > > Why pr_err()? > >> >> > > >> >> > if (true), that means the GTDT table has bugs. > >> >> > > >> >> > >> >> And that's not a very useful piece of information unless you're debugging the > >> >> platform, is it? > >> > > >> > FWIW, I'm not a big fan of printing "your firmware is buggy" type of messages > >> > (especially at the "error" log level or higher) unless they can be clearly > >> > connected to a specific type of functional failure. > >> > > >> > So if you want to pring an error-level message, something like "I cannot do X > >> > because of the firmware bug Y" would be better IMO. > >> > >> So can I do this: > >> pr_err(FW_BUG "Can NOT init platform_timer pointer, because of the > >> GTDT table bug\n"); > >> > >> or pr_debug(FW_BUG "Can NOT init platform_timer_start, because of > >> platform_timer_offset bug in GTDT\n"); > >> > >> or just delete it? > >> > >> which one do you prefer? I think maybe should provide some clue for > >> users to fix the problem :-) > > > > And how exactly would they fix it then? > > > >> > >> any thought ? > > > > If you print variable or function names and the like, the message should be > > a debug one, because that's information that can only be understood by > > developers (some developers are users too, but they are a minority). > > > > If you want to report an error, say what is not working (or not available > > etc) and why (if you know the reason at the time the message is printed). > > Great thanks, I guess I got you point. > > maybe just a very simple message like: > pr_err(FW_BUG "Failed to init table: GTDT table is buggy.\n"); To understand this message one needs to know what "table" means here and what "GTDT" is. Also the prefix already will be something like "ACPI: GTDT:", so repeating part of it is not really useful IMO. Can you tell me please what's not going to work when that message is printed? Will the system boot at all then? If so, the functionality will be limited somehow I suppose. How is it going to be limited? > I will also check other pr_* , if I can update them OK, great! Thanks, Rafael