From mboxrd@z Thu Jan 1 00:00:00 1970 From: "Stephen Biggs" Date: Wed, 09 Feb 2005 07:29:16 +0000 Subject: Re: [KJ] [PATCH] Reduction of compile warnings (Warning: long post Message-Id: <4209D7EC.4134.C3790@localhost> MIME-Version: 1 Content-Type: multipart/mixed; boundary="===============50738308859547909==" List-Id: References: <420953CC.2707.A99E9B@localhost> In-Reply-To: <420953CC.2707.A99E9B@localhost> To: kernel-janitors@vger.kernel.org --===============50738308859547909== Content-type: text/plain; charset=US-ASCII Content-transfer-encoding: 7BIT Content-description: Mail message body On 8 Feb 2005 at 18:43, Jim Nelson wrote: > Stephen Biggs wrote: > > Greetings, > > > > This will be my first attempt at submitting a patch, so please be > > reasonably gentle with me :-) I have tried to comply with all that I > > have read, but I most likely missed something. > > > > I attach 3 files in text/plain Ascii format; I think I understood the > > FAQ correctly that this is the right way to do things? > > > > My build tree consists of > > http://www.kernel.org/pub/linux/kernel/v2.6/testing/linux-2.6.11- > > rc3.tar.bz2 with Andrew Morton's -mm patch: > > http://www.kernel.org/pub/linux/kernel/people/akpm/patches/2.6/2.6.11- > > rc3/2.6.11-rc3-mm1/2.6.11-rc3-mm1.bz2 > > > > I am running Fedora Core 3 with gcc 3.4.2 (I THINK that is the correct > > version number - for me to be sure, I would have to reboot back into > > Linux, and I don't want to do this right now.) In any case, it is a > > fully updated vanilla Fedora Core 3 (slightly customized kernel build). > > > > It seems that 2.6.11-rc3-mm introduced a static function in swsusp.c > > that breaks the compile on my GCC. This . > > > > The error I get when trying to build the original is the first > > attachment to this post. > > The fix for this is part of the following patch. > > > > My sequence of execution for the build is: > > $ cd ~/linux-2.6.11-rc3-mm > > $ make mrproper && \ > > make allyesconfig && \ > > make bzImage 2>&1 | tee ~/warnings.txt > > > > I have managed to reduce the number of warnings from 229 to 88. > > > > I arrived at these numbers by first compiling the original without any > > modifications (except the swsusp.c fix to enable completion of build), > > and then a copy of the original with the following patch applied. I > > then executed: > > $ grep "warning:" ~/warnings.txt | wc -l > > > > The second attachment to this post is the set of build warnings that are > > still produced with this patch, distilled from a full build output, > > warnings.txt by using > > $ grep "^[^ ()]" warnings.txt > > > > The warnings that remain are mostly of the "deprecated" type for pm_* > > type warnings that I need guidence on what to do about, since I was not > > able to find any definitive guide to converting a driver with 2.4 style > > power management to 2.6 style power management. Thus, I left these alone. > > I have not seen too much traffic about this on either the kernel- > > janitors list or lkml, so this would be open territory. Any help on this > > would be appreciated. > > > > Most of the other warnings are about check_region which I see is already > > being address as far as I can tell. > > > > The rest of the warnings left are: deprecated for inter_module_* > > functions which seems to be already a topic of discussion in lkml (or > > was), mca-legacy issues which I have no idea about, a couple of warnings > > about mcdx not being edited which are legitimate, a hardcoded "#warning" > > in drivers/isdn/capi/capidrv.c (?), another hardcoded #warning in > > drivers/isdn/icn/icn.c, two assembly language warnings in > > drivers/scsi/ultrastor.c for a line of assembly language that I can't > > identify (and I can identify quite a few!). > > > > The third attachment to this post is the patch itself. Please let me > > know if this comes through ok... > > > > > > Nope. Got the swsusp-nocompile.txt, but everything else was munched by the > mailing list. Jim, Could you double-check, please? I got a copy of my post back from the mailing list and it is copacetic -- I can read and extract everything; 3 "Text/Plain" attachments with 3 "attachment information" parts put in by Pegasus. > > You'll definitely want to split the patch up by driver, or directory at the very > least. You'll also want to send the patches as inlined text, with the > Signed-off-by: line (see Documentation/SubmittingPatches and > http://www.zip.com.au/~akpm/linux/patches/stuff/tpp.txt for more pointers). Ok, I see that I missed a couple of things... the problem with my patch is that it is very many one or two line changes in many files. I count, using: grep "^diff" linux-2.6.11-rc3-mm_compile-warnings.patch | wc -l 49 different files... that is a bit much to break up, no? What do I do about this? > > There's a few good scripts out there for sending groups of patches to mailing > lists. I've also hacked up something to chop up a large diff by file, but I'm > almost embarassed to show it in public :) Can you provide links to these scripts? > > Jim > --===============50738308859547909== Content-Type: text/plain; charset="iso-8859-1" MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable Content-Disposition: inline _______________________________________________ Kernel-janitors mailing list Kernel-janitors@lists.osdl.org http://lists.osdl.org/mailman/listinfo/kernel-janitors --===============50738308859547909==--