From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from [140.186.70.92] (port=48011 helo=eggs.gnu.org) by lists.gnu.org with esmtp (Exim 4.43) id 1PauL6-0001OD-8V for qemu-devel@nongnu.org; Thu, 06 Jan 2011 13:13:36 -0500 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1PauL2-0000vg-6L for qemu-devel@nongnu.org; Thu, 06 Jan 2011 13:13:31 -0500 Received: from a.painless.aaisp.net.uk ([81.187.30.51]:52088) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1PauL1-0000uR-Us for qemu-devel@nongnu.org; Thu, 06 Jan 2011 13:13:28 -0500 Date: Thu, 6 Jan 2011 18:13:20 +0000 From: Stuart Brady Subject: Re: [Qemu-devel] [PATCH 1/6] softfloat: remove HPPA specific code Message-ID: <20110106181320.GA2180@zubnet.me.uk> References: <1294065273-30274-1-git-send-email-aurelien@aurel32.net> <1294065273-30274-2-git-send-email-aurelien@aurel32.net> <151C200F-E9B1-4BDC-B234-C0B03D3086D2@web.de> <20110104200711.GC3615@hall.aurel32.net> <10FC1308-1CC8-41DF-A48C-AD104BCE6E11@web.de> <20110104235604.GD3615@hall.aurel32.net> <072B9AB0-7B44-4A79-80FE-A3D95914AAD4@web.de> <20110105102106.GD15256@volta.aurel32.net> <20110105231306.GA20254@zubnet.me.uk> MIME-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Disposition: inline In-Reply-To: Content-Transfer-Encoding: quoted-printable List-Id: qemu-devel.nongnu.org List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Peter Maydell Cc: Andreas =?iso-8859-1?Q?F=E4rber?= , Riku Voipio , QEMU Developers , Aurelien Jarno On Thu, Jan 06, 2011 at 08:58:17AM +0000, Peter Maydell wrote: > On 5 January 2011 23:13, Stuart Brady wrote: > > I do have a few concerns regarding SoftFloat, though: > > > > =A0 FIXMEs should be left in the code (or a document maintained on th= e > > =A0 Wiki) to keep track of which architectures have been considered > > =A0 (which I believe are x86, arm, mips, ppc) and which ones haven't. > > =A0 This is in reference to one particular FIXME that was removed, > > =A0 but perhaps shouldn't have been. >=20 > Which one? The only one I know I removed was the one in > the patch that implemented the thing it was complaining about, > but perhaps I took out another by accident... The patch is question was: softfloat: Implement flushing input denormals to zero The comment is: /* FIXME: Flush-To-Zero only effects results. Denormal inputs should also be flushed to zero. */ Do you know whether ARM is the only target architecture supported by QEMU that requires this behaviour? If there are any architectures where we simply don't know whether the current behaviour is correct, we should document that somewhere. For any target-specific behaviour, I really feel that we should have architectures listed explicitly rather than having using a default #else case (and that the #else case should simply contain a #error), as it may be worth guarding against any newly added targets incorrectly picking the default behaviour (as has happened in the past). Presumably, sNaNs on any target where they are indicated by a zero bit should be quietened by simply setting that bit. However, if a target doesn't define SNAN_BIT_IS_ONE, that may simply be because it was forgotten. This is why I don't like using #ifndef in general... much better, IMO, to define as '0' or '1', and then -Werror=3Dundef can catch anything that gets overlooked. (Except now the problem is any erroneous use of #ifdef with such definitions that might creep in...) > > =A0 Is there any plan to deal with use of float*_is_quiet_nan(), wher= e > > =A0 float*_is_any_nan() was intended? =A0These should really either b= e > > =A0 fixed (and tested), or if not, a FIXME should be added. >=20 > I was planning to do the ARM related ones, at least (although > they are in the pretty-much-dead FPE-emulation code for > linux user-mode). >=20 > I would be more inclined to track that sort of thing in a bug tracker > than by sprinkling the code with FIXME comments. I'd definitely agree that FIXMEs are not as good as bug reports, and/or possibly even a Features/SoftFloat page on the Wiki if appropriate. Cheers, --=20 Stuart Brady