Date: Mon, 4 Feb 2013 16:04:49 +1100 (EST) From: Bruce Evans <brde@optusnet.com.au> To: Pedro Giffuni <pfg@FreeBSD.org> Cc: "freebsd-fs@freebsd.org" <freebsd-fs@FreeBSD.org>, Andriy Gapon <avg@FreeBSD.org>, Kirk McKusick <mckusick@FreeBSD.org> Subject: Re: svn commit: r246289 - head/sys/ufs/ffs Message-ID: <20130204155554.I932@besplex.bde.org> In-Reply-To: <1359945154.62069.YahooMailNeo@web162105.mail.bf1.yahoo.com> References: <201302031716.r13HGXNP060303@svn.freebsd.org> <510E9D47.2030403@FreeBSD.org> <20130204062149.U2673@besplex.bde.org> <1359945154.62069.YahooMailNeo@web162105.mail.bf1.yahoo.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Sun, 3 Feb 2013, Pedro Giffuni wrote: > (Moving the discussion to freebsd-fs) >> Da: Bruce Evans >> > ... >>> Just a note that clang actually warned about this one. >>> It has a few more similar warnings for ufs/ffs code. >> >> I wondered how the DIP macro hid the warning. >> > > The comparison is perfectly legal for UFS1 so perhaps > gcc gives the "benefit of the doubt" to avoid false positives. > >> Perhaps the larger bugs pointed to this warning were lost in translation: >> - di_blocks overflows for ffs1. This is now physically possible. >> ... > Surely not anywhere near a complete solution but perhaps it wouldn't > be incompatible to change i_blocks and friends to be unsigned in UFS1. > That is something that remains to be completed in ext2fs, but according > to fsx there are bigger problems there at this time. That only gives 1 more bit, but 7 more are required (for the expansion factor MAXBSIZE / DEV_BSIZE = 2**7). More if someone increases MAXBSIZE. Bruce
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20130204155554.I932>