Date: Tue, 14 Oct 2003 10:39:03 +0200 (CEST) From: Harti Brandt <brandt@fokus.fraunhofer.de> To: Bruce Evans <bde@zeta.org.au> Cc: sparc64@freebsd.org Subject: Re: time_t on sparc64 Message-ID: <20031014103446.U45269@beagle.fokus.fraunhofer.de> In-Reply-To: <20031014035805.F32262@gamplex.bde.org> References: <20031013153219.H45269@beagle.fokus.fraunhofer.de> <20031014035805.F32262@gamplex.bde.org>
next in thread | previous in thread | raw e-mail | index | archive | help
Thanks for this info Bruce,
two questions however:
On Tue, 14 Oct 2003, Bruce Evans wrote:
BE>On Mon, 13 Oct 2003, Harti Brandt wrote:
BE>
BE>> I just discovered that time_t is 32-bit on sparc64. One of the problems
BE>> is that struct timeval is defined by Posix as
BE>>
BE>> struct timeval {
BE>> time_t tv_secs;
BE>> suseconds_t tv_usecs;
BE>> };
BE>
BE>This is a bug in POSIX. In BSD, tv_secs has type long which may be,
BE>and is different from time_t.
Why do you think this is a POSIX bug? Aren't we the odd man out? Now we
(and MacOS 10) require a workaround for thinks like
printf("%s", ctime(tv.tv_secs));
which works perfect on other systems.
BE>
BE>> but _timeval.h has
BE>>
BE>> struct timeval {
BE>> long tv_secs;
BE>> suseconds_t tv_usecs;
BE>> }
BE>>
BE>> This means, that our timeval is not Posix compatible. What is the reason
BE>> for time_t not beeing a long on sparc64?
BE>
BE>time_t was used in some data structures whose layout shouldn't be changed
BE>even for new arches. Mainly in ufs in Lite2:
BE>
BE>%%%
BE>ffs/fs.h: time_t fs_time; /* last time written */
BE>ffs/fs.h: time_t cg_time; /* time last written */
BE>ffs/fs.h: time_t cg_time; /* time last written */
BE>lfs/lfs.h: time_t bi_segcreate; /* origin segment create time */
BE>ufs/quota.h: time_t dqb_btime; /* time limit for excessive disk use */
BE>ufs/quota.h: time_t dqb_itime; /* time limit for excessive files */
BE>%%%
BE>
BE>These are now:
BE>
BE>%%%
BE>ffs/fs.h: int32_t fs_old_time; /* last time written */
BE>ffs/fs.h: ufs_time_t fs_time; /* last time written */
BE>ffs/fs.h: int32_t cg_old_time; /* time last written */
BE>ffs/fs.h: ufs_time_t cg_time; /* time last written */
BE>/dev/null: time_t bi_segcreate; /* origin segment create time */
BE>ufs/quota.h: int32_t dqb_btime; /* time limit for excessive disk use */
BE>ufs/quota.h: int32_t dqb_itime; /* time limit for excessive files */
BE>%%%
BE>
BE>I.e., int32_t is now not mispelled time_t in f^Hufs1 and Y2.038K bugs are
BE>fixed in ffs2 except for quotas.
BE>
BE>ffs2 also parametrizes timestamps in inodes better:
BE>
BE>%%%
BE>ufs/dinode.h:typedef int64_t ufs_time_t;
BE>ufs/dinode.h: ufs_time_t di_atime; /* 32: Last access time. */
BE>ufs/dinode.h: ufs_time_t di_mtime; /* 40: Last modified time. */
BE>ufs/dinode.h: ufs_time_t di_ctime; /* 48: Last inode change time. */
BE>ufs/dinode.h: ufs_time_t di_birthtime; /* 56: Inode creation time. */
BE>ufs/dinode.h: int32_t di_mtimensec; /* 64: Last modified time. */
BE>ufs/dinode.h: int32_t di_atimensec; /* 68: Last access time. */
BE>ufs/dinode.h: int32_t di_ctimensec; /* 72: Last inode change time. */
BE>ufs/dinode.h: int32_t di_birthnsec; /* 76: Inode creation time. */
BE>
BE>[Note that these aren't in a timespec struct, POSIX or otherwise, since the
BE>struct would give MD packing which happens to be inefficient in most cases.]
BE>
BE>ufs/dinode.h: int32_t di_atime; /* 16: Last access time. */
BE>ufs/dinode.h: int32_t di_atimensec; /* 20: Last access time. */
BE>ufs/dinode.h: int32_t di_mtime; /* 24: Last modified time. */
BE>ufs/dinode.h: int32_t di_mtimensec; /* 28: Last modified time. */
BE>ufs/dinode.h: int32_t di_ctime; /* 32: Last inode change time. */
BE>ufs/dinode.h: int32_t di_ctimensec; /* 36: Last inode change time. */
BE>
BE>[Y2.038K bugs are still in ffs1.]
BE>%%%
BE>
BE>To change time_t to 64 bits, all in-use non-transient data structures
BE>need to be changed similarly.
I guess we have to do this work before 2038, don't we? If we don't do it
before 5.2 we have to stick with this until 6.0. Correct?
harti
--
harti brandt,
http://www.fokus.fraunhofer.de/research/cc/cats/employees/hartmut.brandt/private
brandt@fokus.fraunhofer.de, harti@freebsd.org
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20031014103446.U45269>