Date: Mon, 30 Mar 2015 13:33:55 +0300 From: Konstantin Belousov <kostikbel@gmail.com> To: Bruce Evans <brde@optusnet.com.au> Cc: src-committers@freebsd.org, Jilles Tjoelker <jilles@stack.nl>, Don Lewis <truckman@freebsd.org>, delphij@freebsd.org, svn-src-head@freebsd.org, svn-src-all@freebsd.org Subject: Re: svn commit: r280308 - head/sys/fs/devfs Message-ID: <20150330103355.GD2379@kib.kiev.ua> In-Reply-To: <20150330145148.C1660@besplex.bde.org> References: <20150322162507.GD2379@kib.kiev.ua> <201503221825.t2MIP7jv096531@gw.catspoiler.org> <20150329175137.GD95224@stack.nl> <20150329184238.GB2379@kib.kiev.ua> <20150330145148.C1660@besplex.bde.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On Mon, Mar 30, 2015 at 03:14:10PM +1100, Bruce Evans wrote:
> File timestamps use CLOCK_REALTIME, so they are supposed to go backwards
> sometimes. More importantly, if the time is set to a future time (by
> utimes(), etc., not due to a clock step), this prevents it being correted.
> I think you only want to do a null update if tv_nsec is nonzero due to a
> previous setting with vfs_timestamp(), and the new second hasn't arrived
> yet. Something like:
>
> if (tsp->tv_sec != ts) ...
>
> If '<', then as above. If '>', it means a correction by >= 1 second
> (strictly greater if tv_nsec != 0). The initial value tv_nsec is
> irrelevant in both cases.
Below is the updated patch. I also added a comment to devfs_timestamp(),
please reformulate it to your liking.
diff --git a/sys/fs/devfs/devfs_vnops.c b/sys/fs/devfs/devfs_vnops.c
index 941f92c..552ddd2 100644
--- a/sys/fs/devfs/devfs_vnops.c
+++ b/sys/fs/devfs/devfs_vnops.c
@@ -84,7 +84,27 @@ SYSCTL_DECL(_vfs_devfs);
static int devfs_dotimes;
SYSCTL_INT(_vfs_devfs, OID_AUTO, dotimes, CTLFLAG_RW,
- &devfs_dotimes, 0, "Update timestamps on DEVFS");
+ &devfs_dotimes, 0, "Update timestamps on DEVFS with default precision");
+
+/*
+ * Update devfs node timestamp. Note that updates are unlocked and
+ * stat(2) could see partially updated times.
+ */
+static void
+devfs_timestamp(struct timespec *tsp)
+{
+ time_t ts;
+
+ if (devfs_dotimes) {
+ vfs_timestamp(tsp);
+ } else {
+ ts = time_second;
+ if (tsp->tv_sec != ts) {
+ tsp->tv_sec = ts;
+ tsp->tv_nsec = 0;
+ }
+ }
+}
static int
devfs_fp_check(struct file *fp, struct cdev **devp, struct cdevsw **dswp,
@@ -1228,9 +1248,8 @@ devfs_read_f(struct file *fp, struct uio *uio, struct ucred *cred,
foffset_lock_uio(fp, uio, flags | FOF_NOLOCK);
error = dsw->d_read(dev, uio, ioflag);
- if (devfs_dotimes &&
- (uio->uio_resid != resid || (error == 0 && resid != 0)))
- vfs_timestamp(&dev->si_atime);
+ if (uio->uio_resid != resid || (error == 0 && resid != 0))
+ devfs_timestamp(&dev->si_atime);
td->td_fpop = fpop;
dev_relthread(dev, ref);
@@ -1708,9 +1727,8 @@ devfs_write_f(struct file *fp, struct uio *uio, struct ucred *cred,
resid = uio->uio_resid;
error = dsw->d_write(dev, uio, ioflag);
- if (devfs_dotimes &&
- (uio->uio_resid != resid || (error == 0 && resid != 0))) {
- vfs_timestamp(&dev->si_ctime);
+ if (uio->uio_resid != resid || (error == 0 && resid != 0)) {
+ devfs_timestamp(&dev->si_ctime);
dev->si_mtime = dev->si_ctime;
}
td->td_fpop = fpop;
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20150330103355.GD2379>