Re: [PATCH v7 12/13] ext4: switch to multigrain timestamps

[Bruno Haible]

Jeff Layton wrote:
> > Surely this is a safe choice as it moves the responsibility to the sysadmin
> > and the cases where finegrained timestamps are required. But I kind of
> > wonder how is the sysadmin going to decide whether mgtime is safe for his
> > system or not? Because the possible breakage needn't be obvious at the
> > first sight...
> 
> That's the main reason I really didn't want to go with a mount option.
> Documenting that may be difficult.

You could document it like this:

  The mgtime option enables more precise modification times (mtime)
  on some files, together with an optimization that limits the amount
  of metadata changes.
  Note that this option may, in some cases, after writing to file F1
  and then writing to file F2, report a lower mtime for F2 than for F2.
  Enabling this option may be useful on file systems shared via NFS.
  The safe choice is to disable this option.

For me as a user, there's no need to go into more details than that.

It's important to have this mount option, for people who want maximum
reliability. Personally, I always enable the 'strictatime' option on
all ext4 mounts, since 'relatime' optimizes too much for my use-cases.
If I fear wrong results of "make" runs, I will definitely opt for the
safe choice regarding mgtime as well — since I don't want to spend
hours debugging binaries that were built incorrectly from correct
source code.

Bruno