Re: [PATCH] sha1_file.c:create_tmpfile(): Fix race when creating loose object dirs

[Junio C Hamano <gitster@pobox.com>]

Jeff King <peff@peff.net> writes:

> I was almost tempted to say "we do not even need to run
> adjust_shared_perm twice", but we do, or we risk another race: one side
> loses the mkdir race, but wins the open() race, and writes to a
> wrong-permission directory. Of course, that should not matter unless the
> racers are two different users (in a shared repo), and in that case, we
> wins the adjust_shared_perm race, but does not have permission to change
> the mode.

Interesting.

> Agreed. We already leave a wrong-permission directory in place if it
> existed before we started create_tmpfile. The code before your patch,
> when racing with such a wrong-directory creator, would abort the
> tmpfile. Now it will correct the permissions. Either behavior seems fine
> to me (yours actually seems better, but the point is that it does not
> matter because they are dwarfed by the non-race cases where the
> directory is already sitting there).

Agreed.  We may notice the failure to correct the permissions in the
new code, where the old code left existing directories incorrect
permissions as they were.