git.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Michael Haggerty <mhagger@alum.mit.edu>
To: Junio C Hamano <gitster@pobox.com>
Cc: git@vger.kernel.org, Michael Haggerty <mhagger@alum.mit.edu>
Subject: [PATCH v3 09/17] safe_create_leading_directories(): add new error value SCLD_VANISHED
Date: Sat, 18 Jan 2014 23:48:53 +0100	[thread overview]
Message-ID: <1390085341-2553-10-git-send-email-mhagger@alum.mit.edu> (raw)
In-Reply-To: <1390085341-2553-1-git-send-email-mhagger@alum.mit.edu>

Add a new possible error result that can be returned by
safe_create_leading_directories() and
safe_create_leading_directories_const(): SCLD_VANISHED.  This value
indicates that a file or directory on the path existed at one point
(either it already existed or the function created it), but then it
disappeared.  This probably indicates that another process deleted the
directory while we were working.  If SCLD_VANISHED is returned, the
caller might want to retry the function call, as there is a chance
that a new attempt will succeed.

Why doesn't safe_create_leading_directories() do the retrying
internally?  Because an empty directory isn't really ever safe until
it holds a file.  So even if safe_create_leading_directories() were
absolutely sure that the directory existed before it returned, there
would be no guarantee that the directory still existed when the caller
tried to write something in it.

Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu>
---
 cache.h     | 10 +++++++++-
 sha1_file.c | 11 +++++++++++
 2 files changed, 20 insertions(+), 1 deletion(-)

diff --git a/cache.h b/cache.h
index 8030e36..c0a7a8a 100644
--- a/cache.h
+++ b/cache.h
@@ -742,12 +742,20 @@ int adjust_shared_perm(const char *path);
  * Create the directory containing the named path, using care to be
  * somewhat safe against races.  Return one of the scld_error values
  * to indicate success/failure.
+ *
+ * SCLD_VANISHED indicates that one of the ancestor directories of the
+ * path existed at one point during the function call and then
+ * suddenly vanished, probably because another process pruned the
+ * directory while we were working.  To be robust against this kind of
+ * race, callers might want to try invoking the function again when it
+ * returns SCLD_VANISHED.
  */
 enum scld_error {
 	SCLD_OK = 0,
 	SCLD_FAILED = -1,
 	SCLD_PERMS = -2,
-	SCLD_EXISTS = -3
+	SCLD_EXISTS = -3,
+	SCLD_VANISHED = -4
 };
 enum scld_error safe_create_leading_directories(char *path);
 enum scld_error safe_create_leading_directories_const(const char *path);
diff --git a/sha1_file.c b/sha1_file.c
index a065308..8b0849f 100644
--- a/sha1_file.c
+++ b/sha1_file.c
@@ -132,6 +132,17 @@ enum scld_error safe_create_leading_directories(char *path)
 			if (errno == EEXIST &&
 			    !stat(path, &st) && S_ISDIR(st.st_mode))
 				; /* somebody created it since we checked */
+			else if (errno == ENOENT)
+				/*
+				 * Either mkdir() failed because
+				 * somebody just pruned the containing
+				 * directory, or stat() failed because
+				 * the file that was in our way was
+				 * just removed.  Either way, inform
+				 * the caller that it might be worth
+				 * trying again:
+				 */
+				ret = SCLD_VANISHED;
 			else
 				ret = SCLD_FAILED;
 		} else if (adjust_shared_perm(path)) {
-- 
1.8.5.2

  parent reply	other threads:[~2014-01-18 22:56 UTC|newest]

Thread overview: 18+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2014-01-18 22:48 [PATCH v3 00/17] Fix some mkdir/rmdir races Michael Haggerty
2014-01-18 22:48 ` [PATCH v3 01/17] safe_create_leading_directories(): fix format of "if" chaining Michael Haggerty
2014-01-18 22:48 ` [PATCH v3 02/17] safe_create_leading_directories(): reduce scope of local variable Michael Haggerty
2014-01-18 22:48 ` [PATCH v3 03/17] safe_create_leading_directories(): add explicit "slash" pointer Michael Haggerty
2014-01-18 22:48 ` [PATCH v3 04/17] safe_create_leading_directories(): rename local variable Michael Haggerty
2014-01-18 22:48 ` [PATCH v3 05/17] safe_create_leading_directories(): split on first of multiple slashes Michael Haggerty
2014-01-18 22:48 ` [PATCH v3 06/17] safe_create_leading_directories(): always restore slash at end of loop Michael Haggerty
2014-01-18 22:48 ` [PATCH v3 07/17] safe_create_leading_directories(): introduce enum for return values Michael Haggerty
2014-01-18 22:48 ` [PATCH v3 08/17] cmd_init_db(): when creating directories, handle errors conservatively Michael Haggerty
2014-01-18 22:48 ` Michael Haggerty [this message]
2014-01-18 22:48 ` [PATCH v3 10/17] lock_ref_sha1_basic(): on SCLD_VANISHED, retry Michael Haggerty
2014-01-18 22:48 ` [PATCH v3 11/17] lock_ref_sha1_basic(): if locking fails with ENOENT, retry Michael Haggerty
2014-01-18 22:48 ` [PATCH v3 12/17] remove_dir_recurse(): tighten condition for removing unreadable dir Michael Haggerty
2014-01-18 22:48 ` [PATCH v3 13/17] remove_dir_recurse(): handle disappearing files and directories Michael Haggerty
2014-01-18 22:48 ` [PATCH v3 14/17] rename_ref(): extract function rename_tmp_log() Michael Haggerty
2014-01-18 22:48 ` [PATCH v3 15/17] rename_tmp_log(): handle a possible mkdir/rmdir race Michael Haggerty
2014-01-18 22:49 ` [PATCH v3 16/17] rename_tmp_log(): limit the number of remote_empty_directories() attempts Michael Haggerty
2014-01-18 22:49 ` [PATCH v3 17/17] rename_tmp_log(): on SCLD_VANISHED, retry Michael Haggerty

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=1390085341-2553-10-git-send-email-mhagger@alum.mit.edu \
    --to=mhagger@alum.mit.edu \
    --cc=git@vger.kernel.org \
    --cc=gitster@pobox.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).