Date:   Tue,  9 May 2017 11:49:28 -0400
From:   Jeff Layton <jlayton@...hat.com>
To:     linux-fsdevel@...r.kernel.org, linux-kernel@...r.kernel.org,
        linux-btrfs@...r.kernel.org, linux-ext4@...r.kernel.org,
        linux-cifs@...r.kernel.org, linux-nfs@...r.kernel.org,
        linux-mm@...ck.org, jfs-discussion@...ts.sourceforge.net,
        linux-xfs@...r.kernel.org, cluster-devel@...hat.com,
        linux-f2fs-devel@...ts.sourceforge.net,
        v9fs-developer@...ts.sourceforge.net, linux-nilfs@...r.kernel.org,
        linux-block@...r.kernel.org
Cc:     dhowells@...hat.com, akpm@...ux-foundation.org, hch@...radead.org,
        ross.zwisler@...ux.intel.com, mawilcox@...rosoft.com,
        jack@...e.com, viro@...iv.linux.org.uk, corbet@....net,
        neilb@...e.de, clm@...com, tytso@....edu, axboe@...nel.dk,
        josef@...icpanda.com, hubcap@...ibond.com, rpeterso@...hat.com,
        bo.li.liu@...cle.com
Subject: [PATCH v4 25/27] Documentation: flesh out the section in vfs.txt on storing and reporting writeback errors

I waxed a little loquacious here, but I figured that more detail was
better, and writeback error handling is so hard to get right.

Cc: Jan Kara <jack@...e.cz>
Signed-off-by: Jeff Layton <jlayton@...hat.com>
---
 Documentation/filesystems/vfs.txt | 54 ++++++++++++++++++++++++++++++++-------
 1 file changed, 45 insertions(+), 9 deletions(-)

diff --git a/Documentation/filesystems/vfs.txt b/Documentation/filesystems/vfs.txt
index f201a77873f7..382190a872e5 100644
--- a/Documentation/filesystems/vfs.txt
+++ b/Documentation/filesystems/vfs.txt
@@ -576,12 +576,46 @@ should clear PG_Dirty and set PG_Writeback.  It can be actually
 written at any point after PG_Dirty is clear.  Once it is known to be
 safe, PG_Writeback is cleared.
 
-If there is an error during writeback, then the address_space should be
-marked with an error (typically using mapping_set_error), in order to
-ensure that the error can later be reported to the application when an
-fsync is issued.
-
-Writeback makes use of a writeback_control structure...
+Writeback makes use of a writeback_control structure to direct the
+operations.  This gives the the writepage and writepages operations some
+information about the nature of and reason for the writeback request,
+and the constraints under which it is being done.  It is also used to
+return information back to the caller about the result of a writepage or
+writepages request.
+
+Handling errors during writeback
+--------------------------------
+Most applications that utilize the pagecache will periodically call
+fsync to ensure that data written has made it to the backing store.
+When there is an error during writeback, that error should be reported
+when fsync is called.  After an error has been reported to fsync,
+subsequent fsync calls on the same file descriptor should return 0,
+unless further writeback errors have occurred since the previous fsync.
+
+Ideally, the kernel would report it only on file descriptions on which
+writes were done that subsequently failed to be written back.  The
+generic pagecache infrastructure does not track the file descriptions
+that have dirtied each individual page however, so determining which
+file descriptors should get back an error is not possible.
+
+Instead, the generic writeback error tracking infrastructure in the
+kernel settles for reporting errors to fsync on all file descriptions
+that were open at the time that the error occurred.  In a situation with
+multiple writers, all of them will get back an error on a subsequent fsync,
+even if all of the writes done through that particular file descriptor
+succeeded (or even if there were no writes on that file descriptor at all).
+
+Filesystems that wish to use this infrastructure should call
+mapping_set_error to record the error in the address_space when it
+occurs.  The generic vfs code will then handle reporting the error when
+fsync is called, even if the fsync file operation returned 0.
+
+Filesystems are free to track errors internally if they choose (i.e. if
+they do keep track of how the pages were dirtied), but they should aim
+to provide the same (or better) error reporting semantics for when there
+are multiple writers.  Those filesystems should avoid calling
+mapping_set_error in order to ensure that errors stored in the mapping
+aren't improperly reported by the generic filesystem code.
 
 struct address_space_operations
 -------------------------------
@@ -810,7 +844,8 @@ struct address_space_operations {
 The File Object
 ===============
 
-A file object represents a file opened by a process.
+A file object represents a file opened by a process. This is also known
+as an "open file description" in POSIX parlance.
 
 
 struct file_operations
@@ -893,9 +928,10 @@ otherwise noted.
 
   release: called when the last reference to an open file is closed
 
-  fsync: called by the fsync(2) system call. Errors that were previously
+  fsync: called by the fsync(2) system call.  Errors that were previously
 	 recorded using mapping_set_error will automatically be returned to
-	 the application and the file's error sequence advanced.
+	 the application and the struct file's error sequence advanced.
+	 See the section above on handling writeback errors.
 
   fasync: called by the fcntl(2) system call when asynchronous
 	(non-blocking) mode is enabled for a file
-- 
2.9.3

Powered by blists - more mailing lists