linux-kernel - [PATCH] drivers/staging/exfat - explain the fs

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [thread-next>] [day] [month] [year] [list]

Message-ID: <9837.1570042895@turing-police>
Date:   Wed, 02 Oct 2019 15:01:35 -0400
From:   "Valdis Klētnieks" <valdis.kletnieks@...edu>
To:     gregkh@...uxfoundation.org
cc:     devel@...verdev.osuosl.org, linux-kernel@...r.kernel.org
Subject: [PATCH] drivers/staging/exfat - explain the fs_sync() issue in TODO

We've seen several incorrect patches for fs_sync() calls in the exfat driver.
Add code to the TODO that explains this isn't just a delete code and refactor,
but that actual analysis of when the filesystem should be flushed to disk
needs to be done.

Signed-off-by: Valdis Kletnieks <valdis.kletnieks@...edu>

---
diff --git a/drivers/staging/exfat/TODO b/drivers/staging/exfat/TODO
index a3eb282f9efc..77c302acfcb8 100644
--- a/drivers/staging/exfat/TODO
+++ b/drivers/staging/exfat/TODO
@@ -3,6 +3,15 @@ same for ffsWriteFile.
 
 exfat_core.c - fs_sync(sb,0) all over the place looks fishy as hell.
 There's only one place that calls it with a non-zero argument.
+Randomly removing fs_sync() calls is *not* the right answer, especially
+if the removal then leaves a call to fs_set_vol_flags(VOL_CLEAN), as that
+says the file system is clean and synced when we *know* it isn't.
+The proper fix here is to go through and actually analyze how DELAYED_SYNC
+should work, and any time we're setting VOL_CLEAN, ensure the file system
+has in fact been synced to disk.  In other words, changing the 'false' to
+'true' is probably more correct. Also, it's likely that the one current
+place where it actually does an bdev_sync isn't sufficient in the DELAYED_SYNC
+case.
 
 ffsTruncateFile -  if (old_size <= new_size) {
 That doesn't look right. How did it ever work? Are they relying on lazy