On Fri, Apr 2, 2010 at 4:20 PM, Martin Bligh <[email protected]> wrote:
> On Fri, Apr 2, 2010 at 4:09 PM, Akshay Lal <[email protected]> wrote: > > I tested it with etx4 and ext3 and ext2. > > The only change I really made was to force the fsck (-f) and instead of > > doing a "yes|fsck_cmd", I'm passing the '-y' flag thereby forcing the yes > to > > any question. > > We needed to set the -f flag since there were scenarios wherein the > > filesystem had a corrupted inode table (eg: using dbugfs) and since the > > "check_filesystem" flag wasn't set, the fsck wouldn't run. > > Hmm, OK, that's not good. > > > But if you can think of any other file system I should run this against, > > I'll ensure that I check it. > > IIRC, Reiser, jfs and xfs had incompatible option flags. btrfs would be > another one. > > Yuck. but I guess thats what happens since the fscks are individual programs written by people who don't care to maintain a common API. Even the yes|fsck is not guaranteed to work on everything in that case. When this is actually an issue for anyone, we should probably have the fsck code look at what the filesystem type is and explicitly call the correct fs specific fsck with the appropriate options. -gps
_______________________________________________ Autotest mailing list [email protected] http://test.kernel.org/cgi-bin/mailman/listinfo/autotest
