Hello Mark and All,
The implementation code looks a little tricky, but It looks to have to hack like that way in a online file system environment. I also want to find a more graceful/concise way to implement this feature.
Please help to take some time in thinking about this problem, find a compromised way, and make the thing move forward.
Thanks a lot.
Gang
>>>
> Hello Mark,
>
> Thanks for your reviewing, please see my comments inline.
>
>
> >>>
> > On Wed, Aug 26, 2015 at 03:12:22PM -0700, Andrew Morton wrote:
> >> From: Gang He <ghe at suse.com>
> >> Subject: ocfs2: check/fix inode block for online file check
> >>
> >> Implement online check or fix inode block during reading a inode block to
> >> memory.
> >>
> >> Signed-off-by: Gang He <ghe at suse.com>
> >> Reviewed-by: Goldwyn Rodrigues <rgoldwyn at suse.de>
> >> Cc: Mark Fasheh <mfasheh at suse.com>
> >> Cc: Joel Becker <jlbec at evilplan.org>
> >> Signed-off-by: Andrew Morton <akpm at linux-foundation.org>
> >> ---
> >>
> >> fs/ocfs2/inode.c | 196 +++++++++++++++++++++++++++++++++++++--
> >> fs/ocfs2/ocfs2_trace.h | 2
> >> 2 files changed, 192 insertions(+), 6 deletions(-)
> >>
> >> diff -puN fs/ocfs2/inode.c~ocfs2-check-fix-inode-block-for-online-file-check
> > fs/ocfs2/inode.c
> >> --- a/fs/ocfs2/inode.c~ocfs2-check-fix-inode-block-for-online-file-check
> >> +++ a/fs/ocfs2/inode.c
> >> @@ -53,6 +53,7 @@
> >> #include "xattr.h"
> >> #include "refcounttree.h"
> >> #include "ocfs2_trace.h"
> >> +#include "filecheck.h"
> >>
> >> #include "buffer_head_io.h"
> >>
> >> @@ -74,6 +75,13 @@ static int ocfs2_truncate_for_delete(str
> >> struct inode *inode,
> >> struct buffer_head *fe_bh);
> >>
> >> +static int ocfs2_filecheck_read_inode_block_full(struct inode *inode,
> >> + struct buffer_head **bh, int flags, int type);
> >> +static int ocfs2_filecheck_validate_inode_block(struct super_block *sb,
> >> + struct buffer_head *bh);
> >> +static int ocfs2_filecheck_repair_inode_block(struct super_block *sb,
> >> + struct buffer_head *bh);
> >> +
> >> void ocfs2_set_inode_flags(struct inode *inode)
> >> {
> >> unsigned int flags = OCFS2_I(inode)->ip_attr;
> >> @@ -127,6 +135,7 @@ struct inode *ocfs2_ilookup(struct super
> >> struct inode *ocfs2_iget(struct ocfs2_super *osb, u64 blkno, unsigned
> > flags,
> >> int sysfile_type)
> >> {
> >> + int rc = 0;
> >> struct inode *inode = NULL;
> >> struct super_block *sb = osb->sb;
> >> struct ocfs2_find_inode_args args;
> >> @@ -161,12 +170,17 @@ struct inode *ocfs2_iget(struct ocfs2_su
> >> }
> >> trace_ocfs2_iget5_locked(inode->i_state);
> >> if (inode->i_state & I_NEW) {
> >> - ocfs2_read_locked_inode(inode, &args);
> >> + rc = ocfs2_read_locked_inode(inode, &args);
> >> unlock_new_inode(inode);
> >> }
> >> if (is_bad_inode(inode)) {
> >> iput(inode);
> >> - inode = ERR_PTR(-ESTALE);
> >> + if ((flags & OCFS2_FI_FLAG_FILECHECK_CHK) ||
> >> + (flags & OCFS2_FI_FLAG_FILECHECK_FIX))
> >> + /* Return OCFS2_FILECHECK_ERR_XXX related errno */
> >> + inode = ERR_PTR(rc);
> >> + else
> >> + inode = ERR_PTR(-ESTALE);
> >> goto bail;
> >> }
> >>
> >> @@ -494,16 +508,32 @@ static int ocfs2_read_locked_inode(struc
> >> }
> >>
> >> if (can_lock) {
> >> - status = ocfs2_read_inode_block_full(inode, &bh,
> >> - OCFS2_BH_IGNORE_CACHE);
> >> + if (args->fi_flags & OCFS2_FI_FLAG_FILECHECK_CHK)
> >> + status = ocfs2_filecheck_read_inode_block_full(inode,
> >> + &bh, OCFS2_BH_IGNORE_CACHE, 0);
> >> + else if (args->fi_flags & OCFS2_FI_FLAG_FILECHECK_FIX)
> >> + status = ocfs2_filecheck_read_inode_block_full(inode,
> >> + &bh, OCFS2_BH_IGNORE_CACHE, 1);
> >> + else
> >> + status = ocfs2_read_inode_block_full(inode,
> >> + &bh, OCFS2_BH_IGNORE_CACHE);
> >
> > NAK, at first glance this is very hacky - I don't like that we've hidden
> > these checks
> > and fixups down in iget(). If there's a reason it has to be this way that
> > should be explained, but otherwise I would expect the check/repair code to
> > be less intertwined with ocfs2_iget(). Otherwise I fear we're setting
> > ourselves up
> > for finding some ugly bugs down the road.
> Firstly, I want to read inode block separately, but the feature is working
> with a online file system,
> we can't avoid the inodes cache and the cluster environment, I think that I
> should not handle a inode
> block separately without considering the current inodes cache and potential
> cluster access from other node.
> Then I tried to integrate this light-level online check/fix with iget()
> function, make the online check/fix
> operations is compatible with the inodes cache and the cluster environment.
> This is why I use iget() to integrate this feature. if there is a more
> graceful way to implement this feature,
> please help to give some suggestions.
>
> >
> > Btw, how does the code handle the case where the inode is already in our
> > cache? In that case you'd never get to read_locked_inode()...
> This online file check/fix is light-level meta-data block check/fix, the
> check/fix fields usually correspond
> with ocfs2_validate_inode_block(), for more serious problem, we have to use
> fsck by offline.
> If the inode is already in our cache, that means this inode block passed
> ocfs2_validate_inode_block()
> verification during loading inode block from the disk, this inode block is
> sane, we need not check it again.
>
> Thanks
> Gang
>
>
> >
> > Thanks,
> > --Mark
> >
> > --
> > Mark Fasheh