On Thu, Jun 10, 2010 at 5:06 PM, Venkateswararao Jujjuri (JV) <jvrao@xxxxxxxxxxxxxxxxxx> wrote: > Create a symbolic link ... > + > +static int > +v9fs_vfs_symlink_dotl(struct inode *dir, struct dentry *dentry, > + const char *symname) .... > + v9ses = v9fs_inode2v9ses(dir); > + > + dfid = v9fs_fid_lookup(dentry->d_parent); > + if (IS_ERR(dfid)) { > + err = PTR_ERR(dfid); > + P9_DPRINTK(P9_DEBUG_VFS, "fid lookup failed %d\n", err); > + return err; > + } On Mon, Jun 7, 2010 at 7:08 PM, Linus Torvalds <torvalds@xxxxxxxxxxxxxxxxxxxx> wrote: >>And the only thing that makes it ok to access dentry->d_parent - the fact >>that we are in v9fs_create() - is also the thing that should have made >>people look at the arguments to the function and say "hmm". >>We pass in the directory inode pointer as an argument to the create >>function! The code could have used that thing directly, instead of >>mucking around with dentry pointers that it had no business looking at. Please fix this in all the .L mk* patches prior to submission. We will need to fix v9fs_fid_lookup eventually, but we don't need to introduce new (unnecessary) dentry->d_parent references while we figure out how to fix v9fs_fid_lookup. I'll also accept patches fixing the my older code (like v9fs_create) if someone beats me to it. Thanks, -eric -- To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html