Hi Darrick, [...] > I'm confused and going to step through through this code one piece at a time. > > /* > * It won't fit in the shortform, transform to a leaf block. > * GROT: another possible req'mt for a double-split btree op. > */ > xfs_defer_init(args.dfops, args.firstblock); > error = xfs_attr_shortform_to_leaf(&args, &leaf_bp); > > /* hold the leaf buffer locked, when "args.trans" transaction commits */ > if (leaf_bp) > xfs_trans_bhold(args.trans, leaf_bp); > > If we hit an error in xfs_attr_shortform_to_leaf leaf_bp will have been > bjoined to args.trans by xfs_attr_shortform_to_leaf -> > xfs_attr3_leaf_create -> xfs_da_get_buf -> xfs_trans_get_buf_map, so > when we goto out in the error handling clause below, we'll trans_cancel, > which releases leaf_bp, right? And we can't just xfs_buf_relse because > that would unlock a buffer that's still owned by args.trans (like Brian > said)... If xfs_attr_shortform_to_leaf() hits an error, it will not set the caller's leaf_bp pointer. So this pointer will remain NULL, and therefore all the error handing logic at the caller will not touch this xfs_buf. I believe this addresses your concern. Thanks, Alex. -- To unsubscribe from this list: send the line "unsubscribe linux-xfs" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
