[JFFS2] Correct handling of JFFS2_FEATURE_RWCOMPAT_COPY nodes.

We should preserve these when we come to garbage collect them, not let
them get erased. Use jffs2_garbage_collect_pristine() for this, and make
sure the summary code copes -- just refrain from writing a summary for any
block which contains a node we don't understand.
Signed-off-by: David Woodhouse <dwmw2@infradead.org>

fs/jffs2/gc.c

@@ -256,10 +256,14 @@ int jffs2_garbage_collect_pass(struct jffs2_sb_info *c)
 
 	if (!raw->next_in_ino) {
 		/* Inode-less node. Clean marker, snapshot or something like that */
-		/* FIXME: If it's something that needs to be copied, including something
-		   we don't grok that has JFFS2_NODETYPE_RWCOMPAT_COPY, we should do so */
 		spin_unlock(&c->erase_completion_lock);
-		jffs2_mark_node_obsolete(c, raw);
+		if (ref_flags(raw) == REF_PRISTINE) {
+			/* It's an unknown node with JFFS2_FEATURE_RWCOMPAT_COPY */
+			jffs2_garbage_collect_pristine(c, NULL, raw);
+		} else {
+			/* Just mark it obsolete */
+			jffs2_mark_node_obsolete(c, raw);
+		}
 		up(&c->alloc_sem);
 		goto eraseit_lock;
 	}
@@ -533,15 +537,16 @@ static int jffs2_garbage_collect_pristine(struct jffs2_sb_info *c,
 
 	D1(printk(KERN_DEBUG "Going to GC REF_PRISTINE node at 0x%08x\n", ref_offset(raw)));
 
-	rawlen = ref_totlen(c, c->gcblock, raw);
+	alloclen = rawlen = ref_totlen(c, c->gcblock, raw);
 
 	/* Ask for a small amount of space (or the totlen if smaller) because we
 	   don't want to force wastage of the end of a block if splitting would
 	   work. */
-	ret = jffs2_reserve_space_gc(c, min_t(uint32_t, sizeof(struct jffs2_raw_inode) +
-				JFFS2_MIN_DATA_LEN, rawlen), &phys_ofs, &alloclen, rawlen);
-				/* this is not the exact summary size of it,
-					it is only an upper estimation */
+	if (ic && alloclen > sizeof(struct jffs2_raw_inode) + JFFS2_MIN_DATA_LEN)
+		alloclen = sizeof(struct jffs2_raw_inode) + JFFS2_MIN_DATA_LEN;
+
+	ret = jffs2_reserve_space_gc(c, alloclen, &phys_ofs, &alloclen, rawlen);
+	/* 'rawlen' is not the exact summary size; it is only an upper estimation */
 
 	if (ret)
 		return ret;
@@ -605,9 +610,12 @@ static int jffs2_garbage_collect_pristine(struct jffs2_sb_info *c,
 		}
 		break;
 	default:
-		printk(KERN_WARNING "Unknown node type for REF_PRISTINE node at 0x%08x: 0x%04x\n",
-		       ref_offset(raw), je16_to_cpu(node->u.nodetype));
-		goto bail;
+		/* If it's inode-less, we don't _know_ what it is. Just copy it intact */
+		if (ic) {
+			printk(KERN_WARNING "Unknown node type for REF_PRISTINE node at 0x%08x: 0x%04x\n",
+			       ref_offset(raw), je16_to_cpu(node->u.nodetype));
+			goto bail;
+		}
 	}
 
 	nraw = jffs2_alloc_raw_node_ref();
@@ -674,15 +682,16 @@ static int jffs2_garbage_collect_pristine(struct jffs2_sb_info *c,
 	nraw->flash_offset |= REF_PRISTINE;
 	jffs2_add_physical_node_ref(c, nraw);
 
-	/* Link into per-inode list. This is safe because of the ic
-	   state being INO_STATE_GC. Note that if we're doing this
-	   for an inode which is in-core, the 'nraw' pointer is then
-	   going to be fetched from ic->nodes by our caller. */
-	spin_lock(&c->erase_completion_lock);
-        nraw->next_in_ino = ic->nodes;
-        ic->nodes = nraw;
-	spin_unlock(&c->erase_completion_lock);
-
+	if (ic) {
+		/* Link into per-inode list. This is safe because of the ic
+		   state being INO_STATE_GC. Note that if we're doing this
+		   for an inode which is in-core, the 'nraw' pointer is then
+		   going to be fetched from ic->nodes by our caller. */
+		spin_lock(&c->erase_completion_lock);
+		nraw->next_in_ino = ic->nodes;
+		ic->nodes = nraw;
+		spin_unlock(&c->erase_completion_lock);
+	}
 	jffs2_mark_node_obsolete(c, raw);
 	D1(printk(KERN_DEBUG "WHEEE! GC REF_PRISTINE node at 0x%08x succeeded\n", ref_offset(raw)));
 

fs/jffs2/nodelist.h

@@ -97,6 +97,11 @@ struct jffs2_raw_node_ref
 #define ref_obsolete(ref)	(((ref)->flash_offset & 3) == REF_OBSOLETE)
 #define mark_ref_normal(ref)    do { (ref)->flash_offset = ref_offset(ref) | REF_NORMAL; } while(0)
 
+/* NB: REF_PRISTINE for an inode-less node (ref->next_in_ino == NULL) indicates
+   it is an unknown node of type JFFS2_NODETYPE_RWCOMPAT_COPY, so it'll get
+   copied. If you need to do anything different to GC inode-less nodes, then
+   you need to modify gc.c accordingly. */
+
 /* For each inode in the filesystem, we need to keep a record of
    nlink, because it would be a PITA to scan the whole directory tree
    at read_inode() time to calculate it, and to keep sufficient information

fs/jffs2/scan.c

@@ -851,11 +851,22 @@ static int jffs2_scan_eraseblock (struct jffs2_sb_info *c, struct jffs2_eraseblo
 				ofs += PAD(je32_to_cpu(node->totlen));
 				break;
 
-			case JFFS2_FEATURE_RWCOMPAT_COPY:
+			case JFFS2_FEATURE_RWCOMPAT_COPY: {
+				struct jffs2_raw_node_ref *ref;
 				D1(printk(KERN_NOTICE "Unknown but compatible feature node (0x%04x) found at offset 0x%08x\n", je16_to_cpu(node->nodetype), ofs));
-				USED_SPACE(PAD(je32_to_cpu(node->totlen)));
+
+				ref = jffs2_alloc_raw_node_ref();
+				if (!ref)
+					return -ENOMEM;
+				ref->flash_offset = ofs | REF_PRISTINE;
+				ref->next_in_ino = 0;
+				jffs2_link_node_ref(c, jeb, ref, PAD(je32_to_cpu(node->totlen)));
+
+				/* We can't summarise nodes we don't grok */
+				jffs2_sum_disable_collecting(s);
 				ofs += PAD(je32_to_cpu(node->totlen));
 				break;
+				}
 			}
 		}
 	}

fs/jffs2/summary.c

@@ -760,7 +760,14 @@ static int jffs2_sum_write_data(struct jffs2_sb_info *c, struct jffs2_eraseblock
 			}
 #endif
 			default : {
-				BUG();	/* unknown node in summary information */
+				if ((je16_to_cpu(temp->u.nodetype) & JFFS2_COMPAT_MASK)
+				    == JFFS2_FEATURE_RWCOMPAT_COPY) {
+					dbg_summary("Writing unknown RWCOMPAT_COPY node type %x\n",
+						    je16_to_cpu(temp->u.nodetype));
+					jffs2_sum_disable_collecting(c->summary);
+				} else {
+					BUG();	/* unknown node in summary information */
+				}
 			}
 		}