v2.4.13.2 -> v2.4.13.3

  - René Scharfe: random bugfix
  - me: block device queuing low-water-marks, VM mapped tweaking.

Makefile

 VERSION = 2
 PATCHLEVEL = 4
 SUBLEVEL = 14
-EXTRAVERSION =-pre2
+EXTRAVERSION =-pre3
 
 KERNELRELEASE=$(VERSION).$(PATCHLEVEL).$(SUBLEVEL)$(EXTRAVERSION)
 

drivers/block/ll_rw_blk.c

@@ -140,21 +140,23 @@ inline request_queue_t *blk_get_queue(kdev_t dev)
 		return &blk_dev[MAJOR(dev)].request_queue;
 }
 
-static int __blk_cleanup_queue(struct list_head *head)
+static int __blk_cleanup_queue(struct request_list *list)
 {
+	struct list_head *head = &list->free;
 	struct request *rq;
 	int i = 0;
 
-	if (list_empty(head))
-		return 0;
-
-	do {
+	while (!list_empty(head)) {
 		rq = list_entry(head->next, struct request, queue);
 		list_del(&rq->queue);
 		kmem_cache_free(request_cachep, rq);
 		i++;
-	} while (!list_empty(head));
+	};
 
+	if (i != list->count)
+		printk("request list leak!\n");
+
+	list->count = 0;
 	return i;
 }
 
@@ -176,10 +178,8 @@ void blk_cleanup_queue(request_queue_t * q)
 {
 	int count = queue_nr_requests;
 
-	count -= __blk_cleanup_queue(&q->request_freelist[READ]);
-	count -= __blk_cleanup_queue(&q->request_freelist[WRITE]);
-	count -= __blk_cleanup_queue(&q->pending_freelist[READ]);
-	count -= __blk_cleanup_queue(&q->pending_freelist[WRITE]);
+	count -= __blk_cleanup_queue(&q->rq[READ]);
+	count -= __blk_cleanup_queue(&q->rq[WRITE]);
 
 	if (count)
 		printk("blk_cleanup_queue: leaked requests (%d)\n", count);
@@ -331,11 +331,10 @@ static void blk_init_free_list(request_queue_t *q)
 	struct request *rq;
 	int i;
 
-	INIT_LIST_HEAD(&q->request_freelist[READ]);
-	INIT_LIST_HEAD(&q->request_freelist[WRITE]);
-	INIT_LIST_HEAD(&q->pending_freelist[READ]);
-	INIT_LIST_HEAD(&q->pending_freelist[WRITE]);
-	q->pending_free[READ] = q->pending_free[WRITE] = 0;
+	INIT_LIST_HEAD(&q->rq[READ].free);
+	INIT_LIST_HEAD(&q->rq[WRITE].free);
+	q->rq[READ].count = 0;
+	q->rq[WRITE].count = 0;
 
 	/*
 	 * Divide requests in half between read and write
@@ -349,7 +348,8 @@ static void blk_init_free_list(request_queue_t *q)
 		}
 		memset(rq, 0, sizeof(struct request));
 		rq->rq_status = RQ_INACTIVE;
-		list_add(&rq->queue, &q->request_freelist[i & 1]);
+		list_add(&rq->queue, &q->rq[i&1].free);
+		q->rq[i&1].count++;
 	}
 
 	init_waitqueue_head(&q->wait_for_request);
@@ -423,10 +423,12 @@ void blk_init_queue(request_queue_t * q, request_fn_proc * rfn)
 static inline struct request *get_request(request_queue_t *q, int rw)
 {
 	struct request *rq = NULL;
+	struct request_list *rl = q->rq + rw;
 
-	if (!list_empty(&q->request_freelist[rw])) {
-		rq = blkdev_free_rq(&q->request_freelist[rw]);
+	if (!list_empty(&rl->free)) {
+		rq = blkdev_free_rq(&rl->free);
 		list_del(&rq->queue);
+		rl->count--;
 		rq->rq_status = RQ_ACTIVE;
 		rq->special = NULL;
 		rq->q = q;
@@ -443,17 +445,13 @@ static struct request *__get_request_wait(request_queue_t *q, int rw)
 	register struct request *rq;
 	DECLARE_WAITQUEUE(wait, current);
 
+	generic_unplug_device(q);
 	add_wait_queue_exclusive(&q->wait_for_request, &wait);
-	for (;;) {
-		__set_current_state(TASK_UNINTERRUPTIBLE);
-		spin_lock_irq(&io_request_lock);
-		rq = get_request(q, rw);
-		spin_unlock_irq(&io_request_lock);
-		if (rq)
-			break;
-		generic_unplug_device(q);
-		schedule();
-	}
+	do {
+		set_current_state(TASK_UNINTERRUPTIBLE);
+		if (q->rq[rw].count < batch_requests)
+			schedule();
+	} while ((rq = get_request(q,rw)) == NULL);
 	remove_wait_queue(&q->wait_for_request, &wait);
 	current->state = TASK_RUNNING;
 	return rq;
@@ -542,15 +540,6 @@ static inline void add_request(request_queue_t * q, struct request * req,
 	list_add(&req->queue, insert_here);
 }
 
-inline void blk_refill_freelist(request_queue_t *q, int rw)
-{
-	if (q->pending_free[rw]) {
-		list_splice(&q->pending_freelist[rw], &q->request_freelist[rw]);
-		INIT_LIST_HEAD(&q->pending_freelist[rw]);
-		q->pending_free[rw] = 0;
-	}
-}
-
 /*
  * Must be called with io_request_lock held and interrupts disabled
  */
@@ -564,28 +553,12 @@ inline void blkdev_release_request(struct request *req)
 
 	/*
 	 * Request may not have originated from ll_rw_blk. if not,
-	 * asumme it has free buffers and check waiters
+	 * assume it has free buffers and check waiters
 	 */
 	if (q) {
-		/*
-		 * If nobody is waiting for requests, don't bother
-		 * batching up.
-		 */
-		if (!list_empty(&q->request_freelist[rw])) {
-			list_add(&req->queue, &q->request_freelist[rw]);
-			return;
-		}
-
-		/*
-		 * Add to pending free list and batch wakeups
-		 */
-		list_add(&req->queue, &q->pending_freelist[rw]);
-
-		if (++q->pending_free[rw] >= batch_requests) {
-			int wake_up = q->pending_free[rw];
-			blk_refill_freelist(q, rw);
-			wake_up_nr(&q->wait_for_request, wake_up);
-		}
+		list_add(&req->queue, &q->rq[rw].free);
+		if (++q->rq[rw].count >= batch_requests && waitqueue_active(&q->wait_for_request))
+			wake_up(&q->wait_for_request);
 	}
 }
 
@@ -1144,7 +1117,7 @@ int __init blk_dev_init(void)
 	/*
 	 * Batch frees according to queue length
 	 */
-	batch_requests = queue_nr_requests/3;
+	batch_requests = queue_nr_requests/4;
 	printk("block: %d slots per queue, batch=%d\n", queue_nr_requests, batch_requests);
 
 #ifdef CONFIG_AMIGA_Z2RAM

drivers/block/nbd.c

@@ -471,7 +471,7 @@ static int nbd_release(struct inode *inode, struct file *file)
 
 static struct block_device_operations nbd_fops =
 {
-	owner:		THIS_MODULE.
+	owner:		THIS_MODULE,
 	open:		nbd_open,
 	release:	nbd_release,
 	ioctl:		nbd_ioctl,

drivers/char/random.c

@@ -1253,7 +1253,7 @@ static inline void xfer_secondary_pool(struct entropy_store *r,
 			  r == sec_random_state ? "secondary" : "unknown",
 			  r->entropy_count, nbytes * 8);
 
-		extract_entropy(random_state, tmp, nwords, 0);
+		extract_entropy(random_state, tmp, nwords * 4, 0);
 		add_entropy_words(r, tmp, nwords);
 		credit_entropy_store(r, nwords * 32);
 	}

drivers/char/vt.c

@@ -1183,6 +1183,24 @@ void complete_change_console(unsigned int new_console)
 	old_vc_mode = vt_cons[fg_console]->vc_mode;
 	switch_screen(new_console);
 
+	/*
+	 * This can't appear below a successful kill_proc().  If it did,
+	 * then the *blank_screen operation could occur while X, having
+	 * received acqsig, is waking up on another processor.  This
+	 * condition can lead to overlapping accesses to the VGA range
+	 * and the framebuffer (causing system lockups).
+	 *
+	 * To account for this we duplicate this code below only if the
+	 * controlling process is gone and we've called reset_vc.
+	 */
+	if (old_vc_mode != vt_cons[new_console]->vc_mode)
+	{
+		if (vt_cons[new_console]->vc_mode == KD_TEXT)
+			unblank_screen();
+		else
+			do_blank_screen(1);
+	}
+
 	/*
 	 * If this new console is under process control, send it a signal
 	 * telling it that it has acquired. Also check if it has died and
@@ -1209,19 +1227,15 @@ void complete_change_console(unsigned int new_console)
 		 * to account for and tracking tty count may be undesirable.
 		 */
 		        reset_vc(new_console);
-		}
-	}
 
-	/*
-	 * We do this here because the controlling process above may have
-	 * gone, and so there is now a new vc_mode
-	 */
-	if (old_vc_mode != vt_cons[new_console]->vc_mode)
-	{
-		if (vt_cons[new_console]->vc_mode == KD_TEXT)
-			unblank_screen();
-		else
-			do_blank_screen(1);
+			if (old_vc_mode != vt_cons[new_console]->vc_mode)
+			{
+				if (vt_cons[new_console]->vc_mode == KD_TEXT)
+					unblank_screen();
+				else
+					do_blank_screen(1);
+			}
+		}
 	}
 
 	/*

fs/buffer.c

@@ -705,8 +705,12 @@ void __invalidate_buffers(kdev_t dev, int destroy_dirty_buffers)
 
 static void free_more_memory(void)
 {
+	zone_t * zone = contig_page_data.node_zonelists[GFP_NOFS & GFP_ZONEMASK].zones[0];
+	
 	balance_dirty();
 	wakeup_bdflush();
+	try_to_free_pages(zone, GFP_NOFS, 0);
+	run_task_queue(&tq_disk);
 	current->policy |= SCHED_YIELD;
 	__set_current_state(TASK_RUNNING);
 	schedule();

include/linux/blkdev.h

@@ -66,14 +66,17 @@ typedef void (unplug_device_fn) (void *q);
  */
 #define QUEUE_NR_REQUESTS	8192
 
+struct request_list {
+	unsigned int count;
+	struct list_head free;
+};
+
 struct request_queue
 {
 	/*
 	 * the queue request freelist, one for reads and one for writes
 	 */
-	struct list_head	request_freelist[2];
-	struct list_head	pending_freelist[2];
-	int			pending_free[2];
+	struct request_list	rq[2];
 
 	/*
 	 * Together with queue_head for cacheline sharing

include/linux/pagemap.h

@@ -82,6 +82,7 @@ extern struct page * find_or_create_page(struct address_space *mapping,
 extern void lock_page(struct page *page);
 #define find_lock_page(mapping, index) \
 	__find_lock_page(mapping, index, page_hash(mapping, index))
+extern struct page *find_trylock_page(struct address_space *, unsigned long);
 
 extern void add_to_page_cache(struct page * page, struct address_space *mapping, unsigned long index);
 extern void add_to_page_cache_locked(struct page * page, struct address_space *mapping, unsigned long index);

include/linux/swap.h

@@ -145,6 +145,7 @@ extern int swap_duplicate(swp_entry_t);
 extern int swap_count(struct page *);
 extern int valid_swaphandles(swp_entry_t, unsigned long *);
 extern void swap_free(swp_entry_t);
+extern void free_swap_and_cache(swp_entry_t);
 struct swap_list_t {
 	int head;	/* head of priority-ordered swapfile list */
 	int next;	/* swapfile to be used next */

mm/filemap.c

@@ -818,6 +818,24 @@ struct page * __find_get_page(struct address_space *mapping,
 	return page;
 }
 
+/*
+ * Same as above, but trylock it instead of incrementing the count.
+ */
+struct page *find_trylock_page(struct address_space *mapping, unsigned long offset)
+{
+	struct page *page;
+	struct page **hash = page_hash(mapping, offset);
+
+	spin_lock(&pagecache_lock);
+	page = __find_page_nolock(mapping, offset, *hash);
+	if (page) {
+		if (TryLockPage(page))
+			page = NULL;
+	}
+	spin_unlock(&pagecache_lock);
+	return page;
+}
+
 /*
  * Must be called with the pagecache lock held,
  * will return with it held (but it may be dropped

mm/memory.c

@@ -325,7 +325,7 @@ static inline int zap_pte_range(mmu_gather_t *tlb, pmd_t * pmd, unsigned long ad
 			/* This will eventually call __free_pte on the pte. */
 			tlb_remove_page(tlb, ptep, address + offset);
 		} else {
-			swap_free(pte_to_swp_entry(pte));
+			free_swap_and_cache(pte_to_swp_entry(pte));
 			pte_clear(ptep);
 		}
 	}

mm/mmap.c

@@ -72,14 +72,6 @@ int vm_enough_memory(long pages)
 	free += nr_free_pages();
 	free += nr_swap_pages;
 
-	/*
-	 * This double-counts: the nrpages are both in the page-cache
-	 * and in the swapper space. At the same time, this compensates
-	 * for the swap-space over-allocation (ie "nr_swap_pages" being
-	 * too small. 
-	 */
-	free += swapper_space.nrpages;
-
 	/*
 	 * The code below doesn't account for free space in the inode
 	 * and dentry slab cache, slab cache fragmentation, inodes and

mm/swapfile.c

@@ -145,18 +145,13 @@ swp_entry_t get_swap_page(void)
 	return entry;
 }
 
-/*
- * Caller has made sure that the swapdevice corresponding to entry
- * is still around or has not been recycled.
- */
-void swap_free(swp_entry_t entry)
+static struct swap_info_struct * swap_info_get(swp_entry_t entry)
 {
 	struct swap_info_struct * p;
 	unsigned long offset, type;
 
 	if (!entry.val)
 		goto out;
-
 	type = SWP_TYPE(entry);
 	if (type >= nr_swapfiles)
 		goto bad_nofile;
@@ -172,8 +167,37 @@ void swap_free(swp_entry_t entry)
 	if (p->prio > swap_info[swap_list.next].prio)
 		swap_list.next = type;
 	swap_device_lock(p);
-	if (p->swap_map[offset] < SWAP_MAP_MAX) {
-		if (!--(p->swap_map[offset])) {
+	return p;
+
+bad_free:
+	printk(KERN_ERR "swap_free: %s%08lx\n", Unused_offset, entry.val);
+	goto out;
+bad_offset:
+	printk(KERN_ERR "swap_free: %s%08lx\n", Bad_offset, entry.val);
+	goto out;
+bad_device:
+	printk(KERN_ERR "swap_free: %s%08lx\n", Unused_file, entry.val);
+	goto out;
+bad_nofile:
+	printk(KERN_ERR "swap_free: %s%08lx\n", Bad_file, entry.val);
+out:
+	return NULL;
+}	
+
+static void swap_info_put(struct swap_info_struct * p)
+{
+	swap_device_unlock(p);
+	swap_list_unlock();
+}
+
+static int swap_entry_free(struct swap_info_struct *p, unsigned long offset)
+{
+	int count = p->swap_map[offset];
+
+	if (count < SWAP_MAP_MAX) {
+		count--;
+		p->swap_map[offset] = count;
+		if (!count) {
 			if (offset < p->lowest_bit)
 				p->lowest_bit = offset;
 			if (offset > p->highest_bit)
@@ -181,23 +205,45 @@ void swap_free(swp_entry_t entry)
 			nr_swap_pages++;
 		}
 	}
-	swap_device_unlock(p);
-	swap_list_unlock();
-out:
-	return;
+	return count;
+}
 
-bad_nofile:
-	printk(KERN_ERR "swap_free: %s%08lx\n", Bad_file, entry.val);
-	goto out;
-bad_device:
-	printk(KERN_ERR "swap_free: %s%08lx\n", Unused_file, entry.val);
-	goto out;
-bad_offset:
-	printk(KERN_ERR "swap_free: %s%08lx\n", Bad_offset, entry.val);
-	goto out;
-bad_free:
-	printk(KERN_ERR "swap_free: %s%08lx\n", Unused_offset, entry.val);
-	goto out;
+/*
+ * Caller has made sure that the swapdevice corresponding to entry
+ * is still around or has not been recycled.
+ */
+void swap_free(swp_entry_t entry)
+{
+	struct swap_info_struct * p;
+
+	p = swap_info_get(entry);
+	if (p) {
+		swap_entry_free(p, SWP_OFFSET(entry));
+		swap_info_put(p);
+	}
+}
+
+/*
+ * Free the swap entry like above, but also try to
+ * free the page cache entry if it is the last user.
+ */
+void free_swap_and_cache(swp_entry_t entry)
+{
+	struct swap_info_struct * p;
+	struct page *page = NULL;
+
+	p = swap_info_get(entry);
+	if (p) {
+		if (swap_entry_free(p, SWP_OFFSET(entry)) == 1)
+			page = find_trylock_page(&swapper_space, entry.val);
+		swap_info_put(p);
+	}
+	if (page) {
+		page_cache_get(page);
+		delete_from_swap_cache(page);
+		UnlockPage(page);
+		page_cache_release(page);
+	}
 }
 
 /*

mm/vmscan.c

@@ -292,8 +292,7 @@ static int swap_out(unsigned int priority, unsigned int gfp_mask, zone_t * class
 	int counter, nr_pages = SWAP_CLUSTER_MAX;
 	struct mm_struct *mm;
 
-	/* Then, look at the other mm's */
-	counter = mmlist_nr / priority;
+	counter = mmlist_nr;
 	do {
 		if (unlikely(current->need_resched)) {
 			__set_current_state(TASK_RUNNING);
@@ -334,7 +333,7 @@ static int shrink_cache(int nr_pages, zone_t * classzone, unsigned int gfp_mask,
 {
 	struct list_head * entry;
 	int max_scan = nr_inactive_pages / priority;
-	int max_mapped = max_scan / 4;
+	int max_mapped = nr_pages*10;
 
 	spin_lock(&pagemap_lru_lock);
 	while (--max_scan >= 0 && (entry = inactive_list.prev) != &inactive_list) {