Merge branch 'for-linus' of git://git.kernel.dk/linux-2.6-block

* 'for-linus' of git://git.kernel.dk/linux-2.6-block:
  cciss: fix build for PROC_FS disabled
  block: fix amiga and atari floppy driver compile warning
  blk-throttle: Fix calculation of max number of WRITES to be dispatched
  ioprio: grab rcu_read_lock in sys_ioprio_{set,get}()
  xen/blkfront: cope with backend that fail empty BLKIF_OP_WRITE_BARRIER requests
  xen/blkfront: Implement FUA with BLKIF_OP_WRITE_BARRIER
  xen/blkfront: change blk_shadow.request to proper pointer
  xen/blkfront: map REQ_FLUSH into a full barrier

6 files changed, 43 insertions, 52 deletions

diff --git a/block/blk-throttle.c b/block/blk-throttle.c
index 56ad4531b412..004be80fd894 100644
--- a/block/blk-throttle.c
+++ b/block/blk-throttle.c
@@ -645,7 +645,7 @@ static int throtl_dispatch_tg(struct throtl_data *td, struct throtl_grp *tg,
 {
         unsigned int nr_reads = 0, nr_writes = 0;
         unsigned int max_nr_reads = throtl_grp_quantum*3/4;
-        unsigned int max_nr_writes = throtl_grp_quantum - nr_reads;
+        unsigned int max_nr_writes = throtl_grp_quantum - max_nr_reads;
         struct bio *bio;
 
         /* Try to dispatch 75% READS and 25% WRITES */
diff --git a/drivers/block/amiflop.c b/drivers/block/amiflop.c
index a1725e6488d3..7888501ad9ee 100644
--- a/drivers/block/amiflop.c
+++ b/drivers/block/amiflop.c
@@ -1341,7 +1341,7 @@ static struct request *set_next_request(void)
 {
         struct request_queue *q;
         int cnt = FD_MAX_UNITS;
-        struct request *rq;
+        struct request *rq = NULL;
 
         /* Find next queue we can dispatch from */
         fdc_queue = fdc_queue + 1;
diff --git a/drivers/block/ataflop.c b/drivers/block/ataflop.c
index 4e4cc6c828cb..605a67e40bbf 100644
--- a/drivers/block/ataflop.c
+++ b/drivers/block/ataflop.c
@@ -1399,7 +1399,7 @@ static struct request *set_next_request(void)
 {
         struct request_queue *q;
         int old_pos = fdc_queue;
-        struct request *rq;
+        struct request *rq = NULL;
 
         do {
                 q = unit[fdc_queue].disk->queue;
diff --git a/drivers/block/cciss.c b/drivers/block/cciss.c
index a67d0a611a8a..f291587d753e 100644
--- a/drivers/block/cciss.c
+++ b/drivers/block/cciss.c
@@ -66,6 +66,7 @@ MODULE_VERSION("3.6.26");
 MODULE_LICENSE("GPL");
 
 static DEFINE_MUTEX(cciss_mutex);
+static struct proc_dir_entry *proc_cciss;
 
 #include "cciss_cmd.h"
 #include "cciss.h"
@@ -363,8 +364,6 @@ static const char *raid_label[] = { "0", "4", "1(1+0)", "5", "5+1", "ADG",
 #define ENG_GIG_FACTOR (ENG_GIG/512)
 #define ENGAGE_SCSI     "engage scsi"
 
-static struct proc_dir_entry *proc_cciss;
-
 static void cciss_seq_show_header(struct seq_file *seq)
 {
         ctlr_info_t *h = seq->private;
diff --git a/drivers/block/xen-blkfront.c b/drivers/block/xen-blkfront.c
index 255035cfc88a..4f9e22f29138 100644
--- a/drivers/block/xen-blkfront.c
+++ b/drivers/block/xen-blkfront.c
@@ -65,7 +65,7 @@ enum blkif_state {
 
 struct blk_shadow {
         struct blkif_request req;
-        unsigned long request;
+        struct request *request;
         unsigned long frame[BLKIF_MAX_SEGMENTS_PER_REQUEST];
 };
 
@@ -136,7 +136,7 @@ static void add_id_to_freelist(struct blkfront_info *info,
                                unsigned long id)
 {
         info->shadow[id].req.id  = info->shadow_free;
-        info->shadow[id].request = 0;
+        info->shadow[id].request = NULL;
         info->shadow_free = id;
 }
 
@@ -245,14 +245,11 @@ static int blkif_ioctl(struct block_device *bdev, fmode_t mode,
 }
 
 /*
- * blkif_queue_request
+ * Generate a Xen blkfront IO request from a blk layer request.  Reads
+ * and writes are handled as expected.  Since we lack a loose flush
+ * request, we map flushes into a full ordered barrier.
  *
- * request block io
- *
- * id: for guest use only.
- * operation: BLKIF_OP_{READ,WRITE,PROBE}
- * buffer: buffer to read/write into. this should be a
- *   virtual address in the guest os.
+ * @req: a request struct
  */
 static int blkif_queue_request(struct request *req)
 {
@@ -281,7 +278,7 @@ static int blkif_queue_request(struct request *req)
         /* Fill out a communications ring structure. */
         ring_req = RING_GET_REQUEST(&info->ring, info->ring.req_prod_pvt);
         id = get_id_from_freelist(info);
-        info->shadow[id].request = (unsigned long)req;
+        info->shadow[id].request = req;
 
         ring_req->id = id;
         ring_req->sector_number = (blkif_sector_t)blk_rq_pos(req);
@@ -290,6 +287,18 @@ static int blkif_queue_request(struct request *req)
         ring_req->operation = rq_data_dir(req) ?
                 BLKIF_OP_WRITE : BLKIF_OP_READ;
 
+        if (req->cmd_flags & (REQ_FLUSH | REQ_FUA)) {
+                /*
+                 * Ideally we could just do an unordered
+                 * flush-to-disk, but all we have is a full write
+                 * barrier at the moment.  However, a barrier write is
+                 * a superset of FUA, so we can implement it the same
+                 * way.  (It's also a FLUSH+FUA, since it is
+                 * guaranteed ordered WRT previous writes.)
+                 */
+                ring_req->operation = BLKIF_OP_WRITE_BARRIER;
+        }
+
         ring_req->nr_segments = blk_rq_map_sg(req->q, req, info->sg);
         BUG_ON(ring_req->nr_segments > BLKIF_MAX_SEGMENTS_PER_REQUEST);
 
@@ -634,7 +643,7 @@ static irqreturn_t blkif_interrupt(int irq, void *dev_id)
 
                 bret = RING_GET_RESPONSE(&info->ring, i);
                 id   = bret->id;
-                req  = (struct request *)info->shadow[id].request;
+                req  = info->shadow[id].request;
 
                 blkif_completion(&info->shadow[id]);
 
@@ -647,6 +656,16 @@ static irqreturn_t blkif_interrupt(int irq, void *dev_id)
                                 printk(KERN_WARNING "blkfront: %s: write barrier op failed\n",
                                        info->gd->disk_name);
                                 error = -EOPNOTSUPP;
+                        }
+                        if (unlikely(bret->status == BLKIF_RSP_ERROR &&
+                                     info->shadow[id].req.nr_segments == 0)) {
+                                printk(KERN_WARNING "blkfront: %s: empty write barrier op failed\n",
+                                       info->gd->disk_name);
+                                error = -EOPNOTSUPP;
+                        }
+                        if (unlikely(error)) {
+                                if (error == -EOPNOTSUPP)
+                                        error = 0;
                                 info->feature_flush = 0;
                                 xlvbd_flush(info);
                         }
@@ -899,7 +918,7 @@ static int blkif_recover(struct blkfront_info *info)
         /* Stage 3: Find pending requests and requeue them. */
         for (i = 0; i < BLK_RING_SIZE; i++) {
                 /* Not in use? */
-                if (copy[i].request == 0)
+                if (!copy[i].request)
                         continue;
 
                 /* Grab a request slot and copy shadow state into it. */
@@ -916,9 +935,7 @@ static int blkif_recover(struct blkfront_info *info)
                                 req->seg[j].gref,
                                 info->xbdev->otherend_id,
                                 pfn_to_mfn(info->shadow[req->id].frame[j]),
-                                rq_data_dir(
-                                        (struct request *)
-                                        info->shadow[req->id].request));
+                                rq_data_dir(info->shadow[req->id].request));
                 info->shadow[req->id].req = *req;
 
                 info->ring.req_prod_pvt++;
@@ -1067,14 +1084,8 @@ static void blkfront_connect(struct blkfront_info *info)
          */
         info->feature_flush = 0;
 
-        /*
-         * The driver doesn't properly handled empty flushes, so
-         * lets disable barrier support for now.
-         */
-#if 0
         if (!err && barrier)
-                info->feature_flush = REQ_FLUSH;
-#endif
+                info->feature_flush = REQ_FLUSH | REQ_FUA;
 
         err = xlvbd_alloc_gendisk(sectors, info, binfo, sector_size);
         if (err) {
diff --git a/fs/ioprio.c b/fs/ioprio.c
index 2f7d05c89922..7da2a06508e5 100644
--- a/fs/ioprio.c
+++ b/fs/ioprio.c
@@ -103,22 +103,15 @@ SYSCALL_DEFINE3(ioprio_set, int, which, int, who, int, ioprio)
         }
 
         ret = -ESRCH;
-        /*
-         * We want IOPRIO_WHO_PGRP/IOPRIO_WHO_USER to be "atomic",
-         * so we can't use rcu_read_lock(). See re-copy of ->ioprio
-         * in copy_process().
-         */
-        read_lock(&tasklist_lock);
+        rcu_read_lock();
         switch (which) {
                 case IOPRIO_WHO_PROCESS:
-                        rcu_read_lock();
                         if (!who)
                                 p = current;
                         else
                                 p = find_task_by_vpid(who);
                         if (p)
                                 ret = set_task_ioprio(p, ioprio);
-                        rcu_read_unlock();
                         break;
                 case IOPRIO_WHO_PGRP:
                         if (!who)
@@ -141,12 +134,7 @@ SYSCALL_DEFINE3(ioprio_set, int, which, int, who, int, ioprio)
                                 break;
 
                         do_each_thread(g, p) {
-                                int match;
-
-                                rcu_read_lock();
-                                match = __task_cred(p)->uid == who;
-                                rcu_read_unlock();
-                                if (!match)
+                                if (__task_cred(p)->uid != who)
                                         continue;
                                 ret = set_task_ioprio(p, ioprio);
                                 if (ret)
@@ -160,7 +148,7 @@ free_uid:
                         ret = -EINVAL;
         }
 
-        read_unlock(&tasklist_lock);
+        rcu_read_unlock();
         return ret;
 }
 
@@ -204,17 +192,15 @@ SYSCALL_DEFINE2(ioprio_get, int, which, int, who)
         int ret = -ESRCH;
         int tmpio;
 
-        read_lock(&tasklist_lock);
+        rcu_read_lock();
         switch (which) {
                 case IOPRIO_WHO_PROCESS:
-                        rcu_read_lock();
                         if (!who)
                                 p = current;
                         else
                                 p = find_task_by_vpid(who);
                         if (p)
                                 ret = get_task_ioprio(p);
-                        rcu_read_unlock();
                         break;
                 case IOPRIO_WHO_PGRP:
                         if (!who)
@@ -241,12 +227,7 @@ SYSCALL_DEFINE2(ioprio_get, int, which, int, who)
                                 break;
 
                         do_each_thread(g, p) {
-                                int match;
-
-                                rcu_read_lock();
-                                match = __task_cred(p)->uid == user->uid;
-                                rcu_read_unlock();
-                                if (!match)
+                                if (__task_cred(p)->uid != user->uid)
                                         continue;
                                 tmpio = get_task_ioprio(p);
                                 if (tmpio < 0)
@@ -264,6 +245,6 @@ SYSCALL_DEFINE2(ioprio_get, int, which, int, who)
                         ret = -EINVAL;
         }
 
-        read_unlock(&tasklist_lock);
+        rcu_read_unlock();
         return ret;
 }