Merge branch 'for-2.6.25' of git://git.kernel.dk/linux-2.6-block

* 'for-2.6.25' of git://git.kernel.dk/linux-2.6-block:
  block: implement drain buffers
  __bio_clone: don't calculate hw/phys segment counts
  block: allow queue dma_alignment of zero
  blktrace: Add blktrace ioctls to SCSI generic devices

8 files changed, 123 insertions, 24 deletions

diff --git a/block/blktrace.c b/block/blktrace.c
index 9b4da4ae3c7d..568588cd16b2 100644
--- a/block/blktrace.c
+++ b/block/blktrace.c
@@ -235,7 +235,7 @@ static void blk_trace_cleanup(struct blk_trace *bt)
         kfree(bt);
 }
 
-static int blk_trace_remove(struct request_queue *q)
+int blk_trace_remove(struct request_queue *q)
 {
         struct blk_trace *bt;
 
@@ -249,6 +249,7 @@ static int blk_trace_remove(struct request_queue *q)
 
         return 0;
 }
+EXPORT_SYMBOL_GPL(blk_trace_remove);
 
 static int blk_dropped_open(struct inode *inode, struct file *filp)
 {
@@ -316,18 +317,17 @@ static struct rchan_callbacks blk_relay_callbacks = {
 /*
  * Setup everything required to start tracing
  */
-int do_blk_trace_setup(struct request_queue *q, struct block_device *bdev,
+int do_blk_trace_setup(struct request_queue *q, char *name, dev_t dev,
                         struct blk_user_trace_setup *buts)
 {
         struct blk_trace *old_bt, *bt = NULL;
         struct dentry *dir = NULL;
-        char b[BDEVNAME_SIZE];
         int ret, i;
 
         if (!buts->buf_size || !buts->buf_nr)
                 return -EINVAL;
 
-        strcpy(buts->name, bdevname(bdev, b));
+        strcpy(buts->name, name);
 
         /*
          * some device names have larger paths - convert the slashes
@@ -352,7 +352,7 @@ int do_blk_trace_setup(struct request_queue *q, struct block_device *bdev,
                 goto err;
 
         bt->dir = dir;
-        bt->dev = bdev->bd_dev;
+        bt->dev = dev;
         atomic_set(&bt->dropped, 0);
 
         ret = -EIO;
@@ -399,8 +399,8 @@ err:
         return ret;
 }
 
-static int blk_trace_setup(struct request_queue *q, struct block_device *bdev,
-                           char __user *arg)
+int blk_trace_setup(struct request_queue *q, char *name, dev_t dev,
+                    char __user *arg)
 {
         struct blk_user_trace_setup buts;
         int ret;
@@ -409,7 +409,7 @@ static int blk_trace_setup(struct request_queue *q, struct block_device *bdev,
         if (ret)
                 return -EFAULT;
 
-        ret = do_blk_trace_setup(q, bdev, &buts);
+        ret = do_blk_trace_setup(q, name, dev, &buts);
         if (ret)
                 return ret;
 
@@ -418,8 +418,9 @@ static int blk_trace_setup(struct request_queue *q, struct block_device *bdev,
 
         return 0;
 }
+EXPORT_SYMBOL_GPL(blk_trace_setup);
 
-static int blk_trace_startstop(struct request_queue *q, int start)
+int blk_trace_startstop(struct request_queue *q, int start)
 {
         struct blk_trace *bt;
         int ret;
@@ -452,6 +453,7 @@ static int blk_trace_startstop(struct request_queue *q, int start)
 
         return ret;
 }
+EXPORT_SYMBOL_GPL(blk_trace_startstop);
 
 /**
  * blk_trace_ioctl: - handle the ioctls associated with tracing
@@ -464,6 +466,7 @@ int blk_trace_ioctl(struct block_device *bdev, unsigned cmd, char __user *arg)
 {
         struct request_queue *q;
         int ret, start = 0;
+        char b[BDEVNAME_SIZE];
 
         q = bdev_get_queue(bdev);
         if (!q)
@@ -473,7 +476,8 @@ int blk_trace_ioctl(struct block_device *bdev, unsigned cmd, char __user *arg)
 
         switch (cmd) {
         case BLKTRACESETUP:
-                ret = blk_trace_setup(q, bdev, arg);
+                strcpy(b, bdevname(bdev, b));
+                ret = blk_trace_setup(q, b, bdev->bd_dev, arg);
                 break;
         case BLKTRACESTART:
                 start = 1;
diff --git a/block/compat_ioctl.c b/block/compat_ioctl.c
index cae0a852619e..b73373216b0e 100644
--- a/block/compat_ioctl.c
+++ b/block/compat_ioctl.c
@@ -545,6 +545,7 @@ static int compat_blk_trace_setup(struct block_device *bdev, char __user *arg)
         struct blk_user_trace_setup buts;
         struct compat_blk_user_trace_setup cbuts;
         struct request_queue *q;
+        char b[BDEVNAME_SIZE];
         int ret;
 
         q = bdev_get_queue(bdev);
@@ -554,6 +555,8 @@ static int compat_blk_trace_setup(struct block_device *bdev, char __user *arg)
         if (copy_from_user(&cbuts, arg, sizeof(cbuts)))
                 return -EFAULT;
 
+        strcpy(b, bdevname(bdev, b));
+
         buts = (struct blk_user_trace_setup) {
                 .act_mask = cbuts.act_mask,
                 .buf_size = cbuts.buf_size,
@@ -565,7 +568,7 @@ static int compat_blk_trace_setup(struct block_device *bdev, char __user *arg)
         memcpy(&buts.name, &cbuts.name, 32);
 
         mutex_lock(&bdev->bd_mutex);
-        ret = do_blk_trace_setup(q, bdev, &buts);
+        ret = do_blk_trace_setup(q, b, bdev->bd_dev, &buts);
         mutex_unlock(&bdev->bd_mutex);
         if (ret)
                 return ret;
diff --git a/block/elevator.c b/block/elevator.c
index f9736fbdab03..8cd5775acd7a 100644
--- a/block/elevator.c
+++ b/block/elevator.c
@@ -741,7 +741,21 @@ struct request *elv_next_request(struct request_queue *q)
                         q->boundary_rq = NULL;
                 }
 
-                if ((rq->cmd_flags & REQ_DONTPREP) || !q->prep_rq_fn)
+                if (rq->cmd_flags & REQ_DONTPREP)
+                        break;
+
+                if (q->dma_drain_size && rq->data_len) {
+                        /*
+                         * make sure space for the drain appears we
+                         * know we can do this because max_hw_segments
+                         * has been adjusted to be one fewer than the
+                         * device can handle
+                         */
+                        rq->nr_phys_segments++;
+                        rq->nr_hw_segments++;
+                }
+
+                if (!q->prep_rq_fn)
                         break;
 
                 ret = q->prep_rq_fn(q, rq);
@@ -754,6 +768,16 @@ struct request *elv_next_request(struct request_queue *q)
                          * avoid resource deadlock.  REQ_STARTED will
                          * prevent other fs requests from passing this one.
                          */
+                        if (q->dma_drain_size && rq->data_len &&
+                            !(rq->cmd_flags & REQ_DONTPREP)) {
+                                /*
+                                 * remove the space for the drain we added
+                                 * so that we don't add it again
+                                 */
+                                --rq->nr_phys_segments;
+                                --rq->nr_hw_segments;
+                        }
+
                         rq = NULL;
                         break;
                 } else if (ret == BLKPREP_KILL) {
diff --git a/block/ll_rw_blk.c b/block/ll_rw_blk.c
index c16fdfed8c62..1932a56f5e4b 100644
--- a/block/ll_rw_blk.c
+++ b/block/ll_rw_blk.c
@@ -721,6 +721,45 @@ void blk_queue_stack_limits(struct request_queue *t, struct request_queue *b)
 EXPORT_SYMBOL(blk_queue_stack_limits);
 
 /**
+ * blk_queue_dma_drain - Set up a drain buffer for excess dma.
+ *
+ * @q:  the request queue for the device
+ * @buf:        physically contiguous buffer
+ * @size:       size of the buffer in bytes
+ *
+ * Some devices have excess DMA problems and can't simply discard (or
+ * zero fill) the unwanted piece of the transfer.  They have to have a
+ * real area of memory to transfer it into.  The use case for this is
+ * ATAPI devices in DMA mode.  If the packet command causes a transfer
+ * bigger than the transfer size some HBAs will lock up if there
+ * aren't DMA elements to contain the excess transfer.  What this API
+ * does is adjust the queue so that the buf is always appended
+ * silently to the scatterlist.
+ *
+ * Note: This routine adjusts max_hw_segments to make room for
+ * appending the drain buffer.  If you call
+ * blk_queue_max_hw_segments() or blk_queue_max_phys_segments() after
+ * calling this routine, you must set the limit to one fewer than your
+ * device can support otherwise there won't be room for the drain
+ * buffer.
+ */
+int blk_queue_dma_drain(struct request_queue *q, void *buf,
+                                unsigned int size)
+{
+        if (q->max_hw_segments < 2 || q->max_phys_segments < 2)
+                return -EINVAL;
+        /* make room for appending the drain */
+        --q->max_hw_segments;
+        --q->max_phys_segments;
+        q->dma_drain_buffer = buf;
+        q->dma_drain_size = size;
+
+        return 0;
+}
+
+EXPORT_SYMBOL_GPL(blk_queue_dma_drain);
+
+/**
  * blk_queue_segment_boundary - set boundary rules for segment merging
  * @q:  the request queue for the device
  * @mask:  the memory boundary mask
@@ -1374,6 +1413,16 @@ new_segment:
                 bvprv = bvec;
         } /* segments in rq */
 
+        if (q->dma_drain_size) {
+                sg->page_link &= ~0x02;
+                sg = sg_next(sg);
+                sg_set_page(sg, virt_to_page(q->dma_drain_buffer),
+                            q->dma_drain_size,
+                            ((unsigned long)q->dma_drain_buffer) &
+                            (PAGE_SIZE - 1));
+                nsegs++;
+        }
+
         if (sg)
                 sg_mark_end(sg);
 
diff --git a/drivers/scsi/sg.c b/drivers/scsi/sg.c
index 17216b76efdc..aba28f335b88 100644
--- a/drivers/scsi/sg.c
+++ b/drivers/scsi/sg.c
@@ -48,6 +48,7 @@ static int sg_version_num = 30534;	/* 2 digits for each component */
 #include <linux/blkdev.h>
 #include <linux/delay.h>
 #include <linux/scatterlist.h>
+#include <linux/blktrace_api.h>
 
 #include "scsi.h"
 #include <scsi/scsi_dbg.h>
@@ -1067,6 +1068,17 @@ sg_ioctl(struct inode *inode, struct file *filp,
         case BLKSECTGET:
                 return put_user(sdp->device->request_queue->max_sectors * 512,
                                 ip);
+        case BLKTRACESETUP:
+                return blk_trace_setup(sdp->device->request_queue,
+                                       sdp->disk->disk_name,
+                                       sdp->device->sdev_gendev.devt,
+                                       (char *)arg);
+        case BLKTRACESTART:
+                return blk_trace_startstop(sdp->device->request_queue, 1);
+        case BLKTRACESTOP:
+                return blk_trace_startstop(sdp->device->request_queue, 0);
+        case BLKTRACETEARDOWN:
+                return blk_trace_remove(sdp->device->request_queue);
         default:
                 if (read_only)
                         return -EPERM;  /* don't know so take safe approach */
diff --git a/fs/bio.c b/fs/bio.c
index d59ddbf79626..242e409dab4b 100644
--- a/fs/bio.c
+++ b/fs/bio.c
@@ -248,11 +248,13 @@ inline int bio_hw_segments(struct request_queue *q, struct bio *bio)
  */
 void __bio_clone(struct bio *bio, struct bio *bio_src)
 {
-        struct request_queue *q = bdev_get_queue(bio_src->bi_bdev);
-
         memcpy(bio->bi_io_vec, bio_src->bi_io_vec,
                 bio_src->bi_max_vecs * sizeof(struct bio_vec));
 
+        /*
+         * most users will be overriding ->bi_bdev with a new target,
+         * so we don't set nor calculate new physical/hw segment counts here
+         */
         bio->bi_sector = bio_src->bi_sector;
         bio->bi_bdev = bio_src->bi_bdev;
         bio->bi_flags |= 1 << BIO_CLONED;
@@ -260,8 +262,6 @@ void __bio_clone(struct bio *bio, struct bio *bio_src)
         bio->bi_vcnt = bio_src->bi_vcnt;
         bio->bi_size = bio_src->bi_size;
         bio->bi_idx = bio_src->bi_idx;
-        bio_phys_segments(q, bio);
-        bio_hw_segments(q, bio);
 }
 
 /**
diff --git a/include/linux/blkdev.h b/include/linux/blkdev.h
index b71c3900810d..71e7a847dffc 100644
--- a/include/linux/blkdev.h
+++ b/include/linux/blkdev.h
@@ -356,6 +356,8 @@ struct request_queue
         unsigned int            max_segment_size;
 
         unsigned long           seg_boundary_mask;
+        void                    *dma_drain_buffer;
+        unsigned int            dma_drain_size;
         unsigned int            dma_alignment;
 
         struct blk_queue_tag    *queue_tags;
@@ -692,6 +694,8 @@ extern void blk_queue_max_hw_segments(struct request_queue *, unsigned short);
 extern void blk_queue_max_segment_size(struct request_queue *, unsigned int);
 extern void blk_queue_hardsect_size(struct request_queue *, unsigned short);
 extern void blk_queue_stack_limits(struct request_queue *t, struct request_queue *b);
+extern int blk_queue_dma_drain(struct request_queue *q, void *buf,
+                               unsigned int size);
 extern void blk_queue_segment_boundary(struct request_queue *, unsigned long);
 extern void blk_queue_prep_rq(struct request_queue *, prep_rq_fn *pfn);
 extern void blk_queue_merge_bvec(struct request_queue *, merge_bvec_fn *);
@@ -768,12 +772,7 @@ static inline int bdev_hardsect_size(struct block_device *bdev)
 
 static inline int queue_dma_alignment(struct request_queue *q)
 {
-        int retval = 511;
-
-        if (q && q->dma_alignment)
-                retval = q->dma_alignment;
-
-        return retval;
+        return q ? q->dma_alignment : 511;
 }
 
 /* assumes size > 256 */
diff --git a/include/linux/blktrace_api.h b/include/linux/blktrace_api.h
index 7e11d23ac36a..06dadba349ac 100644
--- a/include/linux/blktrace_api.h
+++ b/include/linux/blktrace_api.h
@@ -148,7 +148,7 @@ extern int blk_trace_ioctl(struct block_device *, unsigned, char __user *);
 extern void blk_trace_shutdown(struct request_queue *);
 extern void __blk_add_trace(struct blk_trace *, sector_t, int, int, u32, int, int, void *);
 extern int do_blk_trace_setup(struct request_queue *q,
-        struct block_device *bdev, struct blk_user_trace_setup *buts);
+        char *name, dev_t dev, struct blk_user_trace_setup *buts);
 
 
 /**
@@ -282,6 +282,11 @@ static inline void blk_add_trace_remap(struct request_queue *q, struct bio *bio,
         __blk_add_trace(bt, from, bio->bi_size, bio->bi_rw, BLK_TA_REMAP, !bio_flagged(bio, BIO_UPTODATE), sizeof(r), &r);
 }
 
+extern int blk_trace_setup(request_queue_t *q, char *name, dev_t dev,
+                           char __user *arg);
+extern int blk_trace_startstop(request_queue_t *q, int start);
+extern int blk_trace_remove(request_queue_t *q);
+
 #else /* !CONFIG_BLK_DEV_IO_TRACE */
 #define blk_trace_ioctl(bdev, cmd, arg)         (-ENOTTY)
 #define blk_trace_shutdown(q)                   do { } while (0)
@@ -290,7 +295,10 @@ static inline void blk_add_trace_remap(struct request_queue *q, struct bio *bio,
 #define blk_add_trace_generic(q, rq, rw, what)  do { } while (0)
 #define blk_add_trace_pdu_int(q, what, bio, pdu)        do { } while (0)
 #define blk_add_trace_remap(q, bio, dev, f, t)  do {} while (0)
-#define do_blk_trace_setup(q, bdev, buts)       (-ENOTTY)
+#define do_blk_trace_setup(q, name, dev, buts)  (-ENOTTY)
+#define blk_trace_setup(q, name, dev, arg)      (-ENOTTY)
+#define blk_trace_startstop(q, start)           (-ENOTTY)
+#define blk_trace_remove(q)                     (-ENOTTY)
 #endif /* CONFIG_BLK_DEV_IO_TRACE */
 #endif /* __KERNEL__ */
 #endif