Add the rt linux 4.1.3-rt3 as base
[kvmfornfv.git] / kernel / block / bounce.c
diff --git a/kernel/block/bounce.c b/kernel/block/bounce.c
new file mode 100644 (file)
index 0000000..39d123e
--- /dev/null
@@ -0,0 +1,290 @@
+/* bounce buffer handling for block devices
+ *
+ * - Split from highmem.c
+ */
+
+#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
+
+#include <linux/mm.h>
+#include <linux/export.h>
+#include <linux/swap.h>
+#include <linux/gfp.h>
+#include <linux/bio.h>
+#include <linux/pagemap.h>
+#include <linux/mempool.h>
+#include <linux/blkdev.h>
+#include <linux/init.h>
+#include <linux/hash.h>
+#include <linux/highmem.h>
+#include <linux/bootmem.h>
+#include <linux/printk.h>
+#include <asm/tlbflush.h>
+
+#include <trace/events/block.h>
+
+#define POOL_SIZE      64
+#define ISA_POOL_SIZE  16
+
+static mempool_t *page_pool, *isa_page_pool;
+
+#if defined(CONFIG_HIGHMEM) || defined(CONFIG_NEED_BOUNCE_POOL)
+static __init int init_emergency_pool(void)
+{
+#if defined(CONFIG_HIGHMEM) && !defined(CONFIG_MEMORY_HOTPLUG)
+       if (max_pfn <= max_low_pfn)
+               return 0;
+#endif
+
+       page_pool = mempool_create_page_pool(POOL_SIZE, 0);
+       BUG_ON(!page_pool);
+       pr_info("pool size: %d pages\n", POOL_SIZE);
+
+       return 0;
+}
+
+__initcall(init_emergency_pool);
+#endif
+
+#ifdef CONFIG_HIGHMEM
+/*
+ * highmem version, map in to vec
+ */
+static void bounce_copy_vec(struct bio_vec *to, unsigned char *vfrom)
+{
+       unsigned long flags;
+       unsigned char *vto;
+
+       local_irq_save_nort(flags);
+       vto = kmap_atomic(to->bv_page);
+       memcpy(vto + to->bv_offset, vfrom, to->bv_len);
+       kunmap_atomic(vto);
+       local_irq_restore_nort(flags);
+}
+
+#else /* CONFIG_HIGHMEM */
+
+#define bounce_copy_vec(to, vfrom)     \
+       memcpy(page_address((to)->bv_page) + (to)->bv_offset, vfrom, (to)->bv_len)
+
+#endif /* CONFIG_HIGHMEM */
+
+/*
+ * allocate pages in the DMA region for the ISA pool
+ */
+static void *mempool_alloc_pages_isa(gfp_t gfp_mask, void *data)
+{
+       return mempool_alloc_pages(gfp_mask | GFP_DMA, data);
+}
+
+/*
+ * gets called "every" time someone init's a queue with BLK_BOUNCE_ISA
+ * as the max address, so check if the pool has already been created.
+ */
+int init_emergency_isa_pool(void)
+{
+       if (isa_page_pool)
+               return 0;
+
+       isa_page_pool = mempool_create(ISA_POOL_SIZE, mempool_alloc_pages_isa,
+                                      mempool_free_pages, (void *) 0);
+       BUG_ON(!isa_page_pool);
+
+       pr_info("isa pool size: %d pages\n", ISA_POOL_SIZE);
+       return 0;
+}
+
+/*
+ * Simple bounce buffer support for highmem pages. Depending on the
+ * queue gfp mask set, *to may or may not be a highmem page. kmap it
+ * always, it will do the Right Thing
+ */
+static void copy_to_high_bio_irq(struct bio *to, struct bio *from)
+{
+       unsigned char *vfrom;
+       struct bio_vec tovec, *fromvec = from->bi_io_vec;
+       struct bvec_iter iter;
+
+       bio_for_each_segment(tovec, to, iter) {
+               if (tovec.bv_page != fromvec->bv_page) {
+                       /*
+                        * fromvec->bv_offset and fromvec->bv_len might have
+                        * been modified by the block layer, so use the original
+                        * copy, bounce_copy_vec already uses tovec->bv_len
+                        */
+                       vfrom = page_address(fromvec->bv_page) +
+                               tovec.bv_offset;
+
+                       bounce_copy_vec(&tovec, vfrom);
+                       flush_dcache_page(tovec.bv_page);
+               }
+
+               fromvec++;
+       }
+}
+
+static void bounce_end_io(struct bio *bio, mempool_t *pool, int err)
+{
+       struct bio *bio_orig = bio->bi_private;
+       struct bio_vec *bvec, *org_vec;
+       int i;
+
+       if (test_bit(BIO_EOPNOTSUPP, &bio->bi_flags))
+               set_bit(BIO_EOPNOTSUPP, &bio_orig->bi_flags);
+
+       /*
+        * free up bounce indirect pages used
+        */
+       bio_for_each_segment_all(bvec, bio, i) {
+               org_vec = bio_orig->bi_io_vec + i;
+               if (bvec->bv_page == org_vec->bv_page)
+                       continue;
+
+               dec_zone_page_state(bvec->bv_page, NR_BOUNCE);
+               mempool_free(bvec->bv_page, pool);
+       }
+
+       bio_endio(bio_orig, err);
+       bio_put(bio);
+}
+
+static void bounce_end_io_write(struct bio *bio, int err)
+{
+       bounce_end_io(bio, page_pool, err);
+}
+
+static void bounce_end_io_write_isa(struct bio *bio, int err)
+{
+
+       bounce_end_io(bio, isa_page_pool, err);
+}
+
+static void __bounce_end_io_read(struct bio *bio, mempool_t *pool, int err)
+{
+       struct bio *bio_orig = bio->bi_private;
+
+       if (test_bit(BIO_UPTODATE, &bio->bi_flags))
+               copy_to_high_bio_irq(bio_orig, bio);
+
+       bounce_end_io(bio, pool, err);
+}
+
+static void bounce_end_io_read(struct bio *bio, int err)
+{
+       __bounce_end_io_read(bio, page_pool, err);
+}
+
+static void bounce_end_io_read_isa(struct bio *bio, int err)
+{
+       __bounce_end_io_read(bio, isa_page_pool, err);
+}
+
+#ifdef CONFIG_NEED_BOUNCE_POOL
+static int must_snapshot_stable_pages(struct request_queue *q, struct bio *bio)
+{
+       if (bio_data_dir(bio) != WRITE)
+               return 0;
+
+       if (!bdi_cap_stable_pages_required(&q->backing_dev_info))
+               return 0;
+
+       return test_bit(BIO_SNAP_STABLE, &bio->bi_flags);
+}
+#else
+static int must_snapshot_stable_pages(struct request_queue *q, struct bio *bio)
+{
+       return 0;
+}
+#endif /* CONFIG_NEED_BOUNCE_POOL */
+
+static void __blk_queue_bounce(struct request_queue *q, struct bio **bio_orig,
+                              mempool_t *pool, int force)
+{
+       struct bio *bio;
+       int rw = bio_data_dir(*bio_orig);
+       struct bio_vec *to, from;
+       struct bvec_iter iter;
+       unsigned i;
+
+       if (force)
+               goto bounce;
+       bio_for_each_segment(from, *bio_orig, iter)
+               if (page_to_pfn(from.bv_page) > queue_bounce_pfn(q))
+                       goto bounce;
+
+       return;
+bounce:
+       bio = bio_clone_bioset(*bio_orig, GFP_NOIO, fs_bio_set);
+
+       bio_for_each_segment_all(to, bio, i) {
+               struct page *page = to->bv_page;
+
+               if (page_to_pfn(page) <= queue_bounce_pfn(q) && !force)
+                       continue;
+
+               to->bv_page = mempool_alloc(pool, q->bounce_gfp);
+               inc_zone_page_state(to->bv_page, NR_BOUNCE);
+
+               if (rw == WRITE) {
+                       char *vto, *vfrom;
+
+                       flush_dcache_page(page);
+
+                       vto = page_address(to->bv_page) + to->bv_offset;
+                       vfrom = kmap_atomic(page) + to->bv_offset;
+                       memcpy(vto, vfrom, to->bv_len);
+                       kunmap_atomic(vfrom);
+               }
+       }
+
+       trace_block_bio_bounce(q, *bio_orig);
+
+       bio->bi_flags |= (1 << BIO_BOUNCED);
+
+       if (pool == page_pool) {
+               bio->bi_end_io = bounce_end_io_write;
+               if (rw == READ)
+                       bio->bi_end_io = bounce_end_io_read;
+       } else {
+               bio->bi_end_io = bounce_end_io_write_isa;
+               if (rw == READ)
+                       bio->bi_end_io = bounce_end_io_read_isa;
+       }
+
+       bio->bi_private = *bio_orig;
+       *bio_orig = bio;
+}
+
+void blk_queue_bounce(struct request_queue *q, struct bio **bio_orig)
+{
+       int must_bounce;
+       mempool_t *pool;
+
+       /*
+        * Data-less bio, nothing to bounce
+        */
+       if (!bio_has_data(*bio_orig))
+               return;
+
+       must_bounce = must_snapshot_stable_pages(q, *bio_orig);
+
+       /*
+        * for non-isa bounce case, just check if the bounce pfn is equal
+        * to or bigger than the highest pfn in the system -- in that case,
+        * don't waste time iterating over bio segments
+        */
+       if (!(q->bounce_gfp & GFP_DMA)) {
+               if (queue_bounce_pfn(q) >= blk_max_pfn && !must_bounce)
+                       return;
+               pool = page_pool;
+       } else {
+               BUG_ON(!isa_page_pool);
+               pool = isa_page_pool;
+       }
+
+       /*
+        * slow path
+        */
+       __blk_queue_bounce(q, bio_orig, pool, must_bounce);
+}
+
+EXPORT_SYMBOL(blk_queue_bounce);