RDS: Use page_remainder_alloc() for recv bufs
authorAndy Grover <andy.grover@oracle.com>
Tue, 25 May 2010 03:12:41 +0000 (20:12 -0700)
committerAndy Grover <andy.grover@oracle.com>
Thu, 9 Sep 2010 01:15:20 +0000 (18:15 -0700)
Instead of splitting up a page into RDS_FRAG_SIZE chunks
ourselves, ask rds_page_remainder_alloc() to do it. While it
is possible PAGE_SIZE > FRAG_SIZE, on x86en it isn't, so having
duplicate "carve up a page into buffers" code seems excessive.

The other modification this spawns is the use of a single
struct scatterlist in rds_page_frag instead of a bare page ptr.
This causes verbosity to increase in some places, and decrease
in others.

Finally, I decided to unify the lifetimes and alloc/free of
rds_page_frag and its page. This is a nice simplification in itself,
but will be extra-nice once we come to adding cmason's recycling
patch.

Signed-off-by: Andy Grover <andy.grover@oracle.com>
net/rds/ib.h
net/rds/ib_recv.c
net/rds/page.c

index 282ec69..9bb7a74 100644 (file)
@@ -28,13 +28,9 @@ extern struct list_head rds_ib_devices;
  * try and minimize the amount of memory tied up both the device and
  * socket receive queues.
  */
-/* page offset of the final full frag that fits in the page */
-#define RDS_PAGE_LAST_OFF (((PAGE_SIZE  / RDS_FRAG_SIZE) - 1) * RDS_FRAG_SIZE)
 struct rds_page_frag {
        struct list_head        f_item;
-       struct page             *f_page;
-       unsigned long           f_offset;
-       dma_addr_t              f_mapped;
+       struct scatterlist      f_sg;
 };
 
 struct rds_ib_incoming {
@@ -107,7 +103,6 @@ struct rds_ib_connection {
        struct rds_header       *i_recv_hdrs;
        u64                     i_recv_hdrs_dma;
        struct rds_ib_recv_work *i_recvs;
-       struct rds_page_frag    i_frag;
        u64                     i_ack_recv;     /* last ACK received */
 
        /* sending acks */
index 24d1461..f6dbf16 100644 (file)
@@ -43,17 +43,11 @@ static struct kmem_cache *rds_ib_incoming_slab;
 static struct kmem_cache *rds_ib_frag_slab;
 static atomic_t        rds_ib_allocation = ATOMIC_INIT(0);
 
-static void rds_ib_frag_drop_page(struct rds_page_frag *frag)
-{
-       rdsdebug("frag %p page %p\n", frag, frag->f_page);
-       __free_page(frag->f_page);
-       frag->f_page = NULL;
-}
-
+/* Free frag and attached recv buffer f_sg */
 static void rds_ib_frag_free(struct rds_page_frag *frag)
 {
-       rdsdebug("frag %p page %p\n", frag, frag->f_page);
-       BUG_ON(frag->f_page);
+       rdsdebug("frag %p page %p\n", frag, sg_page(&frag->f_sg));
+       __free_page(sg_page(&frag->f_sg));
        kmem_cache_free(rds_ib_frag_slab, frag);
 }
 
@@ -71,12 +65,8 @@ static void rds_ib_recv_unmap_page(struct rds_ib_connection *ic,
 {
        struct rds_page_frag *frag = recv->r_frag;
 
-       rdsdebug("recv %p frag %p page %p\n", recv, frag, frag->f_page);
-       if (frag->f_mapped)
-               ib_dma_unmap_page(ic->i_cm_id->device,
-                              frag->f_mapped,
-                              RDS_FRAG_SIZE, DMA_FROM_DEVICE);
-       frag->f_mapped = 0;
+       rdsdebug("recv %p frag %p page %p\n", recv, frag, sg_page(&frag->f_sg));
+       ib_dma_unmap_sg(ic->i_cm_id->device, &frag->f_sg, 1, DMA_FROM_DEVICE);
 }
 
 void rds_ib_recv_init_ring(struct rds_ib_connection *ic)
@@ -116,8 +106,6 @@ static void rds_ib_recv_clear_one(struct rds_ib_connection *ic,
        }
        if (recv->r_frag) {
                rds_ib_recv_unmap_page(ic, recv);
-               if (recv->r_frag->f_page)
-                       rds_ib_frag_drop_page(recv->r_frag);
                rds_ib_frag_free(recv->r_frag);
                recv->r_frag = NULL;
        }
@@ -129,16 +117,12 @@ void rds_ib_recv_clear_ring(struct rds_ib_connection *ic)
 
        for (i = 0; i < ic->i_recv_ring.w_nr; i++)
                rds_ib_recv_clear_one(ic, &ic->i_recvs[i]);
-
-       if (ic->i_frag.f_page)
-               rds_ib_frag_drop_page(&ic->i_frag);
 }
 
 static int rds_ib_recv_refill_one(struct rds_connection *conn,
                                  struct rds_ib_recv_work *recv)
 {
        struct rds_ib_connection *ic = conn->c_transport_data;
-       dma_addr_t dma_addr;
        struct ib_sge *sge;
        int ret = -ENOMEM;
 
@@ -161,50 +145,27 @@ static int rds_ib_recv_refill_one(struct rds_connection *conn,
                if (!recv->r_frag)
                        goto out;
                INIT_LIST_HEAD(&recv->r_frag->f_item);
-               recv->r_frag->f_page = NULL;
-       }
-
-       if (!ic->i_frag.f_page) {
-               ic->i_frag.f_page = alloc_page(GFP_NOWAIT);
-               if (!ic->i_frag.f_page)
+               sg_init_table(&recv->r_frag->f_sg, 1);
+               ret = rds_page_remainder_alloc(&recv->r_frag->f_sg,
+                                              RDS_FRAG_SIZE, GFP_NOWAIT);
+               if (ret) {
+                       kmem_cache_free(rds_ib_frag_slab, recv->r_frag);
+                       recv->r_frag = NULL;
                        goto out;
-               ic->i_frag.f_offset = 0;
+               }
        }
 
-       dma_addr = ib_dma_map_page(ic->i_cm_id->device,
-                                 ic->i_frag.f_page,
-                                 ic->i_frag.f_offset,
-                                 RDS_FRAG_SIZE,
-                                 DMA_FROM_DEVICE);
-       if (ib_dma_mapping_error(ic->i_cm_id->device, dma_addr))
-               goto out;
-
-       /*
-        * Once we get the RDS_PAGE_LAST_OFF frag then rds_ib_frag_unmap()
-        * must be called on this recv.  This happens as completions hit
-        * in order or on connection shutdown.
-        */
-       recv->r_frag->f_page = ic->i_frag.f_page;
-       recv->r_frag->f_offset = ic->i_frag.f_offset;
-       recv->r_frag->f_mapped = dma_addr;
+       ret = ib_dma_map_sg(ic->i_cm_id->device, &recv->r_frag->f_sg,
+                           1, DMA_FROM_DEVICE);
+       WARN_ON(ret != 1);
 
        sge = &recv->r_sge[0];
        sge->addr = ic->i_recv_hdrs_dma + (recv - ic->i_recvs) * sizeof(struct rds_header);
        sge->length = sizeof(struct rds_header);
 
        sge = &recv->r_sge[1];
-       sge->addr = dma_addr;
-       sge->length = RDS_FRAG_SIZE;
-
-       get_page(recv->r_frag->f_page);
-
-       if (ic->i_frag.f_offset < RDS_PAGE_LAST_OFF) {
-               ic->i_frag.f_offset += RDS_FRAG_SIZE;
-       } else {
-               put_page(ic->i_frag.f_page);
-               ic->i_frag.f_page = NULL;
-               ic->i_frag.f_offset = 0;
-       }
+       sge->addr = sg_dma_address(&recv->r_frag->f_sg);
+       sge->length = sg_dma_len(&recv->r_frag->f_sg);
 
        ret = 0;
 out:
@@ -247,8 +208,8 @@ int rds_ib_recv_refill(struct rds_connection *conn, int prefill)
                /* XXX when can this fail? */
                ret = ib_post_recv(ic->i_cm_id->qp, &recv->r_wr, &failed_wr);
                rdsdebug("recv %p ibinc %p page %p addr %lu ret %d\n", recv,
-                        recv->r_ibinc, recv->r_frag->f_page,
-                        (long) recv->r_frag->f_mapped, ret);
+                        recv->r_ibinc, sg_page(&recv->r_frag->f_sg),
+                        (long) sg_dma_address(&recv->r_frag->f_sg), ret);
                if (ret) {
                        rds_ib_conn_error(conn, "recv post on "
                               "%pI4 returned %d, disconnecting and "
@@ -281,7 +242,6 @@ static void rds_ib_inc_purge(struct rds_incoming *inc)
 
        list_for_each_entry_safe(frag, pos, &ibinc->ii_frags, f_item) {
                list_del_init(&frag->f_item);
-               rds_ib_frag_drop_page(frag);
                rds_ib_frag_free(frag);
        }
 }
@@ -333,13 +293,13 @@ int rds_ib_inc_copy_to_user(struct rds_incoming *inc, struct iovec *first_iov,
                to_copy = min_t(unsigned long, to_copy, len - copied);
 
                rdsdebug("%lu bytes to user [%p, %zu] + %lu from frag "
-                        "[%p, %lu] + %lu\n",
+                        "[%p, %u] + %lu\n",
                         to_copy, iov->iov_base, iov->iov_len, iov_off,
-                        frag->f_page, frag->f_offset, frag_off);
+                        sg_page(&frag->f_sg), frag->f_sg.offset, frag_off);
 
                /* XXX needs + offset for multiple recvs per page */
-               ret = rds_page_copy_to_user(frag->f_page,
-                                           frag->f_offset + frag_off,
+               ret = rds_page_copy_to_user(sg_page(&frag->f_sg),
+                                           frag->f_sg.offset + frag_off,
                                            iov->iov_base + iov_off,
                                            to_copy);
                if (ret) {
@@ -595,7 +555,7 @@ static void rds_ib_cong_recv(struct rds_connection *conn,
                to_copy = min(RDS_FRAG_SIZE - frag_off, PAGE_SIZE - map_off);
                BUG_ON(to_copy & 7); /* Must be 64bit aligned. */
 
-               addr = kmap_atomic(frag->f_page, KM_SOFTIRQ0);
+               addr = kmap_atomic(sg_page(&frag->f_sg), KM_SOFTIRQ0);
 
                src = addr + frag_off;
                dst = (void *)map->m_page_addrs[map_page] + map_off;
@@ -698,12 +658,12 @@ static void rds_ib_process_recv(struct rds_connection *conn,
                 * the inc is freed.  We don't go that route, so we have to drop the
                 * page ref ourselves.  We can't just leave the page on the recv
                 * because that confuses the dma mapping of pages and each recv's use
-                * of a partial page.  We can leave the frag, though, it will be
-                * reused.
+                * of a partial page.
                 *
                 * FIXME: Fold this into the code path below.
                 */
-               rds_ib_frag_drop_page(recv->r_frag);
+               rds_ib_frag_free(recv->r_frag);
+               recv->r_frag = NULL;
                return;
        }
 
index e5b2527..5e44f5a 100644 (file)
@@ -186,6 +186,7 @@ out:
                 ret ? 0 : scat->length);
        return ret;
 }
+EXPORT_SYMBOL_GPL(rds_page_remainder_alloc);
 
 static int rds_page_remainder_cpu_notify(struct notifier_block *self,
                                         unsigned long action, void *hcpu)