*/
/*
* This file is part of Lustre, http://www.lustre.org/
- * Lustre is a trademark of Sun Microsystems, Inc.
*
* Client IO.
*
#include <lustre_fid.h>
#include <cl_object.h>
#include "cl_internal.h"
+#include <libcfs/crypto/llcrypt.h>
/*****************************************************************************
*
case CIT_GLIMPSE:
break;
case CIT_LADVISE:
+ case CIT_LSEEK:
break;
default:
LBUG();
}
EXPORT_SYMBOL(cl_io_rw_init);
+#ifdef HAVE_LIST_CMP_FUNC_T
+static int cl_lock_descr_cmp(void *priv,
+ const struct list_head *a,
+ const struct list_head *b)
+#else /* !HAVE_LIST_CMP_FUNC_T */
static int cl_lock_descr_cmp(void *priv,
struct list_head *a, struct list_head *b)
+#endif /* HAVE_LIST_CMP_FUNC_T */
{
const struct cl_io_lock_link *l0 = list_entry(a, struct cl_io_lock_link,
cill_linkage);
const struct cl_io_slice *scan;
int result = 0;
- LINVRNT(io->ci_type == CIT_READ || io->ci_type == CIT_FAULT);
+ LINVRNT(io->ci_type == CIT_READ ||
+ io->ci_type == CIT_FAULT ||
+ io->ci_type == CIT_WRITE);
LINVRNT(io->ci_state == CIS_IO_GOING || io->ci_state == CIS_LOCKED);
LINVRNT(cl_io_invariant(io));
ENTRY;
EXPORT_SYMBOL(cl_io_read_ahead);
/**
+ * Called before io start, to reserve enough LRU slots to avoid
+ * deadlock.
+ *
+ * \see cl_io_operations::cio_lru_reserve()
+ */
+int cl_io_lru_reserve(const struct lu_env *env, struct cl_io *io,
+ loff_t pos, size_t bytes)
+{
+ const struct cl_io_slice *scan;
+ int result = 0;
+
+ LINVRNT(io->ci_type == CIT_READ || io->ci_type == CIT_WRITE);
+ LINVRNT(cl_io_invariant(io));
+ ENTRY;
+
+ list_for_each_entry(scan, &io->ci_layers, cis_linkage) {
+ if (scan->cis_iop->cio_lru_reserve) {
+ result = scan->cis_iop->cio_lru_reserve(env, scan,
+ pos, bytes);
+ if (result)
+ break;
+ }
+ }
+
+ RETURN(result);
+}
+EXPORT_SYMBOL(cl_io_lru_reserve);
+
+/**
* Commit a list of contiguous pages into writeback cache.
*
* \returns 0 if all pages committed, or errcode if error occurred.
* \see cl_io_operations::cio_commit_async()
*/
int cl_io_commit_async(const struct lu_env *env, struct cl_io *io,
- struct cl_page_list *queue, int from, int to,
- cl_commit_cbt cb)
+ struct cl_page_list *queue, int from, int to,
+ cl_commit_cbt cb)
{
const struct cl_io_slice *scan;
int result = 0;
}
EXPORT_SYMBOL(cl_io_commit_async);
+void cl_io_extent_release(const struct lu_env *env, struct cl_io *io)
+{
+ const struct cl_io_slice *scan;
+ ENTRY;
+
+ list_for_each_entry(scan, &io->ci_layers, cis_linkage) {
+ if (scan->cis_iop->cio_extent_release == NULL)
+ continue;
+ scan->cis_iop->cio_extent_release(env, scan);
+ }
+ EXIT;
+}
+EXPORT_SYMBOL(cl_io_extent_release);
+
/**
* Submits a list of pages for immediate io.
*
struct cl_sync_io *anchor = &cl_env_info(env)->clt_anchor;
struct cl_page *pg;
int rc;
+ ENTRY;
cl_page_list_for_each(pg, &queue->c2_qin) {
LASSERT(pg->cp_sync_io == NULL);
pg->cp_sync_io = anchor;
}
- cl_sync_io_init(anchor, queue->c2_qin.pl_nr, &cl_sync_io_end);
+ cl_sync_io_init(anchor, queue->c2_qin.pl_nr);
rc = cl_io_submit_rw(env, io, iot, queue);
if (rc == 0) {
/*
cl_page_list_for_each(pg, &queue->c2_qin)
pg->cp_sync_io = NULL;
}
- return rc;
+ RETURN(rc);
}
EXPORT_SYMBOL(cl_io_submit_sync);
/**
- * Cancel an IO which has been submitted by cl_io_submit_rw.
- */
-int cl_io_cancel(const struct lu_env *env, struct cl_io *io,
- struct cl_page_list *queue)
-{
- struct cl_page *page;
- int result = 0;
-
- CERROR("Canceling ongoing page trasmission\n");
- cl_page_list_for_each(page, queue) {
- int rc;
-
- rc = cl_page_cancel(env, page);
- result = result ?: rc;
- }
- return result;
-}
-
-/**
* Main io loop.
*
* Pumps io through iterations calling
*/
int cl_io_loop(const struct lu_env *env, struct cl_io *io)
{
- int result = 0;
+ int result = 0;
+ int rc = 0;
LINVRNT(cl_io_is_loopable(io));
ENTRY;
}
}
cl_io_iter_fini(env, io);
- } while (result == 0 && io->ci_continue);
+ if (result)
+ rc = result;
+ } while ((result == 0 || result == -EIOCBQUEUED) &&
+ io->ci_continue);
+
+ if (rc && !result)
+ result = rc;
- if (result == -EWOULDBLOCK && io->ci_ndelay) {
+ if (result == -EAGAIN && io->ci_ndelay) {
io->ci_need_restart = 1;
result = 0;
}
ENTRY;
plist->pl_nr = 0;
INIT_LIST_HEAD(&plist->pl_pages);
- plist->pl_owner = current;
EXIT;
}
EXPORT_SYMBOL(cl_page_list_init);
/**
* Adds a page to a page list.
*/
-void cl_page_list_add(struct cl_page_list *plist, struct cl_page *page)
+void cl_page_list_add(struct cl_page_list *plist, struct cl_page *page,
+ bool get_ref)
{
ENTRY;
/* it would be better to check that page is owned by "current" io, but
* it is not passed here. */
LASSERT(page->cp_owner != NULL);
- LINVRNT(plist->pl_owner == current);
LASSERT(list_empty(&page->cp_batch));
list_add_tail(&page->cp_batch, &plist->pl_pages);
++plist->pl_nr;
lu_ref_add_at(&page->cp_reference, &page->cp_queue_ref, "queue", plist);
- cl_page_get(page);
+ if (get_ref)
+ cl_page_get(page);
EXIT;
}
EXPORT_SYMBOL(cl_page_list_add);
{
LASSERT(plist->pl_nr > 0);
LASSERT(cl_page_is_vmlocked(env, page));
- LINVRNT(plist->pl_owner == current);
ENTRY;
list_del_init(&page->cp_batch);
struct cl_page *page)
{
LASSERT(src->pl_nr > 0);
- LINVRNT(dst->pl_owner == current);
- LINVRNT(src->pl_owner == current);
ENTRY;
list_move_tail(&page->cp_batch, &dst->pl_pages);
struct cl_page *page)
{
LASSERT(src->pl_nr > 0);
- LINVRNT(dst->pl_owner == current);
- LINVRNT(src->pl_owner == current);
ENTRY;
list_move(&page->cp_batch, &dst->pl_pages);
/**
* splice the cl_page_list, just as list head does
*/
-void cl_page_list_splice(struct cl_page_list *list, struct cl_page_list *head)
+void cl_page_list_splice(struct cl_page_list *src, struct cl_page_list *dst)
{
+#ifdef CONFIG_LUSTRE_DEBUG_LU_REF
struct cl_page *page;
struct cl_page *tmp;
- LINVRNT(list->pl_owner == current);
- LINVRNT(head->pl_owner == current);
-
ENTRY;
- cl_page_list_for_each_safe(page, tmp, list)
- cl_page_list_move(head, list, page);
+ cl_page_list_for_each_safe(page, tmp, src)
+ lu_ref_set_at(&page->cp_reference, &page->cp_queue_ref,
+ "queue", src, dst);
+#else
+ ENTRY;
+#endif
+ dst->pl_nr += src->pl_nr;
+ src->pl_nr = 0;
+ list_splice_tail_init(&src->pl_pages, &dst->pl_pages);
+
EXIT;
}
EXPORT_SYMBOL(cl_page_list_splice);
struct cl_page *page;
struct cl_page *temp;
- LINVRNT(plist->pl_owner == current);
ENTRY;
cl_page_list_for_each_safe(page, temp, plist) {
struct cl_page *page;
struct cl_page *temp;
- LINVRNT(plist->pl_owner == current);
ENTRY;
cl_page_list_for_each_safe(page, temp, plist)
{
struct cl_page *page;
- LINVRNT(plist->pl_owner == current);
cl_page_list_for_each(page, plist)
cl_page_assume(env, io, page);
{
struct cl_page *page;
- LINVRNT(plist->pl_owner == current);
ENTRY;
cl_page_list_for_each(page, plist)
cl_page_discard(env, io, page);
/**
* Add a page to the incoming page list of 2-queue.
*/
-void cl_2queue_add(struct cl_2queue *queue, struct cl_page *page)
+void cl_2queue_add(struct cl_2queue *queue, struct cl_page *page, bool get_ref)
{
- ENTRY;
- cl_page_list_add(&queue->c2_qin, page);
- EXIT;
+ cl_page_list_add(&queue->c2_qin, page, get_ref);
}
EXPORT_SYMBOL(cl_2queue_add);
*/
void cl_2queue_init_page(struct cl_2queue *queue, struct cl_page *page)
{
- ENTRY;
- cl_2queue_init(queue);
- cl_2queue_add(queue, page);
- EXIT;
+ ENTRY;
+ cl_2queue_init(queue);
+ cl_2queue_add(queue, page, true);
+ EXIT;
}
EXPORT_SYMBOL(cl_2queue_init_page);
}
EXPORT_SYMBOL(cl_req_attr_set);
-/*
- * cl_sync_io_end callback is issued as cl_sync_io completes and before
- * control of anchor reverts to model used by the caller of cl_sync_io_init()
- *
- * NOTE: called with spinlock on anchor->csi_waitq.lock
- */
-void cl_sync_io_end(const struct lu_env *env, struct cl_sync_io *anchor)
-{
- /* deprecated pending future removal */
-}
-EXPORT_SYMBOL(cl_sync_io_end);
-
/**
- * Initialize synchronous io wait anchor
+ * Initialize synchronous io wait \a anchor for \a nr pages with optional
+ * \a end handler.
+ * \param anchor owned by caller, initialzied here.
+ * \param nr number of pages initally pending in sync.
+ * \param end optional callback sync_io completion, can be used to
+ * trigger erasure coding, integrity, dedupe, or similar operation.
+ * \q end is called with a spinlock on anchor->csi_waitq.lock
*/
-void cl_sync_io_init(struct cl_sync_io *anchor, int nr,
- void (*end)(const struct lu_env *, struct cl_sync_io *))
+
+void cl_sync_io_init_notify(struct cl_sync_io *anchor, int nr,
+ struct cl_dio_aio *aio, cl_sync_io_end_t *end)
{
ENTRY;
memset(anchor, 0, sizeof(*anchor));
atomic_set(&anchor->csi_sync_nr, nr);
anchor->csi_sync_rc = 0;
anchor->csi_end_io = end;
- LASSERT(end != NULL);
+ anchor->csi_aio = aio;
EXIT;
}
-EXPORT_SYMBOL(cl_sync_io_init);
+EXPORT_SYMBOL(cl_sync_io_init_notify);
/**
* Wait until all IO completes. Transfer completion routine has to call
int cl_sync_io_wait(const struct lu_env *env, struct cl_sync_io *anchor,
long timeout)
{
- struct l_wait_info lwi = LWI_TIMEOUT_INTR(cfs_time_seconds(timeout),
- NULL, NULL, NULL);
- int rc;
+ int rc = 0;
ENTRY;
LASSERT(timeout >= 0);
- rc = l_wait_event(anchor->csi_waitq,
- atomic_read(&anchor->csi_sync_nr) == 0,
- &lwi);
- if (rc < 0) {
+ if (timeout > 0 &&
+ wait_event_idle_timeout(anchor->csi_waitq,
+ atomic_read(&anchor->csi_sync_nr) == 0,
+ cfs_time_seconds(timeout)) == 0) {
+ rc = -ETIMEDOUT;
CERROR("IO failed: %d, still wait for %d remaining entries\n",
rc, atomic_read(&anchor->csi_sync_nr));
+ }
- lwi = (struct l_wait_info) { 0 };
- (void)l_wait_event(anchor->csi_waitq,
- atomic_read(&anchor->csi_sync_nr) == 0,
- &lwi);
- } else {
+ wait_event_idle(anchor->csi_waitq,
+ atomic_read(&anchor->csi_sync_nr) == 0);
+ if (!rc)
rc = anchor->csi_sync_rc;
- }
+
/* We take the lock to ensure that cl_sync_io_note() has finished */
spin_lock(&anchor->csi_waitq.lock);
LASSERT(atomic_read(&anchor->csi_sync_nr) == 0);
}
EXPORT_SYMBOL(cl_sync_io_wait);
+#ifndef HAVE_AIO_COMPLETE
+static inline void aio_complete(struct kiocb *iocb, ssize_t res, ssize_t res2)
+{
+ if (iocb->ki_complete)
+ iocb->ki_complete(iocb, res, res2);
+}
+#endif
+
+static void cl_aio_end(const struct lu_env *env, struct cl_sync_io *anchor)
+{
+ struct cl_dio_aio *aio = container_of(anchor, typeof(*aio), cda_sync);
+ ssize_t ret = anchor->csi_sync_rc;
+
+ ENTRY;
+
+ /* release pages */
+ while (aio->cda_pages.pl_nr > 0) {
+ struct cl_page *page = cl_page_list_first(&aio->cda_pages);
+
+ cl_page_get(page);
+ cl_page_list_del(env, &aio->cda_pages, page);
+ cl_page_delete(env, page);
+ cl_page_put(env, page);
+ }
+
+ if (!aio->cda_no_aio_complete)
+ aio_complete(aio->cda_iocb, ret ?: aio->cda_bytes, 0);
+
+ EXIT;
+}
+
+struct cl_dio_aio *cl_aio_alloc(struct kiocb *iocb, struct cl_object *obj)
+{
+ struct cl_dio_aio *aio;
+
+ OBD_SLAB_ALLOC_PTR_GFP(aio, cl_dio_aio_kmem, GFP_NOFS);
+ if (aio != NULL) {
+ /*
+ * Hold one ref so that it won't be released until
+ * every pages is added.
+ */
+ cl_sync_io_init_notify(&aio->cda_sync, 1, aio, cl_aio_end);
+ cl_page_list_init(&aio->cda_pages);
+ aio->cda_iocb = iocb;
+ if (is_sync_kiocb(iocb))
+ aio->cda_no_aio_complete = 1;
+ else
+ aio->cda_no_aio_complete = 0;
+ cl_object_get(obj);
+ aio->cda_obj = obj;
+ }
+ return aio;
+}
+EXPORT_SYMBOL(cl_aio_alloc);
+
+void cl_aio_free(const struct lu_env *env, struct cl_dio_aio *aio)
+{
+ if (aio) {
+ cl_object_put(env, aio->cda_obj);
+ OBD_SLAB_FREE_PTR(aio, cl_dio_aio_kmem);
+ }
+}
+EXPORT_SYMBOL(cl_aio_free);
+
+
/**
* Indicate that transfer of a single page completed.
*/
LASSERT(atomic_read(&anchor->csi_sync_nr) > 0);
if (atomic_dec_and_lock(&anchor->csi_sync_nr,
&anchor->csi_waitq.lock)) {
+ struct cl_dio_aio *aio = NULL;
+
+ cl_sync_io_end_t *end_io = anchor->csi_end_io;
+
/*
* Holding the lock across both the decrement and
* the wakeup ensures cl_sync_io_wait() doesn't complete
* to immediately reclaim anchor when cl_sync_io_wait()
* completes.
*/
- wake_up_all_locked(&anchor->csi_waitq);
- anchor->csi_end_io(env, anchor);
+ wake_up_locked(&anchor->csi_waitq);
+ if (end_io)
+ end_io(env, anchor);
+
+ aio = anchor->csi_aio;
+
spin_unlock(&anchor->csi_waitq.lock);
- /* Can't access anchor any more */
+ /**
+ * For AIO (!is_sync_kiocb), we are responsible for freeing
+ * memory here. This is because we are the last user of this
+ * aio struct, whereas in other cases, we will call
+ * cl_sync_io_wait to wait after this, and so the memory is
+ * freed after that call.
+ */
+ if (aio && !is_sync_kiocb(aio->cda_iocb))
+ cl_aio_free(env, aio);
}
EXIT;
}
EXPORT_SYMBOL(cl_sync_io_note);
+
+
+int cl_sync_io_wait_recycle(const struct lu_env *env, struct cl_sync_io *anchor,
+ long timeout, int ioret)
+{
+ int rc = 0;
+
+ /*
+ * @anchor was inited as 1 to prevent end_io to be
+ * called before we add all pages for IO, so drop
+ * one extra reference to make sure we could wait
+ * count to be zero.
+ */
+ cl_sync_io_note(env, anchor, ioret);
+ /* Wait for completion of normal dio.
+ * This replaces the EIOCBQEUED return from the DIO/AIO
+ * path, and this is where AIO and DIO implementations
+ * split.
+ */
+ rc = cl_sync_io_wait(env, anchor, timeout);
+ /**
+ * One extra reference again, as if @anchor is
+ * reused we assume it as 1 before using.
+ */
+ atomic_add(1, &anchor->csi_sync_nr);
+
+ return rc;
+}
+EXPORT_SYMBOL(cl_sync_io_wait_recycle);
git://git.whamcloud.com / fs / lustre-release.git / blobdiff