LU-13386 llite: allow current readahead to exceed reservation

[fs/lustre-release.git] / lustre / llite / rw.c

diff --git a/lustre/llite/rw.c b/lustre/llite/rw.c
index 7d5453f..b6559b4 100644 (file)
--- a/lustre/llite/rw.c
+++ b/lustre/llite/rw.c
@@ -373,19 +373,30 @@ ll_read_ahead_pages(const struct lu_env *env, struct cl_io *io,
                        if (ra.cra_end_idx == 0 || ra.cra_end_idx < page_idx) {
                                pgoff_t end_idx;
 
+                               /*
+                                * Do not shrink ria_end_idx at any case until
+                                * the minimum end of current read is covered.
+                                *
+                                * Do not extend read lock accross stripe if
+                                * lock contention detected.
+                                */
+                               if (ra.cra_contention &&
+                                   page_idx > ria->ria_end_idx_min) {
+                                       ria->ria_end_idx = *ra_end;
+                                       break;
+                               }
+
                                cl_read_ahead_release(env, &ra);
 
                                rc = cl_io_read_ahead(env, io, page_idx, &ra);
                                if (rc < 0)
                                        break;
 
-                               /* Do not shrink ria_end_idx at any case until
-                                * the minimum end of current read is covered.
-                                * And only shrink ria_end_idx if the matched
-                                * LDLM lock doesn't cover more. */
-                               if (page_idx > ra.cra_end_idx ||
-                                   (ra.cra_contention &&
-                                    page_idx > ria->ria_end_idx_min)) {
+                                /*
+                                 * Only shrink ria_end_idx if the matched
+                                 * LDLM lock doesn't cover more.
+                                 */
+                               if (page_idx > ra.cra_end_idx) {
                                        ria->ria_end_idx = ra.cra_end_idx;
                                        break;
                                }
@@ -517,13 +528,15 @@ static void ll_readahead_handle_work(struct work_struct *wq)
        __u64 kms;
        int rc;
        pgoff_t eof_index;
+       struct ll_sb_info *sbi;
 
        work = container_of(wq, struct ll_readahead_work,
                            lrw_readahead_work);
-       fd = LUSTRE_FPRIVATE(work->lrw_file);
+       fd = work->lrw_file->private_data;
        ras = &fd->fd_ras;
        file = work->lrw_file;
        inode = file_inode(file);
+       sbi = ll_i2sbi(inode);
 
        env = cl_env_alloc(&refcheck, LCT_NOREF);
        if (IS_ERR(env))
@@ -556,7 +569,7 @@ static void ll_readahead_handle_work(struct work_struct *wq)
 
        ria->ria_end_idx = work->lrw_end_idx;
        pages = ria->ria_end_idx - ria->ria_start_idx + 1;
-       ria->ria_reserved = ll_ra_count_get(ll_i2sbi(inode), ria,
+       ria->ria_reserved = ll_ra_count_get(sbi, ria,
                                            ria_page_count(ria), pages_min);
 
        CDEBUG(D_READA,
@@ -620,6 +633,7 @@ out_put_env:
 out_free_work:
        if (ra_end_idx > 0)
                ll_ra_stats_inc_sbi(ll_i2sbi(inode), RA_STAT_ASYNC);
+       atomic_dec(&sbi->ll_ra_info.ra_async_inflight);
        ll_readahead_work_free(work);
 }
 
@@ -715,9 +729,12 @@ static int ll_readahead(const struct lu_env *env, struct cl_io *io,
 
        /* at least to extend the readahead window to cover current read */
        if (!hit && vio->vui_ra_valid &&
-           vio->vui_ra_start_idx + vio->vui_ra_pages > ria->ria_start_idx)
+           vio->vui_ra_start_idx + vio->vui_ra_pages > ria->ria_start_idx) {
                ria->ria_end_idx_min =
                        vio->vui_ra_start_idx + vio->vui_ra_pages - 1;
+               pages_min = vio->vui_ra_start_idx + vio->vui_ra_pages -
+                               ria->ria_start_idx;
+       }
 
        ria->ria_reserved = ll_ra_count_get(ll_i2sbi(inode), ria, pages,
                                            pages_min);
@@ -996,7 +1013,7 @@ static void ras_detect_read_pattern(struct ll_readahead_state *ras,
 
 void ll_ras_enter(struct file *f, loff_t pos, size_t count)
 {
-       struct ll_file_data *fd = LUSTRE_FPRIVATE(f);
+       struct ll_file_data *fd = f->private_data;
        struct ll_readahead_state *ras = &fd->fd_ras;
        struct inode *inode = file_inode(f);
        unsigned long index = pos >> PAGE_SHIFT;
@@ -1308,7 +1325,7 @@ int ll_writepages(struct address_space *mapping, struct writeback_control *wbc)
 
 struct ll_cl_context *ll_cl_find(struct file *file)
 {
-       struct ll_file_data *fd = LUSTRE_FPRIVATE(file);
+       struct ll_file_data *fd = file->private_data;
        struct ll_cl_context *lcc;
        struct ll_cl_context *found = NULL;
 
@@ -1327,7 +1344,7 @@ struct ll_cl_context *ll_cl_find(struct file *file)
 void ll_cl_add(struct file *file, const struct lu_env *env, struct cl_io *io,
               enum lcc_type type)
 {
-       struct ll_file_data *fd = LUSTRE_FPRIVATE(file);
+       struct ll_file_data *fd = file->private_data;
        struct ll_cl_context *lcc = &ll_env_info(env)->lti_io_ctx;
 
        memset(lcc, 0, sizeof(*lcc));
@@ -1344,7 +1361,7 @@ void ll_cl_add(struct file *file, const struct lu_env *env, struct cl_io *io,
 
 void ll_cl_remove(struct file *file, const struct lu_env *env)
 {
-       struct ll_file_data *fd = LUSTRE_FPRIVATE(file);
+       struct ll_file_data *fd = file->private_data;
        struct ll_cl_context *lcc = &ll_env_info(env)->lti_io_ctx;
 
        write_lock(&fd->fd_lock);
@@ -1357,7 +1374,7 @@ int ll_io_read_page(const struct lu_env *env, struct cl_io *io,
 {
        struct inode              *inode  = vvp_object_inode(page->cp_obj);
        struct ll_sb_info         *sbi    = ll_i2sbi(inode);
-       struct ll_file_data       *fd     = LUSTRE_FPRIVATE(file);
+       struct ll_file_data       *fd     = file->private_data;
        struct ll_readahead_state *ras    = &fd->fd_ras;
        struct cl_2queue          *queue  = &io->ci_queue;
        struct cl_sync_io         *anchor = NULL;
@@ -1453,7 +1470,7 @@ static int kickoff_async_readahead(struct file *file, unsigned long pages)
        struct ll_readahead_work *lrw;
        struct inode *inode = file_inode(file);
        struct ll_sb_info *sbi = ll_i2sbi(inode);
-       struct ll_file_data *fd = LUSTRE_FPRIVATE(file);
+       struct ll_file_data *fd = file->private_data;
        struct ll_readahead_state *ras = &fd->fd_ras;
        struct ll_ra_info *ra = &sbi->ll_ra_info;
        unsigned long throttle;
@@ -1468,7 +1485,8 @@ static int kickoff_async_readahead(struct file *file, unsigned long pages)
         * we do async readahead, allowing the user thread to do fast i/o.
         */
        if (stride_io_mode(ras) || !throttle ||
-           ras->ras_window_pages < throttle)
+           ras->ras_window_pages < throttle ||
+           atomic_read(&ra->ra_async_inflight) > ra->ra_async_max_active)
                return 0;
 
        if ((atomic_read(&ra->ra_cur_pages) + pages) > ra->ra_max_pages)
@@ -1480,6 +1498,7 @@ static int kickoff_async_readahead(struct file *file, unsigned long pages)
        /* ll_readahead_work_free() free it */
        OBD_ALLOC_PTR(lrw);
        if (lrw) {
+               atomic_inc(&sbi->ll_ra_info.ra_async_inflight);
                lrw->lrw_file = get_file(file);
                lrw->lrw_start_idx = start_idx;
                lrw->lrw_end_idx = end_idx;
@@ -1543,7 +1562,7 @@ int ll_readpage(struct file *file, struct page *vmpage)
 
        if (io == NULL) { /* fast read */
                struct inode *inode = file_inode(file);
-               struct ll_file_data *fd = LUSTRE_FPRIVATE(file);
+               struct ll_file_data *fd = file->private_data;
                struct ll_readahead_state *ras = &fd->fd_ras;
                struct lu_env  *local_env = NULL;
                struct vvp_page *vpg;