LU-12142 clio: fix hang on urgent cached pages

[fs/lustre-release.git] / lustre / osc / osc_page.c

diff --git a/lustre/osc/osc_page.c b/lustre/osc/osc_page.c
index 0f009b3..065bebe 100644 (file)
--- a/lustre/osc/osc_page.c
+++ b/lustre/osc/osc_page.c
@@ -794,6 +794,13 @@ static int osc_lru_alloc(const struct lu_env *env, struct client_obd *cli,
                        break;
                if (rc > 0)
                        continue;
+               /* IO issued by readahead, don't try hard */
+               if (oio->oi_is_readahead) {
+                       if (atomic_long_read(cli->cl_lru_left) > 0)
+                               continue;
+                       rc = -EBUSY;
+                       break;
+               }
 
                cond_resched();
                rc = l_wait_event_abortable(
@@ -826,16 +833,23 @@ unsigned long osc_lru_reserve(struct client_obd *cli, unsigned long npages)
        unsigned long reserved = 0;
        unsigned long max_pages;
        unsigned long c;
+       int rc;
 
-       /* reserve a full RPC window at most to avoid that a thread accidentally
-        * consumes too many LRU slots */
-       max_pages = cli->cl_max_pages_per_rpc * cli->cl_max_rpcs_in_flight;
-       if (npages > max_pages)
-               npages = max_pages;
-
+again:
        c = atomic_long_read(cli->cl_lru_left);
        if (c < npages && osc_lru_reclaim(cli, npages) > 0)
                c = atomic_long_read(cli->cl_lru_left);
+
+       if (c < npages) {
+               /*
+                * Trigger writeback in the hope some LRU slot could
+                * be freed.
+                */
+               rc = ptlrpcd_queue_work(cli->cl_writeback_work);
+               if (rc)
+                       return 0;
+       }
+
        while (c >= npages) {
                if (c == atomic_long_cmpxchg(cli->cl_lru_left, c, c - npages)) {
                        reserved = npages;
@@ -843,6 +857,16 @@ unsigned long osc_lru_reserve(struct client_obd *cli, unsigned long npages)
                }
                c = atomic_long_read(cli->cl_lru_left);
        }
+
+       if (reserved != npages) {
+               cond_resched();
+               rc = l_wait_event_abortable(
+                       osc_lru_waitq,
+                       atomic_long_read(cli->cl_lru_left) > 0);
+               goto again;
+       }
+
+       max_pages = cli->cl_max_pages_per_rpc * cli->cl_max_rpcs_in_flight;
        if (atomic_long_read(cli->cl_lru_left) < max_pages) {
                /* If there aren't enough pages in the per-OSC LRU then
                 * wake up the LRU thread to try and clear out space, so