gru: fix bug in allocation of kernel contexts

Fix a bug in the assignment of GRU contexts used for kernel functions.  If
a sleep occurs on the wait for a semaphore, the thread could switch cpus
and allocate resources on the wrong blade.

Signed-off-by: Jack Steiner <steiner@sgi.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

1 files changed, 11 insertions, 3 deletions

diff --git a/drivers/misc/sgi-gru/grukservices.c b/drivers/misc/sgi-gru/grukservices.c
index c13342dda5fd..bfbf8fdbcd37 100644
--- a/drivers/misc/sgi-gru/grukservices.c
+++ b/drivers/misc/sgi-gru/grukservices.c
@@ -221,13 +221,21 @@ static int gru_free_kernel_contexts(void)
 static struct gru_blade_state *gru_lock_kernel_context(int blade_id)
 {
         struct gru_blade_state *bs;
+        int bid;
 
         STAT(lock_kernel_context);
-        bs = gru_base[blade_id];
+again:
+        bid = blade_id < 0 ? uv_numa_blade_id() : blade_id;
+        bs = gru_base[bid];
 
+        /* Handle the case where migration occured while waiting for the sema */
         down_read(&bs->bs_kgts_sema);
+        if (blade_id < 0 && bid != uv_numa_blade_id()) {
+                up_read(&bs->bs_kgts_sema);
+                goto again;
+        }
         if (!bs->bs_kgts || !bs->bs_kgts->ts_gru)
-                gru_load_kernel_context(bs, blade_id);
+                gru_load_kernel_context(bs, bid);
         return bs;
 
 }
@@ -256,7 +264,7 @@ static int gru_get_cpu_resources(int dsr_bytes, void **cb, void **dsr)
 
         BUG_ON(dsr_bytes > GRU_NUM_KERNEL_DSR_BYTES);
         preempt_disable();
-        bs = gru_lock_kernel_context(uv_numa_blade_id());
+        bs = gru_lock_kernel_context(-1);
         lcpu = uv_blade_processor_id();
         *cb = bs->kernel_cb + lcpu * GRU_HANDLE_STRIDE;
         *dsr = bs->kernel_dsr + lcpu * GRU_NUM_KERNEL_DSR_BYTES;