xen/grant-table: log the lack of grants
authorWengang Wang <wen.gang.wang@oracle.com>
Tue, 18 Jul 2017 07:40:35 +0000 (09:40 +0200)
committerJuergen Gross <jgross@suse.com>
Sun, 23 Jul 2017 06:09:43 +0000 (08:09 +0200)
log a message when we enter this situation:
1) we already allocated the max number of available grants from hypervisor
and
2) we still need more (but the request fails because of 1)).

Sometimes the lack of grants causes IO hangs in xen_blkfront devices.
Adding this log would help debuging.

Signed-off-by: Wengang Wang <wen.gang.wang@oracle.com>
Reviewed-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Reviewed-by: Junxiao Bi <junxiao.bi@oracle.com>
Reviewed-by: Juergen Gross <jgross@suse.com>
Signed-off-by: Juergen Gross <jgross@suse.com>
drivers/xen/grant-table.c

index d6786b87e13b2392c366cfa807c35cf2401c8387..2c6a9114d332c74a85e2d679b9c2672a7447e188 100644 (file)
@@ -42,6 +42,7 @@
 #include <linux/delay.h>
 #include <linux/hardirq.h>
 #include <linux/workqueue.h>
+#include <linux/ratelimit.h>
 
 #include <xen/xen.h>
 #include <xen/interface/xen.h>
@@ -1072,8 +1073,14 @@ static int gnttab_expand(unsigned int req_entries)
        cur = nr_grant_frames;
        extra = ((req_entries + (grefs_per_grant_frame-1)) /
                 grefs_per_grant_frame);
-       if (cur + extra > gnttab_max_grant_frames())
+       if (cur + extra > gnttab_max_grant_frames()) {
+               pr_warn_ratelimited("xen/grant-table: max_grant_frames reached"
+                                   " cur=%u extra=%u limit=%u"
+                                   " gnttab_free_count=%u req_entries=%u\n",
+                                   cur, extra, gnttab_max_grant_frames(),
+                                   gnttab_free_count, req_entries);
                return -ENOSPC;
+       }
 
        rc = gnttab_map(cur, cur + extra - 1);
        if (rc == 0)