Reduce max_sectors_kb to avoid crash on some special block devices.
Test-Parameters: alwaysuploadlogs envdefinitions=SLOW=yes,ENABLE_QUOTA=yes mdtfilesystemtype=ldiskfs mdsfilesystemtype=ldiskfs ostfilesystemtype=ldiskfs clientdistro=el7 ossdistro=el7 mdsdistro=el7 mdtcount=1 testlist=runtests
Signed-off-by: Andreas Dilger <andreas.dilger@intel.com>
Change-Id: Ifc7efd169f0db0b54c3f727fc65716c203930e29
Reviewed-on: http://review.whamcloud.com/16116
Reviewed-by: Bob Glossman <bob.glossman@intel.com>
Tested-by: Jenkins
Tested-by: Maloo <hpdd-maloo@intel.com>
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
if (newval == 0 || newval == ULLONG_MAX || end == buf)
goto subdevs;
- /* Don't increase IO request size limit past 32MB. It is about
- * 2x PTLRPC_MAX_BRW_SIZE, but that isn't in a public header. */
- if (newval > 32 * 1024) {
- newval = 32 * 1024;
+ /* Don't increase IO request size limit past 16MB. It is about
+ * PTLRPC_MAX_BRW_SIZE, but that isn't in a public header.
+ * Note that even though the block layer allows larger values,
+ * setting max_sectors_kb = 32768 causes crashes (LU-6974). */
+ if (newval > 16 * 1024) {
+ newval = 16 * 1024;
snprintf(buf, sizeof(buf), "%llu", newval);
}