Testing multi-threaded single shard file write performance has shown
the inode mutex to be a limiting factor when using the
generic_file_write_iter function. To work around this bottle neck, this
change replaces the locked version of that call with the lock less
version, specifically, __generic_file_write_iter.
In order to maintain posix consistency, Lustre must now employ it's
own locking mechanism in the higher layers. Currently writes are
protected using the lli_write_mutex in the ll_inode_info structure.
To protect against simultaneous write and truncate operations, since
we no longer take the inode mutex during writes, we must down the
lli_trunc_sem semaphore.
Unfortunately, this change by itself does not garner any performance
benefits. Using FIO on a single machine with 32 GB of RAM, write
performance tests were ran with and without this change applied; the
results are below:
+---------+-----------+---------+--------+--------+
| fio v2.0.13 | Write Bandwidth (KB/s) |
+---------+-----------+---------+--------+--------+
| # Tasks | GB / Task | Test 1 | Test 2 | Test 3 |
+---------+-----------+---------+--------+--------+
| 1 | 64 | 452446 | 454623 | 457653 |
| 2 | 32 | 850318 | 565373 | 602498 |
| 4 | 16 |
1058900 | 463546 | 529107 |
| 8 | 8 |
1026300 | 468190 | 576451 |
| 16 | 4 |
1065500 | 503160 | 462902 |
| 32 | 2 |
1068600 | 462228 | 466963 |
| 64 | 1 | 991830 | 556618 | 557863 |
+---------+-----------+---------+--------+--------+
* Test 1: Lustre client running
04ec54f. File per process write
workload. This test was used as a baseline for what we
_could_ achieve in the single shared file tests if the
bottle necks were removed.
* Test 2: Lustre client running
04ec54f. Single shared file
workload, each task writing to a unique region.
* Test 3: Lustre client running
04ec54f + this patch. Single shared
file workload, each task writing to a unique region.
In order to garner any real performance benefits out of a single
shared file workload, the lli_write_mutex needs to be broken up into a
range lock. That would allow write operations to unique regions of a
file to be executed concurrently. This work is left to be done in a
follow up patch.
Signed-off-by: Prakash Surya <surya1@llnl.gov>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-1669
Reviewed-on: http://review.whamcloud.com/6672
Reviewed-by: Lai Siyao <lai.siyao@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Jinshan Xiong <jinshan.xiong@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
io = vvp_env_io(env)->vui_cl.cis_io;
LASSERT(io);
- /* 0. Need locking between buffered and direct access. and race with
- * size changing by concurrent truncates and writes.
- * 1. Need inode mutex to operate transient pages.
- */
- if (iov_iter_rw(iter) == READ)
- inode_lock(inode);
-
LASSERT(obj->vob_transient_pages == 0);
while (iov_iter_count(iter)) {
struct page **pages;
}
out:
LASSERT(obj->vob_transient_pages == 0);
- if (iov_iter_rw(iter) == READ)
- inode_unlock(inode);
if (tot_bytes > 0) {
struct vvp_io *vio = vvp_env_io(env);
CDEBUG(D_VFSTRACE, "write: [%lli, %lli)\n", pos, pos + (long long)cnt);
- if (!vio->vui_iter) /* from a temp io in ll_cl_init(). */
+ if (!vio->vui_iter) {
+ /* from a temp io in ll_cl_init(). */
result = 0;
- else
- result = generic_file_write_iter(vio->vui_iocb, vio->vui_iter);
+ } else {
+ /*
+ * When using the locked AIO function (generic_file_aio_write())
+ * testing has shown the inode mutex to be a limiting factor
+ * with multi-threaded single shared file performance. To get
+ * around this, we now use the lockless version. To maintain
+ * consistency, proper locking to protect against writes,
+ * trucates, etc. is handled in the higher layers of lustre.
+ */
+ bool lock_node = !IS_NOSEC(inode);
+
+ if (lock_node)
+ inode_lock(inode);
+ result = __generic_file_write_iter(vio->vui_iocb,
+ vio->vui_iter);
+ if (lock_node)
+ inode_unlock(inode);
+
+ if (result > 0 || result == -EIOCBQUEUED)
+ result = generic_write_sync(vio->vui_iocb, result);
+ }
if (result > 0) {
result = vvp_io_write_commit(env, io);
return 0;
}
-static void vvp_transient_page_verify(const struct cl_page *page)
-{
- struct inode *inode = vvp_object_inode(page->cp_obj);
-
- LASSERT(!inode_trylock(inode));
-}
-
static int vvp_transient_page_own(const struct lu_env *env,
const struct cl_page_slice *slice,
struct cl_io *unused, int nonblock)
{
- vvp_transient_page_verify(slice->cpl_page);
return 0;
}
const struct cl_page_slice *slice,
struct cl_io *unused)
{
- vvp_transient_page_verify(slice->cpl_page);
}
static void vvp_transient_page_unassume(const struct lu_env *env,
const struct cl_page_slice *slice,
struct cl_io *unused)
{
- vvp_transient_page_verify(slice->cpl_page);
}
static void vvp_transient_page_disown(const struct lu_env *env,
const struct cl_page_slice *slice,
struct cl_io *unused)
{
- vvp_transient_page_verify(slice->cpl_page);
}
static void vvp_transient_page_discard(const struct lu_env *env,
{
struct cl_page *page = slice->cpl_page;
- vvp_transient_page_verify(slice->cpl_page);
-
/*
* For transient pages, remove it from the radix tree.
*/
const struct cl_page_slice *slice,
int ioret)
{
- vvp_transient_page_verify(slice->cpl_page);
}
static void vvp_transient_page_fini(const struct lu_env *env,
struct vvp_object *clobj = cl2vvp(clp->cp_obj);
vvp_page_fini_common(vpg);
- LASSERT(!inode_trylock(clobj->vob_inode));
clobj->vob_transient_pages--;
}
} else {
struct vvp_object *clobj = cl2vvp(obj);
- LASSERT(!inode_trylock(clobj->vob_inode));
cl_page_slice_add(page, &vpg->vpg_cl, obj, index,
&vvp_transient_page_ops);
clobj->vob_transient_pages++;
static inline int cl_page_invariant(const struct cl_page *pg)
{
- /*
- * Page invariant is protected by a VM lock.
- */
- LINVRNT(cl_page_is_vmlocked(NULL, pg));
-
return cl_page_in_use_noref(pg);
}
(const struct lu_env *,
const struct cl_page_slice *, int), ioret);
if (anchor) {
- LASSERT(cl_page_is_vmlocked(env, pg));
LASSERT(pg->cp_sync_io == anchor);
pg->cp_sync_io = NULL;
}
projects / cascardo / linux.git / commitdiff