Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]
Groups > linux.kernel > #1535301
| From | James Simmons <jsimmons@infradead.org> |
|---|---|
| Newsgroups | linux.kernel |
| Subject | [PATCH 15/22] staging: lustre: rpc: increase bulk size |
| Date | 2016-12-03 02:10 +0100 |
| Message-ID | <sK6ZX-2TQ-7@gated-at.bofh.it> (permalink) |
| References | <sK6Qi-2BA-3@gated-at.bofh.it> |
| Organization | linux.* mail to news gateway |
From: Jinshan Xiong <jinshan.xiong@intel.com>
To make the ptlrpc be able to size 16MB IO
Signed-off-by: Jinshan Xiong <jinshan.xiong@intel.com>
Signed-off-by: Gu Zheng <gzheng@ddn.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-7990
Reviewed-on: http://review.whamcloud.com/19366
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
---
.../lustre/lustre/include/lustre/lustre_idl.h | 6 +++++-
drivers/staging/lustre/lustre/include/lustre_net.h | 8 ++++++--
drivers/staging/lustre/lustre/ptlrpc/wiretest.c | 2 ++
3 files changed, 13 insertions(+), 3 deletions(-)
diff --git a/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h b/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h
index b9f333b..20f55b7 100644
--- a/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h
+++ b/drivers/staging/lustre/lustre/include/lustre/lustre_idl.h
@@ -1683,8 +1683,12 @@ struct obd_ioobj {
__u32 ioo_bufcnt; /* number of niobufs for this object */
};
+/*
+ * NOTE: IOOBJ_MAX_BRW_BITS defines the _offset_ of the max_brw field in
+ * ioo_max_brw, NOT the maximum number of bits in PTLRPC_BULK_OPS_BITS.
+ * That said, ioo_max_brw is a 32-bit field so the limit is also 16 bits.
+ */
#define IOOBJ_MAX_BRW_BITS 16
-#define IOOBJ_TYPE_MASK ((1U << IOOBJ_MAX_BRW_BITS) - 1)
#define ioobj_max_brw_get(ioo) (((ioo)->ioo_max_brw >> IOOBJ_MAX_BRW_BITS) + 1)
#define ioobj_max_brw_set(ioo, num) \
do { (ioo)->ioo_max_brw = ((num) - 1) << IOOBJ_MAX_BRW_BITS; } while (0)
diff --git a/drivers/staging/lustre/lustre/include/lustre_net.h b/drivers/staging/lustre/lustre/include/lustre_net.h
index 2be135d..411eb0d 100644
--- a/drivers/staging/lustre/lustre/include/lustre_net.h
+++ b/drivers/staging/lustre/lustre/include/lustre_net.h
@@ -69,13 +69,17 @@
#define PTLRPC_MD_OPTIONS 0
/**
- * Max # of bulk operations in one request.
+ * log2 max # of bulk operations in one request: 2=4MB/RPC, 5=32MB/RPC, ...
* In order for the client and server to properly negotiate the maximum
* possible transfer size, PTLRPC_BULK_OPS_COUNT must be a power-of-two
* value. The client is free to limit the actual RPC size for any bulk
* transfer via cl_max_pages_per_rpc to some non-power-of-two value.
+ * NOTE: This is limited to 16 (=64GB RPCs) by IOOBJ_MAX_BRW_BITS.
*/
-#define PTLRPC_BULK_OPS_BITS 2
+#define PTLRPC_BULK_OPS_BITS 4
+#if PTLRPC_BULK_OPS_BITS > 16
+#error "More than 65536 BRW RPCs not allowed by IOOBJ_MAX_BRW_BITS."
+#endif
#define PTLRPC_BULK_OPS_COUNT (1U << PTLRPC_BULK_OPS_BITS)
/**
* PTLRPC_BULK_OPS_MASK is for the convenience of the client only, and
diff --git a/drivers/staging/lustre/lustre/ptlrpc/wiretest.c b/drivers/staging/lustre/lustre/ptlrpc/wiretest.c
index f50f487..e12eb83 100644
--- a/drivers/staging/lustre/lustre/ptlrpc/wiretest.c
+++ b/drivers/staging/lustre/lustre/ptlrpc/wiretest.c
@@ -1576,6 +1576,8 @@ void lustre_assert_wire_constants(void)
(long long)(int)offsetof(struct obd_ioobj, ioo_bufcnt));
LASSERTF((int)sizeof(((struct obd_ioobj *)0)->ioo_bufcnt) == 4, "found %lld\n",
(long long)(int)sizeof(((struct obd_ioobj *)0)->ioo_bufcnt));
+ LASSERTF(IOOBJ_MAX_BRW_BITS == 16, "found %lld\n",
+ (long long)IOOBJ_MAX_BRW_BITS);
/* Checks for union lquota_id */
LASSERTF((int)sizeof(union lquota_id) == 16, "found %lld\n",
--
1.7.1
Back to linux.kernel | Previous | Next — Previous in thread | Next in thread | Find similar | Unroll thread
[PATCH 00/22] Next batch of missing work for upstream client James Simmons <jsimmons@infradead.org> - 2016-12-03 02:00 +0100
[PATCH 06/22] staging: lustre: llog: reset llog bitmap James Simmons <jsimmons@infradead.org> - 2016-12-03 02:00 +0100
[PATCH 13/22] staging: lustre: statahead: set sai_index_wait with lli_sa_lock held James Simmons <jsimmons@infradead.org> - 2016-12-03 02:10 +0100
[PATCH 19/22] staging: lustre: llite: ll_dir_ioctl cleanup of redundant comparisons James Simmons <jsimmons@infradead.org> - 2016-12-03 02:10 +0100
[PATCH 17/22] staging: lustre: clio: remove mtime check in vvp_io_fault_start() James Simmons <jsimmons@infradead.org> - 2016-12-03 02:10 +0100
[PATCH 12/22] staging: lustre: libcfs: report hnode value for cfs_hash_putref James Simmons <jsimmons@infradead.org> - 2016-12-03 02:10 +0100
[PATCH 15/22] staging: lustre: rpc: increase bulk size James Simmons <jsimmons@infradead.org> - 2016-12-03 02:10 +0100
[PATCH 18/22] staging: lustre: import: don't reconnect during connect interpret James Simmons <jsimmons@infradead.org> - 2016-12-03 02:10 +0100
[PATCH 16/22] staging: lustre: llite: Invoke file_update_time in page_mkwrite James Simmons <jsimmons@infradead.org> - 2016-12-03 02:10 +0100
[PATCH 09/22] staging: lustre: llite: Add client mount opt to ignore suppress_pings James Simmons <jsimmons@infradead.org> - 2016-12-03 02:10 +0100
[PATCH 22/22] staging: lustre: libcfs: remove lnet upcall code James Simmons <jsimmons@infradead.org> - 2016-12-03 02:10 +0100
[PATCH 21/22] staging: lustre: remove set but unused variables James Simmons <jsimmons@infradead.org> - 2016-12-03 02:10 +0100
[PATCH 14/22] staging: lustre: obd: add callback for llog_cat_process_or_fork James Simmons <jsimmons@infradead.org> - 2016-12-03 02:20 +0100
Re: [PATCH 14/22] staging: lustre: obd: add callback for llog_cat_process_or_fork Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2016-12-06 11:10 +0100
[PATCH 07/22] staging: lustre: obdclass: lu_site_purge() to handle purge-all James Simmons <jsimmons@infradead.org> - 2016-12-03 02:20 +0100
[PATCH 20/22] staging: lustre: osc: set lock data for readahead lock James Simmons <jsimmons@infradead.org> - 2016-12-03 02:30 +0100
Re: [PATCH 00/22] Next batch of missing work for upstream client Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2016-12-06 11:20 +0100
csiph-web