linux-stable.git/drivers/block/rnbd, branch v5.13.2

block/rnbd: Remove all likely and unlikely

2021-05-03T17:00:11+00:00

The IO performance test with fio after removing the likely and
unlikely macros in all if-statement shows no performance drop.
They do not help for the performance of rnbd.

The fio test did random read on 32 rnbd devices and 64 processes.
Test environment:
- AMD Opteron(tm) Processor 6386 SE
- 125G memory
- kernel version: 5.4.86
- gcc version: gcc (Debian 8.3.0-6) 8.3.0
- Infiniband controller: InfiniBand: Mellanox Technologies MT26428
[ConnectX VPI PCIe 2.0 5GT/s - IB QDR / 10GigE] (rev b0)

before
read: IOPS=549k, BW=2146MiB/s
read: IOPS=544k, BW=2125MiB/s
read: IOPS=553k, BW=2158MiB/s
read: IOPS=535k, BW=2089MiB/s
read: IOPS=543k, BW=2122MiB/s
read: IOPS=552k, BW=2154MiB/s
average: IOPS=546k, BW=2132MiB/s

after
read: IOPS=556k, BW=2172MiB/s
read: IOPS=561k, BW=2191MiB/s
read: IOPS=552k, BW=2156MiB/s
read: IOPS=551k, BW=2154MiB/s
read: IOPS=562k, BW=2194MiB/s
-----------
average: IOPS=556k, BW=2173MiB/s

The IOPS and bandwidth got better slightly after removing
likely/unlikely. (IOPS= +1.8% BW= +1.9%) But we cannot make sure
that removing the likely/unlikely help the performance because it
depends on various situations. We only make sure that removing the
likely/unlikely does not drop the performance.

Signed-off-by: Gioh Kim 
Reviewed-by: Md Haris Iqbal 
Link: https://lore.kernel.org/r/20210428061359.206794-5-gi-oh.kim@ionos.com
Signed-off-by: Jens Axboe

block/rnbd-clt: Check the return value of the function rtrs_clt_query

2021-05-03T17:00:11+00:00

In case none of the paths are in connected state, the function
rtrs_clt_query returns an error. In such a case, error out since the
values in the rtrs_attrs structure would be garbage.

Fixes: f7a7a5c228d45 ("block/rnbd: client: main functionality")
Signed-off-by: Md Haris Iqbal 
Reviewed-by: Guoqing Jiang 
Signed-off-by: Jack Wang 
Signed-off-by: Gioh Kim 
Link: https://lore.kernel.org/r/20210428061359.206794-4-gi-oh.kim@ionos.com
Signed-off-by: Jens Axboe

block/rnbd: Fix style issues

2021-05-03T17:00:11+00:00

This patch fixes some style issues detected by scripts/checkpatch.pl
* Resolve spacing and tab issues
* Remove extra braces in rnbd_get_iu
* Use num_possible_cpus() instead of NR_CPUS in alloc_sess
* Fix the comments styling in rnbd_queue_rq

Signed-off-by: Dima Stepanov 
Signed-off-by: Gioh Kim 
Signed-off-by: Md Haris Iqbal 
Signed-off-by: Jack Wang 
Link: https://lore.kernel.org/r/20210428061359.206794-3-gi-oh.kim@ionos.com
Signed-off-by: Jens Axboe

block/rnbd-clt: Change queue_depth type in rnbd_clt_session to size_t

2021-05-03T17:00:11+00:00

The member queue_depth in the structure rnbd_clt_session is read from the
rtrs client side using the function rtrs_clt_query, which in turn is read
from the rtrs_clt structure. It should really be of type size_t.

Fixes: 90426e89f54db ("block/rnbd: client: private header with client structs and functions")
Signed-off-by: Md Haris Iqbal 
Reviewed-by: Guoqing Jiang 
Signed-off-by: Gioh Kim 
Link: https://lore.kernel.org/r/20210428061359.206794-2-gi-oh.kim@ionos.com
Signed-off-by: Jens Axboe

block/rnbd: Use strscpy instead of strlcpy

2021-04-20T14:59:04+00:00

During checkpatch analyzing the following warning message was found:
  WARNING:STRLCPY: Prefer strscpy over strlcpy - see:
  https://lore.kernel.org/r/CAHk-=wgfRnXz0W3D37d01q3JFkr_i_uTL=V6A6G1oUZcprmknw@mail.gmail.com/
Fix it by using strscpy calls instead of strlcpy.

Signed-off-by: Dima Stepanov 
Signed-off-by: Jack Wang 
Signed-off-by: Gioh Kim 
Reviewed-by: Chaitanya Kulkarni 
Link: https://lore.kernel.org/r/20210419073722.15351-20-gi-oh.kim@ionos.com
Signed-off-by: Jens Axboe

block/rnbd-clt-sysfs: Remove copy buffer overlap in rnbd_clt_get_path_name

2021-04-20T14:59:04+00:00

cppcheck report the following error:
  rnbd/rnbd-clt-sysfs.c:522:36: error: The variable 'buf' is used both
  as a parameter and as destination in snprintf(). The origin and
  destination buffers overlap. Quote from glibc (C-library)
  documentation
  (http://www.gnu.org/software/libc/manual/html_mono/libc.html#Formatted-Output-Functions):
  "If copying takes place between objects that overlap as a result of a
  call to sprintf() or snprintf(), the results are undefined."
  [sprintfOverlappingData]
Fix it by initializing the buf variable in the first snprintf call.

Fixes: 91f4acb2801c ("block/rnbd-clt: support mapping two devices")
Signed-off-by: Dima Stepanov 
Cc: Arnd Bergmann 
Signed-off-by: Jack Wang 
Signed-off-by: Gioh Kim 
Reviewed-by: Chaitanya Kulkarni 
Link: https://lore.kernel.org/r/20210419073722.15351-19-gi-oh.kim@ionos.com
Signed-off-by: Jens Axboe

block/rnbd-clt: Remove max_segment_size

2021-04-20T14:59:04+00:00

We always map with SZ_4K, so do not need max_segment_size.

Cc: Leon Romanovsky 
Cc: linux-rdma@vger.kernel.org
Signed-off-by: Jack Wang 
Reviewed-by: Md Haris Iqbal 
Signed-off-by: Gioh Kim 
Acked-by: Jason Gunthorpe 
Reviewed-by: Leon Romanovsky 
Link: https://lore.kernel.org/r/20210419073722.15351-18-gi-oh.kim@ionos.com
Signed-off-by: Jens Axboe

block/rnbd-clt: Generate kobject_uevent when the rnbd device state changes

2021-04-20T14:59:04+00:00

When an RTRS session state changes, the transport layer generates an event
to RNBD. Then RNBD will change the state of the RNBD client device
accordingly.

This commit add kobject_uevent when the RNBD device state changes. With
this udev rules can be configured to react accordingly.

Signed-off-by: Md Haris Iqbal 
Signed-off-by: Jack Wang 
Signed-off-by: Gioh Kim 
Link: https://lore.kernel.org/r/20210419073722.15351-17-gi-oh.kim@ionos.com
Signed-off-by: Jens Axboe

block/rnbd-srv: Remove unused arguments of rnbd_srv_rdma_ev

2021-04-20T14:59:04+00:00

struct rtrs_srv is not used when handling rnbd_srv_rdma_ev messages, so
cleaned up
rdma_ev function pointer in rtrs_srv_ops also is changed.

Cc: Leon Romanovsky 
Cc: linux-rdma@vger.kernel.org
Signed-off-by: Aleksei Marov 
Signed-off-by: Jack Wang 
Signed-off-by: Gioh Kim 
Reviewed-by: Chaitanya Kulkarni 
Acked-by: Jason Gunthorpe 
Reviewed-by: Leon Romanovsky 
Link: https://lore.kernel.org/r/20210419073722.15351-16-gi-oh.kim@ionos.com
Signed-off-by: Jens Axboe

block/rnbd-clt: Support polling mode for IO latency optimization

2021-04-20T14:59:04+00:00

RNBD can make double-queues for irq-mode and poll-mode.
For example, on 4-CPU system 8 request-queues are created,
4 for irq-mode and 4 for poll-mode.
If the IO has HIPRI flag, the block-layer will call .poll function
of RNBD. Then IO is sent to the poll-mode queue.
Add optional nr_poll_queues argument for map_devices interface.

To support polling of RNBD, RTRS client creates connections
for both of irq-mode and direct-poll-mode.

For example, on 4-CPU system it could've create 5 connections:
con[0] => user message (softirq cq)
con[1:4] => softirq cq

After this patch, it can create 9 connections:
con[0] => user message (softirq cq)
con[1:4] => softirq cq
con[5:8] => DIRECT-POLL cq

Cc: Leon Romanovsky 
Cc: linux-rdma@vger.kernel.org
Signed-off-by: Gioh Kim 
Signed-off-by: Jack Wang 
Acked-by: Jason Gunthorpe 
Reviewed-by: Leon Romanovsky 
Link: https://lore.kernel.org/r/20210419073722.15351-14-gi-oh.kim@ionos.com
Signed-off-by: Jens Axboe