rtrs_clt_sess is used for paths and not sessions on the client side. This
creates confusion so let's rename it to rtrs_clt_path. Also, rename
related variables and functions.
Coccinelle is used to do the transformations for most of the occurrences
and remaining ones were handled manually.
Link: https://lore.kernel.org/r/20220105180708.7774-4-jinpu.wang@ionos.com
Signed-off-by: Vaishali Thakkar <vaishali.thakkar@ionos.com>
Signed-off-by: Jack Wang <jinpu.wang@ionos.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
rtrs_srv_sess is used for paths and not sessions on the server side. This
creates confusion so let's rename it to rtrs_srv_path. Also, rename
related variables and functions.
Coccinelle is used to do the transformations for most of the occurrences
and remaining ones were handled manually.
Link: https://lore.kernel.org/r/20220105180708.7774-3-jinpu.wang@ionos.com
Signed-off-by: Vaishali Thakkar <vaishali.thakkar@ionos.com>
Signed-off-by: Jack Wang <jinpu.wang@ionos.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
With preemption enabled (CONFIG_DEBUG_PREEMPT=y), the following appeared
when rnbd client tries to map remote block device.
BUG: using smp_processor_id() in preemptible [00000000] code: bash/1733
caller is debug_smp_processor_id+0x17/0x20
CPU: 0 PID: 1733 Comm: bash Not tainted 5.16.0-rc1 #5
Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS rel-1.14.0-0-g155821a-rebuilt.opensuse.org 04/01/2014
Call Trace:
<TASK>
dump_stack_lvl+0x5d/0x78
dump_stack+0x10/0x12
check_preemption_disabled+0xe4/0xf0
debug_smp_processor_id+0x17/0x20
rtrs_clt_update_all_stats+0x3b/0x70 [rtrs_client]
rtrs_clt_read_req+0xc3/0x380 [rtrs_client]
? rtrs_clt_init_req+0xe3/0x120 [rtrs_client]
rtrs_clt_request+0x1a7/0x320 [rtrs_client]
? 0xffffffffc0ab1000
send_usr_msg+0xbf/0x160 [rnbd_client]
? rnbd_clt_put_sess+0x60/0x60 [rnbd_client]
? send_usr_msg+0x160/0x160 [rnbd_client]
? sg_alloc_table+0x27/0xb0
? sg_zero_buffer+0xd0/0xd0
send_msg_sess_info+0xe9/0x180 [rnbd_client]
? rnbd_clt_put_sess+0x60/0x60 [rnbd_client]
? blk_mq_alloc_tag_set+0x2ef/0x370
rnbd_clt_map_device+0xba8/0xcd0 [rnbd_client]
? send_msg_open+0x200/0x200 [rnbd_client]
rnbd_clt_map_device_store+0x3e5/0x620 [rnbd_client
To supress the calltrace, let's call get_cpu_ptr/put_cpu_ptr pair in
rtrs_clt_update_rdma_stats to disable preemption when accessing per-cpu
variable.
While at it, let's make the similar change in rtrs_clt_update_wc_stats.
And for rtrs_clt_inc_failover_cnt, though it was only called inside rcu
section, but it still can be preempted in case CONFIG_PREEMPT_RCU is
enabled, so change it to {get,put}_cpu_ptr pair either.
Link: https://lore.kernel.org/r/20211128133501.38710-1-guoqing.jiang@linux.dev
Signed-off-by: Guoqing Jiang <guoqing.jiang@linux.dev>
Reviewed-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
When testing with poll mode, it will fail and lead to warning below on
client side:
$ echo "sessname=bla path=gid:fe80::2:c903:4e:d0b3@gid:fe80::2:c903:8:ca17 device_path=/dev/nullb2 nr_poll_queues=-1" | \
sudo tee /sys/devices/virtual/rnbd-client/ctl/map_device
rnbd_client L597: Mapping device /dev/nullb2 on session bla, (access_mode: rw, nr_poll_queues: 8)
WARNING: CPU: 3 PID: 9886 at drivers/infiniband/core/cq.c:447 ib_cq_pool_get+0x26f/0x2a0 [ib_core]
The problem is in case of poll queue, we need to still call
ib_alloc_cq/ib_free_cq, we can't use cq_poll api for poll queue.
As both client and server use shared function from rtrs, set irq_con_num
to con_num on server side, which is number of total connection of the
session, this way we can differ if the rtrs_con requires pollqueue.
Following up patches will replace the duplicate code with helpers.
Link: https://lore.kernel.org/r/20210922125333.351454-4-haris.iqbal@ionos.com
Signed-off-by: Jack Wang <jinpu.wang@ionos.com>
Reviewed-by: Gioh Kim <gi-oh.kim@ionos.com>
Signed-off-by: Md Haris Iqbal <haris.iqbal@cloud.ionos.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
There are mis-match at counting inflight IO after changing the multipath
policy.
For example, we started fio test with round-robin policy and then we
changed the policy to min-inflight. IOs created under the RR policy is
finished under the min-inflight policy and inflight counter only
decreased. So the counter would be negative value. And also we started
fio test with min-inflight policy and changed the policy to the
round-robin. IOs created under the min-inflight policy increased the
inflight IO counter but the inflight IO counter was not decreased because
the policy was the round-robin when IO was finished.
So it should count IOs only if the IO is created under the min-inflight
policy. It should not care the policy when the IO is finished.
This patch adds a field mp_policy in struct rtrs_clt_io_req and stores the
multipath policy when an object of rtrs_clt_io_req is created. Then
rtrs-clt checks the mp_policy of only struct rtrs_clt_io_req instead of
the struct rtrs_clt.
Link: https://lore.kernel.org/r/20210806112112.124313-6-haris.iqbal@ionos.com
Signed-off-by: Gioh Kim <gi-oh.kim@ionos.com>
Signed-off-by: Jack Wang <jinpu.wang@ionos.com>
Reviewed-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Md Haris Iqbal <haris.iqbal@ionos.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
The IO performance test with fio after swapping the likely and unlikely
macros in all if-statement shows no difference. They do not help for the
performance of rtrs.
Thanks to Haakon Bugge for the test scenario.
The fio test did random read on 32 rnbd devices and 64 processes.
Test environment:
- Intel(R) Xeon(R) Gold 6130 CPU @ 2.10GHz
- 376G memory
- kernel version: 5.4.86
- gcc version: gcc (Debian 8.3.0-6) 8.3.0
- Infiniband controller: Mellanox Technologies MT27800 Family [ConnectX-5]
Test result:
- before swapping: IOPS=829k, BW=3239MiB/s
- after swapping: IOPS=829k, BW=3238MiB/s
- remove all (un)likely: IOPS=829k, BW=3238MiB/s
Link: https://lore.kernel.org/r/20210806112112.124313-5-haris.iqbal@ionos.com
Signed-off-by: Gioh Kim <gi-oh.kim@ionos.com>
Signed-off-by: Jack Wang <jinpu.wang@ionos.com>
Reviewed-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Md Haris Iqbal <haris.iqbal@ionos.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
On idle session, because we do not do signal for heartbeat, it will
overflow the send queue after sometime.
To avoid that, we need to enable the signal for heartbeat. To do that, add
a new member signal_interval in rtrs_path, which will set min of
queue_depth and SERVICE_CON_QUEUE_DEPTH, and track it for both heartbeat
and IO, so the sq queue full accounting is correct.
Fixes: b38041d50a ("RDMA/rtrs: Do not signal for heatbeat")
Link: https://lore.kernel.org/r/20210712060750.16494-4-jinpu.wang@ionos.com
Signed-off-by: Jack Wang <jinpu.wang@ionos.com>
Reviewed-by: Aleksei Marov <aleksei.marov@ionos.com>
Reviewed-by: Gioh Kim <gi-oh.kim@ionos.com>
Reviewed-by: Md Haris Iqbal <haris.iqbal@ionos.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
With fast memory registration in write path, we can reduce
the memory consumption by using less max_send_sge, support IO bigger
than 116 KB (29 segments * 4 KB) without splitting, and it also
make the IO path more symmetric.
To avoid some times MR reg failed, waiting for the invalidation to finish
before the new mr reg. Introduce a refcount, only finish the request
when both local invalidation and io reply are there.
Link: https://lore.kernel.org/r/20210621055340.11789-3-jinpu.wang@ionos.com
Signed-off-by: Jack Wang <jinpu.wang@cloud.ionos.com>
Signed-off-by: Md Haris Iqbal <haris.iqbal@ionos.com>
Signed-off-by: Dima Stepanov <dmitrii.stepanov@ionos.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
Currently rtrs when create_qp use a coarse numbers (bigger in general),
which leads to hardware create more resources which only waste memory with
no benefits.
For max_send_wr, we don't really need alway max_qp_wr size when creating
qp, reduce it to cq_size.
For max_recv_wr, cq_size is enough.
With the patch when sess_queue_depth=128, per session (2 paths) memory
consumption reduced from 188 MB to 65MB
When always_invalidate is enabled, we need send more wr, so treat it
special.
Fixes: 9cb8374804 ("RDMA/rtrs: server: main functionality")
Link: https://lore.kernel.org/r/20210614090337.29557-2-jinpu.wang@ionos.com
Signed-off-by: Jack Wang <jinpu.wang@cloud.ionos.com>
Reviewed-by: Md Haris Iqbal <haris.iqbal@cloud.ionos.com>
Signed-off-by: Gioh Kim <gi-oh.kim@ionos.com>
Reviewed-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
drivers/infiniband/ulp/rtrs/rtrs-clt.c:1786:19: warning: result of comparison of
constant 'MAX_SESS_QUEUE_DEPTH' (65536) with expression of type 'u16'
(aka 'unsigned short') is always false [-Wtautological-constant-out-of-range-compare]
To fix it, limit MAX_SESS_QUEUE_DEPTH to u16 max, which is 65535, and
drop the check in rtrs-clt, as it's the type u16 max.
Link: https://lore.kernel.org/r/20210531122835.58329-1-jinpu.wang@ionos.com
Signed-off-by: Jack Wang <jinpu.wang@ionos.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
sess->stats and sess->stats->pcpu_stats objects are freed
when sysfs entry is removed. If something wrong happens and
session is closed before sysfs entry is created,
sess->stats and sess->stats->pcpu_stats objects are not freed.
This patch adds freeing of them at three places:
1. When client uses wrong address and session creation fails.
2. When client fails to create a sysfs entry.
3. When client adds wrong address via sysfs add_path.
Fixes: 215378b838 ("RDMA/rtrs: client: sysfs interface functions")
Link: https://lore.kernel.org/r/20210528113018.52290-21-jinpu.wang@ionos.com
Signed-off-by: Gioh Kim <gi-oh.kim@ionos.com>
Signed-off-by: Jack Wang <jinpu.wang@ionos.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
The queue_depth is a module parameter for rtrs_server. It is used on the
client side to determing the queue_depth of the request queue for the RNBD
virtual block device.
During a reconnection event for an already mapped device, in case the
rtrs_server module queue_depth has changed, fail the reconnect attempt.
Also stop further auto reconnection attempts. A manual reconnect via
sysfs has to be triggerred.
Fixes: 6a98d71dae ("RDMA/rtrs: client: main functionality")
Link: https://lore.kernel.org/r/20210528113018.52290-20-jinpu.wang@ionos.com
Signed-off-by: Md Haris Iqbal <haris.iqbal@ionos.com>
Signed-off-by: Gioh Kim <gi-oh.kim@ionos.com>
Signed-off-by: Jack Wang <jinpu.wang@ionos.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
When closing a session, currently the rtrs_srv_stats object in the
closing session is freed by kobject release. But if it failed
to create a session by various reasons, it must free the rtrs_srv_stats
object directly because kobject is not created yet.
This problem is found by kmemleak as below:
1. One client machine maps /dev/nullb0 with session name 'bla':
root@test1:~# echo "sessname=bla path=ip:192.168.122.190 \
device_path=/dev/nullb0" > /sys/devices/virtual/rnbd-client/ctl/map_device
2. Another machine failed to create a session with the same name 'bla':
root@test2:~# echo "sessname=bla path=ip:192.168.122.190 \
device_path=/dev/nullb1" > /sys/devices/virtual/rnbd-client/ctl/map_device
-bash: echo: write error: Connection reset by peer
3. The kmemleak on server machine reported an error:
unreferenced object 0xffff888033cdc800 (size 128):
comm "kworker/2:1", pid 83, jiffies 4295086585 (age 2508.680s)
hex dump (first 32 bytes):
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................
backtrace:
[<00000000a72903b2>] __alloc_sess+0x1d4/0x1250 [rtrs_server]
[<00000000d1e5321e>] rtrs_srv_rdma_cm_handler+0xc31/0xde0 [rtrs_server]
[<00000000bb2f6e7e>] cma_ib_req_handler+0xdc5/0x2b50 [rdma_cm]
[<00000000e896235d>] cm_process_work+0x2d/0x100 [ib_cm]
[<00000000b6866c5f>] cm_req_handler+0x11bc/0x1c40 [ib_cm]
[<000000005f5dd9aa>] cm_work_handler+0xe65/0x3cf2 [ib_cm]
[<00000000610151e7>] process_one_work+0x4bc/0x980
[<00000000541e0f77>] worker_thread+0x78/0x5c0
[<00000000423898ca>] kthread+0x191/0x1e0
[<000000005a24b239>] ret_from_fork+0x3a/0x50
Fixes: 39c2d639ca ("RDMA/rtrs-srv: Set .release function for rtrs srv device during device init")
Link: https://lore.kernel.org/r/20210528113018.52290-18-jinpu.wang@ionos.com
Signed-off-by: Gioh Kim <gi-oh.kim@ionos.com>
Signed-off-by: Md Haris Iqbal <haris.iqbal@ionos.com>
Signed-off-by: Jack Wang <jinpu.wang@ionos.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
If two clients try to use the same session name, rtrs-server generates a
kernel error that it failed to create the sysfs because the filename
is duplicated.
This patch adds code to check if there already exists the same session
name with the different UUID. If a client tries to add more session,
it sends the UUID and the session name. Therefore it is ok if there is
already same session name with the same UUID. The rtrs-server must fail
only-if there is the same session name with the different UUID.
Link: https://lore.kernel.org/r/20210528113018.52290-17-jinpu.wang@ionos.com
Signed-off-by: Gioh Kim <gi-oh.kim@ionos.com>
Signed-off-by: Aleksei Marov <aleksei.marov@ionos.com>
Signed-off-by: Jack Wang <jinpu.wang@ionos.com>
Signed-off-by: Md Haris Iqbal <haris.iqbal@ionos.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
When re-connecting, it resets hb_missed_max to 0.
Before the first re-connecting, client will trigger re-connection
when it gets hb-ack more than 5 times. But after the first
re-connecting, clients will do re-connection whenever it does
not get hb-ack because hb_missed_max is 0.
There is no need to reset hb_missed_max when re-connecting.
hb_missed_max should be kept until closing the session.
Fixes: c0894b3ea6 ("RDMA/rtrs: core: lib functions shared between client and server modules")
Link: https://lore.kernel.org/r/20210528113018.52290-16-jinpu.wang@ionos.com
Signed-off-by: Gioh Kim <gi-oh.kim@ionos.com>
Signed-off-by: Jack Wang <jinpu.wang@ionos.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
ids_inflight is used to track the inflight IOs. But the use of atomic_t
variable can cause performance drops and can also become a performance
bottleneck.
This commit replaces the use of atomic_t with a percpu_ref structure. The
advantage it offers is, it doesn't check if the reference has fallen to 0,
until the user explicitly signals it to; and that is done by the
percpu_ref_kill() function call. After that, the percpu_ref structure
behaves like an atomic_t and for every put call, checks whether the
reference has fallen to 0 or not.
rtrs_srv_stats_rdma_to_str shows the count of ids_inflight as 0
for user-mode tools not to be confused.
Fixes: 9cb8374804 ("RDMA/rtrs: server: main functionality")
Link: https://lore.kernel.org/r/20210528113018.52290-14-jinpu.wang@ionos.com
Signed-off-by: Md Haris Iqbal <haris.iqbal@ionos.com>
Signed-off-by: Jack Wang <jinpu.wang@ionos.com>
Signed-off-by: Gioh Kim <gi-oh.kim@ionos.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
When get_next_path_min_inflight is called to select the next path, it
iterates over the list of available rtrs_clt_sess (paths). It then reads
the number of inflight IOs for that path to select one which has the least
inflight IO.
But it may so happen that rtrs_clt_sess (path) is no longer in the
connected state because closing or error recovery paths can change the status
of the rtrs_clt_Sess.
For example, the client sent the heart-beat and did not get the
response, it would change the session status and stop IO processing.
The added checking of this patch can prevent accessing the broken path
and generating duplicated error messages.
It is ok if the status is changed after checking the status because
the error recovery path does not free memory and only tries to
reconnection. And also it is ok if the session is closed after checking
the status because closing the session changes the session status and
flush all IO beforing free memory. If the session is being accessed for
IO processing, the closing session will wait.
Fixes: 6a98d71dae ("RDMA/rtrs: client: main functionality")
Link: https://lore.kernel.org/r/20210528113018.52290-13-jinpu.wang@ionos.com
Signed-off-by: Md Haris Iqbal <haris.iqbal@ionos.com>
Reviewed-by: Gioh Kim <gi-oh.kim@ionos.com>
Signed-off-by: Gioh Kim <gi-oh.kim@ionos.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
