When the size of input buffer is larger than 16 and not divisible
by 16, the final line will keep part of the characters of the last
line. For example, if I want to dump "spdk dump 16 more chars",the
output is:
00000000 "hexadecimal" spdk dump 16 mor
00000010 "hexadecimal" e charsmp 16 mor
But the correct should be:
00000000 "hexadecimal" spdk dump 16 mor
00000010 "hexadecimal" e chars
Signed-off-by: Richael Zhuang <richael.zhuang@arm.com>
Change-Id: Iaf3f33c5ce68920ada83c59277f89a547e8030d7
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/1739
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
We should be giving completions for all requests when we destroy a qpair.
Signed-off-by: Seth Howell <seth.howell@intel.com>
Change-Id: I802f5120f2e8289aa825872f8085ac21b5fce0f3
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/1756
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom CI
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Paul Luse <paul.e.luse@intel.com>
That is done to make sure that scenario described in github
issue #1292 won't happen
Change-Id: Ie2ad001da701e25ef984ae57da850fb84d51b734
Signed-off-by: Alexey Marchuk <alexeymar@mellanox.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/1771
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: <dongx.yi@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
In some situations we may get a completion of RDMA_RECV before
completion of RDMA_SEND and this can lead to a bug described in #1292
To avoid such situations we must complete nvme_request only when
we received both RMDA_RECV and RDMA_SEND completions.
Add a new field to spdk_nvme_rdma_req to store response idx -
it is used to complete nvme request when RDMA_RECV was completed
before RDMA_SEND
Repost RDMA_RECV when both RDMA_SEND and RDMA_RECV are completed
Side changes: change type of spdk_nvme_rdma_req::id to uint16_t,
repack struct nvme_rdma_qpair
Fixes#1292
Change-Id: Ie51fbbba425acf37c306c5af031479bc9de08955
Signed-off-by: Alexey Marchuk <alexeymar@mellanox.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/1770
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: <dongx.yi@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Add scsi_blk_init function to replace the common
code of task initlization. The same to process_vq
function.
Change-Id: Ied0582ad7f087990c581e6e7aacde6ec201964dc
Signed-off-by: Jin Yu <jin.yu@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/1613
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Xiaodong Liu <xiaodong.liu@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
NVMe ctrlr init state machine shall be async whenever possible so it
is not blocking other code from processing. It can result in deadlock
when cmd producer and consumer are sharing the same thread.
This patch is making identify active ns async by introducing new
state to wait for completions.
Signed-off-by: Jacek Kalwas <jacek.kalwas@intel.com>
Change-Id: I346d35bab4733d3941e023602854fdd5b1ef23b5
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/1463
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Community-CI: Broadcom CI
It is a prework for changes related to ctrlr init state machine.
Signed-off-by: Jacek Kalwas <jacek.kalwas@intel.com>
Change-Id: If289580f65ae27468b659a7ea07a4e4298876e77
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/1489
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Community-CI: Broadcom CI
The timeout is designed for REVERT action, the REVERT needs
even several minutes based on different drives. Due to the
REVERT is low level API, it does need a response from the drive,
even we returned the timeout to the user, the drive is still in
the REVERT state, and it can't be used anymore, so the timeout
here doesn't help anything, just remove it.
Change-Id: I8f5119630d52f40e5bacea8ef8f3c771ee001aeb
Signed-off-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/1665
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Also rename them to security send/recv to reflect the fact.
Change-Id: Icdeb7f15849a9d1aacf5936c5954bb39875f4cd9
Signed-off-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/1651
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Function opal_cmd_finalize() is OPAL command related, it's
part of the OPAL command construction, and opal_send_recv()
is low level security send/receive channel, so it's better
to use them separately.
Change-Id: I0648585726d5caa7b37a4bc6783bcb870c028bd2
Signed-off-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/1650
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Previously the OPAL library uses the global session which
is OK, because the drive can only support 1 session, for
now, we can change the code to use allocated session, this
can make the library to be used in asynchronous way.
Change-Id: Ie033f905bd41c1171b8222d59925dc25729df84e
Signed-off-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/1649
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
The function inside opal_finalize_and_send() will be executed
synchronously, so remove the callback will make the code more
clear. Also rename the completion function with "_done" suffix.
No code logic change from this patch.
Change-Id: I03c5875457e52009768410ad29a89730a7df1c8b
Signed-off-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/1648
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Currently run count of poller has been incremented per execution.
It will be helpful for us to know how poller is busy by adding busy
count which is incremented only when some work is done.
spdk_thread_poll() has used the same timestamp in it, and so this is
the maximum we can do for now.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I0adfbf9a62c959499978124ecc97d377c96c3769
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/1713
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Set 5 seconds timeout to wait until exiting thread is exited into
spdk_thread_poll(). After the timeout, collect error log and then
move the thread to exited forcefully.
Add necessary unit test case accordingly.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: Ied8f58a2023a3bbe098530810fd3288bef93c3e7
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/1644
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Extract _spdk_thread_exit() from spdk_thread_exit() and
_spdk_thread_poll() calls _spdk_thread_exit() if the thread is in
the exiting state. spdk_thread_exit() changes to move the state to
the exiting state. The spdk_thread_poll() loop will end after the
thread moves to the exited state because the caller of
spdk_thread_poll() will check if the thread is in the exited state,
and break the loop if true.
If the user does not call spdk_thread_exit() explicitly, the reactor
has to terminate all existing threads at its shutdown. In this case,
multiple threads may have some dependency to release I/O channels or
unregister pollers. So the reactor has the large two loops, the first
loop calls spdk_thread_exit() on all threads, the second loop calls
spdk_thread_destroy() if exited or spdk_thread_poll() otherwise for
each thread until all threads are destroyed.
Besides, change the return value of spdk_thread_exit() to return
always 0. Keep it for ABI compatibility. Change ERRLOG to INFOLOG
for _spdk_thread_exit() because it is called repeatedly now. Remove
the check of I/O reference count from _spdk_thread_exit() because
_free_thread() cannot free I/O channel. Refine the unit test
accordingly.
Fixes issue #1288.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: Iee5fb984a96bfac53110fe991dd994ded31dffa4
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/1423
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Add enum spdk_thread_state made of RUNNING, EXITING, and EXITED, and
the current state to struct spdk_thread.
The state EXITING is not actually used in this patch yet.
Replace the flag exit simply by the state EXITED.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I6e5dc7184d50ae6d00e6ba00f5e2cf6045e5d48d
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/1630
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Change spdk_thread_exit() to nest a wrapper function, spdk_thread_exit()
and a static function made of function body. Include the check if the
state is running into the wrapper function.
These will make the following patches easier.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I81a57407acb772cd869819f2fac6665f61935369
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/1642
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Following patches will require thread termination so that thread_exit()
will move thread to exiting, thread_poll() will move thread from
exiting to exited, thread_is_exited() will detect the threadis exited,
and then call thread_destroy().
So change all places as a preparation.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I6b2e8aee5ed7cd160a88b4c9aaed7d90bd9dac07
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/1640
Community-CI: Broadcom CI
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Previously the caller had to check if thread is not exited when
it calls spdk_thread_exit().
Subsequent patches will change return type of spdk_thread_exit()
to void, and so include the check int spdk_thread_exit() in this
patch.
If spdk_thread_exit() is called when the thread is already exited,
collect INFOLOG and return normally.
This will make the next patch a little easier.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I8b94261575e770485b33c0b37e76e770b77b417c
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/1639
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Following the idea of thread CPU stats, add reactor CPU stats.
Reactor CPU stats accumulates run time of spdk_thread_poll() calls
to idle TSC or busy TSC according to their return codes.
Add necessary unit tests.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Signed-off-by: Maciej Szwed <maciej.szwed@intel.com>
Change-Id: I1a1391e79d74387c68f1651a61c8900e4c6faf66
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/1501
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
The code used to do this but it was removed when the buffering was
shifted down to the posix layer. Add a way for users of sockets
to still properly size the buffers.
This also means that by default, the receive buffering is not enabled
on sockets. That matches the behavior of the previous release.
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Change-Id: I20ce875be2efd841fe3a900047b4655a317d7799
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/1560
Community-CI: Broadcom CI
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
It's actually faster to process them until you run out of data.
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Change-Id: I9e81babdb9bdc405a8dbf03b2f701fe50bcc70f6
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/1559
Community-CI: Broadcom CI
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Currently nvme_completion_poll_status object is allocated using
malloc, so it may cotnain some garbage. In some scenarious
nvme_completion_poll_cb can be triggered before we enter
spdk_nvme_wait_for_completion_*. In that case status object
will be freed by nvme_completion_poll_cb if it contains a
garbage in `timed_out` field. Later spdk_nvme_wait_for_completion
will work with already freed memory.
Fix - allocate nvme_completion_poll_status object using
calloc and explicitly zerofy it before usage
Fixes#1292
Change-Id: Iac39653a6cd102471de16e65814f0760bbeda7d9
Signed-off-by: Alexey Marchuk <alexeymar@mellanox.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/1373
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: <dongx.yi@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
This API will allow us to simplify the polling mechanism for qpairs on a single
thread. It also will pave the way for doing transport specific aggregation of
qpair polling to increase performance.
The generic implementation is included. The transport specific calls
have yet to be implemented.
Change-Id: If07b4170b2be61e4690847c993ec3bde9560b0f0
Signed-off-by: Seth Howell <seth.howell@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/579
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom CI
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Also modify some api documentation to indicate how the
new API should be used.
Signed-off-by: Seth Howell <seth.howell@intel.com>
Change-Id: Icdbfb09aceda28635fdd191c520b36c692c2c100
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/1340
Community-CI: Broadcom CI
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
connect_io_qpair essentially allows us to split the qpair allocation process
in half which will make it possible for us to do more sophisticated things
with RDMA qpairs in poll groups. as a companion to this new API, a connect_only
option has been added to the io_qpair_opts struct which instructs alloc_io_qpair
to only allocate the qpair and not connect it.
Signed-off-by: Seth Howell <seth.howell@intel.com>
Change-Id: I9ba9502dd39436006a9ac71436dd1871d648ed1c
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/1123
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom CI
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Contidion previous to this should already verify that
md page is not an extent page.
All extent pages are not part of the chain (sequence_num == 0),
and their location (ctx->cur_page) cannot be the root of
md chain (page->id).
Yet during development it could appen, so adding assert
to verify further that the md page is not extent page.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: I6d5dc2ae965f8f9a388cd1c8e186145f8ca91db4
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/1667
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Before this patch reading out the extent pages during
blobstore replay was serialized. Only issuing reads for next
extent page when previous operation finished.
This was done by continously calling _spdk_bs_load_replay_extent_page_cpl()
and decreasing ctx->num_extent_pages.
This patch changes spdk_bs_sequence_* to spdk_bs_batch_*.
All the reads are submitted at once, and only when all of them
finish we proceed to next valid md chain.
Goal of this change is improving efficiency and readability.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: I807cdb98166e04706fedb494363f5776e3151827
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/1540
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
This variable currently holds single extent page.
Further patch will utilize it to use multiple.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: If86ffd57cecf5d3bfd0812a767c784d7bf503fb6
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/1538
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
This field does not hold actual pages, but just md page numbers
which hold the extent pages.
Rename as prepartation to adding new one that will hold actual
extent pages.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: I1fb85a58c92a93b968e1fad22e421252399e9281
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/1537
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Add function nvme_rdma_get_key to get either lkey
or rkey, use it in request building functions
Change-Id: Ic9e3429e07a10b2dddc133b553e437359532401d
Signed-off-by: Alexey Marchuk <alexeymar@mellanox.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/1462
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Cache payload type and in-capsule data transfer support
Change-Id: Id40a6e86d1f29235ca3e0189d7fbcf19baa30ffe
Signed-off-by: Alexey Marchuk <alexeymar@mellanox.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/1461
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
%*s prints at least * characters and may add padding to the string.
%.*s prints at most * characters.
In a few places we used the first instead of the second and
printed some garbage to screen (in the best case...).
Change-Id: I97a862be61a5e43aa61e8230044dbd64a9db33bd
Signed-off-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/1569
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
We have been intermittently hitting the assert where
we check sock->cb_fn != NULL in spdk_sock_group_impl_poll_count.
The only way we could be hitting this specific error is if we
wereremoving a socket from a sock group within after receiving
an event for it.
Specifically, we are seeing this error on the NVMe-oF TCP target
which relies on posix sockets using epoll.
The man page for epoll states the following:
If you use an event cache or store all the file descriptors
returned from epoll_wait(2), then make sure to provide
a way to mark its closure dynamically (i.e., caused by
a previous event's processing). Suppose you receive 100 events
from epoll_wait(2), and in event #47 a condition causes event
#13 to be closed. If you remove the structure and close(2)
the file descriptor for event #13, then your event cache might
still say there are events waiting for that file descriptor
causing confusion.
One solution for this is to call, during the processing
of event 47, epoll_ctl(EPOLL_CTL_DEL) to delete file
descriptor 13 and close(2), then mark its associated data
structure as removed and link it to a cleanup list. If
you find another event for file descriptor 13 in your batch
processing, you will discover the file descriptor had
been previously removed and there will be no confusion.
Since we do store all of the file descriptors returned from
epoll_wait, we need to implement the tracking mentioned above.
fixes issue #1294
Signed-off-by: Seth Howell <seth.howell@intel.com>
Change-Id: Ib592ce19e3f0b691e3a825d02ebb42d7338e3ceb
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/1589
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Here destruct contrllers are in one function, and we can
remove the duplicated codes using goto.
It can save several lines of codes.
Signed-off-by: yidong0635 <dongx.yi@intel.com>
Change-Id: Ibf3cb9fe2ea4bfc65d42603a7b13aaf575854580
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/1638
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
When removing large number of devices (>8) in parallel,
the 20ms timeout is not long enough.
As part of spdk_detach_cb, DPDK calls into the VFIO driver
which may get delayed due to multiple hot removes being
processed by pciehp driver (pciehp IRQ thread function
is handling the actual removal of a device in paralle but
all of the IRQ thread function compete for a global mutex
increasing processing time and race conditions).
Signed-off-by: Michael Haeuptle <michael.haeuptle@hpe.com>
Change-Id: I470fbbee92dac9677082c873781efe41e2941cd5
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/1588
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Recent improvements to blobstore code, allow proper
reporting whenever error occured.
Callback for blob_create, should not refer to the blob
pointer when the operation failed.
Defer checking spdk_blob_get_id(blob), til after checking
error code.
This has been encountered in "vhost_boot" test:
15:20:54 # /var/jenkins/workspace/vhost-autotest/spdk/scripts/rpc.py -s /home/sys_sgci/vhost_test/vhost/0/rpc.sock bdev_lvol_create -u d605539d-2186-486f-bc3e-745e33d37072 lvb0 20000
vhost: blobstore.c:5101: spdk_blob_get_id: Assertion `blob != NULL' failed.
https://ci.spdk.io/results/autotest-per-patch/builds/8135/archive/vhost-autotest/build.log
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: Id926576e10a72e0d50fb65105be07c8ae8e40096
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/1604
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
This patch adds more ways to back off when parts of
blob persist fails.
Otherwise the process would proceed as if nothing happened.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: I7cff73e1dc3066d0c822d1e3dac4bd35e27cd54a
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/1263
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Darek Stojaczyk <dariusz.stojaczyk@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
After opening the blob for deletion, in _spdk_bs_delete_open_cpl(),
the blob is removed from list of blobs in blobstore.
This is to prevent future _spdk_blob_lookup()s from referencing
blob while it is deleted.
In usual blob deletion path, next step is proceeding with deletion
of the blob by reducing its size to 0 and syncing the blob.
Changes from this point forward are persisted.
Meanwhile in special case of deleting snapshot which has single clone
on it, before above occurs additional steps are performed.
Each of the blobs are opened and their attributes changed.
Failures on those steps are fully recoverable on any errors,
and in such case blob should be added back to the bs list of blobs.
Original code had condition on how many references there were
to blob being deleted, which is incorrect.
Any error on that path should clean up after itself (revert attributes
and close blobs) and re-add the blob.
This change is tested with blob_delete_snapshot_power_failure() UT,
by adding error path in persist - which triggers error in aforementioned
blob delete code path.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: I926e7cbf3cb86170c69f31231399535859f290dd
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/985
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
When creating snapshot, 'original' blob will end up being a thin provisioned clone.
Before that first thin_provisioned 'newblob' is created during this process.
If the first md sync for 'newblob' fails, it means that only valid references to
clusters are still only present in 'original' blob. The 'newblob' can be safely
cleaned up.
Unfortunetly 'newblob' inherited some of 'original' blob properties before sync.
Cluster maps were already swaped in current cleanup code. But during blob close
of 'newblob' - persist blob code expects clusters to be 0 only for thin_provisioned
blobs. If original blob was thick, then it triggers an assert within persist code.
This patch makes sure to set thin_provision to 'newblob', to align with its creation.
Added asserts to verify that clusters maps are 0's, which should be the case
as I/O to origblob is frozen.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: I5420617792aefe8a3ef4e5989b2056504cdd1850
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/1394
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Previously the SPDK NVMe driver always set PSDT to 01b for hardware SGLs
which is aligned to the Linux NVMe driver, for this case the metadata length
is not required when filling the NVMe command fields. There is no alignment
nor granularity requirement for Data Blocks for PSDT 01b case. And if the
drive reported that it needs dword alignment with SGL, for this case, when
using spearate metadata, it needs a length parameter to fill the SGL descriptor.
Change-Id: I56ffaada775fe66de7637dae15b509ee9556e80a
Signed-off-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/1351
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
When the drives report that SGL dword alignment is required in
Identify Controller data structure, when using separate metadata,
PSDT should only be set with 10b. The specification says: If PSDT
01b was used, Metadata Pointer (MPTR) contains an address of a
single contiguous physical buffer that is byte aligned.
For supporting this case, SPDK driver needs a metadata SGL entry,
so we can reserve one entry in the tracker data structure.
Change-Id: I2d86a58b0395c3000626f922e56d7f2212c8a752
Signed-off-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/1316
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Add FC listen address to target listeners table before attaching the
same to the subsystem listener list. Without this fix, subsystem allow
any listener provision which is very important to FC-NVMe is broken.
Signed-off-by: Anil Veerabhadrappa <anil.veerabhadrappa@broadcom.com>
Change-Id: I06436c0a73e65cb5f7bb3280658fcf200b975989
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/1443
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
