These can produce a lot of output, which doesn't really give any
additional information.
Signed-off-by: Konrad Sztyber <konrad.sztyber@intel.com>
Change-Id: I572cd203d61c717ce6400f67ef27ec1d7bb54c0c
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/10414
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Jacek Kalwas <jacek.kalwas@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
adding transport to tgt should be the last step
also there is an issue before change i.e. if calloc failed then
transport remains on the list
Signed-off-by: Jacek Kalwas <jacek.kalwas@intel.com>
Change-Id: Iaf29cfc7b0f535d40160c6fdf9ef6a7e6bfb127c
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/10429
Community-CI: Mellanox Build Bot
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
For error case, just set ctrlr->num_aers to 0, and
then the loop won't execute at all. This avoids an
extra call to nvme_ctrlr_set_state() and simplifies
the code a bit.
Signed-off-by: Jim Harris <james.r.harris@intel.com>
Change-Id: Iff7bbf6e03d18b5f553b9e8527b4c803db583917
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/10330
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Discovery services using the SPDK nvme driver may
use long-lasting connections that detect AER completions
to determine when there are changes in the discovery
log. This means that we still need to send keep alives
on discovery controller admin queues. So move the
SET_KEEP_ALIVE_TIMEOUT state immediately after
IDENTIFY, and run the SET_KEEP_ALIVE_TIMEOUT state
even for discovery controllers.
Note, we need the IDENTIFY's KAS value to properly
set the keep alive timeout, so we have to keep the
IDENTIFY state before SET_KEEP_ALIVE_TIMEOUT.
Signed-off-by: Jim Harris <james.r.harris@intel.com>
Change-Id: I5c6403c28fb72d42629c5f9009a89c4bfd44d162
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/10329
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Keep alive is valid for discovery controllers, so don't overwrite
the value requested with zero in nvme_fabric_ctrlr_scan().
Signed-off-by: Jim Harris <james.r.harris@intel.com>
Change-Id: I7dcda6ebf4ab1c8a9085e4e3a02b814d8e586a97
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/10328
Community-CI: Mellanox Build Bot
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Reviewed-by: Dong Yi <dongx.yi@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Next patch in series will look up the struct spdk_sock_group_impl
from spdk_sock_group in spdk_sock_get_optimal_sock_group().
Since this is third place it will be used, make this function common.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: I43d6472016782e78709c1d52aa74abf594e5bfe6
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/10347
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
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>
When no optimal poll group exists for a qpair,
assignment for round robin happens in spdk_nvmf_tgt_new_qpair().
RDMA transport implments the logic for this assignment in
nvmf_rdma_get_optimal_poll_group().
TCP relied on the spdk_nvmf_tgt_new_qpair() instead.
This resulted in race condition when looking up and assigning
optimal poll groups - see #2113.
To remedy that, TCP now follows the same pattern as RDMA.
Next patch will improve the sock map lookup to fix the #2113.
Signed-off-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: I672d22ac15d06309edf87ece5d30f8e8d1095fbb
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/10270
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
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>
There is a bad memory corruption where the code for CUSE attempts to write
one byte (with value 0x1) after the memory is freed.
Context:
When the CUSE device is unregistered, the poller thread is signaled with
fuse_session_exit(), which writes the value 1 to fuse_session::exited.
The poller thread then detects with fuse_session_exited() that it must exit
the routing and finally destroys its own fuse session with
cuse_lowlevel_teardown() before it exits.
However, FUSE may also call fuse_session_exit() for its internal purposes.
I'm not sure exactly under what conditions that happens, but I added
some trace messages and I could clearly see that the CUSE thread exits
before it was requested to exit in cuse_nvme_ns_stop().
If the poller thread early-exits, it would destroy its own FUSE session
(and free the memory) before fuse_session_exit() gets executed, causing
the memory to be corrupted with a single byte of value 0x1.
Reproducer:
The bug can be reproduced by resetting the FUSE session to NULL after it
is destroyed. This will cuse_nvme_ns_stop() to crash with a segmentation
fault in fuse_session_exit() because it tries to access a NULL pointer.
static void *
cuse_thread(void *arg)
{
[...]
free(buf.mem);
fuse_session_reset(cuse_device->session);
+ cuse_device->session = NULL;
pthread_exit(NULL);
}
This fix:
The fix I suggest is to destroy the FUSE session with
cuse_lowlevel_teardown() after the thread is joined.
Signed-off-by: Sylvain Didelot <sdidelot@ddn.com>
Change-Id: I47202891a358f139506845110b012f840974b6fc
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9931
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
spdk_nvme_ctrlr_free_io_qpair can be called when
qpair is already disconnected. In that case qpair's
state is changed to NVME_QPAIR_DESTROYING and
transport's ctrlr_delete_io_qpair callback is
called. RDMA and TCP transports call
nvme_transport_ctrlr_disconnect_qpair in
the callback and since qpair's state is
not DISCONNECTED or DISCONNECTING, qpair
is disconnected for the second time.
If spdk_nvme_ctrlr_free_io_qpair is called
when qpair is in ENABLED state than nothing
changes, qpair will be disconnected before destroy.
PCIE/vfio_user don't implement transport disconnect
callback, so they are not affected.
Signed-off-by: Alexey Marchuk <alexeymar@mellanox.com>
Change-Id: I23e11856ecafb51669acf4a3118be049c11eecda
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/10326
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Dong Yi <dongx.yi@intel.com>
Reviewed-by: Konrad Sztyber <konrad.sztyber@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Change-Id: I58794b7b946eeb8ff82512905af0a296e3b534aa
Signed-off-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9817
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: John Kariuki <John.K.Kariuki@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
For DSM command, the NVMe drive may take a long time to finish it,
if we set a small timeout value for DSM command, the bdev/nvme module
will try to reset the IO queue pair when timeout happens,
in `spdk_nvme_ctrlr_free_io_qpair`, we will abort the outstanding
IO requests first, then in the `nvme_pcie_ctrlr_delete_io_qpair`,
we will poll the CQ for any requests that have been completed by
the NVMe controller, if there are NVMe completions in the CQ,
we will finish them again, thus double completions happened.
Here we rename `nvme_qpair_abort_reqs` to `nvme_qpair_abort_all_queued_reqs`,
so the common layer will just abort queued request, and let each
transport to abort outstanding requests case by case.
Fix#2233.
Change-Id: Icae6214239160c615418cb514fc51cfe77b59211
Signed-off-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/10233
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Add a parameter which determines the owner of the
map - target or initiator. It allows to set different
access flags when creating Memory Regions
Signed-off-by: Alexey Marchuk <alexeymar@mellanox.com>
Change-Id: I0016847fe116e193d0954db1c8e65066b4ff82bf
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/10283
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
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>
The unit test test_nvme_cuse_stop() manually creates 2 cuse devices
and executes nvme_cuse_stop(). Problem is that the Fuse session is
never initialized for those 2 cuse devices, causing cuse_nvme_ns_stop()
to access 'ns_device->session', which is a NULL pointer.
This bug is detected by ASAN as follows:
==77298==ERROR: AddressSanitizer: SEGV on unknown address 0x000000000180 (pc 0x7fdac6d7d40e bp 0x000000000000 sp 0x7fff74768320 T0)
==77298==The signal is caused by a READ memory access.
==77298==Hint: address points to the zero page.
0 0x7fdac6d7d40e in fuse_session_destroy (/usr/lib64/libfuse3.so.3+0x1640e)
1 0x40dc7a in cuse_nvme_ns_stop /home/vagrant/spdk_repo/spdk/lib/nvme/nvme_cuse.c:851
2 0x40df59 in cuse_nvme_ctrlr_stop /home/vagrant/spdk_repo/spdk/lib/nvme/nvme_cuse.c:923
3 0x40f103 in nvme_cuse_stop /home/vagrant/spdk_repo/spdk/lib/nvme/nvme_cuse.c:1094
4 0x415803 in test_nvme_cuse_stop /home/vagrant/spdk_repo/spdk/test/unit/lib/nvme/nvme_cuse.c/nvme_cuse_ut.c:393
5 0x7fdac724c1a6 (/usr/lib64/libcunit.so.1+0x41a6)
6 0x7fdac724c528 (/usr/lib64/libcunit.so.1+0x4528)
7 0x7fdac724d456 in CU_run_all_tests (/usr/lib64/libcunit.so.1+0x5456)
8 0x415a4e in main /home/vagrant/spdk_repo/spdk/test/unit/lib/nvme/nvme_cuse.c/nvme_cuse_ut.c:415
9 0x7fdac62351e1 in __libc_start_main (/usr/lib64/libc.so.6+0x281e1)
10 0x403ddd in _start (/home/vagrant/spdk_repo/spdk/test/unit/lib/nvme/nvme_cuse.c/nvme_cuse_ut+0x403ddd)
The fix is to call fuse_session_destroy() only if the fuse session is != NULL.
Signed-off-by: Sylvain Didelot <sdidelot@ddn.com>
Change-Id: I41881243227d83e8d1e6b90e72c1b6d62ccd98d3
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/10225
Community-CI: Mellanox Build Bot
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
We allocate from the head, so it's better to free to
the head too for better cache utilization.
Signed-off-by: Jim Harris <james.r.harris@intel.com>
Change-Id: I5c32244f446bd7a1df12eefc81245b3ef7e24070
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/10193
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Reviewed-by: John Kariuki <John.K.Kariuki@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
We allocate tasks from the head, so it's better to
free them to the head too for better cache utilization.
Signed-off-by: Jim Harris <james.r.harris@intel.com>
Change-Id: I67c23e3d89cda16f94b1770eada5465015ddb6ff
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/10192
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: John Kariuki <John.K.Kariuki@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
The NOTICELOGs really clutter the output during
application start - it's better to make these DEBUGLOGs
instead.
Signed-off-by: Jim Harris <james.r.harris@intel.com>
Change-Id: I3ae37d5d057d7b972017befbc0834de414b9710b
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/10191
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
This limitation doesn't really take effect currently,
since the typical number of slots per channel isn't
bigger than MAX_COMPLETIONS_PER_POLL. But there's
no reason for this limit anymore - we should always
poll as many completions as we find.
It's better to remove this now, in case we have
configs in the future with higher number of slots
per channel.
Signed-off-by: Jim Harris <james.r.harris@intel.com>
Change-Id: I5d90d41e5142622b79d9765fbc62da1516e2b8be
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/10189
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: John Kariuki <John.K.Kariuki@intel.com>
Reviewed-by: Monica Kenguva <monica.kenguva@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
spdk_bdev_io_get_nvme_status() does not follow NVMe abort command
about return values.
NVMe abort command sets completion status to SUCCESS both for success and
failure cases and differentiates only the bit 0 of cdw0.
lib/nvmf do not use spdk_bdev_io_get_nvme_status() but checks only
success or failure at completion.
So there is no issue now but let spdk_bdev_io_get_nvme_status()
follow NVMe abort command. In future, the user of spdk_bdev_abort()
may use spdk_bdev_io_get_nvme_status().
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: Ic4a08056bd8a1aee4c400f72ef5de7c68e23990b
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9977
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
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: Konrad Sztyber <konrad.sztyber@intel.com>
The pointer to struct spdk_nvmf_ctrlr is used to save mandatory
controller registers to the migration region.
Also rename some ctrlr/qpiar to vu_ctrlr/vu_qpiar.
Change-Id: Ifcb862bf4543a9df3c62d3b0a57b7f93228ccaba
Signed-off-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/7629
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.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>
Reviewed-by: Thanos Makatos <thanos.makatos@nutanix.com>
1. use the transport lock to protect transport endpoints list.
2. don't use mixed errno and -1 as the return value, use -1 for all error cases.
Change-Id: I657fa06a6d82ee8dbeefaa3397df2285ba574b80
Signed-off-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9579
Community-CI: Mellanox Build Bot
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
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 destroying controller, we will disconnect each connected qpair,
and in the spdk_nvmf_qpair_disconnect() call, qpair_fini() will also
try to hold the same lock, so existing vfio-user implementation assume
that qpair_fini() will not be called in the same context. Patch
https://review.spdk.io/gerrit/c/spdk/spdk/+/8963 remind me that
vfio-user has this issue. While here, we add one more thread poll
to avoid such issue.
Change-Id: I83b82ddcce3eb54c724291223e794dcb53a08059
Signed-off-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9998
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
For live migration support in vfio-user transport, we need to pause
the subsystem when starting migration in source VM, then after
migration, the subsystem is in paused state, when exiting the
application, we will call spdk_nvmf_subsystem_stop() at last,
and existing code will assert this case.
Change-Id: If5214c45973b27f6092c4a6d71ede336e54d89e8
Signed-off-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9407
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
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: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Dong Yi <dongx.yi@intel.com>
The recent changes merged multiple Data-OUT PDUs within the same
sequence into a single subtask up to 64KB.
However, they were not enough.
For a large write operation, the hardware iSCSI HBA host sent an immediate
data whose size was not block size multiples and then more solicit
data through R2T exchanges.
One example for a 64KB write operation was as follows:
host sent SCSI Write with 5792 bytes and F = 1
target replied a R2T
host sent Data-OUT with 15880 bytes
host sent Data-OUT with 11536 bytes
host sent Data-OUT with 2848 bytes
host sent Data-OUT with 11536 bytes
host sent Data-OUT with 5744 bytes
host sent Data-OUT with 12200 bytes and F = 1
The hardware iSCSI HBA host can decide the size of the unsolicited data
but the SPDK iSCSI target can require the host to send the solicited data
whose size is block size multiples.
Hence we merge immediate data to the following R2T data if the immediate
data is not more than 64KB and more R2T data come.
Add another test case to check if the fix works for the above example.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I4906b4e1a8b61e08862f4ccc27a6caf165126530
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9708
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Mellanox Build Bot
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
This clean up will make the following patches easier.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I1ad288ec16aec69a168e0f3019b68e11132b65c9
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9707
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Mellanox Build Bot
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Issue: spdk_top tracked pollers by the poller name string and the
thread_id they are running on. This shows incorrect stats when
multiple pollers exist on the same thread with the same name.
Solution: Added a unique poller id for each poller on a thread and
to allow spdk_top to track pollers by thread_id and poller_id.
Signed-off-by: Michael Piszczek <mpiszczek@ddn.com>
Change-Id: I1879e2afc9a929d1df9e8e35510f0092c5443bdc
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/5868
Community-CI: Mellanox Build Bot
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Wait until the namespace is attached, where it does this operation
again. As of this commit it doesn't really matter because it is just
filling in some values in a structure and if it does it twice it's not a
problem. But later when we only allocate active namespaces, we do not
want to allocate the namespace twice.
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Change-Id: Ie28653b178975d1ca80bf71ca6b5095224f1c5d1
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/10026
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Mellanox Build Bot
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Eugene Kochetov <ekochetov@yandex.ru>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
This function destructs a single namespace and removes it from the
controller.
Change-Id: I4b7b3576beda85c9ddad4e0f2db6d1964fa72b82
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/10024
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Eugene Kochetov <ekochetov@yandex.ru>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
This was sometimes used as the maximum array index and sometimes as the
maximum count. Make it consistent everywhere and give it a better name.
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Change-Id: I518efd99a7d36584624490b0b3497bb6e81ce9ac
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/10101
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Aleksey Marchuk <alexeymar@mellanox.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
list
The list should always be null terminated, but add an additional layer
of buffer overrun protection.
Change-Id: Iee31057fdca5ec4a6177615dff5171e5cb07984e
Signed-off-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/10027
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Eugene Kochetov <ekochetov@yandex.ru>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
This quirk was already applied to the 0x0A53 SSD, but is
likely needed on 0x0A54 as well.
Possible fix for issue #2231.
Signed-off-by: Jim Harris <james.r.harris@intel.com>
Change-Id: I36a48ab92d1698a411472f714b5108413bbc3c56
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/10162
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Dong Yi <dongx.yi@intel.com>
Reviewed-by: Tomasz Zawadzki <tomasz.zawadzki@intel.com>
Community-CI: Mellanox Build Bot
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
In the bdev-zone API, there are a few functions that takes a zone_id:
spdk_bdev_get_zone_info(), spdk_bdev_zone_management(), and the
spdk_bdev_zone_append() functions.
The way a zoned application is usually written is that it starts off
by getting the zone report for all zones (zone_id will be sent in as 0),
and then the application will keep the whole zone report in memory.
Therefore, an application usually have access to the zone_id/zslba for
all zones. However, there are cases, e.g. when getting an error on write,
where the completion callback will only have the lba of the write that
failed.
Add a helper function that can be used to get the zone_id/slba for a
given lba. Having this helper in bdev-zone will avoid SPDK applications
needing to provide their own implementation for this.
Signed-off-by: Niklas Cassel <niklas.cassel@wdc.com>
Change-Id: I978335f87f7d49bc33aed81afcaa6d9f0af8a1e4
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/10180
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
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>
DPDK vhost will call `new_device` when the VRINGs are
queue paired(virtio-net) or all the VRINGs are started.
However, for virtio-blk/scsi, SeaBIOS will only use one
VRING queue, DPDK added a workaround patch to add
`pre_msg_handle` and `post_msg_handle` callbacks to let
devices other than virtio-net to process such scenarios.
In SPDK, we will start the device when there is one valid
VRING, so there is a case that SPDK and DPDK have different
state for one device. For a virtio-scsi device, SeaBIOS will
only start the request queue, and in the BIOS stage, SPDK will
start the device but DPDK doesn't think so. If users killed
SPDK vhost target at the moment, in `session_shutdown`, SPDK
will expect DPDK to call `destroy_device` to do the cleanup,
but DPDK won't do that as it thinks the device isn't started.
Here in `session_shutdown`, SPDK will do this first, it's OK
that DPDK will call another `destroy_device` for devices that
have the same state both in SPDK and DPDK.
Fix issue #2228.
Change-Id: Ib76dd54c8fa302ffe6da9b13498312b7d344bbfe
Signed-off-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/10143
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
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>
_stop_session() is called while holding the global vhost lock,
and in the caller we do release the vhost lock, so even for the
error return from device backend, we don't need to release it
in _stop_session().
Change-Id: I08fef64f900bb42ee68bf02b4c4f1406e903a8a6
Signed-off-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/10142
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Dong Yi <dongx.yi@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
`struct spdk_vhost_dev vdev` in `struct spdk_vhost_scsi_dev` can be
unregistered in `vhost_scsi_dev_remove`, so we can't use it
anymore in other places after `vhost_dev_unregister`.
Ideally `state->remove_cb` should not take the `vdev` as
the input parameter either, but I don't find it's used
anywhere, so leave it unchanged.
==29555==ERROR: AddressSanitizer: heap-use-after-free on address 0x602000006df0
READ of size 2 at 0x602000006df0 thread T0 (reactor_0)
#0 0x7f3c246c0f0a (/lib64/libasan.so.5+0x9cf0a)
#1 0x7f3c246c3c15 in vsnprintf (/lib64/libasan.so.5+0x9fc15)
#2 0xa55cfa in spdk_vlog /spdk/lib/log/log.c:158
#3 0xa5596f in spdk_log /spdk/lib/log/log.c:110
#4 0x842e43 in remove_scsi_tgt /spdk/lib/vhost/vhost_scsi.c:208
#5 0x851508 in vhost_scsi_dev_remove_tgt_cpl_cb /spdk/lib/vhost/vhost_scsi.c:1149
#6 0x8383f1 in foreach_session_finish_cb /spdk/lib/vhost/vhost.c:1144
#7 0x9d3223 in msg_queue_run_batch /spdk/lib/thread/thread.c:703
#8 0x9d73fe in thread_poll /spdk/lib/thread/thread.c:919
#9 0x9d7c3b in spdk_thread_poll /spdk/lib/thread/thread.c:979
#10 0x8812fe in _reactor_run /spdk/lib/event/reactor.c:920
#11 0x881bf1 in reactor_run /spdk/lib/event/reactor.c:958
#12 0x88292b in spdk_reactors_start /spdk/lib/event/reactor.c:1060
#13 0x873ff9 in spdk_app_start /spdk/lib/event/app.c:585
#14 0x408044 in main /spdk/app/vhost/vhost.c:105
#15 0x7f3c23691f42 in __libc_start_main (/lib64/libc.so.6+0x23f42)
#16 0x407add in _start (/spdk/build/bin/vhost+0x407add)
0x602000006df0 is located 0 bytes inside of 8-byte region [0x602000006df0,0x602000006df8)
freed by thread T0 (reactor_0) here:
#0 0x7f3c2473191f in __interceptor_free (/lib64/libasan.so.5+0x10d91f)
#1 0x8369f2 in vhost_dev_unregister /spdk/lib/vhost/vhost.c:1024
#2 0x84f32d in vhost_scsi_dev_remove /spdk/lib/vhost/vhost_scsi.c:913
#3 0x83cdb7 in spdk_vhost_dev_remove /spdk/lib/vhost/vhost.c:1494
#4 0x83ed66 in vhost_fini /spdk/lib/vhost/vhost.c:1644
#5 0x9d3223 in msg_queue_run_batch /spdk/lib/thread/thread.c:703
#6 0x9d73fe in thread_poll /spdk/lib/thread/thread.c:919
#7 0x9d7c3b in spdk_thread_poll /spdk/lib/thread/thread.c:979
#8 0x8812fe in _reactor_run /spdk/lib/event/reactor.c:920
#9 0x881bf1 in reactor_run /spdk/lib/event/reactor.c:958
#10 0x88292b in spdk_reactors_start /spdk/lib/event/reactor.c:1060
#11 0x873ff9 in spdk_app_start /spdk/lib/event/app.c:585
#12 0x408044 in main /spdk/app/vhost/vhost.c:105
#13 0x7f3c23691f42 in __libc_start_main (/lib64/libc.so.6+0x23f42)
Change-Id: I511c4316a838cd92961d57c9193d384acd49d760
Signed-off-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/10141
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Dong Yi <dongx.yi@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Previously task->current_data_offset was updated by add_transfer_task().
However, the following patches will merge unsolicited data and solicited
data into a single subtask. It will be possible that add_transfer_task()
is called but subtask is not submitted. As a preparation, extract
updating task->current_data_offset into iscsi_pdu_payload_op_scsi_write().
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I5262bb883fa2a081be1f087181de98d4c3c24d69
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9706
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
When data segment size is 64KB and data digest is enabled, if
data segment and data digest are split into different two packets,
- pdu->mobj[0] became full first when reading data semgment,
- pdu->mobj[1] was allocated but unused and data digest was read.
In this case, two SCSI write tasks were submitted by mistake and
the second SCSI write task had no data.
Fix the bug in this patch.
When iscsi_pdu_payload_read() is called and pdu->mobj[0] is full,
allocate pdu->mobj[1] only if any of data segment remains to read.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I9a0c36c05f90092c3c2122a7eb91e10976830b40
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9965
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
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>
We had not considered a case that incoming data to the second data
buffer was split into multiple TCP packets when merging incoming data
up to 64KB.
We do not change the unit test because we already have data check
and it is very hard to include partial read into the data check.
However, it is very usual that incoming data is split into multple
TCP packets. The feature to merge incoming data up to 64KB will be
actually enabled in the following patches. So we rely on the I/O test
to verify this fix.
Signed-off-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Change-Id: I50d702d6c118bc16f0767845136e14414ccdf813
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/9736
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
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>
The internal device list isn't used anywhere, and will cause ASAN
error because we didn't remove the entry from the device list when
destructing controller.
Change-Id: Ie97bf10ca44ff773a8bc5f0476611b3844ef901a
Signed-off-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/10109
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
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>
There is no need to sum SPDK_CEIL_DIV(length, io_unit_size) and
req->iovcnt as the later is always zero (assignment in spdk_nvmf_request_get_buffers).
Checking SPDK_CEIL_DIV(length, io_unit_size) is enough.
Signed-off-by: Maciej Szulik <maciej.szulik@intel.com>
Change-Id: I0fea1d86706d83e4dd9083e6f0ce09e4e4b033a8
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/10089
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Reviewed-by: Jacek Kalwas <jacek.kalwas@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
Reviewed-by: Dong Yi <dongx.yi@intel.com>
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
The specification says "host specifies an offset (i.e., LPOL and LPOU)
that is greater than the size of the log page requested, then the
controller shall abort the command with a status of Invalid Field
in Command."
Offset is used (if needed) to retrieve specific records of
Discovery Log Page, so we don't check it for Discovery Log Page.
Change-Id: I76ce929600b9f2ca9b69397d25f339d55729e6d3
Signed-off-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/10093
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
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>
This can be used for multipath validation.
Signed-off-by: Jim Harris <james.r.harris@intel.com>
Change-Id: Iba62c5e90b22a9a85078fbc783d3a7273c029cde
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/10137
Community-CI: Mellanox Build Bot
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Reviewed-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-by: Dong Yi <dongx.yi@intel.com>
Reviewed-by: Shuhei Matsumoto <shuhei.matsumoto.xt@hitachi.com>
CAP.CQR (Contiguous Queues Required) is always 1, so we should
return invalid field when PC bit is 1 and return Invalid Interrupt
Vector if Interrupt Vector is too big.
Also fix the issue that just creat/delete a CQ.
Change-Id: I7cc64a5946a1ed3161448fca8b433d08e5fee715
Signed-off-by: Changpeng Liu <changpeng.liu@intel.com>
Reviewed-on: https://review.spdk.io/gerrit/c/spdk/spdk/+/10080
Reviewed-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: Ben Walker <benjamin.walker@intel.com>
Tested-by: SPDK CI Jenkins <sys_sgci@intel.com>
Community-CI: Broadcom CI <spdk-ci.pdl@broadcom.com>
Community-CI: Mellanox Build Bot
