This commit reworks the port functions to retrieve information
about the port, like the enq or deq depths. Note that "port count"
is a device attribute, and is added in a later patch for dev attributes.
Signed-off-by: Harry van Haaren <harry.van.haaren@intel.com>
Update doxygen to make it clear that RTE_EVENT_OP_FORWARD and
RTE_EVENT_OP_RELEASE must only be enqueued to the same port that the
original event was dequeued from.
Signed-off-by: Tim McDaniel <timothy.mcdaniel@intel.com>
Acked-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
Events sent through single-link queues are naturally in-order and
atomic, without reordering or atomic scheduling. Logically the
nb_atomic_flows and nb_atomic_order_sequences arguments don't apply to a
single link queue, but applications must set these (depending on the queue
config type) to bypass the is_valid_{ordered, atomic}_queue_conf() checks
in the eventdev layer.
This commit updates those is_valid_* functions to ignore queues with the
SINGLE_LINK flag, to simplify their configuration.
Signed-off-by: Gage Eads <gage.eads@intel.com>
Acked-by: Jerin Jacob <jerin.jacob@caviumnetworks.com>
This patch adds an union in VhostUserMsg to distinguish between
master and slave initiated requests, instead of casting slave
requests as master request.
Signed-off-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Acked-by: Yuanhan Liu <yliu@fridaylinux.org>
Added new callbacks to notify about socket connection status.
As destroy_device is used for virtqueue processing *pause* as well as
connection close, the user has no distinction between those.
Consider the following scenario:
rte_vhost: received SET_VRING_BASE message,
calling destroy_device() as usual
user: end-user asks to remove the device (together with socket file),
OK, device is not *in use* - that's NOT the behavior we want
calling rte_vhost_driver_unregister() etc.
Instead of changing new_device/destroy_device callbacks and breaking
the ABI, a set of new functions new_connection/destroy_connection
has been added.
Signed-off-by: Dariusz Stojaczyk <dariuszx.stojaczyk@intel.com>
Reviewed-by: Jens Freimann <jfreimann@redhat.com>
As soon as a page used by a ring is invalidated, the access_ok flag
is cleared, so that processing threads try to map them again.
Signed-off-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Acked-by: Yuanhan Liu <yliu@fridaylinux.org>
Translating the start addresses of the rings is not enough, we need to
be sure all the ring is made available by the guest.
It depends on the size of the rings, which is not known on SET_VRING_ADDR
reception. Furthermore, we need to be be safe against vring pages
invalidates.
This patch introduces a new access_ok flag per virtqueue, which is set
when all the rings are mapped, and cleared as soon as a page used by a
ring is invalidated. The invalidation part is implemented in a following
patch.
Signed-off-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Acked-by: Yuanhan Liu <yliu@fridaylinux.org>
When IOMMU is enabled, the ring addresses set by the
VHOST_USER_SET_VRING_ADDR requests are guest's IO virtual addresses,
whereas Qemu virtual addresses when IOMMU is disabled.
When enabled and the required translation is not in the IOTLB cache,
an IOTLB miss request is sent, but being called by the vhost-user
socket handling thread, the function does not wait for the requested
IOTLB update.
The function will be called again on the next IOTLB update message
reception if matching the vring addresses.
Signed-off-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Acked-by: Yuanhan Liu <yliu@fridaylinux.org>
This patch postpones rings addresses translations and checks, as
addresses sent by the master shuld not be interpreted as long as
ring is not started and enabled[0].
When protocol features aren't negotiated, the ring is started in
enabled state, so the addresses translations are postponed to
vhost_user_set_vring_kick().
Otherwise, it is postponed to when ring is enabled, in
vhost_user_set_vring_enable().
[0]: http://lists.nongnu.org/archive/html/qemu-devel/2017-05/msg04355.html
Signed-off-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Acked-by: Yuanhan Liu <yliu@fridaylinux.org>
numa_realloc() reallocates the virtio_net device structure and
updates the vhost_devices[] table with the new pointer if the rings
are allocated different NUMA node.
Problem is that vhost_user_msg_handler() still dereferences old
pointer afterward.
This patch prevents this by fetching again the dev pointer in
vhost_devices[] after messages have been handled.
Fixes: af295ad4698c ("vhost: realloc device and queues to same numa node as vring desc")
Cc: stable@dpdk.org
Signed-off-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Acked-by: Yuanhan Liu <yliu@fridaylinux.org>
When VHOST_USER_F_PROTOCOL_FEATURES is negotiated, the ring is not
enabled when started, but enabled through dedicated
VHOST_USER_SET_VRING_ENABLE request.
When not negotiated, the ring is started in enabled state, at
VHOST_USER_SET_VRING_KICK request time.
Signed-off-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Acked-by: Yuanhan Liu <yliu@fridaylinux.org>
Replace rte_vhost_gpa_to_vva() calls with vhost_iova_to_vva(), which
requires to also pass the mapped len and the access permissions needed.
Signed-off-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Acked-by: Yuanhan Liu <yliu@fridaylinux.org>
This patch introduces vhost_iova_to_vva() function to translate
guest's IO virtual addresses to backend's virtual addresses.
When IOMMU is enabled, the IOTLB cache is queried to get the
translation. If missing from the IOTLB cache, an IOTLB_MISS request
is sent to Qemu, and IOTLB cache is queried again on IOTLB event
notification.
When IOMMU is disabled, the passed address is a guest's physical
address, so the legacy rte_vhost_gpa_to_vva() API is used.
Signed-off-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Acked-by: Yuanhan Liu <yliu@fridaylinux.org>
Vhost-user device IOTLB protocol extension introduces
VHOST_USER_IOTLB message type. The associated payload is the
vhost_iotlb_msg struct defined in Kernel, which in this was can
be either an IOTLB update or invalidate message.
On IOTLB update, the virtqueues get notified of a new entry.
Signed-off-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Acked-by: Yuanhan Liu <yliu@fridaylinux.org>
The per-virtqueue IOTLB cache init is done at virtqueue
init time. init_vring_queue() now takes vring id as parameter,
so that the IOTLB cache mempool name can be generated.
Signed-off-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Acked-by: Yuanhan Liu <yliu@fridaylinux.org>
In order to be able to handle other ports or queues while waiting
for an IOTLB miss reply, a pending list is created so that waiter
can return and restart later on with sending again a miss request.
Signed-off-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Acked-by: Yuanhan Liu <yliu@fridaylinux.org>
These defines and enums have been introduced in upstream kernel v4.8,
and backported to RHEL 7.4.
Signed-off-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Acked-by: Yuanhan Liu <yliu@fridaylinux.org>
Currently, only QEMU sends requests, the backend sends
replies. In some cases, the backend may need to send
requests to QEMU, like IOTLB miss events when IOMMU is
supported.
This patch introduces a new channel for such requests.
QEMU sends a file descriptor of a new socket using
VHOST_USER_SET_SLAVE_REQ_FD.
Signed-off-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Acked-by: Yuanhan Liu <yliu@fridaylinux.org>
send_vhost_message() is currently only used to send
replies, so it modifies message flags to perpare the
reply.
With upcoming channel for backend initiated request,
this function can be used to send requests.
This patch introduces a new send_vhost_reply() that
does the message flags modifications, and makes
send_vhost_message() generic.
Signed-off-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Acked-by: Yuanhan Liu <yliu@fridaylinux.org>
In the non-mergeable receive case, when copy_mbuf_to_desc()
call fails the packet is skipped, the corresponding used element
len field is set to vnet header size, and it continues with next
packet/desc. It could be a problem because it does not know why
it failed, and assume the desc buffer is large enough.
In mergeable receive case, when copy_mbuf_to_desc_mergeable()
fails, packets burst is simply stopped.
This patch makes the non-mergeable error path to behave as the
mergeable one, as it seems the safest way. Also, doing this way
will simplify pending IOTLB miss requests handling.
Signed-off-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Acked-by: Yuanhan Liu <yliu@fridaylinux.org>
This reverts commit 04d81227960b ("vhost: workaround MQ fails to
startup").
As agreed when this workaround was introduced, it can be reverted
as Qemu v2.10 that fixes the issue is now out.
The reply-ack feature is required for vhost-user IOMMU support.
Signed-off-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Acked-by: Yuanhan Liu <yliu@fridaylinux.org>
This patch adaptively batches the small guest memory copies.
By batching the small copies, the efficiency of executing the
memory LOAD instructions can be improved greatly, because the
memory LOAD latency can be effectively hidden by the pipeline.
We saw great performance boosts for small packets PVP test.
This patch improves the performance for small packets, and has
distinguished the packets by size. So although the performance
for big packets doesn't change, it makes it relatively easy to
do some special optimizations for the big packets too.
Signed-off-by: Tiwei Bie <tiwei.bie@intel.com>
Signed-off-by: Zhihong Wang <zhihong.wang@intel.com>
Signed-off-by: Zhiyong Yang <zhiyong.yang@intel.com>
Reviewed-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Tested-by: Santosh Shukla <santosh.shukla@caviumnetworks.com>
Acked-by: Yuanhan Liu <yliu@fridaylinux.org>
Split pci_vfio_map_resource for primary and secondary processes.
Save all relevant mapping data in primary process to allow
the secondary process to perform mappings.
Signed-off-by: Jonas Pfefferle <jpf@zurich.ibm.com>
Acked-by: Anatoly Burakov <anatoly.burakov@intel.com>
DMA window size needs to be big enough to span all memory segment's
physical addresses. We do not need multiple levels of IOMMU tables
as we already span ~70TB of physical memory with 16MB hugepages.
Signed-off-by: Jonas Pfefferle <jpf@zurich.ibm.com>
Acked-by: Anatoly Burakov <anatoly.burakov@intel.com>
Reviewed-by: Alexey Kardashevskiy <aik@ozlabs.ru>
Normally, command line argument strings are considered immutable, but
SPDK [1] and urdma [2] construct argv arrays to pass to rte_eal_init().
These strings are allocated using malloc() and freed after DPDK
initialization with free(). However, in the case of --file-prefix and
--huge-dir, DPDK takes the pointer to these strings in argv directly. If
a secondary process calls rte_eal_pci_probe() after rte_eal_init()
returns, as is done by SPDK, this causes a use-after-free error because
the strings have been freed by the calling code immediately after
rte_eal_init() returns.
This problem was observed when running SPDK example programs as a
secondary process and causes the secondary processes to fail:
Starting DPDK 16.11.1 initialization...
[ DPDK EAL parameters: identify -c 4 --file-prefix=spdk3260 --base-virtaddr=0x1000000000 --proc-type=auto ]
EAL: Detected 40 lcore(s)
EAL: Auto-detected process type: SECONDARY
EAL: Probing VFIO support...
EAL: VFIO support initialized
EAL: PCI device 0000:81:00.0 on NUMA socket 1
EAL: probe driver: 8086:953 spdk_nvme
EAL: cannot connect to primary process!
EAL: Error - exiting with code: 1
Cause: Requested device 0000:81:00.0 cannot be used
Running strace shows that the file prefix has been zero'd out by the
time that the secondary process attempts to probe the NVMe device.
The use-after-free errors can be easily detected with valgrind:
==8489== Invalid read of size 1
==8489== at 0x4C30D22: strlen (in /usr/lib/valgrind/vgpreload_memcheck-amd64-linux.so)
==8489== by 0x58DB955: vfprintf (vfprintf.c:1637)
==8489== by 0x59A4685: __vsnprintf_chk (vsnprintf_chk.c:63)
==8489== by 0x59A45E7: __snprintf_chk (snprintf_chk.c:34)
==8489== by 0x1246AB: get_socket_path.constprop.0 (in /home/pmacarth/src/spdk/examples/nvme/identify/identify)
==8489== by 0x124B09: vfio_mp_sync_connect_to_primary (in /home/pmacarth/src/spdk/examples/nvme/identify/identify)
==8489== by 0x123BE4: vfio_get_group_fd.part.1 (in /home/pmacarth/src/spdk/examples/nvme/identify/identify)
==8489== by 0x124366: vfio_setup_device (in /home/pmacarth/src/spdk/examples/nvme/identify/identify)
==8489== by 0x126C8A: pci_vfio_map_resource (in /home/pmacarth/src/spdk/examples/nvme/identify/identify)
==8489== by 0x12B115: pci_probe_all_drivers.part.0 (in /home/pmacarth/src/spdk/examples/nvme/identify/identify)
==8489== by 0x12B596: rte_eal_pci_probe (in /home/pmacarth/src/spdk/examples/nvme/identify/identify)
==8489== by 0x11D5B5: spdk_pci_enumerate (pci.c:147)
==8489== Address 0x63f362e is 14 bytes inside a block of size 32 free'd
==8489== at 0x4C2ED5B: free (in /usr/lib/valgrind/vgpreload_memcheck-amd64-linux.so)
==8489== by 0x11E6FB: spdk_free_args (init.c:136)
==8489== by 0x11EBF5: spdk_env_init (init.c:309)
==8489== by 0x10D2AA: main (identify.c:976)
==8489== Block was alloc'd at
==8489== at 0x4C2DB2F: malloc (in /usr/lib/valgrind/vgpreload_memcheck-amd64-linux.so)
==8489== by 0x11E7D7: _sprintf_alloc (init.c:76)
==8489== by 0x11EA78: spdk_build_eal_cmdline (init.c:251)
==8489== by 0x11EA78: spdk_env_init (init.c:282)
==8489== by 0x10D2AA: main (identify.c:976)
==8489==
Fix this by using strdup() to create separate memory buffers for these
strings. Note that this patch will cause valgrind to report memory
leaks of these buffers as there is nowhere to free them. Using static
buffers is an option but would make these strings have a fixed maximum
length whereas there is currently no limit defined by the API.
[1] http://spdk.io
[2] https://github.com/zrlio/urdma
Fixes: af75078fece3 ("first public release")
Cc: stable@dpdk.org
Signed-off-by: Patrick MacArthur <patrick@patrickmacarthur.net>
Acked-by: Sergio Gonzalez Monroy <sergio.gonzalez.monroy@intel.com>
If mmap fails, it will return the value MAP_FAILED. Checking for this
return code allows us to properly identify mmap failures and report
them as such to the calling function.
Signed-off-by: Seth Howell <seth.howell@intel.com>
Acked-by: Sergio Gonzalez Monroy <sergio.gonzalez.monroy@intel.com>
malloc_elem_free() is clearing(setting to 0) the trailer cookie when
RTE_MALLOC_DEBUG is enabled. In case of joining free neighbor element,
part of joined memory is not getting cleared due to missing the length
of trailer cookie in the middle.
This patch fixes calculation of free memory length to be cleared in
malloc_elem_free() by including trailer cookie.
Fixes: af75078fece3 ("first public release")
Cc: stable@dpdk.org
Signed-off-by: Xueming Li <xuemingl@mellanox.com>
Acked-by: Sergio Gonzalez Monroy <sergio.gonzalez.monroy@intel.com>
Currently, enabling assertion have to set CONFIG_RTE_LOG_LEVEL to
RTE_LOG_DEBUG. CONFIG_RTE_LOG_LEVEL is the default log level of control
path, RTE_LOG_DP_LEVEL is the log level of data path. It's a little bit
hard to understand literally that assertion is decided by control path
LOG_LEVEL, especially assertion used on data path.
On the other hand, DPDK need an assertion enabling switch w/o impacting
log output level, assuming "--log-level" not specified.
Assertion is an important API to balance DPDK high performance and
robustness. To promote assertion usage, it's valuable to unhide
assertion out of COFNIG_RTE_LOG_LEVEL.
In one word, log is log, assertion is assertion, debug is hot pot :)
Rationale of this patch is to introduce an dedicate switch of
assertion: RTE_ENABLE_ASSERT
Signed-off-by: Xueming Li <xuemingl@mellanox.com>
Acked-by: Gaetan Rivet <gaetan.rivet@6wind.com>
We remove xen-specific code in EAL, including the option --xen-dom0,
memory initialization code, compiling dependency, etc.
Related documents are removed or updated, and bump the eal library
version.
Signed-off-by: Jianfeng Tan <jianfeng.tan@intel.com>
Previously, to get MFN address in dom0, this API is a wrapper to
obtain the "physical address".
As we will removed xen dom0 support, this API is not necessary.
Signed-off-by: Jianfeng Tan <jianfeng.tan@intel.com>
Acked-by: Bruce Richardson <bruce.richardson@intel.com>
New functions added to cfgfile library make it possible
to significantly simplify the code of rte_cfgfile_load_with_params()
This patch shows the new body of this function.
Signed-off-by: Jacek Piasecki <jacekx.piasecki@intel.com>
Acked-by: Bruce Richardson <bruce.richardson@intel.com>
Extend existing cfgfile library with providing new API functions:
rte_cfgfile_create() - create new cfgfile object
rte_cfgfile_add_section() - add new section to existing cfgfile
object
rte_cfgfile_add_entry() - add new entry to existing cfgfile
object in specified section
rte_cfgfile_set_entry() - update existing entry in cfgfile object
rte_cfgfile_save() - save existing cfgfile object to INI file
This modification allows to create a cfgfile on
runtime and opens up the possibility to have applications
dynamically build up a proper DPDK configuration, rather than having
to have a pre-existing one.
Signed-off-by: Jacek Piasecki <jacekx.piasecki@intel.com>
Acked-by: Bruce Richardson <bruce.richardson@intel.com>
Change to flat arrays in cfgfile struct force slightly
different data access for most of cfgfile functions.
This patch provides necessary changes in existing API.
Signed-off-by: Jacek Piasecki <jacekx.piasecki@intel.com>
Acked-by: Bruce Richardson <bruce.richardson@intel.com>
This patch removes the dependency to EAL in cfgfile library.
Signed-off-by: Jacek Piasecki <jacekx.piasecki@intel.com>
Acked-by: Bruce Richardson <bruce.richardson@intel.com>
For key search, the signatures of all entries are compared against
the signature of the key that is being looked up. Since all
signatures are contiguously put in a bucket, they can be compared
with vector instructions (AVX2), achieving higher lookup performance.
This patch adds AVX2 implementation in a separate header file.
Signed-off-by: Yipeng Wang <yipeng1.wang@intel.com>
Reviewed-by: Pablo de Lara <pablo.de.lara.guarch@intel.com>
Bloom Filter (BF) [1] is a well-known space-efficient
probabilistic data structure that answers set membership queries.
Vector of Bloom Filters (vBF) is an extension to traditional BF
that supports multi-set membership testing. Traditional BF will
return found or not-found for each key. vBF will also return
which set the key belongs to if it is found.
Since each set requires a BF, vBF should be used when set count
is small. vBF's false positive rate could be set appropriately so
that its memory requirement and lookup speed is better in certain
cases comparing to HT based set-summary.
This patch adds the vBF implementation.
[1]B H Bloom, “Space/Time Trade-offs in Hash Coding with Allowable
Errors,” Communications of the ACM, 1970.
Signed-off-by: Yipeng Wang <yipeng1.wang@intel.com>
Reviewed-by: Pablo de Lara <pablo.de.lara.guarch@intel.com>
One of the set-summary structures is hash-table based
set-summary (HTSS). One example is cuckoo filter [1].
Comparing to a traditional hash table, HTSS has a much more
compact structure. For each element, only one signature and
its corresponding set ID is stored. No key comparison is required
during lookup. For the table structure, there are multiple entries
in each bucket, and the table is composed of many buckets.
Two modes are supported for HTSS, "cache" and "none-cache" modes.
The non-cache mode is similar to the cuckoo filter [1].
When a bucket is full, one entry will be evicted to its
alternative bucket to make space for the new key. The table could
be full and then no more keys could be inserted. This mode has
false-positive rate but no false-negative. Multiple entries
with same signature could stay in the same bucket.
The "cache" mode does not evict key to its alternative bucket
when a bucket is full, an existing key will be evicted out of
the table like a cache. Thus, the table will never reject keys when
it is full. Another property is in each bucket, there cannot be
multiple entries with same signature. The mode could have both
false-positive and false-negative probability.
This patch adds the implementation of HTSS.
[1] B Fan, D G Andersen and M Kaminsky, “Cuckoo Filter: Practically
Better Than Bloom,” in Conference on emerging Networking
Experiments and Technologies, 2014.
Signed-off-by: Yipeng Wang <yipeng1.wang@intel.com>
Reviewed-by: Pablo de Lara <pablo.de.lara.guarch@intel.com>
Membership library is an extension and generalization of a traditional
filter (for example Bloom Filter and cuckoo filter) structure.
In general, the Membership library is a data structure that provides a
"set-summary" and responds to set-membership queries of whether a
certain element belongs to a set(s). A membership test for an element
will return the set this element belongs to or not-found if the
element is never inserted into the set-summary.
The results of the membership test are not 100% accurate. Certain
false positive or false negative probability could exist. However,
comparing to a "full-blown" complete list of elements, a "set-summary"
is memory efficient and fast on lookup.
This patch adds the main API definition.
Signed-off-by: Yipeng Wang <yipeng1.wang@intel.com>
Reviewed-by: Pablo de Lara <pablo.de.lara.guarch@intel.com>
IGB_UIO compilation recently got enabled for ARM64 by default
The igb_uio compilation against ARM64 based stock 4.x (e.g. 4.13)
kernel is giving compilation warnings:
igb_uio.c: In function ‘igbuio_pci_irqcontrol’:
igb_uio.c:115:25: error: implicit declaration of function
‘irq_get_irq_dat ’ [-Werror=implicit-function-declaration]
struct irq_data *irq = irq_get_irq_data(udev->info.irq);
^
igb_uio.c:115:25: error: initialization makes pointer from integer without
a cast [-Werror=int-conversion]
Fixes: d196343a258e ("igb_uio: use kernel functions for masking MSI-X")
Signed-off-by: Hemant Agrawal <hemant.agrawal@nxp.com>
Tested-by: Santosh Shukla <santosh.shukla@caviumnetworks.com>
Use rte_bsf32 and fast bit unset operation to optimize the
softrss computation.
The following measurements shows improvement over the default
softrss computation function.
tuple lens old(cycles) new(cycles)
3 1225 337
9 3743 992
Signed-off-by: Yangchao Zhou <zhouyates@gmail.com>
Reviewed-by: Vladimir Medvedkin <medvedkinv@gmail.com>
When adding a new entry in a hash table, there is
a maximum number of evictions that can be
performed. When the counter of these evictions reaches
this maximum, the entry cannot be added, as it is considered
that the algorithm has encountered an infinite loop.
The problem with the current implementation, is that this
counter was declared as a static variable.
If there are multiple threads adding entries in the same table
or in different tables, they should access different counters,
one per core and per table.
Therefore, the variable has been modified to be non-static.
Fixes: 243e93a5046f ("hash: fix unlimited cuckoo path")
Cc: stable@dpdk.org
Signed-off-by: Pablo de Lara <pablo.de.lara.guarch@intel.com>
Acked-by: Bruce Richardson <bruce.richardson@intel.com>
This is not bugfix, but it's convenient to help developer
to review and maintain the igbuio codes.
Signed-off-by: Tonghao Zhang <xiangxia.m.yue@gmail.com>
This patch adds MSI IRQ mode in a way, that should
also work on older kernel versions. The base for my patch
was an attempt to do this in cf705bc36c which was later
reverted in d8ee82745a. Compilation was tested on Linux 3.2,
4.10 and 4.12.
Signed-off-by: Markus Theil <markus.theil@tu-ilmenau.de>
Acked-by: Ferruh Yigit <ferruh.yigit@intel.com>
