vhost_poll_enqueue_completed() assumes some inflight
packets could have been completed in a previous call but
not returned to the application. But this is not the case,
since check_completed_copies callback is never called with
more than the current count as argument.
In other words, async->last_pkts_n is always 0. Removing it
greatly simplifies the function.
Signed-off-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Reviewed-by: Jiayu Hu <jiayu.hu@intel.com>
Now that IO vectors iterator have been simplified, the
rte_vhost_async_desc struct only contains a pointer on
the iterator array stored in the async metadata.
This patch removes it, and pass directly the iterators
array pointer to the transfer_data callback. Doing that,
we avoid declaring the descriptor array in the stack, and
also avoid the cost of filling it.
Signed-off-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Reviewed-by: Jiayu Hu <jiayu.hu@intel.com>
IO vectors and their iterators arrays were part of the
async metadata but not their indexes.
In order to makes this more consistent, the patch adds the
indexes to the async metadata. Doing that, we can avoid
triggering DMA transfer within the loop as it IO vector
index overflow is now prevented in the async_mbuf_to_desc()
function.
Note that previous detection mechanism was broken
since the overflow already happened when detected, so OOB
memory access would already have happened.
With this changes done, virtio_dev_rx_async_submit_split()
and virtio_dev_rx_async_submit_packed() can be further
simplified.
Signed-off-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Reviewed-by: Jiayu Hu <jiayu.hu@intel.com>
Offset and count fields are unused and so can be removed.
The offset field was actually in the Vhost example, but
in a way that does not make sense.
Signed-off-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Reviewed-by: Chenbo Xia <chenbo.xia@intel.com>
Reviewed-by: Jiayu Hu <jiayu.hu@intel.com>
This patch introduces rte_vhost_iovec struct that contains
both source and destination addresses since we always have
a 1:1 mapping between source and destination. While using
the standard iovec struct might have seemed better, having
to duplicate IO vectors and its iterators is memory
inefficient and make the implementation more complex.
Signed-off-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Reviewed-by: Chenbo Xia <chenbo.xia@intel.com>
Reviewed-by: Jiayu Hu <jiayu.hu@intel.com>
Reaching the async batch threshold was one of the condition
to trigger the DMA transfer. However, this condition was
never met since the threshold value is 32, same as the
MAX_PKT_BURST value.
Signed-off-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Reviewed-by: Chenbo Xia <chenbo.xia@intel.com>
Reviewed-by: Jiayu Hu <jiayu.hu@intel.com>
This patch splits the iterator arrays in two, one for
source and one for destination. The goal is make the code
easier to understand.
Signed-off-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Reviewed-by: Chenbo Xia <chenbo.xia@intel.com>
Reviewed-by: Jiayu Hu <jiayu.hu@intel.com>
IO vectors implementation is unnecessarily complex, mixing
source and destinations vectors in the same array.
This patch declares two arrays, one for the source and one
for the destination. It also gets rid of seg_awaits variable
in both packed and split implementation, which is the same
as iovec_idx.
Signed-off-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Reviewed-by: Chenbo Xia <chenbo.xia@intel.com>
Reviewed-by: Jiayu Hu <jiayu.hu@intel.com>
This patch moves async_inflight_info struct to internal
header since it should not be part of the API.
Signed-off-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Reviewed-by: Chenbo Xia <chenbo.xia@intel.com>
Reviewed-by: Jiayu Hu <jiayu.hu@intel.com>
This patch moves async-related metadata from vhost_virtqueue
to a dedicated struct. It makes it clear which fields are
async related, and also saves some memory when async feature
is not in use.
Signed-off-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Reviewed-by: Chenbo Xia <chenbo.xia@intel.com>
Reviewed-by: Jiayu Hu <jiayu.hu@intel.com>
Since some vdevs like virtio and vhost do not support rxq_info_get and
queue state inquiry, the error return value -ENOTSUP need to be ignored
when queue_stopped cannot get rx queue information and rx queue state.
This patch changes the return value of queue_stopped when
rte_eth_rx_queue_info_get return -ENOTSUP to support vdevs which cannot
provide rx queue information and rx queue state enable power management.
Fixes: 209fd58545 ("power: make ethdev power management thread unsafe")
Cc: stable@dpdk.org
Signed-off-by: Miao Li <miao.li@intel.com>
Acked-by: Anatoly Burakov <anatoly.burakov@intel.com>
Reviewed-by: Chenbo Xia <chenbo.xia@intel.com>
This commit defines rte_vhost_power_monitor_cond which is used to pass
some information to vhost driver. The information is including the
address to monitor, the expected value, the mask to extract value read
from 'addr', the value size of monitor address, the match flag used to
distinguish the value used to match something or not match something.
Vhost driver can use these information to fill rte_power_monitor_cond.
Signed-off-by: Miao Li <miao.li@intel.com>
Reviewed-by: Chenbo Xia <chenbo.xia@intel.com>
Acked-by: David Hunt <david.hunt@intel.com>
Async DMA map status flag was added to prevent the unnecessary unmap
when DMA devices bound to kernel driver. This brings maintenance cost
for a lot of code. This patch removes the DMA map status by using
rte_errno instead.
This patch relies on the following patch to fix a partial
unmap check in vfio unmapping API.
[1] https://www.mail-archive.com/dev@dpdk.org/msg226464.html
Signed-off-by: Xuan Ding <xuan.ding@intel.com>
Reviewed-by: Maxime Coquelin <maxime.coquelin@redhat.com>
An external project that wants to do additional checks on fields
endianness can remap rte_beXX types to instrumented types and use
sparse.
The current code breaks OVS build with sparse:
../../lib/ofp-packet.c: note: in included file (through
.../ovs/dpdk-dir/build/include/rte_flow.h, ../../lib/netdev-dpdk.h,
../../lib/dp-packet.h):
.../ovs/dpdk-dir/build/include/rte_l2tpv2.h:92:37:
error: invalid bitfield specifier for type restricted ovs_be16.
.../ovs/dpdk-dir/build/include/rte_l2tpv2.h:93:37:
error: invalid bitfield specifier for type restricted ovs_be16.
.../ovs/dpdk-dir/build/include/rte_l2tpv2.h:94:40:
error: invalid bitfield specifier for type restricted ovs_be16.
.../ovs/dpdk-dir/build/include/rte_l2tpv2.h:95:37:
error: invalid bitfield specifier for type restricted ovs_be16.
.../ovs/dpdk-dir/build/include/rte_l2tpv2.h:96:40:
error: invalid bitfield specifier for type restricted ovs_be16.
.../ovs/dpdk-dir/build/include/rte_l2tpv2.h:97:37:
error: invalid bitfield specifier for type restricted ovs_be16.
.../ovs/dpdk-dir/build/include/rte_l2tpv2.h:98:37:
error: invalid bitfield specifier for type restricted ovs_be16.
.../ovs/dpdk-dir/build/include/rte_l2tpv2.h:99:40:
error: invalid bitfield specifier for type restricted ovs_be16.
.../ovs/dpdk-dir/build/include/rte_l2tpv2.h💯39:
error: invalid bitfield specifier for type restricted ovs_be16.
make[3]: *** [lib/ofp-packet.lo] Error 1
Use simple uint16_t types for bitfields in L2TPv2 struct.
Fixes: 3a929df1f2 ("ethdev: support L2TPv2 and PPP procotol")
Signed-off-by: David Marchand <david.marchand@redhat.com>
Reviewed-by: Ferruh Yigit <ferruh.yigit@intel.com>
Warning continuously is a pain when developping or if a unit test
is/gets broken.
It could also be a problem if application behaves badly only in some
corner cases and a DoS results of those logs being continuously displayed.
Let's warn once per port and per rx/tx.
Getting such a log is scary, but let's make it more eye catching by
dumping a backtrace with it.
Tested by introducing a bug in testpmd:
static int
eth_dev_start_mp(uint16_t port_id)
{
- if (is_proc_primary())
+ if (!is_proc_primary())
return rte_eth_dev_start(port_id);
return 0;
Then, running a basic null test:
$ ./devtools/test-null.sh
...
Start automatic packet forwarding
io packet forwarding - ports=2 - cores=1 - streams=2 - NUMA support
enabled, MP allocation mode: native
Logical Core 1 (socket 0) forwards packets on 2 streams:
RX P=0/Q=0 (socket 0) -> TX P=1/Q=0 (socket 0) peer=02:00:00:00:00:01
RX P=1/Q=0 (socket 0) -> TX P=0/Q=0 (socket 0) peer=02:00:00:00:00:00
lcore 0 called rx_pkt_burst for not ready port 0
8: [build/app/dpdk-testpmd() [0x59e839]]
7: [/lib64/libc.so.6(__libc_start_main+0xf5) [0x7ff481b69555]]
6: [build/app/dpdk-testpmd(main+0x54b) [0x662d24]]
5: [build/app/dpdk-testpmd(start_packet_forwarding+0x263) [0x65e795]]
4: [build/app/dpdk-testpmd() [0x65e1be]]
3: [build/app/dpdk-testpmd() [0x65a996]]
2: [build/app/dpdk-testpmd() [0xa6cbc7]]
1: [build/app/dpdk-testpmd(rte_dump_stack+0x27) [0xaee796]]
lcore 0 called rx_pkt_burst for not ready port 1
8: [build/app/dpdk-testpmd() [0x59e839]]
7: [/lib64/libc.so.6(__libc_start_main+0xf5) [0x7ff481b69555]]
6: [build/app/dpdk-testpmd(main+0x54b) [0x662d24]]
5: [build/app/dpdk-testpmd(start_packet_forwarding+0x263) [0x65e795]]
4: [build/app/dpdk-testpmd() [0x65e1be]]
3: [build/app/dpdk-testpmd() [0x65a996]]
2: [build/app/dpdk-testpmd() [0xa6cbc7]]
1: [build/app/dpdk-testpmd(rte_dump_stack+0x27) [0xaee796]]
io packet forwarding packets/burst=32
nb forwarding cores=1 - nb forwarding ports=2
port 0: RX queue number: 1 Tx queue number: 1
Rx offloads=0x0 Tx offloads=0x0
Fixes: c87d435a4d ("ethdev: copy fast-path API into separate structure")
Signed-off-by: David Marchand <david.marchand@redhat.com>
Acked-by: Konstantin Ananyev <konstantin.ananyev@intel.com>
Tested-by: Konstantin Ananyev <konstantin.ananyev@intel.com>
Acked-by: Thomas Monjalon <thomas@monjalon.net>
Since its introduction in 2018, the SIGBUS handler was never registered,
and all related functions were unused.
A SIGBUS can be received by the application when accessing to hugepages
even if mmap() was successful, This happens especially when running
inside containers when there is not enough hugepages. In this case, we
need to recover. A similar scheme can be found in eal_memory.c.
Fixes: 582bed1e1d ("mem: support mapping hugepages at runtime")
Cc: stable@dpdk.org
Signed-off-by: Olivier Matz <olivier.matz@6wind.com>
Reviewed-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Acked-by: David Marchand <david.marchand@redhat.com>
When using rte_malloc() from a thread which is not bound to a numa
socket (the typical case is a control thread, but it can also happen
on a dataplane thread if its cpu affinity is on cores attached to
several sockets), the used heap is the one from numa socket 0, which
may not have available memory.
Fix this by selecting the first socket which has available memory.
Note: malloc_get_numa_socket() is only used from one .c file, so move
it there, and remove the inline keyword.
Fixes: b94580d688 ("malloc: avoid unknown socket id")
Cc: stable@dpdk.org
Signed-off-by: Ilyes Ben Hamouda <ilyes.ben_hamouda@6wind.com>
Signed-off-by: Olivier Matz <olivier.matz@6wind.com>
Acked-by: David Marchand <david.marchand@redhat.com>
If the user requests to use an lcore above 128 using -l,
the eal will exit with "EAL: invalid core list syntax" and
very little else useful information.
This patch adds some extra information suggesting to use --lcores
so that physical cores above RTE_MAX_LCORE (default 128) can be
used. This is achieved by using the --lcores option by mapping
the logical cores in the application to physical cores.
For example, if "-l 12-16,130,132" is used, we see the following
additional output on the command line:
EAL: lcore 132 >= RTE_MAX_LCORE (128)
EAL: lcore 133 >= RTE_MAX_LCORE (128)
EAL: To use high physical core ids, please use --lcores to map them
to lcore ids below RTE_MAX_LCORE,
EAL: e.g. --lcores 0@12,1@13,2@14,3@15,4@16,5@132,6@133
The same is added to -c option parsing.
For example, if "-c 0x300000000000000000000000000000000" is
used, we see the following additional output on the command line:
EAL: lcore 128 >= RTE_MAX_LCORE (128)
EAL: lcore 129 >= RTE_MAX_LCORE (128)
EAL: To use high physical core ids, please use --lcores to map them
to lcore ids below RTE_MAX_LCORE,
EAL: e.g. --lcores 0@128,1@129
Signed-off-by: David Hunt <david.hunt@intel.com>
Acked-by: Bruce Richardson <bruce.richardson@intel.com>
This API has been around for more than a year (and is in LTS 20.11).
It did not receive negative feedback and will be used in a next OVS
release.
Mark it stable.
Signed-off-by: David Marchand <david.marchand@redhat.com>
Acked-by: Kevin Traynor <ktraynor@redhat.com>
Acked-by: Thomas Monjalon <thomas@monjalon.net>
rte_bpf_convert() implementation depends on libpcap.
Right now it is defined only when this library is installed and
RTE_PORT_PCAP is defined.
Fix that by providing for such case stub rte_bpf_convert()
implementation that will always return an error.
To draw user attention, if proper implementation is disabled,
warning will be thrown at meson configure stage.
Also move stub for another function (rte_bpf_elf_load) into
the same place (bpf_stub.c).
Fixes: 2eccf6afbe ("bpf: add function to convert classic BPF to DPDK BPF")
Signed-off-by: Konstantin Ananyev <konstantin.ananyev@intel.com>
Fix typo in doxygen comments for rte_bpf_convert().
Fixes: 2eccf6afbe ("bpf: add function to convert classic BPF to DPDK BPF")
Signed-off-by: Konstantin Ananyev <konstantin.ananyev@intel.com>
A previous change blamed it on compiler/ASan, while this is a real
(yet minor) issue.
This return -EINVAL is never reached since we test all combinations of
fidx and fcin booleans.
All branches end up with a return 0, factorize them.
Fixes: 84f5ac9418 ("pipeline: fix build with ASan")
Fixes: f38913b7fb ("pipeline: add meter array to SWX")
Signed-off-by: David Marchand <david.marchand@redhat.com>
Acked-by: Cristian Dumitrescu <cristian.dumitrescu@intel.com>
Implement PIE based congestion management based on rfc8033.
The Proportional Integral Controller Enhanced (PIE) algorithm works
by proactively dropping packets randomly.
PIE is implemented as more advanced queue management is required to
address the bufferbloat problem and provide desirable quality of
service to users.
Tests for PIE code added to test application.
Added PIE related information to documentation.
Signed-off-by: Wojciech Liguzinski <wojciechx.liguzinski@intel.com>
Acked-by: Cristian Dumitrescu <cristian.dumitrescu@intel.com>
Acked-by: Jasvinder Singh <jasvinder.singh@intel.com>
The device event interrupt handler was always freed.
Bugzilla ID: 845
Fixes: c2bd9367e1 ("lib: remove direct access to interrupt handle")
Signed-off-by: David Marchand <david.marchand@redhat.com>
Tested-by: Yan Xia <yanx.xia@intel.com>
Reviewed-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Caught with ASan:
==9727==ERROR: AddressSanitizer: stack-buffer-overflow on address
0x7f0daa2fc0d0 at pc 0x7f0daeefacb2 bp 0x7f0daa2fadd0 sp 0x7f0daa2fa578
READ of size 1 at 0x7f0daa2fc0d0 thread T1
#0 0x7f0daeefacb1 (/lib64/libasan.so.5+0xbacb1)
#1 0x115eba1 in dev_uev_parse ../lib/eal/linux/eal_dev.c:167
#2 0x115f281 in dev_uev_handler ../lib/eal/linux/eal_dev.c:248
#3 0x1169b91 in eal_intr_process_interrupts
../lib/eal/linux/eal_interrupts.c:1026
#4 0x116a3a2 in eal_intr_handle_interrupts
../lib/eal/linux/eal_interrupts.c:1100
#5 0x116a7f0 in eal_intr_thread_main
../lib/eal/linux/eal_interrupts.c:1172
#6 0x112640a in ctrl_thread_init
../lib/eal/common/eal_common_thread.c:202
#7 0x7f0dade27159 in start_thread (/lib64/libpthread.so.0+0x8159)
#8 0x7f0dadb58f72 in clone (/lib64/libc.so.6+0xfcf72)
Address 0x7f0daa2fc0d0 is located in stack of thread T1 at offset 4192
in frame
#0 0x115f0c9 in dev_uev_handler ../lib/eal/linux/eal_dev.c:226
This frame has 2 object(s):
[32, 48) 'uevent'
[96, 4192) 'buf' <== Memory access at offset 4192 overflows this
variable
HINT: this may be a false positive if your program uses some custom
stack unwind mechanism or swapcontext
(longjmp and C++ exceptions *are* supported)
Thread T1 created by T0 here:
#0 0x7f0daee92ea3 in __interceptor_pthread_create
(/lib64/libasan.so.5+0x52ea3)
#1 0x1126542 in rte_ctrl_thread_create
../lib/eal/common/eal_common_thread.c:228
#2 0x116a8b5 in rte_eal_intr_init
../lib/eal/linux/eal_interrupts.c:1200
#3 0x1159dd1 in rte_eal_init ../lib/eal/linux/eal.c:1044
#4 0x7a22f8 in main ../app/test-pmd/testpmd.c:4105
#5 0x7f0dada7f802 in __libc_start_main (/lib64/libc.so.6+0x23802)
Bugzilla ID: 792
Fixes: 0d0f478d04 ("eal/linux: add uevent parse and process")
Cc: stable@dpdk.org
Signed-off-by: David Marchand <david.marchand@redhat.com>
Tested-by: Yan Xia <yanx.xia@intel.com>
Reviewed-by: Maxime Coquelin <maxime.coquelin@redhat.com>
clang-13 rightfully complains that the total_mem variable in
eal_parse_socket_arg is set but not used, since the final
accumulated total_mem result isn't used anywhere.
So just remove the total_mem variable.
Fixes: 0a703f0f36 ("eal/linux: fix parsing zero socket memory and limits")
Signed-off-by: Jim Harris <james.r.harris@intel.com>
Reviewed-by: David Marchand <david.marchand@redhat.com>
This patch adds a new parameter to the FIB configuration to specify
the size of the extension for internal RIB structure.
Signed-off-by: Vladimir Medvedkin <vladimir.medvedkin@intel.com>
Tested-by: Conor Walsh <conor.walsh@intel.com>
Devargs used in device iterator initialization wasn't set to zero, random
data like bus string lead to invalid address access.
This patch initializes devargs.
Bugzilla ID: 862
Fixes: c99a2d4c6b ("eal: implement device iteration initialization")
Cc: stable@dpdk.org
Signed-off-by: Xueming Li <xuemingl@nvidia.com>
This patch fixes use after free in thash library, reported by ASAN.
Bugzilla ID: 868
Fixes: 28ebff11c2 ("hash: add predictable RSS")
Cc: stable@dpdk.org
Signed-off-by: Vladimir Medvedkin <vladimir.medvedkin@intel.com>
Reviewed-by: David Marchand <david.marchand@redhat.com>
This patch enables new GFNI Toeplitz hash in
predictable RSS library.
Signed-off-by: Vladimir Medvedkin <vladimir.medvedkin@intel.com>
Acked-by: Konstantin Ananyev <konstantin.ananyev@intel.com>
This patch adds a bulk version for the Toeplitz hash implemented
with Galios Fields New Instructions (GFNI).
Signed-off-by: Vladimir Medvedkin <vladimir.medvedkin@intel.com>
Acked-by: Konstantin Ananyev <konstantin.ananyev@intel.com>
This patch add a new Toeplitz hash implementation using
Galios Fields New Instructions (GFNI).
Signed-off-by: Vladimir Medvedkin <vladimir.medvedkin@intel.com>
Acked-by: Konstantin Ananyev <konstantin.ananyev@intel.com>
FreeBSD EAL selected IOVA mode PA even in --no-huge mode
where PA are not available. Memory zones were created with IOVA
equal to RTE_BAD_IOVA with no indication this field is not usable.
Change IOVA mode detection:
1. Always allow to force --iova-mode=va.
2. In --no-huge mode, disallow forcing --iova-mode=pa, and select VA.
3. Otherwise select IOVA mode according to bus requests, default to PA.
In case contigmem is inaccessible, memory initialization will fail
with a message indicating the cause.
Fixes: c2361bab70 ("eal: compute IOVA mode based on PA availability")
Cc: stable@dpdk.org
Signed-off-by: Dmitry Kozlyuk <dkozlyuk@nvidia.com>
Acked-by: Bruce Richardson <bruce.richardson@intel.com>
Instead of polling for bufptr64 to be updated, use
wait until scheme for this case.
Signed-off-by: Feifei Wang <feifei.wang2@arm.com>
Reviewed-by: Ruifeng Wang <ruifeng.wang@arm.com>
Acked-by: Jerin Jacob <jerinj@marvell.com>
Instead of polling for cbi->use to be updated, use wait until scheme.
Signed-off-by: Feifei Wang <feifei.wang2@arm.com>
Reviewed-by: Ruifeng Wang <ruifeng.wang@arm.com>
Acked-by: Konstantin Ananyev <konstantin.ananyev@intel.com>
Instead of polling for mcslock to be updated, use wait until scheme
for this case.
Signed-off-by: Feifei Wang <feifei.wang2@arm.com>
Reviewed-by: Ruifeng Wang <ruifeng.wang@arm.com>
Acked-by: Jerin Jacob <jerinj@marvell.com>
Instead of polling for read pflock update, use wait until scheme for
this case.
Signed-off-by: Feifei Wang <feifei.wang2@arm.com>
Reviewed-by: Ruifeng Wang <ruifeng.wang@arm.com>
Acked-by: Jerin Jacob <jerinj@marvell.com>
Add a new generic helper which is a macro for wait until scheme.
Furthermore, to prevent compilation warning in arm:
----------------------------------------------
'warning: implicit declaration of function ...'
----------------------------------------------
Delete 'undef' constructions for '__LOAD_EXC_xx', '__SEVL' and '__WFE'.
And add ‘__RTE_ARM’ for these macros to fix the namespace.
This is because original macros are undefine at the end of the file.
If the new macro calls them in other files, they will be seen as
'not defined'.
Signed-off-by: Feifei Wang <feifei.wang2@arm.com>
Reviewed-by: Ruifeng Wang <ruifeng.wang@arm.com>
Acked-by: Konstantin Ananyev <konstantin.ananyev@intel.com>
Acked-by: Jerin Jacob <jerinj@marvell.com>
rte_pdump_init() always allocates new memzone for pdump_stats.
Though rte_pdump_uninit() never frees it.
So the following combination will always fail:
rte_pdump_init(); rte_pdump_uninit(); rte_pdump_init();
The issue was caught by pdump_autotest UT.
While first test run successful, any consecutive runs
of this test-case will fail.
Fix the issue by calling rte_memzone_free() for statistics memzone.
Fixes: 10f726efe2 ("pdump: support pcapng and filtering")
Signed-off-by: Konstantin Ananyev <konstantin.ananyev@intel.com>
Acked-by: Reshma Pattan <reshma.pattan@intel.com>
Acked-by: Stephen Hemminger <stephen@networkplumber.org>
If packet dump was enabled via pdump_enable_by_deviceid
the packet snapshot length was not being set.
Bugzilla ID: 840
Fixes: 10f726efe2 ("pdump: support pcapng and filtering")
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
RTE_ prefix was added by
commit 295968d174 ("ethdev: add namespace")
Fixes: 8d23ce8f5e ("pcapng: add new library for writing pcapng files")
Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
This patch adds necessary hooks in the memory allocator for ASan.
This feature is currently available in DPDK only on Linux x86_64.
If other OS/architectures want to support it, ASAN_SHADOW_OFFSET must be
defined and RTE_MALLOC_ASAN must be set accordingly in meson.
Signed-off-by: Xueqin Lin <xueqin.lin@intel.com>
Signed-off-by: Zhihong Peng <zhihongx.peng@intel.com>
Acked-by: Anatoly Burakov <anatoly.burakov@intel.com>
Code changes to avoid the following build error:
"Control reaches end of non-void function".
Signed-off-by: Xueqin Lin <xueqin.lin@intel.com>
Signed-off-by: Zhihong Peng <zhihongx.peng@intel.com>
Acked-by: Cristian Dumitrescu <cristian.dumitrescu@intel.com>
Partial unmap support was introduced in commit c13ca4e81c
("vfio: fix DMA mapping granularity for IOVA as VA"), and with it
was added a check that dereferenced the IOMMU type to determine whether
partial ummapping is supported for currently configured IOMMU type. In
certain circumstances (such as when VFIO is supported, but no devices
were bound to the VFIO driver), the IOMMU type pointer can be NULL.
However, dereferencing of IOMMU type was guarded by access to the user
maps list - that is, we were always checking the user map list first,
and then, if we found a memory region that encloses the one we're trying
to unmap, we would have performed the IOMMU type check.
This ensured that the IOMMU type check will not cause any NULL pointer
dereferences, because in order for an IOMMU type check to have been
performed, there necessarily must have been at least one memory region
that was previously mapped successfully, and that implies having a
defined IOMMU type.
When commit 56259f7fc0 ("vfio: allow partially unmapping adjacent
memory") was introduced, the IOMMU type check was moved to
before we were traversing the user mem maps list, thereby introducing a
potential NULL dereference, because the IOMMU type access was no longer
guarded by the user mem maps list traversal.
Fix the issue by moving the IOMMU type check to after the user mem maps
traversal, thereby ensuring that by the time the check happens, the
IOMMU type is always valid.
Fixes: 56259f7fc0 ("vfio: allow partially unmapping adjacent memory")
Signed-off-by: Anatoly Burakov <anatoly.burakov@intel.com>
Reviewed-by: David Marchand <david.marchand@redhat.com>
Tested-by: Xuan Ding <xuan.ding@intel.com>
Remove the usage of pthread barrier and replace it with
synchronization using atomic variable.
This also removes the use of reference count required to synchronize
freeing the memory.
Signed-off-by: Honnappa Nagarahalli <honnappa.nagarahalli@arm.com>
Reviewed-by: Olivier Matz <olivier.matz@6wind.com>
Dynamically allocating the efds and elist array of intr_handle
structure, based on size provided by user. Eg size can be
MSIX interrupts supported by a PCI device.
Signed-off-by: Harman Kalra <hkalra@marvell.com>
Signed-off-by: David Marchand <david.marchand@redhat.com>
Acked-by: Dmitry Kozlyuk <dmitry.kozliuk@gmail.com>
Tested-by: Raslan Darawsheh <rasland@nvidia.com>
VFIO/UIO are mutually exclusive, storing file descriptor in a single
field is enough.
Signed-off-by: Harman Kalra <hkalra@marvell.com>
Signed-off-by: David Marchand <david.marchand@redhat.com>
Tested-by: Raslan Darawsheh <rasland@nvidia.com>
Moving interrupt handle structure definition inside a EAL private
header to make its fields totally opaque to the outside world.
Signed-off-by: Harman Kalra <hkalra@marvell.com>
Signed-off-by: David Marchand <david.marchand@redhat.com>
Tested-by: Raslan Darawsheh <rasland@nvidia.com>
Removing direct access to interrupt handle structure fields,
rather use respective get set APIs for the same.
Making changes to all the drivers access the interrupt handle fields.
Signed-off-by: Harman Kalra <hkalra@marvell.com>
Acked-by: Hyong Youb Kim <hyonkim@cisco.com>
Signed-off-by: David Marchand <david.marchand@redhat.com>
Tested-by: Raslan Darawsheh <rasland@nvidia.com>
Removing direct access to interrupt handle structure fields,
rather use respective get set APIs for the same.
Making changes to all the libraries access the interrupt handle fields.
Signed-off-by: Harman Kalra <hkalra@marvell.com>
Signed-off-by: David Marchand <david.marchand@redhat.com>
Tested-by: Raslan Darawsheh <rasland@nvidia.com>
Removing direct access to interrupt handle structure fields,
rather use respective get set APIs for the same.
Making changes to all the libraries access the interrupt handle fields.
Implementing alarm cleanup routine, where the memory allocated
for interrupt instance can be freed.
Signed-off-by: Harman Kalra <hkalra@marvell.com>
Signed-off-by: David Marchand <david.marchand@redhat.com>
Tested-by: Raslan Darawsheh <rasland@nvidia.com>
Making changes to the interrupt framework to use interrupt handle
APIs to get/set any field.
Signed-off-by: Harman Kalra <hkalra@marvell.com>
Signed-off-by: David Marchand <david.marchand@redhat.com>
Tested-by: Raslan Darawsheh <rasland@nvidia.com>
Prototype/Implement get set APIs for interrupt handle fields.
User won't be able to access any of the interrupt handle fields
directly while should use these get/set APIs to access/manipulate
them.
Internal interrupt header i.e. rte_eal_interrupt.h is rearranged,
as APIs defined are moved to rte_interrupts.h and epoll specific
definitions are moved to a new header rte_epoll.h.
Later in the series rte_eal_interrupt.h will be removed.
Signed-off-by: Harman Kalra <hkalra@marvell.com>
Acked-by: Ray Kinsella <mdr@ashroe.eu>
Acked-by: Dmitry Kozlyuk <dmitry.kozliuk@gmail.com>
Signed-off-by: David Marchand <david.marchand@redhat.com>
Tested-by: Raslan Darawsheh <rasland@nvidia.com>
Windows EAL did not detect IOVA mode and worked incorrectly
if physical addresses could not be obtained
(if virt2phys driver was missing or inaccessible).
In this case, rte_mem_virt2iova() reported RTE_BAD_IOVA for any address.
Inability to obtain IOVA, be it PA or VA, should cause a failure
for the DPDK allocator, but it was hidden by the implementation,
so allocations did not fail when they should.
The mode when DPDK cannot obtain PA but can work is IOVA-as-VA mode.
However, rte_eal_iova_mode() always returned RTE_IOVA_DC
(while it should only ever return RTE_IOVA_PA or RTE_IOVA_VA),
because IOVA mode detection was not implemented.
Implement IOVA mode detection:
1. Always allow to force --iova-mode=va.
2. Allow to force --iova-mode=pa only if virt2phys is available.
3. If no mode is forced and virt2phys is available,
select the mode according to bus requests, default to PA.
4. If no mode is forced but virt2phys is unavailable, default to VA.
Fix rte_mem_virt2iova() by returning VA when using IOVA-as-VA.
Fix rte_eal_iova_mode() by returning the selected mode.
Fixes: 2a5d547a4a ("eal/windows: implement basic memory management")
Cc: stable@dpdk.org
Reported-by: Tal Shnaiderman <talshn@nvidia.com>
Signed-off-by: Dmitry Kozlyuk <dkozlyuk@nvidia.com>
Tested-by: Pallavi Kadam <pallavi.kadam@intel.com>
Acked-by: Pallavi Kadam <pallavi.kadam@intel.com>
This patch fixes buffer overflow reported by ASAN,
please reference https://bugs.dpdk.org/show_bug.cgi?id=819
The rte_lpm6 keeps routing information for control plane purpose
inside the rte_hash table which uses rte_jhash() as a hash function.
From the rte_jhash() documentation: If input key is not aligned to
four byte boundaries or a multiple of four bytes in length,
the memory region just after may be read (but not used in the
computation).
rte_lpm6 uses 17 bytes keys consisting of IPv6 address (16 bytes) +
depth (1 byte).
This patch increases the size of the depth field up to uint32_t
and sets the alignment to 4 bytes.
Bugzilla ID: 819
Fixes: 86b3b21952 ("lpm6: store rules in hash table")
Cc: stable@dpdk.org
Signed-off-by: Vladimir Medvedkin <vladimir.medvedkin@intel.com>
Acked-by: Bruce Richardson <bruce.richardson@intel.com>
Fixes: 7574c3ef74 ("hash: add toeplitz algorithm used by RSS")
Cc: stable@dpdk.org
Signed-off-by: Vladimir Medvedkin <vladimir.medvedkin@intel.com>
Acked-by: John McNamara <john.mcnamara@intel.com>
Ensure that the memory operations before the call to
rte_eal_remote_launch are visible to the worker thread.
Use the function pointer to execute in worker thread
as the guard variable.
Ensure that the memory operations in worker thread, that happen
before it returns the status of the assigned function, are
visible to the main thread. Use the variable containing the
lcore's state as the guard variable.
Signed-off-by: Honnappa Nagarahalli <honnappa.nagarahalli@arm.com>
Reviewed-by: Ola Liljedahl <ola.liljedahl@arm.com>
Reviewed-by: Feifei Wang <feifei.wang2@arm.com>
FINISHED state seems to be used to indicate that the worker's update
of the 'state' is not visible to other threads. There seems to be no
requirement to have such a state.
Since the FINISHED state is removed, the API rte_eal_wait_lcore
is updated to always return the status of the last function that
ran in the worker core.
Signed-off-by: Honnappa Nagarahalli <honnappa.nagarahalli@arm.com>
Reviewed-by: Ola Liljedahl <ola.liljedahl@arm.com>
Reviewed-by: Feifei Wang <feifei.wang2@arm.com>
In the rte_eal_remote_launch function, the lcore function
pointer is checked for NULL. However, the pointer is never
reset to NULL. Reset the lcore function pointer and argument
after the worker has completed executing the lcore function.
Fixes: af75078fec ("first public release")
Cc: stable@dpdk.org
Signed-off-by: Honnappa Nagarahalli <honnappa.nagarahalli@arm.com>
Reviewed-by: Ruifeng Wang <ruifeng.wang@arm.com>
Reviewed-by: Feifei Wang <feifei.wang2@arm.com>
Functions and macros in x86 rte_memcpy.h may cause cast-align warnings,
when using strict cast align flag with supporting gcc:
gcc (Ubuntu 9.3.0-17ubuntu1~20.04) 9.3.0
CFLAGS="-Wcast-align=strict" make V=1 -C examples/l2fwd clean static
For example:
In file included from main.c:24:
/dpdk/build/include/rte_memcpy.h: In function 'rte_mov16':
/dpdk/build/include/rte_memcpy.h:306:25: warning: cast increases
required alignment of target type [-Wcast-align]
306 | xmm0 = _mm_loadu_si128((const __m128i *)src);
| ^
As the code assumes correct alignment, add first a (void *) or (const
void *) castings, to avoid the warnings.
Fixes: 9484092baa ("eal/x86: optimize memcpy for AVX512 platforms")
Cc: stable@dpdk.org
Signed-off-by: Eli Britstein <elibr@nvidia.com>
In rte_pktmbuf_mtod_offset macro, there is a casting from char * to type
't', which may cause cast-align warning when using strict cast align
flag with supporting gcc:
gcc (Ubuntu 9.3.0-17ubuntu1~20.04) 9.3.0
CFLAGS="-Wcast-align=strict" make V=1 -C examples/l2fwd clean static
main.c: In function 'l2fwd_mac_updating':
/dpdk/build/include/rte_mbuf_core.h:719:3: warning: cast increases
required alignment of target type [-Wcast-align]
719 | ((t)((char *)(m)->buf_addr + (m)->data_off + (o)))
| ^
/dpdk/build/include/rte_mbuf_core.h:733:32: note: in expansion of macro
'rte_pktmbuf_mtod_offset'
733 | #define rte_pktmbuf_mtod(m, t) rte_pktmbuf_mtod_offset(m, t, 0)
| ^~~~~~~~~~~~~~~~~~~~~~~
As the code assumes correct alignment, add first a (void *) casting, to
avoid the warning.
Fixes: af75078fec ("first public release")
Cc: stable@dpdk.org
Signed-off-by: Eli Britstein <elibr@nvidia.com>
Acked-by: Olivier Matz <olivier.matz@6wind.com>
In rte_vlan_insert there is a casting of rte_pktmbuf_prepend returned
value to (struct rte_ether_hdr *), which causes cast-align warning when
using strict cast align flag with supporting gcc:
gcc (Ubuntu 9.3.0-17ubuntu1~20.04) 9.3.0
CFLAGS="-Wcast-align=strict" make V=1 -C examples/l2fwd clean static
In file included from main.c:35:
/dpdk/build/include/rte_ether.h:370:7: warning: cast increases required
alignment of target type [-Wcast-align]
370 | nh = (struct rte_ether_hdr *)
| ^
As the code assumes correct alignment, add first a (void *) casting, to
avoid the warning.
Fixes: c974021a59 ("ether: add soft vlan encap/decap")
Cc: stable@dpdk.org
Signed-off-by: Eli Britstein <elibr@nvidia.com>
Acked-by: Olivier Matz <olivier.matz@6wind.com>
When mempool had been created with RTE_MEMPOOL_F_NO_IOVA_CONTIG flag
but later populated with valid IOVA, RTE_MEMPOOL_F_NON_IO was unset,
while it should be kept. The unit test did not catch this
because rte_mempool_populate_default() it used was populating
with RTE_BAD_IOVA.
Keep setting RTE_MEMPOOL_NON_IO at an empty mempool creation
and add an assert for it in the unit test (remove the separate case).
Do not reset the flag if RTE_MEMPOOL_F_ON_IOVA_CONTIG is set.
Fixes: 11541c5c81 ("mempool: add non-IO flag")
Signed-off-by: Dmitry Kozlyuk <dkozlyuk@nvidia.com>
Acked-by: Olivier Matz <olivier.matz@6wind.com>
This API was introduced in 18.05, therefore removing
experimental tag to promote it to stable state
Signed-off-by: Jasvinder Singh <jasvinder.singh@intel.com>
Acked-by: Cristian Dumitrescu <cristian.dumitrescu@intel.com>
Acked-by: Ray Kinsella <mdr@ashroe.eu>
Enable restricting the scope of an action to regular table entries or
to the table default entry in order to support the P4 language
tableonly or defaultonly annotations.
Signed-off-by: Yogesh Jangra <yogesh.jangra@intel.com>
Acked-by: Cristian Dumitrescu <cristian.dumitrescu@intel.com>
Add support for configurable number of loops through the input PCAP
file for the source port. Added an additional parameter to source
port CLI command.
Signed-off-by: Yogesh Jangra <yogesh.jangra@intel.com>
Acked-by: Cristian Dumitrescu <cristian.dumitrescu@intel.com>
The instruction_data array was incorrectly indexed, which resulted in
the array index getting out of bounds and sometimes segfault.
Fixes: a1711f (“pipeline: add SWX Rx and extract instructions“)
Cc: stable@dpdk.org
Signed-off-by: Yogesh Jangra <yogesh.jangra@intel.com>
Acked-by: Cristian Dumitrescu <cristian.dumitrescu@intel.com>
Build is broken on RHEL7 following introduction of this new protocol.
Fixes: 3a929df1f2 ("ethdev: support L2TPv2 and PPP procotol")
Signed-off-by: David Marchand <david.marchand@redhat.com>
Tested-by: Raslan Darawsheh <rasland@nvidia.com>
Fix the mbuf offload flags namespace by adding an RTE_ prefix to the
name. The old flags remain usable, but a deprecation warning is issued
at compilation.
Signed-off-by: Olivier Matz <olivier.matz@6wind.com>
Acked-by: Andrew Rybchenko <andrew.rybchenko@oktetlabs.ru>
Acked-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
Acked-by: Somnath Kotur <somnath.kotur@broadcom.com>
The flags PKT_TX_VLAN_PKT and PKT_TX_QINQ_PKT are
marked as deprecated since commit 380a7aab1a ("mbuf: rename deprecated
VLAN flags") (2017). But they were not using the RTE_DEPRECATED
macro, because it did not exist at this time. Add it, and replace
usage of these flags.
Signed-off-by: Olivier Matz <olivier.matz@6wind.com>
Acked-by: Andrew Rybchenko <andrew.rybchenko@oktetlabs.ru>
Acked-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
The flags PKT_RX_L4_CKSUM_BAD and PKT_RX_IP_CKSUM_BAD are defined
twice with the same value. Remove one of the occurrence, which was
marked as "deprecated".
Signed-off-by: Olivier Matz <olivier.matz@6wind.com>
Acked-by: Andrew Rybchenko <andrew.rybchenko@oktetlabs.ru>
Set correct tunnel type telemetry text - tunnel type
was wrongly set as IPv4-UDP for all types.
Fixes: bf5b65a8e781 ("ipsec: support SA telemetry")
Signed-off-by: Radu Nicolau <radu.nicolau@intel.com>
Acked-by: Akhil Goyal <gakhil@marvell.com>
The device specific structures - rte_cryptodev
and rte_cryptodev_data are moved to cryptodev_pmd.h
to hide it from the applications.
Signed-off-by: Akhil Goyal <gakhil@marvell.com>
Tested-by: Rebecca Troy <rebecca.troy@intel.com>
Acked-by: Fan Zhang <roy.fan.zhang@intel.com>
Acked-by: Konstantin Ananyev <konstantin.ananyev@intel.com>
Rework fast-path cryptodev functions to use rte_crypto_fp_ops[].
While it is an API/ABI breakage, this change is intended to be
transparent for both users (no changes in user app is required) and
PMD developers (no changes in PMD is required).
Signed-off-by: Akhil Goyal <gakhil@marvell.com>
Acked-by: Fan Zhang <roy.fan.zhang@intel.com>
Acked-by: Konstantin Ananyev <konstantin.ananyev@intel.com>
Added a rte_cryptodev_pmd_probing_finish API which
need to be called by the PMD after the device is initialized
completely. This will set the fast path function pointers
in the flat array for secondary process. For primary process,
these are set in rte_cryptodev_start.
Signed-off-by: Akhil Goyal <gakhil@marvell.com>
Acked-by: Konstantin Ananyev <konstantin.ananyev@intel.com>
Acked-by: Fan Zhang <roy.fan.zhang@intel.com>
Move fastpath inline function pointers from rte_cryptodev into a
separate structure accessed via a flat array.
The intention is to make rte_cryptodev and related structures private
to avoid future API/ABI breakages.
Signed-off-by: Akhil Goyal <gakhil@marvell.com>
Tested-by: Rebecca Troy <rebecca.troy@intel.com>
Acked-by: Fan Zhang <roy.fan.zhang@intel.com>
Acked-by: Konstantin Ananyev <konstantin.ananyev@intel.com>
At queue_pair config stage, allocate memory for maximum
number of queue pair pointers that a device can support.
This will allow fast path APIs(enqueue_burst/dequeue_burst) to
refer pointer to internal QP data without checking for currently
configured QPs.
This is required to hide the rte_cryptodev and rte_cryptodev_data
structure from user.
Signed-off-by: Akhil Goyal <gakhil@marvell.com>
Acked-by: Fan Zhang <roy.fan.zhang@intel.com>
Acked-by: Konstantin Ananyev <konstantin.ananyev@intel.com>
A new header file rte_cryptodev_core.h is added and all
internal data structures which need not be exposed directly to
application are moved to this file. These structures are mostly
used by drivers, but they need to be in the public header file
as they are accessed by datapath inline functions for
performance reasons.
Signed-off-by: Akhil Goyal <gakhil@marvell.com>
Tested-by: Rebecca Troy <rebecca.troy@intel.com>
Acked-by: Fan Zhang <roy.fan.zhang@intel.com>
Acked-by: Konstantin Ananyev <konstantin.ananyev@intel.com>
The macros RTE_BIT32 and RTE_BIT64 are used to replace single bit masks.
Do not switch VLAN offload flags since type is not fixed size.
Signed-off-by: Andrew Rybchenko <andrew.rybchenko@oktetlabs.ru>
Reviewed-by: Ferruh Yigit <ferruh.yigit@intel.com>
Add 'RTE_ETH' namespace to all enums & macros in a backward compatible
way. The macros for backward compatibility can be removed in next LTS.
Also updated some struct names to have 'rte_eth' prefix.
All internal components switched to using new names.
Syntax fixed on lines that this patch touches.
Signed-off-by: Ferruh Yigit <ferruh.yigit@intel.com>
Acked-by: Tyler Retzlaff <roretzla@linux.microsoft.com>
Acked-by: Andrew Rybchenko <andrew.rybchenko@oktetlabs.ru>
Acked-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
Acked-by: Jerin Jacob <jerinj@marvell.com>
Acked-by: Wisam Jaddo <wisamm@nvidia.com>
Acked-by: Rosen Xu <rosen.xu@intel.com>
Acked-by: Chenbo Xia <chenbo.xia@intel.com>
Acked-by: Hemant Agrawal <hemant.agrawal@nxp.com>
Acked-by: Somnath Kotur <somnath.kotur@broadcom.com>
rte_eth_dev_configure() always sets MTU to either dev_conf.rxmode.mtu
or RTE_ETHER_MTU if application doesn't provide the value.
So, there is no point to allow rte_eth_dev_set_mtu() before since
set value will be overwritten on configure anyway.
Fixes: 1bb4a528c4 ("ethdev: fix max Rx packet length")
Signed-off-by: Ivan Ilchenko <ivan.ilchenko@oktetlabs.ru>
Signed-off-by: Andrew Rybchenko <andrew.rybchenko@oktetlabs.ru>
Reviewed-by: Ferruh Yigit <ferruh.yigit@intel.com>
This patch adds API to return name of device capability.
Signed-off-by: Xueming Li <xuemingl@nvidia.com>
Reviewed-by: Andrew Rybchenko <andrew.rybchenko@oktetlabs.ru>
Acked-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
Acked-by: Thomas Monjalon <thomas@monjalon.net>
In current DPDK framework, each Rx queue is pre-loaded with mbufs to
save incoming packets. For some PMDs, when number of representors scale
out in a switch domain, the memory consumption became significant.
Polling all ports also leads to high cache miss, high latency and low
throughput.
This patch introduces shared Rx queue. Ports in same Rx domain and
switch domain could share Rx queue set by specifying non-zero sharing
group in Rx queue configuration.
Shared Rx queue is identified by share_rxq field of Rx queue
configuration. Port A RxQ X can share RxQ with Port B RxQ Y by using
same shared Rx queue ID.
No special API is defined to receive packets from shared Rx queue.
Polling any member port of a shared Rx queue receives packets of that
queue for all member ports, port_id is identified by mbuf->port. PMD is
responsible to resolve shared Rx queue from device and queue data.
Shared Rx queue must be polled in same thread or core, polling a queue
ID of any member port is essentially same.
Multiple share groups are supported. PMD should support mixed
configuration by allowing multiple share groups and non-shared Rx queue
on one port.
Example grouping and polling model to reflect service priority:
Group1, 2 shared Rx queues per port: PF, rep0, rep1
Group2, 1 shared Rx queue per port: rep2, rep3, ... rep127
Core0: poll PF queue0
Core1: poll PF queue1
Core2: poll rep2 queue0
PMD advertise shared Rx queue capability via RTE_ETH_DEV_CAPA_RXQ_SHARE.
PMD is responsible for shared Rx queue consistency checks to avoid
member port's configuration contradict each other.
Signed-off-by: Xueming Li <xuemingl@nvidia.com>
Reviewed-by: Andrew Rybchenko <andrew.rybchenko@oktetlabs.ru>
Acked-by: Ajit Khaparde <ajit.khaparde@broadcom.com>
In secondary process, rte_eth_dev_close() doesn't clear eth_dev->data.
If calling rte_dev_remove() after rte_eth_dev_close(), in
rte_eth_dev_pci_generic_remove() function, the released eth device still
can be found by its name in shared memory. As a result, the eth device
will be released repeatedly. The state of the eth device is modified to
RTE_ETH_DEV_UNUSED after rte_eth_dev_close(). So this state can be used
to avoid this problem.
Fixes: dcd5c8112b ("ethdev: add PCI driver helpers")
Cc: stable@dpdk.org
Signed-off-by: Huisong Li <lihuisong@huawei.com>
Reviewed-by: Ferruh Yigit <ferruh.yigit@intel.com>
The use of IOMMU has many advantages, such as isolation and address
translation. This patch extends the capability of DMA engine to use
IOMMU if the DMA engine is bound to vfio.
When set memory table, the guest memory will be mapped
into the default container of DPDK.
Signed-off-by: Xuan Ding <xuan.ding@intel.com>
Tested-by: Yvonne Yang <yvonnex.yang@intel.com>
Reviewed-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Currently, if we map a memory area A, then map a separate memory area B
that by coincidence happens to be adjacent to A, current implementation
will merge these two segments into one, and if partial unmapping is not
supported, these segments will then be only allowed to be unmapped in
one go. In other words, given segments A and B that are adjacent, it
is currently not possible to map A, then map B, then unmap A.
Fix this by adding a notion of "chunk size", which will allow
subdividing segments into equally sized segments whenever we are dealing
with an IOMMU that does not support partial unmapping. With this change,
we will still be able to merge adjacent segments, but only if they are
of the same size. If we keep with our above example, adjacent segments A
and B will be stored as separate segments if they are of different
sizes.
Signed-off-by: Anatoly Burakov <anatoly.burakov@intel.com>
Signed-off-by: Xuan Ding <xuan.ding@intel.com>
Tested-by: Yvonne Yang <yvonnex.yang@intel.com>
Reviewed-by: Maxime Coquelin <maxime.coquelin@redhat.com>
The index in rte_vhost_set_last_inflight_io_split is from
the frontend driver, check if it's in the virtqueue range.
Fixes: bb0c2de960 ("vhost: add APIs to operate inflight ring")
Cc: stable@dpdk.org
Signed-off-by: Li Feng <fengli@smartx.com>
Reviewed-by: Maxime Coquelin <maxime.coquelin@redhat.com>
Added flow pattern items and header formats of L2TPv2 and PPP.
Signed-off-by: Wenjun Wu <wenjun1.wu@intel.com>
Signed-off-by: Jie Wang <jie1x.wang@intel.com>
Acked-by: Ori Kam <orika@nvidia.com>
Acked-by: Andrew Rybchenko <andrew.rybchenko@oktetlabs.ru>
Reviewed-by: Ferruh Yigit <ferruh.yigit@intel.com>
Full stop at the end of short comment just make line longer. It
should be either everywhere or nowhere to be consistent.
Signed-off-by: Andrew Rybchenko <andrew.rybchenko@oktetlabs.ru>
Reviewed-by: Ferruh Yigit <ferruh.yigit@intel.com>
Add empty lines to separate fields commented using different styles.
Signed-off-by: Andrew Rybchenko <andrew.rybchenko@oktetlabs.ru>
Reviewed-by: Ferruh Yigit <ferruh.yigit@intel.com>
Signed-off-by: Andrew Rybchenko <andrew.rybchenko@oktetlabs.ru>
Acked-by: Ori Kam <orika@nvidia.com>
Reviewed-by: Ferruh Yigit <ferruh.yigit@intel.com>
Signed-off-by: Andrew Rybchenko <andrew.rybchenko@oktetlabs.ru>
Acked-by: Ori Kam <orika@nvidia.com>
Reviewed-by: Ferruh Yigit <ferruh.yigit@intel.com>
Fix both in one changeset since they share line in a number of cases.
Signed-off-by: Andrew Rybchenko <andrew.rybchenko@oktetlabs.ru>
Reviewed-by: Ferruh Yigit <ferruh.yigit@intel.com>
