Age | Commit message (Collapse) | Author | Files | Lines |
|
Change-Id: I0b437ac5fecc81c7762d9cad0f33e977fcf3aa27
Signed-off-by: Neale Ranns <nranns@cisco.com>
(cherry picked from commit 60c1c7c0397eaeb201f0fe300285bda4ae3ef707)
|
|
Change-Id: Ie57b81f8743f14182813558887d84d6667c81d43
Signed-off-by: Filip Tehlar <ftehlar@cisco.com>
(cherry picked from commit 0a62e5a3d817c3400be122c58d0311c298047580)
|
|
- Interactive commands like "ping" read extra input from the
input stream.
- In the case of "ping" it is simply a signal to cease the current
operation.
- "vppctl", in non-interactive mode, will issue a "quit" immediately
after the requested command to queue up closing of the session.
- This resulted in "ping" thinking a keypress was seen and returning
control to the CLI; the "quit" command however is consumed by the
keypress event handler and thus the session does not close.
- This patch reworks vppctl slightly to only issue "quit" after the
command has completed. In particular it uses the fact that VPP issues
NUL bytes as a surrogate prompt between output of commands to signal
acknowledgement that the command has completed; vppctl now flags
that the quit should be issued after the next such acknowledgement.
- Since input it still accepted, the user can still terminate the
"ping" early, if desired.
Change-Id: I7e3dbe767f32f8e364ccb5f81799759b311585df
Signed-off-by: Chris Luke <chrisy@flirble.org>
|
|
Dynamically calculate the required buffer size to pack into based on
message definition. Also add input parameter length checking.
Change-Id: I7633bec596e4833bb328fbf63a65b866c7985de5
Signed-off-by: Ole Troan <ot@cisco.com>
(cherry picked from commit 895b6e8b4408108a9b5cea99dcb378c3524b18b2)
|
|
VPP-1036
Change-Id: I96dae6a4a40c0142d03bdf750ef23c9ce5b6edfa
Signed-off-by: Sergio Gonzalez Monroy <sergio.gonzalez.monroy@intel.com>
|
|
VPP-1034
Change-Id: Ib13cf0395bc2e42d24de51347e0aeed8a137f2f2
Signed-off-by: Sergio Gonzalez Monroy <sergio.gonzalez.monroy@intel.com>
|
|
We need to push/pop the rx pthread's heap without affecting other
thread(s).
Search clib_per_cpu_mheaps, locate an unused slot. Duplicate the main
thread heap pointer in that slot, and set __os_thread_index
appropriately.
Don't bail out of vpp_api_test with results pending, e.g. at the end
of a vpp_api_test script. Even though vpp will eventuallly
garbage-collect them, We don't want to leave allocated reply messages
lurking in the api message allocation rings...
This patch is a manual cherry-pick from gerrit 8862
Change-Id: If3c84abe61496905432bfa36767068bad4bd243b
Signed-off-by: Dave Barach <dave@barachs.net>
|
|
Change-Id: I0c1671f3eaf2dad084e3ac9fb124c9ed78273f50
Signed-off-by: Dave Barach <dave@barachs.net>
|
|
Change-Id: Ief8c3d3bec116e9f884981fb52af528f98b5f6ff
Signed-off-by: Filip Tehlar <ftehlar@cisco.com>
|
|
Change-Id: I20ce799c9dd57332c06003b466ee7c36169bce98
Signed-off-by: Dave Barach <dave@barachs.net>
(cherry picked from commit fa77e8fb1cc8a87a4a31792698e7200c4195583e)
|
|
Change-Id: I198f58a84c4692408f9205052af24ee22df7aeaa
Signed-off-by: Filip Tehlar <ftehlar@cisco.com>
|
|
Change-Id: Ia31b978c6c1619c3e0075a84fcbbb6ccbf1c0076
Signed-off-by: Florin Coras <fcoras@cisco.com>
(cherry picked from commit d172a62365523427aebd34444736632c9a68a463)
|
|
It's way too easy to imagine leaving a mutex or a spin-lock held in
the /vpe-api shared-memory segment, or elsewhere. Set a volatile
variable and check it in a safe place...
Change-Id: I9d91c38cffeb921143c272162d055c9c24a6c312
Signed-off-by: Dave Barach <dave@barachs.net>
(cherry picked from commit 903651caf320dfdaabd20a0e6f3cd0ffc843e02f)
|
|
VPP-1025
Change-Id: I87bb8f38974a7be274c1b1d205f5513e7d068e48
Signed-off-by: Georgina <georgina.sheehan@intel.com>
|
|
Change-Id: I2446c646de7f227f9438dd7ef93a455ba5af0102
Signed-off-by: Matus Fabian <matfabia@cisco.com>
|
|
Change-Id: I92b0dbd2d795e3ba9050797ede2b08638e55b6f5
Signed-off-by: Florin Coras <fcoras@cisco.com>
|
|
Change-Id: I19fdf13a4848306ee3841d822b832cba96c5bce5
Signed-off-by: Neale Ranns <nranns@cisco.com>
(cherry picked from commit 8edad036f1f8b920a6643a026f50a5d3aff8cbf1)
|
|
Change-Id: I011dda118f37cb31a37dda270027612d0af57ca0
Signed-off-by: Florin Coras <fcoras@cisco.com>
|
|
Change-Id: If5e0fb8a1b60c7181b5b3d6dd923f52a24dda9b8
Signed-off-by: Florin Coras <fcoras@cisco.com>
|
|
With heavy traffic, tx code path may crash due to memory corruption
Thread 5 "vpp_wk_2" received signal SIGSEGV, Segmentation fault.
[Switching to Thread 0x7fff3995c700 (LWP 2505)]
0x00007ffff73675e8 in vhost_user_if_input (vm=0x7fffb5f5bf9c,
vum=0x7ffff7882a40 <vhost_user_main>, vui=0x7fffb65570c4, qid=0,
node=0x7fffb6577dac, mode=VNET_HW_INTERFACE_RX_MODE_POLLING)
at /home/sluong/vpp-master/vpp/build-data/../src/vnet/devices/virtio/vhost-user.c:1610
1610 bi_current = (vum->cpus[thread_index].rx_buffers)
[vum->cpus[thread_index].rx_buffers_len];
(gdb) p vum->cpus[thread_index].rx_buffers_len
$2 = 793212607
(gdb)
Apparently, some code accidentally wrote the bad value in rx_buffers_len.
rx_buffers_len should never be greater than 1024 since that is how many buffers
we request each time.
After debugging many hours, I discovered that the memory corruption happens
in the tx code path right here on line 2176.
{
vhost_copy_t *cpy = &vum->cpus[thread_index].copy[copy_len];
copy_len++;
cpy->len = bytes_left;
cpy->len = (cpy->len > buffer_len) ? buffer_len : cpy->len;
cpy->dst = buffer_map_addr;
cpy->src = (uword) vlib_buffer_get_current (current_b0) +
current_b0->current_length - bytes_left;
(gdb) p cpy
$3 = (vhost_copy_t *) 0x7fffb554077c
(gdb) p copy_len
$4 = 1025
(gdb) p &vum->cpus[3].rx_buffers_len
$8 = (u32 *) 0x7fffb5540784
copy_len is picking up the index entry 1024 before it was incremented. copy array has only
1024 members (0 - 1023 are valid).
The assignment here in cpy surely causes memory corruption. It is only discovered later
when the memory location that it corrupted is used.
The condition for the crash is to transmit jumbo frames under heavy volume. Since ring
size is 1024, with one packet taking up one index for frame size (less 2048), it does
not cause overflow. With jumbo frames, it requires multiple indices for one packet,
it can cause the overflow under heavy traffic.
The fix is to do copy out when we have 1000 entries in the array to avoid
overflow.
Change-Id: Iefbc739b8e80470f1cf13123113f8331ffcd0eb2
Signed-off-by: Steven <sluong@cisco.com>
(cherry picked from commit aa5df48cb233b377b5910694e2440a16e5973864)
|
|
Change-Id: Ie3245b96c511cc30915e70e8c881f445291a38c2
Signed-off-by: Matus Fabian <matfabia@cisco.com>
|
|
https://gerrit.fd.io/r/#/c/8551/ decoupled the global variable,
namely tm->iovecs from TX and RX. However, to support multi-threads,
we have to eliminate the use of this global variable with per thread
variable. I notice that rx_buffers must also be per thread variable.
So, we introduce per thread struct to contain rx_buffers and iovecs.
Each thread will find the per thread struct with thread_index.
Change-Id: I61abf2fdace8d722525a382ac72f0d04a173b9ce
Signed-off-by: Steven <sluong@cisco.com>
(cherry picked from commit 4cd257667406d0500a81323ef91f5c7c8c902b25)
|
|
A bug was reported where a jumbo packet would stay in vhost
queue forever or until a large enough number of other packets
arrived in the queue too.
This is due to a bug in vhost input node buffer allocation.
The fix is to make sure that vhost always allocates at least
enough buffers for one single big packet. '40' is used to
account for 65kB frames.
Change-Id: I1d293028854165083e30cd798fab9d4140230b78
Signed-off-by: Pierre Pfister <ppfister@cisco.com>
|
|
Change-Id: I99c2c1d0d5b96f33efdb58dd3a2897a752e65349
Signed-off-by: Klement Sekera <ksekera@cisco.com>
|
|
Crash was seen on recent image with this BT on top of the stack
(gdb) bt full
(mif=0x7fffb6226568) at
/vpp/build-data/../src/plugins/memif/memif.c:297
ring = 0x0 <<<<<<<<<<
i = 0
j = 0
buffer_offset = 65792
r = 0x7fffb5e59f80
alloc = {flags = 1, name = 0x7fffb449f965 "memif region",
size = 4260096, numa_node = 0, addr = 0x7fff41dac000,
fd = 11,
log2_page_size = 12, n_pages = 1041}
err = 0x0
__FUNCTION__ = "memif_init_regions_and_queues"
The crash happened at this line.
ring = memif_get_ring (mif, MEMIF_RING_S2M, i);
ring=>head = ring->tail = 0; <=====
Please note that the crash is caused by dereferencing NULL rinng.
Put breakpoint into the function. I notice that
mif->regions[0].shm is not initialized.
(gdb) p mif->regions[0].shm
$8 = (void *) 0x0
It looks like we forgot to set shm after clib_mem_vm_ext_alloc().
Add the missing cide and the crash is fixed.
Change-Id: Ib722a6c241c77acfa8e33962106b57faa50e1ea7
Signed-off-by: Steven <sluong@cisco.com>
(cherry picked from commit 9fefa9a697daf0e949ea7a2700ecaf2ba4d1d2cb)
|
|
Change-Id: I492bea060ba5c219ea75e19ebfdad79b1074e04b
Signed-off-by: Florin Coras <fcoras@cisco.com>
|
|
Identify and complain when the same IP prefix is assigned
to two different SW interfaces:
vpp# set int ip address TenGigabitEthernet6/0/0 1.2.3.4/32
vpp# set int ip address TenGigabitEthernet6/0/1 1.2.3.4/32
set interface ip address: Prefix 1.2.3.4/32 already found on
interface TenGigabitEthernet6/0/0
Change-Id: I1aee1b6a7ddd00d3109a53d8e1b6ce97bf45e372
Signed-off-by: Jon Loeliger <jdl@netgate.com>
(cherry picked from commit 35ffa3e8f6b032f6e324234d495f769049d8feea)
|
|
Change-Id: I7b175d57b85e626aab00221b6dac0498aebcbeae
Signed-off-by: Neale Ranns <nranns@cisco.com>
(cherry picked from commit d867a7cf6baffcebbf1b6e408272ec22dc55dd68)
|
|
Change-Id: I58772a83e22885a9ea8a7a981d2bcb4b31a050d2
Signed-off-by: Neale Ranns <nranns@cisco.com>
(cherry picked from commit 7b7ba572ab486d57b59c12af521175a6bcd7a52b)
|
|
Change L2 learning path so it update stale timestamp in MAC entry
only if aging is enabled on the BD for the MAC entry.
Change-Id: I7babe986ceef3c030d8ef9185076c42b405f7b0f
Signed-off-by: John Lo <loj@cisco.com>
(cherry picked from commit 5a6508d7269266b4a3ecacdd197ea3514a0c0e28)
|
|
Change-Id: I3e6276e6829dfee5a7aeae1b4ab4c3d2f2e932a4
Signed-off-by: Florin Coras <fcoras@cisco.com>
|
|
fixes an issue where events were not sent if BD doesn't enable mac aging
Change-Id: Iddc53cb5c45e560633e6c5cff2731dccfc70ad5b
Signed-off-by: Eyal Bari <ebari@cisco.com>
(cherry picked from commit 24db0ec78fb651c4c585ebf30e07108240574045)
|
|
When making a call to vlib_packet_template_get_packet(), it
is possible to get back a NULL if the system runs out of buffer.
This can happen when there is buffer leaks. But don't crash
just because we run out of buffers, just punt.
Change-Id: Ie90ea41f3dda6e583d48959cbd18ff124158d7f8
Signed-off-by: Steven <sluong@cisco.com>
(cherry picked from commit 0ff5c563d5048991dbd02a3892dccde8305a7e30)
|
|
It was observed that under heavy traffic, VPP accidentally sent traffic
with the wrong source and destination to the tun/tap interface. Traffic
appears to be sent to the wrong direction. This problem is only
seen when worker thread is configured.
When worker thread is used, TX and RX may reside in different
core. Yet both TX and RX threads are sharing the same global variable,
namely iovecs without any mutex or memory barrier protection.
This creates a race condition when heavy traffic is blasted to VPP,
like 1000 pps.
We could create a mutex or memory barrier to ensure atomic memory access.
But why bother? It is a lot cheaper to just decouple the iovecs such
that TX and RX have their own iovecs.
Change-Id: I86a5a19bd8de54d54f32e1f0845bae6a81bbf686
Signed-off-by: Steven <sluong@cisco.com>
(cherry picked from commit 4ff586d1c6fc5c40e1548cd6f221a8a7f3ad033b)
|
|
- We now have several developer-focused docs, so create an index page
for them.
- Rework several docs to fit into the index structure.
- Experiment with code highlighting; tweak the CSS slightly to make
it slightly nicer to look at.
Change-Id: I4185a18f84fa0764745ca7a3148276064a3155c6
Signed-off-by: Chris Luke <chrisy@flirble.org>
(cherry picked from commit 64ebb5ff1338140d94c7f9ee72138fe84d89de2e)
|
|
Change-Id: I5ca5763f0dc0a73cc6f014b855426b7ac180f356
Signed-off-by: Matej Perina <mperina@cisco.com>
|
|
Change-Id: Ib675164c475edcdbe3013df7b847adf5e050c53f
Signed-off-by: Filip Tehlar <ftehlar@cisco.com>
|
|
On host interface if a VLAN tagged packet is received, linux kernel removes
the VLAN header from packet byte stream and adds metadata in tpacket2_hdr.
This patch explicitely checks for the presense of VLAN metadata and adds it
in VPP packet.
Change-Id: I0ba35c1e98dbc008ce18d032f22f2717d610c1aa
Signed-off-by: Akshaya N <akshaya@rtbrick.com>
|
|
UIO binding is not required for Mellanox NIC and calling vlib_pci_bind_to_uio()
should be skipped.
Change-Id: I10ea457bc3c8d4be8117dec51d5bd940ee416a44
Signed-off-by: Steve Shin <jonshin@cisco.com>
|
|
174267: Revisit this string termination issue
174816: Add check for NULL when trace is enabled
177211: Add notation that mutex is not required here
177117: Added check for log2_page_size == 0 and returns an error if so
163697,163698: Added missing sw_if_index validation
Change-Id: I5a76fcf6505c785bfb3269e353360031c6a0fd0f
Signed-off-by: Chris Luke <chrisy@flirble.org>
|
|
The replacement of [] with pool_elt_at_index and subsequent fixing it
was incorrect - it was equivalent to &[], since it returns a pointer to
the element. I've added VPP-993 previously to create a testcase,
so this commit partially fulfills that one as well.
Change-Id: I5b15e3ce48316f0429232aacf885e8f7c63d9522
Signed-off-by: Andrew Yourtchenko <ayourtch@gmail.com>
|
|
Change-Id: I2cf4c4850b9c3c093a7dce0cec89b9f710f69393
Signed-off-by: Florin Coras <fcoras@cisco.com>
|
|
Update ping code to use the new function
Change-Id: Ieb753b23f8402cbe5667c22747896784c8ece937
Signed-off-by: Florin Coras <fcoras@cisco.com>
Signed-off-by: Dave Barach <dave@barachs.net>
|
|
Change-Id: I3873d3e411bf93cac82e73a0b8e3b22563aaf217
Signed-off-by: Matus Fabian <matfabia@cisco.com>
|
|
Change-Id: I7d8f807fb502d61688aa1dee25fa4edcbeb32f41
Signed-off-by: Aequitas <wang.junqi@zte.com.cn>
|
|
Change-Id: I365c31607332a944ef498369881332b515894ed7
Signed-off-by: Florin Coras <fcoras@cisco.com>
|
|
vector elements
bb7f0f644 aimed to fix the coverity issue has incorrectly replaced the previous [] access
with pool_elt_at_index(), for an element of a vector, with predictably interesting result.
VPP-991 has uncovered the issue.
Change-Id: Ifd3fb70332d3fdd1c4ff8570372f394913f7b6c8
Signed-off-by: Andrew Yourtchenko <ayourtch@gmail.com>
|
|
Change-Id: Idad65cbb3765500a66f1097126076a2c5fdb4f1b
Signed-off-by: Jerome Tollet <jtollet@cisco.com>
|
|
For bonded interface in Active/Backup mode (mode 1), we need to
send a GARP/NA packet, if IP address is present, on slave link
state change to up or down to help with route convergence. The
callback from DPDK happens in a separate thread so we need to make
sure RPC call is used to signal the send_garp_na process in the
main thread. Also need to fix DPDK polling so the slave links are
not polled.
Change-Id: If5fd8ea2d28c54dd28726ac403ad366386ce9651
Signed-off-by: John Lo <loj@cisco.com>
|
|
Change-Id: Id1a5da12b13d87bacfa81094f471b95db40c39be
Signed-off-by: Dave Barach <dave@barachs.net>
|