Thanks Dave!!!.
I could see precision was indeed the problem. I think it is because the
slots are missed due to below check in tw_timer_expire_timers_internal(...)
when now time is behind the calculated next_run_time.
* /* Shouldn't happen */*
* if (PREDICT_FALSE (now < tw->next_run_time))*
*
Hi,
I want to share the parameter block across different plugins.
startup.conf has --
foo {
timeout 10
..
..
}
Plugin A - VLIB_CONFIG_FUNCTION(foo_fn_a, "foo")
Plugin B - VLIB_CONFIG_FUNCTION(foo_fn_b, "foo")
Is this possible ? I am not able to get the parameter block -foo shared. Any
Yes, It can definitely be done via mmap the memory and marking the thread
process shared.
*fd=shm_open(...);*
*mmap(NULL, shmSize, PROT_READ|PROT_WRITE, MAP_SHARED, fd, 0);*
*...*
*pthread_mutexattr_setpshared(, PTHREAD_PROCESS_SHARED);*
*pthread_mutex_init(>lock, ); *
*Regards,*
*Shaligram
gt;
> *From:* vpp-dev@lists.fd.io *On Behalf Of *
> shaligram.prakash
> *Sent:* Friday, May 3, 2019 2:48 PM
> *To:* vpp-dev@lists.fd.io
> *Subject:* [vpp-dev] Bihash memory requirement #vpp_stability
>
>
>
> Hello,
>
> Past days, I am trying to understand the VPP Bihash memory requirem
Hi,
I am facing a crash in the below code. It's 12 worker/1 core setup. we are
using bit old VPP - 20.09.
function under suspect is ---
buffer_send_to_node()
* vlib_frame_t *frame;
* u32 *to_next;
*
* frame = vlib_get_frame_to_node (vm, node_index);
* frame->n_vectors = 1;
* to_next =
On Tue, Jan 10, 2023 at 06:04 AM, Zhang, Fan wrote:
>
> vnet_hw_if_tx_frame_t
[Typo - 12 worker/1 main core setup]
This function is used to send packet at buffer_index to destination
node_index[say ip4-lookup]. It works as expected except this case from
production environment .
what happens
