DPDK usage discussions
 help / color / mirror / Atom feed
From: Isaac Boukris <iboukris@gmail.com>
To: users@dpdk.org, Stephen Hemminger <stephen@networkplumber.org>
Subject: dumpcap: timestamp is years ahead when in pcapng format
Date: Tue, 19 Sep 2023 17:40:02 +0300	[thread overview]
Message-ID: <CAC-fF8T630ZAsT4NEJTZM-zxrA809_h0pxs+ObOwsa1wfaa_NA@mail.gmail.com> (raw)

Hi Stephen,

Machine: vm on a laptop with i7 core
rte_get_timer_hz() = 2500000000
rte_get_tsc_hz()   = 2500000000

This issue doesn't occur with pcap format, only with pcapng.
When I capture on my system with dumpcap I get timestamp years ahead:

[root@Rocky8 ~]# dpdk-dumpcap -i 0 -f 'not tcp port 22' -w - | tcpdump -ttttr -
File: -
reading from file -, link-type EN10MB (Ethernet)
Packets captured: 0 dropped privs to tcpdump
2257-07-15 19:27:35.541912 IP _gateway.domain > Rocky8.36824: 42742
NXDomain 0/0/0 (44)
2257-07-15 19:27:35.541912 IP Rocky8.58223 > _gateway.domain: 22616+
PTR? 2.202.168.192.in-addr.arpa. (44)

Looking at the code, I tried the below simplification in
pcapng_tsc_to_ns() and it yielded good results, but I'm still not
clear as the logic is quite intimidating.

diff --git a/lib/pcapng/rte_pcapng.c b/lib/pcapng/rte_pcapng.c
index 80d08e1..96540b4 100644
--- a/lib/pcapng/rte_pcapng.c
+++ b/lib/pcapng/rte_pcapng.c
@@ -79,21 +79,8 @@ static uint64_t pcapng_tsc_to_ns(uint64_t cycles)
         * Currently all TSCs operate below 5GHz.
         */
        delta = cycles - pcapng_time.cycles;
-       if (unlikely(delta >= pcapng_time.tsc_hz)) {
-               if (likely(delta < pcapng_time.tsc_hz * 2)) {
-                       delta -= pcapng_time.tsc_hz;
-                       pcapng_time.cycles += pcapng_time.tsc_hz;
-                       pcapng_time.ns += NSEC_PER_SEC;
-               } else {
-                       secs = rte_reciprocal_divide_u64(delta,
&pcapng_time.tsc_hz_inverse);
-                       delta -= secs * pcapng_time.tsc_hz;
-                       pcapng_time.cycles += secs * pcapng_time.tsc_hz;
-                       pcapng_time.ns += secs * NSEC_PER_SEC;
-               }
-       }

-       return pcapng_time.ns + rte_reciprocal_divide_u64(delta * NSEC_PER_SEC,
-
&pcapng_time.tsc_hz_inverse);
+       return pcapng_time.ns + ((uint64_t)(delta /
pcapng_time.tsc_hz) * NSEC_PER_SEC);

Results:

[root@Rocky8 ~]# dpdk-dumpcap -i 0 -f 'not tcp port 22' -w - | tcpdump -ttttr -
File: -
reading from file -, link-type EN10MB (Ethernet)
Packets captured: 0 dropped privs to tcpdump
2023-09-19 05:33:18.191601 IP _gateway.domain > Rocky8.50821: 37740
NXDomain 0/0/0 (44)
2023-09-19 05:33:18.191601 IP Rocky8.57382 > _gateway.domain: 53422+
PTR? 2.202.168.192.in-addr.arpa. (44)

Any thoughts on this?

Thanks!

             reply	other threads:[~2023-09-19 14:40 UTC|newest]

Thread overview: 10+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-09-19 14:40 Isaac Boukris [this message]
2023-09-19 16:35 ` Isaac Boukris
2023-09-19 18:00   ` Stephen Hemminger
2023-09-20 17:59     ` Isaac Boukris
2023-09-20 18:53       ` Isaac Boukris
2023-09-20 19:55         ` Isaac Boukris
2023-09-21  2:09           ` Stephen Hemminger
2023-09-21  5:14             ` Isaac Boukris
2023-09-21 15:31               ` Stephen Hemminger
2023-09-21 19:00                 ` Isaac Boukris

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=CAC-fF8T630ZAsT4NEJTZM-zxrA809_h0pxs+ObOwsa1wfaa_NA@mail.gmail.com \
    --to=iboukris@gmail.com \
    --cc=stephen@networkplumber.org \
    --cc=users@dpdk.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).