Soft Patch Panel
 help / color / mirror / Atom feed
From: Yasufumi Ogawa <ogawa.yasufumi@lab.ntt.co.jp>
To: spp@dpdk.org
Subject: [spp] Port ID incrementation is changed from DPDK18.02
Date: Thu, 1 Mar 2018 18:28:40 +0900	[thread overview]
Message-ID: <8b08be87-6b64-8a4a-53e0-29269011d81c@lab.ntt.co.jp> (raw)

Hi all,

Before DPDK 18.02 is fixed, behaviour of rte_eth_dev_get_port_by_name() 
in lib/librte_ether/rte_ethdev.c is changed by this commit.
 > 5b7ba3114 ethdev: add port ownership

This update is a desirable change to fix a problem for synchronization 
of port management. However, this function assigns unexpected port_id 
for SPP.

Previsouly, add command of secondary expected to be assigned port_id 
incrementally and independent among each of other secondary processes. 
For instance, adding ring 0 for sec 1 and 2 results like as

spp > sec 1;status
recv:8:{Client ID 1 Idling
clinet_id:1
port_id:0,on,PHY,outport:none
port_id:1,on,PHY,outport:none
port_id:2,on,RING(0),outport:none
}
spp > sec 2;status
recv:9:{Client ID 2 Idling
clinet_id:2
port_id:0,on,PHY,outport:none
port_id:1,on,PHY,outport:none
port_id:2,on,RING(0),outport:none
}

Both of secondary has port_id:2 for ring 0. However, it is different by 
using DPDK 18.02.

spp > sec 1;status
recv:8:{Client ID 1 Idling
clinet_id:1
port_id:0,on,PHY,outport:none
port_id:1,on,PHY,outport:none
port_id:4,on,RING(0),outport:none
}
spp > sec 2;status
recv:9:{Client ID 2 Idling
clinet_id:2
port_id:0,on,PHY,outport:none
port_id:1,on,PHY,outport:none
port_id:5,on,RING(0),outport:none
}

port_id are different, 4 and 5. It is not a serious problem, but would 
be annoying for users. I would like to fix it.

Luckly, we can patch ports without not port IDs and instead of resource 
IDs by recent update. So, I will change port assignment to not to use 
port ID by removing port_id from status.

Here is an updated message example.

spp > sec 1;status
recv:8:{Client ID 1 Idling
clinet_id:1
PHY(0),outport:none
PHY(1),outport:none
RING(0),outport:none
}

Patching ports with resource IDs is done by

spp > patch phy:0 ring:0

Thanks,
Yasufumi

-- 
Yasufumi Ogawa
NTT Network Service Systems Labs

             reply	other threads:[~2018-03-01  9:29 UTC|newest]

Thread overview: 12+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-03-01  9:28 Yasufumi Ogawa [this message]
2018-03-12  5:35 ` [spp] [PATCH 0/9] Update port management ogawa.yasufumi
2018-03-12  5:35   ` [spp] [PATCH 1/9] controller: change displaying status ogawa.yasufumi
2018-03-12  5:35   ` [spp] [PATCH 2/9] shared: update print_active_ports ogawa.yasufumi
2018-03-12  5:35   ` [spp] [PATCH 3/9] spp_nfv: change format displaying status ogawa.yasufumi
2018-03-12  5:35   ` [spp] [PATCH 4/9] spp_vm: " ogawa.yasufumi
2018-03-12  5:35   ` [spp] [PATCH 5/9] controller: add delimiter for topo command ogawa.yasufumi
2018-03-12  5:35   ` [spp] [PATCH 6/9] controller: update Shell for topo_subgraph ogawa.yasufumi
2018-03-12  5:35   ` [spp] [PATCH 7/9] controller: update generating graph ogawa.yasufumi
2018-03-12  5:35   ` [spp] [PATCH 8/9] controller: update topo command ogawa.yasufumi
2018-03-12  5:35   ` [spp] [PATCH 9/9] shared: fix bug for print port status ogawa.yasufumi
2018-03-27 23:51   ` [spp] [PATCH 0/9] Update port management Ferruh Yigit

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=8b08be87-6b64-8a4a-53e0-29269011d81c@lab.ntt.co.jp \
    --to=ogawa.yasufumi@lab.ntt.co.jp \
    --cc=spp@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).