netdev - Re: [PATCH] ip route: timeout for routes has to be set in seconds

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [day] [month] [year] [list]

Message-ID: <20160629092522.494bd502@xeon-e3>
Date:	Wed, 29 Jun 2016 09:25:22 -0700
From:	Stephen Hemminger <shemming@...cade.com>
To:	Xin Long <lucien.xin@...il.com>
CC:	Andrey Vagin <avagin@...nvz.org>,
	Hannes Frederic Sowa <hannes@...essinduktion.org>,
	network dev <netdev@...r.kernel.org>,
	"Andrew Vagin" <avagin@...tuozzo.com>,
	Hangbin Liu <liuhangbin@...il.com>
Subject: Re: [PATCH] ip route: timeout for routes has to be set in seconds


> There's another issue.
> now  the expires transition with HZ happens in kernel when we add
> route with expire, but when we dump,  the expires transition is in the
> user-space:
>                         if (ci->rta_expires != 0)
>                                 fprintf(fp, " expires %dsec",
> ci->rta_expires/hz);
> 
> I'm not sure if  the *hz* value between kernel and user-space are
> always same. I'd like to do this transition both in kernel. just like
> ipaddr's
> valid_lft, preferred_lft.

All API's from applications are supposed to use USER_HZ which is 100hz
on most platforms. Actual kernel internal hz is configurable.

There were some cases legacy of early days where iproute2 had to deal
with kernel hz, mostly in the QoS stuff.