Date: Sun, 11 Apr 2004 02:31:24 -0700 From: Luigi Rizzo <luigi@freebsd.org> To: Ruslan Ermilov <ru@freebsd.org> Cc: net@freebsd.org Subject: Re: Per-interface polling(4) status Message-ID: <20040411023124.A83988@xorpc.icir.org> In-Reply-To: <20040410222112.GA23046@ip.net.ua>; from ru@freebsd.org on Sun, Apr 11, 2004 at 01:21:12AM %2B0300 References: <20040410222112.GA23046@ip.net.ua>
next in thread | previous in thread | raw e-mail | index | archive | help
more or less i am fine with your proposed change. I believe you could add something to state the capability of the hw/driver (polling-capable, like checksum-offloading). For capabilities such as checksum offloading, 'xl' reports them as 'options' rather than 'flags', but other capabilities such as MULTICAST and SIMPLEX are instead reported as 'flags' (well you could think of MULTICAST as a capability that is always enabled so it makes sense to report it as a 'flag'). Actual settings (PROMISC, UP, POLLING, NOARP) are always reported as flags. I'd love if we could sort out this confusion and have capabilities clearly marked/reported in some ifconfig field (e.g. "options"). POLA would suggest to set IFF_POLLING by default on all drivers which support it. You still need the global sysctl to enable it. Finally, while you are at it (and before we branch 6.0), maybe you could store the 'IFF_CANTCHANGE' flags in a separate place from user-configured ones, so one wouldn't have to apply the mask all the times. cheers luigi On Sun, Apr 11, 2004 at 01:21:12AM +0300, Ruslan Ermilov wrote: > Dear networkers, > > Like I mumbled in another email, I've started working on a small > project that (when done) will allow for a per-interface polling(4) > status to be changed in run-time. Attached is a proof of the > concept patch, that currently has support for the dc(4) driver > only (basically because I have this card plugged into my notebook).
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20040411023124.A83988>