Date: Mon, 9 Dec 2013 11:37:08 -0800 From: Adrian Chadd <adrian@freebsd.org> To: Michael Tuexen <Michael.Tuexen@lurchi.franken.de>, John Baldwin <jhb@freebsd.org> Cc: Yong-Hyeon Pyun <pyunyh@gmail.com>, Jack F Vogel <jfv@freebsd.org>, John-Mark Gurney <jmg@funkthat.com>, "freebsd-net@freebsd.org list" <freebsd-net@freebsd.org> Subject: Re: A small fix for if_em.c, if_igb.c, if_ixgbe.c Message-ID: <CAJ-VmomNf2m3Qinxz_vN5gXPcMnJmysamR4LPRcK=vPKJvG_DA@mail.gmail.com> In-Reply-To: <CAJ-Vmok9oy5BTU8f2KrRU5NBaxOZWF11qO2Z7=KTrkFhfBc0cQ@mail.gmail.com> References: <521B9C2A-EECC-4412-9F68-2235320EF324@lurchi.franken.de> <20131202022338.GA3500@michelle.cdnetworks.com> <B9593E83-E687-49E9-ABDC-B2DD615180E9@lurchi.franken.de> <20131203021658.GC2981@michelle.cdnetworks.com> <CAJ-Vmo=kfoPMYjZ0WAtqmoJMz1utXH50SW9N92RA83EMUzY7WA@mail.gmail.com> <B89B1E2D-BAF0-4815-B3AB-EB226F4F76DE@lurchi.franken.de> <CAJ-Vmo=4Zwv5V6ZYDuDLtt%2BowgbvmqyvrnrfnU%2BHeXQ3vAn-KA@mail.gmail.com> <20131205223711.GB55638@funkthat.com> <3576B69E-E943-46E0-83E5-0B2194A44ED0@lurchi.franken.de> <20131206202012.GG55638@funkthat.com> <609C63CD-9332-4EAE-AACE-5B911416DF80@lurchi.franken.de> <CAJ-Vmomnu4VLE0Q8A%2BQS6%2B7LA7ry_kD9j05=TvNZeocRjsuE7A@mail.gmail.com> <9E163DC1-D647-4E19-BE23-44E5DFE2F284@lurchi.franken.de> <CAJ-Vmok9oy5BTU8f2KrRU5NBaxOZWF11qO2Z7=KTrkFhfBc0cQ@mail.gmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
Jack / John - thoughts? -a On 8 December 2013 19:44, Adrian Chadd <adrian@freebsd.org> wrote: > .. I'm saying that we should have if_transmit( )return an error only > on the current packet, and 0 if it was queued. > > We don't have a mechanism to say that a queued packet actually made it > onto the wire. Whether we should is a different discussion. > > Ie, after thinking about this some more, I'd like to: > > * change if_transmit in these drivers to do exactly what you suggest - > it should return an error only if the given packet couldn't be queued. > It shouldn't matter whether it is later transmitted or not - we don't > have a feedback mechanism for that. > * figure out a very specific definition of what xxx_mq_start_locked() > should return - my gut feeling is an error if it couldn't queue a > frame, and 0 if it dispatched a frame to the hardware - and then make > the code match this definition. > > > > -adrian
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?CAJ-VmomNf2m3Qinxz_vN5gXPcMnJmysamR4LPRcK=vPKJvG_DA>