netdev - RE: [PATCH net-next-2.6 02/13] net-caif: add CAIF header files

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

Date:	Fri, 22 Jan 2010 10:43:28 +0100
From:	Sjur Brændeland <sjur.brandeland@...ricsson.com>
To:	"Marcel Holtmann" <marcel@...tmann.org>
Cc:	<netdev@...r.kernel.org>, <davem@...emloft.net>,
	<stefano.babic@...ic.homelinux.org>, <randy.dunlap@...cle.com>
Subject: RE: [PATCH net-next-2.6 02/13] net-caif: add CAIF header files

Marcel Holtmann wrote:
> Hi Sjur,
> 
>>>> CAIFSO_CHANNEL is used for specifying the physical interface to use
>>>> for the CAIF Channel. You can select the type of interface to use
>>>> by setting link_selector: CAIF_PHYPREF_LOW_LAT will typically be
>>>> used for AT (or other control traffic), and CAIF_PHYPREF_HIGH_BW
>>>> for IP traffic.
>>>> When the CAIF interfaces registers itself it will inform about
>>>> their type, (low-latency or high-bandwidth). This approach assumes
>>>> that you have only one modem, but multiple links to it (e.g. USB
>>>> and UART). 
>>>> 
>>>> But you can also specify interface by name using link_name. In this
>>>> case you specify the name of the interface to use. I think this
>>>> would support your use case with multiple modems attached.
>>> 
>>> sounds good, but why using a socket option and not allowing to just
>>> use bind(). Maybe it is just my personal preference, because I am
>>> used to do it like this for TCP and Bluetooth.
>> 
>> We actually considered this when designing the interface, but I don't
>> feel bind() fits in this case. bind() would normally bind an address
>> to the client socket. An interface did not seem like a client socket
>> address to us, so we decided to use sockopt instead.
>> Conceptually I don't think CAIF has a client address at all, it only
>> connects to a server side address.
> 
> seems fair enough to me. My personal preference would just be bind,
> but I can follow your argumentation. I just wanted to make sure that
> we have that option before setting CAIF socket address in stone.  
> 
> Any reason why not just supporting SO_PRIORITY and SO_BINDTODEVICE on
> CAIF sockets then? 

Sounds like a good idea, I'll definitely look into that, thanks.

BR/Sjur
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html