linux-kernel - RE: [PATCH v2 03/10] RapidIO: Use stored ingress port number instead of register read

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] [day] [month] [year] [list]

Date:	Fri, 1 Oct 2010 13:46:06 -0700
From:	"Bounine, Alexandre" <Alexandre.Bounine@....com>
To:	"Micha Nelissen" <micha@...i.hopto.org>
Cc:	"Andrew Morton" <akpm@...ux-foundation.org>,
	<linux-kernel@...r.kernel.org>, <linuxppc-dev@...ts.ozlabs.org>,
	"Thomas Moll" <thomas.moll@...go.com>,
	"Matt Porter" <mporter@...nel.crashing.org>,
	"Li Yang" <leoli@...escale.com>,
	"Kumar Gala" <galak@...nel.crashing.org>
Subject: RE: [PATCH v2 03/10] RapidIO: Use stored ingress port number instead of register read

Hi Micha,

Sorry for delayed reply.

Micha Nelissen <micha@...i.hopto.org> wrote:
> 
> Bounine, Alexandre wrote:
> > struct rio_dev {
> >     struct list_head global_list;
> >     struct list_head net_list;
> >     .....
> >     ..... rest of rio_dev
> >     .....
> >     struct rio_switch switch[0];
> > }
> 
> It makes sense to let rio_dev structures point to the switch they are
> attached to. That can be useful in various situations, but is not
> possible with this setup.
> 
> If a rio_dev is a switch then rdev->rswitch->rdev == rdev holds.
>

But the switch is a RIO device itself and all other parts of rio_dev
structure
are applicable to it as well. If there is situation when a device 
needs to hold a pointer to the attached switch that should be a pointer
to the switch rio_dev and not to its switch-specific extension. 

> > This will remove extra memory allocation, remove overlapping
structure
> > members and clean code sections like one shown below:
> >
> > 	u8 hopcount = 0xff;
> > 	u16 destid = rdev->destid;
> >
> > 	if (rdev->rswitch) {
> > 		destid = rdev->rswitch->destid;
> > 		hopcount = rdev->rswitch->hopcount;
> > 	}
> 
> Note that it is possible for rdev->destid to differ from
> rdev->rswitch->destid even if rswitch->rdev == rdev (for non-hosts
i.e.
> agents). rswitch->destid is the destid by which we discovered the
switch
> (and can reach it) but rdev->destid is the actual id given to the
switch.
> 

My goal is to have one destid storage for device - endpoint or switch.
And destid should be used only for one purpose: to reach corresponding
device.
In your statement above you suggest using rdev->destid instead of
rswitch->switchid.
RIO switches do not have any specific RIO ID that can be assigned to the
switch.
In this case the rswitch->switchid should work well for logical
identification
of the switch.

I think if (for switch) we load rdev->destid with some function
different
from its routing role this may bring unnecessary confusion.

I also will move rswitch->hopcount to rdev->hopcount. For endpoint it
will be set
to 0xff during rio_dev initialization.

Alex.

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/