netdev - RE: [net-next 08/15] i40e/i40evf: organize and re-number feature flags

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]

Message-ID: <063D6719AE5E284EB5DD2968C1650D6DD0055BCA@AcuExch.aculab.com>
Date:   Tue, 15 Aug 2017 12:42:34 +0000
From:   David Laight <David.Laight@...LAB.COM>
To:     "'Keller, Jacob E'" <jacob.e.keller@...el.com>,
        David Miller <davem@...emloft.net>,
        "Kirsher, Jeffrey T" <jeffrey.t.kirsher@...el.com>
CC:     "netdev@...r.kernel.org" <netdev@...r.kernel.org>,
        "nhorman@...hat.com" <nhorman@...hat.com>,
        "sassmann@...hat.com" <sassmann@...hat.com>,
        "jogreene@...hat.com" <jogreene@...hat.com>
Subject: RE: [net-next 08/15] i40e/i40evf: organize and re-number feature
 flags

From: Keller, Jacob E
> Sent: 14 August 2017 23:11
> > From: David Miller [mailto:davem@...emloft.net]
> > Sent: Saturday, August 12, 2017 1:04 PM
> > From: Jeff Kirsher <jeffrey.t.kirsher@...el.com>
> > Date: Sat, 12 Aug 2017 04:08:41 -0700
> >
> > > Also ensure that the flags variable is actually a u64 to guarantee
> > > 64bits of space on all architectures.
> >
> > Why?  You don't need 64-bits, you only need 27.
> >
> > This will be unnecessarily expensive on 32-bit platforms.
> >
> > Please don't do this.
> 
> I suppose a better method would be to switch to using a declare_bitmap instead, so that it
> automatically sizes based on the number of flags we have. The reason we chose 64bits is because we
> will add flags in the future, as we originally had more than 32 flags prior to this patch until we
> moved some into a separate field.
> 
> But now that I think about it, using DECLARE_BITMAP makes more sense, though it's a bit more invasive
> of the code.

And horribly stupid unless you really need dynamic indexes.

	David