netdev - Re: [PATCH V5 vfio 04/10] vfio: Add an IOVA bitmap support

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 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:   Thu, 1 Sep 2022 20:16:59 -0300
From:   Jason Gunthorpe <jgg@...dia.com>
To:     Alex Williamson <alex.williamson@...hat.com>
Cc:     joao.m.martins@...cle.com, Yishai Hadas <yishaih@...dia.com>,
        saeedm@...dia.com, kvm@...r.kernel.org, netdev@...r.kernel.org,
        kuba@...nel.org, kevin.tian@...el.com, leonro@...dia.com,
        maorg@...dia.com, cohuck@...hat.com
Subject: Re: [PATCH V5 vfio 04/10] vfio: Add an IOVA bitmap support

On Thu, Sep 01, 2022 at 02:36:25PM -0600, Alex Williamson wrote:

> > Much of the bitmap helpers don't check that the offset is within the range
> > of the passed ulong array. So I followed the same thinking and the
> > caller is /provided/ with the range that the IOVA bitmap covers. The intention
> > was minimizing the number of operations given that this function sits on the
> > hot path. I can add this extra check.
> 
> Maybe Jason can quote a standard here, audit the callers vs sanitize
> the input.  It'd certainly be fair even if the test were a BUG_ON since
> it violates the defined calling conventions and we're not taking
> arbitrary input, but it could also pretty easily and quietly go into
> the weeds if we do nothing.  Thanks,

Nope, no consensus I know of

But generally people avoid sanity checks on hot paths

Linus will reject your merge request if you put a BUG_ON :)

Turn on a check if kasn is on or something if you think it is really
important?

Jason