linux-kernel - Re: btusb "firmware request while host is not available" at resume

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:   Mon, 11 Sep 2017 12:29:55 -0700
From:   Greg Kroah-Hartman <gregkh@...uxfoundation.org>
To:     "Luis R. Rodriguez" <mcgrof@...nel.org>
Cc:     Marcel Holtmann <marcel@...tmann.org>,
        Linus Torvalds <torvalds@...ux-foundation.org>,
        Gabriel C <nix.or.die@...il.com>, Tejun Heo <tj@...nel.org>,
        "Gustavo F. Padovan" <gustavo@...ovan.org>,
        Sukumar Ghorai <sukumar.ghorai@...el.com>,
        "Rafael J. Wysocki" <rjw@...ysocki.net>,
        Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
        "bluez mailin list (linux-bluetooth@...r.kernel.org)" 
        <linux-bluetooth@...r.kernel.org>
Subject: Re: btusb "firmware request while host is not available" at resume

On Mon, Sep 11, 2017 at 07:11:38PM +0200, Luis R. Rodriguez wrote:
> On Mon, Sep 11, 2017 at 06:46:47AM -0700, Greg Kroah-Hartman wrote:
> > To confirm, reverting this fixes the problem I was seeing in 4.13.  I've
> > queued it up for the next 4.13-stable release as well.
> 
> Commit 81f95076281f ("firmware: add sanity check on shutdown/suspend") may
> seem kludgy but the reason for it was to cleanup the horrible forced and
> required UMH lock even when the UMH lock was *not* even needed, which was later
> removed via commit 06a45a93e7d34aa ("firmware: move umh try locks into the umh
> code").

So what does this mean now that it is reverted?

> Removing the old UMH lock even when the UMH lock was *not* needed was the right
> thing to do but commit 81f95076281f (("firmware: add sanity check on
> shutdown/suspend") was put in place as a safe guard as the lock was also
> placing an implicit sanity check on the API. It ensures the API with the cache
> was used as designed, otherwise you do run the risk of *not getting the
> firmware you may need* -- Marcel seems to acknowledge this possibility.
> 
> It may be possible for us to already have in place safeguards so that upon
> resume we are ensuring the path to the firmware *is* available, so IMHO we
> should remove this *iff* we can provide this guarantee.  Otherwise the check is
> valid. You see, although the UMH lock was bogus, it did implicitly ask the
> question: is it safe for *any* helper to run and make assumptions on the
> filesystem then?
> 
> In lieu of this question being answered the warning is valid given the design
> of the firmware API and the having the cache available as a measure to resolve
> this race.

I don't understand what you are trying to say here at all.

To be specific, what, if anything, is a problem with the current state
of Linus's tree (and the next 4.13-stable release)?

If something needs to be fixed, can you make a patch showing that?  Or
do we also need to revert anything else as well to get back to a "better
working" state?

thanks,

greg k-h