linux-kernel - Re: Re: [PATCH] uapi: update install list after nvme.h rename

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, 19 Feb 2016 16:23:40 -0600
From:	Mario Limonciello <mario_limonciello@...l.com>
To:	Jiri Slaby <jslaby@...e.cz>, linux-kernel@...r.kernel.org,
	hch@....de, keith.busch@...el.com, axboe@...com,
	stable@...r.kernel.org
Cc:	Michal Marek <MMarek@...e.com>,
	Kernel Build Daemon <kbuild@...e.de>,
	"Dominguez, Jared" <Jared_Dominguez@...L.com>
Subject: Re: Re: [PATCH] uapi: update install list after nvme.h rename



On 01/14/2016 02:26 AM, Jiri Slaby wrote:
> On 01/14/2016, 02:17 AM, Mike Frysinger wrote:
> > On 12 Jan 2016 09:45, Jiri Slaby wrote:
> >> On 01/11/2016, 02:14 AM, Mike Frysinger wrote:
> >>> Commit 9d99a8dda154 ("nvme: move hardware structures out of the
> >>> uapi version of nvme.h") renamed nvme.h to nvme_ioctl.h, but
> >>> the uapi list still refers to nvme.h.  People trying to install
> >>> the headers hit a failure as the header no longer exists.
> >>
> >> Wouldn't reverting the rename be nicer to userspace? So that the
> >> filename is preserved and anybody who #included nvme.h still can
> >> do so?
>
> > i have no opinion on either route
>
Can someone with authority please decide on what to do about this? 
It's wrecking havoc in userspace right now on stuff that needs nvme.h.

We can't build efivar in Debian unstable or Ubuntu xenial anymore. 
efivar used to build against kernels older than 4.4.0 no problem, then
this happened and because nvme_ioctl.h isn't installed properly you
can't even fix it by #include nvme_ioctl.h.
Consequently it's impossible to fix unless you do something ugly like
define the macro you need from nvme_ioctl.h in your source.

Fedora has pulled this patch in already in advance of it hitting -stable
so that they could build efivar and similar apps with a fix like this:
https://github.com/rhinstaller/efivar/commit/3a0ae7189fe96355d64dc2daf91cf85282773c66

The whole point of -stable is to avoid distros needing to pull stuff in
that really should benefit everyone..

So please:
* Either pull this commit into stable so that the correct headers are
installed at kernel build.
Or
* Revert the old commit (9d99a8dda154) in 4.4 stable so things can build
again.

> > on a related note, shouldn't headers install be part of the
> > automatic kernel checks that are run against repos on kernel.org ?
> > who runs that thing ?
>
> I have been thinking about at least 'test -f' for all 'headers-y'
> during make all. What do you think, Michal?
>
> thanks,
> > >