linux-kernel - Re: [RFC PATCH 00/19] ide: remove deprecated host drivers (part 1)

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:   Thu, 08 Dec 2016 18:14:33 +0100
From:   Bartlomiej Zolnierkiewicz <b.zolnierkie@...sung.com>
To:     David Miller <davem@...emloft.net>
Cc:     linux-ide@...r.kernel.org, linux-kernel@...r.kernel.org,
        tj@...nel.org, gnomes@...rguk.ukuu.org.uk,
        sergei.shtylyov@...entembedded.com
Subject: Re: [RFC PATCH 00/19] ide: remove deprecated host drivers (part 1)

On Thursday, December 08, 2016 11:39:57 AM David Miller wrote:
> From: Bartlomiej Zolnierkiewicz <b.zolnierkie@...sung.com>
> Date: Thu, 08 Dec 2016 17:15:16 +0100
> 
> > On Thursday, December 08, 2016 10:23:08 AM David Miller wrote:
> >> From: Bartlomiej Zolnierkiewicz <b.zolnierkie@...sung.com>
> >> Date: Thu, 08 Dec 2016 14:42:12 +0100
> >> 
> >> > Ping.. since there was no negative (or any other) feedback I think that
> >> > this can be put into -next for some wider exposure..
> >> 
> >> I'm not going to apply this and break things on people, sorry.
> > 
> > People building their own kernels with their own kernel config files
> > and still using IDE drivers (deprecated in 2009 BTW) will have to
> > update their setups to libata but otherwise nothing is supposed to
> > break.  So could you please explain in more detail what do you mean
> > by "break things on people"?
> 
> There is no proof that the PATA drivers work %100 reliably as well as
> the IDE driver they replace for every possible chip and architecture.

This is why only some IDE host drivers were nominated for removal.

Please see cover-letter and patch descriptions for details.

> Therefore the only safe thing is to keep the IDE drivers around
> forever.

Uh...

> They are not a maintainence burdon, I rarely get more than 1 patch
> each merge window and most of the time those are cleanups or for the
> handling of a kernel wide API change rather than bug fixes.

In the long-term perspective having two drivers for the same hardware
is not good for the whole kernel as:

- testing efforts are divided (while number of PATA systems goes down)

- bugs/missing features in the new subsystem are not getting reported
  and fixed (since it is easier to just switch back to the old stack)

- having duplicated support for the same hardware confuses users

I was hoping for incremental removal of IDE host drivers and keeping
only these that are really needed (while at the same time migrating
them slowly to libata).

I asked you about this in private mail in August 2015, you told me to
bring this on the list.  I did it (with these patches) in February
2016.  After two pings and months of waiting for a reply all I get is
is a quick NAK?

Best regards,
--
Bartlomiej Zolnierkiewicz
Samsung R&D Institute Poland
Samsung Electronics