mbox series

[v2,0/2] Use libata platform drivers to replace deprecated m68k IDE drivers

Message ID cover.1623131194.git.fthain@linux-m68k.org
Headers show
Series Use libata platform drivers to replace deprecated m68k IDE drivers | expand

Message

Finn Thain June 8, 2021, 5:46 a.m. UTC 
This patch series allows m68k platforms to switch from deprecated IDE
drivers to libata drivers.

Changed since v1:
 - Added reviewed-by and tested-by tags.
 - Improved commit logs.


Finn Thain (2):
  m68k/mac: Replace macide driver with generic platform drivers
  m68k/q40: Replace q40ide driver with pata_falcon and falconide

 arch/m68k/atari/config.c          |  12 +--
 arch/m68k/configs/mac_defconfig   |   1 -
 arch/m68k/configs/multi_defconfig |   2 -
 arch/m68k/configs/q40_defconfig   |   2 +-
 arch/m68k/mac/config.c            |  24 +++--
 arch/m68k/q40/config.c            |  37 +++++--
 drivers/ata/Kconfig               |   6 +-
 drivers/ata/pata_falcon.c         |  62 ++++++++---
 drivers/ide/Kconfig               |  32 +-----
 drivers/ide/Makefile              |   2 -
 drivers/ide/falconide.c           |  75 ++++++++-----
 drivers/ide/macide.c              | 161 ----------------------------
 drivers/ide/q40ide.c              | 168 ------------------------------
 13 files changed, 151 insertions(+), 433 deletions(-)
 delete mode 100644 drivers/ide/macide.c
 delete mode 100644 drivers/ide/q40ide.c

Comments

Christoph Hellwig June 8, 2021, 2:37 p.m. UTC | #1 
On Tue, Jun 08, 2021 at 03:46:34PM +1000, Finn Thain wrote:
> This patch series allows m68k platforms to switch from deprecated IDE
> drivers to libata drivers.

Thanks, this looks good to me:

Reviewed-by: Christoph Hellwig <hch@lst.de>

Jens, any chance you could pick this up quickly to prepare for the
legacy IDE removal series?
Jens Axboe June 8, 2021, 9:04 p.m. UTC | #2 
On 6/8/21 8:37 AM, Christoph Hellwig wrote:
> On Tue, Jun 08, 2021 at 03:46:34PM +1000, Finn Thain wrote:
>> This patch series allows m68k platforms to switch from deprecated IDE
>> drivers to libata drivers.
> 
> Thanks, this looks good to me:
> 
> Reviewed-by: Christoph Hellwig <hch@lst.de>
> 
> Jens, any chance you could pick this up quickly to prepare for the
> legacy IDE removal series?

I can, but had assumed it'd go through the IDE tree. But not that
important I guess. I've queued it up for now.
Michael Schmitz June 8, 2021, 9:50 p.m. UTC | #3 
Hi Jens,

please note that Finn's patch depends on one of mine currently under 
review. Without that one, Q40 support may break in certain cases.

Cheers,

     Michael Schmitz

On 9/06/21 9:04 am, Jens Axboe wrote:
> On 6/8/21 8:37 AM, Christoph Hellwig wrote:
>> On Tue, Jun 08, 2021 at 03:46:34PM +1000, Finn Thain wrote:
>>> This patch series allows m68k platforms to switch from deprecated IDE
>>> drivers to libata drivers.
>> Thanks, this looks good to me:
>>
>> Reviewed-by: Christoph Hellwig <hch@lst.de>
>>
>> Jens, any chance you could pick this up quickly to prepare for the
>> legacy IDE removal series?
> I can, but had assumed it'd go through the IDE tree. But not that
> important I guess. I've queued it up for now.
>
Finn Thain June 9, 2021, 3:53 a.m. UTC | #4 
On Wed, 9 Jun 2021, Michael Schmitz wrote:

> please note that Finn's patch depends on one of mine currently under review.
> Without that one, Q40 support may break in certain cases.
> 

AFAIK, that's only true for a configuration that enables both CONFIG_Q40 
and CONFIG_ATARI. In my tests with stock v5.12, such a configuration 
presently doesn't work correctly on Atari.

This patch series doesn't make this issue any worse.

Besides, I think it is a non-issue because the problem is so rare that 
no-one noticed it before I did, and the workaround for CONFIG_Q40 users is 
trivial anyway (disable CONFIG_ATARI).

Is there really some benefit in holding up this patch series?
Michael Schmitz June 9, 2021, 4:05 a.m. UTC | #5 
Hi Finn,

Am 09.06.2021 um 15:53 schrieb Finn Thain:
> On Wed, 9 Jun 2021, Michael Schmitz wrote:
>
>> please note that Finn's patch depends on one of mine currently under review.
>> Without that one, Q40 support may break in certain cases.
>>
>
> AFAIK, that's only true for a configuration that enables both CONFIG_Q40
> and CONFIG_ATARI. In my tests with stock v5.12, such a configuration
> presently doesn't work correctly on Atari.
>
> This patch series doesn't make this issue any worse.

You're right, it's not a regression (that combination was broken before) 
, and there's no patch dependency or conflict either.

> Besides, I think it is a non-issue because the problem is so rare that
> no-one noticed it before I did, and the workaround for CONFIG_Q40 users is
> trivial anyway (disable CONFIG_ATARI).
>
> Is there really some benefit in holding up this patch series?

On second thought - no point in holding this up.

Cheers,

	Michael
Jens Axboe June 9, 2021, 3:23 p.m. UTC | #6 
On 6/8/21 3:50 PM, Michael Schmitz wrote:
> Hi Jens,
> 
> please note that Finn's patch depends on one of mine currently under 
> review. Without that one, Q40 support may break in certain cases.

Can you point me at it?
Michael Schmitz June 9, 2021, 7:26 p.m. UTC | #7 
Hi Jens,

it's this one:

https://lore.kernel.org/linux-m68k/1623223322-4242-1-git-send-email-schmitzmic@gmail.com/

But see my reply to Finn's mail - my patch is entirely orthogonal to 
his, and simply fixes a long standing issue that has recently surfaced 
with Finn's efforts to unify Atari and Q40 IDE drivers.

Cheers,

     Michael

On 10/06/21 3:23 am, Jens Axboe wrote:
> On 6/8/21 3:50 PM, Michael Schmitz wrote:
>> Hi Jens,
>>
>> please note that Finn's patch depends on one of mine currently under
>> review. Without that one, Q40 support may break in certain cases.
> Can you point me at it?
>