Re: [GIT PULL] Core block IO bits for 2.6.39

[Jens Axboe <jaxboe@fusionio.com>]

On 2011-03-26 09:34, Geert Uytterhoeven wrote:
> On Sat, Mar 26, 2011 at 09:25, Jens Axboe <jaxboe@fusionio.com> wrote:
>> On 2011-03-26 08:21, Geert Uytterhoeven wrote:
>>> On Sat, Mar 26, 2011 at 07:29, Jens Axboe <jaxboe@fusionio.com> wrote:
>>>> On 2011-03-25 22:35, Geert Uytterhoeven wrote:
>>>>> On Thu, Mar 24, 2011 at 14:43, Jens Axboe <jaxboe@fusionio.com> wrote:
>>>>>> Jens Axboe (20):
>>>>>>      block: remove per-queue plugging
>>>>>
>>>>> This one (commit 7eaceaccab5f40bbfda044629a6298616aeaed50) breaks IDE
>>>>> on Atari/m68k under ARAnyM. It hangs on:
>>>>>
>>>>> | ide: Falcon IDE controller
>>>>> | Probing IDE interface ide0...
>>>>> | hda: Sarge m68k, ATA DISK drive
>>>>> | ide0 at 0xfff00000 on irq 15 (serialized)
>>>>> | ide-gd driver 1.18
>>>>> | hda: max request size: 128KiB
>>>>> | hda: 2118816 sectors (1084 MB) w/256KiB Cache, CHS=2102/16/63
>>>>>
>>>>> The next expected line is the partition parsing:
>>>>>
>>>>> | hda: AHDI hda1 hda2
>>>>
>>>> Geert, does this work for you?
>>>
>>> Yep.Thanks!
>>
>> Great! I think we should place those blk_delay_queue() calls under the
>> if (rq), that should workd to and be more optimal. Can I ask you to
>> check that, too?
>>
>> So:
>>
>> if (rq) {
>>        blk_requeue_request(q, rq);
>>        blk_delay_queue(q, msecs_to_jiffies(3));
>> }
>>
>> in both locations.
> 
> Works, too.

Great, thanks for testing! I'll queue this up and submit it so it makes
2.6.39-rc1.

-- 
Jens Axboe