Re: [PATCH 09/13] sata_mv ncq Use DMA memory pools for hardware memory tables

[Jeff Garzik <jgarzik@pobox.com>]

Mark Lord wrote:
> Jeff Garzik wrote:
>> Mark Lord wrote:
>>> Create host-owned DMA memory pools, for use in allocating/freeing 
>>> per-port
>>> command/response queues and SG tables.  This gives us a way to 
>>> guarantee we
>>> meet the hardware address alignment requirements, and also reduces 
>>> memory that
>>> might otherwise be wasted on alignment gaps.
>>>
>>> Signed-off-by: Mark Lord <mlord@pobox.com>
>>
>> ACK patches 1-13
>>
>> applied patches 1-9 to #upstream.  patch #10 failed, with git-am 
>> reporting it as a corrupt patch.
> ..
> 
> That's weird.  I can save the email from linux-ide here,
> and apply as a patch (after 01-09) with no issues at all.
> 
> Jeff got mail reader problems?
> 
> Here it is again, in case it got corrupted in transit to you.

Nope, not corrupted in transit or on this side.  It falls into a 
familiar pattern:

* git-am(1) fails
* patch(1) succeeds
* when applying patch, patch(1) drops a .orig turd

So while patch(1) succeeds because patch(1) is highly forgiving and 
git-am(1) is more strict, something was definitely strange on that 
incoming email.  patch(1) lets you know by giving you a .orig file, 
something not normally created if the patch operation was 100% sound.

ISTR Linus or Junio explaining why git-am(1) was more strict and why it 
was a good thing... As I did in this case, I usually just run patch(1), 
look carefully at the result using 'git diff HEAD', and then 
commit/resolve the changes.

	Jeff