Re: Fw: [PATCH] ahci: add ACard 8620 support (rough draft)

Re: Fw: [PATCH] ahci: add ACard 8620 support (rough draft)

[Jeff Garzik]

Jim Yeh wrote:
> Dear Jeff,
> 
> This is Jim,an IC PM from ACARD Technology. Thanks for your effort on 
> building Linux driver for ACARD ATP8620. Since ACARD already have a 
> verified AHCI driver for ATP8620, we'd like to submit this one for the 
> kernel built-in driver (please see attachment). This driver support 
> Linux Kernel 2.6.29 AHCI mode for ATP8620 (2ch SATA IC) and ATP8624 (4ch 
> SATA IC). It already passed our own stability and compatibility test 
> with SATA devices like HDD, ODD, or PortMultiplier. If you have any 
> question or need the hardware for testing, please don't hesitate to 
> contact me. Thanks again for your help.

Thank you for your email.

We understand you have verified your driver, but it is standard Linux 
driver policy to avoid code duplication.  Multiple implementations of 
AHCI drivers implies several negative factors:

* each bug fix must be reviewed across multiple AHCI drivers, to 
determine if the bug fix must be copied

* non-uniform user experience, depending on AHCI platform

* Shared AHCI code gets much more testing and verification.

* When a hardware vendor's AHCI chip reaches End Of Life, or the company 
goes out of business, users are not abandoned.

So, for the official Linux kernel drivers that are shipped by Red Hat, 
Novell/SuSE, Canonical and other vendors, we will need to modify the 
existing Linux AHCI driver to support ATP8620 and ATP8624.


And yes -- test hardware would be greatly appreciated!  If you could 
send these, that would be appreciated:

	To Tejun Heo
	------------
	1 x ATP8620
	1 x ATP8624

	To Jeff Garzik
	--------------
	1 x ATP8620
	1 x ATP8624

I assume this is PCI, PCI-X or PCI-Express form factor?

Regards and thanks,

	Jeff

Re: Fw: [PATCH] ahci: add ACard 8620 support (rough draft)

[Jim Yeh]

Dear Jeff,

Thanks for your explanation. We agree with you.

 I'll prepare the hardware for your testing. The form factor of PCB will 
be a 32bit PCI-base add-on card. Please advise the address for both of 
yours. It is expected to be sent by next week.

Regards,
 
Jim Yeh
Product Manager
Marketing Dept.
ACARD Technology Corp.
TEL:886-2-85122290#3244



Jeff Garzik 提到:
> Jim Yeh wrote:
>> Dear Jeff,
>>
>> This is Jim,an IC PM from ACARD Technology. Thanks for your effort on 
>> building Linux driver for ACARD ATP8620. Since ACARD already have a 
>> verified AHCI driver for ATP8620, we'd like to submit this one for 
>> the kernel built-in driver (please see attachment). This driver 
>> support Linux Kernel 2.6.29 AHCI mode for ATP8620 (2ch SATA IC) and 
>> ATP8624 (4ch SATA IC). It already passed our own stability and 
>> compatibility test with SATA devices like HDD, ODD, or 
>> PortMultiplier. If you have any question or need the hardware for 
>> testing, please don't hesitate to contact me. Thanks again for your 
>> help.
>
> Thank you for your email.
>
> We understand you have verified your driver, but it is standard Linux 
> driver policy to avoid code duplication.  Multiple implementations of 
> AHCI drivers implies several negative factors:
>
> * each bug fix must be reviewed across multiple AHCI drivers, to 
> determine if the bug fix must be copied
>
> * non-uniform user experience, depending on AHCI platform
>
> * Shared AHCI code gets much more testing and verification.
>
> * When a hardware vendor's AHCI chip reaches End Of Life, or the 
> company goes out of business, users are not abandoned.
>
> So, for the official Linux kernel drivers that are shipped by Red Hat, 
> Novell/SuSE, Canonical and other vendors, we will need to modify the 
> existing Linux AHCI driver to support ATP8620 and ATP8624.
>
>
> And yes -- test hardware would be greatly appreciated!  If you could 
> send these, that would be appreciated:
>
>     To Tejun Heo
>     ------------
>     1 x ATP8620
>     1 x ATP8624
>
>     To Jeff Garzik
>     --------------
>     1 x ATP8620
>     1 x ATP8624
>
> I assume this is PCI, PCI-X or PCI-Express form factor?
>
> Regards and thanks,
>
>     Jeff
>
>
>
>
>
> __________ Information from ESET NOD32 Antivirus, version of virus 
> signature database 4005 (20090413) __________
>
> The message was checked by ESET NOD32 Antivirus.
>
> http://www.eset.com
>
>
>
>

Re: Fw: [PATCH] ahci: add ACard 8620 support (rough draft)

[Tejun Heo]

Jim Yeh wrote:
> Dear Jeff,
> 
> Thanks for your explanation. We agree with you.
> 
> I'll prepare the hardware for your testing. The form factor of PCB will
> be a 32bit PCI-base add-on card. Please advise the address for both of
> yours. It is expected to be sent by next week.

My address is

 Bekgung Dongyang Paragon B#2107
 Jeongja, Bundang, Seongnam, Kyoungki
 South Korea
 zip: 463842
 cell: +82-10-5488-4679 (please print on top of the package)

Thanks.

-- 
tejun

Re: Fw: [PATCH] ahci: add ACard 8620 support (rough draft)

[Jeff Garzik]

Jim Yeh wrote:
> Hi, Tejun and Jeff,
> 
> The PCB samples was sent early today. There are two PCB inside, the PCB 
> AEC6293 is for IC ATP8620, the other one is for IC ATP8624. The quick 
> installation guide and latest version datasheet was attached. Fedex 
> tracking no. is listed below. Any question please let me know.

(removed linux-kernel mailing list from CC)

Thank you for these cards and documents.  Unfortunately the documents 
will not be much help with the driver, because they do not include 
information on registers and DMA data structures.

Many months ago, you gave me permission to post this in our hardware 
document archive for programmers:
http://gkernel.sourceforge.net/specs/acard/ATP8620pub_Rev0.98.pdf



Do you have updated documents for 8620 and 8624, that include register 
and DMA data structure information?

If yes, may we have permission to post these in our programmer's archive 
on the Web?  http://gkernel.sourceforge.net/specs/

Thanks and best regards,

	Jeff