1/1 
|
OCIDEC-3 in verilog? Do PIO only accesses to hard drives workwithout DMA?
by Unknown on Jan 20, 2004 |
Not available! | ||
|
Hello, Take a look at the FFR16 Core. It has two modules, one processes the ATA-ATAPI protocol and the other the FAT volume. Its purpose is to read files stored into the IDE device. The first one can be used independently. Benefits : Really small area needed, easy to adapt (soft. part), WB compatible. Drawbacks : Target FPGAs. The code is prepared for read (i know that other OpenCores users have done the "write" version). Although it doesn´t macth your needs the code of the ATA protocol processing is commented into the source files and it could be helpful. Armando, At 03:26 20/01/2004 +0100, you wrote:
>
> Hello, > > I am needing an IDE controller for my current SOC design and have been > looking at the Opencores OCIDEC core. I am unfamiliar with IDE > communication so I hope I don't ask a stupid question. There are no stupid questions, only stupid answers.
> We are needing to talk to 1.8" or 2.5" hard drives. All of the drives
> we are looking at support the ATA (2-5) standards. > > First question: > Is there verilog rtl code available for the OCIDEC-3 core? No
> I found the
> VHDL version, but cannot find the verilog version. This core seems the
> best fit our target application, but I don't have time to convert the
> VHDL to verilog. > > Second question: > If there is no verilog version of the OCIDEC-3, will the OCIDEC-1 or 2 > talk to hard drives, or are they just able to talk to Compact Flash / PC
> cards?
They can (and do) talk to harddisks. It's merely the other way around. CF/PCCards can use the same protocol as harddisks. PIO means Progammed IO, or in normal language: the CPU needs to take care of the IO; moving the data from/to the device. If you need to process a lot of data this can be quite a burden. However you can create your own dma engine that handles this task for the CPU. To the DMA controller it would appear is if it is talking to a FIFO. Note that this is not the same as IDE-DMA modes!! Work is in progress to create a newer, more flexible version of the OCIDEC cores. This will be a one in all core, instead of having increasingly complex cores (as is the case with OCIDEC-1,2,3) you simply enable the options you need (and disable what you don't need). PIO modes are working fine (and tested in hardware). DMA modes are still under development. If you can live with the beta-status of this core send me an email and I will send you the files. Richard
> My intuition tells me that I could make the OCIDEC-1 or 2 work,
> we just give up the DMA capability of the harddrive? Obviously this > would increase the CPU overhead to access the harddrive, but we can live
> with this. Basically, I need to know if we can talk to a current
> harddrive using PIO transfers only, ignoring the DMA between the IDE > controller and the drive. PIO Mode 4 @ 16.6MB/s is plenty fast for what
> we need to do. We just need to be able to read/write to the drive.
>
> Thanks,
> Jeff Hanoch
>
> _______________________________________________
> http://www.opencores.org/mailman/listinfo/cores
_______________________________________________
http://www.opencores.org/mailman/listinfo/cores
|
|||
1/1