EmbeddedRelated.com
Forums
Memfault Beyond the Launch

EMIFA and NAND flash

Started by Vladimir Vassilevsky November 20, 2010
According to the meager description in the SPRUFL6E datasheet, the EMIFA 
of TMS 674x/OMAP supports hardware ECC for NAND flash pages up to 512 bytes.

Would it be possible to use the hardware ECC facility for larger pages 
by restarting the ECC calculation at every 512 bytes of transfer? Or 
will I have to do everything in the software?


Vladimir Vassilevsky
DSP and Mixed Signal Design Consultant
http://www.abvolt.com

On Nov 20, 11:45=A0am, Vladimir Vassilevsky <nos...@nowhere.com> wrote:
> According to the meager description in the SPRUFL6E datasheet, the EMIFA > of TMS 674x/OMAP supports hardware ECC for NAND flash pages up to 512 byt=
es.
> > Would it be possible to use the hardware ECC facility for larger pages > by restarting the ECC calculation at every 512 bytes of transfer? Or > will I have to do everything in the software? > > Vladimir Vassilevsky > DSP and Mixed Signal Design Consultanthttp://www.abvolt.com
Sounds like a good question for TI support. They are actually pretty responsive and knowledgeable on their forums: http://e2e.ti.com Jason

Jason wrote:

> On Nov 20, 11:45 am, Vladimir Vassilevsky <nos...@nowhere.com> wrote: > >>According to the meager description in the SPRUFL6E datasheet, the EMIFA >>of TMS 674x/OMAP supports hardware ECC for NAND flash pages up to 512 bytes. >> >>Would it be possible to use the hardware ECC facility for larger pages >>by restarting the ECC calculation at every 512 bytes of transfer? Or >>will I have to do everything in the software? >> > > Sounds like a good question for TI support. They are actually pretty > responsive and knowledgeable on their forums: http://e2e.ti.com
Sounds like insoluble question for TI support. Based on the previous experience, I doubt there will be anything useful from them. VLV
On Nov 22, 5:14=A0pm, Vladimir Vassilevsky <nos...@nowhere.com> wrote:
> Jason wrote: > > On Nov 20, 11:45 am, Vladimir Vassilevsky <nos...@nowhere.com> wrote: > > >>According to the meager description in the SPRUFL6E datasheet, the EMIF=
A
> >>of TMS 674x/OMAP supports hardware ECC for NAND flash pages up to 512 b=
ytes.
> > >>Would it be possible to use the hardware ECC facility for larger pages > >>by restarting the ECC calculation at every 512 bytes of transfer? Or > >>will I have to do everything in the software? > > > Sounds like a good question for TI support. They are actually pretty > > responsive and knowledgeable on their forums:http://e2e.ti.com > > Sounds like insoluble question for TI support. Based on the previous > experience, I doubt there will be anything useful from them. > > VLV
How many devices will you be buying?

John wrote:

> On Nov 22, 5:14 pm, Vladimir Vassilevsky <nos...@nowhere.com> wrote: > >>Jason wrote: >> >>>On Nov 20, 11:45 am, Vladimir Vassilevsky <nos...@nowhere.com> wrote: >> >>>>According to the meager description in the SPRUFL6E datasheet, the EMIFA >>>>of TMS 674x/OMAP supports hardware ECC for NAND flash pages up to 512 bytes. >> >>>>Would it be possible to use the hardware ECC facility for larger pages >>>>by restarting the ECC calculation at every 512 bytes of transfer? Or >>>>will I have to do everything in the software? >> >>>Sounds like a good question for TI support. They are actually pretty >>>responsive and knowledgeable on their forums:http://e2e.ti.com >> >>Sounds like insoluble question for TI support. Based on the previous >>experience, I doubt there will be anything useful from them. >> > How many devices will you be buying?
How many do you think would be enough to get support from real DSP engineers? I tried going through regional FAEs, too. VLV

Vladimir Vassilevsky wrote:

> > According to the meager description in the SPRUFL6E datasheet, the EMIFA > of TMS 674x/OMAP supports hardware ECC for NAND flash pages up to 512 > bytes. > > Would it be possible to use the hardware ECC facility for larger pages > by restarting the ECC calculation at every 512 bytes of transfer? Or > will I have to do everything in the software?
TI support must be parsing comp.dsp. A large page should be read or written as several chunks of 512 bytes. Each chunk should be arranged as complete data transfer with ECC calculation. Daniel Allred from TI pointed me to Linux kernel source as an example. VLV
On Nov 22, 6:37=A0pm, Vladimir Vassilevsky <nos...@nowhere.com> wrote:
> John wrote: > > On Nov 22, 5:14 pm, Vladimir Vassilevsky <nos...@nowhere.com> wrote: > > >>Jason wrote: > > >>>On Nov 20, 11:45 am, Vladimir Vassilevsky <nos...@nowhere.com> wrote: > > >>>>According to the meager description in the SPRUFL6E datasheet, the EM=
IFA
> >>>>of TMS 674x/OMAP supports hardware ECC for NAND flash pages up to 512=
bytes.
> > >>>>Would it be possible to use the hardware ECC facility for larger page=
s
> >>>>by restarting the ECC calculation at every 512 bytes of transfer? Or > >>>>will I have to do everything in the software? > > >>>Sounds like a good question for TI support. They are actually pretty > >>>responsive and knowledgeable on their forums:http://e2e.ti.com > > >>Sounds like insoluble question for TI support. Based on the previous > >>experience, I doubt there will be anything useful from them. > > > How many devices will you be buying? > > How many do you think would be enough to get support from real DSP > engineers? I tried going through regional FAEs, too. > > VLV
Less than 100k is what I always say.

Memfault Beyond the Launch