DDR eMMC write failures

DDR eMMC write failures

[Praveen Krishnan]

Hello all,

I am working on getting the DDR50 mode up and running for the eMMC
cards.  I am running into some weird CRC error cases.  Initially, I
thought that everything works fine, but when I stress test it with
large amounts of data (more than 100MB of writes), I notice that my
console is spewed with venomous error messages.  I get a CRC error and
every now and then, I also see this message "Got data interrupt even
though no data operation was in progress" and "Buffer I/O error" and
so on.

Can anybody give any pointers as to what could be causing this?

Thanks a lot,
Praveen

Re: DDR eMMC write failures

[Praveen Krishnan]

I am seeing some very strange things.

When I do

Dd if=/dev/zero of=/dev/mmcblk0 bs=512k count=200 (it fails when
writing about 100MB of data with a crc error)

When I do
Dd if=/dev/zero of=/dev/mmcblk0 bs=64k count=3000 (it fails only after
125 MB of data with a crc error)

When I do

Dd if=/dev/zero of=/dev/mmcblk0 bs=32k count=16000 (it fails when I
try to write 500MB of data, but good until 375MB of data)

Any ideas what I am running into?

Thanks,
Praveen


On Wed, Nov 2, 2011 at 4:51 PM, Praveen Krishnan <praveen.gk@gmail.com> wrote:
> Hello all,
>
> I am working on getting the DDR50 mode up and running for the eMMC
> cards.  I am running into some weird CRC error cases.  Initially, I
> thought that everything works fine, but when I stress test it with
> large amounts of data (more than 100MB of writes), I notice that my
> console is spewed with venomous error messages.  I get a CRC error and
> every now and then, I also see this message "Got data interrupt even
> though no data operation was in progress" and "Buffer I/O error" and
> so on.
>
> Can anybody give any pointers as to what could be causing this?
>
> Thanks a lot,
> Praveen
>

RE: DDR eMMC write failures

[Subhash Jadavani]

I guess there is some issue with board level signal condition issue on MMC
interface lines. You may try increasing the drive strength of interface
lines from host controller side.

Regards,
Subhash

> -----Original Message-----
> From: linux-mmc-owner@vger.kernel.org [mailto:linux-mmc-
> owner@vger.kernel.org] On Behalf Of Praveen Krishnan
> Sent: Saturday, November 05, 2011 5:57 AM
> To: linux-mmc@vger.kernel.org
> Subject: Re: DDR eMMC write failures
> 
> I am seeing some very strange things.
> 
> When I do
> 
> Dd if=/dev/zero of=/dev/mmcblk0 bs=512k count=200 (it fails when
> writing about 100MB of data with a crc error)
> 
> When I do
> Dd if=/dev/zero of=/dev/mmcblk0 bs=64k count=3000 (it fails only after
> 125 MB of data with a crc error)
> 
> When I do
> 
> Dd if=/dev/zero of=/dev/mmcblk0 bs=32k count=16000 (it fails when I
> try to write 500MB of data, but good until 375MB of data)
> 
> Any ideas what I am running into?
> 
> Thanks,
> Praveen
> 
> 
> On Wed, Nov 2, 2011 at 4:51 PM, Praveen Krishnan <praveen.gk@gmail.com>
> wrote:
> > Hello all,
> >
> > I am working on getting the DDR50 mode up and running for the eMMC
> > cards.  I am running into some weird CRC error cases.  Initially, I
> > thought that everything works fine, but when I stress test it with
> > large amounts of data (more than 100MB of writes), I notice that my
> > console is spewed with venomous error messages.  I get a CRC error
> and
> > every now and then, I also see this message "Got data interrupt even
> > though no data operation was in progress" and "Buffer I/O error" and
> > so on.
> >
> > Can anybody give any pointers as to what could be causing this?
> >
> > Thanks a lot,
> > Praveen
> >
> --
> To unsubscribe from this list: send the line "unsubscribe linux-mmc" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html

RE: DDR eMMC write failures

[Subhash Jadavani]

You may also try by lowering the clock speed (from 50mhz to some lower value
say 40 or 30 Mhz). See if you are still seeing the data crc errors with
lower clock speed in DDR50 mode or not.

Regards,
Subhash

> -----Original Message-----
> From: linux-mmc-owner@vger.kernel.org [mailto:linux-mmc-
> owner@vger.kernel.org] On Behalf Of Subhash Jadavani
> Sent: Sunday, November 13, 2011 10:05 PM
> To: 'Praveen Krishnan'; linux-mmc@vger.kernel.org
> Subject: RE: DDR eMMC write failures
> 
> I guess there is some issue with board level signal condition issue on
> MMC
> interface lines. You may try increasing the drive strength of interface
> lines from host controller side.
> 
> Regards,
> Subhash
> 
> > -----Original Message-----
> > From: linux-mmc-owner@vger.kernel.org [mailto:linux-mmc-
> > owner@vger.kernel.org] On Behalf Of Praveen Krishnan
> > Sent: Saturday, November 05, 2011 5:57 AM
> > To: linux-mmc@vger.kernel.org
> > Subject: Re: DDR eMMC write failures
> >
> > I am seeing some very strange things.
> >
> > When I do
> >
> > Dd if=/dev/zero of=/dev/mmcblk0 bs=512k count=200 (it fails when
> > writing about 100MB of data with a crc error)
> >
> > When I do
> > Dd if=/dev/zero of=/dev/mmcblk0 bs=64k count=3000 (it fails only
> after
> > 125 MB of data with a crc error)
> >
> > When I do
> >
> > Dd if=/dev/zero of=/dev/mmcblk0 bs=32k count=16000 (it fails when I
> > try to write 500MB of data, but good until 375MB of data)
> >
> > Any ideas what I am running into?
> >
> > Thanks,
> > Praveen
> >
> >
> > On Wed, Nov 2, 2011 at 4:51 PM, Praveen Krishnan
> <praveen.gk@gmail.com>
> > wrote:
> > > Hello all,
> > >
> > > I am working on getting the DDR50 mode up and running for the eMMC
> > > cards.  I am running into some weird CRC error cases.  Initially, I
> > > thought that everything works fine, but when I stress test it with
> > > large amounts of data (more than 100MB of writes), I notice that my
> > > console is spewed with venomous error messages.  I get a CRC error
> > and
> > > every now and then, I also see this message "Got data interrupt
> even
> > > though no data operation was in progress" and "Buffer I/O error"
> and
> > > so on.
> > >
> > > Can anybody give any pointers as to what could be causing this?
> > >
> > > Thanks a lot,
> > > Praveen
> > >
> > --
> > To unsubscribe from this list: send the line "unsubscribe linux-mmc"
> in
> > the body of a message to majordomo@vger.kernel.org
> > More majordomo info at  http://vger.kernel.org/majordomo-info.html
> 
> --
> To unsubscribe from this list: send the line "unsubscribe linux-mmc" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: DDR eMMC write failures

[Praveen Krishnan]

On Sun, Nov 13, 2011 at 8:48 AM, Subhash Jadavani
<subhashj@codeaurora.org> wrote:
>
> You may also try by lowering the clock speed (from 50mhz to some lower value
> say 40 or 30 Mhz). See if you are still seeing the data crc errors with
> lower clock speed in DDR50 mode or not.
Figured there was an hardware issue.  The DMA was not able to service
the FIFO to keep pace with the DDR transfer rate.  There was a FIFO
underrun happening, because of which the clock used to get gated in
the middle of a transfer.  This was leading to a CRC issue.  Adjusted
the drive strength as well (and this fixes the initialization issue
when the cards used to randomly fail during init). And yes, have seen
that it works at lower frequencies. Anyway, thanks for the
suggestions.
> Regards,
> Subhash
>
> > -----Original Message-----
> > From: linux-mmc-owner@vger.kernel.org [mailto:linux-mmc-
> > owner@vger.kernel.org] On Behalf Of Subhash Jadavani
> > Sent: Sunday, November 13, 2011 10:05 PM
> > To: 'Praveen Krishnan'; linux-mmc@vger.kernel.org
> > Subject: RE: DDR eMMC write failures
> >
> > I guess there is some issue with board level signal condition issue on
> > MMC
> > interface lines. You may try increasing the drive strength of interface
> > lines from host controller side.
> >
> > Regards,
> > Subhash
> >
> > > -----Original Message-----
> > > From: linux-mmc-owner@vger.kernel.org [mailto:linux-mmc-
> > > owner@vger.kernel.org] On Behalf Of Praveen Krishnan
> > > Sent: Saturday, November 05, 2011 5:57 AM
> > > To: linux-mmc@vger.kernel.org
> > > Subject: Re: DDR eMMC write failures
> > >
> > > I am seeing some very strange things.
> > >
> > > When I do
> > >
> > > Dd if=/dev/zero of=/dev/mmcblk0 bs=512k count=200 (it fails when
> > > writing about 100MB of data with a crc error)
> > >
> > > When I do
> > > Dd if=/dev/zero of=/dev/mmcblk0 bs=64k count=3000 (it fails only
> > after
> > > 125 MB of data with a crc error)
> > >
> > > When I do
> > >
> > > Dd if=/dev/zero of=/dev/mmcblk0 bs=32k count=16000 (it fails when I
> > > try to write 500MB of data, but good until 375MB of data)
> > >
> > > Any ideas what I am running into?
> > >
> > > Thanks,
> > > Praveen
> > >
> > >
> > > On Wed, Nov 2, 2011 at 4:51 PM, Praveen Krishnan
> > <praveen.gk@gmail.com>
> > > wrote:
> > > > Hello all,
> > > >
> > > > I am working on getting the DDR50 mode up and running for the eMMC
> > > > cards.  I am running into some weird CRC error cases.  Initially, I
> > > > thought that everything works fine, but when I stress test it with
> > > > large amounts of data (more than 100MB of writes), I notice that my
> > > > console is spewed with venomous error messages.  I get a CRC error
> > > and
> > > > every now and then, I also see this message "Got data interrupt
> > even
> > > > though no data operation was in progress" and "Buffer I/O error"
> > and
> > > > so on.
> > > >
> > > > Can anybody give any pointers as to what could be causing this?
> > > >
> > > > Thanks a lot,
> > > > Praveen
> > > >
> > > --
> > > To unsubscribe from this list: send the line "unsubscribe linux-mmc"
> > in
> > > the body of a message to majordomo@vger.kernel.org
> > > More majordomo info at  http://vger.kernel.org/majordomo-info.html
> >
> > --
> > To unsubscribe from this list: send the line "unsubscribe linux-mmc" in
> > the body of a message to majordomo@vger.kernel.org
> > More majordomo info at  http://vger.kernel.org/majordomo-info.html
>