simple sd card performance bug fix on OMAP

simple sd card performance bug fix on OMAP

[Stephen Schwarm, CSDP]

[-- Attachment #1: Type: text/plain, Size: 922 bytes --]

I have not had time to make a formal patch for this but I thought it was 
significant enough just to send it out.  The problem is on omap systems 
that use omap_hsmmc.c to run sd or mmc cards.  If the system you are 
using has an 8 wire interface, it will only use a one wire interface to 
4 wire cards (eg, sd class 4 and sd class 6).

In the file driver/mmc/omap_hsmmc.c in the function omap_hsmmc_probe:
at about line 1739 change:

    if (mmc_slot(host).wires >= 8)
        mmc->caps |= MMC_CAP_8_BIT_DATA;
    else if (mmc_slot(host).wires >= 4)
        mmc->caps |= MMC_CAP_4_BIT_DATA;

to:

    if (mmc_slot(host).wires >= 8)
        mmc->caps |= MMC_CAP_8_BIT_DATA;
    if (mmc_slot(host).wires >= 4)
        mmc->caps |= w;

just delete the word "else".  This sets the MMC_CAP_4_BIT_DATA bit on 
interfaces that have 8 wire interfaces.

We have seen a BIG performance improvement on our systems.

Steve Schwarm



[-- Attachment #2: w3eve.vcf --]
[-- Type: text/x-vcard, Size: 227 bytes --]

begin:vcard
fn:Stephen  Schwarm, CSDP
n:Schwarm, CSDP;Stephen 
adr;dom:;;30 Hayden Woods;Wrentham;MA;02093
email;internet:schwarm@ieee.org
tel;home:508-384-7697
tel;cell:508-922-9688
x-mozilla-html:FALSE
version:2.1
end:vcard

RE: simple sd card performance bug fix on OMAP

[Madhusudhan]

> -----Original Message-----
> From: linux-omap-owner@vger.kernel.org [mailto:linux-omap-
> owner@vger.kernel.org] On Behalf Of Stephen Schwarm, CSDP
> Sent: Friday, May 28, 2010 2:05 PM
> To: linux-omap@vger.kernel.org
> Subject: simple sd card performance bug fix on OMAP
> 
> I have not had time to make a formal patch for this but I thought it was
> significant enough just to send it out.  The problem is on omap systems
> that use omap_hsmmc.c to run sd or mmc cards.  If the system you are
> using has an 8 wire interface, it will only use a one wire interface to
> 4 wire cards (eg, sd class 4 and sd class 6).
> 
A patch to fix this was already posted on LO by Kishore.

Regards,
Madhu

> In the file driver/mmc/omap_hsmmc.c in the function omap_hsmmc_probe:
> at about line 1739 change:
> 
>     if (mmc_slot(host).wires >= 8)
>         mmc->caps |= MMC_CAP_8_BIT_DATA;
>     else if (mmc_slot(host).wires >= 4)
>         mmc->caps |= MMC_CAP_4_BIT_DATA;
> 
> to:
> 
>     if (mmc_slot(host).wires >= 8)
>         mmc->caps |= MMC_CAP_8_BIT_DATA;
>     if (mmc_slot(host).wires >= 4)
>         mmc->caps |= w;
> 
> just delete the word "else".  This sets the MMC_CAP_4_BIT_DATA bit on
> interfaces that have 8 wire interfaces.
> 
> We have seen a BIG performance improvement on our systems.
> 
> Steve Schwarm
> 

RE: simple sd card performance bug fix on OMAP

[Gadiyar, Anand]

Stephen Schwarm, CSDP wrote:
> 
> I have not had time to make a formal patch for this but I thought it was
> significant enough just to send it out.  The problem is on omap systems
> that use omap_hsmmc.c to run sd or mmc cards.  If the system you are
> using has an 8 wire interface, it will only use a one wire interface to
> 4 wire cards (eg, sd class 4 and sd class 6).
> 
> In the file driver/mmc/omap_hsmmc.c in the function omap_hsmmc_probe:
> at about line 1739 change:
> 
>     if (mmc_slot(host).wires >= 8)
>         mmc->caps |= MMC_CAP_8_BIT_DATA;
>     else if (mmc_slot(host).wires >= 4)
>         mmc->caps |= MMC_CAP_4_BIT_DATA;
> 
> to:
> 
>     if (mmc_slot(host).wires >= 8)
>         mmc->caps |= MMC_CAP_8_BIT_DATA;
>     if (mmc_slot(host).wires >= 4)
>         mmc->caps |= w;
> 
> just delete the word "else".  This sets the MMC_CAP_4_BIT_DATA bit on
> interfaces that have 8 wire interfaces.
> 
> We have seen a BIG performance improvement on our systems.
> 

Thanks for reporting this.

A similar patch was posted a while ago and is currently marked as "awaiting upstream", although it looks like it got lost again.

<https://patchwork.kernel.org/patch/78713/>

- Anand

RE: simple sd card performance bug fix on OMAP

[Madhusudhan]

> -----Original Message-----
> From: linux-omap-owner@vger.kernel.org [mailto:linux-omap-
> owner@vger.kernel.org] On Behalf Of Gadiyar, Anand
> Sent: Friday, May 28, 2010 2:55 PM
> To: Stephen Schwarm, CSDP; linux-omap@vger.kernel.org
> Subject: RE: simple sd card performance bug fix on OMAP
> 
> Stephen Schwarm, CSDP wrote:
> >
> > I have not had time to make a formal patch for this but I thought it was
> > significant enough just to send it out.  The problem is on omap systems
> > that use omap_hsmmc.c to run sd or mmc cards.  If the system you are
> > using has an 8 wire interface, it will only use a one wire interface to
> > 4 wire cards (eg, sd class 4 and sd class 6).
> >
> > In the file driver/mmc/omap_hsmmc.c in the function omap_hsmmc_probe:
> > at about line 1739 change:
> >
> >     if (mmc_slot(host).wires >= 8)
> >         mmc->caps |= MMC_CAP_8_BIT_DATA;
> >     else if (mmc_slot(host).wires >= 4)
> >         mmc->caps |= MMC_CAP_4_BIT_DATA;
> >
> > to:
> >
> >     if (mmc_slot(host).wires >= 8)
> >         mmc->caps |= MMC_CAP_8_BIT_DATA;
> >     if (mmc_slot(host).wires >= 4)
> >         mmc->caps |= w;
> >
> > just delete the word "else".  This sets the MMC_CAP_4_BIT_DATA bit on
> > interfaces that have 8 wire interfaces.
> >
> > We have seen a BIG performance improvement on our systems.
> >
> 
> Thanks for reporting this.
> 
> A similar patch was posted a while ago and is currently marked as
> "awaiting upstream", although it looks like it got lost again.
> 
> <https://patchwork.kernel.org/patch/78713/>
> 
A revised version of the same patch was posted.

https://patchwork.kernel.org/patch/93519/

Tony, Can you please push this?

Regards,
Madhu

> - Anand
> --
> To unsubscribe from this list: send the line "unsubscribe linux-omap" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: simple sd card performance bug fix on OMAP

[Tony Lindgren]

* Madhusudhan <madhu.cr@ti.com> [100602 20:13]:
> 
> 
> > -----Original Message-----
> > From: linux-omap-owner@vger.kernel.org [mailto:linux-omap-
> > owner@vger.kernel.org] On Behalf Of Gadiyar, Anand
> > Sent: Friday, May 28, 2010 2:55 PM
> > To: Stephen Schwarm, CSDP; linux-omap@vger.kernel.org
> > Subject: RE: simple sd card performance bug fix on OMAP
> > 
> > Stephen Schwarm, CSDP wrote:
> > >
> > > I have not had time to make a formal patch for this but I thought it was
> > > significant enough just to send it out.  The problem is on omap systems
> > > that use omap_hsmmc.c to run sd or mmc cards.  If the system you are
> > > using has an 8 wire interface, it will only use a one wire interface to
> > > 4 wire cards (eg, sd class 4 and sd class 6).
> > >
> > > In the file driver/mmc/omap_hsmmc.c in the function omap_hsmmc_probe:
> > > at about line 1739 change:
> > >
> > >     if (mmc_slot(host).wires >= 8)
> > >         mmc->caps |= MMC_CAP_8_BIT_DATA;
> > >     else if (mmc_slot(host).wires >= 4)
> > >         mmc->caps |= MMC_CAP_4_BIT_DATA;
> > >
> > > to:
> > >
> > >     if (mmc_slot(host).wires >= 8)
> > >         mmc->caps |= MMC_CAP_8_BIT_DATA;
> > >     if (mmc_slot(host).wires >= 4)
> > >         mmc->caps |= w;
> > >
> > > just delete the word "else".  This sets the MMC_CAP_4_BIT_DATA bit on
> > > interfaces that have 8 wire interfaces.
> > >
> > > We have seen a BIG performance improvement on our systems.
> > >
> > 
> > Thanks for reporting this.
> > 
> > A similar patch was posted a while ago and is currently marked as
> > "awaiting upstream", although it looks like it got lost again.
> > 
> > <https://patchwork.kernel.org/patch/78713/>
> > 
> A revised version of the same patch was posted.
> 
> https://patchwork.kernel.org/patch/93519/
> 
> Tony, Can you please push this?

This needs to get merged via linux-mmc list.

Regards,

Tony

RE: simple sd card performance bug fix on OMAP

[Madhusudhan]

> -----Original Message-----
> From: Tony Lindgren [mailto:tony@atomide.com]
> Sent: Thursday, June 10, 2010 1:36 AM
> To: Madhusudhan
> Cc: 'Gadiyar, Anand'; 'Stephen Schwarm, CSDP'; linux-omap@vger.kernel.org
> Subject: Re: simple sd card performance bug fix on OMAP
> 
> * Madhusudhan <madhu.cr@ti.com> [100602 20:13]:
> >
> >
> > > -----Original Message-----
> > > From: linux-omap-owner@vger.kernel.org [mailto:linux-omap-
> > > owner@vger.kernel.org] On Behalf Of Gadiyar, Anand
> > > Sent: Friday, May 28, 2010 2:55 PM
> > > To: Stephen Schwarm, CSDP; linux-omap@vger.kernel.org
> > > Subject: RE: simple sd card performance bug fix on OMAP
> > >
> > > Stephen Schwarm, CSDP wrote:
> > > >
> > > > I have not had time to make a formal patch for this but I thought it
> was
> > > > significant enough just to send it out.  The problem is on omap
> systems
> > > > that use omap_hsmmc.c to run sd or mmc cards.  If the system you are
> > > > using has an 8 wire interface, it will only use a one wire interface
> to
> > > > 4 wire cards (eg, sd class 4 and sd class 6).
> > > >
> > > > In the file driver/mmc/omap_hsmmc.c in the function
> omap_hsmmc_probe:
> > > > at about line 1739 change:
> > > >
> > > >     if (mmc_slot(host).wires >= 8)
> > > >         mmc->caps |= MMC_CAP_8_BIT_DATA;
> > > >     else if (mmc_slot(host).wires >= 4)
> > > >         mmc->caps |= MMC_CAP_4_BIT_DATA;
> > > >
> > > > to:
> > > >
> > > >     if (mmc_slot(host).wires >= 8)
> > > >         mmc->caps |= MMC_CAP_8_BIT_DATA;
> > > >     if (mmc_slot(host).wires >= 4)
> > > >         mmc->caps |= w;
> > > >
> > > > just delete the word "else".  This sets the MMC_CAP_4_BIT_DATA bit
> on
> > > > interfaces that have 8 wire interfaces.
> > > >
> > > > We have seen a BIG performance improvement on our systems.
> > > >
> > >
> > > Thanks for reporting this.
> > >
> > > A similar patch was posted a while ago and is currently marked as
> > > "awaiting upstream", although it looks like it got lost again.
> > >
> > > <https://patchwork.kernel.org/patch/78713/>
> > >
> > A revised version of the same patch was posted.
> >
> > https://patchwork.kernel.org/patch/93519/
> >
> > Tony, Can you please push this?
> 
> This needs to get merged via linux-mmc list.
> 
Sure. We will repost it.

Regards,
Madhu

> Regards,
> 
> Tony