at91sam9263 sdio + wlan ti wl1271

at91sam9263 sdio + wlan ti wl1271

[Olivier Sobrie]

Hi Ludovic,

I'm trying to use a wlan sdio TI wl1271 module with an at91sam9263
processor. The wlan card is correctly recognized by the kernel but
once the driver tries to initialize the card, then I observe sdio
timeouts.

I enabled the traces in the atmel-mci driver and here is what I see
when the problem occurs (kernel v3.11):

[   31.690000] atmel_mci fff84000.mmc: FSM: state=0
[   31.900000] atmel_mci fff84000.mmc: MRQ: cmd 53
[   31.900000] atmel_mci fff84000.mmc: start request: cmd 53
[   31.900000] atmel_mci fff84000.mmc: IRQ: tx buffer empty
[   31.900000] atmel_mci fff84000.mmc: (atmci_pdc_complete) set pending xfer complete
[   31.900000] atmel_mci fff84000.mmc: IRQ: cmd ready
[   31.900000] atmel_mci fff84000.mmc: set pending cmd rdy
[   31.900000] atmel_mci fff84000.mmc: FSM: state=1
[   31.900000] atmel_mci fff84000.mmc: FSM: cmd ready?
[   31.900000] atmel_mci fff84000.mmc: set completed cmd ready
[   31.900000] atmel_mci fff84000.mmc: command with data transfer
[   31.900000] atmel_mci fff84000.mmc: FSM: state=2
[   31.900000] atmel_mci fff84000.mmc: FSM: xfer complete?
[   31.900000] atmel_mci fff84000.mmc: (atmci_tasklet_func) set completed xfer complete
[   31.900000] atmel_mci fff84000.mmc: FSM: state=3
[   31.900000] atmel_mci fff84000.mmc: FSM: not busy?
[   33.900000] atmel_mci fff84000.mmc: software timeout
[   33.900000] atmel_mci fff84000.mmc: FSM: state=5
[   33.900000] atmel_mci fff84000.mmc: FSM: state=0

The driver uses PDC mode to perform the transfers, the chip version is
0x210. The data transferred in cmd53 is 4 bytes long.

It looks like the chip never sets the 'ATMCI_BLKE' flag.
After the problem occured I observe this trace:
[   37.650000] mmc_host mmc0: WARNING: IMR=0x00000008

The same wlan module on an at91sam9g45 (mci version: 0x410) in DMA mode
works without problem.

I also tested a Marvell 8385 (which uses the libertas driver) on
at91sam9263 and it looks to work, I don't see any SDIO timeout.
For this chip, the datasize sent is always >= 16 bytes.

I'm wondering if the Errata "MMC data write operation with less than 12
bytes is impossible.", mentionned in the sources (see function
atmci_prepare_data()), also applies for this version of the chip?
I didn't find this errata in the at91sam9263 datasheet...

If anyone has info or suggestions to make it work, please let me know.

Thanks,

-- 
Olivier

Re: at91sam9263 sdio + wlan ti wl1271

[Ludovic Desroches]

Hi Olivier,

On Tue, Dec 03, 2013 at 10:59:28AM +0100, Olivier Sobrie wrote:
> Hi Ludovic,
> 
> I'm trying to use a wlan sdio TI wl1271 module with an at91sam9263
> processor. The wlan card is correctly recognized by the kernel but
> once the driver tries to initialize the card, then I observe sdio
> timeouts.
> 
> I enabled the traces in the atmel-mci driver and here is what I see
> when the problem occurs (kernel v3.11):
> 
> [   31.690000] atmel_mci fff84000.mmc: FSM: state=0
> [   31.900000] atmel_mci fff84000.mmc: MRQ: cmd 53
> [   31.900000] atmel_mci fff84000.mmc: start request: cmd 53
> [   31.900000] atmel_mci fff84000.mmc: IRQ: tx buffer empty
> [   31.900000] atmel_mci fff84000.mmc: (atmci_pdc_complete) set pending xfer complete
> [   31.900000] atmel_mci fff84000.mmc: IRQ: cmd ready
> [   31.900000] atmel_mci fff84000.mmc: set pending cmd rdy
> [   31.900000] atmel_mci fff84000.mmc: FSM: state=1
> [   31.900000] atmel_mci fff84000.mmc: FSM: cmd ready?
> [   31.900000] atmel_mci fff84000.mmc: set completed cmd ready
> [   31.900000] atmel_mci fff84000.mmc: command with data transfer
> [   31.900000] atmel_mci fff84000.mmc: FSM: state=2
> [   31.900000] atmel_mci fff84000.mmc: FSM: xfer complete?
> [   31.900000] atmel_mci fff84000.mmc: (atmci_tasklet_func) set completed xfer complete
> [   31.900000] atmel_mci fff84000.mmc: FSM: state=3
> [   31.900000] atmel_mci fff84000.mmc: FSM: not busy?
> [   33.900000] atmel_mci fff84000.mmc: software timeout
> [   33.900000] atmel_mci fff84000.mmc: FSM: state=5
> [   33.900000] atmel_mci fff84000.mmc: FSM: state=0
> 
> The driver uses PDC mode to perform the transfers, the chip version is
> 0x210. The data transferred in cmd53 is 4 bytes long.
> 
> It looks like the chip never sets the 'ATMCI_BLKE' flag.
> After the problem occured I observe this trace:
> [   37.650000] mmc_host mmc0: WARNING: IMR=0x00000008
> 
> The same wlan module on an at91sam9g45 (mci version: 0x410) in DMA mode
> works without problem.

Unfortunately I am aware of some issues with sdio modules and PDC mode
but I didn't have the time to correct it yet. I have done some fixes but
only for the DMA mode.

> 
> I also tested a Marvell 8385 (which uses the libertas driver) on
> at91sam9263 and it looks to work, I don't see any SDIO timeout.
> For this chip, the datasize sent is always >= 16 bytes.
> 

In fact I did some tests with Marvell 8686 and it works well but I have
recently heard about issues with other sdio wifi modules...

I will work on this topic as soon as possible.

> I'm wondering if the Errata "MMC data write operation with less than 12
> bytes is impossible.", mentionned in the sources (see function
> atmci_prepare_data()), also applies for this version of the chip?
> I didn't find this errata in the at91sam9263 datasheet...

>From the top of my head, I don't remember until which IP version this
errata applies. It only involves a controller reset.

> 
> If anyone has info or suggestions to make it work, please let me know.
> 

Regards

Ludovic

[PATCH] atmel-mci: fix transfers < 16 bytes problem in PDC mode

[Olivier Sobrie]

This patch attempts to fix a problem when trying to transfer
small data.

Without the patch, when there is a transfer of less than 16 bytes,
it hangs. Here is what I observed for a transfer of 4 bytes with
the traces enabled in the driver:

[   31.690000] atmel_mci fff84000.mmc: FSM: state=0
[   31.900000] atmel_mci fff84000.mmc: MRQ: cmd 53
[   31.900000] atmel_mci fff84000.mmc: start request: cmd 53
[   31.900000] atmel_mci fff84000.mmc: IRQ: tx buffer empty
[   31.900000] atmel_mci fff84000.mmc: (atmci_pdc_complete) set pending xfer complete
[   31.900000] atmel_mci fff84000.mmc: IRQ: cmd ready
[   31.900000] atmel_mci fff84000.mmc: set pending cmd rdy
[   31.900000] atmel_mci fff84000.mmc: FSM: state=1
[   31.900000] atmel_mci fff84000.mmc: FSM: cmd ready?
[   31.900000] atmel_mci fff84000.mmc: set completed cmd ready
[   31.900000] atmel_mci fff84000.mmc: command with data transfer
[   31.900000] atmel_mci fff84000.mmc: FSM: state=2
[   31.900000] atmel_mci fff84000.mmc: FSM: xfer complete?
[   31.900000] atmel_mci fff84000.mmc: (atmci_tasklet_func) set completed xfer complete
[   31.900000] atmel_mci fff84000.mmc: FSM: state=3
[   31.900000] atmel_mci fff84000.mmc: FSM: not busy?
[   33.900000] atmel_mci fff84000.mmc: software timeout
[   33.900000] atmel_mci fff84000.mmc: FSM: state=5
[   33.900000] atmel_mci fff84000.mmc: FSM: state=0

The BLKE bit is never set by the hardware.

By setting the ATMEL_PDC_TCR register with a value equals to 4 when the
datasize sent is smaller than 16 bytes, it looks to solve the problem.

Note that in the old at91_mci driver, the value was set to 3 but in my
case it doesn't work...
http://lxr.free-electrons.com/source/drivers/mmc/host/at91_mci.c?v=3.7#L580
So I made the assumption that the transfer size should be >= 16 bytes.

This patch has been tested with a TI wl1271 sdio module.

Signed-off-by: Olivier Sobrie <olivier@sobrie.be>
---

Hi Ludovic,

With this patch I'm able to use the wlan sdio module wl1271.
When you've time, it would be nice to validate or fix this
patch.

Thanks for your help,

Olivier

 drivers/mmc/host/atmel-mci.c |    6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)

diff --git a/drivers/mmc/host/atmel-mci.c b/drivers/mmc/host/atmel-mci.c
index b86b482..c599731 100644
--- a/drivers/mmc/host/atmel-mci.c
+++ b/drivers/mmc/host/atmel-mci.c
@@ -45,6 +45,7 @@
 
 #define ATMCI_DATA_ERROR_FLAGS	(ATMCI_DCRCE | ATMCI_DTOE | ATMCI_OVRE | ATMCI_UNRE)
 #define ATMCI_DMA_THRESHOLD	16
+#define ATMCI_PDC_MIN_DATASIZE	16
 
 enum {
 	EVENT_CMD_RDY = 0,
@@ -765,7 +766,10 @@ static void atmci_pdc_set_single_buf(struct atmel_mci *host,
 	}
 
 	if (host->data_size <= buf_size) {
-		if (host->data_size & 0x3) {
+		if ((host->data_size < ATMCI_PDC_MIN_DATASIZE) &&
+		    (dir == XFER_TRANSMIT)) {
+			atmci_writel(host, counter_reg, 4);
+		} else if (host->data_size & 0x3) {
 			/* If size is different from modulo 4, transfer bytes */
 			atmci_writel(host, counter_reg, host->data_size);
 			atmci_writel(host, ATMCI_MR, host->mode_reg | ATMCI_MR_PDCFBYTE);
-- 
1.7.9.5

Re: [PATCH] atmel-mci: fix transfers < 16 bytes problem in PDC mode

[Ludovic Desroches]

Hi Olivier,

On Thu, Dec 05, 2013 at 10:45:16AM +0100, Olivier Sobrie wrote:

[...]

> 
> Hi Ludovic,
> 
> With this patch I'm able to use the wlan sdio module wl1271.
> When you've time, it would be nice to validate or fix this
> patch.

Thanks for your patch, before giving my ack I would like to test it on
other devices. I hope I could manage that next week.

Regards

Ludovic

> 
> Thanks for your help,
> 
> Olivier
> 
>  drivers/mmc/host/atmel-mci.c |    6 +++++-
>  1 file changed, 5 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/mmc/host/atmel-mci.c b/drivers/mmc/host/atmel-mci.c
> index b86b482..c599731 100644
> --- a/drivers/mmc/host/atmel-mci.c
> +++ b/drivers/mmc/host/atmel-mci.c
> @@ -45,6 +45,7 @@
>  
>  #define ATMCI_DATA_ERROR_FLAGS	(ATMCI_DCRCE | ATMCI_DTOE | ATMCI_OVRE | ATMCI_UNRE)
>  #define ATMCI_DMA_THRESHOLD	16
> +#define ATMCI_PDC_MIN_DATASIZE	16
>  
>  enum {
>  	EVENT_CMD_RDY = 0,
> @@ -765,7 +766,10 @@ static void atmci_pdc_set_single_buf(struct atmel_mci *host,
>  	}
>  
>  	if (host->data_size <= buf_size) {
> -		if (host->data_size & 0x3) {
> +		if ((host->data_size < ATMCI_PDC_MIN_DATASIZE) &&
> +		    (dir == XFER_TRANSMIT)) {
> +			atmci_writel(host, counter_reg, 4);
> +		} else if (host->data_size & 0x3) {
>  			/* If size is different from modulo 4, transfer bytes */
>  			atmci_writel(host, counter_reg, host->data_size);
>  			atmci_writel(host, ATMCI_MR, host->mode_reg | ATMCI_MR_PDCFBYTE);
> -- 
> 1.7.9.5
> 

Re: [PATCH] atmel-mci: fix transfers < 16 bytes problem in PDC mode

[Ludovic Desroches]

Hi Olivier,

On Thu, Dec 05, 2013 at 10:45:16AM +0100, Olivier Sobrie wrote:

[...]

> With this patch I'm able to use the wlan sdio module wl1271.
> When you've time, it would be nice to validate or fix this
> patch.

I have tested it and it doesn't solve issue I have with Atheros module.
 
> Thanks for your help,
> 
> Olivier
> 
>  drivers/mmc/host/atmel-mci.c |    6 +++++-
>  1 file changed, 5 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/mmc/host/atmel-mci.c b/drivers/mmc/host/atmel-mci.c
> index b86b482..c599731 100644
> --- a/drivers/mmc/host/atmel-mci.c
> +++ b/drivers/mmc/host/atmel-mci.c
> @@ -45,6 +45,7 @@
>  
>  #define ATMCI_DATA_ERROR_FLAGS	(ATMCI_DCRCE | ATMCI_DTOE | ATMCI_OVRE | ATMCI_UNRE)
>  #define ATMCI_DMA_THRESHOLD	16
> +#define ATMCI_PDC_MIN_DATASIZE	16
>  
>  enum {
>  	EVENT_CMD_RDY = 0,
> @@ -765,7 +766,10 @@ static void atmci_pdc_set_single_buf(struct atmel_mci *host,
>  	}
>  
>  	if (host->data_size <= buf_size) {
> -		if (host->data_size & 0x3) {
> +		if ((host->data_size < ATMCI_PDC_MIN_DATASIZE) &&
> +		    (dir == XFER_TRANSMIT)) {
> +			atmci_writel(host, counter_reg, 4);
> +		} else if (host->data_size & 0x3) {

I have feedback from people who have issue with atmel-mci driver but not
with at91-mci which has been removed unfortunately (it seems I had the
only sdio wifi module working with atmel-mci driver in pdc mode...). So I
think the 12 bytes value in the errata is correct.

Moreover I am afraid that this patch could cause some other bugs. I mean the
transfer size is 4 bytes, the sg length is 4 bytes too and we will ask
PDC to perform a bigger transfer... With the previous driver it was not a
problem because we didn't use sg directly, they were copied into another
buffer.

Regards

Ludovic

>  			/* If size is different from modulo 4, transfer bytes */
>  			atmci_writel(host, counter_reg, host->data_size);
>  			atmci_writel(host, ATMCI_MR, host->mode_reg | ATMCI_MR_PDCFBYTE);
> -- 
> 1.7.9.5
> 

Re: [PATCH] atmel-mci: fix transfers < 16 bytes problem in PDC mode

[Olivier Sobrie]

Hi Ludovic,

On Fri, Dec 13, 2013 at 12:09:13PM +0100, Ludovic Desroches wrote:
> Hi Olivier,
> 
> On Thu, Dec 05, 2013 at 10:45:16AM +0100, Olivier Sobrie wrote:
> 
> [...]
> 
> > With this patch I'm able to use the wlan sdio module wl1271.
> > When you've time, it would be nice to validate or fix this
> > patch.
> 
> I have tested it and it doesn't solve issue I have with Atheros module.

What kind of errors do you see with this module?
Timeouts on some SDIO commands with small datasize?

>  
> > Thanks for your help,
> > 
> > Olivier
> > 
> >  drivers/mmc/host/atmel-mci.c |    6 +++++-
> >  1 file changed, 5 insertions(+), 1 deletion(-)
> > 
> > diff --git a/drivers/mmc/host/atmel-mci.c b/drivers/mmc/host/atmel-mci.c
> > index b86b482..c599731 100644
> > --- a/drivers/mmc/host/atmel-mci.c
> > +++ b/drivers/mmc/host/atmel-mci.c
> > @@ -45,6 +45,7 @@
> >  
> >  #define ATMCI_DATA_ERROR_FLAGS	(ATMCI_DCRCE | ATMCI_DTOE | ATMCI_OVRE | ATMCI_UNRE)
> >  #define ATMCI_DMA_THRESHOLD	16
> > +#define ATMCI_PDC_MIN_DATASIZE	16
> >  
> >  enum {
> >  	EVENT_CMD_RDY = 0,
> > @@ -765,7 +766,10 @@ static void atmci_pdc_set_single_buf(struct atmel_mci *host,
> >  	}
> >  
> >  	if (host->data_size <= buf_size) {
> > -		if (host->data_size & 0x3) {
> > +		if ((host->data_size < ATMCI_PDC_MIN_DATASIZE) &&
> > +		    (dir == XFER_TRANSMIT)) {
> > +			atmci_writel(host, counter_reg, 4);
> > +		} else if (host->data_size & 0x3) {
> 
> I have feedback from people who have issue with atmel-mci driver but not
> with at91-mci which has been removed unfortunately (it seems I had the
> only sdio wifi module working with atmel-mci driver in pdc mode...). So I
> think the 12 bytes value in the errata is correct.

Does the atheros module work with the at91-mci driver ?

> 
> Moreover I am afraid that this patch could cause some other bugs. I mean the
> transfer size is 4 bytes, the sg length is 4 bytes too and we will ask
> PDC to perform a bigger transfer... With the previous driver it was not a
> problem because we didn't use sg directly, they were copied into another
> buffer.

Ok. I'll have a look at the old at91-mci driver.

Thanks,

-- 
Olivier

Re: [PATCH] atmel-mci: fix transfers < 16 bytes problem in PDC mode

[Ludovic Desroches]

Hi Olivier,

On Wed, Dec 18, 2013 at 09:09:41AM +0100, Olivier Sobrie wrote:
> Hi Ludovic,
> 
> On Fri, Dec 13, 2013 at 12:09:13PM +0100, Ludovic Desroches wrote:
> > Hi Olivier,
> > 
> > On Thu, Dec 05, 2013 at 10:45:16AM +0100, Olivier Sobrie wrote:
> > 
> > [...]
> > 
> > > With this patch I'm able to use the wlan sdio module wl1271.
> > > When you've time, it would be nice to validate or fix this
> > > patch.
> > 
> > I have tested it and it doesn't solve issue I have with Atheros module.
> 
> What kind of errors do you see with this module?
> Timeouts on some SDIO commands with small datasize?

Yes so same kind of errors than yours but your patch doesn't solve the
issue. The one I have done based on yours is not better in my case...

> 
> >  
> > > Thanks for your help,
> > > 
> > > Olivier
> > > 
> > >  drivers/mmc/host/atmel-mci.c |    6 +++++-
> > >  1 file changed, 5 insertions(+), 1 deletion(-)
> > > 
> > > diff --git a/drivers/mmc/host/atmel-mci.c b/drivers/mmc/host/atmel-mci.c
> > > index b86b482..c599731 100644
> > > --- a/drivers/mmc/host/atmel-mci.c
> > > +++ b/drivers/mmc/host/atmel-mci.c
> > > @@ -45,6 +45,7 @@
> > >  
> > >  #define ATMCI_DATA_ERROR_FLAGS	(ATMCI_DCRCE | ATMCI_DTOE | ATMCI_OVRE | ATMCI_UNRE)
> > >  #define ATMCI_DMA_THRESHOLD	16
> > > +#define ATMCI_PDC_MIN_DATASIZE	16
> > >  
> > >  enum {
> > >  	EVENT_CMD_RDY = 0,
> > > @@ -765,7 +766,10 @@ static void atmci_pdc_set_single_buf(struct atmel_mci *host,
> > >  	}
> > >  
> > >  	if (host->data_size <= buf_size) {
> > > -		if (host->data_size & 0x3) {
> > > +		if ((host->data_size < ATMCI_PDC_MIN_DATASIZE) &&
> > > +		    (dir == XFER_TRANSMIT)) {
> > > +			atmci_writel(host, counter_reg, 4);
> > > +		} else if (host->data_size & 0x3) {
> > 
> > I have feedback from people who have issue with atmel-mci driver but not
> > with at91-mci which has been removed unfortunately (it seems I had the
> > only sdio wifi module working with atmel-mci driver in pdc mode...). So I
> > think the 12 bytes value in the errata is correct.
> 
> Does the atheros module work with the at91-mci driver ?

Yes. State machine is not exactly the same. I thought I have found one
compatible with all mci versions but it seems it's not the case with
some sdio modules.

> 
> > 
> > Moreover I am afraid that this patch could cause some other bugs. I mean the
> > transfer size is 4 bytes, the sg length is 4 bytes too and we will ask
> > PDC to perform a bigger transfer... With the previous driver it was not a
> > problem because we didn't use sg directly, they were copied into another
> > buffer.
> 
> Ok. I'll have a look at the old at91-mci driver.

If you can, use it. I have tried to find small workaround but it seems
it's not possible. I'll have to rework state machine so it will be a bit
long.

Regards

Ludovic