Linux SPI slave support

Linux SPI slave support

[Marek Vasut]

Hello,

is there currently any ongoing effort to support SPI slave mode in Linux?

If there is none, could you please share any hints that might help me better 
choose and go in the direction of implementing the SPI slave support?

I had this quick idea of simply extending the flags field of struct spi_master 
{} by something like SPI_MASTER_CAN_BE_SWITCHED_TO_SLAVE_MODE (shortened of 
course). Then there'd have to be a function present that can switch the mode, 
also in struct spi_master it seems. And then, any driver willing to use the 
slave support will simply call this switch function. Something amongst those 
lines maybe? Basically, the idea would be to have SPI slave mode drivers and 
master mode drivers using the same API, just make the bus configured 
differently.

Please advise. Thanks in advance!

Best regards,
Marek Vasut

------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and 
threat landscape has changed and how IT managers can respond. Discussions 
will include endpoint security, mobile security and the latest in malware 
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/

Re: Linux SPI slave support

[Shubhrajyoti Datta]

Hello,

On Sat, Jun 30, 2012 at 8:45 AM, Marek Vasut <marex-ynQEQJNshbs@public.gmane.org> wrote:
> Hello,
>
> is there currently any ongoing effort to support SPI slave mode in Linux?
>
> If there is none, could you please share any hints that might help me better
> choose and go in the direction of implementing the SPI slave support?


May be you may want to check the archives.
You may want to see below before designing.

http://permalink.gmane.org/gmane.linux.kernel.spi.devel/7495

>
> I had this quick idea of simply extending the flags field of struct spi_master
> {} by something like SPI_MASTER_CAN_BE_SWITCHED_TO_SLAVE_MODE (shortened of
> course). Then there'd have to be a function present that can switch the mode,
> also in struct spi_master it seems. And then, any driver willing to use the
> slave support will simply call this switch function. Something amongst those
> lines maybe? Basically, the idea would be to have SPI slave mode drivers and
> master mode drivers using the same API, just make the bus configured
> differently.
>
> Please advise. Thanks in advance!
>
> Best regards,
> Marek Vasut
>
> ------------------------------------------------------------------------------
> Live Security Virtual Conference
> Exclusive live event will cover all the ways today's security and
> threat landscape has changed and how IT managers can respond. Discussions
> will include endpoint security, mobile security and the latest in malware
> threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
> _______________________________________________
> spi-devel-general mailing list
> spi-devel-general-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f@public.gmane.org
> https://lists.sourceforge.net/lists/listinfo/spi-devel-general

------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and 
threat landscape has changed and how IT managers can respond. Discussions 
will include endpoint security, mobile security and the latest in malware 
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/

Re: Linux SPI slave support

[Marek Vasut]

Dear Shubhrajyoti Datta,

> Hello,
> 
> On Sat, Jun 30, 2012 at 8:45 AM, Marek Vasut <marex-ynQEQJNshbs@public.gmane.org> wrote:
> > Hello,
> > 
> > is there currently any ongoing effort to support SPI slave mode in Linux?
> > 
> > If there is none, could you please share any hints that might help me
> > better choose and go in the direction of implementing the SPI slave
> > support?
> 
> May be you may want to check the archives.
> You may want to see below before designing.
> 
> http://permalink.gmane.org/gmane.linux.kernel.spi.devel/7495

Oh, didn't find this one. Thanks a lot!

[...]

Best regards,
Marek Vasut

------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and 
threat landscape has changed and how IT managers can respond. Discussions 
will include endpoint security, mobile security and the latest in malware 
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/

Re: Linux SPI slave support

[Marek Vasut]

Dear Shubhrajyoti Datta,

> Hello,
> 
> On Sat, Jun 30, 2012 at 8:45 AM, Marek Vasut <marex-ynQEQJNshbs@public.gmane.org> wrote:
> > Hello,
> > 
> > is there currently any ongoing effort to support SPI slave mode in Linux?
> > 
> > If there is none, could you please share any hints that might help me
> > better choose and go in the direction of implementing the SPI slave
> > support?
> 
> May be you may want to check the archives.
> You may want to see below before designing.
> 
> http://permalink.gmane.org/gmane.linux.kernel.spi.devel/7495

I also noticed this stuff below, the intel moorestown SPI slave work. Did it get 
anywhere?

http://www.mail-archive.com/spi-devel-
general-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f@public.gmane.org/msg03514.html

Best regards,
Marek Vasut

------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and 
threat landscape has changed and how IT managers can respond. Discussions 
will include endpoint security, mobile security and the latest in malware 
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/

Re: Linux SPI slave support

[Marek Vasut]

Dear Shubhrajyoti Datta,

> Hello,
> 
> On Sat, Jun 30, 2012 at 8:45 AM, Marek Vasut <marex-ynQEQJNshbs@public.gmane.org> wrote:
> > Hello,
> > 
> > is there currently any ongoing effort to support SPI slave mode in Linux?
> > 
> > If there is none, could you please share any hints that might help me
> > better choose and go in the direction of implementing the SPI slave
> > support?
> 
> May be you may want to check the archives.
> You may want to see below before designing.
> 
> http://permalink.gmane.org/gmane.linux.kernel.spi.devel/7495
[...]

So, idea V2:

1) Separate out common parts from struct spi_master
2) Create struct spi_slave
2a) Implement spi_slave_send_message for spi_slave
2b) Implement <?> netlink interface to push and pull messages to/from userland ?

The protocol won't be done in kernel, but completely in userland.

Is this good approach to take? Or maybe 2b can be replaced by blocking read on 
character device? I think netlink interface might actually be suited quite well 
for this task.

Best regards,
Marek Vasut

------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and 
threat landscape has changed and how IT managers can respond. Discussions 
will include endpoint security, mobile security and the latest in malware 
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/

Re: Linux SPI slave support

[Shubhrajyoti Datta]

On Mon, Jul 16, 2012 at 5:56 AM, Marek Vasut <marex-ynQEQJNshbs@public.gmane.org> wrote:
> Dear Shubhrajyoti Datta,
>
>> Hello,
>>
>> On Sat, Jun 30, 2012 at 8:45 AM, Marek Vasut <marex-ynQEQJNshbs@public.gmane.org> wrote:
>> > Hello,
>> >
>> > is there currently any ongoing effort to support SPI slave mode in Linux?
>> >
>> > If there is none, could you please share any hints that might help me
>> > better choose and go in the direction of implementing the SPI slave
>> > support?
>>
>> May be you may want to check the archives.
>> You may want to see below before designing.
>>
>> http://permalink.gmane.org/gmane.linux.kernel.spi.devel/7495
> [...]
>
> So, idea V2:
>
> 1) Separate out common parts from struct spi_master
> 2) Create struct spi_slave
> 2a) Implement spi_slave_send_message for spi_slave
> 2b) Implement <?> netlink interface to push and pull messages to/from userland ?
>
> The protocol won't be done in kernel, but completely in userland.
>
> Is this good approach to take? Or maybe 2b can be replaced by blocking read on
> character device? I think netlink interface might actually be suited quite well
> for this task.


The other option is to have something like a gpio  for synchronization.



>
> Best regards,
> Marek Vasut

------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and 
threat landscape has changed and how IT managers can respond. Discussions 
will include endpoint security, mobile security and the latest in malware 
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/

Re: Linux SPI slave support

[Marek Vasut]

Dear Shubhrajyoti Datta,

> On Mon, Jul 16, 2012 at 5:56 AM, Marek Vasut <marex-ynQEQJNshbs@public.gmane.org> wrote:
> > Dear Shubhrajyoti Datta,
> > 
> >> Hello,
> >> 
> >> On Sat, Jun 30, 2012 at 8:45 AM, Marek Vasut <marex-ynQEQJNshbs@public.gmane.org> wrote:
> >> > Hello,
> >> > 
> >> > is there currently any ongoing effort to support SPI slave mode in
> >> > Linux?
> >> > 
> >> > If there is none, could you please share any hints that might help me
> >> > better choose and go in the direction of implementing the SPI slave
> >> > support?
> >> 
> >> May be you may want to check the archives.
> >> You may want to see below before designing.
> >> 
> >> http://permalink.gmane.org/gmane.linux.kernel.spi.devel/7495
> > 
> > [...]
> > 
> > So, idea V2:
> > 
> > 1) Separate out common parts from struct spi_master
> > 2) Create struct spi_slave
> > 2a) Implement spi_slave_send_message for spi_slave
> > 2b) Implement <?> netlink interface to push and pull messages to/from
> > userland ?
> > 
> > The protocol won't be done in kernel, but completely in userland.
> > 
> > Is this good approach to take? Or maybe 2b can be replaced by blocking
> > read on character device? I think netlink interface might actually be
> > suited quite well for this task.
> 
> The other option is to have something like a gpio  for synchronization.

Synchronisation of what exactly? With my hardware (that doesn't exist yet I 
guess), the idea is to simply cross-connect the MISO/MOSI pins and connect the 
CLK signal. Is that possible btw?

Thanks!

> > Best regards,
> > Marek Vasut

Best regards,
Marek Vasut

------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and 
threat landscape has changed and how IT managers can respond. Discussions 
will include endpoint security, mobile security and the latest in malware 
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/

Re: Linux SPI slave support

[Shubhrajyoti Datta]

On Tue, Jul 17, 2012 at 2:22 PM, Marek Vasut <marex-ynQEQJNshbs@public.gmane.org> wrote:
> Dear Shubhrajyoti Datta,
>
>> On Mon, Jul 16, 2012 at 5:56 AM, Marek Vasut <marex-ynQEQJNshbs@public.gmane.org> wrote:
>> > Dear Shubhrajyoti Datta,
[..]
>> The other option is to have something like a gpio  for synchronization.
>
> Synchronisation of what exactly? With my hardware (that doesn't exist yet I
> guess), the idea is to simply cross-connect the MISO/MOSI pins and connect the
> CLK signal. Is that possible btw?

IIUC , the clock is to be driven by master. The slave gets the clock.
Master knows when to start and stop the transaction.

However for the slave the options are to
1) Keep listening always.
2) Use something to syncronize( I may have used the word loosely)
master and slave.
So that it doesnt have to listen always.

Btw it is just a suggestion. I am not opposed to any other implementation.

Thanks and Regards,

>
> Thanks!
>

------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and 
threat landscape has changed and how IT managers can respond. Discussions 
will include endpoint security, mobile security and the latest in malware 
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/