[PATCH] SPI bit bang: Deactivate CS before restoring speed & wordsize

[PATCH] SPI bit bang: Deactivate CS before restoring speed & wordsize

[Joakim Tjernlund]

Restoring speed & wordsize on mpc83xx disables the controller temporarily
which makes the SPI outputs floating. This can cause a glich on
the SPI CLK which makes the slave in the other end unhappy.

Make sure that CS is disabled before restoring speed & wordsize
so any gliches are ignored by the SPI slave.

Signed-off-by: Joakim Tjernlund <Joakim.Tjernlund-SNLAxHN9vbcOP4wsBPIw7w@public.gmane.org>
---

diff --git a/drivers/spi/spi_bitbang.c b/drivers/spi/spi_bitbang.c
index 0c85c98..53649cd 100644
--- a/drivers/spi/spi_bitbang.c
+++ b/drivers/spi/spi_bitbang.c
@@ -371,10 +371,6 @@ static void bitbang_work(struct work_struct *work)
 		m->status = status;
 		m->complete(m->context);
 
-		/* restore speed and wordsize */
-		if (setup_transfer)
-			setup_transfer(spi, NULL);
-
 		/* normally deactivate chipselect ... unless no error and
 		 * cs_change has hinted that the next message will probably
 		 * be for this chip too.
@@ -385,6 +381,10 @@ static void bitbang_work(struct work_struct *work)
 			ndelay(nsecs);
 		}
 
+		/* restore speed and wordsize */
+		if (setup_transfer)
+			setup_transfer(spi, NULL);
+
 		spin_lock_irqsave(&bitbang->lock, flags);
 	}
 	bitbang->busy = 0;


-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2008.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/

Re: [PATCH] SPI bit bang: Deactivate CS before restoring speed & wordsize

[David Brownell]

On Friday 07 March 2008, Joakim Tjernlund wrote:
> Restoring speed & wordsize on mpc83xx disables the controller temporarily
> which makes the SPI outputs floating. This can cause a glich on
> the SPI CLK which makes the slave in the other end unhappy.

This will also cause trouble with any driver using these protocol
tweaking options:

 - per-transfer overrides of speed and/or wordsize
 - leaving chip selected between messages

The mpc83xx driver is not overriding master->transfer, so all of
those are behaviors this driver claims to support...


> Make sure that CS is disabled before restoring speed & wordsize
> so any gliches are ignored by the SPI slave.

I think the correct fix hast to be in mpc83xx_spi_setup_transfer().
If that routine is fixed, the bitbang core won't need to change,
and those protocol tweaking options will just work.

Or else, this patch would need to be accompanied by a patch which
updates spi_mpc83xx to override master->transfer and filter out all
messages which use those protocol tweaking options, so that all the
places this glitch can appear will be addressed.

- Dave


-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2008.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/