Thread (1 message) 1 message, 1 author, 2010-02-16

Re: [PATCH/RFC 1/2] 5200: improve i2c bus error recovery

From: Albrecht Dre� <hidden>
Date: 2010-02-16 20:14:00
Also in: linux-devicetree

Hi Ira:

[snip]
I see this exact hang on a MPC8349EA board. I poll my i2c sensors every
500ms, and it takes around 12 hours to produce a hang. The usual hang
has (CF | BB) set, however I have seen a hang with just BB (only once so
far in about 2 weeks).
=20
I think the fixup should be run on 8349 as well, if not all processors.
I'm happy to test patches. I have a way to reliably trigger a lockup,
using another master on the i2c bus.
See my other post - hopefully a new patch is ready tomorrow/Thursday...

[snip]
quoted
 static void mpc_i2c_fixup(struct mpc_i2c *i2c)
 {
-=09writeccr(i2c, 0);
-=09udelay(30);
-=09writeccr(i2c, CCR_MEN);
-=09udelay(30);
-=09writeccr(i2c, CCR_MSTA | CCR_MTX);
-=09udelay(30);
-=09writeccr(i2c, CCR_MSTA | CCR_MTX | CCR_MEN);
-=09udelay(30);
-=09writeccr(i2c, CCR_MEN);
-=09udelay(30);
+=09if (i2c->real_clk =3D=3D 0) {
+=09=09writeccr(i2c, 0);
+=09=09udelay(30);
+=09=09writeccr(i2c, CCR_MEN);
+=09=09udelay(30);
+=09=09writeccr(i2c, CCR_MSTA | CCR_MTX);
+=09=09udelay(30);
+=09=09writeccr(i2c, CCR_MSTA | CCR_MTX | CCR_MEN);
+=09=09udelay(30);
+=09=09writeccr(i2c, CCR_MEN);
+=09=09udelay(30);
+=09} else {
+=09=09int k;
+=09=09u32 delay_val =3D 1000000 / i2c->real_clk + 1;
+
+=09=09if (delay_val < 2)
+=09=09=09delay_val =3D 2;
+
+=09=09for (k =3D 9; k; k--) {
+=09=09=09writeccr(i2c, 0);
+=09=09=09writeccr(i2c, CCR_MSTA | CCR_MTX | CCR_MEN);
+=09=09=09udelay(delay_val);
+=09=09=09writeccr(i2c, CCR_MEN);
+=09=09=09udelay(delay_val << 1);
+=09=09}
+=09}
 }
=20
=20
The old sequence has always un-hung the bus for me. Yours might be
better. I'll try it next time the bus wedges up on me.
The easiest way to verify the /waveforms/ is to simply call the function so=
mewhere else (e.g. in a i2c write call), and then record SCK and SDA with a=
 scope.  I saw *huge* unnecessary delays in the original code, and by some =
creative interpretation of the (not very clear) 5200 data sheet I came to t=
he write's and delays above.  If you have a scope, you might want to repeat=
 the test with your '8349.

Best, Albrecht.

Immer auf dem Laufenden! Sport, Auto, Reise, Politik und Promis. Von uns f=
=FCr Sie: der neue Arcor.de-Newsletter!
Jetzt anmelden und einfach alles wissen: http://www.arcor.de/rd/footer.news=
letter
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help