| Message ID | 20180713152017.2207-1-alexander.sverdlin@nokia.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
On Friday 13 July 2018 08:50 PM, Alexander Sverdlin wrote: > If CLKH is set to 0 I2C clock is not generated at all, so avoid this value > and stretch the clock in this case. This must be by observation since specification does not seem to state this. > > Signed-off-by: Alexander Sverdlin <alexander.sverdlin@nokia.com> In any case, the patch is fine. Acked-by: Sekhar Nori <nsekhar@ti.com> Thanks, Sekhar > --- > drivers/i2c/busses/i2c-davinci.c | 8 ++++++-- > 1 file changed, 6 insertions(+), 2 deletions(-) > > diff --git a/drivers/i2c/busses/i2c-davinci.c b/drivers/i2c/busses/i2c-davinci.c > index 75d6ab177055..7379043711df 100644 > --- a/drivers/i2c/busses/i2c-davinci.c > +++ b/drivers/i2c/busses/i2c-davinci.c > @@ -237,12 +237,16 @@ static void i2c_davinci_calc_clk_dividers(struct davinci_i2c_dev *dev) > /* > * It's not always possible to have 1 to 2 ratio when d=7, so fall back > * to minimal possible clkh in this case. > + * > + * Note: > + * CLKH is not allowed to be 0, in this case I2C clock is not generated > + * at all > */ > - if (clk >= clkl + d) { > + if (clk > clkl + d) { > clkh = clk - clkl - d; > clkl -= d; > } else { > - clkh = 0; > + clkh = 1; > clkl = clk - (d << 1); > } > > -- > 2.18.0 >
On 07/17/2018 06:57 AM, Sekhar Nori wrote: > On Friday 13 July 2018 08:50 PM, Alexander Sverdlin wrote: >> If CLKH is set to 0 I2C clock is not generated at all, so avoid this value >> and stretch the clock in this case. > > This must be by observation since specification does not seem to state > this. It will be good to know (mention) on what hw was it observed? > >> >> Signed-off-by: Alexander Sverdlin <alexander.sverdlin@nokia.com> > > In any case, the patch is fine. > > Acked-by: Sekhar Nori <nsekhar@ti.com> > [...]
On Fri, Jul 20, 2018 at 01:43:46PM -0500, Grygorii Strashko wrote: > > > On 07/17/2018 06:57 AM, Sekhar Nori wrote: > > On Friday 13 July 2018 08:50 PM, Alexander Sverdlin wrote: > > > If CLKH is set to 0 I2C clock is not generated at all, so avoid this value > > > and stretch the clock in this case. > > > > This must be by observation since specification does not seem to state > > this. > > It will be good to know (mention) on what hw was it observed? Some more info would also help me to decide if this is stable material or not. Is it a bug fix or more theoretical (but still valid, of course)?
On Fri, Jul 13, 2018 at 05:20:17PM +0200, Alexander Sverdlin wrote: > If CLKH is set to 0 I2C clock is not generated at all, so avoid this value > and stretch the clock in this case. > > Signed-off-by: Alexander Sverdlin <alexander.sverdlin@nokia.com> Applied to for-current, thanks! I did not add stable because Alexander told me this is very likely not to be observed on HW out there. But TI people are investigating more. I suggest they resend this patch to stable if they see fit. D'accord everyone?
On Monday 23 July 2018 11:37 PM, Wolfram Sang wrote: > On Fri, Jul 13, 2018 at 05:20:17PM +0200, Alexander Sverdlin wrote: >> If CLKH is set to 0 I2C clock is not generated at all, so avoid this value >> and stretch the clock in this case. >> >> Signed-off-by: Alexander Sverdlin <alexander.sverdlin@nokia.com> > > Applied to for-current, thanks! > > I did not add stable because Alexander told me this is very likely not > to be observed on HW out there. But TI people are investigating more. > I suggest they resend this patch to stable if they see fit. D'accord > everyone? Sounds good to me. Thanks, Sekhar
diff --git a/drivers/i2c/busses/i2c-davinci.c b/drivers/i2c/busses/i2c-davinci.c index 75d6ab177055..7379043711df 100644 --- a/drivers/i2c/busses/i2c-davinci.c +++ b/drivers/i2c/busses/i2c-davinci.c @@ -237,12 +237,16 @@ static void i2c_davinci_calc_clk_dividers(struct davinci_i2c_dev *dev) /* * It's not always possible to have 1 to 2 ratio when d=7, so fall back * to minimal possible clkh in this case. + * + * Note: + * CLKH is not allowed to be 0, in this case I2C clock is not generated + * at all */ - if (clk >= clkl + d) { + if (clk > clkl + d) { clkh = clk - clkl - d; clkl -= d; } else { - clkh = 0; + clkh = 1; clkl = clk - (d << 1); }
If CLKH is set to 0 I2C clock is not generated at all, so avoid this value and stretch the clock in this case. Signed-off-by: Alexander Sverdlin <alexander.sverdlin@nokia.com> --- drivers/i2c/busses/i2c-davinci.c | 8 ++++++-- 1 file changed, 6 insertions(+), 2 deletions(-)