[U-Boot] [PATCH v2 09/14] mx5 clocks: Fix get_ipg_per_clk()

[Benoît Thébaudeau]

Hi Stefano,

On Friday, September 28, 2012 12:45:31 PM, Stefano Babic wrote:
> On 28/09/2012 12:42, Benoît Thébaudeau wrote:
> > Hi Stefano,
> > 
> > On Friday, September 28, 2012 11:31:11 AM, Stefano Babic wrote:
> >> On 27/09/2012 22:23, Benoît Thébaudeau wrote:
> >>> This fixes the "IPG PERCLK" frequency printed by the clocks
> >>> command. The issue
> >>> was that get_ipg_per_clk() used periph_clk instead of lp_apm in
> >>> the
> >>> case
> >>> CCM.CBCMR.perclk_lp_apm_sel is set.
> >>>
> >>> It also fixes I²C support.
> >>>
> >>
> >> Hi Benoît,
> >>
> >> I understand "clocks" reports a wrong value only if
> >> CCM.CBCMR.perclk_lp_apm_sel is set, not always.
> > 
> > Correct.
> > 
> >> Can you better explain me which is wrong and which is the fix for
> >> I2C
> >> ?
> >> It seems unrelated, and I do not get the reason checking the patch
> > 
> > It's only because mxc_get_clock(MXC_IPG_CLK) (or
> > mxc_get_clock(MXC_I2C_CLK)
> > after [1]) both return get_ipg_per_clk() for the I²C driver's
> > clock.
> 
> Ah, thanks - it is clear now.

For the full story, I²C was completely broken on my custom i.MX51 platform
because of this issue. The command "i2c probe" only detected address 0x00, and
if run again, it hung.

Best regards,
Benoît