Hi, Philipp
        The table you listed is VDDARM_IN and VDDSOC_IN which is for external 
PMIC and we normally do NOT adjust them when cpufreq change, we only adjust the 
internel LDO output, my table is for VDDARM_CAP and VDDSOC_CAP which are in the 
comment column, for LDO output.

For LDO enabled mode, VDDARM/SOC_CAP is the output of LDO, we pass this table 
to kernel opp.
For LDO bypassed mode, VDDARM/SOC_IN = VDDARM/SOC_CAP.
 
Best regards!
Anson Huang


-----Original Message-----
From: Philipp Zabel [mailto:p.za...@pengutronix.de] 
Sent: 2014-12-05 5:39 PM
To: Huang Yongcai-B20788
Cc: shawn....@linaro.org; ker...@pengutronix.de; devicetree@vger.kernel.org
Subject: Re: [PATCH 2/2] arm: dts: imx6dl: correct cpufreq volt/freq table

Am Freitag, den 05.12.2014, 16:23 +0800 schrieb Anson Huang:
> Currently the cpufreq volt/freq table we used is for LDO enable mode, 
> according to latest datasheet Rev. 3, 03/2014, the volt/freq table is 
> as below:
> 
> LDO enabled(min value):
> 996MHz: VDDARM: 1.225V, VDDSOC: 1.150V;
> 792MHz: VDDARM: 1.150V, VDDSOC: 1.150V;
> 396MHz: VDDARM: 1.050V, VDDSOC: 1.150V;

From which datasheet is this? I find it unlikely that the LDO enabled values 
are not larger than the LDO bypassed ones. I just had a look at the 
"i.MXSolo/6DualLite Automotive and Infotainment Applications Processors" 
Datasheet (IMX6SDLAEC) Rev. 3, 03/2014 and the "i.MX 6Solo/6DualLite 
Applications Processors for Consumer Products" Datasheet
(IMX6SDLCEC) Rev. 3, 03/2014, and they both have different values in Table 9 
(Operating Ranges) for the LDO enabled case:

LDO enabled(min value):
996MHz: VDDARM: 1.350V, VDDSOC: 1.275V;
792MHz: VDDARM: 1.275V, VDDSOC: 1.275V;
396MHz: VDDARM: 1.175V, VDDSOC: 1.275V;

> LDO bypassed(min value):
> 996MHz: VDDARM: 1.250V, VDDSOC: 1.150V;
> 792MHz: VDDARM: 1.150V, VDDSOC: 1.150V;
> 396MHz: VDDARM: 1.050V, VDDSOC: 1.150V;
> 
> Adding 25mV to cover board IR drop, for LDO enabled mode of 996MHz, 
> VDDARM should be 1.250V, so this patch updates it.
> 
> Signed-off-by: Anson Huang <b20...@freescale.com>

regards
Philipp

N�����r��y����b�X��ǧv�^�)޺{.n�+���z��z��z)����w*jg��������ݢj/���z�ޖ��2�ޙ����&�)ߡ�a�����G���h��j:+v���w��٥

Reply via email to