lkml.org 
[lkml]   [2023]   [Jan]   [26]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
SubjectRe: [PATCH v4 0/6] Add MSM8939 SoC support with two devices
From


On 26.01.2023 16:29, Bryan O'Donoghue wrote:
> On 23/01/2023 12:49, Stephan Gerhold wrote:
>>> - Adds gcc dsi1pll and dsi1pllbyte to gcc clock list.
>>>    Reviewing the silicon documentation we see dsi0_phy_pll is used to clock
>>>    GCC_BYTE1_CFG_RCGR : SRC_SEL
>>>    Root Source Select
>>>    000 : cxo
>>>    001 : dsi0_phy_pll_out_byteclk
>>>    010 : GPLL0_OUT_AUX
>>>    011 : gnd
>>>    100 : gnd
>>>    101 : gnd
>>>    110 : gnd
>>>    111 : reserved - Stephan/Bryan
>>>
>> I'm confused. Are you not contradicting yourself here? You say that
>> dsi0_phy_pll (dsi ZERO) is used to clock GCC_BYTE1_CFG_RCGR. Then why
>> do you add dsi1_phy_pll (dsi ONE) to the gcc clock list?
>
> So my understanding of the clock tree here is that dsi0_phy_pll_out_byteclk is a legacy name.
>
> Its perfectly possible to have DSI0 and DSI0_PHY switched off and to have DSI1/DSI1_PHY operable.
>
> dsi0_phy_pll_out_byteclk is perhaps an unfortunate name and probably should have been renamed.
>
>> To me this looks like a confirmation of what downstream does, that both
>> DSI byte clocks are actually sourced from the dsi0_phy and the PLL of
>
> A better name would have been dsiX_phy_pll_out_byteclk.
I believe Stephan is just confused what the clock source of both
pairs of GCC DSI clocks are, as you're suggesting that:

phy_clock0
|_gcc_clock0

and

phy_clock0 (yes, zero)
|_gcc_clock1

whereas on most other SoCs the following is true:

phy_clock0
|_gcc_clock0

phy_clock1
|_gcc_clock_1

Konrad
>
> ---
> bod

\
 
 \ /
  Last update: 2023-03-26 23:57    [W:0.470 / U:0.200 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site