sc8280xp: Add DDR clock #18
Conversation
Signed-off-by: Konrad Dybcio <konrad.dybcio@linaro.org>
| { "disp1_cc_mdss_vsync_clk", &gcc, 0x82, &disp1_cc, 0x17 }, | ||
| { "disp1_cc_sleep_clk", &gcc, 0x82, &disp1_cc, 0x46 }, | ||
| { "disp1_cc_xo_clk", &gcc, 0x82, &disp1_cc, 0x45 }, | ||
| { "measure_only_mccc_clk", &gcc, 0xfeedbeef, &mc_cc, 0x50 }, |
There was a problem hiding this comment.
So far downstream always defined a mux value for gcc and the current code even configures this 0xfeedbeef "mux value" to gcc before calling the custom measure() function:
Are you sure this is correct, since you gave absolutely zero context in the commit message whatsoever?
There was a problem hiding this comment.
I guess @andersson is indifferent about this too 😉
There was a problem hiding this comment.
@MarijnS95 @konradybcio @andersson
I see that all the downstream debugcc drivers also program some logical value for the mc_cc into gcc's debug mux. Do we know the value that should be used for sc8280xp?
There was a problem hiding this comment.
In other words, I see why it is not required now (because for reading mccc we bypass the gcc's debug mux). But there still should be some more correct value, but without the downstream driver or docs we can not deduce it.
There was a problem hiding this comment.
In this case I'd rather see a PR that allows setting measure_clk::primary to NULL to communicate this, and just set it to "measure_only_mccc_clk", NULL, 0, ... everywhere.
There was a problem hiding this comment.
@MarijnS95 The patch should be quite easy
debugcc.txt
However I suppose that we should fix it instead. @andersson do you possibly have a more correct value?
There was a problem hiding this comment.
I never said it would be hard 😉
However I suppose that we should fix it instead. @andersson do you possibly have a more correct value?
That also works, but it's weird if this value is "not needed" (and GCC shouldn't / doesn't need to be touched at all).
4 participants
No description provided.