]> git.ipfire.org Git - thirdparty/kernel/stable.git/commitdiff
clk: qcom: gcc-qcs404: fix initial rate of GPLL3
authorGabor Juhos <j4g8y7@gmail.com>
Tue, 22 Oct 2024 09:45:56 +0000 (11:45 +0200)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Sat, 14 Dec 2024 18:47:37 +0000 (19:47 +0100)
commit 36d202241d234fa4ac50743510d098ad52bd193a upstream.

The comment before the config of the GPLL3 PLL says that the
PLL should run at 930 MHz. In contrary to this, calculating
the frequency from the current configuration values by using
19.2 MHz as input frequency defined in 'qcs404.dtsi', it gives
921.6 MHz:

  $ xo=19200000; l=48; alpha=0x0; alpha_hi=0x0
  $ echo "$xo * ($((l)) + $(((alpha_hi << 32 | alpha) >> 8)) / 2^32)" | bc -l
  921600000.00000000000000000000

Set 'alpha_hi' in the configuration to a value used in downstream
kernels [1][2] in order to get the correct output rate:

  $ xo=19200000; l=48; alpha=0x0; alpha_hi=0x70
  $ echo "$xo * ($((l)) + $(((alpha_hi << 32 | alpha) >> 8)) / 2^32)" | bc -l
  930000000.00000000000000000000

The change is based on static code analysis, compile tested only.

[1] https://git.codelinaro.org/clo/la/kernel/msm-5.4/-/blob/kernel.lnx.5.4.r56-rel/drivers/clk/qcom/gcc-qcs404.c?ref_type=heads#L335
[2} https://git.codelinaro.org/clo/la/kernel/msm-5.15/-/blob/kernel.lnx.5.15.r49-rel/drivers/clk/qcom/gcc-qcs404.c?ref_type=heads#L127

Cc: stable@vger.kernel.org
Fixes: 652f1813c113 ("clk: qcom: gcc: Add global clock controller driver for QCS404")
Signed-off-by: Gabor Juhos <j4g8y7@gmail.com>
Link: https://lore.kernel.org/r/20241022-fix-gcc-qcs404-gpll3-v1-1-c4d30d634d19@gmail.com
Signed-off-by: Bjorn Andersson <andersson@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
drivers/clk/qcom/gcc-qcs404.c

index a7a9884799cd3c832fe2cc5e6c9630f72aa814d0..ef61c5801f1d2388190ff1e35813fe1812eeef62 100644 (file)
@@ -337,6 +337,7 @@ static struct clk_alpha_pll gpll1_out_main = {
 /* 930MHz configuration */
 static const struct alpha_pll_config gpll3_config = {
        .l = 48,
+       .alpha_hi = 0x70,
        .alpha = 0x0,
        .alpha_en_mask = BIT(24),
        .post_div_mask = 0xf << 8,