Re: [PATCH 04/13] clk: qcom: gpucc-sa8775p: Remove the CLK_IS_CRITICAL and ALWAYS_ON flags

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On 6/2/2024 8:58 PM, Krzysztof Kozlowski wrote:
On 02/06/2024 06:12, Bjorn Andersson wrote:
On Fri, May 31, 2024 at 11:59:04AM GMT, Krzysztof Kozlowski wrote:
On 31/05/2024 11:02, Taniya Das wrote:
The gpu clocks and GDSC have been marked critical from the clock driver
which is not desired for functionality. Hence remove the CLK_IS_CRITICAL
and ALWAYS_ON flags.

Fixes: 0afa16afc36d ("clk: qcom: add the GPUCC driver for sa8775p")
Signed-off-by: Taniya Das <quic_tdas@xxxxxxxxxxx>
---
  drivers/clk/qcom/gpucc-sa8775p.c | 27 +++++++++++----------------
  1 file changed, 11 insertions(+), 16 deletions(-)

diff --git a/drivers/clk/qcom/gpucc-sa8775p.c b/drivers/clk/qcom/gpucc-sa8775p.c
index 1167c42da39d..f965babf4330 100644
--- a/drivers/clk/qcom/gpucc-sa8775p.c
+++ b/drivers/clk/qcom/gpucc-sa8775p.c
@@ -1,6 +1,6 @@
  // SPDX-License-Identifier: GPL-2.0-only
  /*
- * Copyright (c) 2021-2022, Qualcomm Innovation Center, Inc. All rights reserved.
+ * Copyright (c) 2021-2022, 2024, Qualcomm Innovation Center, Inc. All rights reserved.

That's not a fix.

   * Copyright (c) 2023, Linaro Limited
   */
@@ -280,7 +280,7 @@ static struct clk_branch gpu_cc_ahb_clk = {
  				&gpu_cc_hub_ahb_div_clk_src.clkr.hw,
  			},
  			.num_parents = 1,
-			.flags = CLK_SET_RATE_PARENT | CLK_IS_CRITICAL,
+			.flags = CLK_SET_RATE_PARENT,

I fail to see why this is a fix. They were marked as critical on
purpose. It was needed, wasn't it?

Provide jsutification for commits, not just sprinkle Fixes tag all around.


This is indeed a fix, as marking clocks CLK_IS_CRITICAL prevents any
power-domain associated with the clock controller from suspending. In
other words, the current behavior is broken.

@Taniya, "not desired for functionality" does not carry any useful
information explaining why this change is made. Please update the commit
message.

Then please provide some sort of bug explanation in the commit msg. I
assume the clocks were marked critical to solve some particular problem,
e.g. missing parents, so marking this as fix sounds like both incorrect
and error-prone when backported. Maybe that's not the case, so this is
why there is commit msg to explain such details...


Thanks for your review. Sure, I will update the next series with the below details.

The GPU clocks/GDSCs have been marked critical from the clock driver
but the GPU driver votes on these resources as per the HW requirement.
We don't require these clocks and GDSC's to be kept always ON which would have power impact and also GPU stability/corruptions.
Hence remove the CLK_IS_CRITICAL and ALWAYS_ON flags.

But I am not sure why the original author left the clocks/GDSCs always ON.


Best regards,
Krzysztof


--
Thanks & Regards,
Taniya Das.




[Index of Archives]     [Linux ARM Kernel]     [Linux ARM]     [Linux Omap]     [Fedora ARM]     [Linux for Sparc]     [IETF Annouce]     [Security]     [Bugtraq]     [Linux MIPS]     [ECOS]     [Asterisk Internet PBX]     [Linux API]

  Powered by Linux