linux-kernel - Re: [PATCH] soc: qcom: Rework BCM_TCS

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <8bd63441-7414-4bc1-9044-b3f1e7b21384@linaro.org>
Date: Mon, 28 Oct 2024 14:22:38 +0200
From: Eugen Hristev <eugen.hristev@...aro.org>
To: Konrad Dybcio <konrad.dybcio@....qualcomm.com>
Cc: andersson@...nel.org, konradybcio@...nel.org,
 linux-kernel@...r.kernel.org, linux-arm-msm@...r.kernel.org
Subject: Re: [PATCH] soc: qcom: Rework BCM_TCS_CMD macro



On 10/25/24 20:03, Konrad Dybcio wrote:
> On 25.10.2024 2:06 PM, Eugen Hristev wrote:
>>
>>
>> On 10/25/24 12:03, Konrad Dybcio wrote:
>>> On 25.10.2024 10:48 AM, Eugen Hristev wrote:
>>>> Reworked BCM_TCS_CMD macro in order to fix warnings from sparse:
>>>>
>>>> drivers/clk/qcom/clk-rpmh.c:270:28: warning: restricted __le32 degrades to integer
>>>> drivers/clk/qcom/clk-rpmh.c:270:28: warning: restricted __le32 degrades to integer
>>>>
>>>> While at it, used u32_encode_bits which made the code easier to
>>>> follow and removed unnecessary shift definitions.
>>>>
>>>> Signed-off-by: Eugen Hristev <eugen.hristev@...aro.org>
>>>> ---
>>>>    include/soc/qcom/tcs.h | 19 ++++++++-----------
>>>>    1 file changed, 8 insertions(+), 11 deletions(-)
>>>>
>>>> diff --git a/include/soc/qcom/tcs.h b/include/soc/qcom/tcs.h
>>>> index 3acca067c72b..130ed2582f37 100644
>>>> --- a/include/soc/qcom/tcs.h
>>>> +++ b/include/soc/qcom/tcs.h
>>>> @@ -60,22 +60,19 @@ struct tcs_request {
>>>>        struct tcs_cmd *cmds;
>>>>    };
>>>>    -#define BCM_TCS_CMD_COMMIT_SHFT        30
>>>>    #define BCM_TCS_CMD_COMMIT_MASK        0x40000000
>>>> -#define BCM_TCS_CMD_VALID_SHFT        29
>>>>    #define BCM_TCS_CMD_VALID_MASK        0x20000000
>>>> -#define BCM_TCS_CMD_VOTE_X_SHFT        14
>>>>    #define BCM_TCS_CMD_VOTE_MASK        0x3fff
>>>> -#define BCM_TCS_CMD_VOTE_Y_SHFT        0
>>>> -#define BCM_TCS_CMD_VOTE_Y_MASK        0xfffc000
>>>> +#define BCM_TCS_CMD_VOTE_Y_MASK        0x3fff
>>>> +#define BCM_TCS_CMD_VOTE_X_MASK        0xfffc000
>>>>      /* Construct a Bus Clock Manager (BCM) specific TCS command */
>>>>    #define BCM_TCS_CMD(commit, valid, vote_x, vote_y)        \
>>>> -    (((commit) << BCM_TCS_CMD_COMMIT_SHFT) |        \
>>>> -    ((valid) << BCM_TCS_CMD_VALID_SHFT) |            \
>>>> -    ((cpu_to_le32(vote_x) &                    \
>>>> -    BCM_TCS_CMD_VOTE_MASK) << BCM_TCS_CMD_VOTE_X_SHFT) |    \
>>>> -    ((cpu_to_le32(vote_y) &                    \
>>>> -    BCM_TCS_CMD_VOTE_MASK) << BCM_TCS_CMD_VOTE_Y_SHFT))
>>>> +    (u32_encode_bits(commit, BCM_TCS_CMD_COMMIT_MASK) |    \
>>>> +    u32_encode_bits(valid, BCM_TCS_CMD_VALID_MASK) |    \
>>>> +    u32_encode_bits((__force u32)cpu_to_le32(vote_x),    \
>>>> +            BCM_TCS_CMD_VOTE_X_MASK) |        \
>>>> +    u32_encode_bits((__force u32)cpu_to_le32(vote_y),    \
>>>> +            BCM_TCS_CMD_VOTE_Y_MASK))
>>>
>>> FIELD_PREP/GET?
>>>
>>> Konrad
>>
>> What would be the difference/advantage in using FIELD_PREP/GET instead of u32_encode_bits ?
> 
> Probably none. I thought it was a function and not another magic macro,
> as it's lowercase..
> 
> Doesn't le32_encode_bits do what you need then?

That works too. It would move the `__force u32` to a higher level, but 
works. I am sending a v2 with that.

> 
> Konrad