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] [thread-next>] [day] [month] [year] [list]
Message-ID: <e6dca8e2-d550-4819-9833-ecbbd1a8f376@linaro.org>
Date: Thu, 10 Apr 2025 17:06:21 +0100
From: Bryan O'Donoghue <bryan.odonoghue@...aro.org>
To: neil.armstrong@...aro.org, Dikshita Agarwal <quic_dikshita@...cinc.com>,
 Vikash Garodia <quic_vgarodia@...cinc.com>,
 Abhinav Kumar <quic_abhinavk@...cinc.com>,
 Mauro Carvalho Chehab <mchehab@...nel.org>, Rob Herring <robh@...nel.org>,
 Krzysztof Kozlowski <krzk+dt@...nel.org>, Conor Dooley
 <conor+dt@...nel.org>, Philipp Zabel <p.zabel@...gutronix.de>
Cc: linux-media@...r.kernel.org, linux-arm-msm@...r.kernel.org,
 devicetree@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH v4 6/6] media: platform: qcom/iris: add sm8650 support

On 10/04/2025 16:40, neil.armstrong@...aro.org wrote:
> On 10/04/2025 15:07, Dikshita Agarwal wrote:
>>
>>
>> On 4/10/2025 2:43 PM, Vikash Garodia wrote:
>>>
>>> On 4/10/2025 2:31 PM, Neil Armstrong wrote:
>>>> On 09/04/2025 18:57, Vikash Garodia wrote:
>>>>> Hi Neil,
>>>>>
>>>>> On 4/9/2025 8:08 PM, Neil Armstrong wrote:
>>>>>> Add support for the SM8650 platform by re-using the SM8550
>>>>>> definitions and using the vpu33 ops.
>>>>>>
>>>>>> The SM8650/vpu33 requires more reset lines, but the H.264
>>>>>> decoder capabilities are identical.
>>>>>>
>>>>>> Tested-by: Bryan O'Donoghue <bryan.odonoghue@...aro.org> # x1e Dell
>>>>>> Signed-off-by: Neil Armstrong <neil.armstrong@...aro.org>
>>>>>> ---
>>>>>>    .../platform/qcom/iris/iris_platform_common.h      |  1 +
>>>>>>    .../platform/qcom/iris/iris_platform_sm8550.c      | 64 +++++++ 
>>>>>> +++++++++++++++
>>>>>>    drivers/media/platform/qcom/iris/iris_probe.c      |  4 ++
>>>>>>    3 files changed, 69 insertions(+)
>>>>>>
>>>>>> diff --git a/drivers/media/platform/qcom/iris/iris_platform_common.h
>>>>>> b/drivers/media/platform/qcom/iris/iris_platform_common.h
>>>>>> index
>>>>>> fdd40fd80178c4c66b37e392d07a0a62f492f108..6bc3a7975b04d612f6c89206eae95dac678695fc 100644
>>>>>> --- a/drivers/media/platform/qcom/iris/iris_platform_common.h
>>>>>> +++ b/drivers/media/platform/qcom/iris/iris_platform_common.h
>>>>>> @@ -35,6 +35,7 @@ enum pipe_type {
>>>>>>      extern struct iris_platform_data sm8250_data;
>>>>>>    extern struct iris_platform_data sm8550_data;
>>>>>> +extern struct iris_platform_data sm8650_data;
>>>>>>      enum platform_clk_type {
>>>>>>        IRIS_AXI_CLK,
>>>>>> diff --git a/drivers/media/platform/qcom/iris/iris_platform_sm8550.c
>>>>>> b/drivers/media/platform/qcom/iris/iris_platform_sm8550.c
>>>>>> index
>>>>>> 35d278996c430f2856d0fe59586930061a271c3e..d0f8fa960d53367023e41bc5807ba3f8beae2efc 100644
>>>>>> --- a/drivers/media/platform/qcom/iris/iris_platform_sm8550.c
>>>>>> +++ b/drivers/media/platform/qcom/iris/iris_platform_sm8550.c
>>>>>> @@ -144,6 +144,10 @@ static const struct icc_info 
>>>>>> sm8550_icc_table[] = {
>>>>>>      static const char * const sm8550_clk_reset_table[] = { "bus" };
>>>>>>    +static const char * const sm8650_clk_reset_table[] = { "bus", 
>>>>>> "core" };
>>>>>> +
>>>>>> +static const char * const sm8650_controller_reset_table[] = 
>>>>>> { "xo" };
>>>>>> +
>>>>>>    static const struct bw_info sm8550_bw_table_dec[] = {
>>>>>>        { ((4096 * 2160) / 256) * 60, 1608000 },
>>>>>>        { ((4096 * 2160) / 256) * 30,  826000 },
>>>>>> @@ -264,3 +268,63 @@ struct iris_platform_data sm8550_data = {
>>>>>>        .dec_op_int_buf_tbl = sm8550_dec_op_int_buf_tbl,
>>>>>>        .dec_op_int_buf_tbl_size = 
>>>>>> ARRAY_SIZE(sm8550_dec_op_int_buf_tbl),
>>>>>>    };
>>>>>> +
>>>>>> +/*
>>>>>> + * Shares most of SM8550 data except:
>>>>>> + * - vpu_ops to iris_vpu33_ops
>>>>>> + * - clk_rst_tbl to sm8650_clk_reset_table
>>>>>> + * - controller_rst_tbl to sm8650_controller_reset_table
>>>>>> + * - fwname to "qcom/vpu/vpu33_p4.mbn"
>>>>>> + */
>>>>>> +struct iris_platform_data sm8650_data = {
>>>>>> +    .get_instance = iris_hfi_gen2_get_instance,
>>>>>> +    .init_hfi_command_ops = iris_hfi_gen2_command_ops_init,
>>>>>> +    .init_hfi_response_ops = iris_hfi_gen2_response_ops_init,
>>>>>> +    .vpu_ops = &iris_vpu33_ops,
>>>>>> +    .set_preset_registers = iris_set_sm8550_preset_registers,
>>>>>> +    .icc_tbl = sm8550_icc_table,
>>>>>> +    .icc_tbl_size = ARRAY_SIZE(sm8550_icc_table),
>>>>>> +    .clk_rst_tbl = sm8650_clk_reset_table,
>>>>>> +    .clk_rst_tbl_size = ARRAY_SIZE(sm8650_clk_reset_table),
>>>>>> +    .controller_rst_tbl = sm8650_controller_reset_table,
>>>>>> +    .controller_rst_tbl_size = 
>>>>>> ARRAY_SIZE(sm8650_controller_reset_table),
>>>>>> +    .bw_tbl_dec = sm8550_bw_table_dec,
>>>>>> +    .bw_tbl_dec_size = ARRAY_SIZE(sm8550_bw_table_dec),
>>>>>> +    .pmdomain_tbl = sm8550_pmdomain_table,
>>>>>> +    .pmdomain_tbl_size = ARRAY_SIZE(sm8550_pmdomain_table),
>>>>>> +    .opp_pd_tbl = sm8550_opp_pd_table,
>>>>>> +    .opp_pd_tbl_size = ARRAY_SIZE(sm8550_opp_pd_table),
>>>>>> +    .clk_tbl = sm8550_clk_table,
>>>>>> +    .clk_tbl_size = ARRAY_SIZE(sm8550_clk_table),
>>>>>> +    /* Upper bound of DMA address range */
>>>>>> +    .dma_mask = 0xe0000000 - 1,
>>>>>> +    .fwname = "qcom/vpu/vpu33_p4.mbn",
>>>>>> +    .pas_id = IRIS_PAS_ID,
>>>>>> +    .inst_caps = &platform_inst_cap_sm8550,
>>>>>> +    .inst_fw_caps = inst_fw_cap_sm8550,
>>>>>> +    .inst_fw_caps_size = ARRAY_SIZE(inst_fw_cap_sm8550),
>>>>>> +    .tz_cp_config_data = &tz_cp_config_sm8550,
>>>>>> +    .core_arch = VIDEO_ARCH_LX,
>>>>>> +    .hw_response_timeout = HW_RESPONSE_TIMEOUT_VALUE,
>>>>>> +    .ubwc_config = &ubwc_config_sm8550,
>>>>>> +    .num_vpp_pipe = 4,
>>>>>> +    .max_session_count = 16,
>>>>>> +    .max_core_mbpf = ((8192 * 4352) / 256) * 2,
>>>>>> +    .input_config_params =
>>>>>> +        sm8550_vdec_input_config_params,
>>>>>> +    .input_config_params_size =
>>>>>> +        ARRAY_SIZE(sm8550_vdec_input_config_params),
>>>>>> +    .output_config_params =
>>>>>> +        sm8550_vdec_output_config_params,
>>>>>> +    .output_config_params_size =
>>>>>> +        ARRAY_SIZE(sm8550_vdec_output_config_params),
>>>>>> +    .dec_input_prop = sm8550_vdec_subscribe_input_properties,
>>>>>> +    .dec_input_prop_size = 
>>>>>> ARRAY_SIZE(sm8550_vdec_subscribe_input_properties),
>>>>>> +    .dec_output_prop = sm8550_vdec_subscribe_output_properties,
>>>>>> +    .dec_output_prop_size =
>>>>>> ARRAY_SIZE(sm8550_vdec_subscribe_output_properties),
>>>>>> +
>>>>>> +    .dec_ip_int_buf_tbl = sm8550_dec_ip_int_buf_tbl,
>>>>>> +    .dec_ip_int_buf_tbl_size = 
>>>>>> ARRAY_SIZE(sm8550_dec_ip_int_buf_tbl),
>>>>>> +    .dec_op_int_buf_tbl = sm8550_dec_op_int_buf_tbl,
>>>>>> +    .dec_op_int_buf_tbl_size = 
>>>>>> ARRAY_SIZE(sm8550_dec_op_int_buf_tbl),
>>>>>> +};
>>>>> While i was extending the data for QCS8300 (one another iris-v3 
>>>>> variant), i
>>>>> realize that this file iris_platform_sm8550.c is getting dumped 
>>>>> with all SOC
>>>>> platform data. It would be a good idea at this point to split it 
>>>>> into something
>>>>> like this
>>>>> 1. Introduce SOC specific c file and move the respective SOC 
>>>>> platform data to
>>>>> it, for ex, in this case sm8650_data
>>>>> 2. Move the common structs from iris_platform_sm8550.c to
>>>>> iris_platform_common.h. This way more SOCs getting added in future, 
>>>>> can include
>>>>> the common header to reuse them, otherwise it would end up using 
>>>>> 8550.c for all
>>>>> future SOC.
>>>>>
>>>>> Share your comments if you have any better approach to manage/re- 
>>>>> use these
>>>>> platform data considering more SOCs getting added.
>>>>
>>>> Right, yes the architecture is fine, but I don't feel 
>>>> iris_platform_common is
>>>> the right
>>>> place, perhaps we could introduce a platform_catalog.c where we 
>>>> could place all
>>>> the common
>>>> platform data and reuse them from the platform_<soc>.c files ?
>>> Common structs would certainly need to be part of a header which can be
>>> included. Where do you plan to keep common struct to be used across 
>>> SOC specific
>>> file in your approach ?
>>>>
>>>> I can design prototype on top of this patchset as an RFC.
>>> I was thinking that the changes are not that big, and can be done in 
>>> existing
>>> series though.
>>>
>> For now, I think you can introduce a platform_sm8650.c as part of this
>> series and use the common structure from platform_sm8550.c.
>> Shouldn't be a big change.
> 
> I tried but I don't how to solve this, you need a build-time ARRAY_SIZE for
> the arrays, so they need to be in the same .c file to allow a build-time
> evaluation of ARRAY_SIZE. If he common tables are moved into a header
> they will be duplicated into both platform_sm8650 & platform_sm8550 objects
> which is not what we want.
> 
> The solution would be to write all the platform tables & iris_platform_data
> into headers, with common headers, then include those into a 
> platform_catalog.c
> like is done for the drm/msm/dpu1 catalog.
> 
> Neil
> 
>>
>> Later you can post a separate patch series to add platform_catalog.c and
>> have common struct placed there which can be used across different SOC
>> platform files.
>>
>> or other way is,
>> Post a patch series to introduce platform_catalog.c with common struct 
>> and
>> then rebase your 8650 series on top of it.
>>
>> Thanks,
>> Dikshita
>>> Thanks,
>>> Vikash
>>>>
>>>> Neil
>>>>
>>>>>
>>>>> Regards,
>>>>> Vikash
>>>>>
>>>>>> diff --git a/drivers/media/platform/qcom/iris/iris_probe.c
>>>>>> b/drivers/media/platform/qcom/iris/iris_probe.c
>>>>>> index
>>>>>> 4f8bce6e2002bffee4c93dcaaf6e52bf4e40992e..7cd8650fbe9c09598670530103e3d5edf32953e7 100644
>>>>>> --- a/drivers/media/platform/qcom/iris/iris_probe.c
>>>>>> +++ b/drivers/media/platform/qcom/iris/iris_probe.c
>>>>>> @@ -345,6 +345,10 @@ static const struct of_device_id 
>>>>>> iris_dt_match[] = {
>>>>>>                .data = &sm8250_data,
>>>>>>            },
>>>>>>    #endif
>>>>>> +    {
>>>>>> +        .compatible = "qcom,sm8650-iris",
>>>>>> +        .data = &sm8650_data,
>>>>>> +    },
>>>>>>        { },
>>>>>>    };
>>>>>>    MODULE_DEVICE_TABLE(of, iris_dt_match);
>>>>>>
>>>>
> 

Eh as I read the suggestion about putting the structs - instantiating 
the structs in the header, I wondered if that would link but anyway.

One way to solve this without going the dpu1 route right now is to 
rename the platform files

iris_platform_sm8250.c -> iris_platform_common_hfi_gen1.c
iris_platform_sm8550.c -> iris_platform_common_hfi_gen2.c

The differentiation around HFI into "generations" instead of 
incrementing the already existing HFIXXX version is unfortunate.

At least this way we have a repository of common HFI code located in 
each file, in expectation of HFI GEN3 whenever.

---
bod

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ