[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Fri, 5 May 2023 21:02:10 +0200
From: Konrad Dybcio <konrad.dybcio@...aro.org>
To: Vikash Garodia <quic_vgarodia@...cinc.com>,
Stanimir Varbanov <stanimir.k.varbanov@...il.com>,
Andy Gross <agross@...nel.org>,
Bjorn Andersson <andersson@...nel.org>,
Mauro Carvalho Chehab <mchehab@...nel.org>,
Dikshita Agarwal <dikshita@....qualcomm.com>,
Bryan O'Donoghue <bryan.odonoghue@...aro.org>,
Mansur Alisha Shaik <mansur@...eaurora.org>,
Jonathan Marek <jonathan@...ek.ca>,
Hans Verkuil <hans.verkuil@...co.com>,
Dikshita Agarwal <quic_dikshita@...cinc.com>
Cc: Mauro Carvalho Chehab <mchehab+huawei@...nel.org>,
Stanimir Varbanov <stanimir.varbanov@...aro.org>,
linux-media@...r.kernel.org, linux-arm-msm@...r.kernel.org,
linux-kernel@...r.kernel.org,
Marijn Suijten <marijn.suijten@...ainline.org>,
stable@...r.kernel.org
Subject: Re: [PATCH v2 03/18] media: venus: Remap bufreq fields on HFI6XX
On 5.05.2023 14:38, Vikash Garodia wrote:
> On 5/4/2023 1:30 PM, Konrad Dybcio wrote:
>> Similarly to HFI4XX, the fields are remapped on 6XX as well. Fix it.
>>
>> Cc: stable@...r.kernel.org # v5.12+
>> Fixes: 7ed9e0b3393c ("media: venus: hfi, vdec: v6 Add IS_V6() to existing IS_V4() if locations")
>> Signed-off-by: Konrad Dybcio <konrad.dybcio@...aro.org>
>> ---
>> drivers/media/platform/qcom/venus/hfi_helper.h | 9 ++++++---
>> 1 file changed, 6 insertions(+), 3 deletions(-)
>>
>> diff --git a/drivers/media/platform/qcom/venus/hfi_helper.h b/drivers/media/platform/qcom/venus/hfi_helper.h
>> index 105792a68060..e0c8f15644df 100644
>> --- a/drivers/media/platform/qcom/venus/hfi_helper.h
>> +++ b/drivers/media/platform/qcom/venus/hfi_helper.h
>> @@ -1170,11 +1170,14 @@ struct hfi_buffer_display_hold_count_actual {
>> /* HFI 4XX reorder the fields, use these macros */
>> #define HFI_BUFREQ_HOLD_COUNT(bufreq, ver) \
>> - ((ver) == HFI_VERSION_4XX ? 0 : (bufreq)->hold_count)
>> + ((ver) == HFI_VERSION_4XX || (ver) == HFI_VERSION_6XX \
>> + ? 0 : (bufreq)->hold_count)
>> #define HFI_BUFREQ_COUNT_MIN(bufreq, ver) \
>> - ((ver) == HFI_VERSION_4XX ? (bufreq)->hold_count : (bufreq)->count_min)
>> + ((ver) == HFI_VERSION_4XX || (ver) == HFI_VERSION_6XX \
>> + ? (bufreq)->hold_count : (bufreq)->count_min)
>> #define HFI_BUFREQ_COUNT_MIN_HOST(bufreq, ver) \
>> - ((ver) == HFI_VERSION_4XX ? (bufreq)->count_min : 0)
>> + ((ver) == HFI_VERSION_4XX || (ver) == HFI_VERSION_6XX \
>> + ? (bufreq)->count_min : 0)
>
> This patch is not correct. The existing code handles the disparity of buffer requirement payload received from firmware.
>
> Its applicable only for V4.
>
> For V6, driver does not rely on firmware to get the buffer requirement. Refer the buffer platform code for more details.
OK right I can see, downstream has a condition for IRIS2/IRIS2_1,
thanks for pointing this out!
Konrad
>
> -Vikash
>
>> struct hfi_buffer_requirements {
>> u32 type;
>>
Powered by blists - more mailing lists