Re: [PATCH v4 24/26] media: iris: Allocate and queue internal buffers for encoder video device
From: Dikshita Agarwal
Date: Fri Aug 29 2025 - 14:07:18 EST
On 8/29/2025 2:35 PM, Bryan O'Donoghue wrote:
> On 25/08/2025 08:00, Dikshita Agarwal wrote:
>> +static inline
>> +u32 size_enc_single_pipe(u32 rc_type, u32 bitbin_size, u32 num_vpp_pipes,
>> + u32 frame_width, u32 frame_height, u32 lcu_size)
>> +{
>> + u32 size_aligned_height = ALIGN((frame_height), lcu_size);
>> + u32 size_aligned_width = ALIGN((frame_width), lcu_size);
>> + u32 size_single_pipe_eval = 0, sao_bin_buffer_size = 0;
>> + u32 padded_bin_sz;
>> +
>> + if ((size_aligned_width * size_aligned_height) > (3840 * 2160))
>> + size_single_pipe_eval = (bitbin_size / num_vpp_pipes);
>> + else if (num_vpp_pipes > 2)
>> + size_single_pipe_eval = bitbin_size / 2;
>> + else
>> + size_single_pipe_eval = bitbin_size;
>> +
>> + sao_bin_buffer_size = (64 * ((((frame_width) + 32) * ((frame_height)
>> + 32)) >> 10)) + 384;
>> + padded_bin_sz = ALIGN(size_single_pipe_eval, 256);
>> + size_single_pipe_eval = sao_bin_buffer_size + padded_bin_sz;
>> +
>> + return ALIGN(size_single_pipe_eval, 256);
>> +}
>
> Applying your full series, this function is unused and causes a warning in CI.
>
> Assuming tests pass for me, I'll be dropping this function in the PR and
> you can resubmit it if/when you find a use for it.
you're right, this is unused. Strangely I didn't see any such warnings
during my runs, but yes, this is good to be dropped.
Thanks,
Dikshita
>
> ---
> bod