In function "libav_internal.h:pl_map_avframe_vulkan()", the "repr.bits.bit_shift" member of the output pl_frame does not seem to be set correctly.

[EliteWeapon]

For the 10-bit pixel format, in "pl_map_avframe_vulkan()" function, the "repr.bits.bit_shift" member of the output pl_frame isn't be set, whereas it is correctly set in the pl_map_avframe_drm() function.

`
static bool pl_map_avframe_vulkan(pl_gpu gpu, struct pl_frame *out,
const AVFrame *frame)
{
...

out->acquire = pl_acquire_avframe;
out->release = pl_release_avframe;
pl_fix_hwframe_sample_depth(out);
return true;

}
`

`
static bool pl_map_avframe_drm(pl_gpu gpu, struct pl_frame *out,
const AVFrame *frame)
{
...

pl_fix_hwframe_sample_depth(out);

switch (hwfc->sw_format) {
case AV_PIX_FMT_P010: out->repr.bits.bit_shift = 6; break;
default: break;
}

return true;

}

`

[EliteWeapon]

When I used ffplay to play a 10-bit video on an NVIDIA RTX5080 GPU, the displayed image is abnormal.
The specific ffplay command used is as follows:
"ffplay -hwaccel cuda -i INPUT_FILE"