Re: [PATCH 2/3] perf record: Avoid overrunning the zstd output buffer

[Dmitry Ilvokhin <d@ilvokhin.com>]

On Thu, Jun 18, 2026 at 04:20:59PM +0000, sashiko-bot@kernel.org wrote:
> [Severity: High]
> Does this check introduce an overly strict requirement that causes compression 
> to fail prematurely?
> 
> By requiring dst_size < hdr_size + max_record_size + sizeof(u64), this 
> condition unconditionally fails compression if the remaining buffer is less 
> than the maximum record size (typically ~64KB).

Fair point.

After thinking about it a bit more, there is also another problem: the
check is bypassed for the first iteration of the loop. I think it is
correct, because of how the output buffer is sized, but it is not
immediately obvious from the code and needs to be traced through few
call levels.

Probably a better solution would be to have 'header_size' argument in
zstd_compress_stream_to_records(): then we would not need to rely on
process_header()'s return value, the check could run on every iteration,
and it would only need room for a header rather than a full
'max_record_size'.