Re: [v4 PATCH 10/13] ubifs: Use crypto_acomp interface

[Zhihao Cheng <chengzhihao1@huawei.com>]

在 2025/3/14 20:22, Herbert Xu 写道:
> Replace the legacy crypto compression interface with the new acomp
> interface.
> 
> Remove the compression mutexes and the overallocation for memory
> (the offender LZO has been fixed).

Hi, Herbert. Can you show me which patch fixed the problem in LZO?
> 
> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
> ---
>   fs/ubifs/compress.c | 116 ++++++++++++++++++++++++++++----------------
>   fs/ubifs/journal.c  |   2 +-
>   fs/ubifs/ubifs.h    |  15 +-----
>   3 files changed, 77 insertions(+), 56 deletions(-)
> 

[...]
> diff --git a/fs/ubifs/ubifs.h b/fs/ubifs/ubifs.h
> index 3375bbe0508c..7d0aaf5d2e23 100644
> --- a/fs/ubifs/ubifs.h
> +++ b/fs/ubifs/ubifs.h
> @@ -124,13 +124,6 @@
>   #define OLD_ZNODE_AGE 20
>   #define YOUNG_ZNODE_AGE 5
>   
> -/*
> - * Some compressors, like LZO, may end up with more data then the input buffer.
> - * So UBIFS always allocates larger output buffer, to be sure the compressor
> - * will not corrupt memory in case of worst case compression.
> - */
> -#define WORST_COMPR_FACTOR 2

Does LZO guarantee the output data length smaller than input buffer 
length? Which commit fixed the issue?
> -
>   #ifdef CONFIG_FS_ENCRYPTION
>   #define UBIFS_CIPHER_BLOCK_SIZE FSCRYPT_CONTENTS_ALIGNMENT
>   #else