linux-kernel - Re: [PATCH] ubifs: prevent corrupted data write via error_remove

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <f800d3d3-4d57-f28c-bece-d95af0d22159@huawei.com>
Date: Wed, 11 Feb 2026 16:46:07 +0800
From: Zhihao Cheng <chengzhihao1@...wei.com>
To: <fan.yu9@....com.cn>, <richard@....at>
CC: <cui.jinpeng2@....com.cn>, <zhang.yue5@....com.cn>, <xu.xin16@....com.cn>,
	<yang.tao172@....com.cn>, <yang.yang29@....com.cn>,
	<hu.shengming@....com.cn>, <linux-mtd@...ts.infradead.org>,
	<linux-kernel@...r.kernel.org>
Subject: Re: [PATCH] ubifs: prevent corrupted data write via
 error_remove_folio

在 2026/2/11 15:40, fan.yu9@....com.cn 写道:
> From: Jinpeng Cui <cui.jinpeng2@....com.cn>
> 
> When a memory corruption event triggers memory_failure, ext4 releases references to the affected dirty pages,
> ensuring that no corrupted data is written to disk. However, ubifs currently lacks the error_remove_folio capability
> and therefore cannot release references to faulty private memory. As a result, corrupted data may eventually be
> written to flash storage.
> 
> Add generic_error_remove_folio support to ubifs, preventing corrupted memory data from being persisted to flash.
> 
> Signed-off-by: Jinpeng Cui <cui.jinpeng2@....com.cn>
> Signed-off-by: Fan Yu <fan.yu9@....com.cn>
> Cc: Shengming Hu <hu.shengming@....com.cn>
> ---
>   fs/ubifs/file.c | 1 +
>   1 file changed, 1 insertion(+)
> 

The 'error_remove_folio' was introduced in commit 
257187362123f15d9d1("HWPOISON: Define a new error_remove_page address 
space op for async truncation") to distinguish the error handling 
behavior beteween metadata page and data page. Files' datapage will be 
truncated if memory error happens during page copy(eg. migration, 
mmap_cow). Then several filesystems(with '.migrate_page' implemtented) 
supportted it. So, I'm fine with this patch.

Reviewed-by: Zhihao Cheng <chengzhihao1@...wei.com>
> diff --git a/fs/ubifs/file.c b/fs/ubifs/file.c
> index 3dc3ca1..34445b7 100644
> --- a/fs/ubifs/file.c
> +++ b/fs/ubifs/file.c
> @@ -1631,6 +1631,7 @@ static int ubifs_symlink_getattr(struct mnt_idmap *idmap,
>      .dirty_folio    = ubifs_dirty_folio,
>      .migrate_folio  = filemap_migrate_folio,
>      .release_folio  = ubifs_release_folio,
> +   .error_remove_folio = generic_error_remove_folio,
>   };
> 
>   const struct inode_operations ubifs_file_inode_operations = {
> --
> 1.8.3.1
> .
>