[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <ddc0f1ab-03bb-2e66-280d-0dcea9042b5c@gmail.com>
Date: Sun, 6 Dec 2020 19:55:34 +0100
From: Bodo Stroesser <bostroesser@...il.com>
To: Xiaohui Zhang <ruc_zhangxiaohui@....com>,
martin.petersen@...cle.com, linux-scsi@...r.kernel.org,
target-devel@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH 1/1] target: Make sure no zero value in the buffer
On 03.12.20 12:40, Xiaohui Zhang wrote:
> From: Zhang Xiaohui <ruc_zhangxiaohui@....com>
>
> The fix makes sure no zero value in the buffer, by comparing the
> strlen() of the original buffer with the size variable.
What problem do you want to fix here?
I think, presently iblock_set_configfs_dev_params() just ignores input after a possible zero byte, which AFAICS is the usual way to handle such 'abnormal' input.
Of course, strictly taken it is a bug to return 'count' without having processed count bytes. OTOH, iblock_set_configfs_dev_params also silently ignores unknown parameter strings.
If you want to have strict error handling in such sys- or config-FS callbacks, you have to do a lot of work ...
>
> Signed-off-by: Zhang Xiaohui <ruc_zhangxiaohui@....com>
> ---
> drivers/target/target_core_iblock.c | 2 ++
> 1 file changed, 2 insertions(+)
>
> diff --git a/drivers/target/target_core_iblock.c b/drivers/target/target_core_iblock.c
> index f2bd2e207..b23e92449 100644
> --- a/drivers/target/target_core_iblock.c
> +++ b/drivers/target/target_core_iblock.c
> @@ -537,6 +537,8 @@ static ssize_t iblock_set_configfs_dev_params(struct se_device *dev,
> int ret = 0, token;
> unsigned long tmp_readonly;
>
> + if (strlen(page) < count)
> + return -EOVERFLOW;
> opts = kstrdup(page, GFP_KERNEL);
> if (!opts)
> return -ENOMEM;
>
Powered by blists - more mailing lists