2021-05-11 06:20:52

by Zou Wei

Subject: [PATCH -next] crypto: hisilicon -: switch to memdup_user_nul()

Use memdup_user_nul() helper instead of open-coding to
simplify the code.

Reported-by: Hulk Robot <[email protected]>
Signed-off-by: Zou Wei <[email protected]>
---
drivers/crypto/hisilicon/qm.c | 11 ++---------
1 file changed, 2 insertions(+), 9 deletions(-)

diff --git a/drivers/crypto/hisilicon/qm.c b/drivers/crypto/hisilicon/qm.c
index ce439a0..83a5d30 100644
--- a/drivers/crypto/hisilicon/qm.c
+++ b/drivers/crypto/hisilicon/qm.c
@@ -1570,17 +1570,10 @@ static ssize_t qm_cmd_write(struct file *filp, const char __user *buffer,
if (count > QM_DBG_WRITE_LEN)
return -ENOSPC;

- cmd_buf = kzalloc(count + 1, GFP_KERNEL);
- if (!cmd_buf)
+ cmd_buf = memdup_user_nul(buffer, count);
+ if (IS_ERR(cmd_buf))
return -ENOMEM;

- if (copy_from_user(cmd_buf, buffer, count)) {
- kfree(cmd_buf);
- return -EFAULT;
- }
-
- cmd_buf[count] = '\0';
-
cmd_buf_tmp = strchr(cmd_buf, '\n');
if (cmd_buf_tmp) {
*cmd_buf_tmp = '\0';
--
2.6.2

2021-05-11 07:07:58

by Zhou Wang

Subject: Re: [PATCH -next] crypto: hisilicon -: switch to memdup_user_nul()

On 2021/5/11 14:37, Zou Wei wrote:
> Use memdup_user_nul() helper instead of open-coding to
> simplify the code.
>
> Reported-by: Hulk Robot <[email protected]>
> Signed-off-by: Zou Wei <[email protected]>
> ---
> drivers/crypto/hisilicon/qm.c | 11 ++---------
> 1 file changed, 2 insertions(+), 9 deletions(-)
>
> diff --git a/drivers/crypto/hisilicon/qm.c b/drivers/crypto/hisilicon/qm.c
> index ce439a0..83a5d30 100644
> --- a/drivers/crypto/hisilicon/qm.c
> +++ b/drivers/crypto/hisilicon/qm.c
> @@ -1570,17 +1570,10 @@ static ssize_t qm_cmd_write(struct file *filp, const char __user *buffer,
> if (count > QM_DBG_WRITE_LEN)
> return -ENOSPC;
>
> - cmd_buf = kzalloc(count + 1, GFP_KERNEL);
> - if (!cmd_buf)
> + cmd_buf = memdup_user_nul(buffer, count);
> + if (IS_ERR(cmd_buf))
> return -ENOMEM;
>
> - if (copy_from_user(cmd_buf, buffer, count)) {
> - kfree(cmd_buf);
> - return -EFAULT;
> - }
> -
> - cmd_buf[count] = '\0';
> -
> cmd_buf_tmp = strchr(cmd_buf, '\n');
> if (cmd_buf_tmp) {
> *cmd_buf_tmp = '\0';
>

It is fine to me, thanks!

BTW, normally we use crypto: hisilicon - switch to memdup_user_nul() as the subject line,
just like other patches :)