linux-kernel - Re: [PATCH mm] maccess: fix strncpy_from_user

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: <CAEf4BzbVPQ=BjWztmEwBPRKHUwNfKBkS3kce-Rzka6zvbQeVpg@mail.gmail.com>
Date: Thu, 17 Apr 2025 13:44:48 -0700
From: Andrii Nakryiko <andrii.nakryiko@...il.com>
To: Mykyta Yatsenko <mykyta.yatsenko5@...il.com>, rostedt@...dmis.org
Cc: akpm@...ux-foundation.org, linux-mm@...ck.org, mhiramat@...nel.org, 
	andrii@...nel.org, kernel-team@...a.com, linux-kernel@...r.kernel.org, 
	Mykyta Yatsenko <yatsenko@...a.com>
Subject: Re: [PATCH mm] maccess: fix strncpy_from_user_nofault empty string handling

On Thu, Apr 17, 2025 at 8:28 AM Mykyta Yatsenko
<mykyta.yatsenko5@...il.com> wrote:
>
> From: Mykyta Yatsenko <yatsenko@...a.com>
>
> strncpy_from_user_nofault should return the length of the copied string
> including the trailing NUL, but if the argument unsafe_addr points to
> an empty string ({'\0'}), the return value is 0.
>
> This happens as strncpy_from_user copies terminal symbol into dst
> and returns 0 (as expected), but strncpy_from_user_nofault does not
> modify ret as it is not equal to count and not greater than 0, so 0 is
> returned, which contradicts the contract.
>
> Signed-off-by: Mykyta Yatsenko <yatsenko@...a.com>
> ---
>  kernel/trace/trace_events_filter.c | 10 ++++++++--
>  mm/maccess.c                       |  2 +-
>  2 files changed, 9 insertions(+), 3 deletions(-)
>
> diff --git a/kernel/trace/trace_events_filter.c b/kernel/trace/trace_events_filter.c
> index 0993dfc1c5c1..86b7e5a4e235 100644
> --- a/kernel/trace/trace_events_filter.c
> +++ b/kernel/trace/trace_events_filter.c
> @@ -800,6 +800,7 @@ static __always_inline char *test_string(char *str)
>  {
>         struct ustring_buffer *ubuf;
>         char *kstr;
> +       int cnt;
>
>         if (!ustring_per_cpu)
>                 return NULL;
> @@ -808,7 +809,9 @@ static __always_inline char *test_string(char *str)
>         kstr = ubuf->buffer;
>
>         /* For safety, do not trust the string pointer */
> -       if (!strncpy_from_kernel_nofault(kstr, str, USTRING_BUF_SIZE))
> +       cnt = strncpy_from_kernel_nofault(kstr, str, USTRING_BUF_SIZE);
> +       /* Return null if empty string or error */
> +       if (cnt <= 1)
>                 return NULL;

I wouldn't touch this part and leave it up to Steven to fix (if he
agrees it needs fixing). Current logic seems wrong already, as it
won't correctly handle -EFAULT. And, on the other hand, there is
nothing wrong or special about empty string, so I don't think it needs
special handling. Let's drop these changes in trace_events_filter.c?

>         return kstr;
>  }
> @@ -818,6 +821,7 @@ static __always_inline char *test_ustring(char *str)
>         struct ustring_buffer *ubuf;
>         char __user *ustr;
>         char *kstr;
> +       int cnt;
>
>         if (!ustring_per_cpu)
>                 return NULL;
> @@ -827,7 +831,9 @@ static __always_inline char *test_ustring(char *str)
>
>         /* user space address? */
>         ustr = (char __user *)str;
> -       if (!strncpy_from_user_nofault(kstr, ustr, USTRING_BUF_SIZE))
> +       cnt = strncpy_from_user_nofault(kstr, ustr, USTRING_BUF_SIZE);
> +       /* Return null if empty string or error */
> +       if (cnt <= 1)
>                 return NULL;

ditto

>
>         return kstr;
> diff --git a/mm/maccess.c b/mm/maccess.c
> index 8f0906180a94..831b4dd7296c 100644
> --- a/mm/maccess.c
> +++ b/mm/maccess.c
> @@ -196,7 +196,7 @@ long strncpy_from_user_nofault(char *dst, const void __user *unsafe_addr,
>         if (ret >= count) {
>                 ret = count;
>                 dst[ret - 1] = '\0';
> -       } else if (ret > 0) {
> +       } else if (ret >= 0) {
>                 ret++;
>         }
>

This part looks good and does indeed fix the issue. Good catch!

Reviewed-by: Andrii Nakryiko <andrii@...nel.org>

> --
> 2.49.0
>