Re: [Qemu-devel] [PATCH] hw/9pfs/virtio-9p-local.c: use snprintf() instead of sprintf()

[Markus Armbruster]

Chen Gang <address@hidden> writes:

> On 02/03/2014 06:39 PM, Chen Gang wrote:
>> On 02/03/2014 06:34 PM, Daniel P. Berrange wrote:
>>> On Mon, Feb 03, 2014 at 06:00:42PM +0800, Chen Gang wrote:
>>>> We can not assume "'path' + 'ctx->fs_root'" must be less than MAX_PATH,
>>>> so need use snprintf() instead of sprintf().
>>>>
>>>> And also recommend to use ARRAY_SIZE instead of hard code macro for an
>>>> array size in snprintf().
>>>
>>> In the event that there is overflow this will cause the data to be
>>> truncated, potentially causing QEMU to access the wrong file on the
>>> host. Both snprintf and sprintf are really bad because of their
>>> use of fixed buffers. Better to change it to g_strdup_printf which
>>> dynamically allocates buffers.
>>>
>
> After check the details, I guess we can not change to g_strdup_printf or
> others (e.g. v9fs_string_*).
>
> v9fs need use "mkdir, remove ..." which have MAX_PATH limitation. So if
> the combined path is longer than MAX_PATH, before it passes to "mkdir,
> remove ...", it has to be truncated just like what rpath() has done.

What good could truncating possibly do?

If the pathname is too long for mkdir(), truncating won't make it work,
it will make it do the wrong thing.

Second guessing when a pathname is too long for a system call is not a
good idea.  If it's too long, the system call will tell you.  As Dan
noted, PATH_MAX is *not* a hard limit.

    {PATH_MAX}
        Maximum number of bytes the implementation will store as a
        pathname in a user-supplied buffer of unspecified size,
        including the terminating null character. Minimum number the
        implementation will accept as the maximum number of bytes in a
        pathname.

[...]