Stanislav Yakovlev <stas.yakovlev@xxxxxxxxx> writes:
> On Tue, 2 Mar 2021 at 15:16, Dan Carpenter <dan.carpenter@xxxxxxxxxx> wrote:
>>
>> The "ext->key_len" is a u16 that comes from the user. If it's over
>> SCM_KEY_LEN (32) that could lead to memory corruption.
>>
>> Fixes: e0d369d1d969 ("[PATCH] ieee82011: Added WE-18 support to
>> default wireless extension handler")
>> Signed-off-by: Dan Carpenter <dan.carpenter@xxxxxxxxxx>
>> ---
>> drivers/net/wireless/intel/ipw2x00/libipw_wx.c | 6 ++++--
>> 1 file changed, 4 insertions(+), 2 deletions(-)
[...]
>> --- a/drivers/net/wireless/intel/ipw2x00/libipw_wx.c
>> +++ b/drivers/net/wireless/intel/ipw2x00/libipw_wx.c
>> @@ -633,8 +633,10 @@ int libipw_wx_set_encodeext(struct libipw_device *ieee,
>> }
>>
>> if (ext->alg != IW_ENCODE_ALG_NONE) {
>> - memcpy(sec.keys[idx], ext->key, ext->key_len);
>> - sec.key_sizes[idx] = ext->key_len;
>> + int len = min_t(int, ext->key_len, SCM_KEY_LEN);
>> +
>> + memcpy(sec.keys[idx], ext->key, len);
>> + sec.key_sizes[idx] = len;
>> sec.flags |= (1 << idx);
>> if (ext->alg == IW_ENCODE_ALG_WEP) {
>> sec.encode_alg[idx] = SEC_ALG_WEP;
In another thread Linus gave a good tip about clamp_val(), I think it
should be used it here as well to make the check safer and more
readable. And also elsewhere in wireless code with similar limits.
--
https://patchwork.kernel.org/project/linux-wireless/list/
https://wireless.wiki.kernel.org/en/developers/documentation/submittingpatches
